Anda di halaman 1dari 587

Design Hydrology

and Sedimentology
for Small Catchments

C . T. H a a n
Biosystems and Agricultural Engineering Department
Oklahoma State University
Stillwater, Oklahoma

B. J. Barfield
Biosystems and Agricultural Engineering Department
Oklahoma State University
Stillwater, Oklahoma

J. C . H a y e s
Agricultural and Biological Engineering Department
Clemson University
Clemson, South Carolina

A c a d e m i c P r e s s
An Imprint of Elsevier
Amsterdam Boston Heidelberg London New York Oxford
Paris San Diego San Francisco Singapore Sydney Tokyo
This book is printed on acid-free paper. ©

Copyright © 1994,1981 by ACADEMIC PRESS, INC.


All Rights Reserved.
No part of this publication may be reproduced or transmitted in any form or by any
means, electronic or mechanical, including photocopy, recording, or any information
storage and retrieval system, without permission in writing from the publisher.

Permissions may be sought directly from Elsevier's Science and Technology Rights Department in
Oxford, UK. Phone: (44) 1865 843830, Fax: (44) 1865 853333, e-mail: permissions@elsevier.co.uk.
You may also complete your request on-line via the Elsevier homepage: http://www.elsevier.com by
selecting "Customer Support" and then "Obtaining Permissions".

A c a d e m i c Press, Inc.
An Imprint of Elsevier
525 Β Street, Suite 1900, San Diego, California 92101-4495

United Kingdom Edition published by


Academic Press Limited
24-28 Oval Road, London NW1 7DX

Library of Congress Cataloging-in-Publication Data

Haan, C. T. (Charles Thomas), Date


D e s i g n hydrology and s e d i m e n t o l o g y for small catchments / C. T.
Haan, B. J. Barfield, J. C. Hayes,
p. cm.
Includes bibliographical references and index.
ISBN-13: 978-0-12-312340-4 ISBN-10: 0-12-312340-2
1. Watershed management. 2. Sediment control. 3. Runoff.
4. Hydrology. I. Barfield, Billy J. II. H a y e s , J. C. III. Title
TC409.H3 1993
627--dc20
93-11165
CIP
ISBN-13: 978-0-12-312340-4
ISBN-10: 0-12-312340-2

PRINTED IN THE UNITED STATES OF AMERICA

0 9 10 11 12 13 13 12 11 10
Jan for 25 years of love and support;
Patti, Chris, and Pam for helping me grow;
Mom and Dad for getting me started.
C T H

Annette for being there for me for 25-plus years,


Michelle and Will for teaching me about myself
Merrell and Ola for believing in me when it mattered.
BJB

Mary Anne for sharing my life and for showing love and faith,
Jay and Matthew for always making life exciting,
Mama and Daddy for providing strength and comfort.
JCH

To our Heavenly Father for His continual guidance.


Preface

In recent years a number of excellent books on hydrol­ In practice, c o m p u t e r software would indeed be used
ogy and related topics have been published. O n e might to carry out many of the required computations. As
wonder about the need for yet a n o t h e r book on the microcomputers become more powerful, computer
topic. In the authors' experiences working with practic­ codes are being continually improved. User-friendly
ing engineers and hydrologists, we have found a defi­ interfaces for c o m p u t e r programs are making it possi­
nite need for a treatment of the design aspects of ble to use many hydrologic programs with little knowl­
hydrology and sedimentology, especially for small edge of the hydrology being simulated. This book pro­
catchments. Most practicing engineers and hydrologists vides the background required to understand most of
work on relatively small watersheds, designing storm the techniques used in current hydrology software and
water control facilities, drainage facilities, erosion and represents an excellent companion to program user
sediment control practices, detention ponds, small manuals, which often contain almost no explanation of
channels and storm drains, and the like. This book the hydrologic techniques being employed.
attempts to provide a single source of design proce­ T h e book contains many solved example problems as
dures for most aspects of runoff and sediment control well as n u m e r o u s problems for solution at the end of
in small catchments. Sections 208 and 319 of " T h e each chapter. T h e s e problems will assist in developing
Clean W a t e r A c t " and their emphasis on storm water a fuller understanding of design procedures. For use in
control in urban and rural areas have m a d e the appli­ classroom or continuing education settings, the prob­
cation of this technology imperative. lems can be easily a d o p t e d to local conditions by using
T h e approach used in the book is to present* state-of- local rainfall, soil, and o t h e r types of information.
the-art design methodologies with enough explanation This book has evolved over the past 15 years from a
of basic principles to ensure understanding of the set of mimeographed notes used in continuing educa­
rationale behind the methodology. T h e mathematical tion courses for engineers interested in learning how to
and theoretical aspects are fully developed only w h e n meet design requirements for permitting of areas to be
required for an understanding of the methodology. surface mined. During its evolution it has appeared as
A d e q u a t e data are presented in tables and charts for two privately published books entitled "Hydrology and
many designs; however, the book does not attempt to Sedimentology of Surface Mined L a n d s " by H a a n and
replace design manuals currently being used by many Barfield and " A p p l i e d Hydrology and Sedimentology
local, state, and federal agencies. References to more for Disturbed A r e a s " by Barfield, W a r n e r , and Haan.
extensive data tabulations are given where required. In the latter form it was widely used in the surface
T h e authors have taught basic hydrology and sedi­ mining industry to design water and sediment control
mentology courses to thousands of practicing engi­ facilities. T h e current version is a complete rewrite of
neers, as well as courses on the application of com­ the previous texts in nearly all aspects. T h e material on
puter models to the analysis and design of hydrologic erosion and sediment control presents extensive new
systems. It has been our experience that those who use technology that has evolved since the previous publica­
computer models without an understanding of the the­ tion.
ory and principles behind them do a poor job of C h a p t e r 1 provides an overview to the volume.
applying them. Inappropriate designs are the frequent C h a p t e r 2 deals with hydrologic frequency analysis.
result. This book was written to provide a knowledge C h a p t e r 3 covers the estimation of runoff rates, vol­
base for practitioners. umes, and hydrographs. C h a p t e r s 4 and 5 deal with the

xiii
xiv Preface

hydraulics of open channel flow and hydraulic control w h o have reviewed or discussed many aspects of t h e
structures. T h e design of channels in stable and erodi- book. A m o n g t h e m are Bruce Wilson, Alex Fogle, R o n
ble materials as well as the design of small hydraulic Elliott, D a n Storm, and Flint Holbrook. G r a d u a t e
structures are covered here. C h a p t e r 6 deals with flow students in t h e Agricultural Engineering programs
routing in channels, ponds, and reservoirs. C h a p t e r s 7, at O k l a h o m a State University, t h e University of
8, and 9 examine sediment properties and transport, Kentucky, and Clemson University have also con­
the principles of erosion and sediment yield, and the tributed directly t h r o u g h their reading of p a r t s of t h e
design of practices to reduce erosion and control sedi­ book and indirectly through t h e stimulation they p r o ­
ment. C h a p t e r 10 discusses channel morphology and vided to the a u t h o r s ' research. A n u m b e r of profes­
the natural equilibrium of erodible channels. G r o u n d sionals w h o have c o m m e n t e d on earlier versions of t h e
water is covered in C h a p t e r 11, monitoring hydrologic work have also unknowingly m a d e valuable contribu­
systems in Chapter 12, and hydrologic modeling in tions. Of course, we have benefitted from the intellec­
Chapter 13. T h e appendices for all of the chapters are tual a t m o s p h e r e and t h e working conditions provided
at the end of the book and contain design information by o u r t h r e e universities. In particular, we are grateful
too voluminous to include in the body of the text. to J o h n Walker, who, as C h a i r m a n of t h e Agricultural
Every effort has been m a d e to eliminate textual Engineering D e p a r t m e n t at t h e University of Ken­
errors. In an undertaking of this magnitude, however, tucky, encouraged and s u p p o r t e d T o m H a a n and Bill
errors inevitably c r e e p in. T h e authors would appreci­ Barfield through t h e publication of t h e first volume.
ate notification of any errors so that they may be Finally, we would b e remiss if we did not express
corrected in future printings. appreciation to Charles A r t h u r of Academic Press for
W e must acknowledge t h e patience and support of his patience with us as deadlines went u n m e t .
our wives and families through the long process of
bringing this book to fruition. Without this support, we C. T. H a a n
would have been forced to a b a n d o n the project long B. J . Barfield*
ago. W e are also grateful to a n u m b e r of colleagues J . C. Hayes

* Formerly of the Agricultural Engineering Deparment, University of


Kentucky, Lexington, Kentucky.
1
Introduction

THE PROBLEM disturbed condition, the sediment yield over a 2-year


period goes from essentially nothing to about 1300 tons
T h e h u m a n propensity to occupy areas subject to for this 20-acre catchment in Maryland having West
occasional flooding, to alter natural watercourses, to Maryland silt loam soil on a 10% slope. N o conserva­
alter land forms, and to engage in other activities that tion practices were used except for the continuously
impact natural hydrologic and sedimentologic pro­ disturbed condition, which, in some cases, used mulch
cesses creates a need to offset those impacts thought to and sediment detention ponds. T h e figure illustrates
be detrimental through designed flow and sediment the effectiveness of sediment control practices in that
control systems. Land clearing, agricultural activities, the sediment yield was reduced on the continuously
construction, mining, urban and industrial develop­ disturbed area from about 1300 tons with no control
ment, and similar activities can have a major impact on practices to about 350 tons with mulch and a sediment
the quantity and rate of water runoff and on the rates detention pond. Procedures presented in this book can
of erosion and sediment transport that take place. be used to make comparative assessments like this for
Environmental concerns often make it desirable, if not many different situations.
necessary, to provide means of controlling runoff and Environmental regulations present great challenges
erosion from altered land areas to the level that would to engineers. They must design water and sediment
be present if no alterations were m a d e or to other control facilities that will e n h a n c e functional and aes­
legally specified levels. thetic aspects of projects, will not h a m p e r construction
Such an approach assumes that the hydrologic and or operational activities, and yet will meet exacting
sedimentologic responses of an area to climatic events state and federal regulations. T h u s a detailed knowl­
can be quantified for both the unaltered and the al­ edge of the principles and practices of hydrology and
tered state and that techniques for limiting the differ­ sediment control is required.
ence in responses to these two states are available.
Further it assumes that quantitative m e t h o d s for evalu­
ating control techniques are available. SCOPE AND OBJECTIVES OF COVERAGE
Figure 1.1, adapted from McBurnie et al. (1990),
illustrates the impact of land use changes on peak This book has b e e n written to acquaint engineers
runoff rates and sediment yields. This figure shows that with hydrologic, hydraulic, and sedimentation princi­
as the land use changes from forest to a continuously ples that will be useful in designing water and sediment

1
2 1. Introduction

σ
condition:
1 forest
2 meadow
> 3 mixed land use
ο
4 d i s t u r b e d , m u l c h e d , s e d i m e n t pond
5 d i s t u r b e d , no m u l c h , s e d i m e n t pond
6 d i s t u r b e d , mulched
0 7 c o n v e n t i o n a l a g r i c u l t u r e , corn
0)
8 d i s t u r b e d , no mulch
>
ο
4 5 6
• sediment runoff

condition
Figure 1.1 Land use impacts on runoff and erosion.

control facilities. Some of the material will be a review with an understanding of:
to some readers; however, nearly all engineers will find
1. hydrologic principles and techniques sufficient to
material in this treatment that supplements their cur­
estimate runoff rates, volumes, and hydrographs from a
rent knowledge.
variety of scenarios;
T h e treatment presented here is not theoretical;
2. open channel hydraulics in d e p t h sufficient to
however, a d e q u a t e theory is presented to develop a
design stable channels in erodible and nonerodible
firm understanding of principles. T h e effort is directed
materials;
toward practical design methodologies. T h e procedures
3. hydraulics in d e p t h sufficient for the design of
are presented in sufficient detail so that the m e t h o d s
simple d r o p structures, pipe spillways, emergency spill­
can be applied directly to actual situations. N u m e r o u s
ways, and culverts;
realistic, solved problems using the m e t h o d s p r e s e n t e d
4. flow routing in detail sufficient for the design of
are included.
water retention and sediment detention basins;
T h e book is not intended to replace or compete with
5. the principles of sediment production and sedi­
federal and state publications regarding acceptable de­
m e n t control, including both structural and nonstruc­
sign procedures. It is not a design manual as such but
tural methods;
presents design techniques that will apply in many
6. principles for evaluating stable alluvial channels
situations. Often more than one solution technique
and for predicting the impact of changes in channel
may be possible and appropriate. T h e design engineer
properties d u e to anthropogenic and natural changes;
must select the best technique to use u n d e r a given set
7. basic definitions and principles of ground water
of circumstances.
hydrology;
T o prevent the book from becoming excessively long,
8. requirements and techniques for monitoring hy­
considerable material has been included by reference
drologic systems; and
only. For instance, state manuals, U.S. Soil Conserva­
9. the basis for and use of hydrologic models.
tion Service reports, U.S Corps of Engineers manuals,
and U.S. Environmental Protection Agency publica­ T h e selection of the notation to use in the book
tions contain much valuable information but are in p r e s e n t e d special problems since information from sev­
themselves voluminous documents and thus not repro­ eral disciplines including hydrology, hydraulics, sedi­
duced herein. m e n t transport, erosion, geomorphology, and statistics
In view of the uncertain and dynamic n a t u r e of is included. T h e decision was m a d e to h o n o r tradi­
federal and state environmental regulations regarding tional notation in these various fields as m u c h as possi­
water and sediment control from disturbed areas, it is ble to simplify supplementing the material from other
not the purpose of this treatment to tell the engineer reference sources. T h u s the notation from chapter to
what must be done to comply with existing laws and chapter may differ d e p e n d i n g on the topic u n d e r dis­
regulations. Rather, the purpose of this t r e a t m e n t is to cussion. Within a chapter, a consistent notation was
provide those in charge of sediment and water control used if possible.
General Considerations 3

T h e r e are many worked problems throughout the cross-drainage on slopes, e m b a n k m e n t s , and haul roads
text. Often the calculations for the problems were may result in gullies. Delays in vegetating exposed
carried out using computer programs and spread sheets slopes can result in substantial sheet and rill erosion,
with the final results r o u n d e d to the n u m b e r of signifi­ which, if unchecked, may lead to gullies.
cant figures shown in the problems after all calcula­ It should be kept in mind that natural streams have
tions were completed. In some cases, this rounding at developed over the centuries a state of dynamic equi­
the end of the calculations produces slightly different librium with t h e a m o u n t of sediment and water they
results in the third significant figure than is obtained by carry. W h e n this equilibrium is disturbed, the stream
hand calculations when rounding is d o n e after each attempts to adjust to the new conditions. Thus, in­
step in the calculation. This is especially noticeable creased water and sediment loads may result in stream
when logarithmic or exponential relationships are in­ channel erosion in t h e form of bed and bank erosion. It
volved. may result in sediment deposition within the channel
and thus a reduction in channel water-carrying capac­
ity. It may result in a combination of these things
d e p e n d i n g on the local situation.
GENERAL CONSIDERATIONS Effective water and sediment m a n a g e m e n t is greatly
aided by
Water and sediment m a n a g e m e n t must be consid­
1. preplanning of water and sediment control strate­
ered from the very beginning in developing plans for
gies;
altering the physical setting of a drainage area. Cer­
2. installing diversions and vegetated waterways well
tainly the final desired configuration of the area will
a h e a d of the actual operation so that t h e vegetation
play a dominate role in t h e design. T h e p r e o p e r a t i o n
can be established prior to disturbance;
plan must consider such things as the installation of
3. keeping all reclamation activities current;
perimeter controls to prevent excessive water from
4. exposing and working as small an area at any one
entering the site and from leaving the site in an uncon­
time as practical;
trolled fashion while the site is undergoing change. T h e
5. controlling w a t e r and sediment on the site;
rate and extent of vegetation removal a h e a d of the
6. using good engineering practice in designing wa­
operation, the placement of spoil and topsoil, and the
ter conveyances; a n d
amount of packing or sealing of the final graded fill
must all b e considered. 7. having an aggressive m a i n t e n a n c e program.
W h e n developing plans, every opportunity should be Complete p r e o p e r a t i o n investigation and planning
taken to control water where it falls and sediment are essential. Not only must the natural topography
where it lies as this will generally be m o r e effective, and drainage system be inventoried, but the desired
more permanent, and cheaper than control at some final topography, drainage configuration, and land use
other point. Preventing erosion or accelerated runoff is anticipated. T h e a m o u n t s and types of spoil and soil
preferred to trying to reduce runoff rates and sediment material and w h e r e it is to b e stockpiled or finally
concentrations at later stages in the runoff process. placed must be d e t e r m i n e d . T h e location of roadways
Consideration must be given to controlling sediment and diversions must be specified. T h e n a t u r e and fre­
production on facilities constructed in conjunction with quency of road culverts must be determined as well as
the site itself. For instance, haul and construction the type of road ditches and erosion protection to be
roads are major contributors of sediment and runoff. used in conjunction with these ditches.
Likewise, access roads and construction activities T h e entire operation should be scheduled so that
around sediment detention basins, staging areas, and vegetated channels a n d diversions can b e established
other high traffic areas often result in substantial sedi­ before major land disturbances start and so that all
ment production. reclamation activities can remain current. T h e amount
Water control facilities such as channels, diversions of sediment production from an exposed site is some­
and culverts must be properly designed and main­ what proportional to the length of time the site is
tained. Improper design often results in failures caused exposed. T h u s , a slope exposed for 2 months will likely
by excessive flows or by erosion and sedimentation yield twice the sediment as o n e exposed for 1 month.
associated with normal flows. I n a d e q u a t e consideration W a t e r and sediment control practices should be in­
of the dissipation of energy at the outfall of a culvert stalled at the problem location as much as possible. It
can result in a scour hole and eventual failure of the is considerably c h e a p e r a n d m o r e p e r m a n e n t to do this
culvert. Excessive flow velocities in channels and diver­ as opposed to an a t t e m p t at a more downstream con­
sions can result in the formation of gullies. I n a d e q u a t e trol. In general, water control is also an effective sedi-
4 1 . Introduction

ment control. Erosion tends to increase as the peak ACCEPTED DESIGN PRACTICE VERSUS STATE
flow rate and the runoff volume increase. OF THE ART
Downstream sediment control measures, largely sed­
iment basins, should be installed early before t h e oper­ Accepted design practices a r e those practices that
ation begins. This gives time for a good job in con­ have come into general usage because of their simplic­
structing the facility and for stabilizing all slopes before ity and relative accuracy. Accepted design practices are
the facility is asked to perform the duty for which it emphasized in this t r e a t m e n t . In new areas, sufficient
was designed. In some instances it may be necessary to time has not elapsed for accepted design practices to
introduce chemicals into the s e d i m e n t - w a t e r mixture e m e r g e . In such cases, state of the art p r o c e d u r e s can
in order to cause the sediment particles to aggregate be used. State of the art p r o c e d u r e s are based on t h e
and thus be more easily removed by a sediment basin. latest research but are not yet in general practice. F o r
All water conveyances must be designed to have example, for small detention structures accepted d e ­
a d e q u a t e capacity, to be stable over the range of flows sign practices are available for certain aspects such as
u n d e r which they will be expected to function, and to flood retardation. However, accepted design proce­
have a d e q u a t e energy dissipation. dures for the design of these basins based on detention
It must be recognized that every hydrologic design is time are not available. In this book, a state of t h e art
subject to the random vagaries of natural weather. p r o c e d u r e is proposed. Similarly, state of the art proce­
Regardless of the design used, a certain level of proba­ dures for sediment yield and sediment and erosion
bility exists that the design condition will be exceeded. problems are given.
Determining the acceptable risk of such a failure be­
comes a part of the design process.
Reference
Finally, importance of timely, effective, and routine
maintenance cannot be overemphasized. T h e first sign McBurnie, J. C , Barfield, B. J., Clar, M. L., and Shaver, E. (1990).
of a developing gully or of scour around a structure is Maryland sediment detention pond design criteria and perfor­
the sign for immediate and effective m a i n t e n a n c e . T h e mance. Appl. Eng. Agric. 6(2):167-173.

maintenance operations and procedures must go on for


some time after the completion of the actual operation.
Hydrologic Frequency Analysis

In any discussion of hydrology o n e constantly hears Britain. A five-volume set of reports details the study
such terms as the 100-year flood or the 50-year rainfall. and the resulting recommendations. T h e procedures
Many times these terms are used rather loosely, and used and the general conclusions reached in that study
rarely are they understood by the layman. Frequently, are of general interest. T h e procedures and relation­
the person using these terms does not fully appreciate ships will likely have to be adjusted for catchments
their meaning, the implications associated with them, outside the geographical region covered by the reports.
the difficulty of estimating the magnitude of events T h r o u g h o u t this chapter, a generalized notation is
associated with the terms, and the uncertainty or vari­ used to d e n o t e the events of interest. Γ-year event
ability of an estimate for the magnitude of an event denotes an event with a return period of Γ years
associated with the terms. (return period is yet to be defined). Q denotes the
T

Hydrologic literature is filled with discussions con­ magnitude of peak discharge of a Γ-year flood; Q is T

cerning flood frequency analysis. A review of this liter­ never known with certainty. O n e must always deal with
ature would require a book u n t o itself. A four-volume an estimate for Q . T

set of papers edited by Singh (1987a, b, c, d) provides a All of the statistical procedures, tables, and relation­
comprehensive treatment of many aspects of flood ships that are used in this chapter are independent of
frequency analysis and provides references to h u n d r e d s the units employed. T h u s any consistent set of units
of other works. W h a t follows h e r e is a basic t r e a t m e n t may be used. It does not m a t t e r if flows are in cubic
of frequency analysis and its application to flood flow feet per second, cubic meters per second, or acre-feet
estimation. A user of these techniques must keep in p e r day; the equations and tables in this chapter may
mind that the statistical techniques set forth are hydro- be used without employing any conversion factors. For
logic tools and not hydrologic laws. T h e section this reason, all of t h e example computations in this
"Discussion of Flood Frequency D e t e r m i n a t i o n s " ap­ chapter are carried out using only one set of units.
pearing in this chapter should be read prior to the
actual application of the techniques set forth. H a a n
(1977) can be consulted for a more detailed treatment
of the application of statistics in hydrology. RETURN PERIOD AND PROBABILITY
Perhaps the most comprehensive study on flood flow
estimation was conducted u n d e r the auspices of the It is well known that maximum observed streamflow
Natural Environment Research Council (1975) of G r e a t (the peak flow) observed on any stream over a period

5
6 2. Hydrologic Frequency Analysis

of 1 year varies from year t o year in an apparently event cannot h a p p e n , while a probability of 1 m e a n s
random fashion. This randomness has led t o t h e use of the event will certainly h a p p e n . Sometimes probability
probability and statistics in selecting t h e hydraulic ca­ is expressed as a p e r c e n t a g e chance, in which case the
pacity of storm water facilities. Reference should b e true probability is multiplied by 100.
m a d e to H a a n (1977) for a more complete t r e a t m e n t of So far we have m a d e several assumptions that must
this topic. T h e following is a generalized t r e a t m e n t of be emphasized. T h e assumptions involve t h e variable
hydrologic frequency analysis. (?, t h e peak flow in any year. First, we have assumed
A Γ-year event is formally defined as an event of that t h e p e a k flows from year-to-year a r e i n d e p e n d e n t
such magnitude that over a long period of time (much of each other. This m e a n s that t h e magnitude of a peak
much longer than Γ years), t h e average time between in any year is unaffected by t h e magnitude of a p e a k in
events having a magnitude greater than t h e Γ-year any o t h e r year. Second, we have assumed that t h e
event is Γ years. T h u s t h e expected n u m b e r of occur­ statistical properties of t h e p e a k flows a r e n o t changing
rences of a Γ-year event in an N-year period is N/T. with time. This m e a n s that t h e r e a r e n o changes going
For example, Stillwater, Oklahoma, has a 25-year, 24-hr on within t h e watershed that result in changes in t h e
rainfall of 6.8 in. O n e would expect four occurrences of p e a k flow characteristics of t h e watershed. It further
this 25-year event in a period of 100 years. In a m e a n s that t h e watershed characteristics have re­
100-year record of annual maximum 24-hr rainfalls at mained constant over t h e period of time producing t h e
Stillwater, the expectation is that in 4 of t h e years, t h e data we a r e using. In t h e language of statistics, we
24-hr maximum rainfall would exceed 6.8 in. This is assume that t h e data a r e from a stationary time series.
another way of saying that on t h e average, o n e expects
a Γ-year event t o occur once every Γ years. It is t o b e
emphasized that there is n o regularity associated with a RISK ANALYSIS
Γ-year event. It is not t o b e implied that a Γ-year event
occurs once every Γ years, n o r taken that in any U n d e r t h e assumptions set forth above, t h e occur­
Γ-year period t h e r e will always b e o n e a n d only o n e rence of a Γ-year event is a r a n d o m process meeting
occurrence of a Γ-year event, nor assumed that t h e the requirements of a particular stochastic process
Γ-year event will occur exactly N/T times in Ν years. known as a Bernoulli process. T h e probability of Q T

T h e s e a r e the expectations in a statistical sense b u t a r e being exceeded in any year is p for all time and is T

not certainties. In fact, later we show that t h e r e is a unaffected by any prior history of occurrence of Q . T

chance that in any Γ-year period, a Γ-year event can Let us now d e n o t e any event exceeding Q as β * . W e T

occur 0 , 1 , 2 , . . . , Γ times. Further, we show how to d o not know t h e actual magnitude of Q*; we know
calculate t h e probabilities of these various possibilities. only that it exceeds Q (Q$ > Q ). Q is a Bernoulli
T T T

T h e return period of a Γ-year event as defined above r a n d o m variable. T h e probability of k occurrences of


is Γ years. Often the actual time between occurrences Qj in η years can b e evaluated from t h e binomial
of a Γ-year event is called t h e recurrence interval. T h e distribution 1

average value of recurrence interval is equal to t h e


return period. Most discussions of return period a n d f(k;p ,n)= "l P TV-PT) ~ ,
k n k
(2-2)
T { n ) l k l

recurrence interval assume that t h e two terms a r e


synonymous. In most instances, when o n e uses t h e where f(k; p , n) is t h e probability of exactly k occur­
T

term recurrence interval, t h e average recurrence inter­ rences of Q* in η years if t h e probability of Q* in any
val is meant. single year is p . F o r example, t h e probability of two
T

Since t h e average time between occurrences of a occurrences of a 20-year event (p = 0.05) in 30 years T

Γ-year event is Γ years, t h e probability of a Γ-year is


event in any given year is 1 / Γ . T h u s we have t h e
relationship 30!
/(2;.05,30) = ^ Τ 0 0 5 2
°· 9 5 2 8
= °· 2 6
·

Pr-1/Γ. (2-1) T h e interpretation of this is that in a large n u m b e r


of 30-year records, we would expect 2 6 % of t h e record
where Γ is the return period associated with an event to contain exactly 2 peaks that exceed β ο · T h e other 2

Q and p is t h e probability of Q in any given year.


T T T
7 4 % of t h e 30-year records would contain 0 , 1 , 3 , 4 , . . . ,
Probability is expressed as a n u m b e r between 0 a n d 1 or 30 peaks that equal or exceed β ο · T h e probabilities 2

inclusively. For example, t h e probabilities associated


with 10-, 25-, and 50-year events a r e 0.10, 0.04, a n d
0.02, respectively. A probability of 0 m e a n s that t h e !
Λ ! = Λ ( / Ι - I X " - 2 ) . . . 1 ; 0 ! = 1.
Risk Analysis 7

of the latter number of exceedances can b e evaluated 1000

from Eq. (2.2) also. If this is d o n e , the summation of


the probabilities of 0 , 1 , 2 , 3 , . . . , 30 peaks in 30 years
equal to or greater than Q must equal 1.00 since all
2Q

possibilities have been exhausted.


Equation (2.2) can be used to calculate the probabil­
ity that a Γ-year event will be exceeded at least once in
an η-year period by noting that " a t least o n c e " m e a n s
* 100
so/
v:
one or more. T h e probability of one or m o r e ex­
ceedances plus the probability of n o exceedances must
equal 1.00. Therefore the probability of at least o n e
exceedance is given by 1 minus t h e probability of n o
exceedances or

l-/(0;p ,n) r = /v)".

Since p T = 1 / Γ and 0! = 1, this relationship reduces


to

f(p ,n)
T = l - ( l - l / T ) " , (2.3)

where f(p ,n)T is the probability that a Γ-year event 1 10 100


will be exceeded at least once in an η-year period. If η Design Period, T (yrs)
d

is equal to Γ in Eq. (2.3), it can b e shown that f(p , T T)


Figure 2.1 Design return period required as a function of design
approaches the constant 0.632 for large Γ (for Γ = 10,
life to be given percentage confident (curve parameter) that the
f(p ,T)
T = / ( 0 . 1 , 1 0 ) = 0.65). W h a t this m e a n s is that design condition is not exceeded.
if a structure having a design life of Γ years is designed
on the basis of a Γ-year event, t h e probability is about
0.63 that the design capacity will be exceeded at least
b e acceptable since what is m e a n t by exceeded is
once during t h e design life.
failure of the structure to handle the resulting flow in
By specifying the acceptable probability of the design
t h e m a n n e r t h e structure was designed to operate.
capacity being exceeded during the design life of the
Failure in this sense does not necessarily mean that
structure, Eq. (2.3) can be used to calculate t h e re­
the structure will b e destroyed. For example, the fail­
quired design return period. For example, if o n e wants
ure of a road culvert to pass a p e a k flow may result in
to be 9 0 % sure of not exceeding the design capacity of
only minor flooding of a roadway or adjacent area and
a structure in a 25-year period, f(p , 25) would be
may b e acceptable on a fairly frequent basis. O n the
T

1 - 0.90 = 0.10. Thus from Eq. (2.3),


other hand, failure of a storm water detention basin
0.10 = 1 - ( 1 - 1/Γ) , 2 5 may result in overtopping of the structure with consid­
erable d a m a g e to property and high risk of loss of life
which can be solved to yield Γ = 238 years. T o b e 9 0 % downstream. T h u s the selection of the acceptable risk
sure of not exceeding the design capacity in a 25-year and design r e t u r n period d e p e n d s on the consequences
period, the design capacity must be based on an event of the design capacity being exceeded. Building the
with a return period of 238 years. In this case, t h e structure large e n o u g h to protect against extremely
acceptable risk was 10%, the degree of confidence was rare events is quite expensive, while allowing the de­
9 0 % , the design life was 25 years, and t h e required sign capacity to b e exceeded on a frequent basis may
design return period was 238 years. Calculations like result in an accumulation of considerable economic
this can be carried out for various design lives, design loss. T h u s , in addition to social and political considera­
return periods, and acceptable risks. Figure 2.1 is based tions, t h e selection of t h e p r o p e r design return period
on such calculations and can be used to quickly deter­ is a problem in economic optimization.
mine the required design return period based on the Figure 2.2 illustrates t h e selection of a design return
design life and acceptable risk or probability of having period based on economic optimization. T h e vertical
the design capacity exceeded. scale contains average annual costs or benefits and
In these discussions, it should be kept in mind that a the horizontal scale contains return period in years.
high risk of having the design capacity exceeded may Average a n n u a l costs should reflect all costs such as
8 2. Hydrologic Frequency Analysis

exceeded. F o r example, in rural areas, r o a d culverts


might b e based on a 10-year r e t u r n period. Minor
structures in u r b a n areas might b e b a s e d on t h e 25-year
event, and major structures and flood plain delin­
eations might b e based on t h e 100-year event.

FREQUENCY DETERMINATIONS

Assigning a flood m a g n i t u d e to a given r e t u r n period


requires knowledge of t h e flood flow characteristics of
t h e basin of concern. T h e a p p r o a c h that is used to
d e t e r m i n e this relationship d e p e n d s largely on t h e
type, quantity, and quality of hydrologic d a t a available
and on the importance of t h e determination. If a minor
culvert or channel is to b e designed, o n e cannot justify
a time-consuming, expensive flood frequency analysis.
RETURN PERIOD, Τ O n t h e o t h e r h a n d , if a major c o m p o n e n t of a drainage
Figure 2.2 Determination of economically optimum design return system is u n d e r construction, t h e best possible flow
period. estimates a r e desired.
In this t r e a t m e n t , five cases or situations a designer
might b e faced with a r e considered:
Case I: A reasonably long record of streamflow is
planning, right-of-way, construction, operation, and available at or n e a r t h e point of interest on
maintenance. T h e horizontal dashed line indicates the the stream of interest.
average annual damages that are occurring without t h e Case II: A reasonably long record of streamflow is
project. Obviously as the design return period in­ available on the stream of interest, but at a
creases, the design capacity would increase, and aver­ point somewhat removed from t h e location
age annual costs would increase as well. Average of interest.
annual benefits are taken as damages prevented on an Case III: A short streamflow record is available on
average annual basis. As the capacity of the system the stream of interest.
increases, the level of protection against d a m a g e in­ Case IV: N o records are available on t h e stream of
creases until presumably all damages could be elimi­ interest, but records are available on nearby
nated. T h e net effect is that average annual benefits streams.
generally are low for low-return period designs and Case V: N o streamflow records are available in t h e
asymptotically approach a constant value as t h e design vicinity.
return period becomes very large.
T h e cases are listed in t h e o r d e r they are considered.
T h e average annual net benefits is the difference in They a r e also listed in t h e o r d e r of increasing difficulty.
average annual costs and benefits. A s shown in Fig. 2.2, Unfortunately, they are listed in t h e inverse o r d e r of
a definite maximum average annual net benefit is ap­ their frequency of occurrence. T h a t is, t h e designer is
parent and represents the economically o p t i m u m de­ m o r e likely to be faced with Case V than with Case I,
sign return period. T h e s e concepts are difficult to apply especially for small watersheds. In spite of this, we
if damages include intangibles such as loss of life or devote a major part of o u r attention to the t r e a t m e n t
destruction of nonreplaceable items. of Case I, because it is essential that t h e Case I
O n e of the factors that inhibit the application of the procedures and their limitations b e understood before
economic approach to the selection of design r e t u r n o n e can appreciate t h e problems associated with any of
periods is the sensitivity of the approach to t h e interest the o t h e r cases. T h e Case I analysis is basic to any
rate used in the analysis. Many designers are reluctant flood frequency analysis. Case V is t r e a t e d extensively
to let the prevailing interest rate d e t e r m i n e the capac­ in the next chapter.
ity of a structure and thus its risk of failure.
Many governmental units have regulations governing
C a s e I. L o n g Flow R e c o r d a t Site
the design period to be used. Often these return peri­
ods are based on the size of the structure and the Several agencies of t h e U n i t e d States G o v e r n m e n t
consequences of the structural hydraulic capacity being sponsored a study to develop a uniform technique for
Frequency Determinations 9

flood frequency analysis. T h e result was a publication the equations


"Guidelines for Determining Flood Flow Frequency,"
Bulletin 17B (Interagency Advisory C o m m i t t e e on W a ­ - ^ x
i
(2.4)
ter D a t a , 1981). Some of the material in this section
comes from that report. This guide states, "Major
problems are encountered when developing guides for Σ A', -2
nX 2

(2-5)
flood flow frequency determinations. T h e r e is no pro­ η - 1
cedure or set of procedures that can b e a d o p t e d which,
when rigidly applied to the available data, will accu­
(2-6)
rately define the flood potential of any given water­ v
X
shed. Statistical analysis alone will not resolve all flood
frequency problems." nZ(X i - X)*
(2.7a)
If one is extremely fortunate, a relatively long record
C s
~ (n - l)(n - 2)S%
of peak flows may be available on t h e stream at t h e
point where an estimate for a flood peak of a given n ZXf
2
- 3ηΣΧ,ΣΧ? + 2(ΣΛ' ) ί
3

, (2.7b)
frequency is desired. Such a listing might a p p e a r as in n(n - l)(n - 2)S X

Table 2.1 for the Middle Fork of Beargrass C r e e k at


C a n n o n s Lane in Louisville, Kentucky. Any collection w h e r e Χ represents t h e ith data value, η is the sample
έ

of data such as contained in T a b l e 2.1 represents a size, and all summations are from 1 to n. Applying
sample of data from a population and u n d e r certain these equations to the Beargrass C r e e k data results in
assumptions can be treated using probability and statis­ X = 1599 cfs, S = 1006 cfs, C = 0.619, and C =
x v s

tics. T h e population in this case would be the maxi­ 2.13.


m u m annual flood p e a k for all time, both past and T h e m e a n is simply a m e a s u r e of the central location
future. T h e data of Table 2.1 r e p r e s e n t s a sample from of a group of data. T h e s t a n d a r d deviation is a measure
this population. of t h e spread of t h e data. T h e larger the standard
Quantities descriptive of a population are known as deviation, the g r e a t e r the spread in the data. T h e
parameters. Population p a r a m e t e r s are never known in square of the s t a n d a r d deviation is known as the vari­
a flood frequency study and must b e estimated from ance. T h e units on t h e s t a n d a r d deviation are the same
the sample of data. Estimates of population p a r a m e ­ as t h e units on t h e raw d a t a . A dimensionless measure
ters are known as sample statistics. Some p a r a m e t e r s of the spread of a set of data is desirable so that
of interest are the mean, μ ; the standard deviation,
χ comparisons of relative variability can be m a d e among
σ \ the coefficient of variation, C \ and the skewness,
χ v variables having widely differing m e a n s or among mea­
γ . Sample estimates for μ , σ , C , a n d γ are given by
χ χ y sures having different units. For example, the variance
X, S , C , and C , respectively, and calculated from
x v s of p e a k flows in t h e Mississippi River is much greater
than for a small tributary stream, yet the relative
variability (relative to the m e a n ) of flow in the tributary
Table 2.1 Peak Discharge (cfs): M i d d l e Fork, Beargrass Creek, would be larger t h a n that of the Mississippi. O n e such
C a n n o n s Lane, L o u i s v i l l e , Kentucky m e a s u r e is the coefficient of variation which is the
standard deviation divided by the m e a n . A compact
Peak Peak Peak data set will have a smaller coefficient of variation than
Year flow Year flow Year flow
will a wide ranging set of data.
1945 1810 1956 1060 1966 874 T h e skewness is a m e a s u r e of the symmetry of a
1946 791 1957 1490 1967 712 distribution. T h e normal distribution has a skewness of
1450
zero. If the data t e n d s to spread, or tail, to the right
1947 839 1958 884 1968
m o r e than it does to t h e left with respect to its mean,
1948 1750 1959 1320 1969 707
the data are positively skewed and C will be positive. s
1949 898 1960 3300 1970 5200
D a t a tailing to the left m o r e t h a n to the right is
1950 2120 1961 2400 1971 2150
negatively skewed and C will b e negative.
s

1951 1220 1962 976 1972 1170


Equations (2.4) through (2.7) indicate that statistics
1952 1290 1963 918 1973 2080 are descriptive and not causal. Any statistical anlaysis
1953 768 1964 3920 1974 1250 is an a t t e m p t to describe, often in a probabilistic man­
1954 1570 1965 1150 1975 2270 ner, the behavior of a set of d a t a . Obviously this
1955 1240 description must be u p d a t e d as new information be­
comes available. As additional years of data are incor-
10 2. Hydrologic! Frequency Analysis

porated into an analysis, the statistics given by Eqs. In some locations, flood flows may b e t h e result of
(2.4) through (2.7) will change, and thus any flow two distinct meteorologic causes. F o r example, winter
estimates that d e p e n d on these statistics will change as flows may be the result of frontal storms and s u m m e r
well. In general, as the n u m b e r of observations in­ flows may be t h e results of convective thunderstorms.
creases, the statistics become better estimates of the If this type of nonhomogeneity is present and flows can
population p a r a m e t e r s . be easily divided into two groups according to t h e
If data such as contained in Table 2.1 m e e t certain storm type, it may b e desirable to treat the two storm
assumptions, we can consider them to be i n d e p e n d e n t types separately and t h e n combine t h e results proba­
random variables and subject them to a frequency bilistically. T h e problems with this approach are the
analysis. T h e main assumptions are that the d a t a are difficulty of actually dividing t h e flows along causative
i n d e p e n d e n t of each o t h e r and are from a stationary lines, and the length of record available in each part of
time series. A stationary time series is a d a t a series the divided record may b e too short to provide reliable
collected over time and having statistical properties estimates of the required statistical p a r a m e t e r s . H a a n
that do not change over time. (1977) can b e consulted for m o r e details on t h e use of
In hydrologic terms, the statistical assumptions re­ mixed populations a n d mixed distributions in flow fre­
quire: quency analysis.
In any d a t a analysis, m e a s u r e m e n t errors are of
1. T h e r e are no trends in the data.
concern. In flood frequency determinations, it is gener­
2. T h e data represent i n d e p e n d e n t hydrologic events.
ally assumed that the d a t a are m e a s u r e d without error.
3. T h e r e is one underlying meteorologic/hydrologic
If actual m e a s u r e m e n t errors are i n d e p e n d e n t from
cause for the flows so that the flows can be assumed
o n e m e a s u r e m e n t to another, tend to overestimate
to be from a single population.
flow as well as u n d e r e s t i m a t e flow so as to have a m e a n
4. M e a s u r e m e n t errors are random, unbiased, and have
error of zero, and are small in comparison to the flow
a relatively small variance.
itself, t h e assumption of n o m e a s u r e m e n t errors is
T r e n d s in data may be caused by climatic shifts, generally acceptable from a hydrologic standpoint. O b ­
natural events, or h u m a n activities. Hydrologists gener­ viously if m e a s u r e m e n t s always p r o d u c e low estimates
ally consider the time scale of climatic change to be or high estimates a n d / o r are grossly in error, any
vary large in comparison to the period of concern in analysis based on the d a t a will be in e r r o r as well.
any analysis and thus do not consider possible climatic Two types of d a t a series are commonly used in flow
shifts. Major natural events such as e a r t h q u a k e s , land frequency analysis—the annual series and t h e partial
slides, and forest fires can cause changes in t h e hydro- duration series sometimes known as the " p e a k s over
logic regime of a catchment and thus introduce n o n h o - t h r e s h o l d " series. In t h e annual series, t h e d a t a consist
mogeneity into the flow record. If such is t h e case, this of the largest observed p e a k flow for each year of data.
nonhomogeneity must be dealt with prior to any fre­ For t h e partial duration series, t h e d a t a consists of all
quency analysis. peak flows greater t h a n some base or threshold value.
T h e most common cause of changes in the flow T h e annual series p r o d u c e s o n e d a t a value per year.
regime of a catchment is h u m a n activity. This may be T h e partial duration series may p r o d u c e n o n e , o n e , or
in the form of land-use changes such as urbanization, m o r e t h a n o n e d a t a value in any year d e p e n d i n g on t h e
deforestation, or surface mining activities. It may be in flows for t h e year a n d t h e m a g n i t u d e of t h e threshold
the form of reservoir construction, stream diversions, value.
or channel work. Sometimes these changes are s u d d e n For r e t u r n periods g r e a t e r than about 10 years, t h e
and easily detected. Closure of a major reservoir can r e t u r n period flow estimate for the two series are
have immediate and obvious impacts on flood flow practically t h e same. F o r m o r e frequent but smaller
magnitude. Sometimes the changes are gradual. Ur­ floods, the relationship b e t w e e n t h e estimates from t h e
banization may show up in the data as a trend toward two series is somewhat d e p e n d e n t on t h e probability
higher peak flows. Generally trends of this type are distribution selected. In this t r e a t m e n t , the annual
difficult to detect over a short period of time d u e to the series is used.
r a n d o m nature of flood flows. If changes of this type
are present, the data must be adjusted for the changes Probability Plotting
before they can be treated in a straightforward way Summarizing the d a t a in t h e form of a probability
using statistics. References can be m a d e to H a a n (1977) plot is often the first step in a frequency analysis. A n
for possible ways to adjust for nonhomogeneity in the intuitive estimate for t h e m a g n i t u d e of frequent floods
flow record. on Beargrass Creek can be m a d e based on our u n d e r -
Frequency Determinations 11

0.7

0.6 h

Ο
c 0.5
0)
σ­
α) 0.4

> °· 3

σ
» 0 . 2 Ι­
ο:

0.1

0.0
0 15 30 45 1000 2000 3000 4000 5000 6000
Flow ( 1 0 0 ' s c f s ) Flow ( c f s )
Figure 2.3 Frequency histogram—Beargrass Creek data. Figure 2.4 Empirical flood frequency—Beargrass Creek data.

standing of the concept of r e t u r n period. For example, W h e n considerable data are available, this is a rea­
the 5-year flood is one that is equaled or exceeded on sonable p r o c e d u r e to use for estimating low-return
the average once every 5 years or about 2 0 % of the periods floods. Inspection of Fig. 2.4 shows that the
time. Looking at Table 2.1, we see that about 2 0 % (six d a t a exhibit some " r o u g h n e s s " and that p e r h a p s a
events) of the peaks exceed 2120 cfs. Therefore, we b e t t e r estimate for low-return period floods could be
might estimate the m a g n i t u d e of t h e 5-year flood as obtained by drawing a smooth curve through the data
2120 cfs. Similarly 10% of t h e flows exceed 2400 cfs so a n d t h e n using t h e curve t o define t h e magnitude of
we can estimate the 10-year event as 2400 cfs. floods with various r e t u r n periods.
A difficulty with the intuitive approach to flood fre­ Unfortunately a plot such as Fig. 2.4 is generally not
quency analysis is that the m a g n i t u d e of events having sufficient for estimating t h e m a g n i t u d e of a longer
return periods longer than the length of the available r e t u r n period flood. F o r example, t h e 25-year flood can
record cannot be estimated. Also the m a g n i t u d e of be d e t e r m i n e d from Fig. 2.4 by reading the smooth
events having return periods close to the record length curve at the 4 % point. This is not a very reliable
is d e p e n d e n t on very few observations a n d is t h u s estimate, however, because it d e p e n d s almost entirely
somewhat uncertain. For example, the 10-year event in on t h e m a g n i t u d e of the two largest events in the
the above example d e p e n d s on only t h r e e observations. record. If t h e largest flood event in t h e record had
W h a t is n e e d e d is a p r o c e d u r e for using all of t h e d a t a b e e n 7000 or 4200 cfs or some o t h e r value, this would
to describe the probabilistic n a t u r e of t h e p e a k flows. have greatly altered o u r estimate for t h e 25-year flood.
A start in this direction can be m a d e by plotting the F u r t h e r m o r e , t h e estimation of a 100-year flood
data in the form of a frequency histogram. This is based on these d a t a requires the smooth curve be
merely a plot of the frequency of occurrence of p e a k extrapolated to the 1% point. This extrapolation, and
flows in some class interval versus the class interval. indeed the entire smooth curve, would be extremely
Figure 2.3 is such a plot using a class interval of d e p e n d e n t o n t h e whims of t h e individual doing the
750 cfs. Similarly a plot of t h e p e r c e n t a g e of t h e values extrapolation. Different individuals would estimate
greater than or equal to a given value versus the different values for t h e 100-year flood, and the values
magnitude of the value can be m a d e . Figure 2.4 is a could differ by 5 0 % or m o r e .
plot of this n a t u r e for the Beargrass C r e e k data. W h a t is n e e d e d is an analytic m e t h o d for placing a
From Fig. 2.4, the magnitude of the 5-year flood curve t h r o u g h t h e plotted points. This analytic curve
(/? = 1/T = 1/5 = 0.20 or 2 0 % chance of occurrence) could t h e n be used to estimate the magnitude of floods
can be estimated as about 2150 cfs and t h e 10-year with various r e t u r n periods. Before discussing analytic
flood ( 1 0 % chance of occurrence) is about 3250 cfs. techniques for flood frequency analysis, the matter of
12 2. Hydrologic Frequency Analysis

plotting r a n d o m data (flood peaks) requires further Table 2.2 Plotting Position: Middle Fork,
Beargrass Creek, Cannons Lane, Louisville,
attention.
Kentucky
T h e procedure arrived at in preparing Fig. 2.4 re­
sults in the point 707 cfs being plotted at the 100% Plotting
point. This is equivalent to stating that 100% of all Year Discharge Rank position
annual flood peaks on this stream will be greater than
707 cfs. Even though this is true for the particular 1945 1810 9 0.281
31-year record that is available, we d o not know that it 1946 791 28 0.875
is true for all time and would suspect that t h e r e is a 1947 839 27 0.844
chance that in some future year an annual p e a k of less 1948 1750 10 0.313
than 707 cfs might occur. T h u s we would like to avoid 1949 898 24 0.750
assigning a 100% chance or probability of 1 to any
1950 2120 7 0.219
event.
1951 1220 18 0.563
A second consideration in plotting flood peaks
1952 1290 15 0.469
against probability is that w h e n arithmetic graph p a p e r
1953 768 29 0.906
is used as in Fig. 2.4, the points generally form an
1954 1570 11 0.344
extremely curved pattern with the larger floods widely
spaced. T o overcome this inconvenience, special p a p e r 1955 1240 17 0.531

known as probability p a p e r has b e e n developed. Sev­ 1956 1060 21 0.656


eral kinds of probability p a p e r are available. T h e most 1957 1490 12 0.375
widely available are normal probability p a p e r and log- 1958 884 25 0.781
normal probability paper. Lognormal probability p a p e r 1959 1320 14 0.438
is used in this treatment. 1960 3300 3 0.094
T h e steps to be followed in plotting r a n d o m d a t a on 1961 2400 4 0.125
probability p a p e r are to
1962 976 22 0.688
1. rank the data from the largest to the smallest; 1963 918 23 0.719
2. calculate the plotting position, p, based on the 1964 3920 2 0.063
rank, m , and the n u m b e r of years of data, n\ and 1965 1150 20 0.625
3. plot the observation on probability p a p e r with ρ 1966 874 26 0.813
along the probability scale and m a g n i t u d e along the 1967 712 30 0.938
variable scale.
1968 1450 13 0.406
Several plotting position relationships are in use.
1969 707 31 0.969
A general relationship is
1970 5200 1 0.031
m - a 1971 2150 6 0.188
η - a - b + 1 1972 1170 19 0.594
1973 2080 8 0.250
where a and b are constants. T h e California (Cali­
1974 1250 16 0.500
fornia State D e p a r t m e n t or Public Works, 1923) plot­
ting position is ρ = m/n. T h e H a z e n (1930) relation is 1975 2270 5 0.156

ρ = (2m - l)/2n. T h e Natural Environment R e s e a r c h


Council (1975) of the United Kingdom used ρ =
( m - 0 . 4 4 ) / ( M + 0.12). T h e most widely used relation­
ship in the U.S. is the Weibull (1939) relationship
given by tion of t h e values greater t h a n or equal to t h e corre­
sponding value of t h e d a t a . T h e d a t a d o not plot as a
straight line on lognormal p a p e r , b u t the curvature is
greatly r e d u c e d over that shown in Fig. 2.4.
As an example of probability plotting, consider t h e At this point a smooth curve can b e sketched through
Beargrass Creek data. These data are ranked and the t h e d a t a or we can use analytical m e t h o d s to "fit" a
plotting positions determined in Table 2.2. Figure 2.5 is line through the points. In this latter approach, an
a plot of the data on lognormal probability p a p e r . equation having unknown p a r a m e t e r s is used to d e ­
Since the data were ranked from the largest to the scribe t h e data much like t h e straight line y = a + bX
smallest, the plotting position, p, represents t h e frac- is fitted through plotted points on regular graph p a p e r .
Frequency Determinations 13

Return Period (yrs)


5 0 2 0 10 5 2 1.11 1.02
7000 ι i i ι ι ι ι I I I I I I
6000 Flood frequency
5000 ο
Middle F o r k B e a r g r a s s Creek]
4000
Μ
^ 3000

s
OOo
° 2000

1000 ο
Cv
900
800
700 ο ο
600 J 1 I I I I I
2 5 10 30 50 70 90 95 99
Figure 2.6 Normal distribution.
Probability ( X )
Figure 2.5 Probability plot—Beargrass Creek data.

pdf is given by

T h e difficulty we now face is selecting the " e q u a t i o n "


exp (2.10)
to use and in estimating the p a r a m e t e r s of this equa­ σ ^2π
χ 2(σ ) 2
χ

tion.
T h e normal distribution is symmetrical about the
Probability Distributions mean μ χ and ranges from — oo to oo. T h e normal
Equations for describing the probability of occur­ distribution is generally not used in flood frequency
rence of random events are known as probability determinations because it permits negative values and
density functions (pdf) and cumulative distribution because flood frequency distributions are generally not
functions (cdf). A pdf can be used to evaluate the symmetrical. For example, the Beargrass Creek data in
probability of a random event in a specified interval. Fig. 2.3 exhibit a p r o n o u n c e d tailing off to the right,
A cdf can b e used to evaluate the probability of an which is typical of flood p e a k data. Even though the
event being equal to or less than a given value. W e use normal distribution is generally not used in flood fre­
the notation p (x)x and P (x) to d e n o t e the pdf and
x
quency analyses, we continue to consider it since an
cdf of the random variable X evaluated at X = x. understanding of it is essential for statistical work.
These two are related by T h e cdf of the normal distribution is

f X 1
p
x(x) = / — 7 ^ = - exp dx (2.11)
(2.9)
y

2(σχΫ

which gives the probability that X < x.


T h e r e are a limitless n u m b e r of functions that can be
P (x) = prob(Z <JC).
used for p d f s. R e q u i r e m e n t s for a function to b e a pdf x

are T h e probability that X is between a and b can be


evaluated from
P {x) > 0 for all χ
x
p r o b ( a < X < b) = p r o b ( X < b) - p r o b ( X < a)

/ P (x)dx=l.
x
= Px{b)-P {a)x
(2.13)
J
— 00
fb
= / Ρχ( ) dx.
χ

P d f s may take on any number of shapes. T h e most


familiar is the bell-shaped curve of t h e normal proba­ T h e normal distribution is a two-parameter distribu­
bility density function shown in Fig. 2.6. T h e normal tion with the p a r a m e t e r s being μ , the m e a n of X and
χ
14 2. Hydrologic Frequency Analysis

σ , the standard deviation of X. For any application of


χ standard deviation of 1006 cfs, t h e n 81.4% of the
the normal distribution, we must estimate the popula­ annual p e a k s should b e less t h a n 2500 cfs a n d 18.6% of
tion m e a n and standard deviation, μ a n d σ , by their χ χ t h e annual p e a k s should b e greater t h a n 2500 cfs. T h u s
sample estimates, X and S . x 2500 cfs is assigned a r e t u r n period of 1/0.186 or 5.4
Using Eqs. (2.4) and (2.5), the m e a n and standard years u n d e r t h e normality assumption. U n d e r this as­
deviation of the Beargrass Creek data are found to b e sumption, we expect 2500 cfs to b e exceeded on the
1599 and 1006 cfs, respectively. Now if t h e normal average once every 5.4 years. T h e d a t a tabulation actu­
distribution was an a d e q u a t e representation of the ally shows that 28 of t h e 31 values or 9 0 . 3 % are less
Beargrass Creek data, it could be used to m a k e proba­ than or equal to 2500 cfs.
bilistic statements concerning the data. For example, Looking back at Eq. (2.13), it is a p p a r e n t that
the probability of a peak less than or equal to 2500 cfs prob(fl < X < b) is the area u n d e r the pdf, p (x\ x

could be evaluated as b e t w e e n X = a and X = b. T h u s , t h e probability of a


r a n d o m observation falling in t h e interval a to b is t h e
prob(<2 < 2500) = Ρ ( 2 5 0 0 ) β area u n d e r t h e pdf b e t w e e n a and b. In a sense, the
relative frequency histogram of Fig. 2.3 gives similar
2500 1 -(x - 1599) 2

exp dx. information. Based on t h e d a t a in hand, we would


1006V2ir 2(1006) 2
estimate for example, t h a t t h e probability that a ran­
(2.14) d o m annual p e a k would fall in the interval 1500 to
2250 cfs is 0.190. T h e r e is apparently a relationship
Unfortunately this latter expression cannot be analyti­ between relative frequency and probability. D e n o t e by
cally evaluated, and numerical procedures must be f (Xi)
x the relative frequency of observations in an
used. T o overcome the problem of requiring a s e p a r a t e interval of width Δ * c e n t e r e d on T h e probability of
numerical integration for the normal distribution for an observation falling in this interval is
every possible combination of the p a r a m e t e r s μ and χ

σ , a transformation of variables is defined as


χ prob(jt, - Ax/2 <X <x + l; Ax/2)

(Χ- μ)
χ /·*,· + Δ * / 2
Ζ = (2.15) = / p (x)dx,
x (2.18)
J
x-kx/2

Z is called a standardized normal variable and has t h e


property that μ = 0 and σ = 1. T h e expression which is the a r e a u n d e r p (x) b e t w e e n x - Ax/2 and
x i
ζ ζ

x + Ax/2.
k This a r e a can b e approximated by
-z /2
2
Αχρ {χ^ χ which is t h e width of t h e interval times t h e
(2.16) height of p (x)
x evaluated at JC, (Fig. 2.7).
Therefore, t h e relationship b e t w e e n t h e relative fre­
is known as the standard normal distribution. Equation quency of observations in an interval Δ * and the pdf is
(2.14) can now be evaluated as
/*(*«) = */>*(*/)· Δ
(2.19)
prob(£? < x) = p r o b ( Z < (χ - μ )/σ ) χ χ (2.17)

or As a side n o t e , since probability is related to t h e


area u n d e r t h e pdf, it is a p p a r e n t that p r o b i ^ = x) for
p r o b ( Q < 2500) = p r o b ( Z < (2500 - 1599)/1006)
a continuous r a n d o m variable must b e zero since
= p r o b ( Z < 0.896)

rO.S96e~ z2/2 χ
p (x) dx. (2.20)
dz. / x

T h e latter expression can be evaluated using tables W e can use Eq. (2.19) and Fig. 2.3 to visually judge
of the standard normal distribution such as that con­ the appropriateness of using t h e normal distribution to
tained in Appendix 2A. Appendix 2A gives the desired describe the Beargrass C r e e k data. T a b l e 2.3 shows,
probability, p r o b ( ( Z < 0.896), as 0.814. This corre­ u n d e r the assumption of a normal distribution, the
sponds to the p r o b ( g < 2500). observed and expected frequency of observations in
T h e interpretation of this calculation is that if the several classes. T h e d a t a are plotted in Fig. 2.8. Entries
flood peaks on Beargrass Creek can be described by a in t h e expected relative frequency column of Table 2.3
normal distribution with a m e a n of 1599 cfs a n d a a r e based on Eq. (2.19) and t h e normal distribution.
Frequency Determinations 15

Table 2.3 Observed and Expected Frequency:


Beargrass Creek Data (Normal Distribution)

Class Observed Expected


interval relative frequency relative frequency

0-750 0.064 0.141


750-1500 0.581 0.267
1500-2250 0.194 0.286
2250-3000 0.064 0.177
3000-3750 0.032 0.063
3750-4500 0.032 0.012
4500-5250 0 032 0.001

0.999 0.947

A second visual comparison between the observed


data and their assumed distribution (the normal distri­
Figure 2.7 Calculation of probU, - Δ * / 2 < χ < χ, +; Δχ/2).
bution) can be m a d e by using the normal distribution
as t h e equation for t h e line describing the data in Fig.
2 . 4 . Equation ( 2 . 1 1 ) can be used to draw a line through
t h e points of Fig. 2 . 4 , assuming the points are from a
For example, for the second class normal distribution. All that is required is to calculate
the p r o b ( Q > q) for the various values of Q, and then

1 -(1125 - 1599)
to plot this probability versus q. T h e prob((2 q) is >
equal to 1 - p r o b ( Q < q) since prob(£) = q) is zero
2

f(
x 1125) = 750 exp
1006V2tt 2(1006) 2 and Q must either be < q or > q. T o obtain
p r o b ( g > q), we first evaluate prob(£) < q). Equations
= 0.267. ( 2 . 1 4 ) and ( 2 . 1 7 ) show such a calculation for Q =
2 5 0 0 cfs. Table 2 . 4 shows the results of similar calcula­
tions for several values of Q. T h e prob((? > q) is
0.6 plotted in Fig. 2 . 9 .

0.5
HI expected
υ Table 2.4 Comparison of Observed and Expected
I I observed Cumulative Probabilities (Normal Distribution)
» 0.4
XT
Observed Expected Expected
ul 0.3 percentage percentage percentage
Q *Q
>
0.2 700 100.0 18.7 81.3
0) 1000 67.7 28.8 71.2
0.1 1500 35.5 46.0 54.0
2000 25.8 65.6 34.4

0.0 2500 9.6 81.6 18.4


0 15 30 45 3000 9.6 91.8 8.2
Flow ( 1 0 0 ' s c f s ) 4000 3.2 99.2 0.8
Figure 2.8 Comparison of observed and expected flow frequency 5000 3.2 100.0 0.0
(under assumption of normal distribution)—Beargrass Creek data.
6000 0 100.0 0.0
16 2. Hydrologic Frequency Analysis

T h e pdf for the Pearson type III distribution is

a Px(x) = P 0 e - ^ s
( x / a ) a / S

D
σ­
ο with the m o d e at X = a and the lower b o u n d at
i_
o X = 0. T h e difference in t h e m o d e and t h e m e a n is δ
c and p is p (a). Most applications of t h e Pearson type
σ Q x

III distributions in hydrologic frequency analysis have


b e e n based on the log Pearson type HI distribution
ο obtained by converting all of t h e observations to their
logarithms and t h e n applying t h e Pearson type III to
these logarithms. Resulting estimates must of course
C
Ο be transformed back to their original units.
υ
k. As noted earlier, the Beargrass C r e e k data, when
β
Q.
plotted in the form of a relative frequency histogram,
tailed off to t h e right much m o r e t h a n to t h e left. This
0 1000 2000 3000 4000 5000 6000
tailing off to t h e right results in a positive skewness.
Flow ( c f s )
T h e normal distribution is symmetrical about the m e a n
Figure 2.9 Flood frequency comparison for Beargrass Creek data and, as such, has a skewness of zero. T h e LN, E V I ,
—observed and normal distribution.
and LP3 distributions can all a c c o m m o d a t e positively
skewed data.
For the L N distribution, the skewness y and the
F r o m either Fig. 2.8 or 2.9 it is a p p a r e n t that the coefficient of variation C are related by v

normal distribution is not a satisfactory approximation


y = 3C + C .
3
(2.21)
to the observed data of Beargrass Creek. A n o t h e r V W

probability distribution must be found to describe the F o r the E V I distribution γ is a constant 1.139. T h e r e
data. This involves finding a n o t h e r mathematical func­ are no restrictions on y for t h e L P 3 distribution since
tion to use as a pdf and cdf in place of Eqs. (2.10) and it can take on any value t h e sample of data yields. T h e
(2.11) used for the normal distribution. A large n u m b e r skewness is an important identifier of potential distri­
of such expressions are available. Again, these expres­ butions that might be used to describe a set of data.
sions are known as probability distributions. F o r the distributions we have considered, t h e skewness
T h e t h r e e probability distributions that receive the is given by
most attention for describing flood frequencies are the
lognormal (LN), extreme value type I (EVI), and log
Distribution Skewness
Pearson type III (LP3). This t r e a t m e n t is restricted to
these three distributions. O t h e r distributions are dis­ Normal 0
cussed in H a a n (1977). Lognormal 3C
V
T h e pdf for the lognormal distribution is Extreme value I 1.139
Log Pearson III Any value
1 -(In χ - μ)
γ

Px(x) = j = exp
χσ \2π 2{σ ) 2

Chow (1951) has shown that many types of frequency


γ
γ

analyses can b e reduced to


where μ and σ are the m e a n and standard deviation
γ Υ

of the natural logarithms of X. (2.22)


XT = X(1 + C K ),
V T
T h e pdf for the extreme value type I distribution for
maximums is w h e r e X j s t h e m a g n i t u d e of the event with r e t u r n
T

period Τ, X is the m e a n of t h e original data, C is the v

1 (χ-β) I -(χ-β) coefficient of variation of t h e original data, a n d K is a


Ρχ{ ) χ
= ~ e x
P - exp T

a a a frequency factor that is a function of t h e probability


distribution selected and properties of t h e original
where a and β can be estimated as data.
A comparison of E q s . (2.15) and (2.22) shows that
K for the normal distribution is equal to the standard­
a = β = 0.45**. T

ized normal variate Z . T h u s , A p p e n d i x 2 A serves as a


Frequency Determinations 17

table of frequency factors for the normal distribution. Table 2.5 Frequency Factors for Lognormal Distribution
(Chow, 1964)
T h e frequency factors for the LN distribution as a
function of C are contained in Table 2.5. Table 2.6
v
Return period
contains the frequency factors for the EVI distribution.
1.01 2 5 20 100 Corresponding C
All that is required for selecting K for this distribu­ T
v

tion is knowledge of the sample size and the desired


-2.33 0 0.84 1.64 2.33 0
return period.
-2.25 -0.02 0.84 1.67 2.40 0.033
T h e steps in using the LP3 distribution are:
-2.18 -0.04 0.83 1.70 2.47 0.067
1. transform the η original observations, X h to their -2.11 -0.06 0.82 1.72 2.55 0.100
logarithmic values, Y by the relationh -2.04 -0.07 0.81 1.75 2.62 0.136
-1.98 -0.09 0.80 1.77 2.70 0.166
(2.23)
-1.91 -0.10 0.79 1.79 2.77 0.197
2. compute the m e a n logarithm, Y; -1.85 -0.11 0.78 2.84
1.81 0.230
3. compute the standard deviation of the logarithm,
-1.79 -0.13 0.77 1.82 2.90 0.262
-1.74 -0.14 0.76 1.84 2.97 0.292
4. compute the coefficient of skewness C s from
-1.68 -0.15 0.75 1.85 3.03 0.324

η ΣΥ?
2
- 3ηΣΥ£Υ? + 2(ΣΥ>Ϋ -1.63 -0.16 0.73 1.86 3.09 0.351
(2.24) -1.58 -0.17 0.72 1.87 3.15 0.381
n(n - l)(n - 2)S Y

-1.54 -0.18 0.71 1.88 3.21 0.409


or -1.49 -0.19 0.69 1.88 3.26 0.436
-1.45 -0.20 0.68 1.89 3.31 0.462
ηΣ{Υ, ~ Y)
-1.41 -0.21 0.67 1.89 3.36 0.490
(n - 1)(« - 2)S$' -1.38 -0.22 0.65 1.89 3.40 0.517

5. compute -1.34 -0.22 0.64 1.89 3.44 0.544


-1.31 -0.23 0.63 1.89 3.48 0.570
Yf — Υ "Γ" S-yKj*, (2.25)
-1.28 -0.24 0.61 1.89 3.52 0.596
where K is from Table 2.7. This relationship is
T
-1.25 -0.24 0.60 1.89 3.55 0.620
identical to Eq. (2.22) except it is based on loga­ -1.22 -0.25 0.59 1.89 3.59 0.643
rithms; and -1.20 -0.25 0.58 1.88 3.62 0.667
calculate -1.17 -0.26 0.57 1.88 3.65 0.691
XT = antilog Y. T (2.26) -1.15 -0.26 0.56 1.88 3.67 0.713
-1.12 -0.26 0.55 1.87 3.70 0.734
T h e skew coefficient is sensitive to extreme flood
-1.10 -0.27 0.54 1.87 3.72 0.755
values and thus difficult to estimate from small samples
-1.08 -0.27 0.53 1.86 3.74 0.776
typically available for many hydrologic studies. Figure
-1.06 -0.27 0.52 1.86 3.76 0.796
2.10 presents a m a p of generalized skew coefficients for
the logs of peak flows taken from Bulletin 17B of the -1.04 -0.28 0.51 1.85 3.78 0.818
Interagency Committee. T h e station skew coefficient -1.01 -0.28 0.49 1.84 3.81 0.857
calculated from observed data and generalized skew -0.98 -0.29 0.47 1.83 3.84 0.895
coefficients can be combined to improve t h e overall -0.95 -0.29 0.46 1.81 3.87 0.930
estimate for the skew coefficient. U n d e r the assump­ -0.92 -0.29 0.44 1.80 3.89 0.966
tion that the generalized skew is unbiased and inde­ -0.90 -0.29 0.42 1.78 3.91 1.000
p e n d e n t of the station skew, the m e a n square error -0.84 -0.30 9.39 1.75 3.93 1.081
(MSE) of the weighted estimate is minimized by
-0.80 -0.30 0.37 1.71 3.95 1.155
weighting the station and generalized skew in inverse
proportion to their individual m e a n square errors ac­
cording to the equation (Tasker, 1978)

MSE (G) + MSE (G)


G C

G = (2.27)
w

MSE + MSE station skew [from Eq. (2.24)], G is the generalized


G G

skew (from Fig. 2.10), M S E is the m e a n square error


G

where G w is the weighted skew coefficient, G is the of the generalized skew, and M S E is the mean square
G
18 2. Hydrologic Frequency Analysis

Table 2.6 Frequency Factors for Extreme Value Type I Distribution

Sample
size
η 5 10 15 20 25 50 75 100 1000

15 0.967 1.703 2.117 2.410 2.632 3.321 3.721 4.005 6.265


20 0.919 1.625 2.023 2.302 2.517 3.179 3.563 3.836 6.006
25 0.888 1.575 1.963 2.235 2.444 3.088 3.463 3.729 5.842

30 0.866 1.541 1.922 2.188 2.393 3.026 3.393 3.653 5.727


35 0.851 1.516 1.891 2.152 2.354 2.979 3.341 3.598
40 0.838 1.495 1.866 2.126 2.326 2.943 3.301 3.554 5.576

50 0.820 1.466 1.831 2.086 2.283 2.889 3.241 3.491 5.478

60 0.807 1.446 1.806 2.059 2.253 2.852 3.200 3.446


70 0.797 1.430 1.788 2.038 2.230 2.824 3.169 3.413 5.359

80 0.788 1.417 1.773 2.020 2.212 2.802 3.145 3.387

90 0.782 1.409 1.762 2.007 2.198 2.785 3.125 3.367


100 0.779 1.401 1.752 1.998 2.187 2.770 3.109 3.349 5.261

oo 0.719 1.305 1.635 1.866 2.044 2.592 2.911 3.137 4.936

error of the station skew. M S E is taken as a constant


G examined and possibly greater weight given to the
0.302 when the generalized skew is estimated from station skew.
Fig. 2.10. M S E can be estimated from (Wallis et al.,
C

1974): Confidence Intervals


Any streamflow record is but a sample of all possible
MSE = antilog [^ - Β log, (iV/10)], (2.28) such records. How well the sample represents the
C 1 0 ()

population d e p e n d s on the sample size and the un­


derlying population probability distribution that is
where unknown. Both the form and p a r a m e t e r s of the under­
lying distribution must b e estimated. If a second sam­
ple of data were available, certainly different estimates
A = - 0 . 3 3 + 0.08IGI if |G| < 0.90 (2.29) would result for the p a r a m e t e r s of the distribution
even if the same distribution were selected. Different
= - 0 . 5 2 + 0.30IGI if |G| > 0.90 p a r a m e t e r estimates will obviously result in different
return period flow estimates. If many samples were
Β = 0.94 - 0.26|G| if |G| < 1.50 available, many estimates could be m a d e of the distri­
bution p a r a m e t e r s and consequently many estimates
= 0.55 if |G| > 1.50
could be m a d e of return period flows—say β ο · O 1 0
n e

Ν = record length. could then examine the probabilistic behavior of these


estimates of Q . m) T h e fraction of the (? o's that fell
10

between certain limits could be d e t e r m i n e d .


It is recommended that if the generalized and sta­ In actuality we have just o n e sample of data from
tion skews differ by more than 0.5, the data and flood which to make estimates of Q . Statistical procedures
T

producing characteristics of the watershed should be are available for estimating confidence intervals about
Frequency Determinations 19

Table 2.7 Frequency Factors for Pearson Type III Distribution


(Interagency Advisory Committee on Water Data, 1981)

Recurrence interval in years

Skew
coefficient
c
s 1.0101 2 5 10 25 50 100 200

3.0 -0.667 -0.396 0.420 1.180 2.278 3.152 4.051 4.970


2.8 -0.714 -0.384 0.460 1.210 2.275 3.114 3.973 4.847
2.6 -0.769 -0.368 0.499 1.238 2.267 3.071 3.889 4.718
2.4 -0.832 -0.351 0.537 1.262 2.256 3.023 3.800 4.584
2.2 -0.905 -0.330 0.574 1.284 2.240 2.970 3.705 4.444
2.0 -0.990 -0.307 0.609 1.302 2.219 2.912 3.605 4.298
1.8 -1.087 -0.282 0.643 1.318 2.193 2.848 3.499 4.147
1.6 -1.197 -0.254 0.675 1.329 2.163 2.780 3.388 3.990
1.4 -1.318 -0.225 0.705 1.337 2.128 2.706 3.271 3.828
1.2 -1.449 -0.195 0.732 1.340 2.087 2.626 3.149 3.661
1.0 -1.588 -0.164 0.758 1.340 2.043 2.542 3.022 3.489
0.8 -1.733 -0.132 0.780 1.336 1.993 2.453 2.891 3.312
0.6 -1.880 -0.099 0.800 1.328 1.939 2.359 2.755 3.132
0.4 -2.029 -0.066 0.816 1.317 1.880 2.261 2.615 2.949
0.2 -2.178 -0.033 0.830 1.301 1.818 2.159 2.472 2.763
0 -2.326 0 0.842 1.282 1.751 2.054 2.326 2.576
-0.2 -2.472 0.033 0.850 1.258 1.680 1.945 2.178 2.388
-0.4 -2.615 0.066 0.855 1.231 1.606 1.834 2.029 2.201
-0.6 -2.755 0.099 0.857 1.200 1.528 1.720 1.880 2.016
-0.8 -2.891 0.132 0.856 1.166 1.448 1.606 1.733 1.837
-1.0 -3.022 0.164 0.852 1.128 1.366 1.492 1.588 1.664
-1.2 -3.149 0.195 0.844 1.086 1.282 1.379 1.449 1.501
-1.4 -3.271 0.225 0.832 1.041 1.198 1.270 1.318 1.351
-1.6 -3.388 0.254 0.817 0.994 1.116 1.166 1.197 1.216
-1.8 -3.499 0.282 0.799 0.945 1.035 1.069 1.087 1.097
-2.0 -3.605 0.307 0.777 0.895 0.959 0.980 0.990 0.995
-2.2 -3.705 0.330 0.752 0.844 0.888 0.900 0.905 0.907
-2.4 -3.800 0.351 0.725 0.795 0.823 0.830 0.832 0.833
-2.6 -3.889 0.368 0.696 0.747 0.764 0.768 0.769 0.769
-2.8 -3.973 0.384 0.666 0.702 0.712 0.714 0.714 0.714
-3.0 -4.051 0.396 0.636 0.660 0.666 0.666 0.667 0.667

estimated values of Q that will give a m e a s u r e of


T tion of L and U d e p e n d on the underlying parent
T T

uncertainty associated with Q . Confidence limits give


T population. Bulletin 17B of the Interagency Committee
a probability that the confidence limits contain the true presents some approximate relationships for confi­
value for Q . A 9 5 % confidence limit indicates that
T dence intervals,
9 5 % of the time intervals so calculated will contain the
true estimate for Q . T LT — X + SK
X T L

Letting L and U be the lower and u p p e r confi­


T T
(2.31)
dence intervals LJ = X + S K
T X T u,

prob(L r < Q T < U) T = a, (2.30)


w h e r e X and S are the sample m e a n s and standard
X

where a is the degree of confidence. Exact determina­ deviations and Κ and Κ υ are the lower and
7\ L τ
20 2. Hydrologic Frequency Analysis

Figure 2.10 Map of generalized skew coefficient based on the logs of peak flows ( M S E ^ = 0.302).

upper confidence coefficients. If a distribution like the Example Analytical Frequency Analysis
LP3 distribution is used, X and S are based on the x
As an example of applying these t h r e e distributions,
logarithms of the data and L and U are the loga­
T T
again consider the d a t a of T a b l e 2.1. T h e m e a n and
rithms of the confidence limits. standard deviation of the original d a t a w e r e found to
Approximations for K Tand Κ L based on large τ α
be 1599 and 1006, respectively. T h e C is 0.629. Val­ v

samples and the noncentral /-distribution are ues of K for various values of Τ for the lognormal
T

K - JK 2
- ab distribution are selected from Table 2.5, and Eq. (2.22)
T

(2.32) gives t h e corresponding values of X . T h e s e results are


T

shown in Table 2.8.


K+
T TJK$ - ab Values of K T for various r e t u r n periods for the
extreme value type I distribution are selected from
Table 2.6, and X again comes from Eq. (2.22). T h e s e
where
T

results are shown in Table 2.8. _


In applying the log Pearson type III distribution, Y
a = l- (2.33a)
2(n - 1) based on the natural logarithms of the flow data is
found to be 7.237, S is 0.507, a n d C from Eq. (2.24)
Y s

b = Kl . (2.33b) is 0.87. K values are t h e n selected from Table 2.7, Y


T T

η calculated from Eq. (2.25), a n d X T from Eq. (2.26).


In these relationships, K is the frequency factor of
T
T h e results of these calculations are in T a b l e 2.8.
Eq. (2.22), Z a is the standard normal deviate with This example shows that t h e distribution that is
cumulative probability a (Appendix 2A), and η is the selected can have a substantial affect on t h e estimated
sample size. Confidence limits can be placed on fre­ flood magnitude for a given frequency. This is espe­
quency curves plotted on probability p a p e r by making cially a p p a r e n t for the longer r e t u r n periods. T o m a k e
calculations such as above for several values of T. this point even more emphatic, if a normal distribution
Frequency Determinations 21

Table 2.8 Flood Frequency Analysis for Return Period (yrs)


Middle Fork, Beargrass Creek, Cannons Lane, 50 20 10 5 2 1.11 1.02
Louisville, Kentucky 7000 —ι 1—Γ"— ι — ι — ι ι ι—r~ τ—Γ τ

Flood frequency
Return period (years): 5 25 100 5000
Middle F o r k B e a r g r a s s Creek
Lognormal distribution 4000

0.60 2.11 3.57


01 3000
χτ
2202 3718 5190 »·-
υ
2000

low
Extreme value type I
0.863 2.385 3.642
x T 2467 3998 5262

1000
Log Pearson type III 900
800
0.772 2.011 2.937
700
Y T 7.628 8.257 8.726 600
2 5 10 30 50 70 90 95 99
X 2056 3853 6161
T
Probability ( X )
Figure 2.11 Comparison of several probability distributions for the
Beargrass Creek data.

had been used, the 100-year estimate would have b e e n Confidence intervals on the LP3 can be calculated
3943 cfs, clearly an inferior estimate. T h e best fitting from Eqs. (2.30) through (2.33). T h e 9 5 % confidence
lines according to t h r e e of t h e distributions used are intervals are being calculated. Z = Z is found from
a 0 9 5

shown in Fig. 2.11. Appendix 2 A to be 1.645. From Eq. (2.33)


T h e plotting p a p e r used to construct Fig. 2.11 is
Zl 1.645 2

known as lognormal probability p a p e r . It consists of a a = 1


" or η =
" ο7ΐϊ
1
ϊΤ °· · = 9 5 5

logarithmic scale and a normal probability scale. Any


2(n - 1) 2 ( 3 1 - 1)
distribution may be plotted on lognormal probability
Table 2.9 contains the calculations of the confidence
paper, but only the lognormal distribution will plot as a
limits. T h e s e limits are plotted in Fig. 2.12. T h e inter­
straight line on this paper. O t h e r types of probability
pretation of this plot is that, on t h e basis of the
p a p e r are available, including normal and extreme
assumptions m a d e , o n e can be 9 5 % confident that the
value probability paper.
calculated confidence intervals will contain the true
F r o m Fig. 2.11 it is a p p a r e n t that it is difficult to flood frequency relationship for Beargrass Creek.
select which of the three distributions best describe As discussed later, o n e problem with these data is
these data. Considering the coefficient of skew, γ , t h e t h e possible nonstationarity of peak flows d u e to
EVI might be discarded since γ for the E V I is 1.139 changing watershed conditions. T h e difficulties experi­
while it is 2.13 for these data. F o r the LN, y and C v
enced with this set of d a t a d e m o n s t r a t e s why many
are related through Eq. (2.21). T h e estimated value of have been led to the r e c o m m e n d a t i o n that single short
y is 2.14, which agrees quite well with the requirement records are not reliable, a n d regional frequency analy­
of the LN. O n e discouraging factor concerning t h e L N sis p r o c e d u r e s as discussed u n d e r Case IV should be
is that y of the logarithms should be zero for the LN used for single sites.
distribution, while for these data it is 0.87. Apparently In summary, the Case I situation is that where a
the LN is not a precise approximation for the data. relatively long record of p e a k flows is available on the
T h e skewness cannot be used to m a k e decisions stream of interest at the point of interest. T h e method
concerning the LP3 since the LP3 is a t h r e e - p a r a m e t e r of analysis is to select a probability density function to
distribution that uses y to estimate these p a r a m e t e r s . describe the data, extract certain statistics from the
Looking at Fig. 2.11, it does a p p e a r that t h e L P 3 is a data, and use Eq. (2.22) along with appropriate fre­
better approximation to the d a t a at the u p p e r end of quency factors to estimate flows within the desired
the frequency curve. return period.
22 2. Hydrologic Frequency Analysis

Table 2.9 95% C o n f i d e n c e Intervals o n LP3 for Beargrass Creek

Τ­ b h

V Οι/

Ι.01 -1.64 -2.60 -2.19 -1.24 6.13 6.61 458 740


2 -0.15 -0.06 -0.46 0.15 7.00 7.31 1100 1498
5 0.77 0.51 0.46 1.15 7.47 7.82 1754 2495
10 1.34 1.71 0.98 1.83 7.73 8.16 2283 3510
25 2.02 3.99 1.57 2.66 8.04 8.58 3087 5344
50 2.48 6.06 1.97 3.23 8.24 8.87 3771 7131
100 2.94 8.56 2.36 3.80 8.43 9.16 4596 9535

From Table 2.7.


fl

''From Eq. (2.33).


'From Eq. (2.32).
''From Eq. (2.31).
' ( 2 = exp(L ).
L T

/ 0 = exp(i/ ).
U T

Return Period ( y r s )

95% confidence Interval

log P e a r e o n III

10 J
...I....I . . I . • . .1.........I . a . . 1 . . . . . . . . . I . . . . . . . . . I . . . . t . . ι I ι ι l n n l i m i i i I ι

.1 1 5 10 30 50 70 90 95 99 99.9

E x c e e d a n c e Probability ( X )
Figure 2.12 Confidence on the LP3 for Beargrass Creek.
Frequency Determinations 23

Some of the assumptions m a d e in the Case I situa­ eooo


tion are:

1. T h e data are sufficient in quantity and quality to


produce reliable estimates for t h e p a r a m e t e r s of the
probability distribution selected.
2. T h e flow characteristics of t h e stream have not b e e n
changing over time (stationary d a t a series).
3. T h e peak flow observations are statistically indepen­
dent from year to year.
4. T h e data are representative of the flow behavior
expected during the life of t h e project being con­
sidered.

Assumption 4 merely extends the stationarity as­


sumption to future flows. In watersheds with changing
land use, assumptions 2 and 4 are especially trouble­
some in that t h e changing land use alters streamflow 60
characteristics. Many times this nonhomogeneity in Year
streamflow data is very subtle and only becomes appar­
Figure 2.13 Changing flow condition resulting in nonhomogeneous
ent over a long period of time or when sudden and time series.
large-scale changes occur. For example, Fig. 2.13 is a
plot of the 31-year record for t h e Beargrass C r e e k
data. T h e bulk of t h e p e a k flows a r e in t h e range 750
to 2000 cfs. It appears as though the frequency of
ation of t h e 15 g e n e r a t e d values. This entire process
occurrence of peaks in excess of 2000 cfs is increasing
was r e p e a t e d 100 times, producing 100 estimates for
with time. T h e r a n d o m n a t u r e of the d a t a makes it
t h e 100-year flood. T h e probabilistic behavior of these
difficult to make firm statements in this regard.
100 estimates was t h e n analyzed.
In closing the discussion on t h e Case I flood fre­
T h e following tabulation shows the frequency of
quency analysis, a word of caution is offered concern­
occurrence of the estimate for the 100-year flood. T h e
ing the extrapolation of frequency data to estimate t h e
t r u e value is 5190 cfs. T h e m e a n estimated value was
magnitude of an event with a r e t u r n period much
4945 cfs. T h e following tabulation can b e used to get
greater than the period of record. In looking at
some idea of the possible errors involved in using a
Fig. 2.12, it appears that the extrapolation of t h e
short record to estimate a rare flood.
frequency lines from t h e 31-year record to estimate the
100-year or even 500-year event is not much of an
extrapolation. In the sense of t h e physical distance on Flow (cfs) No. of estimates
the probability paper, the extrapolation is not very
great; however, in the sense of extrapolating the d a t a 2500-3000 7
to 3 or possibly 15 times its original length, it is a very 3000-3500 8
significant extrapolation. T h e n a t u r e of r a n d o m d a t a 3500-4000 13
makes an extrapolation of this kind very speculative 4000^500 10
and produces estimates of low reliability or ones that 4500-5000 21
possess considerable uncertainty. This is a p p a r e n t from
5000-5500 15
the width of the confidence limits at high r e t u r n peri­
5500-6000 7
ods. H a a n (1977) gives a p r o c e d u r e for evaluating t h e
6000-6500 5
uncertainty that is present.
6500-7000 6
T o illustrate this point, a simulation was m a d e as­
7000-7500 3
suming a lognormal distribution with a m e a n of 1599
cfs and a standard deviation of 1006 cfs. T h e p r o c e d u r e 7500-8000 2

was to randomly select 15 observations from this log- 8000-8500 2


normal distribution and then use these 15 observations 14000-oo 1
to estimate the magnitude of the 100-year event. T h e
100-year event was estimated on the basis of the log- Total 100
normal distribution using the m e a n and standard devi-
24 Chapter 2. Hydrologic Frequency Analysis

Note that only 36 values or about one-third of the Table 2.10 Hypothetical Flood Frequency Data
estimates are within 10% (500 cfs) of the actual 100-year
flood. Nearly half (47 values) of the estimates are in Area (mile ) 2

error by more than 2 0 % , 1 5 % of the values are in error 0.7 2.3


by more than 4 0 % , and 1 estimate was actually high by Τ
(years) Q (cfs): a b
a factor of nearly 3. These numbers illustrate what can T

happen when a short record (15 years in this case) is 1041 0.923
2 750 2250
used to estimate a rare flood (100-year flood in this
5 950 2800 1318 0.909
case). This is the reason that extrapolation is specula­
10 1100 3300 1527 0.923
tive. In any real situation, the variability would be even
greater than shown here, because the true underlying
probability distribution would not be known. W e as­
sumed a lognormal distribution in this example and
based our calculations on this assumption as though it example, consider the d a t a in Table 2.10. H e r e , infor­
were the population distribution. It is understood that mation on the 2-, 5-, and 10-year floods is available at
one must use procedures like those set forth here to two locations. T h e exponents a and b of Eq. (2.34) can
make flood frequency estimates. T h e purpose of this be estimated for each r e t u r n period by substitution.
latter illustration is to shed some light on the uncer­ Consider t h e 10-year data:
tainty that is inherent in dealing with r a n d o m data.
Gio = aA b

C a s e II. L o n g R e c o r d n e a r Site (i) 3300 = a (2.3)* for the larger basin,


In some instances, flow data are available on the (ii) 1100 = a (0.7)* for the smaller basin.
stream of interest but at a location some distance
T h e ratio of (i) and (ii),
either upstream or downstream from the point of inter­
est. In this situation, there are several procedures that 3300 (2.3\ b

can be used to estimate the flood frequency relation­ 1100 =


\077/ '
ship at the point of interest. O n e m e t h o d is to perform
can be solved for b using logarithms resulting in b =
a flood frequency analysis on the available record as
0.923. Substituting this estimate for b into (i) results in
described in Case I, and then adjust the record to the
point of interest. 3300 = a ( 2 . 3 ) 0 , 9 2 3
,
T h e adjustment of the flow record from one location
which gives an estimate for a of 1527. Similar calcula­
on a stream to another can be done in a n u m b e r of
tions for the 2- and 5-year floods result in the estimates
ways. If the flow record includes the entire flood hy-
shown in Table 2.10. F r o m this, an average value of b
drograph, this hydrograph could be routed to the point
is 0.92. T h e coefficient a is seen to b e a function of
of interest making the proper adjustments for local
r e t u r n period T.
inflows along the routing reach. This m e t h o d is a very
T o estimate the 10-year flood at a point on the
good one, but requires more data than are generally
stream w h e r e t h e drainage area is 1.5 m i l e , Eq. (2.34) 2

available and is quite time consuming.


and the estimated 10-year coefficients are used:
A second method of flow adjustment is to correlate
flood peaks with drainage basin characteristics and Ql0 = 1527 ( 1 . 5 ) ' 0 9 2
= 2217 cfs.
then use this correlation to adjust the flow rate. T h e
If data at only o n e other location on the stream are
most common characteristic used in a situation like this
available, o n e can still estimate t h e flow at a second
is the basin area. Quite frequently the peak discharge
location if an assumption as to t h e coefficient b is
for a given frequency is related to the basin area by an
m a d e . For small differences in area, a reasonable as­
equation of the form
sumption for b is 1.0. For example, considering the
10-year d a t a for the 0.7 m i l e basin only and taking b
2

Q = aA ,
b
(2.34)
as 1.0 results in an estimate for a of
T

where Q is the Γ-year flood magnitude, A is the basin


T
a = 1 1 0 0 / 0 . 7 = 1571.
area, and a and b are constants. T h e coefficient b T h e 10-year flood can then b e estimated on the 1.5
generally ranges from 0.5 to 1.0. m i l e basin as
2

If data on a stream are available at two locations, the


coefficients can be estimated from those data. For Q l0 = 1571 ( 1 . 5 ) = 2357 cfs. 1
Frequency Determinations 25

W h e n these approaches are used, several precau­ estimating peak flow only. T h e type of model selected
tions must be exercised. First and most importantly, will d e p e n d on t h e quantity and quality of available
the basic flood-producing characteristics of all the data and the p u r p o s e of the analysis.
basins must be the same. T h e r e cannot be a mix of Regardless of t h e type of model selected, the model
drastically differing land uses unless some type of land- will have empirical coefficients associated with it that
use variable is included in the prediction equation for must b e estimated. In general, the n u m b e r of empirical
a and b. Returning to the above example, it has b e e n coefficients required is proportional to the complexity
assumed that all t h r e e watersheds are similar. F o r of the model, with the continuous simulation models
many watersheds, this is a severe limitation and gener­ having the most coefficients. T h e importance of actual
ally m e a n s that there cannot be a very large difference data on the stream of interest in estimating the empiri­
in the areas of watersheds considered or t h e land use cal coefficients of these models cannot be overempha­
on the watersheds. sized.
If the watershed characteristics are changing along O n c e the model coefficients have b e e n estimated, a
the steam, then calculations such as shown here can be long-term rainfall record can be processed through the
used as an aid in estimating a flow of a given r e t u r n model to p r o d u c e a long-term streamflow record. This
period, but will not give the final estimate. If t h e simulated, long-term streamflow record can then be
available data represent a mixture of land uses and t h e subjected to a frequency analysis as discussed under
point of interest is below only one land use within the Case I if necessary.
basin, the flow estimated by a straightforward applica­ In t h e event that a long-term rainfall record is not
tion of the techniques presented here will most likely available for the site u n d e r study, one can use records
have to be adjusted to reflect the fact that it is below a from the nearest raingage. Fortunately, in many parts
single land use while the coefficients a and b were of the country, rainfall records can be transferred a few
estimated on the basis of a mixed land-use basin. miles without introducing significant effects on the
estimated p e a k runoff rates from major events. T h e
long-term records from nearby raingages may not be
C a s e III. Short S t r e a m R e c o r d
usable for runoff p a r a m e t e r estimation since the
It is not uncommon to find that a streamflow record recorded rainfall may have been considerably different
at the point of interest may b e too short to use in a from what actually fell on the watershed. T h e record
flood frequency analysis. This may be the result of a can be used for simulation, however, because the long-
newly installed gage or a gaging program that was only term statistical properties may be the same as those of
recently changed so that much of the earlier portion of rains that actually fell on the watershed.
the streamflow record is no longer representative of In the absence of any applicable long-term rainfall
the basin. records, it may b e possible to use a stochastic rainfall
A short streamflow record can be a great aid in generation model to p r o d u c e a synthetic rainfall record
checking calculations and procedures used in flood to use in simulations with the runoff model.
frequency estimation. If a major drainage project is to A n o t h e r a p p r o a c h to using the information con­
be planned, the local governing body would b e wise to tained in a short record is to use the short record to
install a streamgage early in the feasibility part of the estimate a low-return period index flood. This might be
project planning process. In this way, by the time the the 2-year flood, for example. Knowing the magnitude
final design is m a d e , some streamflow data would be of this index flood, a regionalized relationship between
available. For relatively large drainage projects, this the ratio of the index flood and floods of a greater
short-term gaging approach is relatively inexpensive r e t u r n period can be used to estimate the magnitude of
and can easily pay for itself through t h e resulting less frequent events. T h e determination of the region­
improvement in the design of the drainage system. alized relationship is covered u n d e r the Case IV situa­
A short streamflow record is one of less t h a n 10 tion discussed below.
years in length. A record such as this will contain a A third option for using a short stream record is to
great deal of information, but will be insufficient for a correlate the annual peak flows from the short record
Case I frequency analysis. Presumably, a record of the with peak flows from a n o t h e r station in the vicinity
rainfall that produced the recorded runoff will be avail­ with a longer record. T h e longer record and the corre­
able or can be estimated from nearby raingages. T h e s e lation can then be used to extend the shorter record.
records on rainfall and runoff can now be used to Let Y and X represent peak flows on two streams.
estimate the empirical coefficients in an approximate Let n , be the n u m b e r of observations in common on Y
model. T h e model might be a continuous simulation a n d X. Let n b e the n u m b e r of additional observa­
2

model, an event or hydrograph model, or a model for tions on X not in common with Y. Assume the obser-
26 Chapter 2. Hydrologic Frequency Analysis

vations on each stream are i n d e p e n d e n t from year to I-PU)


98 .t9 .90 βΟ .70 .60 .50 .40 .30 .20 .10 .05 .02
year (no serial correlation on either stream). Consider
ι . 1 I I I I I I I ' ' I
ρY χas the correlation coefficient between Υ and X
for the λ? ι common observations. T h e correlation co­
β L B A S E D ON OATA FROM DURANT AND B L A C K W E L L (1959)
efficient ρ γ χ is estimated by r given by Y χ
I FOR STATIONS IN P A R T S OF A L B E R T A AND
I S A S K A T C H E W A N , C A N A D A . O i l - 1956 ( R E G I O N A '

i/E(*i-*) E(r,-y) 2 2

V /=i i-l

W e can estimate n values of Υ from the n values


2 2

of X for which Υ values are not available from

Γ ν V S \r( Χ X )
Υ = ' + y, (2.36)

where y , A", S , and 5 ^ are based on the n common


y ]

observations.
Let y, and Y represent the m e a n based on the
2 •02 III 2 5 10 20 50
original n observations and the n estimated observa­
x 2 R E T U R N PERIOD (yrs.)
tions, respectively. A new weighted m e a n for Υ based Figure 2.14 Regional flood frequency analysis.
on η ι + n observations is given by
2

T h e first step in this regional a p p r o a c h is to select


r- ' · (· )
2 37

several streamflow records from nearby locations that


are "hydrologically similar" to t h e basin of interest. At
For the j i added observations to improve the esti­
2 each of these locations, a Case I flood frequency analy­
mate for y , it is necessary that r be greater than Y x sis is m a d e . A n index flood is t h e n defined. This might
l/(n - 2) according to Matalas and Langbein (1962).
l be the 2-year flood. T h e ratio of t h e magnitude of t h e
If y and X contain significant serial correlation, the Γ-year flood to t h e index flood is c o m p u t e d for several
situation is somewhat more complex ( H a a n , 1977). values of Τ at each location. T h e ratio is t h e n plotted
versus Τ a n d a smooth curve drawn through the points
(Fig. 2.14).
C a s e IV. R e g i o n a l Analysis
T h e next step is to relate the index flood to water­
Often one finds that streamflow records are available shed characteristics. T h e area of the watershed is gen­
at several nearby locations, while none or a very short erally used along with o t h e r geomorphic, physical, and
record is available at the point of interest. Several meteorological factors. This step is generally d o n e
methods for using the information on the nearby sta­ through a regression analysis to p r o d u c e an equation of
tions to augment whatever is known at the site of the form
interest are available. T h e s e m e t h o d s generally fall
u n d e r the heading of "regional flood frequency anal­ <2 = 7 (2.38a)
ysis."
or
O n e widely used method of regional flood frequency
analysis is discussed by Dalrymple (1960). T h e m e t h o d Qj = a + bX x + cX 2 + · · · +qX H9 (2.38b)
consists of computing a base flood frequency relation­
ship in terms of the return period and the ratio of t h e where Q is t h e m a g n i t u d e of t h e index flood;
f

peak flow for a given return period to an index flood a,b,...,q are estimated coefficients; and X ,X , X 2

(usually the mean annual flood) for several streams in ..., X nare watershed and climatic factors. Regres­
the region. T h e median value of this ratio is then sions of this type are discussed in m o r e detail by
plotted versus the return period. Figure 2.14 is such a H a a n (1977).
plot for 18 stations in Alberta an Saskatchewan, T h e third step is to use Eq. (2.38) to estimate t h e
Canada, as reported by D u r a n t and Blackwell (1959). index flood for the location of interest. Alternatively, if
Special Considerations 27

a short record is available at the location of interest, sarily the same because of the effect of such factors as
the index flood can be estimated from that record. T h e a n t e c e d e n t soil water content and annual variation in
final step is to use the regional flood frequency curve land use. However, over the long run, the expected or
and the estimated Q to calculate Q for t h e desired
l T average r e t u r n period of t h e runoff will nearly equal
values of T. the r e t u r n period of t h e rainfall.
A variation of the above technique for regional flood Since the Case V situation is really a modeling effort
frequency analysis is to estimate Q for several values
T or requires the use of hydrologic techniques not gener­
of Τ at each gaged location as explained above, a n d ally thought of as being frequency analysis, its treat­
then relate Q to watershed factors and climatic d a t a
T m e n t is deferred t o C h a p t e r 3, which provides a d e ­
by regression to produce an equation like (2.38) with tailed t r e a t m e n t of p e a k flow estimation. C h a p t e r 13
Q replaced by Q . A separate equation is n e e d e d for
l T contains a discussion of hydrologic models of various
each value of T. These equations can then be used to types.
estimate the desired value of Q at the study location.
T

O n e disadvantage of this a p p r o a c h is t h e possibility of


not retaining the p r o p e r relation a m o n g t h e Q 's for
T SPECIAL CONSIDERATIONS
different values of T. T h a t is, o n e could conceivably
estimate Q 25 as being less t h a n Q . H a a n (1977)
lQ Historic D a t a
discussed the use of multivariate multiple regression to
Occasionally flood information outside of the sys­
overcome this difficulty.
tematic flow record is available from historical sources
such as newspaper reports or earlier flood investiga­
C a s e V. N o Flow R e c o r d s tions. Such d a t a contain valuable information that
should not b e ignored in a frequency analysis. Bulletin
W h e n there are no streamflow records available on
17B of the Interagency Advisory Committee on W a t e r
the stream of interest or on nearby streams or w h e n
D a t a (1981) discusses this topic in detail. Basically,
available records are from basins whose characteristics
what o n e does is to c o m p u t e t h e plotting position of
are considerably different from those on the basin of
t h e historical observations on t h e basis of the historical
interest, one is forced to use some type of empirical
record length. Likewise t h e plotting position of the
procedure for estimating the magnitude of runoff events
systematic d a t a is c o m p u t e d on t h e basis of the historic
of the desired frequency. W e are now out of the realm
record length except t h e rank used in the calculation is
of frequency analysis in the sense of determining t h e
adjusted by a factor W d e p e n d i n g on the historic
frequency of occurrence of events based on a proba­
record length, H, the n u m b e r of historic flows, Z , and
bilistic analysis of data.
t h e length of t h e systematic record, N. T h e s e are
For this situation, a hydrologic model of some type
related by
must be employed. T h e model may range in complexity
from the Rational Equation to a complete, continuous Η — Ζ
simulation model. T h e type of model selected will W=
—£T- (· )
2 39

d e p e n d on the data available for model fitting, the


user's familiarity with various models, the purpose of T h e adjusted rank for the systematic data is
the modeling effort, the time and money available for
completing the modeling effort, and the importance of m a = Wm - (W - 1 ) ( Z + 0 . 5 ) , (2.40)
flow estimates.
If a continuous simulation model is selected and with m being the unadjusted rank of the total record
several years of streamflow are generated, these gener­ (systematic plus historic).
ated data can be subjected to a flood frequency analy­ T h u s , if 20 years of systematic data and two historic
sis. If an event-based model (such as a unit hydrograph observations are available from a 50-year period p r e ­
approach) is selected, the most severe rainfall events ceding the systematic record, the plotting position for
each year can be analyzed, with these data subse­ the two largest values would be 1 / 7 1 = 0.014 and
quently subjected to a flood frequency analysis. 2 / 7 1 = 0.028. T h e weighting factor would be
If an approach like the Rational Equation is used,
W= (70 - 2 ) / 2 0 = 3.40.
one is assuming that the frequency of the estimated
flow peak is the same as the frequency of the rainfall T h e remaining plotting positions would be calculated
used in the equation. This is not a bad assumption over from t h e adjusted rank given by
the long run. For individual events, the r e t u r n period
of the rainfall and the resulting runoff are not neces­ ma = 3.40m - (3.40 - 1 ) ( 2 + 0.5) = 3.40m - 6.0.
28 Chapter 2. Hydrologic Frequency Analysis

T h e adjusted rank is then used in the plotting position m e t h o d is to use the t h e o r e m of total probability
relationship [Eq. (2.8)]. T h u s for m = 3 (the largest
prob(X > x) = p r o b ( X > x\X = 0 ) p r o b ( * = 0)
systematic flow observation), the plotting position
would be [3.40(3) - 6 ] / 7 1 or 0.0592 and for m = 22 + p r o b ( X > x\X Φ 0 ) p r o b ( JT Φ 0 ) .
(the smallest value) the plotting position would be Since probiX > x\X = 0) is zero, t h e relationship re­
[3.40(22) - 6 ] / 7 1 or 0.9690. This compares to plotting duces to
positions of 1/21 or 0.0476 and 2 0 / 2 1 or 0.9523,
respectively, if the historic data had been ignored. If p r o b ( Z > x) = p r o b ( Χ Φ 0 ) p r o b ( X > x\X Φ 0).
the historic data had simply been used to augment the In this relationship probiA" Φ 0) would be estimated by
systematic record without using the weighting factor, the fraction of n o n z e r o values and t h e p r o b i A ' >
the plotting positions for these two events would have x\X Φ 0) would be estimated by a standard analysis of
been 3 / 2 3 or 0.1304 and 2 2 / 2 3 or 0.9565, respectively. the n o n z e r o values with the sample size taken as equal
Clearly a plotting position of 0.1304 assigns too high a to the n u m b e r of n o n z e r o values. This relation can be
probability of occurrence to the largest systematic value. written as a function of cumulative distributions,
It is also apparent that the weighting p r o c e d u r e adjusts
the plotting position toward a m o r e frequent occur­ \ - P {x) x = k[\ - P*{x))
rence for the largest systematic value, thus taking into or
account the fact that two flows of magnitude greater
P (x) = l - k + kP*(x), (2.44)
than that of the largest systematic flow occurred. x

Bulletin 17B also suggests the flow statistics be com­ w h e r e P (x) xis the cumulative probability distribution
puted by weighting the contribution of the systematic of all X (pvob(X <x\X> 0)), k is the probability X is
record to the various statistics by the factor W. Thus, not zero, a n d P (x) is the cdf of t h e n o n z e r o values of
x

the adjusted m e a n is X (i.e., probiA" < x\X Φ 0). This type of mixed distri­
bution with a finite probability that X = 0 and a con­
tinuous distribution of probability for X > 0 has b e e n
Xa = jj £
, (2.41)
shown by Jennings a n d Benson (1969) to b e applicable
where X represents the systematic record and X the z
for flow frequencies with zeros present.
historic data. Similarly the variance and skew can be Equation (2.44) can be used to estimate the magni­
determined from t u d e of an event with a r e t u r n period Τ by solving first
for P (x) x and then using t h e inverse transformation of
— \2
r 2 WX(X-X ) a +l{X -X )
z a Ρχ(χ) to obtain the value of X. F o r example the
(2.42) 10-year event (P (x) = 0.90) with k = 0.95 is found to
Η - 1 x

be the value of X satisfying


and
P (x)
- 1 + k 0.90 - 1 + 0.95
Η x

Px(x) = , = — = 0.89.
c =
Sa
(Η-ί)(Η-2) x y
' k 0.95
ζ-. ,3 T o d e t e r m i n e the corresponding value for X, this
WZ(X-X ) a + X(X -X )
z a
equation must be solved for X based on a probability
Χ (2.43)
plot or the assumed probability distribution. N o t e that
it is possible to g e n e r a t e negative estimates for P (x) x

If the LP3 distribution is being used, the X's and from Eq. (2.44). For example, if k = 0.50 and P (x) = x

Xs
z would be based on logarithms. 0.05, the estimate for P (x) is x

0.05 - 1 + 0.50
T r e a t m e n t o f Zeros

T h e following is taken from H a a n (1977). Most hy­


ΡΗχ) = 53 -0.9.
This merely m e a n s that the value of X corresponding
drologic variables are b o u n d e d on the left by zero.
A zero in a set of data that is being logarithmically to P (x) is zero.
x

transformed requires special handling. O n e solution is


Outliers
to add a small constant to all of the observations.
A n o t h e r method is to analyze the nonzero values and Occasionally, a systematic record of p e a k flows may
then adjust the relation to the full period of record. contain one or m o r e observations that deviate greatly,
This method biases the results as the zeros are essen­ either high or low, from the rest of the data. For
tially ignored. A third and theoretically more sound example, it is entirely possible that a 100-year event is
Discussion of Flood Frequency Determinations 29

contained in 10 years of record. If this is the case, then generally extrapolated by the analyst to frequen­
assigning a normal plotting position of 1 / 1 1 to this cies of occurrence well beyond that contained in the
value would not be reflective of its true r e t u r n period. original set of n u m b e r s . From these extrapolations, the
Bulletin 17B suggests that outliers can be identified flows having return periods of 25, 50, 100, or even 500
from years a r e d e t e r m i n e d . T h e straightforward application
of hydrologic frequency analysis as generally employed
X H - X + KS
n x
uses no or very little hydrologic knowledge. In actual­
(2.45)
ity, rare flows are d e t e r m i n e d by the hydrologic condi­
tions that exist at t h e time of these flows and not by the
where X H and are threshold values for high a n d statistical behavior of a sample of maximum peak flows
low outliers and K n can b e approximated from that may have occurred some time in the past. Resolv­
K * 1.055 + 0 . 9 8 1 1 o g n , (2.46) ing t h e a p p a r e n t conflict between these statements is
n 10

what separates the hydrologist from the statistician.


where η is the n u m b e r of observations. Statistics are descriptive of a set of observed data.
If a peak in the record exceeds X Hand historical Statistics do not define a cause and effect relationship
information of the type discussed earlier is available or a physical relationship. Any conclusion drawn on
regarding that peak, it should b e removed from the the basis of a statistical frequency analysis assumes that
systematic record and treated as historical observation the sample of data on h a n d is representative of a wider
as discussed u n d e r Historical D a t a . If historical infor­ range of data known as the population. In hydrologic
mation on the flow is not available, it should be re­ terms, this m e a n s that if we have a sample of 15 years
tained as a part of the systematic data. or so of observed annual maximum peak flows and use
If a flow is less than X , L that value should be these d a t a to estimate the 100-year flood, we are
deleted from the record and conditional probability assuming that t h e hydrologic behavior of the basin
procedures as explained in the section T r e a t m e n t of during the 100-year flood is somehow imprinted in the
Zeros should be employed. 15 years of observed data and that the statistical tech­
M o r e detail on the t r e a t m e n t of outliers is contained nique being used can uncover this imprint and use it.
in Bulletin 17B. Equation (2.46) is an approximation T o d e t e r m i n e if this is truly the case, the hydrology of
for a table in Bulletin 17B. For 10 < η < 149, t h e the basin must be examined. Some of the questions
maximum error in the equation is less than 2 . 5 % and that must b e answered are:
averages about 1%.
a. Is the type of storm that is likely to produce the
100-year flow r e p r e s e n t e d in the observed sample?
DISCUSSION OF FLOOD FREQUENCY b. Is the contributing area of the basin the same for
DETERMINATIONS extreme floods as it is for small ones?
c. A r e there p o n d s and reservoirs that may discharge
T h e foundation of any frequency analysis is the at high rates during rare floods and not during
procedure described u n d e r Case I w h e r e a particular smaller flows? W h a t is the possibility of a dam or
probability distribution is selected for describing a set levee breach and what would be the resulting flow?
of data. T h e p a r a m e t e r s of this distribution are esti­ d. A r e t h e channel flow and storage characteristics the
mated, and the magnitude of events for various r e t u r n same for extreme flows as they are for smaller
periods are calculated. M e t h o d s for plotting the ob­ flows?
served data on probability p a p e r and for constructing e. A r e land-use and soil characteristics such that flows
the best fitting line according to the selected distribu­ from rare storms may relate to precipitation in a
tion have also been discussed. m a n n e r different from m o r e common storms?
At this point, it should be clear that t h e r e is nothing f. A r e there seasonal effects such that rare floods are
inherently hydrologic about frequency analysis proce­ m o r e likely to occur in a different season than the
dures. They are simply statistical techniques that oper­ m o r e common floods?
ate on numbers. T h e fact that the n u m b e r s being used g. Is t h e r a r e flood r e p r e s e n t e d in the sample of data?
are peak flows is of no concern to the technique. It If so, how is it t r e a t e d ? Is it assigned a return period
should b e of great concern to t h e analyst, however. of 15 years where in fact its return period may be
Statistical frequency analysis simply a t t e m p t s to ex­ much greater than that?
tract information about the probabilistic behavior of a h. A r e there changes going on within the basin that
set of numbers from the numbers themselves. In hydro- may cause change in the hydrologic response of the
logic frequency analysis, this probabilistic behavior is basin to rainstorms?
30 Chapter 2. Hydrologic Frequency Analysis

These last few paragraphs paint a discouraging pic­ Table 2.11 Peak Discharge Data: Rose Creek
ture for flood frequency analysis. T h a t need not be the at Nebo, Kentucky
case as long as one does not discard hydrologic knowl­
Discharge Discharge
edge in the process. Often the questions posed can be
Year (cfs) Year (cfs)
answered in such a way as to make the statistical
analysis valid. A t other times, when problems with the 1952 624 1962 730
statistical procedures are recognized, adjustments can 1953 722 1963 680
be m a d e in the resulting flow estimates to m o r e accu­
1954 358 1964 800
rately reflect the hydrology of the situation.
1955 500 1965 622
Hydrologic frequency analysis should be used as an
1956 884 1966 571
aid in estimating rare floods. Sometimes the estimates
1957 689 1967 350
m a d e on the basis of the statistical frequency analysis
1958 1230 1968 920
can be taken as the final estimate. Sometimes the
statistical estimate may need to be adjusted to b e t t e r 1959 1000 1969 1240

reflect the hydrology of the situation. 1960 900 1970 818


It should be kept in mind that other hydrologic 1961 860
estimation techniques suffer from some of the same
difficulties as do the statistical techniques. For example
if a hydrologic model is being employed, the p a r a m e ­
ters of the model must be estimated in some way. This
Solution:
is generally d o n e on the basis of observed data from
Calculation of sums and means
the basin in question, from observed d a t a from a
similar basin, or from so-called physical relationships Σβ ρ = 14498 Σ(1η β ) ρ
2
- 825.92
such as Manning's equation, infiltration p a r a m e t e r s ,
etc., and a set of accompanying tables. Regardless of XQl = 1.2131 Χ ΙΟ 7
Σ(1η β ) ρ
3
= 5466.13
how the p a r a m e t e r s are estimated, the same type of
Σ 1η β = 125.11 β = lQ /n = 763 cfs
questions regarding these estimates and the n a t u r e ρ ρ p

of the hydrologic model itself must be answered as 1200 - β ρ \


outlined above for frequency analysis estimates. W e a. p r o b ( f i > 1200 cfs) = prob Ζ >
p

cannot substitute mathematical and empirical relation­


ships for hydrologic knowledge any more than we can S g g - nQ 2
p

substitute statistics for hydrologic knowledge.


η - 1
Based on these last few paragraphs, o n e might con­
clude that the magnitude of rare events should not be 1.2131 Χ 10 - 19(763) 7 2

estimated since the estimates may be so uncertain. = 244


18
Generally, however, this is not one of the options
1200 - 763
available. A n estimate must be m a d e . Hydrology must p r o b ( g > 1200) = p r o b | z >
p

not be ignored in making this estimate. Statistical, 244


modeling, a n d / o r empirical flow estimates should be = prob(Z > 1.79).
m a d e and then adjusted, if required, to reflect t h e
From the table of the standard normal distribution in Ap­
hydrologic situation. This is not to say a factor of safety
pendix 2A the prob(Z > 1.79) is found to be 0.037. There­
is to be applied. Adjustments should be based on fore, the desired probability is
hydrology, not rules of thumb.
p r o b ( e > 1200) = 0.037
p

b. p r o b ( O < 1200) = 1 - p r o b ( G > 1200)


p p

Example Problem 2 . 1 . Flow probabilities


= 1 - 0.037 = 0.963
If we assume the data on peak discharge for Rose Creek c. Τ = 1/p = l / p r o b ( G > 1200) p

(Table 2.11) can be described by a normal distribution, in any = 1/0.037 = 27 years


year what is the:
d. Τ = 20; therefore ρ = l/T = 0.05
a. Prob(G > 1200 cfs),
p

b. prob((? < 1200 cfs), p r o b ( G p > e o ) =0.05 2


p

c. return period for a flow of 1200 cfs, p r o b ( G < C ) =0.95


p 2 0

d. magnitude of a 20-year flood, and


e. prob(800 < ( ? < 1000)?
p
prob((2 P <Q) 20 = prob(Z < Z ) - 0.95 2 0
Discussion of Flood Frequency Determinations 31

from the standard normal distribution (See Appendix 2A) A plot of Q versus ρ on lognormal paper is shown in Fig.
2.15.
b.
Z 2 0 = 1.645
14498
= 763
Q20 ~ Go 19
= 1.645.
Σ 6 , ~ nQl
2

η - 1
From part a, S Q = 244 and Q = 763. p

1.2131 Χ 10 - 19(763) 7 2

= 244
Q20 = Z S + Q = 1.645(244) + 763 = 1164 cfs. 18
2 0 C p

244
e. prob(800 < Q < 1000)
Γ = — = = 0.320.
p
763

800 -Q p 1000 -Qp


To draw the best fitting lines, calculate several points from
= prob| < Ζ <
Q =Q (\
T P + C K ).
y T

Lognormal (K from Table 2.5):


= prob(0.15 < Ζ < 0.97)
T

Return period
= prob(Z < 0.97) - prob(Z < 0.15)
Τ 2 5 20
= 0.8334 - 0.5596 = 0.274.
*r -0.15 0.75 1.85

Qt 726 946 1215

Example Problem 2.2. Frequency analysis Plot Q versus l/T. Since lognormal probability paper is
T

used, the resulting plot should be a straight line.


Using the Rose Creek data in Table 2.11, Extreme value type I (n = 20 was used in determining K T

from Table 2.6):


a. Plot the data on lognormal probability paper.
b. Draw in the best fitting line according to the Τ 5 10 15 20 25 50
(i) lognormal distribution
(ii) extreme value type I distribution Κγ 0.919 1.625 2.023 2.302 2.517 3.179
(iii) log Pearson type III distribution. Qt 987 1159 1256 1325 1378 1539
c. Estimate the 100-year flood using the
(i) lognormal distribution Plot Q versus \/T. The plot is not necessarily a straight line
T

(ii) extreme value type I distribution since lognormal paper, not EVI paper, is being used. Note
(iii) log Pearson type III distribution. that the LN and EVI produce results that are close to each
Solution: other. The EVI and the LN are nearly identical if the
a. From Eq. (2.8), the plotting position is given by coefficient of skew of the data is 1.139. For these data,
m/(n + 1), where m is rank and η = 19, the number of data = n ^Ql
2
- 3*2e 2Q p p + 2(Sg ) p
3

values ( ρ = m/20). 5
n(n - \)(n - 2)S* Q

(19) 1.0954 Χ 1 0
2 1 0
- 3(19)(14498)1.2131 Χ 1 0 + 2 ( 1 4 4 9 8 ) 7 3

m Q Ρ m Q Ρ
19(18)(17)(244) 3

1 1240 0.05 11 722 0.55 = 0.29.

2 1230 0.10 12 689 0.60 On the basis of the coefficient of skew, the two distributions
are expected to differ. This difference will start to show up in
3 1000 0.15 13 680 0.65
the estimation of ( ? o - The fact that the coefficient of skew
10

4 920 0.20 14 624 0.70


for the extreme value type I is a constant 1.139 suggests the
5 900 0.25 15 622 0.75 EVI may not be a good descriptor of this data.
6 884 0.30 16 571 0.80 Similarly for the LN, γ and C are related by v

7 860 0.35 17 500 0.85 y = 3C + Cl.


K

8 818 0.40 18 358 0.90


Substituting C = 0.32 into this relationship results in an
v

9 800 0.45 19 350 0.95 estimate for γ of 0.99 compared to the calculated 0.29. This
10 730 0.50 deviates from the requirements for the LN to some extent
but not as severely as for the EVI.
Chapter 2. Hydrologic Frequency Analysis

Τ (years)
1.11
K T is determined from Table 2.7.
50 20 10 5 2
3000 ' I 1

Τ 2 5 10 25 50

0.008 0.844 1.276 1.730 2.027


yT 6.587 6.873 7.021 7.186 7.278

Qt 726 966 1120 1310 1448

Q is plotted against l/T. The LP3 results and the LN


T

results are very close to each other. Note that the skewness
of the logarithms is - 0.058 or close to zero. The skewness of
the logarithms for a LN distribution is zero since the loga­
rithms of the values are normally distributed and the normal
ο ο distribution has zero skew.
300 c. Q is estimated in the same manner as the points on
m

the straight line were calculated. The results are:

200 * ' 1 1
'"' ·'•••·' 90

95
11

98
LN EVI LP3
Probability ( X )
1502 1699 1584
Figure 2.15 Frequency plot for Problem 2.2. @100

Note that the LN and the EVI differ by nearly 200 cfs, while
the LN and LP3 differ by only 83 cfs.
Log Pearson type III: A good choice for the best estimate of Q would be lO0

To draw the best fitting straight line, calculate around 1550 cfs for this stream. If designing a facility on the
basis of a 100-year return period and failure of the facility to
properly handle the design flow were serious (high economic
Q = exp(y ),
T r
loss or loss of life), the more conservative estimate of around
1700 cfs could be used.
where

Y = XK<T» + S K
Example Problem 2.3. Regional analysis
T lnQ T

Based on results in Example Problem 2.1, It is desired to estimate the 50-year flood on a 4 mile 2

watershed. No streamflow data on the watershed are avail­


able. The investigator has collected flow data on six nearby
, Σΐηβη 125.11 streams. An analysis of the flood frequency relationship is
IHT£ = = —rz- = 6.585 shown in the following table. Use these data and a regional
μ
η 19 flood frequency approach to estimate the 50-year flows on
Σ(1ηβ ) -η(ϊτΓ^) ρ
2 2
the 4 mile watershed.
2

η - 1
Stream

825.92 - 19(6.585) 2
A Β C D Ε F
= 0.34.
18
050 52,000 26,000 6,400 1,800 8,300 4,000
34,000 15,000 4,600 1,400 6,400 2,950
The C of the logarithms is 0.34/6.585 or 0.052 indicating
v

Co 24,000 9,200 3,400 1,150 5,100 2,200


that the logarithms are nearly symmetrically distributed. K T

Qs 15,000 5,100 2,400 860 3,800 1,580


is a function of C of In Q . Let Y = In Q :
s p p

6,300 1,640 1,200 520 2,600 810

Qui 1,700 295 420 240 960 300


n XY 2 3
- 3nXYlY 2
+ 2(XYf
Cc = ; — —τ = -0.058. Area(mile ) 2
20 5 3 1 7 2
n(n - l)(n - 2)S Y
Discussion of Flood Frequency Determinations 33

Solution: 20 I
First develop a plot of Q /Q T X
versus T. Use Qi
for Q v

Regional Flood Curve Ο

Watershed
σ
A Β C D Ε F σ
10 -
ο Ο
8.25 15.85 5.33 3.46 3.19 4.94 «·* Ο
Q3JQ2 σ
5.40 9.15 3.83 2.69 2.46 3.64 κ
Q7JQ1
Ο Ο
QxJQi 3.81 5.61 2.83 2.21 1.96 2.72
Ο
Qs/Q 2.38 3.11 2.00 1.65 1.46 1.95 θ
2

Q IQ
2 2
1.00 1.00 1.00 1.00 1.00 1.00
0.27 0.18 0.35 0.46 0.37 0.37
fl1.11/fl2 1.11 5 10 50
Return Period (yre)
Figure 2.16 Regional flood frequency curve for Problem 2.3.
Next plot Q /Q versus Τ and draw a smooth line connect­
T 2

ing the median of the data (Fig. 2.16). It appears as though


watershed Β does not fit the other watersheds, so it is not 8000
used in computing this median. Plot Q versus watershed 2

area (Fig. 2.17). Use this relationship to determine the index


flood (Q ). From this information for a 4 mile watershed,
2
2

Q is estimated as 1500 cfs. The ratio Q /Q


2 is 5. There­ 50 2

fore, 3000

β 5 0 = X Q = 5 X 1500 = 7500.
2

1000
The estimated 50-year flow is 7500 cfs.

Example Problem 2 . 4 . Weighted skew coefficient 400

Area (ml*)
Data from a stream near St. Louis, Missouri, yielded the
following statistics based on the natural logarithms of the Figure 2.17 Q —Area
2 relationship for Problem 2.3.
flow. Note that Y is equal to the log of the peak flow
( y - l n β ). ρ From Eq. (2.28) and (2.29),
M S E = antilog [,4 - Β l o g ( N / 1 0 ) ]
C 1() 1()

Ϋ= 7.313, C = -0.108
A = - 0 . 3 3 + 0.08IGI = - 0 . 3 3 + 0.08(0.108) = -0.321
s

SY = 0.250, Ν = 25. Β = 0.94 - 0.26IGI = 0.94 - 0.26(0.108) = 0.912


M S E = a n t i l o g [ - 0 . 3 2 1 - 0.912log (25/10)] = 0.207
G 10 10
Estimate the 100-year peak flow based on the log Pearson
type III distribution using the weighting procedure on the From eq. (2.27)
generalized and data derived skews. MSE^(G) + MSE (G) G

Solution: G = w
MSE a + MSE G
A weighted skew coefficient based on Eq. (2.27) is calcu­
lated. From Fig. 2.10, 0.302(-0.108) + 0.207(-0.4)
= -0.227.
0.302 + 0.207
G = -0.4. From Table 2.7 using a skew of -0.227, K is inter­ m)

polated as
The mean square error for Fig. 2.10 is 2.178 - K - 0 . 2 + 0.227 h

2.178 - 2.029 - 0 . 2 + 0.4


M S E = 0.302.
G

^100
=
2.158
The data derived estimate for the skew is y l 0 0 = Ϋ + SK Y T = 7.313 + 0.25(2.158) = 7.853
£?ioo = e ' 7 8 5 3
= 2573 cfs.
G = -0.108.
34 Chapter 2. Hydrologic Frequency Analysis

type III distribution for t h e data of Problem 2.1 assum­


ing t h e data a r e annual peak flows.
Example Problem 2 . 5 . Historical flow data
(2.4) T h e data given below a r e annual peak flows for
Assume that in 1923 a peak flow of 1800 cfs was estimated the period 1936 through 1971 for t h e Chikaskia River
for Rose Creek at Nebo, Kentucky. Estimate the 100-year n e a r Blackwell, O k l a h o m a .
flow according to the log Pearson type III distribution. a. Plot t h e d a t a o n lognormal probability paper.
Solution: b. Plot t h e estimated flood frequency curve for
Equations (2.39) through (2.43) are used to adjust the return periods of 2 through 100 years based on t h e
observed flow data for the historical observation. Η = lognormal, extreme value type I and log Pearson
1961-1923 = 48 yrs. type III distributions o n t h e same graph as part a.
Η 48 - 1 c. Which distribution gives t h e best estimate of
W= = 2.47. the flood frequency relationship for this location?
Ν 19
d. W h a t is t h e magnitude of t h e 100-year peak
Based on data from Example Problem 2.1 and Eq. (2.41), flow for this location?
WZY + XY 2.47 X 125.11 + In 1800
Z

Yn = ~ = = 6.594.
Η 48
Flow (cfs)
Equation (2.42) gives
10,800 35,800 3,120 15,500
Wl(Y-Y ) a
2
+ l(Y -Y )z a
2

12,900 6,200 39,300 8,460


Η- 1 26,800 31,000 14,600 64,000

Σ( Y - Y ) 2
= X(ln Q - 6.594) = 2.145 2
8,340 23,100 55,000 1,200
a p

6,040 13,300 9,050 5,650


Σ(Υ-Υ,) 3
= Σ(1η β - 6.594) =3
-0.486 20,000 31,000
ρ
8,820 8,070

„ 2.47(2.145) + (In 1800 - 6.594V 85,000 53,000 48,000 26,000


S =2
= 0.132 12,200 8,130 36,500 45,200
47
S = 0.363 82,000 7,280 27,000 5,350
n

Η Wl(Y-Y f a + 1(Y -Y f
Z a

(Η- !)(//-2)
(2.5) Should t h e value of 1200 cfs in t h e data of
48 2.47( -0.486) + (In 1800 - 6.594) 3
Problem 2.4 b e considered an outlier?
(47)(46) (0.363) 3 (2.6) It was established through interviews of local
residents that in 1923 a flood with an estimated p e a k
= -0.217. flow of 100,000 cfs occurred on t h e stream of Problem
From Eqs. (2.25) and (2.26) and Table 2.7 with K « 2.165 2.4. Based on t h e log Pearson type III distribution,
m)

estimate t h e magnitude of t h e 100-year p e a k flow in­


Y\m = a + S K
y
a m = 6.594 + 0.363(2.165) = 7.380 corporating this additional information in t h e estimate.
G,oo = e 73 8
° = 1603 cfs. (2.7) T h e following data represent t h e gage height
and annual p e a k discharge for t h e streamgaging station
on t h e Arkansas River at Ralston, Oklahoma. T h e s e
gage heights a r e in feet, a n d t h e discharges a r e in cfs.
Problems T h e period of record is 1923 through 1971. Based on
(2.1) Using a hand calculator, d e t e r m i n e t h e mean, this data:
standard deviation, coefficient of variation, and skew­ a. plot discharge versus stage,
ness of t h e following data: b. develop an equation that fits t h e plot of
part a,
20, 45, 13, 80, 12, 30, 18, 22, 17, 32, 22. c. plot t h e discharge data on lognormal probabil­
ity paper,
(2.2) Using a calculator and tables in the text, esti­ d. plot a log Pearson type III distribution for the
mate t h e 20-year peak flow based on the lognormal discharge data on t h e plot of part c.
and extreme value type I distribution for the data of e. A local resident claims that in t h e early 1900's a
Problem 1 assuming t h e data are annual peak flows. flood that would correspond to a stage of 30 ft
(2.3) Using a calculator a n d tables in t h e text, esti­ occurred at this location. Estimate t h e return period
mate t h e 25-year peak flow based on t h e log Pearson of t h e flow associated with this r e p o r t e d event.
Discussion of Flood Frequency Determinations 35

River, respectively. Develop a regional flood frequency


Stage Discharge Stage Discharge
relationship relating Q /Q to drainage area based on
T 2

23.0 200,000 18.50 114,000 these three streams. Use Τ ranging u p to 100 years.
11.8 42,000 14.93 70,200
6.4 11,300 15.30 70,700
17,800 4,280 3,880 2,520
10.4 32,400 17.60 92,800
17,500 4,790 6,890 4,100
18.7 108,000 21.45 135,000
4,900 2,610 30,200 1,780
15.0 73,000 10.48 25,800
9,390 2,810 15,400 3,040
15.3 76,500 8.80 17,500
4,890 8,720 8,880 4,440
12.1 47,800 9.07 18,700
4,790 5,430 3,620 3,490
9.5 28,200 12.71 36,300
3,830 12,200 2,250 3,050
10.6 33,700 14.64 49,200
9.3 25,700 21.41 120,000
6.4 56,800
(2.12) T h e following data are for the Salt Fork
11,700 14.86
Arkansas River n e a r C h e r o k e e , Oklahoma, for the
16.0 77,800 14.65 54,800
period of 1941 to 1950. T h e drainage area is 2439
9.9 26,600 21.62 158,000
m i l e . O n the basis of these data and the regional
2

13.0 47,500 21.22 165,000


relationship developed for Problem 2.11, estimate the
16.44 75,600 17.83 103,000 100-year flow at this location. (Note: In practice con­
8.48 19,200 8.76 19,700 siderably m o r e d a t a should b e used to develop a re­
10.26 27,800 9.00 21,100 gional frequency curve.)
13.59 51,000 22.60 171,000
18.54 94,000 6.74 10,400
4,680 14,000 23,300
18.12 97,200 12.54 42,000
35,000 5,760 32,300
22.82 179,000 14.10 52,800
10,300 13,800 9,380
19.55 124,000 16.42 77,000
14,800
19.48 110,000 18.33 101,000
8.14 17,100
(2.13) T h e following data are for the Cimarron River
running through O k l a h o m a . O n the basis of these data,
estimate the 2-, 5-, 10-, 25-, 50-, and 100-year flows at
(2.8) Blackwell, Oklahoma, is in north central Okla­ points w h e r e the drainage area is 2050 m i l e . 2

homa. Figure 2.10 indicates a generalized skew coeffi­


cient of - 0.2 for this area. D e t e r m i n e a weighted skew
Area ( m i l e )
2

0m Q25
coefficient on the basis of the generalized value and Qi Q$0 £?l(Xi

the value calculated for Problem 2.4. How much does 1,038 7,830 17,800 27,000 42,100 55,800 71,600
this affect the estimate for 100-year peak flow? Would 1,879 8,880 20,000 31,700 53,300
you prefer an estimate based on the skew coefficient
determined only from the data, only from Fig. 2.10, or
a combination? Why? (2.14) A n analysis of 56 years of data indicated that
(2.9) Calculate the 9 0 % confidence intervals for the the probability of a flood peak exceeding 1500 cfs was
log Pearson type III flood frequency curve of Problem 0.02. During a 10-year period two such peaks occurred.
2.1. Plot the data, the log Pearson frequency curve, and If the original estimate of exceedance probability was
the confidence intervals on lognormal probability pa­ correct, what is t h e probability of getting two such
per. exceedances in 10 years?
(2.10) R e p e a t Problem 2.9 using the d a t a of Prob­ (2.15) Forty-five years of peak flow data are avail­
lem 2.7. able from an a n n u a l d a t a series. Six of the values are
(2.11) T h e following data are for Black Bear Creek zero. T h e remaining 39 values have a m e a n of 2150 cfs
at Pawnee, Oklahoma, for the period 1943 to 1971. and a standard deviation of 1200 cfs and follow a
This stream and the streams of Problems 2.4 and 2.7 lognormal distribution.
are in the same general region of O k l a h o m a . T h e a. W h a t is the probability of a peak flow exceed­
drainage areas are 576, 1859, and 46,850 m i l e for 2
ing 2750 cfs?
Black Bear Creek, Chikaskia River, and the A r k a n s a s b. Estimate the 20-year peak flow.
36 Chapter 2. Hydrologic Frequency Analysis

(2.16) A project is designed on the basis of a 25-year (2.29) W h a t is the difference between analytical and
return period peak flow. graphical flood frequency analysis? Describe the ad­
a. W h a t is the probability that the design will be vantages and disadvantages of each.
exceeded during the first year?
b. W h a t is the probability that the design will be
exceeded (at least once) during the first 10 years? References
c. W h a t is the probability that the design will be
exceeded (at least once) during the second 10 years? California State Department of Public Works (1923). "Flow in Cali­
fornia Streams," Bulletin 5, Chap. 5. [Original not seen, cited in
d. W h a t is the probability that the design will be
Chow (1964)]
exceeded (at least once) during the first 20 years?
Chow, V. T. (1951). A generalized formula for hydrologic frequency
e. What is the probability that the design will be analysis. Trans. Am. Geophys. Union 32(2): 2 3 1 - 2 3 7 .
exceeded exactly once during the first 10 years? Chow, V. T., ed. (1964). "Handbook of Applied Hydrology."
(2.17) W h a t design return period should be used to McGraw-Hill, N e w York.
ensure a 9 5 % chance that a design will not be ex­ Dalrymple, T. (1960). Flood frequency analysis, U.S. Geological
Survey Water Supply Paper 1543-A. In "Manual of Hydrology,"
ceeded in (a) 10 years, (b) 25 years, (c) 50 years, and
Part 4, "Flood Flow Techniques." U.S. Government Printing
(d) 100 years? Office, Washington, D C .
(2.18) R e p e a t Problem 2.17 using a 5 0 % chance. Durant, E. F., and Blackwell, S. R. (1959). The magnitude and
(2.19) W h a t design return period should b e used t o frequency of floods o n the Canadian prairies in spillway design
be 9 0 % confident of no more than one exceedance in a floods, in "Proceedings, Hydrology Symposium 1, Ottawa,
10-year period? Canada, 1959."
Haan, C. T. (1977). "Statistical Methods in Hydrology." Iowa State
(2.20) Starting at 0 and using class intervals of 40,000
Univ. Press, A m e s , IA.
cfs, plot a frequency histogram of the flow d a t a in Hazen, A. (1930). "Flood Flows, A Study of Frequencies and Magni­
Problem 2.7. Superimpose on the frequency histogram tudes." N e w York.
a lognormal distribution. Does the lognormal a p p e a r to Interagency Advisory Committee on Water Data (1981). "Guidelines
be a good approximation for the histogram? for Determining Flood Flow Frequency," Bulletin 17B of the
Hydrology Subcommittee. U.S. Department of Interior, Geologi­
(2.21) A flood detention structure has its spillway
cal Survey, Office of Water Data Coordination, Reston, V A .
designed on the basis of an estimated 1000-year flood.
Jennings, Μ. E., and Benson, M. A. (1969). Frequency curves for
W h a t is the probability that the design flow will b e annual series with some zero events or incomplete data. Water
exceeded in a (a) 10-year period, (b) 50-year period, Resources Res. 5(1): 2 7 6 - 2 8 0 .
(c) 100-year period or (d) 1000-year period? Matalas, N. C , and Langbein, W. B. (1962). Information content of
(2.22) Why do you think it is common to perform the mean. J. Geophys. Res. 67 (9): 3 4 4 1 - 3 4 4 8 .
Natural Environment Research Council (1975). "Flood Studies R e ­
flood frequency analysis on the logarithm of p e a k flows
port." Natural Environment Research Council, London.
rather than the peak flows themselves? Singh, V. P. (1987a). Hydrologic frequency modeling. In "Proceed­
(2.23) W h a t is the difference b e t w e e n an a n n u a l ings, International Symposium on Flood Frequency and Risk
series of peak flows and a partial duration series? Analysis, 1 4 - 1 7 May 1986, Baton Rouge, LA." Riedel, Dor­
(2.24) W h a t is the reason for using regional informa­ drecht, Holland.
tion on the flood peak skewness coefficient r a t h e r than Singh, V. P. (1987b). Regional flood frequency analysis. In "Proceed­
ings, International Symposium on Flood Frequency and Risk
an estimate based strictly on observed d a t a ?
Analysis, 1 4 - 1 7 May 1986, Baton Rouge, LA." Riedel, Dor­
(2.25) W h a t assumptions are m a d e in applying flood drecht, Holland.
frequency analysis techniques to estimate design flows? Singh, V. P. (1987c). Flood hydrology. In "Proceedings, International
(2.26) For a normal distribution with a m e a n of 25 Symposium on Flood Frequency and Risk Analysis, 1 4 - 1 7 May
and a variance of 400: 1986, Baton Rouge, LA." Riedel, Dordrecht, Holland.
Singh, V. P. (1987d). Application of frequency and risk in water
a. W h a t is the probability of exceeding 35?
resources. In "Proceedings, International Symposium Flood Fre­
b. W h a t is the probability of a value less quency and Risk Analysis, 1 4 - 1 7 May 1986, Baton Rouge, LA."
than 30? Riedel, Dordrecht, Holland.
c. What is the probability of a negative value? Tasker, G. D . (1978). Flood frequency analysis with a generalized
d. What value has a 1% chance of being ex­ skew coefficient. Water Resources Res. 14(2): 3 7 3 - 3 7 6 .
ceeded? Wallis, J. R., Matalas, N. C , and Slack, J. R. (1974). Just a moment.
Water Resources Res. 10(2): 2 1 1 - 2 1 9 .
(2.27) Discuss when a regional analysis would be
Weibull, W. (1939). A statistical study of the strength of materials.
beneficial to a frequency analysis? Ing. Vetenskaps Akad. Handl. {Stockholm). [Original not seen,
(2.28) Define the 50-year flood. cited in Chow (1964)]
Rainfall-Runoff Estimation
in Storm Water Computations

A hydrologist has an obligation to make the best hydro- logic data. T h e r e is no substitute for real, locally appli­
logic estimates possible, commensurate with the cost cable data.
and scope of a particular water m a n a g e m e n t problem. M e t h o d s for estimating hydrologic quantities such as
Hydrologic calculations are estimates, with the error in streamflow range from the relatively simple and widely
these estimates increasing as the degree of approxima­ used Rational Equation to complex, computer simula­
tion increases or as the estimation procedure is applied tion methods. T h e widespread availability of computers
beyond the range of conditions for which it is intended. m e a n s that one can employ more detailed and tedious
T h e hydrologist must determine if the scope, cost, or m e t h o d s than were formerly feasible. Again, the
importance of a particular project justifies collecting method selected should be able to provide the informa­
more data and using less approximate m e t h o d s or tion n e e d e d to design the system in question. If the
whether less precise techniques can be applied. T h e design involves a small drain pipe or culvert, possibly
hydrologist must make the best possible hydrologic all that is required is an estimate of the maximum flow
estimate and then proceed on that basis. Hydrologists the facility will be called upon to carry. If the design
should constantly check back on the projects they have involves storage and delay of the runoff, a complete
completed to determine the adequacy of the proce­ runoff hydrograph for the storm may be required. If
dures they have used. N e e d e d changes can then be the storage facility is large and will not empty to
incorporated in the estimation technique that is used. " p e r m a n e n t p o o l " elevation between storms, then pos­
T h e empirical and approximate nature of hydrologic sibly a continuous simulation of flow will be required.
estimation methods has led to the development and R e c e n t interest in water quality, especially ground
use of a great number of procedures for estimating water quality, has m a d e it even more important to
runoff, whether it be peak flows, runoff volumes, or accurately model actual flow paths. T h e quality of
complete hydrographs. It is difficult to say that one water in various phases of the hydrologic cycle is de­
method is absolutely better than another m e t h o d or p e n d e n t on the flow paths taken and materials the
that there is a best method. O n e can talk of a simple water has come in contact with. Simply modeling
method, but not necessarily of a best method. W h a t is ground water as a lumped storage or even treating all
best in one location may produce poor estimates in ground water flow as Darcian in n a t u r e can be mislead­
another location. W h a t is required to evaluate the ing in terms of water quality. Non-Darcian flow or
adequacy of a hydrologic procedure is actual hydro- large-pore (macropore) flow or flow in subsurface

37
38 3. Rainfall-Runoff Estimation in Storm Water Computations

cracks and channels is known to exist and generally cesses have a role to play in hydrology; however, pre­
occurs at flow rates several orders of magnitude larger cipitation, infiltration, overland flow, surface storage,
than Darcian flow. Thus, transported pollutants may detention storage, and of course streamflow are of
move into and within ground water systems much major importance in storm water hydrology.
quicker than anticipated based on a strictly Darcian T h e most basic equation in hydrology is the continu­
analysis. This is just o n e example of why it is necessary ity equation, which states that over any time interval
to match procedures used in hydrologic analysis with and for any hydrologic system the difference in the
the purpose of the analysis. volume of water entering the system, / , and leaving the
system, O , must equal the change in the volume of
water stored in the system, 5 :

HYDROLOGIC CYCLE
I - Ο = AS. (3.1)
T h e concepts of the hydrologic cycle are well known
and covered by many excellent texts on the subject. If the hydrologic system is a small catchment, the
T h e treatment of hydrology in this chapter is largely inflow to the system would be precipitation. T h e out­
limited to those parts of the hydrologic cycle of major flow from the system would be streamflow, d e e p seep­
importance in storm water management. This m e a n s age, and evapotranspiration. Storage within the water­
that primary emphasis will be placed on precipitation, shed would include soil water, ground water, ponds,
abstractions from precipitation, and the runoff process. lakes, reservoirs, channel storage, surface storage, de­
Such things as vegetal interception of precipitation, tention storage, and interception.
evaporation, transpiration, and soil water movement For short time intervals (hours), evapotranspiration
will be covered in less detail. G r o u n d water is covered and d e e p seepage can generally be ignored. For long
in a separate chapter. These latter parts of the hydro- time intervals (weeks), surface storage, surface deten­
logic cycle are of extreme importance when o n e is tion, and interception can often be ignored. In the
attempting to continuously simulate streamflow. T h e absence of ponds, lakes, or reservoirs, the hydrologic
emphasis in this chapter is on the analysis of single equation is further simplified.
events. This c h a p t e r primarily addresses hydrologic prob­
Figure 3.1 depicts a portion of a watershed during a lems that require a time scale of hours. Storm water
precipitation event. Shown in this figure are the pro­ computations are the major concern. M e t h o d s for esti­
cesses of rainfall, interception, evaporation, transpira­ mating storm rainfall a m o u n t s , intensities, and time
tion, infiltration, percolation, ground water flow, over­ distributions are presented. Also presented are meth­
land flow, subsurface flow, surface storage, detention ods for estimating runoff p e a k flows, volumes, and
storage, and channel precipitation. All of these pro- hydrographs from storm rainfall. This type of hydrol­
ogy is generally thought of as event-based hydrology.
A n overwhelming majority of storm water and erosion
control facilities are designed on t h e basis of single
events or single rainstorms of a given frequency. For
this reason, event-based hydrology is emphasized in
this chapter. Possibly the two most important factors
determining the runoff hydrograph from a small catch­
m e n t for a given volume of rainfall are the time distri­
bution of the rainfall a n d the infiltration characteristics
of the catchment. A major part of this chapter is
devoted to these two considerations. T h e third impor­
tant factor in hydrograph development, which also
receives considerable attention, is t h e routing of a
rainfall excess from the point of its generation to the
catchment outlet.
Also covered in t h e last major section of the chapter
but to a lesser d e g r e e are m e t h o d s for computing
hydrologic balances on p o n d s or subareas within water­
sheds. Such calculations are useful for sizing ponds,
Water Flow estimating the volume of waste water that can be
Figure 3.1 Hydrologic diagram. disposed of via irrigation of vegetated plots, estimating
Precipitation 39
the volume of water that might be required to maintain
a preset water level in a pond, and other problems
where long-term (weeks) hydrologic balances are of
concern. C h a p t e r 13 deals with continuous simulation
models or models for simulating continuous records of
runoff including storm water runoff as well as runoff
that occurs between rainstorms.

PRECIPITATION
LEGEND

T h e starting point for most storm water m a n a g e m e n t X Raingage location


studies is a consideration of precipitation. O u r concern -2.5 - isohyetal line
is limited to precipitation in the form of rainfall. It will T h i e s s e n polygon line
be further limited to storm rainfall. Rainfall d a t a are Figure 3.2 Estimating area depth of precipitation.
much m o r e widely available t h a n streamflow data and
much less affected by land-use changes. In the U.S.,
the National W e a t h e r Service maintains a system of same weight as o n e on the boundary or outside the
raingages and has an extensive system of publications catchment.
for disseminating the data collected. T h e network of T h e Thiessen m e t h o d is second in ease of applica­
gages consists of both recording and nonrecording tion and accuracy. T h e Thiessen m e t h o d weighting
gages. T h e nonrecording or standard gages are used factor is the area of t h e watershed nearest to the
primarily for 24-hr rainfall. T h e recording gages p r o ­ particular gage. This a r e a is d e t e r m i n e d by drawing
vide a complete time-intensity history of rainfall events. polygons whose sides a r e the perpendicular bisectors of
For small watershed storm water studies, relatively lines connecting t h e gages. T h e Thiessen method has
short duration but high-intensity rainfalls are of ex­ the advantage that as long as the same raingages are
treme importance. Thus, data derived from recording used, t h e weighting factors remain t h e same for each
raingages are generally required. rainstorm. T h u s , many storms can be analyzed without
recalculating t h e weighting factors.
T h e isohyetal m e t h o d weighting factor is the area of
M e a n A r e a Precipitation t h e w a t e r s h e d enclosed b e t w e e n adjacent isohyetes
(lines of constant rainfall). Isohyetal lines are drawn on
If one is attempting to reconstruct a historical rain­
a watershed in t h e same m a n n e r as topographic lines
fall on a watershed, interest may exist in m e t h o d s for
are drawn on a topographical m a p except rain depth
estimating the average rainfall on t h e watershed. T h e
r a t h e r t h a n elevation is t h e controlling variable. With
three most common m e t h o d s for computing average
t h e isohyetal m e t h o d , R in Eq. (3.2) is the average
watershed rainfall are the arithmetic m e a n , the
i

rainfall d e p t h associated with the weighting factor W


Thiessen polygon method, and the isohyetal m e t h o d .
{

and is generally taken as the average of the two enclos­


These m e t h o d s are illustrated in Fig. 3.2. All of the
ing isohyetal lines. Exceptions t o this are along water­
methods compute the average watershed rainfall as a
shed b o u n d a r i e s . T h e accuracy of this m e t h o d is better
weighted average of nearby raingages. T h e equation
t h a n the previous two m e t h o d s ; however, the amount
used is
of work is greater. T h e isohyetal m a p , and conse­
quently t h e weighting factors, must b e d e t e r m i n e d for
(3.2) each rainstorm.

where R is the average watershed rainfall, W is the t

weighting factor, and R> is the i t h rainfall a m o u n t . Example Problem 3.1 Mean Areal Rainfall
T h e arithmetic mean, or station average m e t h o d , is
the easiest but least accurate of t h e t h r e e methods. Compute the mean areal rainfall for the situation depicted
T h e arithmetic average is t h e average of all applicable in Fig. 3.2.
raingages. Thus, the weighting factors for E q . (3.2) are Solution: Table 3.1 shows the calculations that are in­
all unity, and account of factors such as the placement volved. The weighting factors shown in the table are the
of the gage with respect to the catchment b o u n d a r i e s is areas enclosed by the isohyetal lines or Thiessen polygons.
not taken. A gage central in the catchment gets the The total area is 175 mile . 2
40 3. Rainfall-Runoff Estimation in Storm Water Computations

Table 3.1 Calculation of Mean Annual Rainfall in Inches As discussed earlier, many hydrologic designs are
and Square Miles d o n e on t h e basis of a selected frequency. This m e a n s
o n e must be able to d e t e r m i n e rainfall d e p t h s for the
Arithmetic average
selected frequency. In specifying a rainfall, it is neces­
R = (2.30 + 2.90 + 2.80 + 3.25 + 2.75 + 1.50)/6 = 2.58 in.
a

sary to specify t h e d e p t h of t h e rainfall, the duration of


the rainfall, and the frequency of occurrence of the
rainfall. Alternatively, t h e intensity, duration, and fre­
Thiessen polygon
quency can be specified since the intensity, i, duration,
2.3 0 0
Γ, and d e p t h , D, are related by
2.9 48 139.2
2.8 27 75.6
D = iT. (3.3)
3.25 21 68.3 /? = 479.1/175
a

2.75 62 170.5 = 2.74 in.


D D F information thus indicates how much rain is
1.5 17 25.5
expected, t h e duration or period of time over which the
rain occurs, and how frequently the rain can be ex­
Total 175 479.1 pected. T h e D D F relationship for point rainfall has
b e e n expressed by t h e general relationship
Isohyetal method
3.1 34 105.4
2.75 104 286.0 /? = 471.1/175
a
i = KF /(T
X
+ b)\ (3.4)
2.25 29 65.3 = 2.69 in.
1.8 8 14.4 w h e r e F is frequency and K, b, x, and η are constants
for a particular location (Bernard, 1942). Values for K,
Total 175 471.1
b, x, and η can be estimated by nonlinear regression if
I D F or D D F d a t a are available. T h e s e coefficients
show considerable variation throughout the U.S.
Bernard p r e s e n t e d some typical values for these con­
stants. E q u a t i o n (3.4) is dimensional, and thus the
Comment: Simple visual inspection indicates that either the
coefficient Κ d e p e n d s on t h e dimensions of i.
Thiessen polygon or isohyetal estimate is superior to the
In the U.S., the use of Eq. (3.4) has been largely
arithmetic average. If one were to sketch the 2.70-in. iso­
hyetal line representing approximately the average rainfall replaced by U.S. W e a t h e r B u r e a u T P 40 (Hershfield,
based on either the Theissen polygon or isohyetal method, it 1961) and similar documents. T P 40 contains D D F
would traverse the basin in a manner approximating the d a t a for the U.S. for durations of 30 min to 24 hr and
center of the catchment. The same cannot be said of the frequencies of 1 to 100 years. A m o r e recent publica­
arithmetic average 2.58-in. isohyetal line, which would only tion (Frederick et al. 1977), H Y D R O - 3 5 , gives D D F
9

catch the upper right part of the catchment. data for the eastern U.S. for durations of 5 to 60 min
and frequencies to 100 years. H Y D R O - 3 5 has largely
superceded T P 40 for durations of 1 hr or less. Many
Point Precipitation Patterns local drainage authorities, highway d e p a r t m e n t s , water
resources agencies, etc., have p r e p a r e d D D F d a t a for
Rainfall Depth-Duration-Frequency use in specific localities. Rainfall d e p t h s for 24-hr
Often interest exists in rainfalls that can be expected durations and various frequencies for the U.S. can be
in the future rather than what has h a p p e n e d in the found in A p p e n d i x 3A. For t h e western U.S., reference
past. Historical rainfalls certainly guide us in estimat­ should be m a d e to National Oceanic and Atmospheric
ing future rainfalls, but only on a probabilistic basis. Administration ( N O A A ) Atlases as listed in t h e refer­
Many types of hydrologic analyses require estimates ences at the end of this chapter. T h e s e atlases give
of rainfall depths (or intensities) for certain durations detailed rainfall information reflective of t h e rapid
and frequencies of occurrence. Rainfall depth- variability in the rainfall regime in the western U.S.
d u r a t i o n - f r e q u e n c y ( D D F ) or i n t e n s i t y - d u r a t i o n - d u e to topographic effects.
frequency ( I D F ) data are generally available in the T h e p r o c e d u r e used to develop T P 40 was to p r e p a r e
form of tables, graphs, or maps on which isohyetal lines four key base m a p s showing the 2-year, 1-hr; 2-year,
are drawn. U.S. W e a t h e r Bureau T P 40 and similar 24-hr; 100-year, 1-hr; and 100-year, 24-hr rainfalls.
documents provide this information (Hershfield, 1961; A n n u a l series d a t a consisting of t h e maximum 60-min
Frederick et aL, 1977) in the form of maps for the U.S. and 24-hr rainfall d e p t h s converted to a partial dura-
Precipitation 41

tion series were used. T h e 2-year rainfall amounts were 10 Γ


determined by plotting on log-log p a p e r the return
period versus rainfall depth, drawing a smooth curve
through the data, and reading the 2-year value.
T h e 100-year depths were d e t e r m i n e d by using the
Extreme Value type I distribution for selected stations
with long rainfall records. T h e ratio of the 100- to
2-year depths was determined for these stations and a
m a p prepared showing this ratio. T h e 100-year rainfall
depths for the stations with short records were esti­
mated by the 100- to 2-year ratio.
Rainfall depths for other return periods were deter­ 0.1 I I I I I I I I II I I I I I I I Μ I 1 1 I Mill
.5 1 2 3 6 12 24
mined by plotting the 2- and 100-year depths on special DURATION (hours)
paper, connecting the points by a straight line, and
Figure 3.4 Smoothed intensity-duration-frequency curves, Still­
reading off the desired depths for other return periods.
water, Oklahoma, based on Eq. (3.4) with Κ = 1.75, χ = 0.21, b =
T h e spacing along the return period scale of this spe­ 0.12, and η = 0.80.
cial paper was empirical from 1 to 10 years and theo­
retical based on the Extreme Value type I distribution
from 20 to 100 years.
Rainfall depths for durations other than 1 hr or the curves is due to the difficulty of reading values from TP
24 hr were obtained by plotting the 1- and 24-hr values 40. Equation (3.4) was used as a smoothing relationship. The
coefficients K, b, x, and η were estimated by nonlinear
on a second special paper and connecting the points
regression. Nonlinear regression programs are widely avail­
with a straight line. This p a p e r was obtained empiri­
able for microcomputers. A program by Wilkinson (1987) was
cally from an analysis of records from 200 fist-order used to estimate these coefficients for Stillwater, Oklahoma,
U.S. W e a t h e r Bureau stations. T h e d e p t h for the 30- with the result that Κ = 1.75 if units are in inches, b = 0.12,
min duration was obtained by multiplying the 1-hr χ = 0.21, and η = 0.80. Using these coefficients, the Stillwa­
value by 0.79. ter data were smoothed and replotted in Fig. 3.4.

Intensity-Duration-Frequency Data for Stillwater, Oklahoma

Example Problem 3 . 2 IDF curves T(hr)

0.5 2 3 6 12 24
The data tabulated below were obtained from TP 40 for F (years) 1

Stillwater, Oklahoma. Use these data to construct a smoothed / (iph)


set of IDF curves. 1 2.20 1.45 0.85 0.62 0.33 0.22 0.13

Solution: The curves in Fig. 3.3 were plotted directly from 2 2.90 1.80 1.07 0.75 0.45 0.27 0.16

the TP 40 without any smoothing. The apparent roughness in 5 3.70 2.35 1.37 1.05 0.60 0.36 0.21

10 4.40 2.75 1.63 1.20 0.71 0.42 0.24

25 5.10 3.20 1.90 1.42 0.83 0.49 0.28

50 5.80 3.60 2.15 1.58 0.92 0.55 0.32


10
Curve parameter is frequency (years) 100 6.50 4.10 2.40 1.77 1.10 0.62 0.36

- 100
/ so
- Comment: These curves represent a consistant set of IDF
\ \ ^ ^ ^ / 10 information for Stillwater, Oklahoma, that can be used for a
:
ζ number of storm water design procedures. Figure 3.5 shows
how well Eq. (3.4) predicts rainfall intensities for Stillwater.
Such agreement makes it possible to replace the TP 40 maps
with Eq. (3.4) in computer programs if the coefficients for the
equation are known. Of course, the coefficients must be
I I 1 I I I I 1 1 1 1 1 1 1 1 1 1 1
estimated for each particular location to which the equation
.5 1 2 3 6 12 24 is to be applied. One of the problems at the end of this
DURATION (hours) chapter suggests a method for estimating the coefficients of
Eq. (3.4) without using a nonlinear regression approach.
Figure 3.3 Raw intensity-duration-frequency curves, Stillwater,
Oklahoma.
42 3. Rainfall-Runoff Estimation in Storm Water Computations

<
«ΙΟΟ

24-hr
3 90

I Ν: 6-hr

\
j 80 I - hr

^ 1- h r

& 60

tf 0 50 IOO ISO 200 250 300 350 400


AREA (SQUARE MILES)
Figure 3.6 Reduction factor for areal precipitation (Hershfield,
1961).

10 1
10° 10 1

If only a short record of rainfall is available, it may


Observed
be desirable to use a partial duration series r a t h e r t h a n
Figure 3.5 Predicted versus observed rainfall intensities, Stillwa­ an annual series of d a t a so that m o r e t h a n 1 data point
ter, Oklahoma.
p e r year may b e included in t h e analysis. T h e results of
a partial duration analysis can b e converted to an
annual series by using t h e multiplying factors of 0.88,
Equation 3.4 can be used to extrapolate I D F curves 0.96, and 0.99 for r e t u r n periods of 2, 5, and 10 years
toward shorter durations, although in the U.S. the use (Hershfield, 1961). F o r longer r e t u r n periods, no ad­
of H Y D R O - 3 5 is preferable where it applies. For j u s t m e n t is required.
example, the 10-min, 10-year rainfall intensity for Still­ F r o m t h e discussion on the development of the data
water (Example Problem 3.2) can be estimated as used to p r o d u c e T P 40, it should be a p p a r e n t that the
data representing any particular frequency (any partic­
1.75(10)° 21 ular line in Fig. 3.4) a r e derived from not o n e , b u t a
i = ^ = 7.71 lph. possible multitude of rain storms. This fact should be
(10/60 + 0.12) ' 0 80

kept in mind w h e n the development of temporal pat­


terns for rainstorms is discussed.
I D F curves can be constructed for any locality in the Rainfall D D F data are derived from point rainfall
U.S. (and many other places as well). From the curves information. W h e n they are applied to an area rainfall
one can easily determine the rainfall intensity (and situation, reduction factors as shown in Fig. 3.6 should
d e p t h ) associated with any duration and frequency. be applied (Hershfield, 1961). This has t h e affect of
Similar data for durations of up to 10 days are avail­ reducing the rainfall d e p t h s . T h e correction is much
able (U.S. D e p a r t m e n t of Commerce, 1964), but these m o r e significant for short duration storms, which might
longer duration storms are seldom used in small water­ typically b e t h u n d e r s t o r m s . F o r small watersheds, areal
shed storm water m a n a g e m e n t work. correction is not required.
T h e determination of D D F or I D F curves from local T h e areal reduction factors shown in Fig. 3.6 are
rainfall data requires several years of data. T h e fre­ generalized curves presenting approximate relation­
quency analysis techniques of C h a p t e r 2 can be used. ships for t h e U.S. It is possible to develop such curves
T h e probability distribution most often used is the for a specific locality. F o r small catchments, t h e correc­
Extreme Value type I distribution. T h e data would tion factor is small a n d generally ignored. R e a s o n s t h e
consist of the annual maximum rainfall d e p t h s for correction factor curves have t h e shape they d o in­
various durations over the period of record. T h e d a t a clude: (1) High intensity, short duration storms t e n d to
for each duration would be analyzed separately and cover a relatively small area; (2) long duration storms
plotted in the form of an I D F curve. Equation (3.4) or generally a r e widespread in areal extent; and (3) d a t a
a similar equation could then be used to smooth the used to p r o d u c e D D F curves a r e maximum observed
data to ensure consistency at a given duration for values for a particular storm and as o n e moves from
various frequencies. the c e n t e r of maximum intensity, t h e recorded rainfall
Precipitation 43

Figure 3.7 Regions of the conterminous U.S. for which P M P estimates are provided in the indicated Hydrometeorological
Reports (U.S. National Weather Service, 1984).

d e p t h decreases, thus lowering the average areal rain­ P M P to the 100-year rainfall for 10 m i l e (25.4 k m )
2 2

fall. and 24-hr durations range from 2 to 6 in the U.S. with


a majority of the ratios being 4 or 5. Many storms that
Probable Maximum Precipitation exceed 5 0 % of the P M P for several combinations of
T h e design of floodwater structures in situations durations and a r e a have b e e n recorded (Riedel and
where failure of the structure might e n d a n g e r h u m a n Schreiner, 1980). P M P estimates for the U.S. have
life, disrupt vital community services, or cause unusual b e e n published in a n u m b e r of different Hydromete­
economic loss must b e done with extreme care. Most orological R e p o r t s by t h e U . S . National W e a t h e r Ser­
engineers and agencies want to minimize the possibility vice (1943-1984) as indicated in Fig. 3.7. Figure 3.8
of failure of structures u n d e r such conditions. They do shows o n e set of estimates for the 6-hr, 6-mile 2

not want to design with a known probability of failure (15.5 k m ) P M P .


2

no matter how small that probability might b e . This has In the U.S., federal and state agencies classify dams
led to the adoption of design rainfalls based on t h e o n t h e basis of h a z a r d criteria. T h e s e criteria generally
concept of the Probable Maximum Precipitation or refer to the consequences of d a m failure in terms of
P M P . T h e P M P is defined as the theoretically greatest potential loss of life, a m o u n t of economic loss, disrup­
depth of precipitation for a given duration that is tion of vital services, d a m a g e to major transportation
physically possible over a given size storm area at a systems, etc. T h e m o r e severe the potential losses in
particular location at a certain time of t h e year t e r m s of these criteria, t h e greater t h e hazard classifi­
(National Research Council, 1985). cation. A n example set of hazard classifications is given
T h e major steps in estimating t h e P M P are: in Table 9.5. Generally the spillway of a structure
receiving a high hazard classification is required to pass
1. Study major rainstorms to d e t e r m i n e maximum
the flood resulting from the P M P . For some agencies,
areal rainfalls and ascertain t h e meteorological factors
this capacity criteria is relaxed somewhat for less haz­
important to the rainfall.
a r d o u s structures so that t h e spillways may be designed
2. Transpose the major storms within topographi­
to pass a fraction of t h e flood resulting from the P M P .
cally and meteorologically h o m o g e n e o u s regions.
A report by the National R e s e a r c h Council (1985)
3. Adjust the rainfall (for each transposed storm) by
summarizes these criteria for the U.S. Note that these
the ratio of maximum atmospheric moisture in t h e
criteria are generally spillway capacity criteria and not
place of occurrence to that which existed during t h e
storage criteria.
storm.
T h e concepts of Probable Maximum Precipitation
4. Smoothly envelope the resulting rainfall values
and Probable Maximum Flood have b e e n criticized
durationally, areally, and if generalized P M P is being
because they are neither probable nor maximum and
developed, regionally. Explanations should be given for
their likelihood of occurrence cannot be stated
discontinuities.
(Yevjevich, 1968). Yet the concepts have found wide
P M P values are quite conservative. T h e ratio of application and are e m b e d d e d in many regulations.
44 3. Rainfall-Runoff Estimation in Storm Water Computations

Linsley et al. (1982) indicate that the world's greatest rainfall intensity is known as the rainfall hyetograph or
observed point rainfalls can be enveloped by the rela­ simply the hyetograph. T h e r e are several possible
tionship m e t h o d s of arriving at a design hyetograph. T h e two
most c o m m o n m e t h o d s are either to adopt a historical
0.48
(3.5) storm that has occurred in the vicinity and has caused
considerable d a m a g e or to develop a synthetic design
where R max is the maximum rainfall observed for dura­ rainstorm.
tion, D , in hours. T h e value for the coefficient α is T h e historical rainfall a p p r o a c h has the advantage of
about 16.4 if rainfall is in inches and 417 if rainfall is in being readily identified a n d explained to engineers and
millimeters. O t h e r references give slightly different es­ to the public. It has t h e distinct disadvantage of an
timates for the coefficient α (Chow et al., 1988). This unknown frequency of occurrence. A s a m a t t e r of fact,
relationship apparently covers durations ranging from it is difficult to associate a r e t u r n period with a particu­
1 min to 2 years. Values from Eq. (3.5) should not be lar rainfall time-intensity p a t t e r n . For example, t h e
confused with P M P values. Equation (3.5) is empirical rainfall depicted in Fig. 3.9 for Stillwater, O k l a h o m a ,
and does not take into account any geographic or
meteorologic factors. Values calculated from this equa­ 40
tion generally exceed P M P values. 1.5

30
Rainfall T i m e Distribution DEPTH
1.0
(mm)
T h e analysis of m o d e r n storm water m a n a g e m e n t DEPTH 20
(inches)
systems often requires hydrographs of storm w a t e r flow
— n o t just peak flow or runoff volume estimates. Hy­ 0.5
10
drographs in turn require knowledge of t h e rainfall
time-intensity pattern that produces the hydrograph.
Thus, not only is it necessary to know the d e p t h of a ΤΊ H120I H~.
60 180
rainfall of a given duration and frequency, but the time
TIME (min)
distribution of the rainfall within its duration must be
known as well. A plot of the time distribution of Figure 3.9 Historical rainfall.
Precipitation 45

Table 3.2 Frequency of Sample catchment. Storm p a t t e r n s of this type are produced by
Historic Rainfall critically " s t a c k i n g " or arranging time increments of
rainfall so that the largest possible d e p t h is obtained
Duration Depth Frequency
(min) (years)
for any duration and frequency. Such a storm pattern is
(in.)
sometimes referred to as a balanced storm. A balanced
15 1.5 8 storm produces a rainfall d e p t h or intensity whose
30 2.9 49 frequency is i n d e p e n d e n t of the duration of the storm.
60 3.3 32 T h u s a 60-min rainfall selected from a balanced storm
90 3.5 25 would have the same r e t u r n period as a 6-hr rainfall
selected from the same balanced storm.
120 3.6 20
180 4.1 24 DDF Rainfall Pattern
O n e m e t h o d of synthetic balanced storm develop­
m e n t has b e e n to read the I D F curve (for example,
Fig. 3.4) for a given frequency at selected durations.
F r o m these intensities, incremental rainfall volumes
has different return periods d e p e n d i n g on the duration and intensities are c o m p u t e d and then rearranged to
being considered. Table 3.2 shows the maximum rain­ form a storm p a t t e r n . Example Problem 3.3 illustrates
fall intensities from this storm and the associated re­ this p r o c e d u r e .
turn period as determined from Fig. 3.4 or Eq. (3.4). It
is apparent that if this storm were a d o p t e d as a design
storm, the resulting runoff would be assigned a differ­
E x a m p l e P r o b l e m 3.3 Time distribution
ent return period depending on the critical flow-time
of rainfall—DDF method
(rainfall duration) p a r a m e t e r s for the watershed.
O n e approach to using historical rainfall data is to Based on data in TP 40, develop a synthetic rainfall time
select severe storms that have occurred on or n e a r the distribution for a 3-hr, 25-year rainfall event at Stillwater,
catchment of interest over a period of several years Oklahoma. Use a time increment of 15 min.
and use each of these storms in a rainfall-runoff model
Solution: Table 3.3 contains the required calculations. The
to estimate the resulting runoff, especially the peak
calculations will be illustrated by considering the 30-min line.
flow. T h e s e estimated flows can then b e analyzed by The intensity in iph is determined from Eq. (3.4) using
standard frequency analysis techniques and the runoff Κ = 1.75, b = 0.12, χ = 0.21, and η = 0.80 as determined in
with the desired return period estimated. This proce­ Example Problem 3.2.
dure eliminates the problem of having to assign a
return period to a rainstorm but does require data on KF X
1.75(25)° 21

severe rainstorms over a long period of time. ~ (T -+- b) n


~~ (30/60 + 0 . 1 2 ) ° 80
~~
T h e difficulty of assigning a return period to a total
rainstorm, or conversely of estimating a rainstorm pat­ Intensities could also be obtained from IDF curves such as
tern given a return period, has led to the development Fig. 3.4. The accumulated depth is from Eq. (3.3).
of synthetic storms to which return periods are as­ D = iT= 5.04(30/60) = 2.52.
signed. These synthetic storms have the advantage of
being a consistent basis for design, but have the disad­ The incremental depth is the difference in the current
vantage of having a very remote possibility of every depth and the depth at the previous time step or
happening exactly as specified. McPherson (1969) also 2.52 - 1.91 = 0.61.
comments on this problem. (It should be noted that
The last column is obtained by rearranging the entries in
any particular historical storm also has a very r e m o t e
column 4 while maintaining the concept of a balanced storm.
chance of being duplicated.)
Synthetic storms represent reasonable approxima­ Comment: The actual pattern or sequence of occurrence of
tions to actual storm patterns that might be expected the blocks of rainfall depth shown in the last column of Table
3.3 is not well defined. Several patterns are possible. What
for a given return period. T h e r e are a n u m b e r of ways
must be done for a balanced storm pattern is to keep the
to develop a synthetic storm pattern. M e t h o d s that
blocks of rainfall depth grouped so that no matter what
produce storms that have the same r e t u r n period re­ duration is being considered, the blocks of greatest depth are
gardless of the duration selected from within the storm contained within that duration.
are appealing from the standpoint of providing a con- The advanced type storm that results without rearranging
sistant basis of design regardless of the size of the the values in Table 3.3 is a storm pattern meeting the
46 3. Rainfall-Runoff Estimation in Storm Water Computations

Table 3.3 Synthetic 3-hr, 25-Year Rainfall, Stillwater, Oklahoma, DDF Method

Accumulated Incremental Depth


Duration Intensity 0
depth' 7
depth 4
increment^
(min) (iph) (in.) (in.) (in.)

15 7.62 1.91 1.91 0.10

30 5.04 2.52 0.61 0.12

45 3.85 2.88 0.36 0.17

60 3.14 3.14 0.26 0.26

75 2.67 3.34 0.20 0.61

90 2.34 3.51 0.17 1.91

105 2.08 3.65 0.14 0.36

120 1.86 3.77 0.12 0.20

135 1.72 3.88 0.11 0.14

150 1.59 3.98 0.10 0.11

165 1.48 4.07 0.09 0.09

180 1.38 4.15 0.08 0.08

Total 4.15 4.15

"From Fig. 3.4 or Eq. (3.4).


* Column (2) χ Column (l)/60.
4
Based on differencing Column 3.
^Column 4 rearranged.

balanced storm criteria. A pattern of this type indicates that curves as shown in Fig. 3.10 are used. Figure 3.11
the highest intensity rainfall occurs at the beginning of the shows the regions of t h e U.S. w h e r e the various type
storm and then gradually decreases to the end of the storm. curves apply. For o t h e r locations from throughout t h e
An alternative would be a delayed pattern obtained by ar­ world, the best fitting type curve can be found by
ranging the rainfall in order of increasing depth so that the developing a storm time distribution using t h e D D F
storm gradually builds in intensity and its highest intensity is
m e t h o d and by using each of t h e type curves as ex­
at the end of the storm.
plained below. T h e Type curve that best describes the
Obviously both the advanced and delayed rainfall patterns
D D F result can be a d o p t e d for that location. T h e type
represent unusual storms in that in one case the rainfall
begins in its most severe state and in the other it ends this II curve is applicable to the majority of the U.S. and
way. A compromise that is widely used is to place the most
intense increment near the center of the storm and pyramid
the values by placing the next highest alternately at the 1.0
beginning and end of the pattern as it develops both forward
and backward in time from the central high intensity.
0.8
The last column of Table 3.3 shows the resulting time
distribution of rainfall depths using this latter or central
pattern. This method of arriving at a time distribution for 0.6 r
rainfall can be used for any frequency and any total storm
duration. In this book this method is referred to as the
0.4
depth-duration-frequency (DDF) method.

0.2 F-

SCS Rainfall Pattern


0.0
20
T h e Soil Conservation Service (SCS) (1973) has 8 12 16

adopted a method similar to the D D F method. Several Time ( h o u r s )

dimensionless rainfall temporal p a t t e r n s called type Figure 3.10 SCS type curves for distribution of 24-hr rainfalls.
represents the most intense storm pattern. T h e type I A corresponds to the time from 10.5 to 13.5 hr on the type II
is the least intense p a t t e r n for short durations. Table curve. The values in column 3 are obtained by multiplying
3.4 contains the coordinates for the various SCS type the column 2 values by the 24-hr rainfall of 6.80 in. The
curves. values of the last column represent differences of successive
values in column 3.
T h e SCS method is based on the 24-hr rainfall of the
desired frequency. This rainfall is p r o p o r t i o n e d
throughout the 24-hr period using the appropriate curve
shown in Fig. 3.10 in the form of percentage mass T o facilitate c o m p u t e r analysis, it would be desirable
curves. T h e s e mass curves were derived so that for the to have an equation that approximates the actual shape
selected frequency, the d e p t h - d u r a t i o n relationship of the type curves. T h e equation
based on the curves would be very close to the 10.75
P(t) Τ 24.04
d e p t h - d u r a t i o n curve developed from a frequency = 0.5 + (3.6)
analysis of actual rainfall. T h u s , the t i m e - d e p t h pat­ 24 24 2\T\ + 0.04
terns based on the SCS method and the D D F m e t h o d where t is time and Τ is time - 12 in hours fits the
should be very similar. Example Problem 3.4 d e m o n ­ type II curve with a slight discrepancy on either side of
strates the SCS method for storm pattern development. 12 hr. T h e relationship is also a very good approxima­
For storm durations less than 24 hr, the steepest part tion of t h e type III curve. T h e discrepancy causes no
of the type curves is selected. For example, using t h e noticeable difference in resulting runoff. This equation
type II curve, a 3-hr storm uses the values extending was furnished by Cronshey (1981) who credited it to
from 10.5 to 13.5 hr in Fig. 3.10. N o r m a n (1981). In some computer-generated runoff
calculations contained later in this chapter, the above
equation was used t o describe t h e type II curve.
Example Problem 3.4 Time distribution T h e SCS type curves can be used to estimate the
of rainfall—SCS method d e p t h of rainfall for any duration and frequency from
t h e 24-hr rainfall for t h e same frequency. This is done
Develop a 3-hr, 25-year rainfall temporal distribution in by taking the difference in the ordinates of the type
15-min time increments using the SCS dimensionless type curve for the steepest part of t h e curve encompassing
curves. t h e desired duration. For example, the 6-hr, 25-year
Solution: Table 3.5 contains the required computations. rainfall for Stillwater can be estimated by multiplying
The 25-year, 24-hr rainfall for Stillwater is estimated from the 24-hr, 25-year rainfall of 6.5 in. (165 m m ) by the
Appendix 3A as 6.80 in. The values in column 2 are from Fig. largest difference in ordinates of the type II curve over
3.10 or Table 3.4. The most intense 3-hr part of the type II a 6-hr period. This difference is found by using the
curve that is applicable to Stillwater (Fig. 3.11) is used. This ordinates at t = 15 hours a n d t = 9 hours to be
48 3. Rainfall-Runoff Estimation in Storm Water Computations

Table 3.4 Coordinates for SCS Type Curves: Table Entries Are P/P24

Time Type I Type la Type II Type III

0.0 0.000 0.000 0.000 0.000


0.5 0.008 0.010 0.005 0.005
1.0 0.017 0.020 0.011 0.010
1.5 0.026 0.035 0.016 0.015
2.0 0.035 0.050 0.022 0.020
2.5 0.045 0.066 0.028 0.025
3.0 0.055 0.082 0.034 0.031
3.5 0.065 0.098 0.041 0.037
4.0 0.076 0.116 0.048 0.043
4.5 0.087 0.135 0.055 0.050
5.0 0.099 0.156 0.063 0.057
5.5 0.111 0.180 0.071 0.064
6.0 0.125 0.206 0.080 0.072
6.5 0.140 0.237 0.089 0.081
7.0 0.156 0.268 0.098 0.091
7.5 0.173 0.310 0.109 0.102
8.0 0.194 0.425 0.120 0.114
8.5 0.219 0.480 0.133 0.128
9.0 0.254 0.520 0.147 0.146
9.5 0.303 0.550 0.163 0.166
10.0 0.515 0.577 0.181 0.189
10.5 0.583 0.601 0.204 0.217
11.0 0.624 0.624 0.235 0.250
11.5 0.654 0.645 0.283 0.298
12.0 0.682 0.664 0.663 0.500
12.5 0.706 0.683 0.735 0.702
13.0 0.727 0.701 0.772 0.750
13.5 0.748 0.719 0.799 0.783
14.0 0.767 0.736 0.820 0.811
14.5 0.785 0.753 0.838 0.834
15.0 0.801 0.769 0.854 0.854
15.5 0.816 0.785 0.868 0.872
16.0 0.830 0.800 0.880 0.886
16.5 0.843 0.815 0.892 0.898
17.0 0.856 0.830 0.903 0.910
17.5 0.868 0.844 0.913 0.919
18.0 0.879 0.858 0.922 0.928
18.5 0.891 0.871 0.930 0.936
19.0 0.902 0.884 0.938 0.943
19.5 0.914 0.896 0.945 0.950
20.0 0.926 0.908 0.952 0.957
20.5 0.936 0.920 0.958 0.963
21.0 0.946 0.932 0.964 0.969
21.5 0.955 0.944 0.970 0.975
22.0 0.964 0.956 0.976 0.981
22.5 0.974 0.967 0.982 0.986
23.0 0.982 0.978 0.988 0.991
23.5 0.991 0.989 0.994 0.996
24.0 1.000 1.000 1.000 1.000
Precipitation 49

0.85 - 0.15 = 0.70. T h e estimated 6-hr, 25-year rain is Table 3.5 Synthetic 3-hr, 25-Year Rainfall,
found to be 0.70 x 6.5 = 4.55 in. (116 mm). This 6 hr Stillwater, Oklahoma, SCS Method
of rain can be distributed in time according to the type
curves by applying t h e relationship Time* Ordinate'* Depth' Increment Depth**
(hrs) (in.) (in.)
P'(t) _ P ( 1 2 + t - D/2) - P(12 - D/2)
(3.7) 10.50 0.204 1.39 0.00
PD ~ P(12 + D/2) - P(12 - D/2) 10.75 0.219 1.49 0.10
11.00 0.235 1.60 0.11
where t is the time within the storm in hours (0 < t <
D); D is the storm duration in hours (D < 24); PD is 11.25 0.257 1.75 0.15

the rainfall volume for the duration D and desired 11.50 0.283 1.92 0.17

frequency; P(t) is the value from the appropriate type 11.75 0.387 2.63 0.71
curve, and P'(t) is the accumulated volume of rainfall 12.00 0.663 4.51 1.88
to time t. 12.25 0.712 4.84 0.33
12.50 0.735 5.00 0.16
12.75 0.758 5.15 0.15
Example Problem 3.5. Synthetic time distribution 13.00 0.772 5.25 0.10
for a 6-hr storm 13.25 0.786 5.35 0.10
13.50 0.799 5.43 0.08
Compute the temporal pattern for a 6-hr, 25-year rain of
4.55 in. (116 mm) for Stillwater, Oklahoma, using the type II
Total 4.04
curve and a time increment of 1 hr.
Solution: Use Eq. (3.7). "Most intense 3-hr part of type II curve.
'Ordinates of type II curve.
P(12 - D/2) = P(12 - 6 / 2 ) = P(9) = 0.147 1
Column 2 χ 25-year, 24-hr rain of 6.80 in.
^Differencing of Column 3.
P(12 + D / 2 ) = P(\2 + 6 / 2 ) = P(15) = 0.854
PD = 4.55 in. (116 mm)
P(12 + t - 3) - 0.147
P'(t) = 4.55 Chicago Hyetograph
0.854 - 0.147
Keifer and C h u (1957) developed a balanced storm
= 6.44[P(9 + 0 - 0 1 4 7
] p a t t e r n known as the Chicago hyetograph by using an
equation similar to Eq. (3.4). They partitioned the
storm p a t t e r n so that at any intensity the time from the
t P(9 + t) P\t)
center of the storm back to the rising limb, t , divided a

0 0.147 0.00 by the time from the rising limb to the falling limb of
1 0.181 0.22 t h e storm, t + t , was a constant r (see Fig. 3.12).
a h

2 0.235 0.57
F r o m Eq. (3.3),
3 0.663 3.32
4 0.772 4.03 where / is the average intensity over the duration 7 ,
a v e

5 0.820 4.33 and D is the total d e p t h of rainfall. T h e instantaneous


6 0.854 4.55 intensity, i, is given by t h e time rate of change of d e p t h
or
dD di.^
Reevaluation of the SCS type curves is ongoing. It is dT ' a v e + T
dT
not unreasonable to expect that modifications for these F r o m Eq. (3.4)
curves will appear. Table 3.6 compares the D D F esti­
mates and SCS type II estimates of the 25-year rainfall KF X
K'
for several durations at a n u m b e r of locations in t h e dVC
(T+b)" (T + b)" 7

U.S. and Canada. Considering the wide range of topo­


graphic and climatic conditions, the type II curve does where K' = KF X
for a fixed frequency, F . Therefore
a reasonably good job of approximating d e p t h - d u r a ­ di ck\>t* „ Ι
tion-frequency data found in T P 40. -nK'(T + b)
dT
50 3. Rainfall-Runoff Estimation in Storm Water Computations

Table 3.6 Comparison of TP 40 and SCS Type II Rainfall Depths for a 25-Year Rainfall of
Various Durations: Type II Values Are First Row and TP 40 Values Are Second Row for Each
Location

Duration (hr)

Location 0.5 1 2 3 6 12 24

Type II ordinate differences 0.37 0.46 0.54 0.59 0.70 0.84 1.00

Lexington, Kentucky 1.85 2.30 2.70 2.95 3.50 4.20 5.00


1.85 2.30 2.75 3.15 3.60 4:20 5.00

St. Louis, Missouri 2.13 2.65 3.11 3.39 4.03 4.83 5.75
2.10 2.65 3.22 3.50 4.30 4.90 5.75

Oklahoma City, Oklahoma 2.52 3.13 3.67 4.01 4.76 5.71 6.80
2.55 3.25 3.90 4.20 5.10 6.10 6.80

Gillette, Wyoming 1.11 1.38 1.62 1.77 2.10 2.52 3.00


1.20 1.40 1.80 1.90 2.30 2.70 3.00

Denver, Colorado 1.18 1.47 1.73 1.89 2.24 2.69 3.20


1.32 1.70 1.90 2.20 2.50 2.75 3.20

and from an earlier relationship or

TnK' K'[(\ - n)T + b] T =


r 1 - r
(3.8)
(T + b) n+l
(T + b)' Using these relationships in Eq. (3.8) for Τ yields for
the rising limb
From the definition of r K'[(i - n)(t /r)
a + b]
(3.10)

(3.9) and for the falling limb


ta + t„ τ
K'[(l - / ! ) ( / * / ( 1 - r ) ) + b]
(3.11)
- Γ ) + 6 Γ '
( , „ / ( !

Note that r = 0.5 produces a symmetrical pattern. A


value of r between 0 and 0.5 produces an advanced
pattern and a value between 0.5 and 1.0 produces a
delayed p a t t e r n .

E x a m p l e P r o b l e m 3 . 6 . Time distribution
of rainfall—Chicago method

Develop a 25-year, 3-hr storm at Stillwater, Oklahoma,


using the Chicago hyetograph method. Use 15-min time
increments.
Solution: Recall from Example Problem 3.2 that Κ = 1.75,
Figure 3.12 Chicago hyetograph. b = 0.12, χ = 0.21, and η = 0.80 for Stillwater. The value for
Precipitation 51

Table 3.7 Synthetic 3-hr, 25-yr rain, Stillwater, Oklahoma,


Chicago Hyetograph Method

Time 'a° 'b


r
Intensity* Depth Modified
(min) (hr) (iph) (in.) depth (in.)
Depth
0 (inches) i.o
1.375 0.35 0.00 0.00
15 1.125 0.41 0.10 0.10
30 0.875 0.52 0.12 0.12
45 0.625 0.72 0.16 0.16
60 0.375 1.19 0.24 0.24
75 0.125 3.50 0.59 0.59
Time (min)
0* 7.35-
Figure 3.13 3-hr, 25-year pattern, Stillwater, Oklahoma.
90 0.125 3.50 1.36 c l.65 d

105 0.375 1.19 0.59 0.59


120 0.625 0.72 0.24 0.24
135 0.875 0.52 0.16 0.16
average intensity for the interval and the duration of the
150 1.125 0.41 0.12 0.12 interval.
165 1.375 0.35 0.10 0.10 If Eq. (3.4) is used to compute the 25-year, 3-hr rainfall for
180 1.625 0.30 0.08 0.08 this location, the result is 4.15 in. The total indicated in the
fourth column of Table 3.7 is only 3.86 in. The difference is
largely due to the assumption used in calculating the inten­
Total 3.86 4.15
sity at t = t = 0. To bring the storm total in line with the
a b

estimate from Eq. (3.4), the depth at the center of the


"From Eq. (3.10) or (3.11) with r = 0.5, K' =3.44, b = 0.12, η = 0.8.
pattern was adjusted by 4.15 - 3.86 — 0.29 inches to 1.65 in.
*Time increment of 0.05 hr used to define storm center.
1Depth for time increment 75 to 90 min.
Sce Example Problem 3.6 for explanation.
d

Figure 3.13 c o m p a r e s the rainstorms synthesized for


Stillwater, O k l a h o m a , in Example Problems 3.3, 3.4,
r is taken as 0.5. K' of Eqs. (3.10) and (3.11) is found to be a n d 3.6 based o n the D D F m e t h o d , the SCS m e t h o d
using t h e type II curve, and t h e Chicago hyetograph.
K' = KF X = 1.75(25) · 0 21 = 3.44 T h e m e t h o d s p r o d u c e essentially identical results. T h e
Chicago p a t t e r n could b e brought m o r e in line with the
for rainfall in inches. Values for i and i are calculated a b
o t h e r two p a t t e r n s by using a value of r less than 0.5.
from Eqs. (3.10) and (3.11) and shown in Table 3.7. So that
In view of t h e uncertainties that exist in the storm
the storm pattern will be in agreement with the patterns
developed using the SCS and DDF methods, the center of p a t t e r n s a n d uncertainties in o t h e r factors that con­
the storm is taken at the center of the time increment from tribute t o runoff, t h e m e t h o d s can b e considered t o b e
75 to 90 min or at 82.5 min. The calculations are illustrated essentially equal. Most examples in this text are based
for the time of 60 min. At 60 min, t is 82.5 - 60 = 22.5 min a on the SCS type II distribution as approximated by
or 0.375 hr. The value of i is determined from Eq. (3.10):
a Eq. (3.6).
T h e r e have b e e n a few a t t e m p t s at studying the
K'[(l - n)(t /r)a + b]
i =
n
actual time distribution of observed rainfalls. Major
{t /r
a + b) n + l
difficulties with studies of this type are obtaining long-
3.44[(1 - 0.8)(0.375/0.5) + 0.12] t e r m records from which short duration intensities can
= 1.19. b e extracted, t h e mass of d a t a required, and t h e vari­
(0.375/0.5 + 0.12)° 8 + 1
ability of n a t u r a l rainfall. Huff (1967) has studied the
The depth increment is the average intensity over the time time distribution of heavy rainfalls with durations of 3
interval times the duration of the interval or to 48 hr in central Illinois. H e divided t h e storms into
four groups d e p e n d i n g on the time quartile in which
depth = 0.25(1.19 + 0.72)/2 = 0.24.
t h e majority of the rain occurred. T h e time distribu­
It can be seen that at t = t = 0, i = i = K'/b
a = 18.8 b a b n
tions of the storm rainfalls within the various quartiles
iph. Since this is unreasonably high and represents an instan­ w e r e c o m p u t e d as w e r e the p e r c e n t a g e of t h e storms
taneous value, a time of 0.05 hr was arbitrarily used to define having t h a t particular time distribution or t h e o n e
the center of the pattern. The depth values are based on the above it. T h e difficulty of assigning a return period to
52 3. Rainfall-Runoff Estimation in Storm Water Computations

A ABSTRACTIONS FROM PRECIPITATION

Abstractions from precipitation are losses from p r e ­


cipitation that d o not show u p as storm water runoff.
T h e volume of storm w a t e r runoff is t h e volume of
precipitation minus t h e volume of abstractions. Precip­
itation minus abstractions is also known as rainfall
excess or effective rainfall. T h u s , t h e volume of storm
t τ w a t e r runoff is equal to t h e volume of effective rainfall.
Abstractions include interception, evapotranspiration,
Β surface storage a n d surface detention, b a n k storage
I i - °
a n d infiltration. T h e combination of interception and
'max - — surface storage is sometimes d e n o t e d as surface
storage.

Interception

t Τ Some rainfall is intercepted by vegetation before it


reaches t h e ground. T h e a m o u n t of interception varies
with the type, density a n d stage of growth of t h e
C
vegetation, intensity of t h e rainfall, and wind speed.
O n an annual basis, interception may involve a signifi­
cant p e r c e n t a g e of the total rainfall. A d e n s e forest
canopy may result in interception values of 2 5 % of t h e
a n n u a l rainfall. This would be t h e case in climates with
frequent, light rainfalls; low wind speeds; and ever­
g r e e n vegetation.
t Τ O n a p e r storm basis, interception storage may range
Figure 3.14 Hypothetical, unbalanced storm patterns. u p to 0.4 in. (10 m m ) for d e n s e forested vegetation but
is generally considerably less t h a n this for a mixed
cover catchment. F o r significant rainfall events likely to
p r o d u c e large stormwater flows from small catchments,
t h e a m o u n t of rainfall going to satisfy interception
the various Huff patterns and the selection of t h e
storage is generally a small p e r c e n t a g e of total storm
proper quartile storm have precluded their widespread
rainfall.
adoption. T h e Huff study definitely d e m o n s t r a t e d t h e
variability of time distributions that occur at particular
location. This in turn means that any particular p a t t e r n
that is adopted is unlikely to actually occur. Design Evapotranspiration
patterns are simply possible, good estimates of reason­ Evapotranspiration or E T is the combination of
able time patterns. evaporation and transpiration. Evaporation is t h e phase
For preliminary designs and rough calculations, storm change of water from a liquid to a vapor. Evaporation
patterns might be approximated as shown in Fig. 3.14. from plant surfaces of water that has traversed from
In this figure, / is the intensity and D r e p r e s e n t s t h e t h e soil t h r o u g h the plant is t e r m e d transpiration. All
rainfall d e p t h for the desired frequency and duration, evaporation from a leaf surface is not transpiration
T. T h e p r o p e r relation for determining / for each of
m a x since intercepted water is also evaporated. O n an an­
the t h r e e assumed patterns is shown. It must b e kept in nual basis, E T generally involves a large fraction of the
mind that these p a t t e r n s are not balanced p a t t e r n s . In total precipitation. In arid climates most of t h e rainfall,
pattern (a) the time of the maximum intensity is arbi­ 9 0 % or m o r e , may be lost t h r o u g h E T . In m o r e humid
trary. In p a t t e r n (c) the duration, T', of the constant climates, E T may account for 40 to 7 0 % of t h e annual
intensity portion and its location within Τ are arbitrary. precipitation.
Both p a t t e r n s (a) and (b) are special cases of pat­ In spite of the high total fraction of rainfall involved
tern (c). in t h e E T process on an a n n u a l basis, for individual
Abstractions from Precipitation 53

rainstorms, E T is generally a minor factor and not Table 3.8 Typical Values for Surface Storage
included in storm water computations. Infiltration, as (Wright-Mclaughlin Engineers, 1969)
we shall see, is a significant abstraction during storm
Land cover Surface storage Recommended value
events. M u c h of t h e water that is abstracted by infiltra­
(in.) (in.)
tion is eventually lost from t h e catchment system via
E T . This loss occurs at a relatively slow r a t e (0 to Impervious
0.4 in. or 10 m m p e r day) and is the most significant Large paved area 0.05-0.15 0.10
during the times between storm events, not during the
Roofs, flat 0.10-0.30 0.10
storm events themselves. W h e n rain falls on w a r m
Roofs, sloping 0.05-0.10 0.05
surfaces, some evaporation occurs. Sometimes an evap­
orative loss of 0.1 in. (2.5 m m ) is used to reflect this
Pervious
loss if a large part of the basin is covered by concrete,
asphalt, roofs, etc. Lawn grass 0.20-0.50 0.30
Wooded area 0.20-0.60 0.40
Open fields 0.20-0.60 0.40
Bank Storage
Bank storage represents losses from streamflow dur­
ing a period of rising stage in a stream and the subse­
q u e n t seepage of streamflow into the banks of t h e
T h e value of the constant K can be estimated by
stream. During t h e falling stage, this water generally d

noting that when Ρ — F is n e a r zero, all of the water


seeps back into t h e stream. T h u s b a n k storage is not
goes to filling depressions so that dV /(d(P - F)) is
actually a loss from runoff but a storage a n d delay in d

essentially 1. Based on this reasoning, K is equal to


the runoff process. For small catchments, b a n k storage d

generally plays a small role in storm water runoff. 1Α,·


Neglecting interception losses, the rate that water
becomes available for surface runoff, σ , is i - f - φ,
Surface Storage and Detention where i is the precipitation rate, / is the infiltration
Surface storage is the volume of water required to rate, a n d ψ is equal to dV /dt. Based on these as­
d

fill depressions and other storages before surface runoff sumptions, the surface runoff supply rate becomes
begins. D e t e n t i o n storage is t h e buildup of small d e p t h s
a=(i-f)(l-e- { p
- F ) / s
<). (3.13)
of water required to support the runoff process. Actual
m e a s u r e m e n t s of surface storage and d e t e n t i o n a r e T h e ratio of surface runoff supply rate to the differ­
extremely difficult to make and consequently are prac­ ence in t h e rainfall a n d infiltration rates becomes
tically nonexistent. Wright-McLaughlin Engineers
(1969) in a special study of u r b a n hydrology in the a/(i-f) = 1 - -( - p f ) / s
* (3.14)
e

Denver, Colorado area, r e c o m m e n d e d t h e values shown


in Table 3.8 for surface storage. T e r s t r i e p a n d Stall and ranges from 0, at t h e beginning of the precipitation
(1974) recommend a value of 0.20 in. (5 m m ) for event (P - F = 0), to 1 when Ρ » F.
detention storage for bluegrass turf. Some investigators Equation (3.14) is plotted in Fig. 3.15 based on a turf
(Linsley et al., 1949) recognized that a watershed sur­ area with an average overall S of 0.25 in. or a pave­
d

face is m a d e u p of depressions of various sizes and that m e n t with S equal to 0.0625 in. T h e vertical dashed
d

as some of t h e smaller depressions were filled, surface line in Fig. 3.15 r e p r e s e n t s the surface runoff supply
runoff could begin even though the larger depressions ratio if it is assumed that t h e overall average surface
were still filling. A n exponential relationship storage must b e filled before any runoff can begin. This
would be the case if t h e abstractions indicated in Table
Vd - S [l
d - e - ^ - ^ ] (3.12) 3.8 were subtracted directly from the beginning of a
storm before any water was allowed to become avail­
has b e e n proposed where V is the volume of w a t e r in
d able for surface runoff.
surface storage, S is the available surface storage,
d Tholin and Keifer (1960) surmised that the actual
Ρ - F is the accumulated mass of surface storage situation might be between that given by Table 3.8
supply (i.e., accumulated rainfall minus infiltration a n d and that given by Eq. (3.14). They found that the curve
other losses except surface storage), and K is a con­ d of t h e normal distribution, as shown in Fig. 3.15,
stant. fell within their desired range. This curve can be
54 3. Rainfall-Runoff Estimation in Storm Water Computations

0.125 0.25 0.375 0.50 (Turf) As might be expected, surface storage is of greater
ο Mass Overland Flow and Depression Storage Supply (P-F) importance on flat surfaces t h a n on steep surfaces.
ο QI25 ( P a v e m e n t s )
100
1.00
Viessman (1967) found the relationship shown in
Fig. 3.16 for four impervious drainage areas. T h e line
i0.90 in Fig. 3.16 should be extrapolated with care. M o r e
a b ι
φ
ο likely the surface storage would decrease exponentially
0.80
c with slope approaching zero at very steep slopes.
ο
tt>
0.70 If long duration rainfalls are being studied, the val­
c
ues of surface storage will not appreciably affect esti­
Δ
0.60 m a t e d runoff rates since the early part of the storm
ο
would fill this storage prior to the occurrence of the
|0.50 major runoff producing part of t h e rainfall. Note that
the values in Table 3.8 d o not include built-in storage
Ηθ.40
in the form of detention basins.
n 0.30

Ηθ.20
Infiltration
T h e major abstraction from rainfall during a signifi­
δ! HO.IO
cant runoff-producing storm is infiltration of water into
σ σ
I 8 pervious areas including soils, infiltration basins, and
0 50 100 150 200 forest litter. T h e process of infiltration of water and
Mean Depth as a Percentage of Overall Depth subsequent water movement is an exceedingly complex
of Depression Storage.
process. In this discussion, soil is used in a general
Figure 3.15 Depth distribution curve of depression storage. Enter
graph from top, read down to selected curve, and project right or left
sense. T h e u p p e r part of Fig. 3.17 shows t h e soil water
as desired (modified from Tholin and Kiefer, I960). content as a function of time and d e p t h during a
rainfall event. T h e curved lines represent water con­
tent at various times with time increasing as the wet­
ting profile advances d e e p e r in t h e soil profile. T h e soil
approximated using a normal distribution with a m e a n was at a uniform initial water content and the soil
equal to S and a standard deviation of S /3
d or d
properties are uniform with d e p t h .
In general, t h e infiltration rate is d e p e n d e n t on soil
physical properties, vegetative cover, a n t e c e d e n t soil
σ
rD/s
dx.
= d
exp water conditions, rainfall intensity, and t h e slope of the
i - f J-ο» 2 \ S /3
infiltrating surface. Referring again to Fig. 3.17, if t h e
d

soil physical properties are not uniform with depth, the


T h e value of S d might be estimated from the d a t a in p a t t e r n of soil wetting may be greatly altered. It is not
Table 3.8. u n c o m m o n to find a soil layer that is less permeable
than t h e surface layer. If this less p e r m e a b l e layer is
located n e a r the sou surface, it restricts the wetting
0.20, front and reduces the infiltration rate. This restricting
layer may b e a shallow rock formation, a soil layer
higher in clay content, a layer compacted by heavy
0.I5
CO
equipment, a plow layer, or a fragipan. Final grading
on construction, landscaping, or mining jobs may in­
^ o.iol volve spreading stockpiled materials. In this way, a
m o r e p e r m e a b l e layer of soil may b e placed over a
0.05I compacted layer. This is especially common when sod­
ding is d o n e , as t h e sod is laid directly on relatively
compact materials. Consequently, water can easily en­
0 I 2 3 4 ter the surface layer, but is restricted by the compacted
Slope (%) zone. Thus, light rainfalls are easily absorbed, but
Figure 3.16 Depression-storage loss versus slope for four impervi­ heavier rains soon s a t u r a t e the surface soil layer. T h e
ous drainage areas (Viessman, 1967). saturated layer has a very low infiltration capacity,
Abstractions from Precipitation 55

WATER CONTENT (cc/cc)


0.08 0.24 0.40

TIME (hr) TIME (hr)


Figure 3.17 Water content profiles.

which results in runoff rates being n e a r t h e rainfall T h e a n t e c e d e n t soil water content also alters the
rates. infiltration r a t e . Generally, a wet soil has a lower
Bare soils tend to have lower infiltration rates t h a n infiltration rate t h a n a dry o n e . T h u s a rain falling on a
soils protected by a vegetative cover. O n b a r e soil, the wet soil will p r o d u c e m o r e runoff at a higher rate than
impacting raindrops tend to puddle the soil. T h e en­ t h e same rain on a dry soil.
ergy of the falling rain breaks down soil aggregates and T h e infiltration opportunity time is a function of the
small particles are carried into the soil pores. T h e net slope of the infiltrating surface. O n a steep slope, the
result is a lowering of t h e infiltration rate. w a t e r t e n d s to run off rapidly and thus have less
56 3. Rainfall-Runoff Estimation in Storm Water Computations

opportunity for infiltration than on a gentle slope. Currently, it is difficult to incorporate this concept into
Also, the soil type found on steeper slopes is generally hydrologic analyses of ungaged areas since p a r a m e t e r s
not the same as on flatter slopes. Often the m o r e defining t h e extent a n d response of t h e variable source
sloping soils, especially if the soils have b e e n used for areas are not available.
agriculture, have experienced more erosion t h a n flatter Runoff estimation based on classical infiltration a p ­
soils. This in turn generally results in lower infiltration p r o a c h e s are known as H o r t o n i a n a p p r o a c h e s and cur­
rates. rently d o m i n a t e in the a r e a of hydrologic analyses. A
Rainfall intensity affects the infiltration rate in two great deal of effort has b e e n e x p e n d e d in developing
ways. For high-intensity rains, the raindrops tend to be the mathematical theory of t h e infiltration of water
larger and have more energy when they strike t h e soil. into soils and t h e subsequent movement of this w a t e r
Thus high-intensity rains are more effective in sealing within t h e soil. T h e physical principles and m a t h e m a t i ­
the soil surface than are low-intensity rains. A good cal relationships are well defined.
vegetative cover can minimize this effect. Obviously,
the infiltration rate cannot exceed the rainfall r a t e for Richards Equation
prolonged periods of time. In the absence of any One-dimensional, u n s a t u r a t e d flow is governed by
p o n d e d water or water flowing over the soil, the maxi­ Darcy's law a n d the continuity equation and is given by
m u m possible infiltration rate is the lesser of t h e rain­
dh
fall rate or the soils infiltration capacity. In t h e pres­
v=
z - K ( h ) - (3.15a)
ence of p o n d e d water or surface flowing water, the
infiltration rate equals the infiltration capacity until and
this surface supply of water is exhausted.
3Θ d dh
T h e combination of all of the factors governing infil­
tration throughout a watershed interact in such a fash­ *-ϊϊ-*<*>ϊί· ( 3 i 5 b
»
ion as to result in a very complex spatial and temporal where v is flow velocity, ζ is t h e coordinate direction
z

distribution of infiltration capacity. A t some locations, positive upward, h is t h e soil water potential, θ is t h e
the infiltration capacity may be so high as to practically soil water content, and K(h) is the hydraulic conduc­
never produce surface runoff, whereas other areas may tivity. T h e soil water potential is t h e sum of two com­
have low infiltration capacities and produce surface p o n e n t s — t h e matric potential ψ a n d the gravitational
flow from light rainfalls. Betson (1964) has t e r m e d potential z. T h e hydraulic conductivity is a strong
these latter areas as source areas. function of soil water potential and may change by 7
In recent years, an alternative theory of streamflow o r d e r s of m a g n i t u d e over the range of soil water poten­
generation known as the variable source area concept tials commonly e n c o u n t e r e d (Fig. 3.17). K(h) is also
has been proposed for hilly terrain (Hewlett and d e p e n d e n t on the soil being considered and may change
Hibbert, 1967). T h e concept particularly applies to with d e p t h . K(h) is very small for dry soils and in­
highly pervious soils. At the beginning of a rainfall creases in value as the soil gets wetter. A t saturation,
event, a water table exists with a capillary zone above. K(h) is t h e saturated hydraulic conductivity. T h e soil
T h e water table near a stream would generally be w a t e r content is also a function of t h e water potential
closer to the soil surface than at some distance from of the soil. Equation (3.15) is based on the continuity
the stream since water table slopes a r e generally flatter equation and Darcy's Law applied to u n s a t u r a t e d flow.
than surface topography slopes in hilly terrain. As the T h e book by Hillel (1971) can be consulted for m o r e
rain continues, the water table n e a r the stream rises detail on t h e relationship. C h a p t e r 11 of this book
faster than that at higher elevations because the perco­ treats saturated ground w a t e r flow.
lating water has less distance to travel. This rise near If o n e considers a uniform, d e e p soil that is initially
the stream may eventually cause the water table to at a constant water content a n d subjected to a constant
reach the soil surface resulting in a saturated condi­ rainfall rate, the change in infiltration rate with time
tion. F u r t h e r rain on the saturated area becomes satu­ can be d e d u c e d . Infiltration is simply equal to v at t h e
z

rated return flow and quickly reaches the stream. Fur­ soil surface c o m p u t e d from Eq. (3.15a). If t h e soil is
ther rain causes the saturated area to grow in size and initially dry, K(h) will b e small. W h e n t h e first rain
to move upstream. As the saturated area grows, a occurs, it will wet u p t h e surface of t h e soil so that
larger portion of the watershed is contributing to satu­ dh/dz will be large. T h e product K{h)dh/dz will thus
rated return flow. W h e n the rainfall rate diminishes or b e relatively high and will result in high infiltration
stops, the saturated areas drain, resulting in a shrink­ rates. As time goes on, the soil will uniformly wet u p
ing of the saturated zone. This growing and shrinking and dh/dz will become small. T h e second t e r m of
saturated zone is known as a variable source area. Eq. (3.15a) will then b e small; however, K(h) will now
Abstractions from Precipitation 57

be higher than it was initially and will control the


infiltration process. T h u s , early in t h e rainfall event,
the water potential gradient at the soil surface governs
the infiltration rate, while later in t h e event, t h e hy­
draulic conductivity governs.
Figure 3.17 shows t h e results of applying Eq. (3.15)
to a Bethany soil having an initial water content of 0.20
c c / c c and a rainfall rate of 15 c m / h r . Within the
figure are shown the water p o t e n t i a l - w a t e r content
functions and water p o t e n t i a l - h y d r a u l i c conductivity
functions. Also shown are water content profiles with
d e p t h at time increments of 0.2 hr. Finally, t h e infiltra­ 0 I ' ' 1 1 ' 1 1 1 1 1 1 1 1 1 ι J
tion rate in c m / h r and t h e cumulative infiltration are 0.0 0.5 1.0 1.5 2.0
shown up to 2 hr. Time ( h o u r s )

Equations (3.15) have found limited application in Figure 3.18 Horton infiltration curves.
design hydrology. Difficulties experienced in using t h e
equation are the nonuniformity of soils, both spatially
and with d e p t h ; the great n u m b e r of m e a s u r e m e n t s
regression, points every 0.2 hr were taken from the
n e e d e d to define the required p a r a m e t e r s ; and the
infiltration curve of Fig. 3.17. T h e s e time-infiltration
difficulty of solving the relationships when t h e required
pairs w e r e subjected to a nonlinear least-squares esti­
data are available. A further difficulty is that of specify­
mation using S Y S T A T (Wilkinson, 1987) with the re­
ing the applicable boundary conditions for t h e equa­
sult that f = 1.391 c m / h r , f = 17.093 c m / h r , and
tion. Finally, the relationship between water potential c 0

k = 15.68 h r . This estimate is plotted in Fig. 3.18 as


- 1

and soil properties is hysteretic, further complicating


curve B. Curve A is t h e infiltration curve of Fig. 3.17.
the solution of the equations.
N o t e that even t h o u g h curve Β is the best overall fit to
Horton's Equation curve A, t h e final infiltration rate is overestimated by
about 0.4 c m / h r .
Because of the difficulty of using theoretically based
By constraining f to b e 1.00 c m / h r and estimating
equations to describe the infiltration process, a great
c

/ a n d k, the results are 16.44 c m / h r and 13.54 h r " , 1

many empirical relationships have b e e n proposed. Hor-


0

respectively. This result is plotted as curve C in


ton (1940) found that an equation of the form
Fig. 3.18 and would b e preferred for hydrologic design
since it fits t h e d a t a b e t t e r for long times or for events
/(0-/c + (/o-/c)*-*' (316) of m o r e t h a n an h o u r or so in length.
If o n e can estimate / from data, t h e o t h e r p a r a m e ­
fit experimentally decreasing infiltration rates as a
c

ters of Eq. (3.16) can be estimated by simple regression


function of time. In this equation, / ( / ) is t h e infiltra­
by transforming t h e equation to
tion rate for any time, t\ f and / are the final and
c 0

initial infiltration rates, respectively; and A: is a mea­


/ ( ' ) - / c = ( / 0 - / c ) e - * '
sure of the rate of decrease in the infiltration rate.
Horton's equation requires knowledge of t h r e e soil and taking natural logarithms to get
parameters, / , / , and k.
ln[/(0-/ ]
0 c

A difficulty with t h e H o r t o n equation is that it makes


c -ln(/o-/c)-*'-
infiltration rate a function of time and does not ac­
count for variations in rainfall intensity. T h e equation Now a linear regression of l n [ / ( f ) - / J versus t yields
has no provision for a recovery of infiltration capacity — A: as its slope a n d / as e x p ( a ) + / , w h e r e a is the
0 c

during periods of low or n o rainfall. intercept of the regression.


T h e r e are not general tables or guidelines for select­ T h e H o r t o n equation really suffers from the same
ing values for the t h r e e p a r a m e t e r s of H o r t o n ' s equa­ difficulties as Eq. (3.15) in that spatially nonhomoge-
tion. Occasionally, locally derived data are used. If n e o u s soils require spatial variability in the p a r a m e t e r
infiltration data are available as from an infiltrometer values. F u r t h e r m o r e , if the soil is n o n h o m o g e n e o u s
study, the p a r a m e t e r s of Eq. (3.16) that best describe with d e p t h , i.e., has a restricting layer at some shallow
the infiltration data can be d e t e r m i n e d . If n o n e of the d e p t h , t h e infiltration rate will not smoothly decrease,
p a r a m e t e r s are known, nonlinear regression techniques but will have a r a t h e r abrupt d r o p in infiltration r a t e as
may be used. T o illustrate the application of nonlinear the wetting front reaches the restricting layer.
58 3. Rainfall-Runoff Estimation in Storm Water Computations

Table 3.9 Infiltration Constants for Bluegrass Turf Table 3.10 Infiltration Calculations Based on Horton
(Terstriep and Stall, 1974) and Holtan Models

HSG A Β C D Horton Holtan


/(hr) y(0(cm/hr) fl
F (mm)»
p / (cm/hr)'

/« iph 1 0.5 0.25 0.1


0.0 12.7 76.0 12.4
fc mm/hr 25.4 12.7 6.3 2.50
0.2 8.7 51.2 7.4
A iph 10 8 5 3
0.4 6.0 36.4 4.8
/o mm/hr 254 203 127 76
0.6 4.3 26.7 3.4
k hH 2 2 2 2
0.8 3.0 20.0 2.4
in. 6 4 3 2
1.0 2.3 15.2 1.9
mm 152 102 76 51
F
P
1.2 1.7 11.4 1.5
η 1.4 1.4 1.4 1.4
1.4 1.4 8.5 1.2
a in./hr 1.0 1.0 1.0 1.0
1.6 1.1 6.1 1.0
a mm/hr 0.274 0.274 0.274 0.274
1.8 1.0 4.1 0.8
2.0 0.9 1.7 0.7

"Based on Eq. (3.16) w i t h / = 0.63 c m / h r , / = 12.7 cm/hr,


c 0

For bluegrass turf, Terstriep a n d Stall (1974) recom­ * = 2hr . 1

m e n d e d values for t h e constants of H o r t o n ' s equation F (t+\) = F (t)-f(t)dt.


h
p p

as shown in Table 3.9. In this table, H S G d e n o t e s t h e B a s e d on Eq. ( 3 . 1 7 ) with f = 0 . 6 3 cm/hr, η = 1.4,


c
c

a = 0.274.
SCS Hydrologic Soil G r o u p designation explained later
in this chapter. T h e values for f shown in t h e table
c

exceed t h e water transmission rates for t h e H S G s .


T h e s e rates a r e given in Table 3.15 for t h e various r e c o m m e n d that a b e taken as 1.0 for / in iph (0.274
H S G s . For soils having less vegetative protection t h a n for / in millimeters p e r hour) a n d η b e taken as 1.4.
provided by a bluegrass turf, values for f less t h a n c
F o r conditions o t h e r t h a n a bluegrass turf, t h e values
those shown in t h e table should be used. of Table 3.9 would have to b e adjusted. N o t e that F is p

t h e unfilled soil w a t e r storage capacity. T h e values in


Holtan's Equation
T a b l e 3.9 a r e for an initially dry soil. A s soon as
Holtan (1961) has advanced an empirical infiltration infiltration begins, F starts decreasing since t h e infil­
p

equation based on t h e concept that t h e infiltration rate trated water fills some of t h e originally unfilled soil
is proportional to t h e unfilled capacity of t h e soil t o w a t e r storage capacity. F decreases continuously as
p

hold water. T h e Holtan model for infiltration is long as infiltration exceeds drainage from t h e soil p r o ­
/=aF "+/ , (3.17) file.
p c

where / is t h e infiltration rate, f is t h e final infiltra­


c

tion rate, F is t h e unfilled capacity of t h e soil t o store


p
E x a m p l e P r o b l e m 3 . 7 . Holtan and Horton
water, a n d a a n d η are constants. T h e exponent η has infiltration
been found to be about 1.4 for many soils. T h e value of
F ranges from a maximum of t h e available water
p Calculate the infiltration curve for a Bethany soil under
capacity ( A W C ) to zero. T h e A W C is a m e a s u r e of t h e bluegrass turf using both the Horton and Holtan equations.
ability of a soil to store water. Values for A W C a r e
Solution: From Appendix 3B the Bethany soil is found to
given for many soils in an Agricultural R e s e a r c h Ser­
be in HSG C. Based on parameters estimated from Table
vice publication (U.S. D e p a r t m e n t of Agriculture, 3.9, infiltration can be computed based on either Eq. (3.16)
1968). or (3.17). Table 3.10 shows the calculations that are illus­
T h e Holtan model for infiltration has t h e advantage trated for a time of 1 hr. The parameters of the Horton
over t h e H o r t o n model in that it has a m o r e physical equation are f = 6.3 m m / h r , f = 127 m m / h r , and k = 2
c 0

basis a n d can describe infiltration a n d t h e recovery of h r . The parameters of the Holtan equation are f = 6.3
- 1
c

infiltration capacity during periods of low or n o rain­ m m / h r , η = 1.4, and a = 0.274. Horton's equation gives
fall.
/ ( < ) = / c + (fo-fc)^' = 6-3 + (127 - 6 3 ) e ^
Table 3.9 contains values for F as r e c o m m e n d e d by
p

Terstriep a n d Stall (1974) for bluegrass turf. They also = 23 m m / h r = 2.3 c m / h r .


Abstractions From Precipitation 59

Θ
A Theory
Β Horton ~~7K >
C Holtan

WETTING FRONT

^^^^^^^^
~H®iF~
-ΔΘ-
- θ β -

1.0 Figure 3.20 Schematic for G r e e n - A m p t infiltration.


Time ( h o u r s )

Figure 3.19 Infiltration comparison.

Skaggs a n d Kahleel (1982) discuss t h e determination of


p a r a m e t e r s of t h e G r e e n a n d A m p t model.
F for the Holtan equation is given as F for the previous
p p
Generally t h e p o n d e d d e p t h , d, is small and may be
time minus the infiltration during the previous time incre­ neglected in Eq. (3.18). T h e change in water content
ment or 20.00 - 0.2(24) = 15.2 mm. Thus the Holtan equa­ across the wetting front, Δ 0 , d e p e n d s on the initial
tion gives w a t e r content, θ \ t h e thoroughly drained or residual
{

w a t e r content, 0 ; t h e effective saturation, s , a n d t h e


r c

/= aF; + / = 0.274(15.2) 14
+ 6.3 total porosity, η. T h e relationship is
c

= 19 m m / h r = 1 . 9 c m / h r . Λ0 = 0 - e . = 0 - 5 0 = ( 1 - 5 ) 0 ,
e e e e e e

w h e r e 0 is t h e effective porosity given by 0 = η - 0


e e r

The results of the calculations are plotted in Fig. 3.19. (Chow et al, 1988). Rawls et al (1983) present values
Special note should be made of the calculations applying to for η, 0 , φ, a n d Κ as shown n Table 3.11.
e
Eqs. (3.17). The F term must be continually decreased by
p
By noting that F(t) = L Ad or L = F ( / ) / A 0 and
the volume of infiltration that has taken place. In Fig. 3.19,
f f

taking d as zero, E q . (3.18) can b e expressed as


curve A is from Fig. 3.17, curve Β is from the Horton model,
and curve C is from the Holtan model. ψΑΘ
+ 1 (3.19)
n o

Green-Ampt. Equation
In 1911, G r e e n a n d A m p t (1911) developed an a p ­ Table 3.11 Green and Ampt Infiltration Parameters
proximate infiltration model based on Darcy's law. (Rawls etal, 1983)°
They assumed vertical flow, a uniform w a t e r content, a
Soil texture η ψ (cm) Κ (cm/hr)
sharp boundary between the w e t t e d soil zone, a n d t h e
soil zone unaffected by infiltration a n d t h a t w a t e r Sand 0.437 0.417 4.95 11.78
movement occurs as " p i s t o n " flow or " s l u g " flow. T h e Loamy sand 0.437 0.401 6.13 2.99
G r e e n and A m p t model can b e " d e r i v e d " by applying Sandy loam 0.453 0.412 11.01 1.09
Eq. (3.15a) to the situation depicted in Fig. 3.20: Loam 0.463 0.434 8.89 0.34
/ = / ^ ( c h a n g e in p o t e n t i a l ) / ( d i s t a n c e ) Silt loam 0.501 0.486 16.68 0.65
or Sandy clay loam 0.398 0.330 21.85 0.15
Clay loam 0.464 0.309 20.88 0.10
f = K(d + L +*)/L ,
f f (3.18)
Silty clay loam 0.471 0.432 27.30 0.10
w h e r e / is the infiltration rate that varies with time as Sandy clay 0.430 0.321 23.90 0.06
t h e wetting front advances into t h e soil, Κ is t h e Silty clay 0.479 0.423 29.22 0.05
hydraulic conductivity of the w e t t e d soil p a r t of the soil
Clay 0.475 0.385 31.63 0.03
profile, L is the d e p t h of t h e wetting front, ψ is t h e
f

pressure head for wetting at the wetting front, a n d d is


"Rawls et al. (1983) contains more information on these parameters
t h e d e p t h of ponding of water on t h e soil surface. including their standard deviation and values for various soil horizons.
60 Chapter 3. Rainfall-Runoff Estimation in Storm Water Computations

where F(t) is the cumulative infiltration at time t. T h e Table 3.12 Green-Ampt Infiltration Calculations
cumulative infiltration is found by integration of Eq.
(3.19) as Fit) (cm)* t (hr)* t (min) £
/ (cm/my

0.0 0.000 0.000 0.000


F(t) = Kt + ψΔΘΙη 1 + (3.20) 0.1 0.006 0.360 8.356
φΔΘ
0.2 0.024 1.431 4.228
Equations (3.19) and (3.20) apply for the case w h e r e 0.3 0.053 3.193 2.852
the p o n d e d d e p t h is negligible. If this is not the case, φ 0.4 0.094 5.633 2.164
should be replaced by φ + d. Since Eq. (3.20) cannot
0.5 0.146 8.734 1.751
be solved explicitly for F(t), iteration is required. A
0.6 0.208 12.481 1.476
trial value for F(t) is substituted into the right-hand
0.7 0.281 16.860 1.279
side of the equation, which is t h e n c o m p a r e d to the
0.8 0.364 21.856 1.132
left-hand side. This process is r e p e a t e d until agree­
ment between the two values is obtained. A good first 0.9 0.458 27.456 1.017

estimate is Kt for F(t). Possibly an easier way to 1.0 0.561 33.648 0.926
calculate / ( / ) from Eq. (3.19) is to solve Eq. (3.20) for t 1.1 0.674 40.418 0.851
for various values of F(t). F{t) can t h e n be used in Eq. 1.2 0.796 47.753 0.788
(3.19) to d e t e r m i n e / ( / ) for the corresponding t. 1.3 0.927 55.643 0.735
1.4 1.068 64.075 0.690
1.5 1.217 73.039 0.650
Example Problem 3.8. Green-Ampt infiltration 1.6 1.375 82.523 0.616
1.7 1.542 92.517 0.586
Calculate the infiltration curve for a silty clay loam at 30% 1.8 1.717 103.010 0.559
effective saturation.
1.9 1.900 113.994 0.535
Solution: Table 3.12 illustrates the approach. Values for
0 , ψ, and Κ are taken from Table 3.11 as 0.432, 27.3 cm, 2.0 2.091 125.458 0.513
e

and 0.1 cm/hr, respectively. F(t) values are assumed. The


calculations are illustrated for F(t) = 1 cm. Δ0 is given by "Assumed.
*From Eq. ( 3 . 2 0 ) with S c 0.3. θ β = 0.432, ψ
Δ0 = (1 - s ) 0 = (1 - 0.3)0.432 = 0.302.
e e
27.3 cm, and = 0.1 cm/hr.
c
Column 2 χ 60 min/hr.
t is computed by rearranging Eq. (3.20) to ''From Eq. (3.19).

F(t) - φΔΘ ln(l + F(O/<AA0)


t =
Κ
1 - 27.3(0.302)ln(l + 1/27.3(0.302)) rainfall r a t e . Sometimes a two-stage constant loss rate
t = — = 0.561 hr. is used. F o r example, for t h e D e n v e r region, Wright-
0.1
McLaughlin E n g i n e e r s (1969) p r o p o s e d that t h e fol­
From Eq. (3.19), lowing constant infiltration loss rates be used in the
' φΔΘ ' 27.3(0.302) absence of m e a s u r e d data:
Κ — — + 1 = 0.1 + 1 = 0.926 c m / h r .
W ) J 1
Storm frequency First half hour Remainder
A plot of column 4 versus column 2 or 3 yields the infiltration
curve. 2 to 5 years 1 iph ϊ iph
2.5 cm/hr 1.3 cm/hr
10 to 100 years 1 iph Ί iph
1.3 cm/hr 1.3 cm/hr
Φ Index
Over the years many other empirical infiltration
models have been proposed. Because of t h e general Wright-McLaughlin E n g i n e e r s (1969) went on to urge
lack of values for the p a r a m e t e r s for these various that each a r e a being considered b e field tested for
models and the nonhomogeneity of soils, these models infiltration rates a n d that t h e m e a s u r e d values be used
have not b e e n widely applied in storm water m a n a g e ­ in preference to those shown in t h e above table. Often
ment. Instead, a steady infiltration loss rate from t h e t h e constant infiltration loss r a t e is t e r m e d t h e Φ
rainfall rate has been defined to obtain the effective index.
Abstractions From Precipitation 61

Effective Rainfall This cyclic interplay between rainfall, surface storage,


In runoff calculations, effective rainfall is t h e rainfall and effective rainfall may b e r e p e a t e d many times
that runs off. Effective rainfall r e p r e s e n t s t h e supply of during a natural rainfall event with fluctuating rainfall
water t o t h e runoff process. T h e volume of effective intensities.
rainfall is equal t o t h e volume of storm water runoff.
Effective rainfall is also known as rainfall excess since
it represents t h e a m o u n t of rainfall in excess of a b ­ Example Problem 3.9. Effective rainfall from the
stractions o r losses from runoff. Considering only t h e Holtan equation
abstractions of surface storage a n d infiltration, t h e
generation of effective rainfall c a n b e envisioned as Estimate the effective rainfall rate based on a constant
follows. All rainfall goes into infiltration until t h e rain­ surface detention of 0.30 in. (7.5 mm) typical of a sod £nd an
fall rate exceeds t h e infiltration rate. A t that point, infiltration curve for a soil in HSG C based on the Holtan
surface storage begins t o fill. A s long as t h e rainfall equation for the rainfall given in Table 3.3.
rate exceeds t h e infiltration rate, surface storage will Solution: The results are tabulated in Table 3.13. The first
continue t o fill. Runoff will begin o n those parts of the two columns defining the rainfall are from Table 3.3. Several
points based on Table 3.13 should be noted. First, the effec­
catchment where t h e surface storage is filled. W h e n
tive rainfall is zero unless surface storage is filled and the
the rainfall rate drops below the infiltration r a t e , water
rainfall rate exceeds the infiltration rate. Second, the actual
in surface storage will begin infiltrating. W h e n t h e
infiltration equals the rainfall rate until the rainfall rate
surface storage capacity at a location is n o longer filled, exceeds the potential infiltration rate as determined from the
no additional water becomes available for surface Holtan relationship. Third, the actual infiltration may exceed
runoff; thus, t h e generation of effective rainfall ceases. the rainfall rate if surface storage is not zero. Fourth, the

Table 3.13 Calculation of Effective Rainfall Based on Holtan Equation

Rain Surface Potential


Time depth detention / infiltration Effective
V
(min) a
(in.) 6
(in.) c

(in.) rf 0ph)< volume ( i n . / (in.)* rain (ίη.) Λ

0
15 0.10 0.00 3.00 4.91 1.23 0.10 0.00
30 0.12 0.00 2.90 4.69 1.17 0.12 0.00
45 0.17 0.00 2.78 4.43 1.11 0.17 0.00
60 0.26 0.00 2.61 4.08 1.02 0.26 0.00
75 0.61 0.00 2.35 3.56 0.89 0.61 0.00
90 1.91 0.30 1.74 2.42 0.61 0.61 1.00
105 0.36 0.30 1.13 1.44 0.36 0.36 0.00
120 0.20 0.26 0.77 0.94 0.24 0.24 0.00
135 0.14 0.23 0.53 0.66 0.17 0.17 000
150 0.11 0.22 0.36 0.49 0.12 0.12 000
165 0.09 0.21 0.24 0.39 0.10 0.10 0.00
180 0.08 0.21 0.14 0.31 0.08 0.08 0.00

Total 4.15 0.21 2.94 1.00

F r o m Table 3.3.
a

*From Table 3.3


c
First 0.3 in. of rain in excess of infiltration goes to detention storage.
F (i)=/ A-i)-/ (/-^).
< /
p
R
P
r
a

'From Eq. (3.17).


/Column 5 x 0.25 hr.
8
Actual incremental infiltration volume.
''Rainfall depth in excess of actual infiltration depth. Note: At end of storm, detention +
total infiltration + total effective rainfall = total rainfall.
62 Chapter 3. Rainfall-Runoff Estimation in Storm Water Computations

value of F is decreased by the actual, not the potential,


p The time increment ending at 135 min has a rainfall
infiltration volume. Consider the calculations for the time of volume of 0.14 in. and a potential infiltration volume of 0.17
60 min. The rainfall depth is 0.26 in. during the 15-min time in. Thus, 0.03 in. are taken from surface storage and no
increment. There is no water in surface storage since the effective rainfall is generated.
potential infiltration rate to this time has always exceeded At the end of the storm, the sum of surface storage, actual
the rainfall rate. F of column 4 is computed from F (60) =
p p infiltration, and effective rainfall must equal the total rainfall
F (45)—actual infiltration during the preceding time interval
p volume:
given in column 7 = 2.78 - 0.17 = 2.61.
Column 5 represents the potential infiltration rate if rain­ 0.21 + 2.94 -I- 1.00 = 4.15.
fall is not limiting. It is determined from Eq. (3.17). The
equation parameters are determined from Table 3.9 as f = c

0.25 iph, F (initial) = 3.00 in., η = 1.4, and a = 1:


p

/ = aF; + f = 1.0(2.61)* 4
+ 0.25 = 4.08 iph.
Example Problem 3.10. Effective rainfall from the
c
Green-Ampt equation
The potential infiltration volume is given by the product of
the potential rate times the time increment or 4.08 X 0.25 = Calculate the effective rainfall pattern for the rain of
1.02 in. Since the rainfall volume during the time increment Table 3.3 using the Green-Ampt model for a silt loam soil
was only 0.26 in., the actual infiltration volume is 0.26 in. F p
with 30% effective saturation. The maximum surface storage
for the next time increment is thus reduced by 0.26 in. is 0.75 cm.
The time increment ending at 90 min is the first to pro­ Solution: The parameters for the Green-Ampt equation
duce any effective rainfall. For this time increment, the [Eq. (3.19)] from Table 3.11 are 0 - 0.486, ψ = 16.68 cm,
e

rainfall volume is 1.91 in. The potential infiltration volume is and Κ = 0.65 cm/hr. The change in water content as a
only 0.61 in. Thus 0.30 in. goes to surface storage, 0.61 in. result of the passing of the wetting front is calculated as
goes to infiltration, and the remainder goes to effective
rainfall. Δ0 = (1 - J ) 0 = (1 - 0.3)0.486 = 0.34.
e e

Table 3.14 Calculation of Effective Rainfall: Green-Ampt Approach

AR AFa
F
p S AR C

(min) (cm) a
(cm/hr)* (cm) c
(cm)* (cm) e
(cm/ (cm)*

0 0.00 0.00 0.00 0.00


15 0.25 h h 0.25 0.25 0.00 0.00
30 0.30 15.40 3.85 0.30 0.55 0.00 0.00
45 0.43 7.36 1.84 0.43 0.98 0.00 0.00
60 0.66 4.42 1.10 0.66 1.64 0.00 0.00
75 1.55 2.90 0.73 0.73 2.37 0.75 0.07
90 4.85 2.21 0.55 0.55 2.92 0.75 4.30
105 0.91 1.91 0.48 0.48 3.40 0.75 0.43
120 0.51 1.74 0.43 0.43 3.83 0.75 0.08
135 0.36 1.61 0.40 0.40 4.23 0.71 0.00
150 0.28 1.52 0.38 0.38 4.61 0.61 0.00
165 0.23 1.45 0.36 0.36 4.97 0.48 0.00
180 0.20 1.39 0.35 0.35 5.32 0.33 0.00

Totals 10.53 5.32 4.88

a
From Table 3.3 converted to centimeters.
^Potential infiltration rate from Green-Ampt equation.
c
Potential infiltration volume / Δ / . ρ

^Actual infiltration volume.


F (t) = F (t-\) + aF .
e
p p a

^Surface storage.
^Effective rainfall.
A
Very large.
Abstractions From Precipitation 63

The infiltration rate, / , and cumulative infiltration are Table 3.15 Definition of SCS Hydrologic Soil Groups
related by Eq. (3.19) as (Soil Conservation Service, 1986)

" ΨΔΘ
Group A soils have low runoff potential and high infiltration rates even
+ 1 = 0.65
F when thoroughly wetted. They consist chiefly of deep, well to excessively
drained sands or gravels and have a high rate of water transmission
Table 3.14 shows the resulting calculations. The rainfall of (greater than 0.30 in./hr).
Table 3.3 is converted to centimeters. The / of column 3
represents the potential infiltration rate calculated from the Group Β soils have moderate infiltration rates when thoroughly wetted
and consist chiefly of moderately deep to deep, moderately well to well
above equation using F from the previous time increment.
p
drained soils with moderately fine to moderately coarse textures. These
AF of column 4 is the potential infiltration volume calcu­
p
soils have a moderate rate of water transmission (0.15-0.30 in./hr).
lated as / Δ ί . AF represents the actual infiltration volume
ρ a

for the time increment. The actual infiltration rate may be Group C soils have low infiltration rates when thoroughly wetted and
less than the potential rate. Early in the storm, the potential consist chiefly of soils with a layer that impedes downward movement of
infiltration rate exceeds the rainfall rate so the actual infil­ water and soils with moderately fine to fine texture. These soils have a
tration volume is limited to the rainfall volume. F at a p
low rate of water transmission (0.05-0.15 in./hr).
particular time increment is equal to F from the previous
p

time increment plus AF for the current time increment. Group D soils have high runoff potential. They have very low infil­
e
tration rates when thoroughly wetted and consist chiefly of clay soils with
In the time increment from 60 to 75 min, the potential
a high swelling potential, soils with a permanent high water table, soils
infiltration rate falls below the rainfall rate so some surface with a claypan or clay layer at or near the surface, and shallow soils over
storage and rainfall excess is generated. Since at this time the nearly impervious material. These soils have a very low rate of water
surface storage is empty, the rainfall is divided with 0.73 cm transmission (0-0.05 in./hr).
going to satisfy infiltration, 0.75 cm going to surface storage,
and 0.07 cm becoming rainfall excess since the surface stor­ Some soils in the list are in group D because of a high water table that
age is filled. The process is continued to the end of the creates a drainage problem. Once these soils are effectively drained, they
storm. It can be seen that for the time interval from 120 to are placed in a different group. For example, Ackerman soil is classified
135 min, AF is 0.40 cm while the rainfall is only 0.36 cm. as A/D. This indicates that the drained Ackerman soil is in group A and
the undrained soil is in group D.
Thus 0.04 cm of water is taken from surface storage. A F
cannot exceed the rainfall for the time increment plus the
depth of water in surface storage.

Curve Number Approach 0.25 before any runoff is g e n e r a t e d . T h u s a rainfall


T h e Soil Conservation Services (SCS) (1972, 1985) of volume of 0.25 must fall before runoff is initiated. It
the U.S. D e p a r t m e n t of Agriculture combines infiltra­ should be n o t e d that Eq. (3.21) is a runoff equation
tion losses with initial abstractions and estimates rain­ and not an infiltration equation. Using it as an infiltra­
fall excess or equivalently the runoff volume by the tion equation can lead to errors.
relationship T h e SCS has classified m o r e than 4000 soils into four
hydrologic soil groups ( H S G ) according t o their mini­
(P - 0.25) m u m infiltration rate obtained for b a r e soil after pro­
Ρ > 0.25, (3.21) longed wetting. Listings of soils and their hydrologic
Ρ + 0.85
soil group may b e found in a n u m b e r of SCS publica­
where Q is the accumulated runoff volume or rainfall tions (SCS, 1972, 1985). Local SCS offices can provide
excess, Ρ is the accumulated precipitation, and 5 is a information on local soils. A p p e n d i x 3B contains the
parameter, sometimes called a maximum soil water H S G classification for a large n u m b e r of soils. T h e four
retention p a r a m e t e r , given by hydrologic soil groups are d e n o t e d by the letters A, B,
C, and D . Table 3.15 contains the definition of these
1000
four soil groups. In determining the hydrologic soil
5 = — -10 ( β , P, Sin.) group, consideration should be given to compaction via
(3.22)
heavy equipment, exposure of subsoil, etc. T h e s e things
25400 may cause a soil that would normally be in one soil
5 = - 254 (Q,P,S mm), g r o u p to be less pervious and thus behave as another
where C N is known as the curve number. Curve num­ soil group. For example, a normal Β soil may behave as
ber tables are available from a n u m b e r of sources. a C soil. W h e n the soil is greatly disturbed but not
Table 3.16 and Appendix 3C contain representative significantly compacted, an estimate of the H S G can be
values. Equation (3.21) indicates that Ρ must exceed m a d e based on the texture of the exposed or surface
64 Chapter 3. Rainfall-Runoff Estimation in Storm Water Computations

soil as follows (Brakensiek and Rawls, 1983): bers includes factors in addition to infiltration, it is, in
fact, a n o n - H o r t o n i a n a p p r o a c h to runoff estimation.
HSG Soil texture As with any d e t e r m i n a t i o n of soil p a r a m e t e r s , spe­
cial attention must b e given to situations w h e r e imper­
A Sand, loamy sand, or sandy loam
vious areas such as streets, buildings, and parking lots
Β Silt loam or loam a r e present. O n e must consider t h e extent of the
C Sandy clay loam impervious area a n d the m a n n e r in which flow from
D Clay loam, silty clay loam, sandy clay, t h e impervious a r e a reaches a drainage channel. T h e
silty clay, or clay extent of t h e impervious a r e a is generally expressed as
a p e r c e n t a g e of t h e total area. T h e p a t h the flow uses
to reach a channel is often specified by stating t h e
T h e curve n u m b e r of an area indicates the runoff impervious area is either "directly c o n n e c t e d " or "indi­
potential of the area. Table 3.16 is a summary of CNs rectly c o n n e c t e d . " Directly connected impervious areas
for various land-use and treatment combinations. A p ­ have flow that travels directly to the drainage system
pendix 3C is a more extensive table of CNs. Impervious (channels, sewers, gutters, etc.) or occurs as concen­
areas and water surfaces are assigned a C N of 9 8 - 1 0 0 . trated flow over a pervious area. Indirectly connected
Recognizing that abstractions from rainfall d e p e n d on impervious areas discharge flow in a diffuse m a n n e r as
the antecedent conditions that exist at the time a overland flow o n t o a pervious area presenting the
rainstorm occurs, three antecedent conditions have runoff an opportunity for infiltration into the pervious
been defined. T h e curve numbers given in Table 3.16 area.
are for antecedent condition II, which is based on T h e CNs of T a b l e 3.16 were developed for typical
median values for C N taken from sample rainfall and land-use relationships a n d specific assumed percent­
runoff data. A n t e c e d e n t condition I is used w h e n t h e r e ages of impervious area. T h e assumptions w e r e that
has been little rainfall preceding the rainfall in ques­ t h e pervious u r b a n area corresponds to a p a s t u r e in
tion and condition III is used where t h e r e has b e e n good hydrologic condition and that t h e impervious area
considerable rainfall prior to the rain in question. was directly connected with a C N of 98.
Curve numbers for antecedent conditions I or III can If all of the impervious area is directly connected but
be estimated by (Chow et al., 1988) t h e pervious area p e r c e n t a g e or t h e pervious land-use
assumptions are not met, t h e following relationship
4.2 C N ( I I ) based on SCS T R - 5 5 (Soil Conservation Service, 1986)
CN(I) =
10 - 0.058 C N ( I I ) can b e used to calculate a composite C N
.(3.23)
23 C N ( I I ) CN = C N
C p + (P i m p /100)(98 - CN ), p (3.24a)
CN(III) =
10 + 0.13 C N ( I I ) '
w h e r e C N is t h e composite C N , C N is the C N for t h e
C p

pervious area, and P is t h e p e r c e n t a g e impervious.


i m p
where CN(I), CN(II), and CN(III) represent curve
If runoff from t h e u n c o n n e c t e d impervious a r e a is
numbers for antecedent conditions I, II, and III, re­
spread out over a pervious a r e a as sheet flow, t h e
spectively.
composite C N should b e c o m p u t e d from
Originally, the CN values were assigned by plotting
observed runoff versus measured rainfall for a n u m b e r CN = C N
C p + (P i m p /100)(98 - CN )(l - p 0.5/?)
of experimental plots scattered t h r o u g h o u t t h e U.S. (3.24b)
T h e CNs were then correlated with land use. T h e
terms good condition or poor condition in C N tables for P i m p less t h a n 3 0 % a n d from Eq. (3.24a) for P i m p

refer to the relative runoff potential. A n area in good greater t h a n 3 0 % . R r e p r e s e n t s the ratio of uncon­
hydrologic condition would have higher infiltration nected impervious area to total impervious area.
rates and lower runoff rates than an area in poor It is important to consider all impervious areas and
condition. Again, note that the C N approach is a o t h e r areas of high runoff potential such as soils in
runoff approach and not an infiltration approach. Cer­ H S G D n o m a t t e r how small since they p r o d u c e high
tainly infiltration is a factor affecting runoff, but so is rates of runoff p e r unit of rainfall.
quick return flow and initial abstractions. Combining A n area-weighted C N for mixed land uses and H S G s
the C N approach with infiltration approaches such as can b e c o m p u t e d from
minimum retention rates carries the C N concept b e ­
M/CN,-
yond its original intent and beyond the data on which CN = (3.25)
the CNs are based. Since the derivation of curve num­ M,
Abstractions From Precipitation 65

Table 3.16 Runoff Curve Numbers for Selected Land Uses (Soil Conservation Service, 1986)

Hydrologic soil group

Land use description A Β C D

Cultivated land*
Without conservation treatment 72 81 88 91
With conservation treatment 62 71 78 81
Pasture or range land
Poor condition 68 79 86 89
Good condition 39 61 74 80
Meadow
Good condition 30 58 71 78
Wood or forest land
Thin stand, poor cover, no mulch 45 66 77 83
Good cover* 25 55 70 77
Open Spaces, lawns, parks, golf courses, cemeteries, etc.
Good condition (grass cover on 75% or more of the area) 39 61 74 80
Fair condition (grass cover on 50 to 75% of the area) 49 69 79 84
Commercial and business areas (85% impervious) 89 92 94 95
Industrial districts (72% impervious) 81 88 91 93
Residential'
Average lot size Average percentage impervious^

β acre or less 65 77 85 90 92
J acre 38 61 75 83 87
5 acre 30 57 72 81 86
j acre 25 54 70 80 85
1 acre 20 51 68 79 84
Paved parking lots, roofs, driveways, e t c / 98 98 98 98
Streets and roads
Paved with curbs and storm sewers'' 98 98 98 98
Gravel 76 85 89 91
Dirt 72 82 87 89

a
F o r a more detailed description of agricultural and land use curve numbers refer to "National
Engineering Handbook," Sect. 4, "Hydrology" Chap. 9, 1972.
''Good cover is protected from grazing, litter, and brush cover soil.
'Curve numbers are computed assuming the runoff from the house and driveway is directed toward the
street with a minimum of roof water directed to lawns where additional infiltrations could occur.
''The remaining pervious areas (lawn) are considered to be in good pasture condition for these curve
numbers.
e
In some warmer climates of the country a curve number of 95 may be used.

where CN, corresponds to the appropriate C N for the In some cases, a straightforward weighting of infil­
part of the catchment having area A r tration indices, whether they are CNs or Φ indices or
Once the p r o p e r CN is obtained, Eq. (3.21) and some o t h e r index, may not b e appropriate. Such would
(3.22) can b e used t o estimate t h e accumulated rainfall be the case when t h e r e is a large difference in the
excess as a function of total accumulated rainfall. Fig­ indices a n d t h e areas with a high runoff potential are
ure 3.21 has b e e n p r e p a r e d to simplify t h e solution of directly connected to the drainage system. In such
Eq. (3.21). cases, runoff from the nearly impervious area may
66 Chapter 3. Rainfall-Runoff Estimation in Storm Water Computations

represent a significant part of the total runoff and 77. Thus, the weighted CN is
should not be diminished by averaging with a m o r e CN = 0.35(91) + 0.35(77) + 0.30(58) = 77.
pervious area. T h e nearly impervious areas may also
respond very quickly to rainfall producing runoff well In this example, it is assumed that the various soils are
randomly and somewhat uniformly scattered throughout the
before the pervious areas. If the nearly impervious area
watershed and an unknown antecedent condition exists. The
drains across a pervious area, then some of the water
total runoff from the 4.04 in. (103 mm) of rain is computed
from the nearly impervious area may infiltrate into the
from Eq. (3.21) and (3.22) as
pervious area. In this case, the assignment of a large
percentage runoff from the nearly impervious a r e a 1000 1000
S = 10 = — — - 10 = 2.99 in.
could overestimate the actual runoff. T h e losses from CN 77
the nearly impervious area runoff would d e p e n d on the or
infiltration rate of the intervening pervious area and 25400 25400
the opportunity for infiltration. S = ^ T T - - 254 = — 254 = 75.9 mm
CN 77
and

(P - 0.2S) 2
(4.04 - 0.2 X 2.99)2

Example Problem 3 . 1 1 . Effective rainfall based on Q = —ζ = —τ—. — = 1.84 in.


Ρ + 0.85 4.04 + 0.8 X 2.99
SCS curve number
or
Assume the rainstorm of Table 3.5 falls on a watershed (P - 0.2S) 2
(102.6 - 0.2 X 75.9)2

that is 35% bare soil in hydrologic soil group D and has 30% Q = — ζ — = „ . = 46.8 mm.
of its soils in hydrologic soil group Β under grass and 35% in Ρ + 0.85 102.6 4- 0.8 X 75.9
hydrologic soil group C under forest. Determine the effective Table 3.17 shows the calculations required to arrive at the
rainfall pattern. effective rainfall pattern. The calculations can be illustrated
Solution: The appropriate CN can be calculated by refer­ by considering the time interval from 11.75 to 12.00 hr. The
ring to Table 3.16 and noting that the bare soil area has a CN accumulated precipitation to 12 hr is obtained by summing
of 91, the Β soil has a CN of 58, and the C soil has a CN of the entries in column 2 up to 12 hr as 3.12 in. Using Ρ = 3.12
Runoff Estimation 67

Table 3.17 Calculation of Effective Rainfall Using contains information on runoff volume as the area
CN Approach u n d e r the hydrograph and p e a k flow rates as the
maximum flow or p e a k of t h e runoff hydrograph as
Accumulated Incremental
well as a complete time history of flow. Hydrographs of
Incremental Accumulated effective effective
rainfall* rainfalF fainfall""
storm water runoff are often required in the design of
rainfall*
Time* (in.) (in.) (in.) (in.) storm water-retarding structures and sediment control
p o n d s . If only an estimate of t h e p e a k flow or runoff
10.50 0.00 0.00 0.00 0.00 volume is n e e d e d , it may not b e necessary to develop
10.75 0.10 0.10 0.00 0.00 the entire runoff hydrograph.
11.00 0.11 0.21 0.00 0.00
11.25 0.15 0.36 0.00 0.00
Storm Water Runoff Volume
11.50 0.17 0.53 0.00 0.00
11.75 0.71 1.24 0.11 0.11 In this section, reference to runoff m e a n s storm
1.04
w a t e r runoff or t h e runoff occurring during and imme­
12.00 1.88 3.12 1.15
diately following a major precipitation event. T h e vol­
12.25 0.33 3.45 1.39 0.24
u m e of storm w a t e r runoff is equal to the volume of
12.50 0.16 3.61 1.51 0.12
rainfall excess or effective precipitation. T h u s runoff
12.75 0.15 3.76 1.63 0.12
volume is rainfall minus abstractions. Any of the meth­
13.00 0.10 3.86 1.70 0.07 ods previously discussed for estimating abstractions or
13.25 0.10 3.96 1.78 0.08 rainfall losses can be used in the computation of the
13.50 0.08 4.04 1.84 0.06 volume of runoff.
T h e infiltration a p p r o a c h to estimating runoff vol­
Total 4.04 4.04 1.84 1.84 u m e consists of estimating initial abstractions and
infiltration and deducting these losses from rainfall.
"From Table 3.5. Several m e t h o d s for estimating these quantities have
^Summing Column 2. already b e e n discussed a n d illustrated in the previous
Based on Eq. (3.21) and a CN of 77 and Ρ of Column 3.
1

sections of this chapter.


differencing Column 4.
W h e n the SCS curve n u m b e r approach is used, Eq.
(3.21) and (3.22) define t h e runoff volume, Q, as a
function of the total precipitation, P, and the curve
n u m b e r , CN, for storms with durations of 24 hr or less.
in. and S = 2.99 in. in Eq. (3.21) results in a β of 1.15 in.
F o r example, if 5 in. (127 m m ) of rain falls in 24 hr on
The Q for the time increment is the difference in the
accumulated runoff at 12.00 and 11.75 hr or 1.15 - 0.11 = a watershed having a C N of 75, t h e estimated runoff
1.04 in. volume is 2.45 in. (62 m m ) . N o t e that this is an esti­
m a t e of storm water runoff and does not include base
flow g e n e r a t e d from ground w a t e r discharges that are
d e p e n d e n t on infiltrated water.

RUNOFF ESTIMATION Development of Runoff Hydrographs


Runoff is the flow resulting from precipitation events. T o this point, the generation of rainfall excess vol­
Some of this flow may occur during and immediately umes and time distributions and runoff volumes have
following the precipitation event and as such is known b e e n discussed. W h a t is often of interest in storm
as storm water runoff or storm flow. Flow occurring water m a n a g e m e n t is the hydrograph of runoff at the
between precipitation events is generally s u p p o r t e d by watershed outlet and possibly at selected points within
seepage and ground water discharge and as such is the watershed. A hydrograph is simply a plot of flow
known as baseflow. T h e r e is not a definable point or rate versus time. T h e hydrograph is a result of a
time at which storm flow ceases and base flow begins. particular effective rainfall hyetograph as modified by
O n e process gradually grades into the other. Generally basin flow characteristics. By definition, t h e volume of
on small catchments t h e magnitude of base flow is w a t e r u n d e r an effective rainfall hyetograph is equal to
quite small and is neglected in the computation of the volume of surface runoff.
storm water runoff. Most of t h e terminology used in discussing runoff
A runoff hydrograph is a continuous record of hydrographs is shown in Fig. 3.22. A n effective rainfall
streamflow over time. A complete runoff hydrograph hyetograph consisting of a single block of rainfall with
68 Chapter 3. Rainfall-Runoff Estimation in Storm Water Computations

- EFFECTIVE RAINFALL OR
RAINFALL E X C E S S HYETOORAPH

• AREA · V

Figure 3.23 Watershed showing isochrones and time area dia-


gram.

hydraulically most r e m o t e point in a basin to t h e basin


outlet.

A Conceptual Model

A conceptual m o d e l of t h e runoff process c a n b e


developed starting with t h e t i m e - a r e a diagram of a
w a t e r s h e d a n d considering only a translation of t h e
TIME
rainfall excess hyetograph t o t h e basin outlet. T h e
Figure 3.22 Hydrograph terminology.
t i m e - a r e a diagram is o b t a i n e d by drawing isochrones
(lines of constant time) o n t h e w a t e r s h e d with t h e
isochrones s e p a r a t e d in time by Δ ί (Fig. 3.23). T h e
duration D is shown in the u p p e r left part of t h e value of time on each isochrone r e p r e s e n t s the travel
figure. T h e runoff hydrograph constitutes the lower time of w a t e r from t h e isochrone to t h e basin outlet.
part of the figure. T h e areas enclosed by t h e hyeto­ This travel time is t h e sum of t h e overland flow travel
graph and by the hydrograph each represent t h e same time and channel flow travel time. T h e t i m e - a r e a
volume, K, of water. Obviously, the scale for rainfall diagram is t h e n a plot of a versus t w h e r e a is t h e
t i9 i

and flow are not the same in Fig. 3.22. T h e maximum a r e a enclosed b e t w e e n isochrone t a n d
t

flow rate on the hydrograph is t h e p e a k flow, q , while p Next visualize a block of effective rainfall of uniform
the time from the start of the hydrograph to q is t h e p intensity r and duration Δ ί falling uniformly on t h e
x

time to peak, i . T h e total time duration of t h e hydro-


p watershed. Also, envision that t h e runoff process is
graph is known as the base time, t . b simply a translation of t h e rainfall from t h e point it
T h e lag time, t is the time from t h e center of mass
L strikes t h e basin to t h e w a t e r s h e d outlet with t h e travel
of the effective rainfall to the p e a k of t h e runoff time defined by t h e t i m e - a r e a diagram. In this event,
hydrograph. It is a p p a r e n t that t h e runoff from a would b e q = 0 at t = 0, increase
x

linearly t o q = a r at t = t at which point t h e rain­


x x v

'P = ' L + D/2, (3.26) fall ceases. T h e runoff would t h e n decrease linearly to
q = 0 at t = t . Similarly, t h e runoff from a would
2 2

using this definition. Some define lag time as t h e time start at q = 0 at t = t since it takes t for t h e runoff
x x

from center of mass of effective rainfall to t h e c e n t e r of to reach t h e outlet from a , increase linearly to q =
2

mass of the runoff hydrograph. M e t h o d s for estimating a r at t = t , a n d d e c r e a s e to q = 0 at t = t . Similar


2 x 2 3

these various hydrograph p a r a m e t e r s are covered in p a t t e r n s result for t h e runoff from t h e remaining areas.
this chapter. T h e total runoff hydrograph from t h e first block of
A time p a r a m e t e r not shown in Fig. 3.22 is t h e time effective rainfall would t h e n equal t h e sum of t h e
of concentration, t . T h e time of concentration is d e ­
c
individual triangular subhydrographs as shown in Fig.
fined as the time it takes water to flow from t h e 3.24. T h e hydrograph p e a k s at t = t . 3
Runoff Estimation 69

o r,
2

o.r,

Figure 3.24 Illustration of runoff hydrograph from conceptual wa­


tershed due to a single block of short duration rainfall.

If a second block of effective rainfall of uniform


intensity r falls uniformly o n t h e basin during t h e
2

time interval t to t , the resulting runoff hydrograph


x 2

from this second block of rainfall would b e obtained in


a similar fashion with the hydrograph starting at t = t l9

peaking at t = t , and ending at t = t . T h e runoff


4 7

hydrograph for t h e two blocks of rainfall r and r x 2

would b e t h e sum of the runoff from t h e individual


blocks of rainfall. Figure 3.25 shows t h e total runoff
hydrograph from five blocks of rainfall o n t h e water­
shed shown in Fig. 3.23.
T h e ordinate for runoff # at time tj is in general
;
\ t t t, t
given by
T e w

Figure 3.25 Hydrograph from rainstorm on conceptual watershed.

r = 0 for i > m
i

Qj = Σ r
i j - i + \
a (3.27)
cii = 0 for ι > η,
The Santa Barbara Urban Hydrograph Method
where m is t h e n u m b e r of rainfall blocks a n d η is t h e T h e Santa B a r b a r a U r b a n Hydrograph ( S B U H )
n u m b e r of area blocks in t h e t i m e - a r e a diagram. m e t h o d is an example of a p r o c e d u r e that produces a
A s can b e seen from Fig. 3.25, the conceptual hydro- runoff hydrograph by routing a rainfall excess hyeto­
graph is a very steeply rising a n d falling hydrograph graph through a conceptual reservoir (Stubchaer, 1975).
with a time base only slightly longer t h a n t h e duration T h e computations consist of applying the equation
of the rainfall excess. A n actual hydrograph would
have a lower p e a k and m u c h slower flow recession. Q, = ( 1 - 2K)Q _ +K{I
T X T + /,_,), (3.28)
This is because in reality t h e flow system on a water­
w h e r e Q is t h e runoff hydrograph o r d i n a t e at time i,
shed contains considerable storage, which r e t a r d s t

I is the d e p t h of rainfall excess in the time interval At,


surface and channel flow. T h e process of simple trans­ t

a n d Κ is a coefficient defined by
lation of t h e water from its point of incidence on t h e
watershed to t h e basin outlet neglects this storage
K= At/(2t c + Δ/), (3.29)
aspect of t h e flow process.
T o obtain a m o r e realistic runoff hydrograph, the w h e r e t is the watershed time of concentration. Any
c

conceptual hydrograph shown in Fig. 3.25 might be consistent set of units can be used. With / in hours and
thought of as the inflow into a hypothetical reservoir I in cfs-hours, l in cfs-hours can be determined from
t

whose outflow would t h e n b e t h e runoff hydrograph. the rainfall excess hyetograph, R in inches, from the
n

T h e effect of the reservoir would be t o delay and lessen relationship


the peak of t h e runoff hydrograph. Such an approach is
developed in the next section. /, = 1.008 AR /At,
t (3.30)
70 Chapter 3. Rainfall-Runoff Estimation in Storm Water Computations

q(x.t) neglecting higher-order differential t e r m s (i.e., (dx) ) 2

and collecting t e r m s result in

dA dv dA
~dt * A
T x + V
~dx~ = 9 ( χ , ί )
· <· )
3 36

E q u a t i o n (3.36) has two unknowns, area and velocity.


A second equation is required. T h e second equation is
the m o m e n t u m equation. Several a u t h o r s (Overton and
Figure 3.26 Control element for derivation of Eq. (3.36).
Meadows, 1976; H e n d e r s o n , 1966; Chow et al., 1988;
Strelkoff, 1969) present t h e developments that lead to
t h e m o m e n t u m equation in t h e form
where A is the watershed area in acres. This m e t h o d is
equivalent to flow routing using t h e Muskingum m e t h o d dv dv dy qv
defined in C h a p t e r 6 with χ = 0 and k = t . c
^7 + V
Tdx + g
T dx +
— y ~ * ( o " S f ) = 0, 5
(3.37)
dt

A Hydrodynamlc Model w h e r e dv/dt + v(dv/dx) is d u e to dynamic waves,


g(dy/dx) is d u e to diffusion waves, qv/y is d u e to
T h e m e t h o d for obtaining runoff hydrographs that is
lateral inflow, and g ( 5 - 5 ) is d u e to kinematic
Q f
the most theoretically elegant is to r o u t e t h e rainfall
waves. Woolhiser and Liggett (1967) proposed a di-
excess hyetograph as overland flow to established chan­
mensionless p a r a m e t e r , k, given by
nels and as channel flow to the basin outlet. T h e
p r o c e d u r e would rely on the continuity equation and a
flow equation, as well as relationships b e t w e e n the k = (3.38)
HF 2 '
various hydraulic elements of slope, roughness, chan­
nel shape, hydraulic radius, etc. w h e r e 5 is t h e slope of t h e flow plane, L is t h e length
Q

T h e continuity equation simply states that in a con­ of the flow plane, Η is t h e equilibrium d e p t h at the
trol element such as shown in Fig. 3.26, the difference b o t t o m of t h e plane, and F is t h e equilibrium F r o u d e
in the rate of inflow and outflow equals t h e rate of n u m b e r at t h e bottom of t h e plane. They found that
change of water stored in the element. T h e volume of for k greater than 10, neglecting all terms in Eq. (3.37)
the element is equal to t h e average area times the except t h e kinematic t e r m had n o appreciable effect on
length or t h e solution. Since S r e p r e s e n t s t h e friction slope, t h e
{

/ dA dx\ implication is that 5 = 5 a n d a uniform flow equa­


Q f

Volume = K = L 4 + - U*- ( · ) 3 31 tion can b e used in place of Eq. (3.37). Such an


equation is given by
\ dx 2 )
Q = aA n
(3.39)
T h e inflow to and outflow from the element are
Inflow rate = / = vA + q(x, t) dx (3.32) This is known as t h e kinematic a p p r o a c h w h e r e a and
η are rating coefficients. Overton a n d Meadows (1976),
I dv
\( dA
\ Wooding (1965a, b , 1966), a n d Woolhiser and Liggett
Outflow rate = 0 = * ; + — dx\\A + —dx\.
(1967) give m o r e information on w h e n the kinematic
\ dx ) J\ dx a p p r o a c h may be used. Essentially t h e kinematic a p ­
(3.33) p r o a c h can b e used w h e n t h e dynamic terms of the
In these expressions, A is the cross-sectional area of m o m e n t u m equation are negligible, backwater effects
the flow, χ is the horizontal coordinate, ν is t h e flow a r e small, a n d a rating equation like Eq. (3.39) applies.
velocity (mean velocity) over the cross-sectional area, In practice, the kinematic a p p r o a c h is widely used.
and q(x, t) is the lateral inflow p e r unit length along
If o n e is considering overland flow over a plane
the flow. T h e continuity equation states
surface, Eq. (3.36) can b e written o n a p e r unit of width
dV
I - Ο = —. (3.34) basis as
Substituting the correct quantities
dt results in ' v

dA dy dv dy
vA+q(x,t)dx- _ - - dx + y— + v— =q(x,t), (3.40)
a M d x
dt dx dx
d[A + (dA/dx)(dx/2)dx] w h e r e q(x, t) is t h e effective rainfall p e r unit area and
(3.35)
dt y is the d e p t h of flow. Several a t t e m p t s have b e e n
Runoff Estimation 71
m a d e to use Eq. (3.39) and (3.40) as a hydrodynamic
model of the overland flow process. 1
Equations (3.36) and (3.39) may also be used to
route flow in channels. In this case, q{x,t) would
represent tributary or o t h e r local inflow to the channel
per unit of channel length. Since Q = uA and

dQ dvA dA dv
— = — + Λ —, (3.41)

dx dx dx dx

Eq. (3.36) can be written

dA dQ
_ + _ = , ( , , , ) . (3.42)
Note that if a channel reach with no lateral inflow
is being considered, q(x, t) = 0 for t h a t reach. V-shaped Planes Cascading Planes
Brakensiek (1967), Overton a n d M e a d o w s (1976), a n d
Figure 3.27 Idealized flow planes.
Viessman et al., (1977) discuss numerical solutions for
Eqs. (3.39) a n d (3.42). T h e books by Eagleson (1970)
and Viessman et al., (1977) give m o r e complete deriva­
tions of t h e routing equations. t h e use of a c o m p u t e r . Some hydrologic models use
A possible modeling p r o c e d u r e using the hydrody­ this a p p r o a c h or have it as an available option.
namic equations is to divide t h e w a t e r s h e d into ideal­
ized planes of overland flow and a series of channels.
Equation (3.39) and (3.40) could be used to r o u t e t h e
The Unit Hydrograph Approach
effective rainfall to t h e channel system. E q u a t i o n (3.39) S h e r m a n (1932) developed the concept of hydro-
and (3.42) could then be used to r o u t e t h e resulting g r a p h estimation via a unit hydrograph. A unit hydro-
overland flow through t h e channel system. T h e lateral graph is a hydrograph of runoff resulting from a unit of
inflow terms, q(x, t) for t h e collector channels would rainfall excess occurring at a uniform rate, uniformly
be the routed overland flow hydrograph. F o r larger distributed over a watershed in a specified duration of
channels such as waterways and natural channels, time. T h e unit hydrograph approach to the develop­
q(x, t) would represent tributary inflows such as flows m e n t of runoff hydrographs is empirical and based on
from smaller channels. This modeling a p p r o a c h has several assumptions. Some of these assumptions are
b e e n used by Brakensiek (1967), W o o d i n g (1965a, b , contained in the definition. T h e s e are the assumptions
1966), Woolhiser (1969), and Woolhiser and Liggett of uniform distribution of rainfall excess over the wa­
(1967). T h e approach is also discussed by Eagleson tershed and uniform r a t e of rainfall excess. Possibly the
(1970), Overton and Meadows (1976), a n d Viessman et most i m p o r t a n t a n d controversial assumption is that of
al. (1977). superposition or linearity. This assumption states that
Advantages of the m e t h o d include minimum reliance the unit hydrograph reflects all basin characteristics to
on any observed runoff data, theoretical appeal, and t h e d e g r e e that t h e runoff rate is simply proportional
the directness with which the p a r a m e t e r s of the equa­ to the runoff volume for a rainfall excess of a given
tions can b e related to land use with t h e consequent duration.
ease of evaluating land-use changes. A unit hydrograph has a t t a c h e d to it the duration of
Disadvantages of the m e t h o d include t h e difficulty of t h e rainfall excess that g e n e r a t e d the unit hydrograph.
idealizing the overland flow planes and d a t a require­ T h u s , o n e might speak of a 20-min unit hydrograph or
ments to define channel geometries. A further concep­ a 1-hr unit hydrograph. Conceptually, an infinite num­
tual difficulty revolves around the fact that it is unlikely b e r of unit hydrographs, each of a different duration,
that overland flow as envisioned by this a p p r o a c h actu­ can be developed for every watershed. Practically, a
ally occurs except on very uniform impervious surfaces. unit hydrograph is applied to rainfall excesses of dura­
Figure 3.27 shows schematically some of t h e ideal­ tions as m u c h as 2 5 % different t h a n t h e duration of the
izations of plane surfaces that have b e e n used by the unit hydrograph. T h u s , a 20-min unit hydrograph might
previously referenced authors. Obviously, use of the be applied to rainfall excess of durations between 15
kinematic approach to routing overland flow requires a n d 25 min. Prior to a discussion of the development of
72 Chapter 3. Rainfall-Runoff Estimation in Storm Water Computations

u m e of rainfall excess. F o r a single block of rainfall


excess, the runoff hydrograph ordinates a r e simply
RAINFALL EXCESS equal to the volume of rainfall excess times the unit
V- « 0 . 7 5 inchtt hydrograph ordinates. N o t e t h a t t h e intensity of the
rainfall excess is 0.75 i n . / D units of time. N o t e also
that t h e rainfall excess a n d not t h e rainfall itself is
UNIT HYDROGRAPH OF DURATION D used.
V· * I inch
Figure 3.29 illustrates t h e development of a runoff
RUNOFF HYDROGRAPH hydrograph from a complex rainfall excess p a t t e r n . T h e
V - 0 . 7 5 inches rainfall excess is divided into blocks of uniform inten­
sity and of d u r a t i o n D . A c o m p o n e n t hydrograph for
each block of rainfall excess is t h e n obtained by multi­
plying the o r d i n a t e s of D-min unit hydrograph by t h e
volume of rainfall excess in each block. Hydrograph
TIME n u m b e r 1 is a D-min unit hydrograph. T h e c o m p o n e n t
Figure 3.28 Runoff hydrograph ordinates from a rainfall excess of
hydrographs a r e plotted making sure that t h e initial
duration D are proportional to the ordinates of a D-min unit
hydrograph. The constant of proportionality is the volume of rainfall
point or beginning of each c o m p o n e n t hydrograph oc­
excess. curs at t h e start of t h e a p p r o p r i a t e block of rainfall
excess. T h u s , hydrograph n u m b e r 2, corresponding to
t h e first block of effective rain of 0.2 in., is simply 0.2
times t h e unit hydrograph. H y d r o g r a p h 3, correspond­
unit hydrographs, their use in calculating runoff hydro- ing to t h e second block of effective rain of 0.15 in., is
graphs is described. 0.15 times t h e unit hydrograph a n d shifted in time
D-min to t h e right. H y d r o g r a p h 4 is 0.35 times the unit
Runoff Hydrographs from Unit Hydrographs hydrograph a n d shifted 2 D time units t o t h e right. This
Figure 3.28 shows how a unit hydrograph is used in process can b e continued for as m a n y blocks of rainfall
t h e development of a runoff hydrograph from a block excess as necessary. T h e resulting total runoff hydro-
of rainfall excess of duration equal to t h e duration of graph, n u m b e r 5 in Fig. 3.29, is t h e n obtained by
the unit hydrograph. U n d e r the unit hydrograph as­ adding t h e ordinates of t h e c o m p o n e n t hydrographs at
sumptions, the runoff rate is proportional to t h e vol­ each point in time.

Time

ι D-minute unit hydrograph V = 1 inch

I Runoff hydrograph ¥ = 0 . 7 inches

) 0.2 times UH v = 0.2 inches

0.35 times UH ¥ = 0.35 inches


0.15 times UH V = 0.15 inches

Figure 3.29 Runoff hydrograph from a complex storm is obtained by summing component
hydrographs from D-min blocks of rainfall excess.
Runoff Estimation 73

Mathematically, the ordinate of the total runoff hy­


graph for rainfall excess durations of any multiple of
drograph at any time t = jD is given by
the duration of the original unit hydrograph. T o obtain
a unit hydrograph of duration nD, simply add η unit
r = 0, i > m
q<= L, i i-i+\
r u
i

η v. (3.43) hydrographs of duration D each displaced by D time


J
. , 1 J 1 + 1
U;; = 0, I > Π, V
' units and divide the sum by n. Keep in mind that the
duration, D , is t h e duration of the blocks of rainfall
where q is the hydrograph ordinate at time / = jD, r
} t excess—not t h e total time base of the unit hydrograph.
is the volume of rainfall excess in the ith block of Using this p r o c e d u r e a 2 D , 3 D , . . . , n D unit hydro-
rainfall excess of duration D , u represents the ordi­ i graph can be derived. Only unit hydrographs of the
nate of the D-min unit hydrograph at time iD, η + 1 is same duration can b e used in this summing process.
the number of unit hydrograph ordinates with u and 0 T h e sum of unit hydrographs of durations D and 2 D
u both being zero, and m is the n u m b e r of blocks of
n cannot b e used to d e t e r m i n e a unit hydrograph of
effective rainfall. Equation (3.43) is t h e same as Eq. duration 3 D .
(3.27), except the unit hydrograph has replaced the A p r o c e d u r e known as the S-curve technique can be
t i m e - a r e a diagram. T h e unit hydrograph has e m b e d ­ used to derive a unit hydrograph of any duration, D ' ,
ded within it the final routing that is lacking in the from a unit hydrograph of any o t h e r duration, D . An
hydrograph generated by Eq. (3.27). T h u s , basin flow S-curve, like a unit hydrograph, has a duration at­
time, lag time, and storage are incorporated into Eq. tached to it. T h e duration of the S-curve is equal to the
(3.43) and not (3.27). Figure 3.30 illustrates how unit duration of the unit hydrograph from which the S-curve
hydrographs of a given duration can be used to develop was derived. Conceptually, an S-curve represents the
runoff hydrographs for storms of any multiple of that results of a steady, continuous excess rainfall rate of
duration. 1 / D units of rain p e r unit of time. A n S-curve can be
In plots such as Fig. 3.29, it is conventional to plot c o m p u t e d by summing the ordinates of a large n u m b e r
the effective rainfall along t h e t o p of the diagram using of D unit hydrographs spaced D time units apart (Fig.
the same time scale as used for the hydrograph, but a 3.31). Computationally the S-curve can be more easily
different rate scale. Many times the rate scale runs obtained from a single D unit hydrograph by comput­
from zero at the top of the diagram with positive values ing t h e cumulative sum of the D unit hydrograph
in the downward direction. In a situation such as that ordinates spaced D time units apart. Table 3.18 shows
shown in Fig. 3.29, the unit hydrograph duration must the calculation of a 30-min S-curve from a 30-min unit
correspond to the duration of rainfall excess. T h a t is, a hydrograph when the ordinates of the unit hydrograph
D-min unit hydrograph times the volume of rainfall are tabulated in 15-min increments. T h e results are
excess received in 3 D min (for example) does not equal shown in Fig. 3.31.
a 3 D-min (or a D-min) runoff hydrograph. Table 3.18 shows that if D is the duration of the unit
hydrograph and S-curve, t h e n the ordinates of the
Converting Unit Hydrographs
S-curve are given by
Figure 3.30 illustrates how a unit hydrograph of a
given duration can be used to develop a unit hydro- S(t) = u(t) for t = 0 , D
V
' V }
(3.44)
5(0 =S(f-D) + u(t) for t>D.

If S(t) represents the ordinates for a D-min S-curve,


the ordinates for a D'-min unit hydrograph can be
c o m p u t e d from

u'(t) = (D/D')[S(t)-S(t-D')}. (3.45)

T h e terms of this relationship are illustrated in Fig.


3.32.
Since the S-curve corresponds to a unit hydrograph
of a specified duration, D , deriving a unit hydrograph
of some o t h e r duration, D ' , requires a volume correc­
tion factor, which is the ratio of the old to the new unit
hydrograph durations or D / D ' as shown in Eq. (3.45).
T h e original unit hydrograph corresponds to a uniform
Figure 3.30 Development of a unit hydrograph of duration nD. rainfall excess r a t e of o n e unit of rain per D time units,
74 Chapter 3. Rainfall-Runoff Estimation in Storm Water Computations

30-minute S curve 15-minute unit hydrograph


400 400

0 50 100 150 200 250 300 350 400 0 50 100 150 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0


Time (minutes) Time (minutes)
Figure 3.31 S-Curve computation. Figure 3.32 Unit hydrograph from an S-curve.

while t h e new o n e corresponds to a r a t e of o n e unit of


rain p e r D' time units.
In computing S-curves, some m a n u a l smoothing is
generally required. T h e unit hydrograph of T a b l e 3.18
Table 3.18 S-Curve Computation from 30-min Unit is from a basin with an a r e a of 190 acres (77 hectares).
Hydrograph
Since t h e unit hydrograph has a duration of 30-min
Time (hr) Time (min) UH (cfs) Sum* S-curve (cfs) a n d a volume of 1 in., t h e equilibrium rate for the
S-curve should correspond to a runoff r a t e of 1 in. p e r
0.00 0 0 0 0 30 min over 190 acres, 2 i n . / h r over 190 acres, or
0.25 15 29 29 29
inches cfs
0.50 30 68 68 68
2- X 190 acres X 1.008
0.75 45 93 93+29 122 hour acre-inch/hour
1.00 60 100 100+68 168 = 383 cfs.
1.25 75 95 95+122 217 T a b l e 3.19 shows t h e c o m p u t a t i o n of a 15-min unit
1.50 90 83 83+168 251 hydrograph based on t h e S-curve of T a b l e 3.18. N o t e
1.75 105 68 68+217 285 t h e required smoothing. T h e smoothing can b e accom­
2.00 120 54 54+251 305 plished by using o n e of t h e unit hydrograph models
2.25 135 46 46+285 331 discussed later in this chapter. T h e volume u n d e r the
2.50 150 32 32+305 337 unit hydrograph should b e checked to e n s u r e that is in
28+331 359
fact a unit hydrograph. If t h e time increment, At, is in
2.75 165 28
minutes, t h e flow, q, in cfs, t h e catchment area, A, in
3.00 180 16 16+337 353
acres, and t h e first a n d last flow values are zero, the
3.25 195 13 13+359 372
volume, V, in inches u n d e r a hydrograph is given by
3.50 210 8 8+353 361
3.75 225 7 7+372 379
(3.46)
4.00 240 5 5+361 366 60.5Λ '
4.25 255 4 4+379 383
w h e r e Σ<?,· r e p r e s e n t s the sum of all of t h e ordinates of
4.50 270 3 3+366 369 the hydrograph spaced At apart. Applying this equa­
4.75 285 2 2+383 385 tion t o t h e last column of T a b l e 3.18 results in a
5.00 300 1 1+369 370 volume of 1.003 in. This sum could b e m a d e even
5.25 315 1 1+385 386 closer to unity by a slight adjustment in t h e hydrograph
5.50 330 0 0+370 370 ordinates; however, 1.003 is sufficiently close to 1.00 for
5.75 345 0 0+386 386 most applications, making this further refinement un­
360 0 0+370 370
necessary.
6.00
Unit hydrographs can b e derived from observed
a
Current UH ordinate plus S-curve ordinate 30 min earlier. records of streamflow and rainfall. D a t a r e q u i r e m e n t s
Runoff Estimation 75

Table 3.19 15-min Unit Hydrograph from S-Curve After presenting p r o c e d u r e s for estimating these at­
tributes, several unit hydrograph models are presented.
Time S-curve Smoothed Displaced UH
(min) (cfs) S-curve S-curve fl
(cfs) smoothed
Estimation of Time Parameters
0 0 0 0 0 0
This section deals with the estimation of the time
15 29 29 0 58 58
p a r a m e t e r s D, t , t , and t as shown in Fig. 3.22 and
L p b
30 68 68 29 78 78
t h e time of concentration, t . Several m e t h o d s for c

45 122 122 68 108 112


estimating these p a r a m e t e r s are available. T h e method
60 168 168 122 92 100
that p r o d u c e s results consistent with good engineering
75 217 217 168 98 96 j u d g e m e n t should b e selected for a particular study
90 251 251 217 68 85 area.
105 285 285 251 68 64 T h e time of concentration is the time it takes for
120 305 305 285 40 44 flow to reach t h e basin outlet from t h e hydraulically
135 331 331 305 52 36 most r e m o t e point on t h e watershed. For some areas,
150 337 342 331 22 28 this p a r a m e t e r can b e estimated by summing the flow
26 20
time for the various flow segments as the water travels
165 359 355 342
toward t h e w a t e r s h e d outlet. T h e s e segments generally
180 353 360 355 10 14
are overland flow, shallow channel flow toward larger
195 372 368 360 16 12
channels, and flow in o p e n channels, both natural and
210 361 375 368 14 10
improved. T h e travel time in these various flow seg­
225 379 377 375 4 6 m e n t s d e p e n d s on the length of travel and the flow
240 366 378 377 2 4 velocity.
255 383 379 378 2 2 O n c e t h e velocity in each flow segment is deter­
270 369 382 379 6 0 mined, the time of concentration is determined from
285 385 383 382 2 0
300 370 383 383 0 0 Ό=Σ^> (3-47)
315 386 383 383 0 0 /=1 u
i

330 370 383 383 0 0 w h e r e η is t h e n u m b e r of flow segments and L , is the


345 386 383 383 0 0 length and u t h e flow velocity for the i t h segment.
t

360 370 383 383 0 0 Flow velocity of overland flow and shallow channel
flow can be estimated using results such as those of
Sum 769
Izzard (1946), R e g a n and D u r u (1972), Overton and
M e a d o w s (1976), or from the relationship
a
S(/-D')
ν = aS l/2
(3.48)
h
u(t) = [5(0 - 5 ( r - D ) ] D / D ' = [S(/) - S(r-15)]30/15.
,

based on information in SCS (1975), w h e r e 5 is in f t / f t


a n d ν is in fps. T h e coefficient a is contained in Table
3.20.
R e g a n and D u r u (1972) present a m e t h o d for esti­
are extensive and generally prevent direct derivation of
mating travel time, t over a plane surface based on
v
unit hydrographs for small catchments. Unit hydro-
the kinematic wave equation [Eq. (3.40)]. T h e equation
graphs represent direct stormwater runoff. Baseflow
is valid for turbulent flow or w h e n the product of the
a n d / o r ground water discharges to streams must b e
rainfall excess intensity, i , in iph and the flow length,
c
removed from the flow record before unit hydrographs
L , in feet is greater t h a n 500. T h e equation is
can be defined from t h e record. Linsley et al. (1982)
can be consulted or details. For small catchments, 0.0155 ( n L )
^04
.503
.
0 6

synthetic unit hydrographs are generally used. Syn­ *t =


' (3-^9)
thetic unit hydrographs are discussed in detail in the
following sections of this chapter. Several synthetic w h e r e t is in hours, η is M a n n i n g ' s n, L is in feet, i
x e

unit hydrograph models have b e e n p r o p o s e d . G e n e r ­ is in iph, and S is the slope in ft/ft. Table 3.21
ally they provide the ordinates of the unit hydrograph p r e s e n t s some values for η for overland flow surfaces.
as a function of the time to peak, t , p e a k flow r a t e , p T h e Soil Conservation Service (1986) presents a rela­
q , and a mathematical or empirical s h a p e description.
p tionship attributed to Overton a n d Meadows (1976) for
76 Chapter 3. Rainfall-Runoff Estimation in Storm Water Computations

Table 3.20 Coefficient a for Eq. (3.48) a


obstacles such as litter, crop residue, ridges, a n d rocks;
and t h e erosion and transport of sediment. T h e s e η
Surface a values are for very shallow flow d e p t h s of about 0 . 1 ft
Overland flow or so. Table 3 . 2 1 gives M a n n i n g ' s η values for these
Forest with heavy ground litter 2.5 conditions. T h e relationship for travel time is
Hay; meadow 2.5
Trash fallow; minimum tillage 5.1 0.007(A*L) 0 8

Contour; strip cropped 5.1 Γ»- P o5co.4 , (3.50)


Woodland 5.1
Short grass 7.0
w h e r e P is t h e 2-year, 24-hr rainfall in inches and t h e
2
Straight row cultivation 8.6
o t h e r terms are as defined for Eq. (3.49). This relation­
Bare; untilled 10.1
ship is based on shallow, steady, uniform flow; a con­
Paved 20.3
stant rainfall excess intensity; and minor effects from
Shallow concentrated flow infiltration.
Alluvial fans 10.1 In u r b a n areas, t h e travel time may have to b e based
Grassed waterways 16.1 on a travel time to a storm drain inlet plus the travel
Small upland gullies 20.3 time t h r o u g h t h e storm drain itself. Inlet travel time
can generally b e c o m p u t e d as t h e sum of overland flow
fl
Results in fps; multiply by 0.305 to get m/sec. a n d shallow channel flow travel times. Flow in storm
drains would b e considered as o p e n channel flow with
the storm drain pipe flowing full. Often large storms
Table 3.21 Manning's η for Travel Time Computations for p r o d u c e runoff rates that exceed t h e capacity of t h e
Flow over Plane Surfaces (Soil Conservation Service, 1986) storm drains a n d some of t h e runoff bypasses t h e
drains in t h e form of c o n c e n t r a t e d surface flow as o p e n
Surface description na

channel flow. Such flow should b e considered in com­


Smooth surfaces (concrete, asphalt, puting the time of concentration.
gravel, or bare soil 0.011 Undersized culverts and bridge openings can cause
Fallow (no residue) 0.05 ponding of flow a n d a reduction in t h e average flow
Cultivated soils velocity. F o r small p o n d s a n d situations w h e r e water is
Residue cover <20% 0.06 passing through t h e p o n d with little or n o storage build
Residue cover >20% 0.17 up, the actual travel time through t h e p o n d may be
Grass very small. If significant storage results, t h e travel time
Short grass prairie 0.15 is lengthened over that for normal channel flow, and
Dense grasses* 0.24
flow routing as discussed in C h a p t e r 6 must b e used.
Bermudagrass s
0.41
Flow velocity for o p e n channels can b e estimated
Range (natural) 0.13
from Manning's equation, which is t r e a t e d in detail in
Woods c
C h a p t e r 4.
Light underbrush < 0.40 O t h e r m e t h o d s are available in t h e form of empirical
Dense underbrush 0.80 equations for estimating t . O n e such relationship that
c

T h e η values are a composite of information compiled by Engman


a is widely used but based on limited d a t a is expressed by
(1986). Kirpich (1940)
^Includes species such as weeping lovegrass, bluegrass, buffalo grass,
blue grama grass, and native grass mixtures.
When selecting n, consider cover to a height of about 0.1 ft. This is the
4 t = 0.0078L
c
0 7 7
(L///) 0 3 8 5
, (3.51)
only part of the plant cover that will obstruct sheet flow.

w h e r e t is in minutes, L is t h e maximum length of


c

flow in feet, and Η is t h e difference in elevation in feet


travel time for sheet flow over plane surfaces based on b e t w e e n t h e outlet of t h e w a t e r s h e d a n d t h e hydrauli-
Manning's equation and a kinematic approximation to cally most r e m o t e point in t h e watershed. Obviously,
the flow equations. T h e equation is for flow lengths of E q . (3.51) does not consider flow resistance in t h e form
less than 300 ft. T h e friction value or Manning's η is of overland flow and channel roughness.
an effective roughness coefficient that includes the Several m e t h o d s for estimating t h e lag time of a
effect of raindrop impact; drag over plane surfaces; watershed are available. O n e simple m e t h o d for lag
Runoff Estimation 77
time estimation is (Soil Conservation Service, 1973) Table 3.22 Φ Values for Eq. (3.55) (Epsey et al, 1977)

f -0.6i .
L c (3.52) Manning's η

T h e SCS (1975) has developed a lag equation based Percentage imp. 0.015 0.03 0.05 0.10 0.15
on natural watersheds
0 0.82 0.86 0.93 1.15 1.30

L 0 8
( 5 + I) ' 0 7
20 0.74 0.80 0.88 1.09 1.27
Κ - ( 5 0 ^ C N < 95), (3.53) 40 0.65 0.72 0.81 1.03 1.22
60 0.60 0.68 0.79 1.00 1.19
where t is the lag in hours, L is the hydraulic length
L

of the watershed in feet, 5 is related to t h e curve


n u m b e r by Eq. (3.22), and Y is t h e average land slope
in percentage. T h e 5 in Eq. (3.53) should be based on
an antecedent condition II curve n u m b e r since it is age impervious a r e a with an assumed minimum value
being used as a m e a s u r e of surface roughness and not of 5 % for an undeveloped area, and Φ is a conveyance
runoff potential. factor that d e p e n d s o n the percentage impervious area
Many local studies relating t or t or t to water­ L p c
a n d Manning's η for t h e main channel. Table 3.22
shed physical characteristics have b e e n conducted. For contains some representative values for Φ.
example, P u t n a m (1972) in a study of 34 watersheds in T h e base time of a unit hydrograph is somewhat
North Carolina, presented t h e relationship arbitrary. Some hydrologists use a base time of five
times the time to peak. Some unit hydrograph models
/ L \ · 0 5 0
have a recession limb that asymptotically approaches
i L = 0.49|-^j Ζ" 0 5 7
, (3.54) q = 0, so that t h e base time is theoretically infinity.

where t is the basin lag in hours, L is t h e length of


L

the main water course in miles, S is the main stream


Estimation of Peak Flow Parameters
slope in feet per mile, and / is fraction of impervious T h e p e a k flow r a t e of a unit hydrograph is often
area. H e r e t was defined as t h e time from the center
L given by an equation of the form
of mass of rainfall to t h e center of mass of runoff.
Before an equation like (3.54) is used, care must be q = KA/t . (3.56)
p p

exercised to see that the conditions u n d e r which the


equation was developed match the conditions of inter­
Based on a triangular unit hydrograph with a base
est.
time of 2 . 6 7 i , t h e SCS (1972) estimates the peak flow
T h e duration, D , of the rainfall excess that is gener­ p

of a unit hydrograph from the equation


ally associated with a unit hydrograph should be o n e -
fifth to one-third of the time to peak. T h e time to p e a k
is given by Eq. (3.26) as 484/1
% = — — , (3.57)
fp = t L + D/2.

Epsey et al. (1977) studied rainfall-runoff records w h e r e q is the p e a k flow in cfs, A is the basin area in
p

from 41 watersheds located in several states (Texas, 16; s q u a r e miles, a n d t is t h e time t o p e a k in hours.
p

North Carolina, 9; Kentucky, 6; Indiana, 4; Colorado, Epsey et al. (1977) r e c o m m e n d that for 10-min unit
2; Mississippi, 2; Tennessee, 1; and Pennsylvania, 1). hydrographs, the relation
T h e watersheds ranged in size from about 9 to 9600
acres (3.5 to 3900 hectares). They developed an estima­ q = 31620(Λ°· 7' · ) 9 07
(3.58)
p Ρ
tion equation for the time to peak of 10-min unit
hydrographs as
b e used w h e r e q is in cfs, A is the drainage area in
p

t p = S.IL - ^- - 0 2 0 2 5
/" 0 1
^ -
1 5 7
, (3.55) square miles, and t is the time to peak in minutes.
p

As was the case for lag time, many studies have been
where t is the time to peak in minutes, L is the main
p conducted in an effort t o relate q to watershed physi­
p

channel length from the u p p e r watershed boundary in cal conditions. Before any of these empirically derived
feet, S is the slope in feet per foot of the lower 8 0 % (in equations are used, their applicability should be care­
terms of length) of the main channel, / is the percent- fully d e t e r m i n e d .
78 Chapter 3. Rainfall-Runoff Estimation in Storm Water Computations

unit hydrograph. T h e curves in Fig. 3.33 give the width


of t h e unit hydrograph at flow rates equal to 0.75 q p

a n d 0.50 q . A s a rule of t h u m b , t h e widths at these


p

points can b e p r o p o r t i o n e d so that one-third of the


width occurs prior to t h e p e a k discharge.
Knowledge of q and t along with Fig. 3.33 make it
p p

possible to sketch a p r o p o s e d unit hydrograph. Any


time a sketching p r o c e d u r e is used, it is essential to
e n s u r e that t h e volume of t h e proposed unit hydro-
graph is 1 in. If it is not, the sketch should be altered
until a volume of 1 in. is achieved.
T h e SCS (1972) uses a dimensionless unit hydro-
graph as shown in Fig. 3.34. They also employ a trian­
gular unit hydrograph derived to have t h e same time to
p e a k a n d p e a k flow rate as their dimensionless unit
0.2 0.4 0.6 0.8 1 2 4 6 8 10 20 40 60 80100 hydrograph. T h e SCS triangular unit hydrograph is
Width of unit hydrograph (hr) also shown in Fig. 3.34. If either of t h e SCS dimension­
Figure 3.33 Unit hydrograph width at 50 and 75% of peak flow less unit hydrographs is used, q a n d t must be
p p

(U.S. Corps of Engineers, 1959). related by E q . (3.57).


H a a n (1970), D e C o u r s e y (1966), a n d others have
proposed dimensionless hydrograph equations. Starting
Shape of Unit Hydrographs
with t h e empirical equation
If a short duration unit hydrograph is applied to a
long rainstorm pattern, the actual shape of the unit q(t) =at e~ b ct

hydrograph is not nearly as important as the time to


peak and the peak flow rate. Some hydrograph proce­ as a hydrograph equation a n d using t h e conditions
dures actually call for a simple sketching of the approx­ t=t,q= q
(1) at p p
imate shape of the unit hydrograph, making sure t h e
volume of the hydrograph is 1 in. dq(t)
Based on the work of Snyder (1938), the U.S. Army () , , ,_A2
2 AT = P =0

Corps of Engineers (1959) produced Fig. 3.33, which


can be used as an aid in determining the shape of a (3) Cq{t)dt = V.

Figure 3.34 Dimensionless unit hydrograph and mass curve with triangular
unit hydrograph superimposed.
Runoff Estimation 79

H a a n (1970) developed the relationship


I *

—e»-'/'» (3.59)

where q(t) is the hydrograph ordinate at any time t, q p

is the peak flow rate (iph), t is the time to peak


p

(hours), and Κ is a p a r a m e t e r defined by the equation

V=q t [e/K] T(K),


p p
K
(3.60)

where V is the runoff volume (inches—1 in. for a unit


hydrograph), e is the base of the natural logarithms,
and Γ represents the gamma function. Figure 3.35
presents a quick solution of Eq. (3.60) for K. T h e
relationship between Κ and q t /V can b e approxi­
p p

mated by
1.92
K = 6.5 ρρ (3.61)
A /inches . x

T h e SCS dimensionless unit hydrograph corresponds Vp (-F5UT x h o u r s


)
to a Κ value of about 3.77. This can be seen by using V (inches)
Eq. (3.57) and making the p r o p e r unit conversions. Figure 3.35 Relationship between qJtJV and K.
Figure 3.36 shows how the shape of the hydrograph
defined by Eq. (3.59) changes as Κ changes. Table 3.23
gives the ordinates for the dimensionless hydrograph of

1.0 r -

Curve parameter is Κ

t/t

Figure 3.36 Comparison of hydrographs for several values of K.


80 Chapter 3. Rainfall-Runoff Estimation in Storm Water Computations

Table 3.23 Coordinates for Dimensionless Unit Hydrographs

1 1.5 2 3 4 5 6 7 8 9 10

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
0.25 0.54 0.39 0.29 0.15 0.08 0.04 0.02 0.01 0.01 0.00 0.00
0.50 0.83 0.75 0.68 0.56 0.46 0.38 0.32 0.26 0.22 0.18 0.15
0.75 0.97 0.95 0.93 0.90 0.87 0.84 0.81 0.78 0.75 0.73 0.70
1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
1.25 0.97 0.96 0.95 0.92 0.90 0.87 0.85 0.83 0.81 0.78 0.76
1.50 0.91 0.87 0.83 0.75 0.69 0.62 0.57 0.52 0.47 0.43 0.39
1.75 0.83 0.75 0.68 0.56 0.47 0.38 0.32 0.26 0.22 0.18 0.15
2.00 0.74 0.63 0.54 0.40 0.29 0.22 0.16 0.12 0.09 0.06 0.04
2.50 0.57 0.43 0.32 0.18 0.10 0.06 0.03 0.02 0.01 0.01
3.00 0.41 0.26 0.16 0.07 0.03 0.01
3.50 0.29 0.15 0.08 0.02 0.01
4.00 0.20 0.09 0.04 0.01
5.00 0.10 0.03 0.01
6.00 0.04 0.01

Eq. (3.59) ior various values of K. T h e derivation of being equal to t h e effective rainfall rate using a time
Eqs. (3.59) and (3.60) is in H a a n (1970). increment of 6 min.
Ardis (1973) developed a double triangle unit hydro-
graph in an a t t e m p t to incorporate both a quick a n d
delayed runoff response. H e assumed that the delayed Example Problem 3.12. Double triangle unit hydro-
response peak would coincide with the time when the graph
quick response e n d e d and that both responses would
start at the same time. Figure 3.37 depicts t h e compo­ Develop a double triangle unit hydrograph for an urban
nents of and the resultant double triangle unit hydro- area of 100 acres with a time to peak of 15 min.
graph. Solution: The points defining the unit hydrograph are
Wilson et al. (1983) have adapted the double triangle computed using the coordinates tabulated above.
unit hydrograph model for small watersheds. They plot
Point a:
the unit hydrograph in the form shown in Fig. 3.38.
t q(t) = 0.756 in.
T h e units used are inches per hour for q(t) and hours p

for time. T h e coordinates for the points labeled 0, b, q(t) = 0.756 in./0.25 hr = 3.02 iph
and c are inches cfs
q = 3.02
Q X 100 acres X 1.008 = 304 cfs :
μ
hour acre-inch
Land use a b c
t = f = 15 min. p

Point b:
Forested 1.0,0.268 4.105,0.054 18.068,0.0
6.982,0.0
0.151
Agricultural 1.0,0.526 2.375,0.113

Urban 1.0,0.756 2.000,0.151 4.333,0.0 q(t) = - y ^ j - X 100 X 1.008 = 61 cfs


Disturbed 1.0,0.756 2.000,0.151 4.333,0.0
t = 2 X 15 = 30 min.
Point c:
«(0 = o
A n effective rainfall pattern is used with a time interval t = 4.33 X 15 = 65 min.
of 3 min. If the time to peak is less than 6 min, n o unit The volume under the hydrograph can be computed by
hydrograph is used and the runoff rate is taken as dividing the hydrograph into three parts at the three defining
Runoff Estimation 81

out G r e a t Britain. O n e aspect of the study involved


developing a p r o c e d u r e for obtaining synthetic unit
hydrographs. T h e prediction equations they developed
were derived from 1631 events from 143 catchments
ranging in size from 3.5 to 500 k m (864 to 123,550 2

acres). T h e result was a triangular unit hydrograph.


T h e volume associated with their unit hydrograph was
10 mm:
V = 10 m m .

T h e time to p e a k for the hydrograph is estimated from

t p = 46.65-°' URBT- "RSMD- - L


3 8 1 0 4 0 1 4
, (3.62)

w h e r e t is in hours, S is the slope of the main channel


p

between the points 10 and 8 5 % of the length of the


channel m e a s u r e d u p from the outlet in meters per
kilometer, U R B T is 1 plus the fraction of the area that
is urbanized, R S M D is t h e 5-year, 24-hr rainfall excess
in millimeters, and L is the length of the main channel
T1 T2 T3
in kilometers. A statistical analysis presented based on
Figure 3.37 Components of double triangle unit hydrograph.
the regressions used to arrive at the above equation
indicates that a simpler version given by
.028
tp = 22.65" 0 5 9 7
URBT- 2
(3.63)

is also an acceptable equation. T h e t given by these p

equations is for a unit hydrograph having a duration of


q(t)t.
1 hr. T h e study r e c o m m e n d s that the unit hydrograph
t should b e about five times t h e duration. T o adjust
p

the t to a new duration, D, in hours, they recommend


p

the relationship

t' = t + (D - l ) / 2 , (3.64)
Figure 3.38 Double triangle unit hydrograph of Wilson et al. p p

(1983).
w h e r e t' is t h e adjusted time to peak.
p

It is r e c o m m e n d e d that if any rainfall and runoff


data are available, the lag time should be evaluated
time points:
empirically and t c o m p u t e d from
v=v _ + v _ + v _
p

Q a a b b c
'P = 0 . 9 / , L
(3.65)

3.02 3.02 + 0.61 w h e r e t is the lag time defined as the time from the
L

V= — X 0.25 + (0.50 - 0.25) center of mass of rainfall to the peak of the unit
0.61 / 65 hydrograph.
\
0.5 O n c e the p r o p e r t is obtained, the q can be
p p

2 \60 / d e t e r m i n e d from
V — 1.009 in.
which is acceptable for a unit hydrograph. q p = 2.20A/i , p (3.66)
Ordinates at other time points can be easily determined
from the straight line relationships defining the unit hydro- w h e r e q is in cubic meters p e r second and A is in
p

graph. square kilometers. T h e base time of the triangular unit


hydrograph is given by

th = 2.52/ . (3.67)
T h e Natural Environmental Research Council (1975)
of G r e a t Britain conducted a comprehensive flood study T h e Flood Studies R e p o r t gives an example for a
involving a large n u m b e r of catchments from through- catchment with A = 63.2 k m , L = 17.22 km, S = 2
82 Chapter 3. Rainfall-Runoff Estimation in Storm Water Computations

14.7 m / k m , and R S M D = 55.2 mm. T h e catchment is time to peak.


5 % urbanized, so U R B T is 1.05. Substituting these Gray (1961), Bleek (1975), W i t t e n b u r g (1975), and
quantities into Eq. (3.62) results in a t of 4.6 hr. If Eq.
p many o t h e r s discuss relating t h e derived hydrograph
(3.63) is used, t is found to be 4.1 hr.
p p a r a m e t e r s to watershed physical characteristics so that
Rainfall and runoff data indicate that for this catch­ they may be applied to u n g a g e d basins.
ment, a lag time of 2.3 hr is appropriate. Based on Eq. It is a p p a r e n t that t h e r e are a n u m b e r of relation­
(3.65), the t is found to be 2.1 hr. T h e a p p r o p r i a t e
p ships that can b e used to estimate t h e p a r a m e t e r s of
duration for the unit hydrograph is thus 2 . 1 / 5 , which unit hydrographs. T h e technique selected should be an
r o u n d e d to a convenient fraction of an hour is 0.5 hr. accepted technique for the location u n d e r study. T h e
t for a 0.5 hr unit hydrograph is c o m p u t e d from Eq.
p conditions of t h e study a r e a should be hydrologically
(3.64) as similar to t h e conditions u n d e r which the technique
was derived.
t' = 2.1 + (0.5 - l ) / 2 = 1.9 hr..
p
Consistency should be maintained in the application
of a unit hydrograph p r o c e d u r e . It was previously stated
T h e peak flow of the unit hydrograph is found from
that if t h e SCS unit hydrograph is used, Eq. (3.57)
Eq. (3.66) as
should b e used to estimate q . O n e should not mix p

qp = 2.20 X 6 3 . 2 / 1 . 9 = 73.2 m / s e c . 3
m e t h o d s . T h a t is, if t is based on a particular m e t h o d ,
p

q should b e based on the same m e t h o d as should t h e


p

T h e base time of the unit hydrograph is unit hydrograph s h a p e . T h e reason for this is that
generally particular hydrograph m e t h o d s are developed
t b = 2.52 X 1.9 = 4.8 h r . as a unified whole. T h e e l e m e n t s of t h e m e t h o d s are
not i n d e p e n d e n t of each other. T h e coefficients in a
T h e volume of this unit hydrograph can be easily
particular estimating equation are somewhat d e p e n ­
computed as the area u n d e r the triangular unit hydro-
d e n t on t h e coefficients in t h e companion equations
graph giving 0.010 m or 10 m m as it should be u n d e r
that m a k e u p t h e m e t h o d .
the conditions of the study.
Some hydrograph equations such as Eq. (3.59) are
Equations (3.62) through (3.67) should be applied
i n d e p e n d e n t of t h e m e t h o d used to estimate q t,
only to conditions, especially meteorological condi­ p9 p

a n d V. However, even in this case, a consistent proce­


tions, similar to those existing in G r e a t Britian. Rain­
d u r e should b e used to estimate these t h r e e p a r a m e ­
fall intensities are generally lower t h a n those prevalent
ters. In general, o n e should not estimate t from o n e
in the U.S. for storms with similar return periods. p

m e t h o d a n d q from a n o t h e r .
Edson (1951) and later Nash (1959) proposed that a p

hydrograph can be represented by an equation of the


form
Example Problem 3.13. Unit hydrograph
Q ( t ) =
rjn)^"''' '" '
1 1 ( 3 6 8 )
Develop a unit hydrograph for a 100-acre watershed with a
main channel length of 5000 ft, Manning's η of 0.025, and a
where η and k are p a r a m e t e r s and Γ represents the slope of 1% in the lower 80% of the channel. The watershed
gamma function. Equation (3.68) is sometimes known is undeveloped.
as the Nash model or the gamma distribution. T h e Solution: The method of Epsey et al. (1977) is used to
equation was derived by assuming that an instanta­ estimate / and q . The choice is dictated by the information
p p

neous rainfall routed through a series of η linear provided in the problem. If land-use and soils data were
reservoirs represents the runoff process. provided, the SCS method might be preferred. From Eq.
(3.55),
In applying Eq. (3.68), it is necessary to b e able to
estimate the p a r a m e t e r s in the equation. Gray (1961) t = 3.1^· 5-°· /-° Φ · .
p
2 3 2 5 1 8 1 5 7

employed the m e t h o d of maximum likelihood. Nash


For η = 0.025 and 0% impervious, Φ is found from Table
(1959) advocated the use of the m e t h o d of m o m e n t s .
3.22 to be 0.83:
Both of these m e t h o d s tend to produce hydrographs
that are " b e s t fits" over the entire range of the hydro- f - 3.1(5000)
p
023
(0.01)" 025
(5)" 018
(0.83) 1 5 7
= 39 min.
graphs and can have substantial deviations from ob­
served hydrographs at the peak. W u (1963) used the qp is determined from Eq. (3.58) as
time to peak and a recession constant to estimate the
p a r a m e t e r s of his runoff equation. Bloomsburg (1960)
q = 31620(Λ · /'Γ)
p
0 96

fit the hydrograph p a r a m e t e r s to the peak flow a n d the = 31620( 1 0 0 / 6 4 0 ) ° ^ ( 2 9 ) ' 1 0 7


= 106 cfs.
Estimation of Peak Runoff Rates 83

U s e Eq. (3.59) to define the unit hydrograph:


F o r small catchments, the p a r a m e t e r s of a 10-min
_ l-'/' |
e P
unit hydrograph can b e estimated from Eqs. (3.55) and
(3.58) and t h e unit hydrograph model of Eq. (3.59)
used.
q p = 106 cfs or a b o u t 1.06 iph
i p = 39 min = 0.65 h r
V = 1 in. ESTIMATION OF PEAK RUNOFF RATES
Κ is c o m p u t e d from Eq. (3.61) as
T h e r e are many occasions w h e n all that is required
/ 1.06 X 0.65 V s 1 9 2

K = 6.5 to design a facility is t h e p e a k flow r a t e that might be


experienced o n a certain frequency. Of course, the
T h e unit hydrograph is defined by hydrograph m e t h o d s discussed in t h e previous sections
3.18 can b e used to estimate a p e a k flow. However, for
t
q(t) = 106 _ (l-//0.65)
6 small structures, it may b e desirable to have quick and
0.65 simple p r o c e d u r e s to estimate t h e p e a k flow.
with q in cfs and t in hours, q at any time can be easily
determined. F o r example, q at t = 1 hr is
3.18
Rational Method
( l - 1/0.65) By far the most c o m m o n m e t h o d used for peak flow
q(l) = 106 e
= 75 cfs.
0.65 estimation is t h e Rational M e t h o d . A joint report by
Note that this is a 10-min unit hydrograph since E q . t h e A m e r i c a n Society of Civil E n g i n e e r s and the W a t e r
(3.55) and (3.58) were used. T h u s t h e effective rainfall Pollution Control F e d e r a t i o n is p e r h a p s t h e most com­
p a t t e r n should be developed in 10-min increments. prehensive t r e a t m e n t of t h e Rational M e t h o d (Water
Pollution Control F e d e r a t i o n , 1969). T h e Rational
M e t h o d has many limitations and shortcomings. A
r e p o r t by M c P h e r s o n (1969) discusses these problems
in considerable detail. In spite of the recognized short­
Runoff Hydrographs from Unit Hydrographs
comings of t h e R a t i o n a l M e t h o d , it continues to be
Several procedures are currently in use for estimat­ widely used because of its simplicity, e n t r e n c h m e n t in
ing runoff hydrographs using synthetic unit hydro- practice, coverage in texts, a n d lack of a comparable
graphs. In general, these procedures consist of devel­ alternative.
oping a design rainstorm, deducting abstractions from T h e Rational E q u a t i o n is
the rain to get the excess rainfall hyetograph, estimat­
ing t h e time to p e a k and t h e p e a k flow r a t e for a unit
dD = CiA, (3.70)
hydrograph, defining the remaining shape of t h e unit
hydrograph, and applying t h e unit hydrograph to t h e
derived rainfall excess hyetograph. T h e Santa B a r b a r a w h e r e q is t h e p e a k flow r a t e in cfs, C is a dimension­
p

U r b a n Hydrograph p r o c e d u r e and t h e routing of over­ less coefficient, i is t h e rainfall intensity in iph with a
land flow using a kinematic a p p r o a c h has b e e n previ­ d u r a t i o n equal to t , a n d A is the drainage area in
c

ously discussed. acres. T o b e dimensionally correct, a conversion factor


T h e Soil Conservation Service (1972) hydrograph of 1.008 should b e included to convert acre-inches per
procedure consists of: (1) estimating rainfall from t h e h o u r to cubic feet p e r second; however, this factor is
24-hr rainfall d e p t h and the a p p r o p r i a t e type curves of generally neglected.
Fig. 3.10, (2) estimating rainfall excess using E q . (3.21), T h e " r a t i o n a l e " b e h i n d t h e Rational Equation is
(3) estimating unit hydrograph timing p a r a m e t e r s from that if a steady rainfall occurs on a watershed, the
Eqs. (3.26) and (3.53) or Eqs. (3.47), (3.48), a n d (3.52), runoff rate will increase until t h e entire watershed is
(4) estimating unit hydrograph p e a k flow r a t e from E q . contributing runoff. If a rainfall of duration less than t c

(3.57), and (5) using a hydrograph s h a p e as shown in occurs, t h e entire basin would not be contributing so
Fig. 3.34 (either the dimensionless hydrograph or t h e t h e resulting runoff r a t e would be less than from a
triangular hydrograph may b e used). A n incremental rainfall with a d u r a t i o n equal to t . If a rainfall of
z

rainfall duration, £>, of t /3 is r e c o m m e n d e d so that


p
duration greater t h a n t occurs, t h e point relationship
c

D is computed from b e t w e e n average rainfall intensity and duration for a


given frequency show that t h e average intensity would
D = 0.4f . L (3.69)
be less than if t h e duration w e r e equal to t . T h u s it is
c
84 3. Rainfall-Runoff Estimation in Storm Water Computations

reasoned that a rainfall with an average p e a k intensity Table 3.24 Runoff Coefficients
of duration t produces the maximum flow rate.
c

T h e Rational Equation is based on certain assump­ Urban areas The use of average coefficients for various surface types,
tions. which are assumed not to vary through the duration of the storm, is
common. The range of coefficients, classified with respect to the general
1. T h e rainfall occurs uniformly over the drainage area. character of the tributary reported in use is:
2. T h e peak rate of runoff can be reflected by t h e Description of area Runoff coefficients
rainfall intensity averaged over a time period equal
to the time of concentration of the drainage area. Business
3. T h e frequency of runoff is the same as t h e fre­ Downtown areas 0.70 to 0.95
quency of the rainfall used in the equation. Neighborhood areas 0.50 to 0.70

T h e coefficient C is called the runoff coefficient and Residential

is the most difficult factor to accurately d e t e r m i n e . C Single-family areas 0.30 to 0.50


must reflect factors such as interception, infiltration, Multiunits, detached 0.40 to 0.60
surface detention, and antecedent conditions. Obvi­ Multiunits, attached 0.60 to 0.75
ously, n o single factor can quantify all of these things Residential (suburban) 0.25 to 0.40
and their effect on peak runoff rates. Apartment dwelling areas 0.50 to 0.70
Several studies have shown that C is not a constant Industrial
but varies with the frequency of the runoff event a m o n g
Light areas 0.50 to 0.80
other things ( H a a n , 1972; H o r n e r , 1910; M c P h e r s o n ,
Heavy areas 0.60 to 0.90
1969). T h e increase in C with r e t u r n period is gener­
Parks, cemeteries 0.10 to 0.25
ally attributed to wetter conditions that would b e ex­
Playgrounds 0.20 to 0.35
pected to exist during m o r e extreme events.
Many tables have b e e n p r e p a r e d for estimating C. Railroad yard areas 0.20 to 0.35

W h e n using these tables, care must be exercised to see Unimproved areas 0.10 to 0.30
that they reflect conditions that exist today. This is
necessary because urbanized areas of years ago t e n d e d Note: It is often desirable to develop a composite runoff coefficient
to have m o r e temporary storage in t h e form of road based on the percentage of different types of surface in the drainage area.
This procedure is often applied to typical 'sample' blocks as a guide to
ditches and many less impervious areas. Table 3.24 was
selection of reasonable values of the coefficient for an entire area.
taken from the W a t e r Pollution Control F e d e r a t i o n Coefficients with respect to surface type currently in use are:
(1969) report for urban areas and from Schwab et al.
Character of surface Runoff coefficients
(1971) for rural areas.
Average coefficients for composite areas may be Streets
calculated on an area weighted basis from
Asphaltic and concrete 0.70 to 0.95
Brick 0.70 to 0.85
(3.71) Roofs 0.75 to 0.95
Lawns; sandy soil

where C, is the coefficient applicable to t h e area A In Flat, 2% 0.05 to 0.10


r

areas where large parts are laid out in typical, r e p e a t ­ Average, 2 to 7% 0.10 to 0.15
ing p a t t e r n s such as subdivisions, the weighting factors Steep, 7% 0.15 to 0.20
and weighted C can be determined by considering a Lawns, heavy soil
single, typical layout. Flat, 2% 0.13 to 0.17
As with any estimation procedure, considerable care Average, 2 to 7% 0.18 to 0.22
should be exercised when applying t h e Rational E q u a ­ Steep, 7% 0.25 to 0.35
tion to estimate p e a k flows. For instance, the location
of relatively impervious areas with respect to the point
Note: The coefficients in these two tabulations are applicable for storms
of flow estimation must be carefully considered. If flow of 5-year to 10-year frequencies. Less frequent higher intensity storms
from an impervious area must cross an infiltrating a r e a will require the use of higher coefficients because infiltration and other
such as grass, the flows may be greatly reduced. If large losses have a proportionally smaller effect on runoff. The coefficients are
impervious areas are present, they should be analyzed based on the assumption that the design storm does not occur when the
ground surface is frozen.
as separate units. T h e reason for this can b e seen by
considering the situation shown in Fig. 3.39. In case A,
Estimation of Peak Runoff Rates 85

Table 3.24—Continued total flow time is 15 min. A weighted C is

Rural areas C = 0.5(0.9 + 0.2) = 0.55.


Soil texture
A 10-year, 15-min rain at Stillwater, Oklahoma, has an
Topography
i of 6.29 iph. T h u s ,
and Open sandy Clay and silt Tight
vegetation loam loam clay
<7totai = 0.55(6.29)^1 = 3 . 1 5 ^ .
Woodland
Considering only t h e impervious area, C is 0.9 and t is
Flat 0-5% slope 0.10 0.30 0.40 c

5 min. T h e corresponding / is 10.15 iph


Rolling 5-10% slope 0.25 0.35 0.50
Hilly 10-30% slope 0.30 0.50 0.60 q = 0 . 9 ( 1 0 . 1 5 ) ^ / 2 = 4.56,4,
imp

Pasture
Flat 0.10 0.30 0.40 or the 10-year flow p e a k from the lower impervious
Rolling 0.16 0.36 0.55 a r e a would exceed the original estimate for the entire
Hilly 0.22 0.42 0.60 area. This illustrates the need to carefully consider the
Cultivated location of impervious or high runoff-producing areas.
Flat 0.30 0.50 0.60
Of course, this same high flow rate would be estimated
to occur at the lower end of the impervious part of case
Rolling 0.40 0.60 0.70
B. H e r e , however, the flow through t h e pervious area
Hilly 0.52 0.72 0.82
would at least partially a t t e n u a t e this peak.
Finally, it must be kept in mind that this is a method
of p e a k flow estimation only. T h e t bears no relation
c

to the time from the beginning of a rainfall. O n e


cannot predict from this m e t h o d when the peak will
the impervious area is next to the outlet, while in case occur. T h e t applies to the time from the beginning of
c

Β the grass area is next to the outlet. Straightforward an intense rainfall with a duration equal to t . This c

application of the Rational Equation would result in intense rainfall may occur anytime during the rain­
the same flow estimate for both cases since t and the c
storm. T h e use of t h e Rational Equation to develop
weighted C would be the same. runoff hydrographs carries the m e t h o d well beyond its
A closer look at case A, however, shows that it is original intent. T a b u l a t e d values of the Rational C
possible that the peak flow from the impervious area have generally b e e n derived based on p e a k flows, not
alone could exceed the p e a k estimated for the whole entire hydrographs, and thus may not be valid when
area based on a weighted C. Using Eq. (3.49), a flow used to develop hydrographs.
time of 10 min is estimated for the 500 ft of grass.
Using a similar procedure, a flow time of only about 5
min is estimated for the impervious area. T h u s , t h e SCS-TR55 Method
T h e Soil Conservation Service (1986) presents a
m e t h o d for estimating p e a k flows from small catch­
Case A Case θ m e n t s based on an analysis of a large n u m b e r of

Τ
c o m p u t e r runs with their T R 2 0 c o m p u t e r program
(SCS, 1983). This p r o g r a m computes runoff hydro-
graphs in a m a n n e r analogous to the p r o c e d u r e de­
500' tailed earlier in this c h a p t e r a n d attributed to the SCS.
T h e results of t h e p e a k flow analysis can b e expressed
as

(3.72)
500'
w h e r e q is t h e p e a k discharge in cfs, q is the unit
p u

p e a k discharge in cfs p e r inch of runoff per square


mile, A is the drainage area in square miles, Q is the
Figure 3.39 Hypothetical runoff situations. runoff in inches from a 24-hr storm of the desired
86 3. Rainfall-Runoff Estimation in Storm Water Computations

Table 3.25 Adjustment Factor, F , for Pond p


Table 3.26 Coefficients for Eq. (3.73)
and Swamp Areas That Are Spread throughout
the Watershed Rainfall type c
o C, c 2

Percentage of pond I 0.10 2.30550 -0.51429 -0.11750


F
P
and swamp areas 0.20 2.23537 -0.50387 -0.08929
0.25 2.18219 -0.48488 -0.06589
0 1.00
0.30 2.10624 -0.45695 -0.02835
0.2 0.97
0.35 2.00303 -0.40769 0.01983
1.0 0.87
0.40 1.87733 -0.32274 0.05754
3.0 0.75
0.45 1.76312 -0.15644 0.00453
5.0 0.72
0.50 1.67889 -0.06930 0.0

IA 0.10 2.03250 -0.31583 -0.13748


0.20 1.91978 -0.28215 -0.07020
frequency, and F is a pond and swamp adjustment
p 0.25 1.83842 -0.25543 -0.02597
factor. 0.30 1.72657 -0.19826 0.02633
Q is computed directly from the curve n u m b e r equa­ 0.50 1.63417 -0.09100 0.0
tion, Eq. (3.21), based on a 24-hr Ρ with a r e t u r n
period equal to the desired return period of the peak
II 0.10 2.55323 -0.61512 -0.16403
flow. F is taken from Table 3.25 assuming t h e p o n d s
p
0.30 2.46532 -0.62257 -0.11657
a n d / o r swampy areas are distributed t h r o u g h o u t the
0.35 2.41896 -0.61594 -0.08820
watershed. T h e value of q is computed from u

0.40 2.36409 -0.59857 -0.05621


0.45 2.29238 -0.57005 -O.02281
l o g ( * ) = C + C log t + C ( l o g t )\
u 0 x c 2 c (3.73)
0.50 2.20282 -0.51599 -0.01259

where the C's come from Table 3.26, a n d logarithms to


III 0.10 2.47317 -0.51848 -0.17083
the base 10 are used. T h e t in hours has limits of 0.1
c

0.30 2.39628 -0.51202 -0.13245


to 10 hr. In Table 3.26, the value of 7 is given by a

/ = 0.25. 5 can be computed from Eq. (3.22) using the 0.35 2.35477 -0.49735 -0.11985
a

appropriate curve number. O n c e I /P is computed, 0.40 2.30726 -0.46541 -0.11094


a

the line in Table 3.26 with the closest value of IJP to 0.45 2.24876 -0.41314 -0.11508
the computed value should be used. Linear interpola­ 0.50 2.17772 -Ό.36803 -0.09525
tion may be used in Table 3.26; however, care must be
exercised to see that all t h r e e C coefficients are consis­
tent with each o t h e r and with other tabulated values. A
graphical solution to Eq. (3.72) from the SCS (1986) is
presented in Fig. 3.40. A c o m p u t e r p r o g r a m that im­
plements T R 5 5 methods has b e e n developed (SCS,
1986). w h e r e q a n d R are the Γ-year p e a k flow and rain­
T T

fall; A is the area, S the slope, / the impervious


fraction, L t h e axial length of the watershed, and a, b,
Frequency Method
c, d, e, and g are coefficients.
Hydrologic frequency m e t h o d s are often combined Generally, it is difficult to find a d e q u a t e streamflow
with regression analyses or correlation studies t o de­ records for estimating q for Τ over about 10 years.
T

velop empirical peak flow prediction equations. T h e Various studies use different combinations of variables
method requires observed streamflow records from in relationships like Eq. (3.74). It is important to deter­
which estimates are m a d e of peak flows for various mine the conditions u n d e r which these relationships
return periods. These peak flows are then related to were developed before they are applied to a particular
watershed physical conditions and rainfall. A typical situation. Often they do not extrapolate well to condi­
relationship might be tions outside those u n d e r which they were developed.
This approach was discussed in C h a p t e r 2 u n d e r " R e ­
q = aA S I L R \
b c d e
(3.74) gional Analysis" w h e r e Eq. (2.38), which is similar in
T T

form to Eq. (3.74), was suggested.


Estimation of Peak Runoff Rates 87

.1 .2 .4 .6 .8 1 2 4 6 8 10
Time of concentration ( T ) , hours
c

Figure 3.40 Graphical solution for Eq. (3.72). SCS rainfall distribution (a) type I, (b) type 1A, (c) type II, and (d)
type III.
88 3. Rainfall-Runoff Estimation in Storm Water Computations

.4 .6 .8 1 2 8 10
Time of concentration ( T ) , hours
c

Time of concentration (T ), hours


c

Figure 3.40 Continued


Estimation of Peak Runoff Rates 89
From Fig. 3.21 for Ρ = 4.25 and CN = 62.5, read Q = 1.03
in. Q can also be computed from Eqs. (3.21) and (3.22).
Therefore, the runoff volume from the 10-year, 24-hr rainfall
before mining is 1.03 in. or 1.03/12 X 100 = 8.6 acre-feet.
After mining, the disturbed area is 180 ft X 4500 ft or 18.6
acres. The undisturbed area is 100 - 18.6 = 81.4 acres.
The CNs on the disturbed area (Table 3.16) are Musk­
ingum (HSG C, CN = 88) and Shelocta (HSG B, CN = 81):
[(18.6/2)(88) + (18.6/2)(81) + (81.4/2)(70) + (81.4/2)(55)]
CN =
100
= 66.6.
From Fig. 3.21 for Ρ = 4.25 and CN = 66.6, read
Q = 1.28 inches
= 10.63 acre-feet.
b. Equation (3.72) will be used to calculate the peak flow
in the diversion. The maximum length of flow within the
diversion depends on the direction the diversion is sloped.
Here we assume the diversion slopes toward the nearest
\ main draw, so the maximum flow length is about 600 ft if the
draws are evenly spaced. We use 700 ft to reflect the uneven
Figure 3.41 Watershed for Example Problem 3.14.
spacing and to be conservative. The maximum area draining
into a diversion is thus
700 X 500
A = ————— = 8 acres.
43560
Example Problem 3.14. Impact of mining The time of concentration is the time it takes overland
on runoff hydrograph flow to travel 500 ft down the slope plus 700 ft to the

The watershed shown in Fig. 3.41 is located in Eastern


Kentucky. One seam of coal is to be surface mined as Table 3.27 Unit Hydrograph for Example
indicated. The entire watershed is 100 acres in size. The Problem 3.14
average width of the area to be stripped is 180 ft. The
stripped area is about 4500 ft long. The average width of the t
area above the stripped area is 500 ft, and it is about 3500 ft
0 0 0 0
in length. The watershed is predominately forested. The soils
are predominately Muskingum and Shelocta with about 50% 5 0.25 0.15 34
of the watershed in each. Average land slopes are around 10 0.50 0.50 114
55% with stream slopes averaging 5%. The maximum eleva­ 15 0.75 0.87 197
tion rise in the watershed is 450 ft, and the maximum flow 20 1.00 1.00 227
length is around 5000 ft. A diversion channel is to be placed
25 1.25 0.90 204
along the upper periphery of the area to be stripped.
30 1.50 0.67 152
a. Estimate the 10-year, 24-hr runoff volume before and 35 1.75 0.43 98
immediately after mining.
40 2.00 0.28 64
b. Estimate the 10-year peak inflow into the diversion chan­
45 2.25 0.19 43
nel.
c. Estimate the 25-year runoff hydrograph both before and 50 2.50 0.13 30
after mining. 55 2.75 0.08 18
60 3.00 0.05 11
Solution:
a. From Appendix 3A, the 10-year, 24-hr rain is found to 65 3.25 0.04 9
be 4.25 in. 70 3.50 0.03 7
Appendix 3B indicates that the hydrologic soil groups are 75 3.75 0.02 5
Muskingum (HSG C) and Shelocta (HSG B). The CNs before
80 4.00 0.01 2
mining (Table 3.16) are Muskingum (HSG C, CN = 70) and
Shelocta (HSG B, CN = 55):
Sum 1215

ΓΝ= [0.5(70) + 0.5(55)]/(0.5 + 0.5) = 62.5.


90 3. Rainfall-Runoff Estimation in Storm Water Computations

diversion outlet. The overland flow velocity is estimated from computed from I = 0.25, where 5 is given by
a

Eq. (3.48). The coefficient a from Table 3.20 is 2.5. The


overland flow velocity is 1000 1000
5
= 7w - 1 0
=^ - 1 0
= · 6

υ = 2.5(0.55) 1/2
= 1.85 or about 2 fps.
Therefore I = 0.2(6) = 1.2. I /P is 1.2/4.25 or 0.28. Figure
a a

3.11 shows Kentucky is in the SCS type II rainfall region.


A reasonable design velocity estimate for the diversion yet Table 3.26 gives C = 2.47, C = - 0 . 6 2 , and C = - 0 . 1 2 .
0 x 2

to be designed is 6 fps. Equation (3.73) gives

500 ft 700 ft l o g ( ? ) = 2.47 - 0.621og(0.1) - 0.12[log(0.1)] = 2.97


u
2

t = —— + — — = 250 + 116 « 6 min.


c
c
2 fps 6 fps
or

The CN for the area above the diversion is 62.5. Therefore q = io


u
29 7
= 933 c f s / i n . / m i l e . 2

the 10-year, 24-hr runoff volume is 1.03 in. as computed


earlier. The pond factor, F , from Table 3.25 is 1.0. I is
p a A similar estimate can be read from Fig. 3.40. Equation

Table 3.28 Effective Rainfall for Example Problem 3.14

Premining Postmining
CN = 62.5 CN = 66.6

Accumulated Incremental Accumulated Incremental


Accumulated effect effect effect effect
Time Ordinate* 1
rain* rain c
rain'' rain c
rain''

11.33 0.26 1.30 0.00 0.00 0.00 0.00

11.42 0.27 1.35 0.00 0.00 0.02 0.02

11.50 0.29 1.45 0.00 0.00 0.04 0.02

11.58 0.31 1.55 0.02 0.02 0.05 0.01


11.67 0.34 1.70 0.04 0.02 0.09 0.04

11.75 0.39 1.95 0.08 0.04 0.15 0.06

11.83 0.51 2.55 0.25 0.17 0.36 0.21

11.92 0.62 3.10 0.46 0.21 0.62 0.26

12.00 0.66 3.30 0.54 0.08 0.72 0.10

12.08 0.68 3.40 0.59 0.05 0.78 0.06

12.17 0.69 3.45 0.61 0.02 0.80 0.02

12.25 0.71 3.55 0.66 0.05 0.86 0.06

12.33 0.72 3.60 0.69 0.03 0.87 0.01

12.42 0.73 3.65 0.71 0.02 0.91 0.04

12.50 0.73 3.65 0.71 0.00 0.91 0.00

12.58 0.74 3.70 0.74 0.03 0.94 0.03

12.67 0.74 3.70 0.74 0.00 0.94 0.00

12.75 0.75 3.75 0.76 0.02 0.97 0.03

12.83 0.75 3.75 0.76 0.00 0.97 0.00

12.92 0.76 3.80 0.79 0.03 1.00 0.03

13.00 0.77 3.85 0.81 0.02 1.03 0.03

0.81 1.03

"From Fig. 2.19.


h
25-year, 24-hr rainfall of 5.00 inches x Column 2.
'From Fig. 2.26 using rain in Column 3 and appropriate CN.
''Difference in successive values in previous column.
Estimation of Peak Runoff Rates 91
(3.72) gives q as p

For computational convenience, a t of 20 min is selected. p

Qp = Qu Q p A f
= 9 3 3 x
(8/640) X 1.03 X 1 = 12 cfs. The peak flow rate is estimated from Eq. (3.57) as
c. Runoff hydrograph before mining: Based on informa­
tion given in the problem statement, the lag time can be 484Λ 484(100/640)
estimated from Eq. (3.53) as Qo = = — ~ = 227 cfs.
0.7
H p
t 20/60
L°*(S + 1)
p

t, =
1900y 05

The unit hydrograph shape is taken from the dimension­


1000 1000 less unit hydrograph in Fig. 3.34.
S
= ~CN - = ^ ? - 1 0 1 0
= 6
The unit hydrograph ordinates are shown in Table 3.27.
L = 5000 ft The volume of the unit hydrograph can be checked by using
Eq. (3.46):
Y = 55%
(5000) (7) 0 8 0 7

5 X 1215
'<•- i90oyo^ = 0 2 5 h r
- V = = 1.00 in.
60.5Λ 60.5 X 100
Use t = 15 min.
L

Since D should be one-fifth to one-third of t and 5 min is p

a convenient time increment, a D of 5 min is selected. The t To get an effective rainfall pattern, a 25-year, 24-hr rainfall
p

can be computed from Eq. (3.26) as pattern is used. With a CN of 62.5, Fig. 3.21 or the relation­
ship Ρ = 0.2S shows that 1.20 in. of rain must occur before
tp - ' L + D/2 any runoff starts. The 25-year, 24-hr rainfall is 5.00 in. The
t p = 15 4- 5 / 2 = 17.5. time runoff begins is determined when 1.20 in. of rain occurs.
This corresponds to a curve ordinate of 1.20/5.00 = 0.24 or

Table 3.29 Calculation of Runoff Hydrograph for Example Problem 3.14

Time: 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120


UH: 0 34 114 197 227 204 152 98 64 43 30 18 11 9 7 5 2 0

Rain

0.02 0 1 2 4 5 4 3 2 1 1 1 0
0.02 0 1 2 4 5 4 3 2 1 1 1 0

0.04 0 1 5 8 9 8 6 4 3 1 1 ! 0
0.17 0 6 19 33 39 35 26 17 11 7 5 3 2 2 1 1 0
0.21 0 7 24 41 48 43 32 21 13 9 6 4 2 2 1 1 0
0.08 0 3 9 16 18 16 12 8 5 3 2 1 1 1 1 0

0.05 0 2 6 10 11 10 8 5 3 2 2 1 1 ! 0
0.02 0 1 2 4 5 4 3 2 1 1 1 1
0.05 0 2 6 10 11 10 8 5 3 2 2 1 1 0
0.03 0 1 3 6 7 6 5 3 2 1 1 1 0

0.02 0 1 2 4 5 4 3 2 1 1 1 0
0.00 0 0 0 0 0 0 0 0 0 0
0.03 0 1 3 6 7 6 5 3 2 1 1 1
0.00 0 0 0 0 0 0 0 0 0 0 0

0.02 0 1 2 4 5 4 3 2 1 1 1 0
0.00 0 0 0 0 0 0 0 0 0 0 0
0.03 0 1 3 6 7 6 5 3 2 1 1 1
0.02
0 1 2 4 5 4 2 2 1 1 1

Total: 0 1 3 7 20 43 76 104 115 107 92 76 60 50 39 32 26 23 22 20 19 14 11 8 5 2 2 2


92 3. Rainfall-Runoff Estimation in Storm Water Computations

180 r 120

100
POST MINING

80
TRIANGULAR UH

60
ο

40

20

0 10 20 30 40 50 60 70 80 90 100
Time (min)

Figure 3.43 Comparison of runoff hydrographs using a curvilinear


and a triangular unit hydrograph.

100 200

TIME (min)
Figure 3.42 Hydrograph for Example Problem 3.14. Equation (3.46) applied to the runoff hydrograph results in

ΔίΣ4, 5 Χ 979
V = = 0.81 in.
60.5Λ =
60.5 Χ 100

a time of about 11 hr on the type II curve. Table 3.28 shows


the calculation of the effective rainfall for the watershed Thus the calculations in Table 3.29 are accurate.
prior to mining. The postmining hydrograph is similarly calculated. The
unit hydrograph parameters change since the C N changes.
Similarly, the effective rainfall pattern for the watershed
For the postmining condition, the following parameters were
after mining is shown in Table 3.28. About 1.00 in. of rain
used a n d / o r determined.
must occur before runoff starts when the CN is 66.6; 1.00/5.0
corresponds to a type II ordinate of 0.20 or a time of 10.33
hr.
In Table 3.28 the effective rainfall is shown as being CN = 66.6 S = 5.01 in.
calculated beginning at a time of 11.33 hr. The reason that
Y = 55% (224= 1.77 in.
the column for CN = 62.5 does not start at a time of 11 hr
L = 5000 ft t = 13.6 min
and for CN = 66.6 a time of 10.33 hr is because the values, L

although finite, are essentially zero for times prior to 11.33 D = 5 min f = 15 min
p

hr. A = 100 acres q = 302.5 cfs


p

The runoff hydrograph is calculated by applying Eq. (3.43) P 2 4 = 5.00 in. Κ = 3.77
to the effective rainfall and unit hydrograph. Table 3.29
shows a tabular procedure for calculating the hydrograph.
The unit hydrograph coordinates run across the top of the
table and the effective rainfall along the left margin. The Based on these parameters, the effective rainfall of Table
table entries are the product of the effective rainfall and the 3.28 and a unit hydrograph defined by Eq. (3.59), the runoff
unit hydrograph ordinates. Each line in the table is moved hydrograph shown in Fig. 3.42 was calculated.
one time increment to the right. The runoff hydrograph is To show the insensitivity of the runoff hydrograph to the
obtained by summing the table entries by columns. Figure shape of the unit hydrograph, t and q were held constant,
p p

3.42 is a plot of the resulting hydrograph. and the runoff hydrograph representing the premining condi­
tion was computed using the SCS triangular unit hydrograph.
Based on an accumulated rainfall of 3.85 in. and a C N =
The results of this calculation are compared with the results
62.5 or 5 = 6, Eq. (3.21) estimates the runoff volume as
of using the curvilinear unit hydrograph in Fig. 3.43. From
this figure, it is apparent that the shape of the unit hydro-
graph had little effect on the runoff hydrograph since the unit
(Ρ - 0.25) (3.85 - 1.2) 2
hydrograph had the same t and q ; and several blocks of
= 0.81 in.
p p

Ρ + 0.85 =
3.85 + 4.8 effective rainfall were included.
Long-Term Water Balances 93

system would be precipitation and applied waste water.


Generally surface runoff would b e prevented from en­
Example Problem 3.15. Estimation
tering or leaving t h e area. Outflow would be evapo-
of peak discharge
transpiration, d e e p seepage, and runoff. T h e change in
Estimate the peak discharge from the watershed described storage would b e t h e change in t h e volume of water
in Example Problem 3.14. Use the premining watershed stored in the soil profile.
condition and a 25-year frequency. T h e s e two situations, ponds and vegetated plots, are
considered in this section. V e g e t a t e d plots are consid­
Solution: Use Eq. (3.72). The ponding factor is 1.0. From
e r e d u n d e r two conditions: (1) T h e soil profile is kept
previous calculations
very wet to p r o m o t e maximum evapotranspiration ( E T )
tL = 0.25 hr. and (2) the soil profile is allowed to dry naturally so
From Eq. (3.52), that E T may be limited by the soil water content.
W a t e r balances can be c o m p u t e d on the basis of
l = i / 0 . 6 = 0.42 hr.
c L
time intervals of minutes u p to 1 year. Using long time
The 25-year, 24-hr rainfall is 5.0 in. The CN is 62.5. The intervals can provide quick estimates of t h e likely water
runoff volume from Fig. 3.21 or Eq. (3.21) is 1.47 in. balance for an a r e a but will not correctly reflect stor­
The values for C , C and C are the same as given in
0 v 2
age and flow processes that have short time scales. For
part b of Example Problem 3.14. q may be taken from Fig. u example, if a time interval of 1 m o n t h is used, then a
3.40 or computed from Eq. (3.73) as rainfall of 5 in. occurring in 1 day will enter the
computations in t h e same way as if the 5 in. were
log<? = 2.47 - 0.621og(0.42) - 0.12(log0.42) = 2.6866 2
u
spread out over several days. Obviously the water bal­
q - 10
u
26 8 6 6
= 486. ance of a vegetated plot will be sensitive to this differ­
ence since 5 in. of rain in 1 day could easily produce
From Eq. (3.72),
considerable runoff, while the same rain spread over
several days might p r o d u c e n o runoff.
= 486(100/640)1.47(1) = 112 cfs. Two terms in the water balance equation that are
very difficult to estimate are evaporation or evapotran­
Note that this peak agrees with the peak of the runoff spiration and seepage. T h e r e is n o general way that
hydrograph of 115 cfs shown in Fig. 3.42. seepage can be c o m p u t e d i n d e p e n d e n t of the other
water balance terms without field data. Often waste
disposal sites and ponds are designed to minimize
seepage. W h e r e seepage is present, it is often com­
LONG-TERM WATER BALANCES p u t e d as a residual t e r m after all o t h e r processes in t h e
water balance are accounted for. T h u s the magnitude
Often in the design of ponds and waste water dis­ of seepage is taken as the flow required to balance the
posal sites it is necessary to c o m p u t e long-term water water balance equation.
balances. Examples would include evaporation p o n d s , Evaporation d a t a in the form of pan evaporation
waste water disposal on vegetated surfaces, and com­ exists for many sites in t h e U . S . and o t h e r parts of the
putation of potential seepage from soil, fill, or waste world. Since evaporation pans are subject to solar
disposal areas. T h e basic equation governing water heating and have m o r e exposure than a body of water,
balances is E q . (3.1) p a n evaporation generally exceeds lake or pond
evaporation. Coefficients are typically applied to pan
/ - Ο = AS,
evaporation d a t a to estimate lake evaporation. Pan
w h e r e / a n d Ο represent all inflows t o a n d outflows coefficients vary seasonally and geographically, but a
from a control volume and AS represents the change reasonable average value is 0.7 for standard U.S.
in the volume of water stored in the control volume. If W e a t h e r B u r e a u Class A pans.
an evaporation pond is the control volume, inflows Figures 3.44 and 3.45 show m e a n annual rainfall and
would consist of precipitation falling on t h e pond, lake evaporation for the continental U.S. For those
runoff entering the pond, and waste water p u m p e d into areas w h e r e evaporation exceeds precipitation, it is
the pond. Outflow would consist of discharges, seepage theoretically possible to dispose of water through the
losses, and evaporation from the pond. use of evaporation ponds. T h e net quantity of water
A control volume might also b e a vegetated plot on that can be evaporated on a long-term average is the
which waste water is being sprinkled. If all subsurface difference in m e a n annual precipitation and evapora­
flows take place in the vertical direction (no horizontal tion. For example at O k l a h o m a City, Oklahoma, the
seepage into or out of the soil), t h e n inflow to t h e m e a n annual precipitation is about 32 in. (810 mm),
94 3. Rainfall-Runoff Estimation in Storm Water Computations

Figure 3.44 Mean annual precipitation in the U.S. (in.).

Figure 3.45 Mean annual lake evaporation in the U.S. (in.).


Long-Term Water Balances 95

Table 3.30 Mean Monthly Precipitation and using several years of data. T h e results should be
Lake Evaporation, Nashville, Tennessee t r e a t e d in a probabilistic m a n n e r such as flood flows
were t r e a t e d in C h a p t e r 2. T h u s , o n e might conclude
Month Precipitation (in.) Evaporation (in.)
that a p o n d of a given size would b e a d e q u a t e in 9 0 %
January 5.49 0.9 of all years. W h e r e the probability of the disposal
February 4.51 1.3
p o n d s being i n a d e q u a t e is appreciable, alternate means
of handling t h e wastes may n e e d to be available.
March 5.19 1.9
Various publications of the U.S. W e a t h e r Bureau
April 3.74 3.3
contain data on precipitation and evaporation. T h e
May 3.72 4.1
d a t a in Table 3.30 w e r e t a k e n from T o d d (1970). This
June 3.25 5.1
type of data is very site specific. For example, average
July 3.72 5.8 annual shallow lake evaporation varies from about 100
August 2.86 5.4 in. (2540 m m ) in southwestern New Mexico to as low as
September 2.87 4.9 15 in. (380 m m ) in n o r t h e r n M a i n e and northwestern
October 2.32 3.7 Washington in t h e U.S.
November 3.28 2.1
December 4.19 1.1

Example Problem 3.16. Estimation of evaporation

For the months of May through October, monthly rainfall


totals of 3.0, 3.5, 2.5, 2.5, 6.4, 3.4, and 4.0 were recorded at
and the m e a n annual lake evaporation is about 66 in. Nashville. How many gallons of waste water could be evapo­
(1675 mm). T h u s a net of about 34 in. (865 m m ) of rated from a 1-acre pond assuming average evaporation and
water is lost from the surface of small p o n d s and no other inflow or outflows?
shallow reservoirs. O n the other hand, n e a r Nashville,
Solution: Monthly evaporation is taken from Table 3.30.
Tennessee, the m e a n annual precipitation is about 45 The change in storage, AS, is calculated from Eq. (3.1),
in. (1140 mm), and t h e m e a n annual evaporation is where precipitation is / and evaporation is O.
about 39 in. (990 mm), which implies that a p o n d or
shallow lake in this vicinity would have a net gain of Month Precipitation Evaporation AS
water from precipitation over evaporation.
Table 3.30 shows m e a n monthly precipitation and May 3.0 4.1 -1.1

evaporation for Nashville, T e n n e s s e e . This table shows June 3.5 5.1 -1.6
that even though the average annual precipitation ex­ July 2.5 5.8 -3.3
ceeds the average annual evaporation, the m o n t h s of Aug 2.5 5.4 -2.9
May through October have the opposite result with Sep 6.4 4.9 1.5
these months showing a net of 10.2 in. (259 m m ) of Oct 4.0 3.7 0.3
evaporation over precipitation. This in t u r n implies
that a 1-acre (0.40 hectare) evaporation p o n d could
Total 21.9 29.0 -7.1
dispose of about 1500 g a l / d a y (5675 l i t e r s / d a y ) of
waste water on t h e average over this 6-month period.
A net loss of 7.1 in. of water occurs. A depth of 7.1 in. of
This is calculated as
water over 1 acre converts to 193,000 gal.
10.2 in. 43560 f t 2
7.48 gal 1 acre Comment: From this problem the need for alternate dis­
X χ X posal systems or large factors of safety are apparent. If
12in./ft acre ft 3
182 days monthly rainfall had been only 1 in. greater each month, a
very likely event, the disposal capacity would be reduced by 6
= 1522 gpd. in. to only about 30,000 gal.
T h e stochastic n a t u r e of rainfall and evaporation
makes the design of evaporation facilities based on
short-term averages extremely risky. In some years,
precipitation may exceed evaporation over the design Evapotranspiration
period (6 months in this case). Generally, a detailed Evaporation from vegetated surfaces is t e r m e d evap­
continuous simulation of t h e water balance in a p o n d otranspiration ( E T ) . E T is m o r e complex than evapora­
should b e done to assess the effectiveness of the p o n d tion since plant and soil factors affect the process. T h e
as a waste disposal facility. Such a simulation could b e r a t e at which E T occurs from a well-watered, actively
done based on monthly precipitation and evaporation growing, completely vegetated surface is t e r m e d poten-
96 3. Rainfall-Runoff Estimation in Storm Water Computations

tial evapotranspiration (PET). P E T equations are soil water content achieved after significant vertical
a b u n d a n t and are generally based on a combination of drainage of water from t h e soil d u e t o gravity has
theoretical and empirical considerations. For example, stopped. A t this point, w starts decreasing toward 0.
one approach to estimating P E T is to set it equal to a Since the function is not well defined, some use a
coefficient times p a n evaporation. T h e coefficient used sigmoidal decrease. Most m a k e t h e decrease in w a
may range from 0.6 to over 1.0 with t h e higher coeffi­ function of t h e actual soil water content. A simple
cient applicable to some forested areas. Theoretical model is to assign a value of 1 to w until the total soil
approaches are based on energy budgets, mass transfer water content in t h e root zone d r o p s 1 in. below field
relationships, or a combination of these approaches. capacity. T h e value of w is t h e n linearly decreased as a
Generally, experimentally determined constants are re­ function of soil water content in t h e root zone to 0 at
quired even for the theoretical approaches. t h e wilting point. B u r m a n et al. (1983) present t h e
W h e n a soil starts to dry, actual evapotranspiration relationship
( A E T ) may be reduced below P E T because the rate of
movement of water through t h e soil into plant roots w = ln(AW + l)/ln(101), (3.75)
may limit the transpiration rate. T h u s
w h e r e A W is t h e available soil water as a p e r c e n t a g e
A E T = P E T X w,
( A W = 100 at field capacity and 0 at t h e wilting point).
where w is a function of the soil water content and Several expressions for P E T a r e available. Many are
plant factors, including such things as stage of growth d e p e n d e n t on plant factors or solar radiation d a t a that
and type of plant. T h e factor w has b e e n extensively a r e not readily available. T h e Soil Conservation Service
investigated, yet a single expression for it has not b e e n (1970) presents the Blaney-Criddle p r o c e d u r e for esti­
developed. For some deep-rooted and well-watered mating water evapotranspiration w h e n water is not
vegetation such as alfalfa and some trees, w may ex­ limiting. Such plant water use is t e r m e d consumptive
ceed 1. For many types of vegetation, the u p p e r limit use and d e p e n d s on t e m p e r a t u r e a n d length of day.
of w is unity. For any vegetation, as t h e soil dries, a Multiplying t h e m e a n monthly t e m p e r a t u r e (f) by t h e
point w h e r e soil factors become limiting a n d A E T falls possible monthly p e r c e n t a g e of daytime hours of t h e
below P E T or w falls below unity will b e reached. T h e year (p) gives a monthly consumptive use factor ( / ) . It
w continues to decline until it approaches a value of is assumed that c r o p consumptive use varies directly
zero, indicating that transpiration has ceased. Some with this factor w h e n an ample water supply is avail­
research has indicated that the point w h e r e A E T b e ­ able. Mathematically, u = kf and U = sum of kf = KF,
comes essentially zero is at or just below the p e r m a ­ w h e r e U is t h e consumptive use in inches for t h e
nent wilting point of the soil. T h e p e r m a n e n t wilting growing season, Κ is t h e empirical consumptive use
point is the soil water content at which plants wilt and crop coefficient for t h e growing season (this coefficient
cannot recover. is crop d e p e n d e n t ) , F is the sum of monthly consump­
From this description of A E T , it is a p p a r e n t that w tive use factors for t h e growing season (sum of prod­
may range from above 1 at soil saturation to 0 for a dry ucts of m e a n monthly t e m p e r a t u r e and monthly per­
soil. Generally the value remains n e a r 1 until t h e soil centage of daylight h o u r s of t h e year), u is the monthly
dries to some value of water content that is generally consumptive use of t h e crop in inches, k is t h e empiri­
below the field capacity of the soil. Field capacity is the cal consumptive use coefficient for a m o n t h (varies by

Table 3.31 Monthly Percentage of Daytime Hours (p)

Latitude
North Jan Feb March April May June July Aug Sep Oct Nov Dec

65 3.52 5.13 7.96 9.97 12.72 14.15 13.59 11.18 8.55 6.53 4.08 2.62

60 4.70 5.67 8.11 9.69 11.78 12.41 12.31 10.68 8.54 6.95 5.02 4.14

50 5.99 6.32 8.24 9.24 10.68 10.92 10.99 9.99 8.46 7.44 6.08 5.65

40 6.75 6.72 8.32 8.93 10.01 10.09 10.22 9.55 8.39 7.75 6.73 6.54

30 7.31 7.02 8.37 8.71 9.54 9.49 9.67 9.21 8.33 7.99 7.20 7.16

20 7.75 7.26 8.41 8.53 9.15 9.02 9.24 8.95 8.29 8.17 7.58 7.65

10 8.14 7.47 8.45 8.37 8.81 8.61 8.85 8.71 8.25 8.34 7.91 8.09

0 8.5 7.67 8.49 8.22 8.49 8.22 8.50 8.49 8.21 8.49 8.22 8.5
Long-Term Water Balances 97

Table 3.32 Monthly Crop Growth Stage Coefficient (k ) c

Crop Jan Feb March April May June July Aug Sept Oct Nov Dec

Alfalfa 0.63 0.73 0.86 0.99 1.08 1.13 1.11 1.06 0.99 0.91 0.78 0.64
Grass 0.49 0.57 0.73 0.85 0.90 0.92 0.92 0.91 0.87 0.79 0.67 0.55

crop), / is the monthly consumptive use factor (prod­ loss from t h e reservoir assuming no inflow or outflow
uct of m e a n monthly t e m p e r a t u r e and monthly per­ o t h e r than as specified in the problem.
centage of daylight hours of t h e year), / = tp/100, t is (3.3) Develop a 3-hr, 25-year rainstorm for Chicago,
the m e a n monthly air t e m p e r a t u r e in degrees F a h r e n ­ Illinois, or a n o t h e r selected location in 15-min time
heit, and ρ is t h e monthly percentage of daylight hours increments based on (a) the appropriate SCS type
in the year (Table 3.31). curve, (b) d e p t h - d u r a t i o n - f r e q u e n c y data, and (c) the
T h e following modifications have b e e n m a d e to the Chicago hyetograph. C o m p a r e results by plotting them
method by the Soil Conservation Service. They define on a single chart.
k as the product of k and k , where k = 0.0173* -
t c t (3.4) Develop a 2-hr, 10-year rainstorm for St. Louis,
0.314 for t between 36 and 100° F. For t < 36° F, k is Missouri, or some o t h e r selected location in 10-min
taken as 0.3. T h e value of k may be taken from Table
c time increments based on (a) t h e appropriate SCS type
3.32. T h e Soil Conservation Service presents a m o r e curve, (b) d e p t h - d u r a t i o n - f r e q u e n c y data, and (c) the
detailed tabulation for k . c Chicago hyetograph. C o m p a r e the results by plotting
them on a single chart.
(3.5) D o p r o b l e m (3.3) for New Orleans, Louisiana,
Example Problem 3 . 1 7 . Monthly water use a n d / o r Seattle, Washington.
(3.6) Secure a nonlinear regression computer pro­
Estimate the monthly water use for a well watered grass at gram and use it to estimate t h e coefficients of Eq. (3.4)
Stillwater, Oklahoma, for the months of April through Octo­ for Chicago, Illinois, or some o t h e r location.
ber. Stillwater is approximately N36°. Average monthly tem­ (3.7) For a constant duration, T, Eq. (3.4) can be
peratures in degrees Farenheit are shown in Table 3.33. written i = K'F . T h e coefficients K' and χ can then
X

Solution: The calculation are shown in Table 3.33. The be estimated by taking logarithms of both sides of the
factor ρ is from Table 3.31, / is rp/100, k = 0.0173* - t
resulting equation a n d using linear regression tech­
0.314, k is from Table 3.32, k = k k , and u = kf. The total
c t c
niques. K' will b e a function of the constant value of Τ
water use for the period is the sum of the monthly values. selected. By repeating this process for several values of
Γ, a plot of K' versus Τ can be constructed. T h e
coefficient χ may also exhibit some variability as Τ
Problems changes. A n average value of χ can be chosen. Rainfall
intensity for any duration and frequency can then be
(3.1) A 927-acre catchment has b e e n monitored for
estimated from the equation by first estimating K'
many years. It has b e e n found that the average annual
rainfall over the catchment is 37 in. and that the
average annual streamflow is 1 cfs. Geologic investiga­
tions reveal a bedrock underlying the catchment that Table 3.33 Calculations for Example Problem 3.17
effectively prevents any d e e p seepage or groundwater
recharge. Estimate the average evapotranspiration in
Month Ρ / Κ k u

inches for this catchment. April 60 8.8 5.28 0.72 0.85 0.62 3.25
(3.2) T h e volume of water stored in a reservoir on May 68 9.8 6.66 0.86 0.90 0.78 5.17
July 1 was 124,000 acre-feet and on August 1 it was June 77 9.8 7.55 1.02 0.92 0.94 7.07
122,000 acre-feet. T h e surface area of the reservoir on
July 82 9.9 8.12 1.10 0.92 1.02 8.25
these two dates was 5650 and 5600 acres, respectively.
Aug 81 9.4 7.61 1.09 0.91 0.99 7.53
Pan evaporation for the period was 7 in. A lake-to-pan
Sep 73 8.4 6.13 0.95 0.87 0.83 5.06
evaporation coefficient of 0.7 is applicable. During the
Oct 61 7.8 4.76 0.74 0.79 0.59 2.79
month of July the average stream flow into t h e reser­
voir from its 150 m i l e drainage area was 25 cfs.
2

Outflow was a constant 30 cfs. T h e catchment and Total 39.12


reservoir received 3.5 in. of rain. Estimate the seepage
98 3. Rainfall-Runoff Estimation in Storm Water Computations

from the K' versus Τ plot and t h e n calculating i from problem (3.3). Surface storage is 0.25 in. Assume that
ι = K'F*. Use this approach to construct a D D F plot t h e soil is initially dry.
for the location used in problem (3.4). (3.14) Based on t h e SCS curve n u m b e r equation,
(3.8) C o m p u t e the defining p a r a m e t e r s for the rain­ using a curve n u m b e r of 80 estimate the effective
fall patterns shown in Fig. 3.14 for a 2-hr, 10-year rainfall p a t t e r n for the rain of problem (3.3).
rainfall for St. Louis, Missouri. Use Τ = 30 min. Plot (3.15) Based on t h e G r e e n - A m p t infiltration rela­
the results on the graph of problem (3.4). W h a t is the tionship for a silty clay loam soil and 3 0 % effective
fundamental difference between these rainfall p a t t e r n s saturation, estimate t h e effective rainfall p a t t e r n for
and the patterns of problem (3.4)? t h e rainstorm of p r o b l e m (3.3). Surface storage is 0.2
(3.9) C o m p u t e the defining p a r a m e t e r s for the rain­ in.
fall p a t t e r n s shown in Fig. 3.14 for a 3-hr, 25-year (3.16) U s e the rainfall p a t t e r n developed for p r o b ­
rainfall at Stillwater, Oklahoma. U s e V = 30 min. lem (3.4a) and the infiltration relation of (a) problem
C o m p a r e the results with Fig. 3.13 by replotting Fig. (3.12), (b) p r o b l e m (3.13), (c) p r o b l e m (3.14),
3.13 with the results of this problem superimposed. (d) p r o b l e m (3.15), to estimate the effective rainfall
(3.10) Assuming you are currently inside, go to the pattern.
nearest outside window. Describe what you see and (3.17) U s e t h e rainfall p a t t e r n developed for p r o b ­
estimate t h e amount of surface storage present as far lem (3.5a) and the infiltration relationship of (a) prob­
as runoff hydrographs are concerned. lem (3.12), (b) p r o b l e m (3.13), (c) problem (3.14), (d)
(3.11) A n infiltrometer applies simulated rainfall to problem (3.15), to estimate t h e effective rainfall pat­
a plot that is 0.001 acres in size. T h e rainfall rate is a tern.
constant 3 i n . / h r . T h e cumulative runoff volume as a Note: F o r the following problems (3.18-3.26) to be
function of time is shown below, (a) Estimate the of maximum benefit a n d interest, two or t h r e e actual
p a r a m e t e r s of the H o r t o n infiltration equation, (b) watersheds in the 5 to 1000 acre size range should be
Estimate the p a r a m e t e r s of the Holton infiltration selected. A watershed that has a significant undevel­
equation assuming the available water capacity is 0.35 o p e d area should b e used for p r o b l e m (3.26). D a t a on
and η is 1.4. This plot is estimated to have a surface topography, soils, and land use should be gathered. If
storage capacity of 0.2 in. and an unfilled soil water this is a class project, different students might use
storage capacity of 3.0 in. different a p p r o a c h e s to t h e problems and t h e n com­
p a r e t h e results.
Time Cumulative runoff (3.18) Develop t h e 25-year, 3-hr rainstorm for the
(min) (in.) catchment using t h e a p p r o p r i a t e time intervals. T h e
D D F , SCS, or Chicago hyetograph might b e used. (See
0 0.00
n o t e above.)
10 0.00
(3.19) Using an a p p r o p r i a t e loss function, develop
20 0.00 t h e effective rainfall p a t t e r n for t h e catchment based
30 0.22 on the rainfall of p r o b l e m (3.18). Loss functions that
40 0.56 might b e used include t h e SCS curve n u m b e r approach
50 0.94 or a surface storage plus infiltration with infiltration
60 1.35 based on the H o r t o n , G r e e n - A m p t , or Φ index ap­
70 1.78 proach. (See n o t e above.)
80 2.21
(3.20) U s e the Santa B a r b a r a U r b a n Hydrograph
m e t h o d to estimate t h e runoff hydrograph. (See n o t e
90 2.65
above.)
100 3.09
(3.21) Develop an a p p r o p r i a t e synthetic unit hydro-
110 3.54
graph for the catchment. O n e of t h e SCS unit hydro-
120 3.99
graphs, t h e H a a n unit hydrograph, or t h e double trian­
gle unit hydrograph can be used. R e q u i r e d p a r a m e t e r s
(3.12) Based on the H o r t o n infiltration relationship can be estimated based on SCS relationships, Epsey
with / o = 6 iph, f = 0.2 iph, and k = 3 h r , estimate
c
- 1
et al. relationships, or by o t h e r m e a n s . (See n o t e above.)
the effective rainfall p a t t e r n for the rain of problem (3.22) Combine t h e effective rainfall of problem
(3.3). Surface storage is 0.25 in. (3.18) with t h e unit hydrograph of p r o b l e m (3.21) to
(3.13) Based on the Holton infiltration relationship develop t h e runoff hydrograph. Plot t h e rainfall, effec­
with f = 0.2 iph, η - 1.4, a = 1, and F = 3.5 in.,
c p tive rainfall, unit hydrograph, a n d runoff hydrograph
estimate the effective rainfall p a t t e r n for the rain of on a single chart. (See n o t e above.)
Long-Term Water Balances 99
(3.23) Estimate the 25-year peak flow from t h e
catchment without developing a runoff hydrograph. (3.31) T h e average monthly t e m p e r a t u r e in degrees
C o m p a r e this estimate to the peak of the runoff hydro- F a r e n h e i t and rainfall in inches for two locations are
graph of problem (3.22). A r e the estimates different? shown below. Assuming n o runoff and no d e e p seep­
Why? (See note above.) age, how much waste water can be disposed of annu­
(3.24) Estimate t h e runoff volume that will result ally through evapotransporation from a 10-acre plot of
from a 25-year, 24-hr rainfall on this catchment. Com­ grass?
pare this volume to the volume u n d e r the runoff hydro-
graph of problem (3.22). A r e t h e estimates different?
Sioux Falls, S D Phoenix, AZ
Why? (See note above.)
(Latitude 42.5° N) (Latitude 32.5° N)
(3.25) Estimate the runoff volume that will occur
from a 100-year, 24-hr rainfall on this catchment. (See Month Temperature Rain Temperature Rain

note above.)
(3.26) Assume that a substantial part (specify exactly
Jan 15.2 0.62 49.7 0.73
how much) of the catchment that is currently undevel­
Feb 19.1 0.93 53.5 0.85
oped is going to be converted to a large shopping mall
and high-density residential use. W h a t will b e the March 30.1 1.54 59.0 0.66

magnitude of the 25-year runoff p e a k a n d volume from April 45.9 2.31 67.2 0.32
the total catchment after this development? Plot t h e May 58.3 3.38 75.0 0.13
pre- and postdevelopment runoff hydrographs on the June 68.1 4.35 83.6 0.09
same chart. (See note above.) July 74.3 2.84 89.8 0.77
(3.27) Plot the d a t a of Table 3.4 for t h e type II curve Aug 71.8 3.59 87.5 1.12
along with Eq. (3.6). Is t h e equation a reasonable Sep 61.8 2.61 82.8 0.73
approximation to the plotted d a t a ? In what region is Oct 50.3 1.25 70.7 0.46
the deviation the greatest? W h a t impact is this likely to
Nov 32.6 1.00 58.1 0.49
have on an estimated runoff hydrograph?
Dec 21.1 0.74 51.6 0.85
(3.28) W h a t is t h e 6-hr, 6-mile P M P for your loca­
2

tion? C o m p a r e this P M P estimate with the value of t h e


greatest observed rainfall for 6-hr as estimated from
Eq. (3.5). (3.32) D o p r o b l e m (3.31) for your current location.
(3.29) Examine the documentation a n d / o r coding (3.33) Develop a 90-min unit hydrograph from the
for an event-based c o m p u t e r program for estimating following 30-min unit hydrograph.
runoff hydrographs. C o m p a r e t h e techniques used in
the program with those discussed in this chapter. M a k e
Time Flow
reference to rainfall, abstractions, and hydrograph d e ­
(min) (cfs)
velopment.
(3.30) Assuming constant rainfall rates during each 0 0
time interval, calculate the uniform loss rate or Φ 30 75
index for the following storm, which p r o d u c e d 2.0 in. of 60 200
direct runoff.
90 250
120 200
Time Accumulated depth 150 125
(hr) of rain (in.)
180 68

0 0.0 210 40

1 0.6 240 25

2 1.0 270 15

3 1.2 300 8

4 1.8 330 5

5 2.9 360 0

6 3.6
7 4.5
(3.34) Develop a 15 min unit hydrograph from the
30 min unit hydrograph of problem (3.33).
100 3. Rainfall-Runoff Estimation in Storm Water Computations

(3.35) Define time of concentration. Accumulated time Accumulated rain


(3.36) List several m e t h o d s for estimating infiltration (min) (in.)
losses from rainfall.
(3.37) W h a t factors influence the time of concentra­ 0 0.00

tion? 10 0.10
(3.38) Discuss the impact of the following catchment 15 0.25
changes on the expected p e a k runoff rate, time t o 30 0.40
peak, runoff volume, and base flow, (a) Urbanization of 60 0.45
an agricultural catchment, (b) urbanization of a natural 70 0.68
or wildland catchment, (c) channel straightening and 90 0.92
other hydraulic improvements within t h e catchment,
105 1.43
(d) conversion of a wildland catchment t o an intensive
NO 1.79
agricultural catchment, (e) installation of an improved
125 2.09
storm drainage system in an existing u r b a n area, (f)
installation of small livestock and recreational ponds 140 2.47

on an agricultural catchment, and (g) installation of 180 2.55


flood detention basins on an u r b a n catchment. 210 2.65
(3.39) Describe how nonrecording rain gage infor­ 240 2.70
mation can be used to supplement recording rain gage
data in determining the time distribution of basin-wide (3.48) Infiltration d a t a reveal the following:
rainfall.
(3.40) Based on a 1200-acre watershed, d e t e r m i n e : Time (hr) /(mm/hr)
(a) the volume in acre-feet of 1.5 in. of runoff,
1
(b) the runoff rate in cfs equivalent to 1 i n . / h r , 2 3.52
(c) the volume in cubic feet equivalent to 0.2 in. of 1 2.16
evapotranspiration, 5 0.51
(d) t h e rate in gallons p e r minute equivalent to 0.2
in. of evapotranspiration per day. Estimate / , / , and k in H o r t o n ' s infiltration equa­
0 c

(3.41) If a steady and prolonged rain of 2 i n . / h r falls tion.


on a 1 m i l e reservoir, what is the equivalent steady
2

(3.49) Estimate t h e m e a n a n n u a l potential evapo-


outflow rate in cfs?
transporation at a location whose latitude is 40° Ν and
(3.42) Develop a two-way table with location along
whose m e a n monthly t e m p e r a t u r e in degrees centi­
the vertical left scale and duration along the u p p e r
grade is:
horizontal scale. Tabulate the 30-min, 1-, 2-, 3-, 6-, 12-,
and 24-hr rainfall for a 10-year frequency for New
Month: J F Μ A Μ J J A S O N D
York City; Miami, Florida; Rapid City, South D a k o t a ;
O k l a h o m a City, Oklahoma; and Seattle, Washington. Temp: 0 5 15 17 19 23 30 28 20 15 11 8

(3.43) Define the probable maximum precipitation.


(3.44) W h a t method of storm rainfall averaging (3.50) A 500-acre c a t c h m e n t experiences a uniform
would be best for determining catchment rainfall in a rainfall of 1.8 iph for 2 hr. W h a t is the maximum
mountainous area? Why? possible runoff r a t e in cfs? W h a t assumptions did you
(3.45) A 1000-hectare catchment has an average an­ m a k e in arriving at your estimate?
nual rainfall of 990 m m and an average annual stream (3.51) A c a t c h m e n t n e a r Atlanta, Georgia, has an
flow of 0.05 m / s e c . T h e estimated average annual
3
average overland flow distance of 350 ft. T h e length of
evapotransportation is 600 mm. T h e r e are n o diver­ the main channel is approximately 5000 ft for this
sions of flow into or out of the catchment. Estimate the 475-acre area. T h e p r e d o m i n a n t land use is wooded
average annual recharge to groundwater. with average land slopes of 10% a n d an average stream
(3.46) Develop a rainfall hyetograph in 30-min incre­ slope of 0 . 5 % . T h e soil is a silt loam and is r a t h e r
ments for a rainfall of 83 m m occurring in 5 hr. F o r the shallow. Estimate t h e 25-year p e a k flow from the
time distribution use (a) the SCS type II curve, (b) t h e catchment u n d e r this new condition. State all assump­
curves of Fig. 3.14, and (c) the Chicago hyetograph. tions that you m a k e .
(3.47) W h a t is the maximum 1-hr a n d 2-hr intensi­ (3.52) Long-range plans call for converting the
ties for the following rainfall? catchment of problem (3.51) to m e d i u m density u r b a n
Long-Term Water Balances 101

housing. This will be accompanied by storm drains and (3.63) A 400-acre watershed has t h e following char­
considerable channel work on the main drainage way. acteristics:
Estimate the 25-year p e a k flow from t h e catchment A. 100 acres of forest Muskingum soil
u n d e r this new condition. State all assumptions that B. 150 acres of forest Shelocta soil
you make. C. 50 acres of surface mined Muskingum soil
(3.53) T h e runoff hydrograph from a 1.7-mile 2
D . 50 acres of surface mined Shelocta soil
catchment can be approximated by a triangle with a E. 50 acres of bottomland P o p e soil.
peak of 600 cfs. If the volume of runoff is 1.3 in., what Calculate t h e a p p r o p r i a t e SCS curve number. Convert
is the time base of t h e hydrograph? to a n t e c e d e n t condition HI.
(3.54) T h e following d a t a o n precipitation and runoff
were collected over a wide range of a n t e c e d e n t condi­
tions from a catchment. Estimate t h e A M C II SCS References
curve number.
Ardis, C. V., Jr. (1973). "Storm Hydrographs Using a Double Trian­
gle Model." T e n n e s s e e Valley Authority, Division of Water Con­
P (in.) 3.4 1.7 2.8 1.9 2.2 4.7 1.5 trol Planning, Knoxville, T N .
Q (in.) 1.4 0.6 0.6 0.5 0.6 2.3 0.4 Bernard, M. (1942). In "Hydrology" ( Ο . E. Meinzer, ed.), Chap. II.
Dover, N e w York.
Betson, R. P. (1964). What is watershed runoff? / . Geogr. Res.
(3.55) C o m p a r e the Κ of Eq. (3.56) based on t h e 69(8):1541-1551.
SCS relationship [Eq. (3.57)] and the relationship of Bleek, J. (1975). Synthetic unit hydrograph procedures in urban
hydrology. In "Proceedings, National Symposium on Urban Hy­
Eq. (3.58) approximating the exponents 0.96 and 1.07
drology and Sediment Control, University of Kentucky, Lexing­
by 1.00. ton, Kentucky, July."
(3.56) Show that the Κ of Eq. (3.59) is about 3.77 Bloomsburg, C. L. (1960). A hydrograph equation, Paper presented
for the SCS unit hydrograph. at Pacific Northwest Section Meeting of American Geophysical
(3.57) C o m p a r e the Κ of Eq. (3.56) based on t h e Union, Moscow, Idaho, October 1 9 - 2 0 .
Borelli, J., Hasfurther, V. R., and Burman, R. D . ed. (1983). Ad­
SCS relationship [Eq. (3.57)] and the relationship of
vances in irrigation and drainage—Surviving external pressures.
Eq. (3.66).
In "Proceedings, American Society of Civil Engineers Specialty
(3.58) Develop a 3-hr, 25-year rainstorm for your Conference, N e w York."
location. U s e 15-min time increments. Brakensiek, D . L. (1967). A simulated watershed flow system for
(3.59) Estimate t h e 6-hr, 50-year rain for your cur­ hydrograph prediction. In "Proceedings, International Hydrology
rent location. Symposium, Fort Collins, C O . "
Brakensiek, D . L., and Rawls, W. J. (1983). Green-Ampt infiltration
(3.60) For the rainfall shown below, develop t h e
model parameters for hydrologic classification of soils. In "Pro­
effective rainfall p a t t e r n assuming a curve n u m b e r of ceedings, American Society of Civil Engineers Specialty Confer­
80. ence, N e w York" (J. Borelli, V. R. Hasfurther, and R. D.
Burman, eds.), pp. 2 2 6 - 2 3 3 .
Time Accumulated rain Time Accumulated rain Burman, R. D . , et al. (1983). Water requirements. In "Design and
(min) (in.) (min) (in.) Operation of Farm Irrigation Systems" (Jensen, ed.), Chap. 6,
Monograph 3. American Society of Agricultural Engineers, St.
0 0 105 2.48 Joseph, MI.
Chow, V. T., ed. (1964). "Handbook of Hydrology." McGraw-Hill,
15 0.06 120 2.63
N e w York.
30 0.13 135 2.74 Chow, V. T., Maidment, D . R., and Mays, L. W. (1988). "Applied
45 0.23 150 2.82 Hydrology." McGraw -Hill, N e w York.
Cronshey, R. (1981). Personal communication. U.S. Department of
60 0.41 165 2.89
Agriculture, Soil Conservation Service, Washington, D C .
75 0.96 180 2.96 DeCoursey, D . G. (1966). " A Runoff Hydrograph Equation." U.S.
90 2.26 Department of Agriculture A R S 41-116, Washington, D C .
Eagleson, P. S. (1970). "Dynamic Hydrology." McGraw-Hill, New
York.
(3.61) A watershed has a time to peak of 60 min and Edson, C. G. (1951). Parameters for relating unit hydrographs to
an area of 600 acres. Write down the unit hydrograph watershed characteristics. Am. Geophys. Union Trans. 32:591-596.
ordinates using a time increment of 15 min and the Engman, Ε. T. (1986). Roughness coefficients for routing surface
SCS dimensionless, curvilinear hydrograph. runoff. / . Irrigat. Drain. Eng. 1 1 2 ( 0 : 3 9 - 5 3 .
Epsey, W. H., Jr., Altman, D . G., and Graves, C. B. (1977). Nomo­
(3.62) C o m p u t e the runoff hydrograph resulting from
graph for ten-minute unit hydrographs for small urban water­
the rain of problem (3.60) and the unit hydrograph of sheds, Technical M e m o N o . 32. Urban Water Resources Re­
problem (3.61). search Program, American Society of Civil Engineers, New York.
102 3. Rainfall-Runoff Estimation in Storm Water Computations

Frederick, R. H., Myers, V. Α., and Auciello, E. P. (1977). Five to 60 Putnam, A . L. (1972). Rainfall and runoff in urban a r e a s — A case
minute precipitation frequency for the eastern and central United study of flooding in the Piedmont of North Carolina. In "Pro­
States, National Oceanic and Atmospheric Administration Tech­ ceedings, Urban Rainfall Management Problems, University of
nical Memorandum N W S H Y D R O - 3 5 . U.S. Department of Com­ Kentucky, Lexington, Kentucky, April 1 7 - 1 8 . "
merce, Washington, D C . Rawls, W. J., Brakensiek, D . L., and Miller, N. (1983). Green-Ampt
Gray, D . M. (1961). Synthetic unit hydrographs for small watersheds. infiltration parameters from soils data. / . Hydraulic Eng., Proc.
Am. Soc. Civil Eng. Proc. 87(HY4):33-54. Am. Soc. Civil Eng. 1 0 9 ( l ) : 6 2 - 7 0 .
Green, W. H., and Ampt, G. A. (1911). Studies on soil physics. I. Regan, R. M., and Duru, J. O. (1972). Kinematic wave nomograph
Flow of air and water through soils. / . Agron. Soc. 4 : 1 - 2 4 . for times of concentration. Proc. Am. Soc. Civil Eng.
Haan, C. T. (1970). A dimensionless hydrograph equation, File 98(HY10):1765-1771.
Report. Agricultural Engineering Department, University of Riedel, J. T., and Schreiner, L. C. (1980). Comparison of generalized
Kentucky, Lexington, KY. estimates of probable maximum precipitation with the greatest
Haan, C. T. (1972). Runoff coefficients for selected small streams in observed rainfall, N O A A Technical Report N W S 25. National
Kentucky, Bulletin 713. University of Kentucky, Agricultural Ex­ Oceanic and Atmospheric Administration, U.S. Department of
periment Station, Department of Agricultural Engineering, Lex­ Commerce, Silver Springs, M D .
ington, KY. Schwab, G. O., Frevert, R., Edminster, T. W., and Barnes, Κ. K.
Henderson, F. M. (1966). "Open Channel Flow." Macmillan, N e w (1971). "Elementary Soil and Water Conservation Engineering."
York. Wiley, N e w York.
Hershfield, D . M. (1961). Rainfall frequency atlas of the United Sherman, L. K. (1932). Streamflow from rainfall by the unit-graph
States, Technical paper 40. U.S. Department of Commerce, method. Eng. News Record 108:501-505.
Weather Bureau, Washington, D.C. Skaggs, R. W., and Khaleel, R. (1982). Infiltration. In "Hydrologic
Hewlett, J. D., and Hibbert, A. R. (1967). Factors affecting the Modeling of Small Watersheds" (Haan et al., eds.). American
response of small watersheds to precipitation in humid regions. Society of Agricultural Engineers, St. Joseph, ML
In "Forest Hydrology" (W. E. Sopper and H. W. Lull, eds.). Snyder, F. F. (1938). Synthetic unit-graphs. Trans. Am. Geophys.
Pergamon Press, Oxford. Union 19:447-454.
Hillel, D . (1971). "Soil and Water Physical Principles and Processes." Soil Conservation Service (1970). "Irrigation Water Requirements,"
Academic Press, New York. TR-21. U.S. Department of Agriculture, Soil Conservation Ser­
Holtan, Η. N. (1961). A concept for infiltration estimates in water­ vice, Washington, D C .
shed engineering. U.S. Department of Agriculture A R S 41-51. Soil Conservation Service (1972, 1985). "Hydrology," Sect. 4, Soil
Horner, W. W. (1910). Modern procedure in district sewer design. Conservation Service National Engineering Handbook. U.S. D e ­
Eng. News 64. partment of Agriculture, Washington, D C .
Horton, R. E. (1940). Approach toward a physical interpretation of Soil Conservation Service (1973). " A Method for Estimating Volume
infiltration capacity. Soil Sci. Soc. Am. Proc. 5:339-417. and Rate of Runoff in Small Watersheds," SCS-TP-149. U.S.
Huff, F. A. (1967). Time distribution of rainfall in heavy storms. Department of Agriculture, Soil Conservation Service, Washing­
Water Resources Res. 3(4): 1007-1019. ton, D C .
Izzard, C. F. (1946). Hydraulics of runoff from developed surfaces. Soil Conservation Service (1975, 1986). Urban hydrology for small
Proc. Highway Res. Board 26:129-150 watersheds, Technical release N o . 55. Soil Conservation Service,
Keifer, C. J., and Chu, Η. H. (1967). Synthetic storm pattern for U.S. Department of Agriculture, Washington, D C .
drainage design. / . Hydraulics Div., Am. Soc. Civil Eng. Soil Conservation Service (1983). Computer program for project
83(HY4):l-25. formulation hydrology (Draft), Technical release N o . 20. Soil
Kirpich, P. Z. (1940). Time of concentration of small agricultural Conservation Service, U.S. Department of Agriculture, Washing­
watersheds. Civil Eng. 10(6). ton, D C .
Linsley, R. K., Kohler, Μ. Α., and Paulhus, J. L. H. (1949). "Applied Strelkoff, T. (1969). One-dimensional equations of open channel
Hydrology." McGraw-Hill, New York. flow. Trans. Hydraulics Div. Am. Soc. Civil Eng.
Linsley, R. K., Kohler, Μ. Α., and Paulhus, J. L. H. (1982). "Hydrol­ 95(HY3):861-876.
ogy for Engineers," 3rd ed. McGraw-Hill, N e w York. Stubchaer, J. M. (1975). T h e Santa Barbara urban hydrology method.
McPherson, Μ. B. (1969). Some notes on the rational method of In "Proceedings, National Symposium on Urban Hydrology and
storm drain design, Technical memorandum N o . 6, A S C E Urban Sediment Control, University of Kentucky, Lexington, Kentucky,
Water Resources Research Program. American Society of Civil July 2 8 - 3 1 . "
Engineers, New York. Terstriep, M. L., and Stall, J. B. (1974). "The Illinois Urban Drainage
Nash, J. E. (1959). Systematic determination of unit hydrograph Area Simulator, I L L U D A S , " ISWS-74-Bul 58. State of Illinois,
parameters. / . Geophys. Res. 64:111-115. Department of Registration and Education, Urbana, IL.
National Research Council (1985). "Safety of D a m s — F l o o d and Tholin, A. L., and Kiefer, C. J. (1960). Hydrology of urban runoff.
Earthquake Criteria." National Academy Press, Washington, D C . Trans. Am. Soc. Civil Eng. 125:1308-1355.
National Oceanic and Atmospheric Administration. Precipitation- Thornthwaite, C. W. (1948). A n approach toward a rational classifi­
frequency atlas of the Western U.S. N O A A atlas II. Superinten­ cation of climate. Geogr. Rev. 3 8 : 5 5 - 6 4 .
dent of Documents, U.S. Government Printing Office, Washing­ Todd, D . K., ed. (1970). "The Water Encyclopedia." Water Informa­
ton, D C . tion Center, Water Research Building, Manhasset Isle, Port
Natural Environmental Research Council (1975). Flood studies re­ Washington, N Y .
port. Natural Environmental Research Council, London. U.S. Army Corps of Engineers (1959). Flood hydrograph analysis and
Norman, D . (1981). Personal communication. Bauer, Borowitz, Mer­ computation, Engineering and Design Manual E M 110-2-1405.
chant, Inc., 2607 1 / 2 N. High St., Columbus, OH. Washington, D C .
Overton, D . E., and Meadows, Μ. E. (1976). "Storm Water Model­ U.S. Department of Agriculture (1968). Moisture-tension data for
ing." Academic Press, New York. selected soils on experimental watersheds, Report 41-144. Agri-
Long-Term Water Balances 103

cultural Research Service, U.S. Government Printing Office, Agricultural Engineering Department, University of Kentucky,
Washington, D C . Lexington, KY.
U.S. Department of Commerce (1964). Two- to ten-day precipitation Wittenberg, H. (1975). A model to predict the effects of urbanization
for return periods of 2 to 100 years in the contiguous United on watershed response. In "Proceedings, National Symposium on
States, Technical paper 49. U.S. Government Printing Office, Urban Hydrology and Sediment Control, University of Kentucky,
Washington, D C . Lexington, Kentucky, July."
U.S. National Weather Service (1943-1984). "National Oceanic and Wooding, R. A . (1965a). A hydraulic model for the catchment-stream
Atmospheric Administration Hydrometeorologic Reports," 1-55. problem, I. / . Hydrol. 3:254-267.
A series of PMP reports for various areas not all of which are Wooding, R. A. (1965b). A hydraulic model for the catchment-stream
published, Washington, D C . problem, II. / . Hydrol. 3:268-282.
Wooding, R. A . (1966). A hydraulic model for the catchment-stream
Viessman, W., Jr. (1967). A linear model for synthesizing hydro-
problem. III. Comparison with runoff observations. / . Hydrol.
graphs for small drainage areas, Paper presented at the 48th
4:21-37.
Annual Meeting American Geophysical Union, Washington, D C .
Woolhiser, D . A . (1969). Overland flow on a converging surface.
Viessman, W., Jr., Knapp, J. W., Lewis, G. L., and Harbaugh, Τ. E.
Trans. Am. Soc. Agricultural Eng. 12(4):460-462.
(1977). "Introduction to Hydrology." Intext Educational Publish­
Woolhiser, D . Α., and Ligget, J. A. (1967). Unsteady one dimen­
ers, N e w York.
sional flow over a p l a n e — T h e rising hydrograph. Water Resources
Water Pollution Control Federation (1969). Design and construction Res. 3(3):753-771.
of sanitary and storm sewers, Manual of practice 9 (Am. Soc. of Wright-McLaughlin Engineers (1969). "Urban Storm Drainage Cri­
Civil Engineers Manual of Engineering Practice N o . 37). Water teria Manual." Denver Regional Council of Governments, D e n ­
Pollution Control Federation, 3900 Wisconsin Avenue, Washing­ ver, Colorado, 1969 and 1975 revision; also a similar manual for
ton, D C . Stillwater, OK, 1979.
Wilkinson, L. (1987). "SYSTAT: The System for Statistics." Wu, I-P. (1963). Design hydrographs for small watersheds in Indiana.
SYSTAT, Inc., Evanston, IL. Am. Soc. Civil Eng. Proc. 89(HY6):35-66.
Wilson, Β. N., et al. (1983). A hydrology and sedimentology water­ Yevjevich, V. (1968). Misconceptions in hydrology and their conse­
shed model, Special publication S E D I M O T II design manual. quences. Water Resources Res. tift'.llS-iyi.
4
Open Channel Hydraulics

T h e subject of o p e n channel hydraulics is extensive val, and change in storage represents the change in the
enough to require a complete text. Obviously, an ex­ volume of water stored within the section from 1 to 2.
haustive coverage cannot b e given in o n e short chapter. T h e continuity equation may also be written in terms
T h e t r e a t m e n t given here is intended only to cover of flow rates as
certain basic principles and to give the details neces­
inflow rate - outflow rate = rate of change in storage,
sary to design stable, o p e n channels; to d o simple
(4.2)
channel routings; and to compute simple backwater
profiles. T h e interested r e a d e r can consult several ex­ w h e r e inflow rate and outflow rate represent the rate
cellent texts for additional details (Chow, 1959; H e n ­ of flow across sections 1 and 2, respectively, and the
derson, 1966). r a t e of change in storage is the rate at which the
volume of water is accumulating or diminishing within
t h e section.
BASIC RELATIONSHIPS T h e flow rate, Q, is generally expressed in cubic feet
p e r second (cfs) or cubic m e t e r s p e r second (cms) and
Continuity Equation may be written
W h e n dealing with the hydraulics of o p e n channel β «i*4, (4.3)
flow, t h e r e are three basic relationships that must b e
w h e r e ν is the average velocity of flow at a cross-sec­
kept in mind. These relationships are the continuity
tion and A is the area of the cross section, ν is
equation, the energy equation, and the m o m e n t u m
generally given in feet p e r second (fps) or m e t e r s per
equation. If we consider a stream with a cross section
second ( m / s e c ) and A in square feet ( f t ) or square
2

as shown in Fig. 4.1, the continuity equation may be


m e t e r s ( m ) . T h r o u g h o u t this chapter, units on symbols
2

written as
appearing in equations will not be given unless n e e d e d
inflow - outflow = change in storage, (4.1) for clarity. Standard units are feet and seconds or
m e t e r s and seconds.
where inflow represents the volume of flow across It should b e kept in mind that ν is the average
section 1 during a time interval, outflow represents the velocity of the flow perpendicular to the cross section.
volume of flow across section 2 during this time inter- T h e actual p a t t e r n of flow velocity can be quite com-

104
Basic Relationships 105

/ / / / / / / / / /
Figure 4.3 Typical velocity profile.

Theoretically, with a solid boundary, t h e flow velocity


at t h e b o u n d a r y is z e r o . Actually for natural channels it
is difficult to d e t e r m i n e precisely w h e r e t h e channel
Section AA
b o u n d a r y is. T h e i m p o r t a n t point is that particles along
Figure 4.1 Typical channel sections. t h e channel b o u n d a r y are subjected to an actual veloc­
ity that is considerably lower t h a n t h e average flow
velocity of t h e cross-section.
plex. Figure 4.2 shows typical distributions of flow
velocity with various channel cross sections. Figure 4.3 Energy
shows a velocity profile for an idealized situation (Chow,
In basic fluid mechanics, t h e energy equation is
1959).
generally written in t h e form of Eq. (4.4). This relation­
Both Figs. 4.2 and 4.3 show that the actual velocity
ship is known as Bernoulli's equation or Bernoulli's
in contact with the channel boundary is quite low.
theorem:

Pi ,
+ ^2 + Z
2 "· ·" " L , l -
2g 2g
(4.4)

Trapezoidal channel
Triangular channel
T h e terms in this equation are shown in Fig. 4.4. T h e
Bernoulli e q u a t i o n r e p r e s e n t s an energy balance b e ­
tween two points along t h e channel. Again, ν is the
average flow velocity, g is t h e gravitational constant, y
is t h e d e p t h of flow, ζ is the elevation of t h e channel
bottom, ρ is a pressure, y is t h e unit weight of water,
a n d h _ r e p r e s e n t s t h e energy loss between sections
L { 2

1 and 2.
Shallow ditch E a c h complete t e r m of Eq. (4.4) has the units of a
length. Since t h e equation is an energy equation, one
should consider that t h e t e r m s represent energy per
unit of flowing fluid. Since t h e units are a length, t h e
t e r m s are commonly associated with a " h e a d " because

Natural irregular channel

Narrow
rectangular
section

Figure 4.2 Typical velocity distributions. Figure 4.4 Terms in Bernoulli equation.
106 4. Open Channel Hydraulics
of the engineer's familiarity with pressure and pressure
heads.
Thus, v /2g
2
is termed the velocity head, y + ζ is
termed the elevation head, and p/y is known as t h e
pressure head. Since the terms represent energy p e r
unit of fluid, we can in a loose sense think of v /2g 2
as
representing kinetic energy, y + ζ as representing po­
tential energy, and p/y as representing stored energy.
T h e sum of t h e velocity head, elevation head, and
pressure h e a d represents the total energy. T h e line
labeled E G L in Fig. 4.4 represents this sum and is
known as t h e energy grade line. T h e sum of the eleva­ Figure 4.5 Specific energy diagrams.
tion head and pressure head is known as t h e hydraulic
grade line ( H G L ) . T h e factor that distinguishes o p e n
channel flow from pipe flow is that in o p e n channel
flow, the free water surface is exposed to t h e atmo­ differential to zero:
sphere so that p/y is zero. Thus, t h e pressure h e a d
dE/dy = -2q /2gy 2 3
+ 1 = 0
term can generally be ignored for o p e n channel prob­
lems, and hence, the H G L coincides with the water or
surface. A rather obvious fact is that the E G L must b e
sloping downward in the direction of flow. T h e E G L yc-W/g- (4.9)
can only go u p if external energy (through a p u m p for
example) is supplied to t h e flow. Since q = uy , we can write Eq. (4.9) as
c

If we consider a channel section in which t h e r e is n o


energy loss, we can write "/V^T = I- (4.10)

v /2g
2
+ y + ζ = constant. (4.5) T h e t e r m v/yfgy^. is known as t h e F r o u d e n u m b e r F.
E q u a t i o n (4.10) shows that when y = y or when the c

If we take the datum elevation to b e the channel flow is at the critical d e p t h , the F r o u d e n u m b e r is o n e .
bottom, we have T h e F r o u d e n u m b e r can be used to classify the flow
into subcritical, critical, and supercritical flow. W h e n
u /2g
2
+ y = constant = E, (4.6)
F < 1, the flow is subcritical and y > y . This corre­ c

where the constant Ε is known as the specific energy. sponds to zone 1 in Fig. 4.5. W h e n F > 1, the flow is
Consider now a wide rectangular channel so that the supercritical and y < y . Supercritical flow is zone 2 in
c

depth all across the channel cross section is y. W e can Fig. 4.5. F = 1 is known as critical flow and corre­
then relate the flow rate on a per unit of width basis sponds to the line y = y = 2E/3 in Fig. 4.5. Equation
c

and the average flow velocity by (4.10) shows that for critical flow, v = y[gy~ . This c c

velocity corresponds to t h e celerity of small gravity


q = uy, (4.7) waves in shallow water.
F o r nonrectangular channels, the F r o u d e n u m b e r is
where q is the flow rate p e r unit of width. defined as
Equation (4.6) can now b e written as
F = U/^, (4.11)
Q /2gy
2 2
+ y = Ε (4.8)
w h e r e d is the hydraulic d e p t h . T h e hydraulic d e p t h is
and a plot of y versus Ε constructed for a constant q. h

defined as t h e area divided by the top width


Figure 4.5 is such a plot and is known as a specific
energy diagram. Some characteristics of the specific d =A/t.
h (4.12)
energy are that for a given Ε there are two possible
d e p t h s of flow, y and y , known as alternate depths,
x 2 Since F is i n d e p e n d e n t of slope, y d e p e n d s only on
c

and t h e r e is a definite minimum Ε for a given q. T h e t h e discharge for a given channel. For a rectangular
d e p t h of flow corresponding to the minimum Ε is channel, this is a p p a r e n t from Eq. (4.9). In general, the
known as the critical d e p t h and is d e n o t e d by y . T h e
c relationship between Q and y can b e d e t e r m i n e d
c

relationship between the flow rate and y can b e c from Eq. (4.11) for any channel by setting F = 1 and
determined by differentiating Eq. (4.8) and setting the noting from Eq. (4.3) that υ = Q/A.
Basic Relationships 107

Example Problem 4.1 Critical depth


A triangular channel with side slopes of 3 :1 is carrying 20
cfs. What is the critical depth for this channel and flow rate?
Solution: Critical depth occurs when F = 1. Equations
(4.11) and (4.12) must be used. Note that a triangular channel Wcos0 W 9

is a special case of a trapezoidal channel with b = 0. The Figure 4.6 Sketch for momentum relationship.
area and top width are given by (see Fig. 4.9)

A = zd = 3d 2 2

t = 2dz = 6d. p e r unit width consideration


Therefore A(mu ) s = pq(u 2 - i^).
d =A/t
h = 3d /6d 2
= 0.5d. T h e forces in t h e s-direction are
S F = P + Wsind - P -R ,
From Eq. (4.11), S x 2 f

where P and P a r e pressure forces p e r unit width


ν QM given by
x 2

yfgd^ y/0.5gd Ρ = yy 7 2 ,
20/3d 2
1.66 R is a frictional resistance, a n d W s i n Θ is t h e s-direc-
{

tion c o m p o n e n t of t h e weight. Combining terms, we


x
^W\d^lF 2

have
d = 1.23 ft.
c

2 2

^ - ψ + Wsine - R = pq(v - v ). (4.14)


As shown in subsequent sections of this chapter, { 2 x

channel roughness, velocity, discharge, a n d slope a r e If a short section is considered so that R is negligi­ {

interrelated. For a given discharge and roughness, the ble a n d t h e channel slope is small so that sin Θ is n e a r
velocity can be increased a n d consequently, t h e d e p t h zero, Eq. (4.14) can be written as
of flow decreased by increasing the channel slope. yy\ yy\
W h e n the channel slope is such that t h e flow d e p t h
resulting in uniform flow equals critical depth, the
or
slope is called the critical slope, S . T h u s for subcritical c

flow, the slope is less than 5 and for supercritical flow,


C
^ + 1±-Z1 + 1J.=m (4.15)
the slope is greater than S . It should b e pointed out
c 2 g 2 g v
'
that critical depth, slope, a n d velocity for a given
w h e r e Μ is t h e specific force plus m o m e n t u m and is a
section change with the discharge.
constant. Again it is possible to plot y versus Μ for a
In designing channels for controlling a n d conveying
constant q in t h e form of a specific force plus momen­
runoff, it is generally desirable to design so that t h e
t u m curve. Figure 4.7 is such a plot again showing two
flow is subcritical. Supercritical flow presents special
possible d e p t h s for a given Μ a n d a definite minimum
problems that are not treated here.

Momentum
T h e m o m e n t u m principle in o p e n channel flow can
be visualized by considering Fig. 4.6 and the basic
relationship from mechanics

XF = A(mz; ),
s s (4.13) ι V Z0NE 2

which states that the sum of the forces in t h e s-direc- 1


tion equals the change in m o m e n t u m in that direction. 1
In Eq. (4.13), F represents forces in the .s-direction
s
M c Μ
and m represents the mass. For a constant mass and a Figure 4.7 Typical specific force plus momentum curve.
108 4. Open Channel Hydraulics
M. T h e two possible depths for a given Μ are known
as sequent depths. It can be shown that y correspond­
ing to the minimum Μ is y . Again zone 1 represents
c

subcritical flow and zone 2 supercritical flow.


0.750 yds 0.750 yds
0.970 Yds
UNIFORM FLOW

O p e n channel flow is generally classified with respect


Trapezoids,
to changes in flow properties with time and with loca­
2 to 1 sideslopes
tion along the channel. If the flow characteristics at a Trapezoids,
point are unchanging with time, the flow is said to b e 2 to 1 and
steady flow; otherwise the flow is unsteady. Similarly, if 1.5 to 1 sideslopes
Trapezoids,
the flow properties are the same at every location 1.5 to 1 sideslopes Rectangles
along the channel, the flow is uniform. Flow with
Trapezoids,
properties that change with channel location is nonuni­
1 to 1 sideslopes
form flow. In natural flow situations, the flow is gener­
Rectangles
ally nonsteady and nonuniform. However, in t h e design
(a) on sides (b) on bottom
of most channels, steady, uniform flow is assumed with JL_I I I I I I I L J I L
the channel design being based on some peak or maxi­ 1 2 3 4 5 6 7 8 9 1 0 1 2 3 4 5 6 7 8 9 10
m u m discharge. Value of the ratio ~
W h e n we speak of uniform flow, steady, uniform
Figure 4.8 Tractive force distribution for trapezoidal channels
flow is generally what is considered. For uniform flow,
(Lane and Carlson, 1953).
y and y and u and v in Fig. 4.6 a r e equal. Thus,
l 2 { 2

Eq. (4.14) reduces to


which u p o n r e a r r a n g e m e n t is
R = Wsin0
{ (4.16)
r-y{A/P)S.
or t h e frictional forces are just equal to t h e down­ T h e t e r m A/P represents the hydraulic radius, R,
stream component of the weight. T h a t is, t h e frictional defined as t h e flow area divided by the wetted perime­
resistance and gravitational forces a r e in equilibrium. ter. T h u s , we have
T h e frictional resistance to flow may be expressed as τ = yRS. (4.20)
a shear, r , p e r unit area times t h e resisting area.
In this equation, τ represents t h e average shear a r o u n d
Neglecting the resistance generated at t h e surface of
t h e periphery of t h e flow. A t some points the actual
the flow between the water and air, t h e resisting area
shear will exceed τ and at o t h e r points it will be less
over which τ operates is the length, L, of a section
t h a n r . L a n e a n d Carlson (1953) found t h e shear on
times the wetted perimeter, P, of t h e channel. T h e
t h e periphery of a trapezoidal channel varied as shown
wetted perimeter is simply the length of t h e boundary
in Fig. 4.8. T h e maximum shear is n e a r ydS rather
between the water and the channel sides and b o t t o m at
t h a n yRS. In designing channels for stability using a
any cross section or the distance a r o u n d the flow cross
critical tractive force approach as shown later, t h e
section starting at o n e edge of the channel and travel­
maximum shear can b e calculated as ydS.
ing along t h e sides and bottom of t h e channel to t h e
Experimental studies on water flow in pipes has
other channel edge.
shown that τ is proportional to t h e D a r c y - W e i s b a c h
T h u s R in Eq. (4.16) can be written as
{
friction factor, / , and t h e square of t h e flow velocity.
R = rPL. (4.17) T h a t is
f

τ=/ρυ /8 2
(4.21)
T h e weight of water in a section of t h e channel is or combining Eqs. (4.20) and (4.21),
simply
υ = y/8y/fρ yfRS.
W = ALy. (4.18)
By letting ^Sy/fp = C, Chezy's equation for open
For small angles 0, sin θ is about equal to t h e slope channel flow is obtained as
of the channel in feet p e r foot. Thus, Eq. (4.16) may b e v = Cy/RS = CR'/W ,
2
(4.22)
written as w h e r e C is a factor related to t h e roughness of the
rPL = ALyS, (4.19) channel.
Uniform Flow 109
A n Irish engineer n a m e d Manning found that t h e w h e r e ν is in fps, R is in feet, and S is in feet p e r foot.
equation Tables of Manning's η a r e widely available. Table 4.1
υ = KR S 2/3 l/2 is such a table taken from several sources, drawing
fit experimental data quite nicely. This equation is heavily on Schwab et al. (1966, 1971). Manning's η is
influenced by many factors, including the physical
known as Manning's equation and differs from Chezy's
roughness of t h e channel surface, the irregularity of
equation only in the exponent on R. So that t h e factor
t h e channel cross section, channel alignment and bends,
related to the channel roughness would increase as
vegetation, silting a n d scouring, and obstruction within
roughness increased, Manning's equation is generally
t h e channel. Chow (1959) displays some photographs
written as
of typical channels and the associated values for
ν = (l/n)R S 2/3 l/2
Manning's n.
in the metric system with υ in meters per second and Figure 4.9 contains some useful relationships for
R in meters. T h e coefficient η is known as Manning's calculating the hydraulic properties of Α, Ρ, Λ, and
n. In the English system of units, Manning's equation is t o p width, Γ , for t h r e e c o m m o n channels. For natural
channels, these properties are best d e t e r m i n e d from
1.49
m e a s u r e m e n t s based o n t h e actual cross sections of the
ν = i? / S
2 3 1 / 2
, (4.23)
channel.

Table 4.1 Typical Values for Manning's η

Type and description η Values 0


Type and description η Values*
of conduits Min. Design Max. of conduits Min. Design Max.

Channels, lined Natural Streams


Asphaltic concrete, machine placed 0.014 (a) Clean, straight bank, full stage,
Asphalt, exposed prefabricated 0.015 no rifts or deep pools 0.025 0.033

Concrete 0.012 0.015 0.018 (b) Same as (a) but some weeds and
stones 0.030 0.040
Concrete, rubble 0.016 0.029
(c) Winding, some pools and shoals,
Metal, smooth (flumes) 0.011 0.015
clean 0.035 0.050
Metal, corrugated 0.021 0.024 0.026
(d) Same as (c), lower stages, more
Plastic 0.012 0.014 ineffective slopes and sections 0.040 0.055
Shotcrete 0.016 0.017 (e) Same as (c), some weeds and
Wood, planed (flumes) 0.009 0.012 0.016 stones 0.033 0.045

Wood, unplaned (flumes) 0.011 0.013 0.015 (f) Same as (d), stony sections 0.045 0.060
(g) Sluggish river reaches, rather
Channels, earth weedy or with very deep pools 0.050 0.080
Earth bottom, rubble sides 0.028 0.032 0.035 (h) Very weedy reaches 0.075 0.150
Drainage ditches, large, no vegetation
Pipe
(a) < 2.5 hydraulic radius 0.040 0.045
Asbestos cement 0.009
(b) 2.5-4.0 hydraulic radius 0.035 0.040
Cast iron, coated 0.011 0.013 0.014
(c) 4.0-5.0 hydraulic radius 0.030 0.035
Cast iron, uncoated 0.012 0.015
(d) > 5.0 hydraulic radius 0.025 0.030
Clay or concrete drain tile ( 4 - 1 2 in.) 0.010 0.0108 0.020
Small drainage ditches 0.035 0.040 0.040
Concrete 0.010 0.014 0.017
Stony bed, weeds on bank 0.025 0.035 0.040
Metal, corrugated 0.021 0.025 0.0255
Straight and uniform 0.017 0.0225 0.025
Steel, riveted and spiral 0.013 0.016 0.017
Winding, sluggish 0.0225 0.025 0.030
Vitrified sewer pipe 0.010 0.014 0.017
Channels, vegetated Wood stave 0.010 0.013
(See subsequent discussion) Wrought iron, black 0.012 0.015
Wrought iron, galvanized 0.013 0.016 0.017

"Selected from numerous sources.


110 4. Open Channel Hydraulics

Noet: Freeboard = Dd- lor al selection


τ Cross-sectional
area a
Wetted
perimeter, ρ
Hydraulic
radius R - —
Ρ
Top width

bd + Zd 2
b + 2d Vz+ 2
b 1
+
b + Zd
2dVZ + 1 2
2

t - b • 2dZ
Τ - b • 2DZ

Trapezodial cross section Zd


t-2dZ
Zd 2\/Z •12

2dVz *1
2 2

or
—• approx.

Traingualr cross section


8d 2
1.5t + 4d
2 z

0.67d
Ο \"2
approx.

Parabocil cross section


Figure 4.9 Properties of typical channels.
T h e expression for the hydraulic radius for wide
shallow channels can be simplified from that shown in T h e angle θ is defined in Fig. 4.11 and m e a s u r e d in
Fig. 4.9. Consider the trapezoidal channel shown in radians. Example Problems 4.2, 4.3, and 4.4 illustrate
Fig. 4.10. If the trapezoid is approximated by a rectan­ the use of Eqs. (4.24) and (4.25) to solve o p e n channel
gle, o n e can write flow problems dealing with circular conduits.
A bd T h e maximum flow capacity of a circular conduit
actually occurs at a d e p t h equal to 0.938 D . Figure 4.12
shows how the hydraulic elements of a circular conduit
If b » d, then the 2 d in the d e n o m i n a t o r can be
change with d e p t h . T h e subscript 0 refers to a d e p t h
ignored leaving
R * bd/b = d.
For a parabolic channel, if t » d, t h e n 4d in the 2

denominator of the expression for R can be ignored


leaving CASE I 0 < Y < £

td
2
2

These approximations can serve as initial estimates


for d in trial and error solutions that often arise in
open channel hydraulics.
T h e hydraulic elements of a circular conduit of di­
ameter D can be calculated from
D 2

A = — ( 0 ~ sin 0) (4.24)

(4.25)

CASE m $<Y<D

Figure 4.10 Approximation of trapezoidal channel with rectangu­


lar channel.
Figure 4.11 Definition of 0,
Uniform Flow 111

0 0.1 02 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3
" i—I I I J ^ / T H . I —
Example Problem 4.3 Flow in circular pipe 2

A circular corrugated metal pipe that is 3 ft in diameter is


flowing 2 ft deep. What is the discharge if the slope of the
pipe is 4%?
Solution: Refer to Fig. 4.11. Since y > D/2,
I 1/2 -
[r -(y-r) Y
2 2

0 = 2tt + 2 t a n - 1

r - y

2 ll/2
[ l . 5 - ( 2 . 0 - 1.5) ]
2 2

6.28 + 2 t a n - 1
= 3.81.
Figure 4.12 Hydraulic properties of a circular conduit.
1.5 - 2

From Eqs. (4.24) and (4.25),


D 2

equal to D. T h e line labeled Q/Q assumes that η is 0


A = — ( 0 - sin0) = 5.00
constant with depth. Even though t h e maximum flow
D I sin 0 \
occurs at 0.938D, it is common to design circular R = —\\ -0.87
conduits to carry maximum flows when flowing full. 4\ 0 j
Wave action and irregularities m a k e it difficult to 1.49
R S A.
2/3 l/2

maintain flow at 0.938D.


From Table 4.1, η = 0.024,
1.49
Example Problem 4.2 Flow in circular pipe 1 ( 0 . 8 7 ) ( 0 . 0 4 ) ( 5 . 0 0 ) = 56.6 cfs.
2/3 1/2

0.024

A circular corrugated metal pipe (CMP) that is 3 ft in


diameter is flowing 1 ft deep. What is the discharge if the
Example Problem 4.4 Flow in circular pipe 3
slope of the pipe is 4%?
Solution: Refer to Fig. 4.11 with the pipe radius, r, equal A circular corrugated metal pipe that is 3 ft in diameter is
to D / 2 . Since y < D/2, carrying 30 cfs. How deep is the water flowing if the slope of
the pipe is 4%?
p - ( r - ) T / 2

Solution:
0 = 2 tan"
y
1

r - y 1.49
Q = R SA 2/3 l/2

1/2
[ l . 5 - (1.5 - 1.0) ]
2 2 η
= 2 tan" 1.49 T D / sin0V 2 / 3 D2

1.5 - 1.0 30 = 1 S —(6 l/2


- sin0).
η 4 θ
= 2 tan" \2.82&) = 2.46. After substituting D = 3, η = 0.024, and S = 0.04, this
equation can be rearranged as
From Eqs. (4.24) and (4.25),
/ sin 0 \ 2 / 3

D 2
9 2.604 = l l — I (0 - s i n 0 ) .
A = — (0 - sind) = 7(2.46 - sin 2.46) = 2.06
8 8
This relationship can be solved by trial by assuming values
D [ sin θ \ 3/ sin 2.46
= 0.56 for 0, comparing the right-hand side of the equation to the
* - . ( ' - — ) - Ϊ ( ' - 2.46 left-hand side and continuing until a match is achieved.
1.49
Q = R S A. 2/3 l/2
Trial θ Right-hand side
η
From Table 4.1, η = 0.024, 3.14 3.14
2.50 1.58
1.49 vl/2
β = ( 0 . 5 6 ) ( 0 . 0 4 ) ( 2 . 0 6 ) = 17.4 cfs.
2/3 l/z
2.90 2.51
0.024
2.94 2.61 OK
112 4. Open Channel Hydraulics

θ = 2.94 is a solution. Since θ < π, y must be less than r


and can be obtained from
80' 4' 100*
[r>-(r-y) ] 2 l/2
3'
θ - 2tan"
50*
r-y
Figure 4.13 Channel section for Example Problem 4.6.
1/2
θ\ [r -(r-y) ]
2 2

tan
r-y

Example Problem 4.6 Compound channel


(?)
2.94 \ [2.25 - ( 1 . 5 - y ) ] 2 For the channel shown in Fig. 4.13, estimate the total flow
tan' for a depth of 8 ft. The channel has a slope of 0.5%.
(1.5 -y)' Manning's η is 0.06 for the overbank area and 0.03 for the
main channel.
When this equation is solved for y, the result is y = 1.35 ft.
Solution: Use Eq. (4.26).
80 Χ 4 = 320, Α2 = 50 Χ 8 = 400,
Example Problem 4.5 Flow in circular pipe 4 100 Χ 5 = 500
Λ = 80 + 4 = 84, Ρ = 4 + 50 + 3 = 57,
Use Fig. 4.12 to solve Example Problems 4.2, 4.3, and 4.4.
2

^3 = 100 + 5 = 105
Solution:
(320/84) 320 2/3
(400/57) 400
2/3

1.49 Q = 1.49(0.005) 1/2


0Μ 0.03
6o=—Rl s^ A . / 3 2
0

η
(500/105) 500 2/3

R = D/4 and A = i r D / 4 ; therefore


0 0
2

0Ότ3
1.49 //33 \ 2 / 3

U
1/2
((0.04) = 72.4 cfs.
0.024 = 9010 cfs.

When y = 1, y / D = 0.33. From Fig. 4.12, Q/Q = 0.23. 0

Therefore Q = 0.23(72.4) = 16.7 cfs. When y = 2, ^ / D =


0.67. From Fig. 4.12, Q/Q = 0.78. Therefore β = 0
DESIGN OF OPEN CHANNELS
0.78(72.4) = 56.5 cfs. When Q = 30, β / β = 0.41. From 0

Fig. 4.12, y / D = 0.44. Therefore y = 0.44(3) = 1.32 ft.


Nonerodible Channels
T h e design of nonerodible o p e n channels can be
d o n e by using Manning's e q u a t i o n [Eq. (4.23)].
Natural channels often have a main channel section
Manning's η should be chosen carefully. A d e q u a t e
and an overbank section. Most flow occurs in the main
consideration should be given to adding a freeboard or
channel; however, during flood events overbank flows
extra d e p t h to the channel as a safety measure to
may occur. T h e usual procedure for calculating such
protect against underestimates of flow or roughness
flows is to break the channel into cross-sectional parts
and wave action. Generally a freeboard of around 2 0 %
and sum the flow calculated for the various parts. In
of the d e p t h or 0.3 to 0.5 ft, whichever is greater,
determining the hydraulic radius for the various parts,
should be a d d e d to t h e channel d e p t h . T h u s , t h e major
only that part of the wetted perimeter in contact with
consideration in t h e design of channels in nonerodible
an actual channel boundary is used. T h u s
material is to ensure a d e q u a t e capacity.

V = i S ' 1 2
(4.26)
ft: Example Problem 4.7 Flow rate concrete channel 1
and
Consider a concrete channel that is trapezoidal with 3 : 1
side slopes and a 6-ft bottom width. The channel is on a 0.5%
Q = Σ ΚΑ,.
i= l slope and is flowing at a depth of 5 ft. What is the flow rate?
Design of Open Channels 113
Solution: S = 0.005 and Table 4.1 gives η = 0.015. From Example Problem 4.7 gives A = 105 ft ; therefore
2

Fig. 4.9, A A 105


dh = = 2.92
b + 2dz 6 + 2 x 5 x 3
bd + zd 6 X 5 + 3 X 5
2 2

13.9
R = , = ; = 2.79 ft F = = 1.43.
b + 2d}/z + 1 6 + 2 x 5V9+T
2

(32.2 Χ 2.92) 1

1 49 1 49 Thus the flow is supercritical. The high flow velocity is an


ν = — R S 2/3 l/2
= ^j(2.79) 2 / 3
(0.005) 1 / 2
= 13.9 fps early indicator of the possibility of supercritical flow.

A =bd + zd 2
= 6 X 5 + 3 X 5 = 105 ft
2 2

Q = uA = 13.9 X 105 = 1459.5 cfs.


Erodible Channels
In designing channels to be constructed in erodible
Example Problem 4.8 Flow depth concrete materials t h e r e are two major considerations. T h e
channel 2 channel must have a d e q u a t e capacity to carry the flow
and it must have a d e q u a t e stability to resist t h e erosive
The channel of Example Problem 4.7 is carrying 75 cfs. action of the flowing water. Erodible channels may be
How deep is the water flowing? either vegetated or nonvegetated. Vegetation tends to
Solution: protect the channel from erosion, thus permitting
higher flow velocities. O n the o t h e r hand, vegetation
1.49 increases t h e roughness of t h e channel. T h e design of
Q = vA = R SA2/3 l/2

nonvegetated channels is considered next followed by


η
2/3 t h e design of vegetated channels. Flexible linings and
1.49 bd + zd 2

riprap linings a r e discussed in subsequent sections.


S ' (bd
l 2
+ zd )
2

η b + Idylz 2
+ 1

1.49 6d + 3 d
22 1 2 / 3 Nonvegetated Channels
75 = (0.005) (6d + 3 d ) 1/2 2

0.015 6 + 6.32Λ Two main design p r o c e d u r e s are used for ensuring


t h e stability of erodible channels. O n e p r o c e d u r e is
2/3
6d + 3d 2
based on a limiting velocity concept and the other on a
10.68 = (6rf 4- 3d ). 2

6 + 6.32J limiting tractive force (boundary shear) concept. Table


4.2 shows allowable velocities and tractive force values
This last relationship must be solved by trial for a d such for several kinds of channels. This table is taken from
that the right-hand side of the equation is equal to 10.68. L a n e (1955) b a s e d on t h e work by Fortier and Scobey
(1926). T h e values a r e for aged, stable channels. For
Trial** Right-hand side newly constructed channels, t h e values shown in Table
4.6 should be used.
1 7.30
W h e n using t h e limiting velocity concept, o n e simply
1.5 15.93 sizes the channel so that it has a d e q u a t e capacity and
1.2 10.32 so that t h e average velocity does not exceed the per­
1.22 10.65 OK missible velocity.
W h e n using the limiting tractive force concept, a
c h a n n e l with a d e q u a t e capacity and having an average
The channel is flowing 1.22 ft deep. shear stress given by Eq. (4.20) that is less than the
values tabulated in T a b l e 4.2 is sought. F o r channels in
noncohesive materials, t h e particles on t h e channel
sides may move d u e to the combined force exerted by
Example Problem 4.9 Froude number
the flowing w a t e r and t h e weight component of the
Calculate the Froude number of the flow in example particles down t h e side of t h e channel. Chow (1959)
problem 4.7. should be referred to for a t r e a t m e n t of tractive force
considerations in noncohesive materials. In cohesive
Solution:
materials, the cohesion generally is much greater than
the gravity c o m p o n e n t so that average shear based on
F = /2 ' Eq. (4.20) can be used in design.
(**h) !
114 4. Open Channel Hydraulics

A n alternative approach to designing stable, unlined Table 4.2 Limiting Velocities and Tractive Forces for Open
channels is to use regime relationships. T h e s e relation­ Channels (Straight after Aging) 0

ships define equilibrium conditions between flow and


Water transporting
the channel boundaries. C h a p t e r 10 discusses this ap­
For Clear Water colloidal silts
proach.
Tractive Tractive
Velocity force Velocity force
Material η (fps) (psf) (fps) (psf)
E x a m p l e Problem 4 . 1 0 Erodible channel design
Fine sand colloidal 0.020 1.50 0.027 2.50 0.075
Design a channel to carry 20 cfs down a 0.5% slope. The Sandy loam noncolloidal 0.020 1.75 0.037 2.50 0.075
channel material is to be an ordinary firm loan. The water Silt loam noncolloidal 0.020 2.00 0.048 3.00 0.110
will be transporting colloidal silts. The channel is to be
Alluvial silts noncol­
trapezoidal with 3 : 1 side slopes. Use (a) the limiting velocity loidal 0.020 2.00 0.048 3.50 0.150
approach and (b) the limiting tractive force approach.
Ordinary firm loam 0.020 2.50 0.075 3.50 0.150
Solution: Volcanic ash 0.020 2.50 0.075 3.50 0.150
(a) Limiting velocity approach. From Table 4.2, t» = 3.5 p
Stiff clay very colloidal 0.025 3.75 0.260 5.00 0.460
fps, η = 0.020,
Alluvial silts colloidal 0.025 3.75 0.260 5.00 0.460
1.49 2/3 1/2 Shales and hardpans 0.025 6.00 0.670 6.00 0.670
Λ 5

Fine gravel 0.020 2.50 0.075 5.00 0.320


3/2
3/2 3.5(0.020) Graded loam to cobbles
R = = 0.54 when noncolloidal 0.030 3.75 0.380 5.00 0.660
1.49S 1/2
1.49(0.005) 1/2

Graded silts to cobbles


Q 20.00 when collodial 0.030 4.00 0.430 5.50 0.800
A = — = = 5.71
3.5 Coarse gravel
noncolloidal 0.025 4.00 0.300 6.00 0.670
bd + zd 2
_ bd + 3d 2

Cobbles and shingles 0.035 5.00 0.910 5.50 1.100


R = 0.54 (a)
b + 2d}/z 2
+ 1 b
+ 6 3 2 d

"From Lane (1955).


A bd + zd 2
= bd + 3d 2
= 5.71. (b)
Substituting Eq. (b) into Eq. (a) yields
5.71 If d = 2.50, then from Eq. (c) we get
= 0.54
b + 6.32d
b = 10.58 - 6.32(2.50) = - 5 . 2 2 ,
or
b = 10.58 - 6.32Λ. (c) which is clearly not possible. If d = 0.69, we get

Substituting this into Eq. (b) yields b = 10.58 - 6.32(0.69) = 6.22.


(10.58 - 6.32d)d + 3d = 5.71
2

Therefore the channel dimensions must be


- 3.32Λ + 10.58J - 5.71 = 0.00.
2

fr = 6.22ft, d = 0.69ft, ζ = 3.0.


This is a quadratic equation of the form
Check:
ax + bx + c = 0,
2

bd + zd 2
6.22(0.69) + 3(0.69) z

which has as a solution R = = 0.54


b + 2d}f?~+7 " 6.22 + 2(0.69)ν Ϊ0 /

-b±}fb ~ 2
4ac
1 49 1 49
2a υ - — R S 2/3 1/2
= — (0.54) (0.005) 2/3 1/2
- 3.5.
Therefore
The velocity is OK.
-10.58 ± V l O . 5 8 - 4 ( - 3 . 3 2 ) ( - 5 . 7 1 )
2

d = A = bd + zd 2
= 6.22(0.69) + 3(0.69) = 5.72 2

2(-3.32)
-10.58 + 6.00 Q = uA = 3.50(5.72) = 20.00.
d = = 2.50; 0.69.
-6.64 The capacity is OK.
Design of Open Channels 115
Add 0.3 ft of freeboard to get the final design of 6 = 6.2 ft in a detention basin, as this flow is likely to be a
and d = 1.0 ft. sustained o n e . A c o m p o u n d channel with a small, lined
(b) Critical tractive force approach. From Table 4.2, T = C
channel in the center to carry base flows and a vege­
0.15, η = 0.020. Figure 4.8 shows that for shallow and wide tated portion to carry storm flows may be used in these
(b/d > 8) trapezoidal channels, the maximum bottom shear
situations.
is ydS. Therefore
Vegetated waterways are somewhat m o r e complex to
T = C ydS design and require m o r e care in their establishment
rc 0.15 t h a n nonvegetated waterways. They carry high flows at
high velocities and require a minimum of maintenance
d =
^5 =
62.4(0.005) =
°' 4 8

if properly constructed.
1.49
T h e additional design consideration for vegetated
η waterways is the variation in roughness (Manning's n)
bd + zd with the height of t h e vegetation and with the type of
__ 0.486 + 0.69
2

g —
vegetation. Typically a tall grass presents a great deal
b + 2d}/z 2
+ 1 b
+ 3
· 0 3

of flow resistance to shallow flow. A s the flow depth


A=bd + zd = 0.486 + 0.69 2
increases, the resistance may decrease. Often the grass
1.49 /0.486 + 0.69 \ . „ 0 6 6 7
will lay over in t h e direction of flow when the flow
2 0
= —: ——— (0.005) (0.486 + 0.69) 1/2
reaches sufficient d e p t h . With the grass in this condi­
0.02 \ b + 3.03 ) tion, t h e resistance is considerably reduced as com­
I 0.486 + 0.69 \ 0 6 6 7

p a r e d to the shallow flow situation.


3797 = — —— (0.486 + 0.69).
Experimental work has shown that Manning's η can
\ 6 + 3.03 ; ' v

b e related to the product of the flow velocity and the


Solving by trial and error, 6 is found to be 12.4 ft. The b/d
ratio is 12.4/0.48 = 26. Thus the assumption that the maxi­ hydraulic radius, vR. This experimental work has also
mum bottom shear is ydS is verified. If b/d had been less shown t h a t different grasses have different n-vR rela­
than 8, r would have been KydS, where Κ would be
c
tionships. As a m a t t e r of fact, the same grass may have
approximated from Fig. 4.8. a different n-vR relationship depending on the height
Upon verifying that a channel with a bottom width of 12.4 of t h e grass.
ft, a depth of 0.48 ft, and 3 : 1 side slopes will have an Grasses have b e e n divided into five retardance
allowable velocity and adequate capacity, a freeboard of 0.3 classes, designated by A , B, C, D , and E. Table 4.3 lists
ft is added giving a final design of 6 = 12.4 ft, d = 0.8 ft, and t h e r e t a r d a n c e class for a n u m b e r of grasses that are
ζ = 3. commonly used. If t h e grass will be mowed part of the
time and long part of the time, both conditions and
r e t a r d a n c e classes must be considered. If a particular
vegetation is n o t listed in T a b l e 4.3, a similar vegeta­
Vegetated Channels tion might be used as a guide in selecting the retar­
From the previous section it can b e seen that the dance. In comparing vegetation, density, stem diame­
allowable velocities and tractive forces for nonvege- ter, stiffness, a n d o t h e r physical characteristics should
tated, erodible channels are quite small, thus requiring be considered. Information in Table 4.4 may be used to
wide shallow channels. Regime theory relationships in estimate the vegetal r e t a r d a n c e if specific information
Chapter 10 also predict wide shallow channels for on the type of vegetation is not known.
these conditions. If the channel can be protected from T h e maximum permissible velocities shown in Table
erosion, the allowable velocities can b e increased, re­ 4.5 should b e used for established sod in good condi­
sulting in deeper and more narrow channels. A n inex­ tion. T h e soil erodibility factor discussed in Chapter 8
pensive and p e r m a n e n t form of protection is vegeta­ can b e used to classify soils as erosion resistant or
tion—specifically grasses. Vegetation protects the easily e r o d e d (see p p . 126). Flow at these maximum
channel material from the erosive action of the flow velocities may require channel maintenance opera­
and binds the channel material together. tions. If poor vegetation exists d u e to shade, climate,
Vegetated waterways generally can b e used to carry soils, or other factors, the design velocity should be
intermittent flows such as storm water runoff. They are about 5 0 % of t h e values of Table 4.5. D a t a in Table 4.6
not recommended for channels having sustained base may b e used t o select permissible velocities when spe­
flows as most vegetation cannot survive continual sub­ cific vegetation and erosion characteristics of soils are
mergence or continual saturation in the root zone. This not known.
means that vegetated waterways would not be used as Figure 4.14 shows the n-vR relationship for the five
the channel carrying the discharge from a pipe spillway r e t a r d a n c e classes. T h e design procedure is to select
116 4. Open Channel Hydraulics

Table 4.3 Vegetal Retardance Classes (Soil Conservation Service, 1969)

Retardance Cover Condition

Reed canary grass Excellent stand, tall (average 36 in.)


Yellow bluestem Ischaemum Excellent stand, tall (average 36 in.)

Smooth bromegrass Good stand, mowed (average 12 to 15 in.)


Bermuda grass Good stand, tall (average 12 in.)
Native grass mixture (little bluestem, blue grams, and other long and
short midwest grasses) Good stand, unmowed
Tall fescue Good stand, unmowed (average 18 in.)
Lespedeza sericea Good stand, not woody, tall (average 19 in.)
Grass-legume mixture — Timothy, smooth bromegrass, or orchard
grass Good stand, uncut (average 20 in.)
Reed canary grass Good stand, mowed (average 12 to 15 in.)
Tall fescue, with bird's foot trefoil or lodino Good stand, uncut (average 18 in.)
Blue grama Good stand, uncut (average 13 in.)

Bahia Good stand, uncut (6 to 8 in.)


Bermuda grass Good stand, mowed (average 6 in.)
Redtop Good stand, headed (15 to 20 in.)
Grass-legume mixture — summer (Orchard grass, redtop, Italian
ryegrass, and common lespedeza) Good stand, uncut (6 to 8 in.)
Centipedegrass Very dense cover (average 6 in.)
Kentucky bluegrass Good stand, headed (6 to 12 in.)

Bermuda grass Good stand, cut to 2.5 in. height


Red fescue Good stand, headed (12 to 18 in.)
Buffalograss Good stand, uncut (3 to 6 in.)
Grass-legume mixture — fall, spring (Orchard grass, i.edtop, Italian
ryegrass, and common lespedeza) Good stand, uncut (4 to 5 in.)
Lespedeza sericea After cutting to 2 in. height, very good stand before cutting

Bermuda grass Good stand, cut to 1.5 in. height

Bermuda grass Burned stubble

the vegetation, determine its retardance class a n d per­ ensures a stable channel with a d e q u a t e capacity re­
missible velocity, and then design the channel based on gardless of t h e condition of t h e vegetation.
the curves of Fig. 4.14. For situations w h e r e two retar­ T e m p l e et al (1987) have developed the following
dance classes are applicable (for example mowed and approximation for the n-vR curves of Fig. 4.14,
unmowed grass), the channel should first b e designed
for stability based on the lower retardance and then η = exp[ 7(0.01329 In(vR)2

additional d e p t h added to the channel to accommodate


-0.09543In(vR) + 0.2971) - 4 . 1 6 ] , (4.27)
the flow when the retardance increases. This p r o c e d u r e
Design of Open Channels 117

where the value of / is Solution: Select Bermuda grass. Bermuda grass is in retar­
dance Β if unmowed and retardance D if mowed. The
Retardance / permissible velocity is selected from Table 4.5 as 6 fps. First
design for the mowed condition
A 10.000
Β 7.643 A = Q/v = 2 5 / 6 = 4.17 ft . 2

C 5.601
D 4.436
Ε 2.876 Table 4.4 Guide to Selection of Vegetal Retardance 0

Stand Length of Retardance


This relationship can be used in c o m p u t e r programs to vegetation (in.) class
make hydraulic computations for vegetated waterways.
T h e relationships should not be used outside the range >30 A
of the curves shown in Fig. 4.14. 11-24 Β
T h e graphs of Fig. 4.15 are solutions to Manning's 6-10 C
equation using the curves in Fig. 4.14. They can b e 2-6 D
used as a design aid for solving Manning's equation for <2 Ε
all retardance classes.
>30 Β
11-24 C
6-10 D
Example Problem 4 . 1 1 Vegetated channel 1 2-6 D
<2 Ε
Design a channel to carry 25 cfs on a 4% slope. Use a
parabolic channel. The soil is easily eroded, and the grass a
S o i l Conservation Service (1979) engineering field manual.
may be mowed to 2.5 in. or it may be uncut.

Table 4.5 Permissible velocities for Vegetated Channels (Ree, 1949)

Permissible velocity (fps)

Erosion-resistant soils Easily eroded soils


(% slope) (% slope)

Cover 0-5 5-10 Over 10 0-5 5-10 Over 10

Bermuda grass

Buffalo grass
Kentucky bluegrass
Smooth brome
Blue grama
Tall fescue

Lespedeza sericea
Weeping lovegrass
Kudzu 3.5 NR* NR 2.5 NR NR
Alfalfa
Crabgrass

Grass mixture 5 4 NR 4 3 NR

Annuals for temporary


protection 3.5 NR NR 2.5 NR NR

''Not recommended.
118 4. Open Channel Hydraulics

Table 4.6 Permissible Velocities (fps) a

Channel vegetation
Soil Bare
texture channel Retardance Poor Fair Good

Sand, silt 1.5 Β 1.5 3 4


Sandy loam 1.5 C 1.5 2.5 3.5
Silty loam 1.5 D 1.5 2 3

Silty clay loam 2 Β 2.5 4 5


Sandy clay loam 2 C 2.5 3.5 4.5
2 D 2.5 3 4

Clay 2.5 Β 3 5 6
2.5 C 3 4.5 5.5
2.5 D 3 4 5

"Soil Conservation (1979) engineering field manual.

.4

-.1 .2 .3 .4 .5 .6 .8 1 2 3 4 5 6 8 10 20

V R , Product of velocity and hydraulic radius


Figure 4.14 n-vR for various retardance classes.
Design of Open Channels 119

HYDRAULIC RADIUS ( F T )

ι « ι I ι ι ι 1 1 ι il 11 HI ι I ι ι ι ι I ι 1 1 il n ulim linilliiiliml I l i I I I I I I I II I
.2 .3 .4 5 & 7 β Β 10 2.0

HYDRAULIC RADIUS (Μ)


Figure 4.15a Solution to Manning's equation retardance class A.

From Fig. 4.15d for retardance class D, R = 0.7 ft. This is too small. Increase d to 1.25 feet, then
/ = 3A/2d = 5.00 and R = 0.714.
td
2

R = 0.7 = Try d = 1.17 ft. Now / = 3A/2d = 5.35 and R = 0.70, which
1.5r + 4d
2 2

is OK.
(see Fig. 4.9)
The design for the short grass condition is
A = 4.17 = hd.
i = 5.35 ft, d= 1.17 ft, R = 0.7 ft.
For small parabolic channels, d » 1.5R. Using this approx­ Now we must add depth using the same basic shape to get
imation, adequate capacity when the grass is long. When grass is long
d = 1.05 ft the retardance class is B. Try D = 1.40 ft. New top width
3A 3(4.17)
t =
2d 2(1.05)
= 5.96 ft.
= 5
· 3
/ 1.40
5
1.40
( Π 7
j 1/2
= 5.85
Check: and
(5.96) (1.05)
2
td
2

R = = 0.646, R = = 0.81.
1.5(5.96) + 4(1.05)
2 2
\.5t 2
+ Ad'
ϋ 03S/W λ!ΙΟ0Π3Λ

CD 03S/W Α1Ι0ΟΊ3Λ
120
1 " t I LIMLI • • • 1 ι T i ι 1 I I I I
JLll__ L l l l l . l l l l l l l l l l l l MILL I III I Μ I l l l l l l l l l I l l Μ I I I I I 1 l l l l j i
ο

—.g
^ L<J CM
5 ο Q ο σ> op Ν φ «Ρ to CVJ

03S/W Α1Ι00Ί3Λ ^21


122 4. Open Channel Hydraulics

From Fig. 4.15b and retardance Β with R = 0.81 and S = Assume R = 0.67, vR = 4.02:
0.04, find ν = 2.9 fps, therefore:
η = exp[4.436(0.01329 ln(4.02) 2

A = 2td/3 = 5.46 ft 2

-0.09543 ln(4.02) + 0.2971) - 4.16] = 0.038


Q = vA = 2.9 X 5.46 = 15.8 cfs too small.
0.298
ν = 0.038 ( 0 . 6 7 ) = 6.00 OK.
2/3
Try D = 1.75 ft:

/1.75\ 1 / 2
_ From this point, the solution follows the procedure of Exam­
7 = 5.35 6.54 with /? = 0.98
[Til) ple Problem 4.11. Note the sensitivity of velocity to hydraulic
radius in these calculations.
ν = 4.5 ft
A = 7.63 ft 2

Q = uA = 35 cfs too big.


Flexible Liners
Try £> = 1.6 ft:
N o r m a n n (1975) presents a uniform p r o c e d u r e for
/ 1.6 \ " l
the design of o p e n channels using flexible liners. Liners
7 = 5 3 5 l _ j =6.26 with R = 0.91 considered are vegetation, temporary liners, and riprap.
T h e p r o c e d u r e for vegetated liners is based on the
υ = 3.9 fps
procedures p r e s e n t e d in the previous section of this
Λ = 6.68 ft 2
book. T h e results for temporary liners are based on
Q = 26 cfs OK. work of M c W h o r t e r et al (1968), and the riprap results
are largely based on A n d e r s o n et al. ( u n d a t e d ) and
Add 0.3 freeboard to get a final design of D = 1.9 ft and
A n d e r s o n (1973). Results are p r e s e n t e d in the form of
/ 1.9
1 9 \\ ^' 2 equations describing t h e maximum permissible d e p t h
Γ =5.35 - — = 6 . 8 ft. of flow for a stable design

(4.28)

w h e r e d is in feet a n d S is in feet per foot and a


Example Problem 4.12 Vegetated channel 2 velocity equation of t h e form

= aR S b c
(4.29)
Work Example Problem 4.11 based on Eq. (4.27). Assume
the grass is always mowed. Vegetated Channels
Solution: Table 4.7 contains values for m and η for Eq. (4.28)
1 49 1 49
0 298 for vegetated channels. Analysis of N o r m a n n ' s results
„ = _ * 2 / 3 S l / 2 == —-—R 1Ζ 2/3= 6. Λ ( 0 .04) 1 / 2 2/3
and the results p r e s e n t e d in the previous section of this
η η η book indicate that b e t t e r agreement is obtained if d max

Assume R, compute vR, compute n, compute v, and if of Eq. (4.28) is replaced by the hydraulic radius, R. For
υ Φ 6, repeat. For retardance D, / = 4.436. Assume R = 0.8, example, for a grass mixture maintained at 6 to 8 in. on
then uR = 6(0.8) - 4.8: an erosion-resistant soil, the maximum hydraulic radius
is given by
η = exp[4.436(0.013291n(4.8) 2

R = 0.125" 0 5 3
. (4.30)
-0.09543 ln(4.8) + 0.2971) - 4.16] - 0.036
F o r vegetation, t h e velocity is d e t e r m i n e d from Figs.
0.298
4.15a-4.15e.
ν - 0.036 ( 0 . 8 ) = 7.13 too high.
2 / 3

Assume R = 0.7, vR = 4.2:


Example Problem 4.13 Flexible liner
η = exp[4.436(0.013291n(4.2) 2

Work Example Problem 4.11 using the Normann proce­


-0.09543 ln(4.2) + 0.2971) - 4.16] = 0.038
dure.
0 298
Solution: For Bermuda grass in retardance D with an
υ = ^5^(°· ) 7 2 / 3 = 6 1 8
slightly too high.
erodible soil, values of m and η are determined a s m = 0.08
Design of Open Channels 123

Table 4.7 m and η for Vegetation

Vegetation Height (in.) Retardance Erosivity m η

Bermuda 12 Β Resistant 0.20 -0.60


good stand Erodible 0.21 -0.51
6 C Resistant 0.13 -0.62
Erodible 0.11 -0.59
2.5 D Resistant 0.10 -0.67
Erodible 0.08 -0.63
1.5 Ε Resistant 0.072 -0.65
Erodible 0.05 -0.67

Grass mix uncut Β Resistant 0.20 -0.51


good stand Erodible 0.18 -0.48
6-8 C Resistant 0.12 -0.53
Erodible 0.11 -0.50
4-5 D Resistant 0.084 -0.58
Erodible 0.063 -0.60

Lespedeza 11 C Resistant 0.12 -0.47


Erodible 0.080 -0.53
4.5 D Resistant 0.13 -0.42
Erodible 0.080 -0.47

and η = - 0 . 6 3 from Table 4.7. Thus T h e d a t a of M c W h o r t e r et al. (1968) as presented by


N o r m a n n (1975) w e r e analyzed, and values for m and
R = 0.08(0.04)"° 6 3
= 0.61ft. η w e r e d e t e r m i n e d . T h e results are p r e s e n t e d in Table
4.9. M c W h o r t e r et al (1968) also found that the
From Fig. 4.15d, ν is determined as 5 fps. The solution
follows the procedures of Example Problem 4.11 from this temporary linings acted m u c h like vegetation; hence
point. M a n n i n g ' s η was not constant. A n equation with a
form similar t o that of Manning's equation was used,
b u t t h e exponents of R and S were not necessarily §
a n d \ . T h e suggested form of t h e velocity equation is
Temporary Channel Linings also given in T a b l e 4.9.
W h e n vegetated linings a r e selected for a channel In o r d e r t o use t h e equations in Table 4.9 for design
and flow can not be diverted from t h e channel during purposes, o n e would typically design t h e channel for
the establishment of vegetation, some form of t e m p o ­ t h e p e r m a n e n t vegetation and t h e n select a temporary
rary lining should b e used to stabilize t h e channel channel lining that would be stable in t h e channel. A
during the period of vegetal establishment. This lining lower-return period storm might b e selected for t h e
should be constructed of materials that will deteriorate design of t h e t e m p o r a r y lining since the exposure time
as vegetation emerges and will not interfere with its is short during vegetal establishment and since dam­
growth. ages during this period can b e easily repaired. Proce­
A selected group of these linings is listed in Table d u r e s for making t h e calculations are given in Example
4.8 along with a brief description of the materials. D a t a Problem 4.14.
on the effectiveness of these channel linings were col­
lected by M c W h o r t e r et al. (1968) at Mississippi State
University. T h e results of these tests indicated that t h e
maximum allowable d e p t h varies inversely with slope, Example Problem 4.14 Temporary channel liner
with different relationships being given for erosion
resistant and easily eroded soils. T h e maximum allow­ Select a temporary lining for use in the channel designed
able depth, d , can be r e p r e s e n t e d as a power func­
max in Example Problem 4.11. Assume that a lower-return period
tion of S according to Eq. (4.28). storm is used for the flexible lining design since it needs to be
124 4. Open Channel Hydraulics

Table 4.8 Description of Temporary Liners (McWhorter et al, 1968)

Excelsior mat Erosionet 315


Excelsior mat is composed of 0.8 pound/yd of excelsior (dried,
2
Erosionet is a paper yarn approximately 0.05 in. in diameter, woven
shredded wood) covered with a fine paper net covering. The paper net, into a net with openings approximately \ in. χ \ in. The material has lit­
reinforced along the edges, has an opening size of approximately 1 x 2 tle erosion prevention capability in itself and is generally used to hold
in. The mat is held in place by steel pins or staples at the rate of five other lining material in place. Erosionet weighs about 0.20 lb/yd and is 2

staples per 6 linear ft of mat, with two staples along each side and one in pinned in the same manner as jute mesh as described later in this table.
the middle. At the start of each roll, four or five staples are spaced
approximately 1 ft apart. Where more than one mat is required, the mats
are butt-joined and securely stapled. Fiberglass roving
Fiberglass roving is delivered as a lightly bound ribbon of continuous
glass fibers. The material is applied to the channel bed using a special
Straw and erosionet venturi nozzle driven by an air compressor, which separates the fibers
This lining consists of straw applied at a rate of 3 tons per acre (1.25 and results in a web-like mat of glass fibers. The glass fibers are tacked
lb/yd ). The straw is covered with Erosionet 315 (See description fol­
2
with asphalt for adhesion to each other and to the soil. The single layer
lowing). This lining is pinned in the same manner as jute mesh, as of fiberglass roving consists of one layer of blown fiberglass fibers
described later in this table. applied at a minimum rate of 0.25 lb/yd tacked with asphalt emulsion or
2

asphalt cement at a minimum rate of 0.25 gal ./yd . The double layer
2

application consists of two alternating layers of fiberglass and asphalt,


\ in. Fiberglass mat each layer consisting of fiberglass roving at 0.25 lb/yd . 2

This lining is fine, loosely woven glass fiber mat similar to furnace air
filter material. It has a weight of 0.11 lb/yd . This material is not to be
2

confused with more dense fiberglass mats used to eliminate plant Jute mesh
growth. Steel pins or staples are placed at the rate of five staples per 6 Jute mesh is a mat lining woven of jute yarn that varies from ? to i in.
linear ft of mat, with two staples along each side and one in the middle. in diameter. The mat weighs approximately 0.80 lb/yd , with open­
2

At the start of each roll four or five staples are spaced approximately 1 ft ings about I in. χ J in. Steel pins or staples are used to hold the jute
apart. Where more than one mat is required, the mats are butt-joined and
mesh in place. The pins or staples should be spaced not more than 3 ft
securely stapled.
apart in three rows for each strip, with one row along each edge and one
row alternately spaced in the center. At the overlapping edges of parallel
\ in. Fiberglass mat strips, staples should be spaced at 2 ft or less. At all anchor slots, junc­
This lining is a fine, loosely woven glass fiber mat, similar to but tion slots, and check slots, spacing should be 6 in. or less.
denser than the f in. fiberglass mat, as it weighs 0.35 lb/yd . The sta­
2

pling procedure is the same as for the f in. fiberglass mat.

Table 4.9 Coefficients for Eqs. (4.28) and (4.29)'

Erodible soil Erosion-resistant soil

Type lining m η m η Velocity equation

Bare soil 0.0030 -0.687 0.0084 -0.687 Κ=22.81/?°· 5 9 1


5 · 0 2 8 6

Fiberglass roving with asphalt tack (single layer) 0.0067 -0.960 0.0141 -0.960 V = 42.45 /?0667 o.5 S

Fiberglass roving with asphalt tack (double layer) 0.0143 -1.01 0.027 -1.01 V = 59.20 /?o.6675o.5

Jute mesh 0.0076 -0.875 0.0202 -0.883 Κ= 61.53Λ 1 0 2 8 1


5°· 4 3 1

Excelsior mat 0.0572 -0.585 0.101 -0.585 V= 32.29 fll-340 0.351 5

Straw and erosionet 0.052 -0.652 0.082 -0.652 V= 70.76 /?14550.529 S

Fiberglass mat \ in. 0.025 -0.670 0.046 -0.670 V= 73.53 /?1·3300.512 5

Fiberglass mat \ in. 0.048 -0.646 0.083 -0.646 V= 14.84 fli.2335O.O86

Erosionet 0.049 -0.642 0.084 -0.642 V = 41.45 /?θ·855ο.4θ 5

a
Adapted from McWhorter et al. (1968).
Design of Open Channels 125

Table 4.10 Initial Calculations for Example Problem 4.14

Maximum Hydraulic Top


depth 0
radius* width* Area* Velocity' Flow
(ft) (ft) (ft) (ft ) 2
(fps) (cfs)

Jute mesh 0.127 0.084 1.75 0.149 1.20 0.18


Excelsior 0.376 0.241 3.02 0.758 1.55 1.17
Straw and erosionet 0.427 0.270 3.21 0.907 1.92 1.74
Fiberglass (two layers) 0.369 0.237 3.00 0.737 4.53 3.34

T r a m Eq. (4.28) and coefficients in Table 4.9.


''From equation in Fig. 4.9.
Trom velocity equations in Table 4.9.

effective for only a short period of time. The design flow is channel, when grass lined only, had to have a Γ of 6.8 ft if
found to be 10 cfs. the D was 1.9 to safely carry 25 cfs. It thus appears that the
channel need not be constructed 1.9 ft deep since the top
Solution: From Problem 4.11, the soil is easily eroded. The
width exceeds the required top width.
slope is 4%, and a parabolic channel is used with 6.8 ft top
Holding the basic channel shape the same, the actual
width and a depth of 1.9 ft. To facilitate selection of the
depth of flow under the long grass condition when carrying
lining, the values shown in Table 4.10 were calculated from
Eq. (4.28) and Table 4.9. 25 cfs can be recalculated. Using retardance class B, a trial
and error procedure can be used to arrive at the flow depth.
Obviously, none of the linings are acceptable since the
Try d = 1.00 ft:
discharge at d is less than the design discharge. The
max

channel will have to be redesigned for stability during the


period of temporary lining. This will require an increase in
the top width without increasing the total depth, thus main­
taining stability. The design is made using the trial and error
i=r
U) = w) = · 2( 14 51

procedure shown in Table 4.11 td


2

What has been shown thus far is that a channel having a R = = 0.658
1.5r + 4d
2 2

shape defined by a parabola with Τ = 20 and D = 1.9 and


lined with a double layer of fiberglass will be stable enough υ = 1.2 fps (Fig. 4.15b)
to carry 10 cfs at a depth of 0.37 ft. A quick calculation shows
that the channel if unlined will be unstable if constructed in A = 2td/3 = 9.67
most soils. We have seen in Example Problem 4.11 that the Q = uA = 11.60 too small.

Table 4.11 Final Calculations for Example Problem 4.14

Top
Hydraulic width
Maximum radius at at Area at Velocity Discharge
Lining depth at J
^max ^max ^max ^max
type (ft) (ft) (ft) (ft ) 2 max (cfs)
(fps)
Top width*, 12ftfiberglass, two layers 0.369 0.243 5.28 1.30 4.61 5.99
too low
Top width*, 15 ft fiberglass, two layers 0.369 0.244 6.61 1.62 4.62 7.49
too low
Top width*, 18 ft fiberglass, two layers 0.369 0.245 7.93 1.95 4.63 9.04
too low
Top width*, 20ftfiberglass, two layers 0.369 0.245 8.81 2.16 4.63 10.00
OK

a
T at a depth of 1.9 ft.
126 4. Open Channel Hydraulics

Try d = 1.3, then t = 16.54, R = 0.853, υ = 3.0, A = 14.33, draglines are satisfactory. Some h a n d work is usually
and Q — 43. The channel is too large. required to ensure a stable and uniform riprap surface.
Try d = 1.15, then r = 15.56, /? = 0.756, ν = 2.1, Λ = 11.93, T h e design of a riprap-lined channel involves the
and Q = 25 cfs. This channel is OK. selection of a rock size large e n o u g h that the force
Therefore, the final channel design with freeboard added attempting to overturn individual rocks is less than the
would be a depth of 1.45 ft and a top width of 17.5 ft. The
gravitational force holding the rocks in place. Since
channel would be sprigged or seeded to Bermuda grass with
riprap is graded, the design p r o c e d u r e s must also in­
a double layer of fiberglass roving with each layer tacked with
asphalt to protect the channel during the establishment of clude a definition of an appropriate gradation of parti­
the vegetation. cle sizes such that erosion of the smaller particles on
A similar procedure could be used to arrive at the channel t h e surface will leave an armored channel that is sta­
design if other liners were used. ble. Finally, the design procedures must include a
methodology for selecting appropriate underlying fil­
ters so that water flowing b e n e a t h t h e riprap will not
e r o d e the base material. Procedures for selecting these
T h e decision to classify a soil as erodible or erosion
materials are included in this section.
resistant is somewhat subjective. N o r m a n n (1975) sug­
gests that the erodibility of the soil, Κ in the Universal Flow on a P l a n e Sloping Bed
Soil Loss equation, can be used as an indicator of
A t the present time, riprap design procedures are
erosion resistance. H e suggests the following classifica­
evolving. T h r e e procedures are presented: (a) a proce­
tion:
d u r e r e p o r t e d by t h e F e d e r a l Highway Administration
Κ = 0.50 erodible ( F H A p r o c e d u r e ) ( N o r m a n , 1975); (b) a p r o c e d u r e in
Κ = 0.17 erosion resistant. t h e Soil Conservation Service (1979) Engineering Field
For Κ values between 0.17 and 0.50, o n e would n e e d M a n u a l (SCS procedure); and (c) a p r o c e d u r e devel­
to interpolate between the values of m a n d η in Table o p e d at Colorado State University (CSU p r o c e d u r e )
4.9. Soil erodibility values are discussed in C h a p t e r 8. and r e p o r t e d by Stevens and Simons (1971) and Simons
and Senturk (1977, 1992). T h e F H A and SCS proce­
d u r e s are similar in that a stone d i a m e t e r is specified in
terms of t h e d e p t h of flow and channel slope. T h e s e
two p r o c e d u r e s are based on experiments and field
Riprap Linings observations. T h e C S U p r o c e d u r e includes a theoreti­
In situations where vegetation is not suitable, riprap cal analysis plus laboratory and field studies. T h e C S U
is often used to stabilize channels. R i p r a p is generally p r o c e d u r e is m o r e complete and allows the specifica­
rocks of various sizes arranged to prevent erosion of tion of a safety factor. Presumably with a safety factor
channel banks and bottom. of 1.0, the rocks are in a state of incipient motion.
Rocks used for riprap should be dense and hard A complication in riprap design is t h e gradation of
enough to resist deterioration d u e to exposure t o air, rock sizes. Rocks u p to some particular size may be
water, and t e m p e r a t u r e extremes, including r e p e a t e d unstable in a flow, but larger rocks might tend to hold
freezing and thawing if necessary. Sometimes rock that t h e m in place. Experimental work with riprap is diffi­
is initially quarried may a p p e a r satisfactory but is not cult and time consuming because of the size of the
able to withstand weathering. If doubt exists as to the rocks involved, t h e many possible gradations of rocks,
suitability of a rock source, a geologist should b e con­ variation in rock shape, materials and handling costs,
sulted. R o u g h angular rocks are generally preferred as a n d the generally high flow rates required. T h e s e fac­
they interlock and resist overturning better t h a n tors have t e n d e d to limit studies on the stability of
smooth, rounded rocks. riprap u n d e r controlled conditions.
Surfaces on which riprap is placed should b e well T h e C S U p r o c e d u r e is the most theoretically com­
compacted and stable. It is especially important to plete and conservative of the t h r e e procedures. It
ensure that the toe sections for channel bank riprap should result in satisfactory designs. C h a n n e l sections
are safe from scour and sloughing, since failure of the lined with riprap should b e carefully monitored, espe­
toe may result in failure of the entire bank. Rocks cially for t h e first few years after completion, to ensure
should be placed in a m a n n e r that prevents segregation that the selected riprap is stable. Any d a m a g e should
by size. D u m p i n g in a m a n n e r that allows excessive b e repaired immediately to prevent much more exten­
rolling of the rocks in a downslope direction and sive d a m a g e from developing.
spreading with a dozer potentially result in segregation. T h e F H A p r o c e d u r e uses a maximum stable d e p t h
Generally front-end loaders or bucket elevators or of flow given by Eq. (4.28) with η = - 1 . 0 and m =
Design of Open Channels 127

5D /y, 50 where £> is t h e riprap diameter in feet such


50
Note:
that 5 0 % of t h e stones have a diameter smaller than dmax = 0.68 ft
D a n d y is t h e unit weight of water (62.4 l b / f t ) . 3
5 0
db 0.68(18)
Thus d is given by
max R= = 0.63 ft.
2d + b ~ 2(0.68) + 18

d max = 5(D /yS). 50 (4.31) Therefore, the assumption that R = d is reasonable. If the
Abt relationship for η is used, the result is υ = 8.4 fps and
T h e velocity of flow is given by Manning's equation D = 0.95 ft.
50

with a roughness, n, given by

η = 0.0395D^ 6
(4.32) T h e SCS p r o c e d u r e is based o n a chart that can b e
approximated by
so that
D75 = 13.5d S ll

37.7 for rock diameter, D , in feet, d e p t h of flow, d, in


ν = — , R 2 /
W \ (4.33) 7 5

^50 feet, a n d S in feet p e r foot. If D is about 1 . 5 D , as 75 50

r e c o m m e n d e d by Simons a n d Senturk (1977, 1992),


This equation is known as t h e M a n n i n g - S t r i c k l e r then
equation. C h a n n e l design is d o n e by computing d max D 50 = 9d S
ll

and υ for a n assumed £> a n d t h e n determining t h e


or
50

appropriate channel dimensions. T h e calculations a r e


0.91
m a d e easier by assuming R = d . max
d =(D /9S)
mm 50

A p a p e r by A b t et al. (1988) suggests that M a n n i n g s T h e SCS also p r e s e n t e d a chart based o n t h e Isbash


η for riprap in steep channels can b e approximated by curves, which c a n b e approximated by

η = 0.0456(D S)
0.159
(4.34) ν = 12MD% . 51

50

This relationship assumes D = 2 D . A n unattrac­ 1 0 0 5 0

where D is in inches a n d 5 is in feet p e r foot.


50 tive theoretical aspect of this p r o c e d u r e is that ν is n o t
Although this relationship h a s n o t b e e n officially expressed as a function of slope a n d thus t h e equation
adopted in any design procedures, t h e d a t a p r e s e n t e d should n o t b e considered a general result. If t h e ex­
by Abt et al. indicate that it better describes M a n n i n g ' s pression £> = 9d S is substituted into t h e relation­
50
ll

η than does E q . 4.32 for t h e conditions they tested. ship, t h e result is υ = 3 9 . 4 r f 5 , which is analogous 0 5 6 0 5 1

to Manning's equation.

Example Problem 4.15 Riprap—FHA procedure


Example Problem 4.16 Riprap—SCS procedure
Determine the D riprap size required to convey 115 cfs
5Q

down a 10% slope in a rectangular channel 18 ft wide. Work Example Problem 4.15 using the SCS approxima­
Riprap is for the bottom only. Use the FHA procedure. tions.
Solution: Assume R = d , max y = 62.4, S = 0.10. Then Solution

5D ID \'
0 91

J
50

<*max = « 0.801D 50
Q = vA = \2MD^\dB) = 12.84Dgj»|-^ 18
115 = 2 5 4 D ^ 42

» - S(^,ax) 2 / 3
S 1 / 2
- ^(0.801Z) ) 5 0
2 / 3
(0.10) 1 / 2
D 50 = 0.57 ft
^50 ^50 0.91
_ / 0.57
υ = 10.28D5 / 1 2
= 0.66 ft.
d
™ x=
\ 9(0Λ)
Q = uA = 10.28 D £ d B l
5
2
m2LX = 10.28D^ (0.801D )(18) 2
50
For this problem, the FHA and SCS criteria result in similar
designs with the FHA procedure resulting in larger estimates
115 = 148.22D Y 2

for the required D . This will generally be the case.


5

50

D 50 = 0.84 ft
128 4. Open Channel Hydraulics

Median diameter, D ^ , in mm
1 2 5 10 20 50 100 200 500

Ο Rounded
Δ Rounded & angular
• Angular

Figure 4.16 Forces on a particle in a channel bed. F , drag force; d

F lift force; PR, point of rotation.


Ll
J I I I
.01 0.02 0.05 1 0.2 0.5 1 2 5 10 32

Median diameter, D ^ , in in.

Simons and Senturk (1977, 1992) have analyzed sev­ Figure 4.17 Angle of repose of dumped riprap (after Simons and
eral procedures for determining t h e required particle Senturk, 1977, 1992).
sizes for stable channel design. They present t h e C S U
procedure, which encompasses a safety factor (SF)
concept. A S F of one represents a point of incipient
is increased until t h e particles just begin to move, t h e
motion or the flow condition where forces holding
particles a r e at their angle of repose φ, and the safety
particles and those tending to move particles are in
factor is 1.0; hence,
exact balance. A S F of 1.5 would be preferred to add
stability for particles smaller than D a n d to recognize 5 0 tan φ = M / M . 1 2 (4.37)
statistical variability and thus prevent the initiation of
localized movement, which might lead to a general Using E q . (4.37) in (4.36),
failure of the riprap protection. cos 0 tan φ
T h e F H A and SCS procedures are found to have SF,=
s i n 0 + (F M /W M ) + (F M /W M )
safety factors of less than 1.0 using t h e C S U criteria
L 4 S 2 d 3 s 2

(presented later). This indicates potential failure prob­ (4.38)


lems at design flows according to t h e C S U criteria.
T h e C S U procedures is developed by considering the or
forces on a particle on a channel b e d sloping at an cos 0 t a n φ
SF„ = (4.39)
angle θ as shown in Fig. 4.16 along with t h e m o m e n t sin 0 + t] tan φ h

arms about the point of rotation, P R . Summing m o ­


ments about P R : where η is a stability p a r a m e t e r given by
ι>

FM + FM + W sin ΘΜ = W cos ΘΜ (4.35)


F M, L F M,d
L 4 d 3 s 2 s ν

These terms are defined in Fig. 4.16. T h e S F for a


given flow situation is the ratio of t h e resisting m o ­ T h e n a t u r e of % can be determined by looking at
ments to the overturning moments, or the safety factor for a plane horizontal bed, where 0 is
equal to zero. U n d e r these conditions, the safety factor
W cos ΘΜ
% λ becomes
cp __ f . . (4.36)
6
W sin ΘΜ + F M
s 2 L 4 + FM
d 3
W Mx s
cp _
p , a n e
FM + FM
T h e key to a stable design is to make the safety L 4 d 3

factor greater than o n e . T o calculate a safety factor, 1


Eq. (4.36) must be manipulated so that it contains ( F j M ^ M , ) + (F.My/WM
p a r a m e t e r s that are readily measurable or can b e d e ­
termined from tables and graphs. or
O n e readily measurable p a r a m e t e r is t h e angle of (4.41)
S F
plane =
^/^b-
repose of a given riprap, given in Fig. 4.17. W h e n there
is n o flow, the lift and drag forces are zero. U n d e r For S F equal to 1.0, t h e b e d is at t h e point of
p l a n e

these conditions, if the angle of the channel bottom, 0, incipient motion, and t h e tractive force is equal to t h e
Design of Open Channels 129
critical tractive force. U n d e r conditions other than Solution: The solution procedure involves a trial and error
incipient motion on a plane bed, it is reasonable to approach of selecting a riprap size, calculating the depth of
assume that the safety factor can b e given by the ratio flow required to convey the flow, and checking the safety
of critical to actual tractive force since there is no factor to ensure that the channel is stable. Assume a D of 5Q

2.5 ft, from Eq. (4.32).


gravity component along the channel bed. Therefore
η = 0.0395D^ = 0.046. 6

± = I i (4.42)
Vb T From Manning's equation, assuming a wide channel,
and 1.49
τ Q=Av = bd d S^
2/3 2

(4.43)

ί 3/5 3/5
0.046(115)
For fully turbulent flow, Gessler (1971) indicates that
d =
.1.49M
n Q
1.
1.49(18)(0.10)
1 / 2 1/2

the Shield's diagram can be reanalyzed to give


d = 0.75 ft depth required to convey the flow.
r = 0.047?(SG -
c 1)D, Checking for stability using Eqs. (4.44) and (4.39),
where SG is the specific gravity of the particles and D τ - ydS = (62.4)(0.75)(0.10) = 4.68 l b / f t 2

is the representative particle diameter, typically the 21τ 21(4.68)


average diameter. Using Gessler's analysis, = 0.382.
y(SG - 1 ) D 50 62.4(2.65 - 1)2.5

Assuming an angular riprap, Fig. 4.17 gives φ = 42°. For a


η
" =
y ( SG -l)D-
2 ( 4
· 4 4 )

10% slope, θ = 5.71°. Hence, from Eq. (4.39),


If τ is given by y dS, these equations can b e used to cos θ tan φ (cos5.71)(tan42)
design a channel if the flow velocity is determined from SF„ =
sin θ + T) tan φ b sin 5.71 + 0.382 tan42
Eq. (4.33). A n illustration of the design procedure is
given in Example Problem 4.17. It should be noted that SF = 2.02
fe over designed.
these equations do not apply to channel banks, but Calculations to select a better design are contained in
only to the channel bottoms. Channel bank stability is Table 4.12. Use a riprap with a D of 1.7 ft on the channel 5 0

considered in the following section. bed. Obviously, there is a problem with stability of the side
slopes. Also the gradation of riprap must be specified and a
filter blanket selected. This is covered in subsequent sections
and examples.
Example Problem 4.17 Riprap—CSU procedure

A channel is being designed to convey a flow of 115 cfs


down a 10% slope. The soil is collodial silt; hence the critical Example Problem 4.18 Riprap—safety factor
tractive force is so small that a lining is needed. Select an
average diameter of riprap needed to stabilize the channel. Calculate SF for Example Problems 4.15 and 4.16.
For this example, neglect the stability problems associated
Solution
with the side slopes. Assume a bottom width of 18 ft, a
specific gravity of 2.65, and a rectangular cross section. De­ cos θ tan φ
SF =
sign for a safety factor of 1.5. sin Θ + t] h log φ

Table 4.12 Calculations for Example Problem 4.17

Depth to Tractive
Angle of convey force Stability Safety
Manning's repose flow τ factor factor
(ft) η (°) (ft) (lb/ft ) 2
(SF )
(%) b

2.5 0.046 42 0.751 4.68 0.382 2.02


2.0 0.044 42 0.734 4.58 0.467 1.72
1.5 0.042 42 0.713 4.45 0.605 1.39
1.7 0.043 42 0.722 4.49 0.541 1.53
130 4. Open Channel Hydraulics
b e e n given by Stevens and Simons (1971) and Simons
and Senturk (1977, 1992) as

SF =
a tan φ
cos
(4.45)
η' tan φ + sin a cos β
cos λ
VteCoeot β = tan" 1 (4.46)
2 sin α/η tan φ + sin λ
CROSS SECTIONAL VIEW 21r„
V = (4.47)
y(SG-l)D 5 0

and
1 + sin(A + β)
V = V (4.48)

TOP OF SIDE SLOPE-J


where r is t h e maximum shear on the channel bank.
m a x

HORIZONTAL LlNE^y
In o r d e r to derive Eqs. (4.45) through (4.48), it was
assumed that t h e ratio of lift to drag forces was one-
WsSin θ half. T h e use of the procedures is illustrated in Exam­
ple Problem 4.19.
W h e n calculating the shear forces on a channel
bank, it is desirable to take into account variations in
RESULTANT OF ORAS channel shear across the channel bed. Figure 4.8 shows
FORCE AND WsSinO
that for a trapezoidal channel, the maximum tractive
force on the channel walls in Ky dS, w h e r e Κ is 0.74
Figure 4.18 Forces on a particle on a stream channel wall. to 0.78 d e p e n d i n g on the channel side slope.

From Problem 4.15, d = 0.68 ft, D 5 0 - 0.84 ft, and φ = 42° Example Problem 4.19 Riprap size—channel bank
r = ydS = 62.4(0.68)0.10 - 4.24 psf
Based on construction considerations and machinery limi­
21r 21(4.24) tations, side slopes of 2.5:1 are selected for the channel in
Vb = 1.03
y(SG - 1 ) D 50 62.4(1.65)0.84 Example Problem 4.17. Select a riprap size that will be stable
on the channel sideslopes.
cos(5.71)tan(42)
SF = V
' 'V
=0.87. Solution: First the safety factor of the riprap selected in
sin(5.71) + 1.03tan(42)
Example Problem 4.17 is calculated assuming the same mate­
From Problem 4.16, d = 0.66 ft, D 5 0 = 0.57 ft, and φ = 42° rial is used on the sides. From Example Problem 4.17,
τ = 62.4(0.66)0.10 = 4.12 D50 = 1.7 ft; n = 0.043; θ = 5.71° ; d = 0.722 ft.
21(4.12) For a trapezoidal channel, the flow depth can be calculated
= 1.47 to be 0.72 ft, which is insignificantly smaller than 0.722 ft for
62.4(1.65)0.57
the rectangular channel in Example 4.17; hence we use 0.722
cos(5.71)tan(42) ft.
SF = = 0.63.
sin(5.71) + 1.47tan(42)
From Fig. 4.8 r m a x is given by 0.76y dS:
Based on the CSU criteria, both of these designs have
TMAX = (0.76)(62.4)(0.722)(0.10) = 3.41 l b / f t 2

SF < 1. 217 21(3.41)


= 0.408.
y(SG - 1)D 62.4(2.65 - 1)1.7
5Q

Channel Bank Stability Assuming uniform flow, the streamlines are parallel to the
channel bottom and
T h e forces on a channel bank are shown in Fig. 4.18.
Λ = θ = 5.71°.
T h e s e forces are different from those in Fig. 4.16 for a
channel bed since the drag forces are not aligned with Also, for a 2.5 :1 sideslope,
the downslope gravitational forces. T h e solution of the 1
a = tan" = 21.8°.
equations describing the safety factor for this case have 2.5
Design of Open Channels 131

Table 4.13 Calculations for Example Problem 4.19

Maximum Safety Maximum


tractive Channel factor tractive Channel Channel
Depth to force on bed for force wall wall
convey channel stability channel on stability safety
Manning's flow bed factor bed walls factor factor
η a
η (ft) (lb/ft )
2

oi )
b
(SF ) b
(lb/ft )
2
(η') (SF)

1.7 0.043 0.72 4.49 0.541 1.53 3.41 0.308 1.36


2.0 0.044 0.73 4.58 0.467 1.72 3.48 0.268 1.45
2.5 0.046 0.75 4.68 0.382 2.02 3.56 0.220 1.56
2.2 0.045 0.74 4.62 0.429 1.84 3.51 0.247 1.50

''Use a riprap with a D 5 0 of 2.2 ft for both channel sides and bottom.

From Eq. (4.46), 100 Ί—I I I I I


90 -
cos λ 80 -
β = tan"
2 sin α/η tan φ + sin λ 70 -
£ 60 -

= tan"
cos(5.71)
1 50 -
2sin(21.8)/0.408tan(42) + sin(5.71) § 4 0

β = 25.1° Q. 30 -
20
From Eq. (4.48), 10
I I I
1 + sin(A + β ) 1 + sin(5.71 + 25.10) 0 0.1 D,
50 0.5D,50 10,50 2D.'50
η = η = 0.408
Figure 4.19 Suggested size distribution of riprap (after Simons and
Senturk, 1977, 1992).
η' = 0.308.
From Eq. (4.45),
cos a tan φ
SF = (1977, 1992) based o n studies at Colorado State Uni­
η' tan φ + sin a cos β versity. T h e proposed gradation is shown in Fig. 4.19.
cos(21.8) tan(42)
Selecting an Underlying Filter
=
0.308(tan(42)) + sin(21.8) cos(25.1)
T h e placement of a properly designed filter blanket
SF = 1.36.
u n d e r n e a t h t h e r i p r a p is necessary w h e n t h e particle
Thus the riprap is stable, but does not have the required
size of the riprap is much larger t h a n that of the base
safety factor of 1.5. The selection of an acceptable riprap for
the channel side slopes will be made using trial and error. material. T h e following criteria have b e e n established
The calculations are in Table 4.13. It is assumed that the for sizing the filter, based o n t h e size distribution of
riprap on the channel bed will be the same as that used on t h e riprap a n d t h e base material:
the side slopes. It would obviously be possible to vary the
side slopes and channel width to obtain a smaller D . The P (filter)
50 D (riprap)
5 0
5 0
(1) —— r < 40 also — — — 7 - < 40
final selection of channel dimensions and riprap size would £> (base)50 D (filter)
50

have to be based on economics.


(2)
D ( filter)
15 D A riprap)
A

5 < L < 40 also 5 < J * 51


/ < 40
Selecting Proper Gradation D (base)
1 5 £> (filter)15

It is important for a riprap to have a gradation such £> (filter)15 £> (riprap)
15

that the voids between the larger particles are filled (3) —— r < 5 also — — τ ζ 7 — — < 5.
D (base) 8 5 D (filter)
85

with smaller particles to reduce flow b e n e a t h the riprap


and t h e formation of o p e n pockets. A suggested grada­ T h e s e criteria were developed for sizing filters
tion for riprap has b e e n m a d e by Simons and Senturk a r o u n d drain pipe to prevent piping of the soil into the
132 4. Open Channel Hydraulics

PARTICLE SIZE (mm)


Figure 4.20 Size distribution determinations of filter material for Example Problem 4.20.
The filter must have a size distribution within the region of overlap.

drain. A filter designed by t h e same criteria should and


prevent piping of t h e parent soil from b e n e a t h riprap.
Filter thickness should be approximately one-half t h e D (filter)
15

< 40 giving D (filter) < 40 X 0.17 = 6.18 mm


15

thickness of t h e riprap, but in n o case less than 6 to D (base)


15

9 in. A n illustration of t h e use of these procedures is


given in Example Problem 4.20. Plastic filter cloth is Criterion (3)
being used in some cases rather than granular filter D (filter)
15

materials. N o r m a n n (1975) should b e consulted for < 5 giving D (filter) < 5 X 1.5 = 7.5 mm.
Z) (base)
15

details.
85

Therefore, with respect to the base parent material, the


following criteria must be satisfied:
Example Problem 4.20 Riprap filter design 0.85 mm < D (filter) < 6.8 mm 15

Select an appropriate riprap gradation for riprap with a and


D value of 1.0 ft. The base parent material on which the
50
D (filter) < 20 mm.
50
riprap is being placed has the properties D = 0.5 mm, 5 0

£> = 1.5 mm, and D = 0.17 mm. Select an appropriate


g5 1 5 These points are plotted as solid dots in Fig. 4.20 and curves
filter blanket for the riprap. approximating these conditions were drawn through the
Solution: Based on Fig. 4.19 with a D of 1.0 ft, the points.
5 0

properties of the riprap are D = 2.0 ft = 610 mm, D = Next, the filter must be sized relative to the riprap.
1 0 0 50

1.0 ft = 305 mm, D = 1.7 ft = 520 mm, D = 0.42 ft = 130


g 5 15 Criterion (1)
mm, and D = 0.10 ft = 30 mm.
Q

These are plotted in Fig. 4.20 along with the size distribu­ D (riprap)
50 305
tion of the parent material. Next the filter blanket must be Z) (filter)
50
< 4 0 β Μ η β D 5 o ( f i l t C r ) >
10" = 7 6 m m

sized. Look first at the requirements of the filter blanket with


respect to the parent material: Criterion (2)
Criterion (1)
D (riprap)
15 130
Z) (filter)
D (filter) " Γ
50 > 5 δ ί ν ί η β D l 5 ( f i l t C r ) < = 2 6 m m

< 40 giving D (filter) < 40 X 0.5 - 20 mm


50
15

D (base)
50

Criterion (2) and

D (filter)
15 D (riprap)
15 130
> 5 giving £> (filter) > 5 X 0.17 = 0.85 mm
D (base)
15
15
D (filter)
15
< 4
° g i v i n g
f i l t e r ) > — = 3.3 mm
Gradually Varied Flow 133
Criterion (3)

D (riprap)
15 130
Z) (filter)
85
<
5 8iYing D
«< > Xfflter > = 26

Therefore the filter must also meet these criteria, or

D (filter) > 7.6 mm


50

3.3 mm < D (filter) < 26 mm


15

£> (filter) > 26 mm.


85

00 02 Q4 Q6 Q8 1.0
These points are also plotted in Fig. 4.20 as solid boxes
and curves drawn through the points. The envelope of points
satisfying both criteria are crosshatched. Any material se­
lected with a size distribution falling within the crosshatched Figure 4.21 Correction factor for shear in flow in a bend (Nor­
area will satisfy the design requirements. mann, 1975).

Flow in Channel Bends


(6) U s e this shear from (5) in t h e stability p a r a m e ­
Because of the curvature in channel bends, the p e a k ters η and η' and d e t e r m i n e the required riprap size
velocity typically occurs on the outside of the center- using p r o c e d u r e s previously discussed.
line, resulting in steeper velocity gradients a n d higher
shear stress values on the outside banks than occur in It must b e pointed out that these procedures have
straight channels. This extra shear must b e considered very limited verification. T h e i r use is still somewhat
when sizing riprap, vegetation, and temporary channel speculative at this point.
linings. A commonly used p r o c e d u r e in riprap-lined
General Comments
channels is to increase the riprap size in t h e channel
bend, or in vegetated lined channels, to line sharp T h e flow range over which differing channel linings
bends with riprap. offer protection d e p e n d s on channel shape and slope.
T h e location of t h e maximum shear varies so much A n example comparison m a d e by N o r m a n n (1975) is
within bends that it is not possible to d e t e r m i n e the given i n Fig. 4.22. A l t h o u g h t h e p r o c e d u r e s used to
exact point at which protection is n e e d e d . Therefore, it calculate the allowable discharge for t h e riprap have
is standard practice to protect the outside bank of the b e e n shown in Simons and Senturk (1977, 1992) to
entire bend. sometimes yield slightly unstable design, the figure
D a t a that can b e used to predict shear in channel gives a reasonable guide to t h e type of channel lining
bends are not abundant. F . J. Watts [as reported by required for varying flow rates and slopes. It should be
Norman (1975)] proposed that a correction factor for pointed out that t h e ranges will change based on side
shear on the channels walls varying from 1.0 to 4.0 slopes a n d erodibility of underlying material.
could be calculated on the basis of v /R , w h e r e υ is
2
d

the average flow velocity in a straight channel and R d

is the radius of the outside bank. A plot of the correc­ GRADUALLY VARIED FLOW
tion factor is given in Fig. 4.21 along with t h e limited
verification data reported by N o r m a n n (1975). T o use T h e relationships p r e s e n t e d in this section are for
the relationship: wide, o p e n channels w h e r e the hydraulic radius may be
approximated by t h e d e p t h of flow. Uniform flow re­
(1) D e t e r m i n e the velocity in a straight channel
quires a channel of constant cross section a n d sufficient
stretch.
length for the gravitational forces to achieve a balance
(2) D e t e r m i n e the radius of curvature of the outside
with t h e frictional resistance. A t changes in slope,
bank, R . d
cross section, or roughness, t h e two forces will not be
(3) Calculate v /R .
2
d
balanced, a n d the flow conditions will adjust toward
(4) D e t e r m i n e the correction factor, fc , from 3
equilibrium. Within the channel reach where this ad­
Fig. 4.21.
j u s t m e n t occurs, t h e flow is said to b e varied flow or
(5) Calculate the corrected bank shear from
nonuniform flow. If t h e change in flow conditions oc­
τ = k ydS.
3
curs gradually over relatively long channel reaches, the
flow is said t o b e gradually varied flow.
134 4. Open Channel Hydraulics
lOOQOf Noting that q /gy 2 3
is F 2
and rearranging the equa­
8000
tion results in
eoao
dy_ Sp-S
(4.53)
dx 1 - F 2

This equation gives the slope of the water surface with


respect to the channel bottom. If dy/dx is positive,
the flow is getting d e e p e r in the downstream direction.
If dy/dx is negative, the flow is getting shallower in
the downstream direction. A dy/dx of zero implies
uniform flow.
A channel is said to have a mild slope if the normal
depth, y , is greater than the critical depth, y . Simi­
n c

larly, if y <y , the slope is a steep slope, and if


n c

y = y , t h e slope is t e r m e d a critical slope. A slope


n c

that is negative or runs uphill in the downstream


direction is known as an adverse slope. Finally a chan­
nel with n o slope is said to be a horizontal channel.
In sketching gradually varied flow profiles, the pro­
files are conventionally labeled with the first letter of
t h e slope type. T h u s Μ denotes a mild slope, S a steep
slope, etc.
If the flow depth exceeds both y and y , the flow is n c

said to b e in zone 1 and is d e n o t e d with the subscript


1. If the d e p t h is between y and y (or between y n c c

and y ) , the zone designation is 2. A d e p t h less than


n

002 003 004 006 010 020 030


b o t h y and y is in zone 3.
n c
CHANNEL SLOPE, S ( f t / f t )
Figure 4.23 depicts possible flow profiles or backwa­
0

Figure 4.22 Comparison of maximum flow rate versus slope for


ter curves. T h e slope of the water surface for the
various channel linings (Norman, 1975).
various situations can b e deduced from Eq. (4.53). T o
d o this, o n e can approximate S as the slope calculated
from Manning's equation using the actual depth of
Equation (4.4) and Fig. 4.4 give the total energy, H, flow. S is the slope in Manning's equation correspond­
Q

as ing to y . T h e appropriate equations are


n

H = y +z + y /2g.
2 (4.49)
qn 2 2

S =
This may be written as 2.22y 10/3

Η = y + ζ + Q /2gA . 2 2
(4.50) and
Differentiation with respect to JC, the distance along 2 2

the channel, yields


2 . 2 2 ^ 1' 0 / 3
dH _ dy dz Q 2
dA
(4.51) If y > y, S < S. If y > y , S > S. Also note that if
n 0 n Q

dx dx dx gA 3
dx S is less t h a n or equal to zero, y is not defined.
0 n

If we consider a rectangular channel or a wide channel, A s an example of determining the slope of the water
the last term of this equation becomes surface, consider an M profile. In this situation y > l n

y , y > y , and y > y . T h u s F < 1 and 5 > 5 . This


c n c 0

Q 2
dA q 2
dy m e a n s that both the n u m e r a t o r and denominator of
gA3
dx gy 3
dx Eq. (4.53) are positive, so dy/dx is positive and the
flow d e p t h increases in the downstream direction.
T h e term dH/dx represents the slope of the energy
A s another example, consider the S profile. H e r e
grade line, S, which is by convention taken as positive 2

yc y > y y >>
* y < y - This m e a n s F > 1 and
> a n c

downward. Similarly dz/dx is the channel slope, 5 , 0


n n c

5 > 5 . T h u s , the n u m e r a t o r of Eq. (4.53) is positive


also positive downward. T h u s 0

and the denominator is negative. T h e S curve has 2

r 2
dy dy/dx negative or the d e p t h decreases in the down­
-5= -S +0 |1 - - . (4.52) stream direction.
Figure 4.23 Possible flow profiles.

T h e above reasoning can b e applied to each of t h e Table 4.14 Slope of Water Surface Profiles with
zones and profiles with the results shown in T a b l e 4.14. Respect to Channel Bottom
Flow profiles develop at changes in channel slope,
Type Designation Slope
roughness, and cross section. Figure 4.24 shows some
typical situations w h e r e profiles develop. Mild Ml +
A n approximate calculation of backwater profiles
can be d o n e by considering Fig. 4.4 and noting
M2 -
M3 +
E\ + Z\ —E + z + h. Steep SI +
2 2 L

S2 -
By definition S3 +
Critical CI +
So = (z -z )/Ax,
l 2
C3 +

where Δ χ is the length of the channel reach. Also


Horizontal H2
-
H3 +
h =dE
L = S Ax, f
Adverse A2
-
A3 +
where S f is the friction slope or slope of t h e energy
136 4. Open Channel Hydraulics

AM ¥

Dam

Figure 4.24 Typical flow profiles.

grade line. Combining these three equations results in


Example Problem 4.21 Flow profile

(4.54) A wide, rectangular channel is carrying 10 cfs/ft down a


0.5% slope. The channel has a Manning's η of 0.025. A 2.5-ft
barrier in the channel causes flow to pass over the barrier at
critical depth. Compute the flow profile upstream from the
S can be approximated from Manning's equation by
f barrier to a point where the depth is within 10% of normal
assuming an average flow depth for the reach. Example depth. Figure 4.25 illustrates the physical situation.
Problem 4.21 illustrates the computation of a back­
water curve. Note that for subcritical flow, backwater
curves should be computed in the upstream direction
and for supercritical flow in the downstream direction.
Profile calculations are started at points of known
water surface elevations such as overfalls from a mild
channel (y = y ) or other types of control sections.
c

Application of Eq. (4.54) is known as the direct step η - 0.025

method.
Figure 4.25 Sketch for Example Problem 4.21.
Channel Transitions 137

Table 4.15 Profile Calculations for Example Problem 4 . 2 1

y v a
V2/2g E b
dx J
(ft) (fps) (ft) (ft) <fty (ft)

3.96 2.525 0.099 4.059 0


3.50 2.857 0.127 3.627 3.730 0.00035 -93 -93
3.25 3.077 0.147 3.397 3.375 0.00048 -51 -143
3.00 3.333 0.173 3.173 3.125 0.00063 -51 -195
2.75 3.636 0.205 2.955 2.875 0.00083 -52 -247
2.50 4.000 0.248 2.748 2.625 0.00112 -53 -300
2.25 4.444 0.307 2.557 2.375 0.00157 -56 -356
2.00 5.000 0.388 2.388 2.125 0.00227 -62 -418
1.75 5.714 0.507 2.257 1.875 0.00345 -85 -503
1.70 5.882 0.537 2.237 1.725 0.00456 -45 -548

a
v = qly.
b
E = yp-llg + y.

d
S = (qn/lA9y ^) .
f m
2

^ = (£,-E )/(5 -5 ). 2 f 0

fx = X\+dx.
2

Solution: From Eq. ( 4 . 9 ) , CHANNEL TRANSITIONS

(Y q
2 /0
( 100 \ * / 3 Changes in channel width, shape, slope, roughness,
b o t t o m elevation, etc., cause changes in the flow regime.
'•-(tJ -""· T h e location of t h e s e changes is known as t h e transi­
tion area. Backwater curves can b e calculated to evalu­
From Manning's equation ( 4 . 2 3 ) using q = vy,
ate changes d u e to channel slope or roughness as
qn \V5 _ / 10(0.025)
indicated in t h e previous section. F o r smooth transi­
= 1.68. tions, energy relationships can b e used to evaluate the
Ϊ 495 / 1 2
j ~ [ 1.49(0.005) 1 / 2

impact of t h e transitions. A smooth transition is one in


which energy losses are minimal.
The depth of flow over the brink in Fig. 4.25 is
Consider the channel transition shown in Fig. 4.26.
Assuming n o energy loss t h r o u g h t h e transition,, Eq.
y = 2.5 + y = 3 . 9 6 ft.
c
(4.4) becomes
The solution is carried out by assuming depths and comput­
ing AJC. Table 4.15 shows the computations.
(4.55)

Equation (4.54) can b e used for channels w h e r e the


showing that with a constant total energy there is a
approximation that y = R is not appropriate. T h e cal­
specific energy loss of Δ ζ . A specific energy diagram
culations are somewhat more cumbersome than those
can b e used to visualize t h e flow change that occurs.
illustrated in Example Problem 4.21. Fortunately, ex­
tensive computer programs are available for calculating
flow profiles in natural channels. C o m p u t e r s are gener­
ally needed because of irregularities in natural chan­ EGL
72--Λ—-
nels and the presence of flow obstructions in the form
of bridges, culverts, low dams, etc. T h e most widely
used program in the U.S. is H E C - 2 , a program devel­
oped by the U.S. Army Corps of Engineers (1982)
γ
ΔΖ datum

Hydrologic Engineering C e n t e r in Davis, California. Figure 4.26 A channel transition.


138 4. Open Channel Hydraulics

F =

A 6 6
7 ~~ b + 2zd =
4 + 2(2)(1)
= 0.75 ft
2.52
F = - 0.51 subcritical
^32.3(0.75)

— + ν + Δζ 2

2g
2
(2.52)
+ 1.0 = + y + 0.1
64.4
2
64.4

0.9986 =
64A +
' 2

Figure 4.27 Specific energy representation in a transition. Q vA x


2.52(6)
A A 2 4y + 2y|
2

3.54
Consider Fig. 4.27. If y is subcritical and represented
x
0.9986 =
by y , y must correspond to y so that the d e p t h of
a 2 b
(4v + 2 2y\)A

flow d u e to a channel bottom rise is decreased. Con­ Solve by trial


versely, if y is supercritical and equal to y , then y
x d 2

must correspond to y . It must b e kept in mind that >2 Right-hand side


e

the specific energy diagram corresponds to a constant


0.90 1.03
unit discharge or is based on a rectangular channel.
0.75 0.958
For nonrectangular channels, Eq. (4.55) is still valid,
0.84 0.996 OK
but the specific energy representation of Fig. 4.27 can
only be used conceptually, not analytically.
y = 0.84 ft
If the flow must pass through critical depth, the 2

assumption of n o energy loss may not b e valid. This is


especially true if the transition is from supercritical to Check the Froude number:
subcritical flow. In such a situation, a hydraulic j u m p Q 2.52(6) 15.1
= 3.16 ft
accompanied by considerable energy loss occurs. Hy­ A2 4(0.84) + 2(0.84) 2
4.77
draulic jumps are considered in the next section.
A 4.77 4.77
= 0.65 ft
t 4 + 2(2)(0.84) =
736

Example Problem 4.22 Channel transition 1 F = = 0.69 still subcritical.

Solution OK.
A trapezoidal channel with 2 : 1 side slopes and a 4-ft
bottom width is flowing at a depth of 1 ft. The channel is
concrete and on a slope of 0.1%. If the channel bottom is Transitions that consist of changes in channel width
raised smoothly by 0.1 ft over a short distance, what will be
can b e t r e a t e d similar to changes in channel bottom
the depth of flow at the exit of the transition?
elevation. Again specific energy curves cannot be used
Solution directly since they are based on a constant flow p e r
0.015 for concrete unit width, q. W h e n the channel width changes, q must
Hfl2/3 l/2 S
change as well.
η
A = fc/ + 2 r f 2
= 4(l) + 2(l) = 6ft
2 2

Example Problem 4.23 Channel transition 2


Ρ = b + 2dy/z + 1 = 4 + 2 ( l ) v 2 ~ 7 T
2 / 2

8.47 ft A rectangular channel 10 ft wide is carrying 75 cfs. The


R = channel smoothly narrows to 8 ft in width. The flow depth in
A/P = 6/8.47 = 0.71 ft
the 10-ft section is 2.5 ft. What is the depth in the 8 ft section
υ, = assuming no energy losses?
^j(0.71) (0.001)
2/3 1/2
= 2.52 fps
Hydraulic Jump 139
Solution

T g
+y
>=T 8
+y
*
Q 75
= 3 fps
A 10 X 2.5
75
Vy =
Sy~ 2
Figure 4 . 2 9 Location of a hydraulic jump.

(75/8y ) 2
1.36
1.
2

+ 2.5= +y = — +y .
64.4 64.4
2 2

y 2

and
The solution may be found by trial to be y = 2.40 ft. Thus 2

the depth in the 8-ft section is 2.40 ft.


i.I(VTT5f-i).
y is known as t h e initial d e p t h , a n d y is t h e sequent
x 2

d e p t h . A hydraulic j u m p from y <y t o y > y oc­


HYDRAULIC JUMP
x c 2 c

curs whenever flow conditions a r e such that y a n d y x 2

are related by E q . (4.15), that is m o m e n t u m is con­


A n example of a flow transition that is a b r u p t a n d
served.
involves considerable energy loss is a hydraulic j u m p
T h e location of a hydraulic j u m p c a n b e found by
that involves a sudden transition from supercritical to
plotting flow profiles a n d superimposing a plot of t h e
subcritical flow. In looking at t h e flow profiles of Fig.
possible sequent d e p t h above t h e supercritical part of
4.23, it can b e seen that t h e profiles a p p r o a c h critical
t h e flow. T h e j u m p occurs whenever t h e s e q u e n t d e p t h
depth nearly vertically. This is also a p p a r e n t from E q .
line intersects t h e d o w n s t r e a m flow profile (assuming
(4.53), w h e r e as y approaches y , F a p p r o a c h e s 1 a n d
c
t h e j u m p h a s zero length). Figure 4.29 illustrates t h e
dy/dx approaches infinity. W h e n y a p p r o a c h e s y as c
procedure.
supercritical flow from below, a hydraulic j u m p m a y
T h e energy loss in a hydraulic j u m p c a n b e com­
occur as shown in Figure 4.28.
p u t e d directly from Bernoulli's equation as
A hydraulic j u m p cannot b e analyzed using t h e en­
ergy equation because t h e r e is a large a n d unknown 1 1
energy loss in t h e j u m p . By assuming that t h e specific Ε - *- 1 " T i l + ( ) Ί - * 2 ) ·
force plus m o m e n t u m is t h e same before a n d after a
1
2 \y\
8

j u m p , E q . (4.15) can b e used:


T h r o u g h algebraic manipulations, this relationship b e ­
comes
_y\_ o_
2 sy2
(4.57)
4
3Ί?2
Through algebraic manipulations, it may b e shown that

^ = ^ ( V T T 8 F ? - 1 ) (4.56) Example Problem 4.24 Hydraulic jump

A rectangular channel is carrying 100 cfs. The channel is


10 ft wide and flowing 0.90 ft deep. Is a hydraulic jump
possible? If so, what will be the sequent depth? How much
energy is lost in the jump?
Solution

ν 100/10(0.9)
F, = - r = / = 2.06.
yfgy v/32.2(0.9)

Figure 4.28 Hydraulic jump. Since the flow is supercritical, a hydraulic jump is possible.
140 4. Open Channel Hydraulics
The sequent depth is computed from Eq. (4.56) as 1.0%. W h a t is the velocity? W h a t is t h e F r o u d e num­
ber?
y = ^(v 71
+ 8(2.06) - l ) = 2.21 ft.
2 (4.2) W h a t will t h e d e p t h of flow in the channel in
2

problem (4.1) b e if t h e flow rate is 50 cfs? W h a t should


be the freeboard?
Note that
(4.3) A channel is being designed to carry 30 cfs
100/10(2.21) through a very colloidal stiff clay soil on a slope of 1%.
F = 2 / = 0.54. D e t e r m i n e t h e design dimensions if t h e side slopes are
^32.2(2.21)
1:1 using both the tractive force and permissible veloc­
ity methods.
The depth after the jump is subcritical as it must be
(4.4) T h e 10-year p e a k flow from a watershed is to
b e channeled through a grassed waterway of bluegrass
(v y _- :jv —
i) 3
(2.21
ν - 0 . 9 i_
) 3

o.29 ft. on a slope of 4 % over erosion resistant soil. T h e grass


2
= = =
1
Ay y 4(2.21)(0.9)
x 2
may be moved (2 to 5 in.) or unmowed (18 in.). T h e
10-year p e a k flow is 100 cfs. Design a grassed waterway
This loss can also be determined directly from Bernoulli's
to convey t h e flow.
equation as
(4.5) If a straw and erosionet liner is used in the
channel of example problem (4.4), will the channel be
stable before t h e vegetation is established u n d e r a flow
of 10 cfs?
[100/10(0.9)1 2
[100/10(2.21)1 2
(4.6) Design a trapezoidal channel with 2 : 1 side
64.4 + 0.9 - - - τ 1
- .
64.4 — - 2.21 slopes to carry 70 cfs down a 1 0 % slope. T h e channel
bottom width must b e limited to 10 ft because of site
= 0.29 ft.
considerations.
(4.7) A trapezoidal channel with 2 : 1 side slopes, an
8-ft b o t t o m width, and a slope of 0.15% is flowing 1.4 ft
Hydraulic jumps are accompanied by a great deal of d e e p . T h e channel is unlined a n d constructed in an
turbulence and energy dissipation. If a hydraulic j u m p erodible sandy loam soil. W h a t is the flow r a t e ? Is the
occurs in an erodible area of a channel, considerable channel stable at this flow r a t e ?
degradation of the channel may occur. Hydraulic j u m p s (4.8) T h e channel of problem (4.7) is vegetated with
are often used to provide energy dissipation below B e r m u d a grass. W h a t is t h e flow rate? Would t h e r e
spillways and channel drop structures. T o ensure that likely b e any problems with this channel?
the j u m p occurs at a controlled location, generally o n a (4.9) Calculate the critical d e p t h for the channel
reinforced concrete apron, stabilizing blocks are used described in problem (4.7) if it carries 150 cfs.
to add drag forces to the flow. In this case,the m o m e n ­ (4.10) If t h e channel of problem 4.7 is concrete
tum equation is modified to lined, what is t h e critical slope for t h e channel at a flow
r a t e of 150 cfs?
(4.11) A n elevated rectangular canal is flowing 3 ft
gy gy d e e p . W h a t is the horizontal force per unit length
2

exerted by the water on the canal side walls?


where F represents the drag force per unit width. T h e
B
(4.12) W h a t size riprap should line the bottom of a
design of energy dissipation devices such as stilling trapezoidal channel with 4 : 1 side slopes, 10-ft bottom
basins is a special area of hydraulics and is discussed in width, and a 7 % slope? T h e channel is to carry 130 cfs.
the next chapter. Extensive model studies are often (4.13) W h a t size riprap should b e used on the side
employed with the results presented in the form of slopes of t h e channel of problem (4.12)?
dimensionless designs. T h e s e designs are t h e n a d a p t e d (4.14) A vegetated channel is to b e used to carry 50
to particular applications by using appropriate scaling cfs down a 4 % slope. T h e vegetation is to b e B e r m u d a
factors. T h e St. Anthony Falls (SAF) stilling basin is an grass, which may b e long or mowed. T h e soil is an
example. easily e r o d e d sandy loam. Design the channel.
(4.15) Will the channel of problem (4.14) be stable
for a flow of 30 cfs prior to establishment of the
Problems vegetation? If not, select a temporary liner that might
(4.1) A trapezoidal concrete-lined ditch has a bot­ b e used during the vegetal establishment period that
tom width of 3 ft, a d e p t h of 2 ft, and side slopes of will permit the safe passage of a flow of 30 cfs. R e ­
2 : 1 . Estimate the discharge if the channel slope is design t h e channel only if necessary.
Hydraulic Jump 141

n=0.025

Κ 50'
Figure 4A

(4.16) A circular, concrete storm sewer 3 ft in diam­ transition? T h e channel properties are:
eter is flowing at a depth of 2.1 ft. T h e sewer is on a
2 % grade. W h a t is the flow rate? Upstream Downstream
(4.17) W h a t is the flow d e p t h in the drain of p r o b ­
lem (4.16) if it carries 25 cfs?
b 10 ft 8ft
(4.18) W h a t size circular, concrete storm drain would
ζ 3:1 2: 1
be required to carry 75 cfs down a 3 % slope without
s 0.05% 0.05%
surcharging the drain (i.e., always flowing as o p e n
channel flow)?
(4.19) Work problem (4.18) for circular corrugated (4.28) Solve problem (4.27) as if the two channels
metal pipe. a r e reversed so t h a t t h e u p s t r e a m channel becomes the
(4.20) Calculate the flow in the channel shown in downstream channel.
Fig. 4A. T h e slope of the channel is 0.05%. (4.29) A rectangular channel narrows from 20 ft to
(4.21) At what depth would the channel of problem 15 ft. T h e b o t t o m elevation simultaneously drops 2 ft.
(4.20) be flowing if it were carrying 6000 cfs? Both changes are smooth with little loss in energy. T h e
(4.22) Design a riprap-lined channel to carry 75 cfs flow rate is 400 cfs. W h a t is the d e p t h of flow down­
down a 7 % slope. Specify the required riprap size as stream from t h e transition if t h e u p s t r e a m d e p t h is
well as the specifications of the filter material. 4 ft?
(4.23) W h a t type of temporary lining should b e used (4.30) W o r k problem (4.29) as if the channel widens
in a road ditch channel required to carry 10 cfs down a from 15 to 20 ft a n d t h e b o t t o m elevation is raised by
7 % slope? 2 ft.
(4.24) A 25 foot wide rectangular channel with a (4.31) A hydraulic j u m p occurs in a wide channel
Manning's η of 0.025 is carrying 5000 cfs. T h e slope of w h e r e the flow is initially at a d e p t h of 1 ft and a flow
the channel is 0.05%. A t station 22 + 50 the slope of velocity of 14 fps. W h a t is t h e d e p t h after the j u m p ?
the channel changed abruptly to 5 % . Calculate t h e W h a t is t h e energy loss within t h e j u m p ?
flow profile in the u p p e r channel from the channel (4.32) A rectangular channel has a Manning's η of
break to a point where the depth is equal to 9 5 % of 0.02, a slope of 0 . 1 % , a n d a flow r a t e of 10 cfs/ft of
normal depth. width. W a t e r enters the channel as supercritical flow.
(4.25) Calculate the flow profile in the lower channel A hydraulic j u m p occurs. W h a t must be the depth
to a point where the depth is equal to 9 5 % of normal before and after t h e j u m p ? H o w much energy is lost?
depth for the situation of problem 4.24. (4.33) Supercritical flow encounters some stabilizing
(4.26) A wide rectangular channel has a slope of 5 % blocks within a stilling basin. T h e drag force intro­
and a Manning's η of 0.02. T h e channel slope changes duced by t h e blocks is given by C pAv /2,
O
2
where C D

abruptly to 0.04%. T h e flow rate is 12 c f s / f o o t of is a drag coefficient (use C = 1), ρ is the density of
D

width. Calculate the resulting flow profile. w a t e r (1.94 s l u g s / f t ) , and A is t h e cross-sectional


3

(4.27) A trapezoidal channel goes through a smooth area of t h e block perpendicular to the flow. T h e blocks
transition. T h e flow depth is originally normal depth. are 1 ft high and occupy 7 5 % of the flow cross section
T h e flow rate is 50 cfs. If there is no loss in energy, at a d e p t h of 1 ft. W a t e r enters the stilling well and
what will b e the depth of flow immediately after the strikes the blocks. T h e d e p t h of flow is initially 2 ft with
142 4. Open Channel Hydraulics
a flow rate of 25 cfs/ft. W h a t is the downstream d e p t h
immediately after the hydraulic j u m p ? H o w much en­
ergy is lost in the jump?
(4.34) Define the following terms:
(a) uniform flow
(b) supercritical flow
(c) subcritical flow
(d) steady flow
(e) gradually varied flow
(f) rapidly varied flow
(g) energy grade line Figure 4C
(h) velocity head
(i) pressure head
(j) flow profiles
(k) F r o u d e number (4.42) A hydraulic j u m p occurs in the channel of
(1) head loss. problem (4.41) with y = 1.0 ft. Use the diagram con­
(4.35) Flow in a wide rectangular channel encoun­ structed for problem (4.41) to d e t e r m i n e y and the
2

ters a barrier and an overflow spillway as shown in Fig. energy loss.


4.B. Calculate the flow profile from the spillway back
(4.43) Show that a generalized F r o u d e n u m b e r may
u p to the channel to a point where the flow is within
b e written as F = ^Q t/gA , w h e r e t is the top width.
0.2 ft of normal depth. T h e flow rate is 25 cfs/ft, the
2 3

channel slope is 0 . 1 % , and Manning's η is 0.015. T h e


barrier is 3 ft high.
References
(4.36) Calculate the flow profiles resulting from the
flow situation shown in Fig. 4.C. T h e underflow gate is Abt, S. R., et al. (1988). Resistance to flow over riprap in steep
channels. Water Resources Bull. 24(6): 1 1 9 3 - 1 2 0 0
1000 feet upstream from the barrier. T h e barrier is 3 ft
Anderson, A . G. (1973). Tentative design procedure for riprap lined
high. T h e channel slope is 0 . 1 % slope, Manning's η is channels—Field evaluation, Project report 146. Prepared for
0.015, and the flow rate is 25 cfs/ft. If a hydraulic j u m p Highway Research Board. [Original not seen, cited in Normann
will occur, locate the j u m p neglecting t h e length of the (1975)]
jump. T h e underflow gate clearance is 1.0 ft. Anderson, A . G., Paintal, Α. Α., and Davenport, J. T. Tentative
design procedure for riprap lined channels, N C H R P Report 108.
(4.37) W h a t are surface water profiles used for?
Highway Research Board, National Academy of Sciences, Wash­
(4.38) W h a t impact might levees used to protect a ington, D C . [Original not seen, cited in Normann (1975).]
particular region have on flood peaks u p s t r e a m a n d Chow, V. T. (1959). " O p e n Channel Hydraulics." McGraw-Hill,
downstream from the protected area? N e w York.
(4.39) W a t e r is flowing at 13 cfs/ft in a wide rectan­ Fortier, S., and Scobey, F. S. (1926). Permissible canal velocities.
Trans. Am. Soc. Civil Eng. 89:940-984.
gular channel. W h a t is the critical d e p t h ?
Gessler, J. (1971). Beginning and ceasing of sediment motion. In
(4.40) A stream has a slope of 0.03%, a hydraulic "River Mechanics" (H. W. Shen, ed.), Chap. 7. Water Resources
radius of 2.2 m, and an average velocity of 1.2 m / s e c . Publications, Ft. Collins, CO.
Estimate Manning's n. If the channel is 50 m wide, Henderson, F. M. (1966). " O p e n Channel Flow." MacMillan, N e w
estimate the discharge in m / s e c .
3
York.
Lane, E. W. (1955). Design of stable channels. Trans. Am. Soc. Civil
(4.41) A rectangular channel is carrying 10 cfs/ft
Eng. 120: 1234-1260.
of width, (a) Construct a specific energy diagram, and Lane, E. W. and Carlson, E. J. (1953). Some factors affecting the
(b) construct a specific force and m o m e n t u m diagram. stability of canals constructed in course granular materials. In
"Proceedings, Minnesota International Hydraulics Convention,
Joint meeting I A H R and Hydraulics Division, A S C E . "
McWhorter, J. C , Carpenter, T. G „ and Clark, R. N . (1968). Erosion
control criteria for drainage channels, Study conducted for the
Mississippi State Highway Department and the Federal Highway-
way Administration. Department of Agricultural Engineering,
Mississippi State University, State College, MS.
Normann, J. M. (1975). Design of stable channels with flexible
linings, Highway engineering circular N o . 15. Federal Highway
Administration, U.S. Department of Transportation, Washington,
DC.
R e e , W. O. (1949). Hydraulic characteristics of vegetation for vege­
Figure 4B tated waterways. Ague. Eng. 30:184-189.
Hydraulic Jump 143
Schwab, G. O., Frevert, R. K., Edminister, T. W., and Barnes, J. K. Soil Conservation Service (1979). "Engineering Field Manual." Soil
(1966). "Soil and Water Conservation Engineering," 2nd ed. Conservation Service, U.S. Department of Agriculture, Washing­
Wiley, N e w York. ton, D C .
Schwab, G. O., Frevert, R. K., Barnes, Κ. K., and Edminister, T. W. Stevens, Μ. Α., and Simons, D . B. (1971). Stability analysis for coarse
(1971). "Elementary Soil and Water Engineering," 3rd ed. Wiley, granular material on slopes. In "River Mechanics" (H. W. Shen,
New York. ed.), Chap. 17. Water Resources Publications, Ft. Collins, CO.
Simons, D . B., and Senturk, F. (1977). "Sediment Transport Tech­ Temple, D . M., et al. (1987). "Stability Design of Grass-Lined Open
nology." Water Resources Publications, Ft. Collins, CO. Channels," A R S Agricultural Handbooks 667. U.S. Department
Simons, D . B., and Senturk, F. (1992). "Sediment Transport Tech­ of Agriculture, Washington, D C .
nology" (2nd ed.). Water Resources Publications, Ft. Collins, CO.
U.S. Army Corps of Engineers (1982). " H E C - 2 Water Surface Pro­
Soil Conservation Service (1969). "Engineering Field Manual for
files Users Manual." U.S. Army Corps of Engineers, Hydrologic
Conservation Practices." Soil Conservation Service, U.S. Depart­
Engineering Center, Davis, CA.
ment of Agriculture, Washington, D C .
5
Hydraulics of Structures

INTRODUCTION engineer's understanding of why they work, many of


the devices utilized today are remarkably similar to
T h e need to measure and control flows u n d e r either those that were in place h u n d r e d s of years ago.
open channel or pipe flow conditions has b e e n a con­ A basic understanding of the hydraulics of flow
cern for engineers dating back h u n d r e d s of years. control devices provides a basis for developing ade­
W h e t h e r the desire was to measure irrigation water q u a t e design. A s h e a d increases on a structure, t h e
being applied to dry croplands or control rampaging flow that is discharged through the structure increases.
streams, structures provided potential solutions. Struc­ Typical h e a d - d i s c h a r g e relationships for idealized flow
tures vary from very small weirs to large spillways with controls are shown in Fig. 5.1. E a c h of these relation­
energy dissipators and inchide both inlet and outlet ships illustrates that discharge increases proportionally
devices. T h e tremendous variability in physical size, to the h e a d on the flow control device. A n example of
flow capacity, and materials contribute to difficulties in a flow control structure is a principal spillway. A n
designing structures. engineer uses a principal spillway as part of a d a m
design to control the rate at which water is discharged.

HYDRAULICS OF FLOW CONTROL DEVICES


Weirs as Flow Control Devices
Introduction Probably the simplest and least expensive flow con­
Flow control devices may o p e r a t e as either o p e n trol device available for installation in a channel is a
channel flow in which the flow has a free water surface weir, which is simply an obstruction placed in a chan­
or pipe flow in which the flow is in a closed conduit nel so that the flow is constricted as it goes over a crest
under pressure. Although the hydraulic principles are (see Fig. 5.1). T h e crest is the edge of t h e weir over
similar for each, the application of these principles for which the water flows. As the water level rises above
describing the hydraulics of flow control devices varies the crest, the flow rate increases dramatically. Weirs
considerably. Simple flow control devices date back at can have a crest that is generally thin, o r s h a r p crested.
least to the ancient Egyptians who sought to control Sharp-crested weirs are constructed from sheet metal
the Nile River's yearly flooding. Although the struc­ or similar thin material so that the flow over the weir,
tures have changed in terms of materials and the or n a p p e , springs free as it leaves the upstream face of

144
Hydraulics of Flow Control Devices 145
SLUICE GATE
Q- KLy (2gH)
2
0 , 5

SHARP CRESTED WOR


Κ dec. o s ( y / H ) i n c
2 Q- CLH
from 0.5 to 0.6 C- 3.27 • 0.4H/W
L - gote width L - Width off W#ir
^ « gate opening

Mutt bt V«otikit«d

BROAD CRESTED WEIR SCC CONTRACTED WOR

>2M L>3H >2H L-Widlh

ΐΤΓ~1 —

1.5 W>2M
Q - 3.087 L H
L - Width

ORIFICE

(o) k~ «4- Q - C'A(2gH)


0.5

Α· Orifice Areo
V-NOTCH Q-Klonf Η **
K-2.5 to 2.7
2

«>-
(a) C'-0.61 6>*s
Q-2.5 H F 0 R β - 9 0 *
2 a

(b) C' - 0 . 9 8
(c) C*-0.80
(d) C ' « 0 . 5 1
(c)

WATERWAY EXPERIMENTAL
STATION STANDARD SPILLWAY

Figure 5.1 Typical head-discharge relationships (Kao, 1975).

the weir. Broad-crested weirs are sometimes used precise flow m e a s u r e m e n t is desired. Conversely, rec­
w h e r e a structure previously existed or w h e r e debris tangular weirs have large capacity but have less sensi­
may damage a sharp-crested weir. Broad-crested weirs tivity for flow m e a s u r e m e n t . A weir can vary in physical
are discussed in a subsequent section. size from quite small to very large. Consequently, the
Sharp-crested weirs can have several shapes, includ­ controlled discharge can vary substantially. T h e dis­
ing rectangular, triangular, trapezoidal, or a combina­ charge across a rectangular weir is defined by the
tion of these, to provide the desired sensitivity at the equation
required flow capacity. A weir is classified according to
Q = CLH ' , 3 2
(5.1)
the shape of its notch. Triangular (also called V-notch)
weirs have greater control u n d e r low flow conditions w h e r e Q is discharge in cubic feet p e r second, C is the
than do rectangular weirs and are often used w h e r e weir coefficient ( d e p e n d e n t u p o n units a n d weir shape),
146 5. Hydraulics of Structures
L is weir length in feet, and Η is head in feet. For a the head on the riser is 1 ft and weir flow exists. Assume the
circular inlet, such as the riser on a d r o p inlet, L is the weir coefficient C equals 3.0.
circumference of the pipe. Solution: Weir flow control will occur at low head. The
Values of C can be found in hydraulic references for governing equation (using C = 3.0) will be
many shapes. Values of C from 3.0 to 3.2 are generally
used for a rectangular weir. T h e length L is the total Q = 3.0L// 3 / 2
,
length over which flow crosses the weir. If the weir where the length L is the circumference of the pipe, which
coefficient and weir length are known, discharge is a equals ttD or π (2 ft) or 6.3 ft. Therefore substituting L
function of head only. H e a d is the difference in eleva­ yields the equation for discharge as a function of head as
tion between the lowest point on the weir crest and the Q = 18.9 H 3 / 2
. (a)
water surface elevation plus the velocity head. T o avoid
having to estimate velocity head, Η should be mea­ With a head of 1 ft, the discharge Q under weir flow would
be 18.9 cfs.
sured at a location at least three times the maximum
design head upstream of the weir at a point w h e r e the
velocity head is negligible. T h e crest should b e located
at least two times the maximum head above the chan­ G r a n t (1978) presents an eloquent discussion of
nel bottom in o r d e r to reduce the likelihood of sub­ weirs, flumes, and o p e n channel flow measurement. H e
mergence. Submergence (tailwater approaching the discusses general requirements that lead to precise
crest of the inlet section) can significantly decrease the flow m e a s u r e m e n t using weirs. T h e s e criteria include
capacity of the structure. T h e designer is cautioned to construction techniques, installation, and head mea­
consider this possibility since it necessitates the use of surement. G r a n t (1978) also provides detailed informa­
different equations or field calibration of the weir tion on the selection of weirs and other control devices,
u n d e r the submerged conditions. If the downstream and he includes equations for other angles of triangu­
channel is not adequate to handle the design flow, flow lar weirs, as well as other shapes.
may be retarded and lead to submergence. Flow should
spring clear of the downstream portion of the weir so
that an air pocket forms b e n e a t h the n a p p e . Air is
Orifices as Flow Control Devices
continuously removed from this pocket by the over­ A n orifice is an opening through which flow occurs.
flowing jet. T h e pressure in this pocket should b e kept Orifices can be used to control flow, as in the case of
constant or the weir will have undesirable characteris­ t h e d r o p inlet shown in Fig. 5.2, or they can be placed
tics. French (1985) stated two such characteristics. As either in a pipe or at the end of a pipe to measure flow.
air pressure in the pocket decreases, the curvature of As water flows through the opening, it can be mea­
the overflowing jet increases, and the value of the sured because its discharged velocity through the open­
coefficient of discharge will increase also. Alternatively, ing is a function of head on t h e orifice. Orifices provide
if the supply of air is irregular, the jet will vibrate. Flow a simple m e a n s to m e a s u r e pipe flow. T h e equation for
over the weir will become unsteady, which can lead to orifice flow is
failure of the structure.
Triangular weirs have angles ranging from 22.5° to
Q = C'A(2gHY , /2
(5.3)
90° with 90° being most prevalent (see Fig. 5.1). T h e w h e r e C is the orifice coefficient, A is the cross-sec­
basic equation for discharge through a triangular weir, tional area of the orifice in square feet, g is the
neglecting the velocity of approach and with no sub­ gravitational constant, and Η is the head on the orifice
mergence, is as shown in Fig. 5.2. T h e leading edge of an orifice can
Q = Ctan(0/2)// 5 / 2
, (5.2a) be r o u n d e d or sharp. C is 0.6 for sharp-edged orifices.
Values for other shapes are provided in Fig. 5.1 and
where Θ is the notch angle. For an angle of 90°,
Hoffman (1974). Streeter (1971) presents a discussion
tan 0 / 2 is 1.0; hence of the theory from which Eq. (5.3) is derived. Equation
Q = CH . 5/2
(5.2b) (5.3) is developed in a less rigorous format in Soil
For a 90° triangular weir, C is typically 2.5. Conservation Service (1984).

Example Problem 5.1 Weir flow Example Problem 5.2 Orifice flow

A 24-in. circular, vertical riser constructed from corru­ Estimate the discharge through a 24-in. circular, vertical
gated metal pipe (CMP) serves as the inlet for the principal riser such as described in Example Problem 5.1 if the head is
spillway of a detention structure. Estimate the discharge if 1 ft and the riser is functioning as an orifice.
Hydraulics of Flow Control Devices 147

Elevation of Water in Reservoir

IT

Elbow <Sc Transition

Figure 5.2 Energy losses for flow in a drop inlet spillway considering bend losses and entrance
losses separately.

Solution: The discharge under orifice flow will equal velocity h e a d , V /2g,
2
and t h e transition and b e n d
h e a d losses a r e c o m b i n e d into a single h e a d loss term,
Q = C'a(2gH) . l/2

t h e n t h e total h e a d H' can b e written as


The area of 24-in. pipe is 3.14 ft . Assuming a value of 0.6 for
2

V 2

C since the riser is corrugated metal pipe and substituting H' = —(l+K c + K b + K L), c (5.4)
values including the gravitational constant, we have

Q = 0.6(3.14)^2(32.2)//, w h e r e H is t h e h e a d o n t h e p i p e as shown in Fig. 5.2,


f

K is t h e e n t r a n c e h e a d loss coefficient, K is t h e b e n d
e b

which reduces to h e a d loss coefficient, K is t h e h e a d loss coefficient


c

d u e to friction, L is t h e length of t h e pipe (including


Q - 15.1// 1 / 2
. t h e riser), a n d V is t h e m e a n velocity in t h e pipe. A
Substituting a head equal to 1 ft into the equation yields schematic showing t h e h e a d loss t e r m s is given in
Q = 15.1 cfs for orifice flow. Fig. 5.2. Since discharge through t h e p i p e is equal to
velocity times area, E q . (5.4) can b e solved for dis­
charge as

a(2gH') l/2

Pipes as Flow Control Devices Q = " Ϊ72 > ( · )


5 5

A d r o p inlet spillway consists of a vertical pipe called (1+1^ + K b + K L) C


1 / 2

a riser and a nearly horizontal pipe called a barrel. w h e r e Q is discharge a n d a is cross-sectional a r e a of


This spillway can serve as a flow control device, even t h e pipe. V a l u e s for K a r e given in T a b l e s 5.1 a n d 5.2
c

when operating u n d e r pipe flow. A schematic showing for circular a n d s q u a r e pipes. Values for K a n d K e b

energy losses with pipe flow is given in Fig. 5.2. W h e n d e p e n d o n t h e configuration of t h e e n t r a n c e a n d t h e


the water level shown in Fig. 5.2 rises to a point such b e n d . Typical values for K a n d K are 1.0 a n d 0.5,
t b

that the pipe flows full, t h e total h e a d causing flow is respectively. B r a t e r a n d King (1976), as well as Hoff­
given by H' (as shown in Fig. 5.2) instead of Η as it m a n (1974), can b e consulted for further details.
was for weir and orifice control. This h e a d is dissipated F o r risers with rectangular inlets, t h e b e n d head
as entrance head loss, transition head loss, b e n d h e a d losses a r e frequently combined with t h e e n t r a n c e head
loss, friction head loss, and velocity head. Frequently, losses into o n e term. T h e total h e a d dissipated through
in pipes used to drain detention reservoirs, t h e only t h e riser can t h e n b e written as
transitions and b e n d s are at t h e connection b e t w e e n
the d r o p inlet and the bottom p i p e . If h e a d losses are
H ' = ^ ( l + K ' e + K C L ) (5.6)
given in terms of a head loss coefficient times t h e
148 5. Hydraulics of Structures

ble 5.1 Head Loss Coefficients for Circular Conduits Flowing FulF

Head loss coefficient, K , for circular pipe flowing full


c

K = 5087 c n /D 2 4/3

(Note: Pipe diameter, D , is in inches)

Manning's coefficient of roughness, η

'ipe
meter Flow area 0.010 0.011 0.012 0.013 0.014 0.015 0.016 0.017 0.018 0.019 0.020 0.021 0.022 0.023 0.024 0.025
in.) (ft )2

6 0.196 0.0467 0.0565 0.0672 0.0789 0.0914 0.1050 0.1194 0.1348 0.1510 0.1680 0.1870 0.2060 0.2260 0.2470 0.2690 0.2920
8 0.349 0.0318 0.0385 0.0458 0.0537 0.0623 0.0715 0.0814 0.0919 0.1030 0.1148 0.1272 0.1400 0.1540 0.1680 0.1830 0.1990
0 0.545 0.0236 0.0286 0.0340 0.0399 0.0463 0.0531 0.0604 0.0682 0.0765 0.0852 0.0944 0.1041 0.1143 0.1249 0.1360 0.1480

2 0.785 0.0185 0.0224 0.0267 0.0313 0.0363 0.0417 0.0474 0.0535 0.0600 0.0668 0.0741 0.0817 0.0896 0.0980 0.1067 0.1157
1.069 0.0151 0.0182 0.0217 0.0255 0.0295 0.0339 0.0386 0.0436 0.0488 0.0544 0.0603 0.0665 0.0730 0.0798 0.0868 0.0942
1.230 0.0138 0.0166 0.0198 0.0232 0.0270 0.0309 0.0352 0.0397 0.0446 0.0496 0.0550 0.0606 0.0666 0.0727 0.0792 0.0859

6 1.400 0.0126 0.0153 0.0182 0.0213 0.0247 0.0284 0.0323 0.0365 0.0409 0.0455 0.0505 0.0556 0.0611 0.0667 0.0727 0.0789
8 1.770 0.01078 0.0130 0.0155 0.0182 0.0211 0.0243 0.0276 0.0312 0.0349 0.0389 0.0431 0.0476 0.0522 0.0570 0.0621 0.0674
:l 2.410 0.00878 0.01062 0.0126 0.0148 0.0172 0.0198 0.0225 0.0254 0.0284 0.0317 0.0351 0.0387 0.0425 0.0464 0.0506 0.0549

:4 3.140 0.00735 0.00889 0.01058 0.0124 0.0144 0.0165 0.0188 0.0212 0.0238 0.0265 0.0294 0.0324 0.0356 0.0389 0.0423 0.0459
:7 3.980 0.00628 0.00760 0.00904 0.01061 0.0123 0.0141 0.0161 0.0181 0.0203 0.0227 0.0251 0.0277 0.0304 0.0332 0.0362 0.0393
.0 4.910 0.00546 0.00660 0.00786 0.00922 0.01070 0.01228 0.0140 0.0158 0.0177 0.0197 0.0218 0.0241 0.0264 0.0289 0.0314 0.0341

16 7.070 0.00428 0.00518 0.00616 0.00723 0.00839 0.00963 0.01096 0.0124 0.0139 0.0154 0.0171 0.0189 0.0207 0.0226 0.0246 0.0267

a 9.620 0.00348 0.00422 0.00502 0.00589 0.00683 0.00784 0.00892 0.01007 0.01129 0.0126 0.0139 0.0154 0.0169 0.0184 0.0201 0.0218

is 12.570 0.00292 0.00353 0.00420 0.00493 0.00572 0.00656 0.00747 0.00843 0.00945 0.01053 0.01166 0.0129 0.0141 0.0154 0.0168 0.0182

A 15.900 0.00249 0.00302 0.00359 0.00421 0.00488 0.00561 0.00638 0.00720 0.00808 0.00900 0.00997 0.01099 0.0121 0.0132 0.0144 0.0156
i0 19.630 0.00217 0.00262 0.00312 0.00366 0.00424 0.00487 0.00554 0.00622 0.00702 0.00782 0.00866 0.00955 0.01048 0.0115 0.0125 0.0135

a
From Soil Conservation Service (1951).

or and 5.2. It is being used as the principal spillway for a


1/2 detention structure. The pipe is 60 ft long and has one 90°
a(2gH')
(5.7) bend. The top of the inlet riser is 15 ft above the bottom of
( l + ^ + ^ L ) 1 7 2
' the outlet. Assume a free outfall and estimate the discharge
under pipe flow if the water elevation 30 ft from the inlet is
where K' is the combined entrance a n d b e n d h e a d
e
1 ft higher than the top of the riser.
loss term. By providing a smooth transition, t h e value
for K' can b e reduced. Typical values of K' are given Solution: For pipe flow, we have
e e

in Table 5.3.
1/2
Frequently when the drop inlet is the same size as a(2gH')
the remainder of the pipe, orifice flow will control, and (1 + K + K + K L)
1/2
c h C

the pipe will never flow full. In this case, it may b e


necessary to increase the size of the d r o p inlet in o r d e r
where K « 1.0 for most entrances of interest and K = 0.5.
to utilize the full capacity of the pipe.
e h

Manning's η for CMP is approximately 0.024 (see Table 4.1


for a range of values for CMP). Using this value in Table 5.1,
K = 0.042. Head for pipe flow is the distance from the water
e

Example Problem 5.3 Pipe flow surface to a point 0.6D above the outlet as shown in Fig. 5.2
and 5.3. Η' then is given in terms of the stage, H, by
An 24-in.-diameter corrugated metal pipe (CMP) is at­
tached to the 24-in. vertical riser described in Problems 5.1 / / ' = / / + 15 - 0.6(2.0) = Η + 13.8.
Hydraulics of Flow Control Devices 149

Table 5.2 Head Loss Coefficients for Square Conduits Flowing FulF

K =
c 29A6n /R2 4/3

Manning coefficient of roughness, η


Conduit size R o w area
(ft) (ft )
2
0.012 0.013 0.014 0.015 0.016

2x2 4.00 0.01058 0.01212 0.01440 0.01653 0.01880


2[ x2{ 6.25 0.00786 0.00922 0.01070 0.01228 0.01397
3x3 9.00 0.00616 0.00723 0.00839 0.00963 0.01096
3^x3i 12.25 0.00502 0.00589 0.00683 0.00784 0.00892
4x4 16.00 0.00420 0.00493 0.00572 0.00656 0.00746
4J x 4 ^ 20.25 0.00359 0.00421 0.00488 0.00561 0.00638
5x5 25.00 0.00312 0.00366 0.00425 0.00487 0.00554
5£ x 5 ^ 30.25 0.00275 0.00322 0.00374 0.00429 0.00488
6x6 36.00 0.00245 0.00287 0.00333 0.00382 0.00435
6j x6j 42.25 0.00220 0.00258 0.00299 0.00343 0.00391
7x7 49.00 0.00199 0.00234 0.00271 0.00311 0.00354
7Jχ 1\ 56.25 0.00182 0.00213 0.00247 0.00284 0.00323
8x8 64.00 0.00167 0.00196 0.00227 0.00260 0.00296
%\ x 8 J 72.25 0.00154 0.00180 0.00209 0.00240 0.00273
9x9 81.00 0.00142 0.00167 0.00194 0.00223 0.00253
9\x9{ 90.25 0.00133 0.00156 0.00180 0.00207 0.00236
10 χ 10 100.00 0.00124 0.00145 0.00168 0.00193 0.00220

"From Soil Conservation Service (1951).

Table 5.3 Entrance Loss Coefficients in Drop Inlet Spillways with Rectangular
Drop Inlets 0

Minimum Maximum
clear water with debris
Description of Spillway Κ
1. Round conduit and standard covered top riser, except
with special elbow and transition
D x 1.5D Riser 0.65 0.75*
D χ 2D Riser 0.41 0.50*
D χ 3D Riser 0.25 0.35*
Dx5D Riser 0.17 0.30*
2. Round conduit and standard covered top riser, with
round bottom and square-edged entrance to conduit
D x 3D Riser 0.60* 0.70*
3. Round conduit and standard rectangular open top riser,
with round bottom and square-edged entrance to conduit
D x 3D Riser 0.50* 0.90*
4. Round conduit and standard rectangular open top riser,
with flat bottom and square-edged entrance to conduit
D χ 3D Riser 0.60* 1.10*
5. Round conduit and standard square open top riser, with
flat bottom and square-edged entrance to conduit
( D + 1 2 ) x ( D + 12) Riser 1.20 2.00*

°Soil Conservation Service (1969).


*Estimated values.
150 5. Hydraulics of Structures

The area a of the 24-in. pipe is 3.14 ft (computed or found


2
the water level in the reservoir increases, water flowing
in Table 5.1). Substituting into the pipe flow equation in from all sides of t h e inlet interferes so that the inlet
begins to act like an orifice. A s the h e a d continues to
3.14[2(32.2)(//+ 1 3 . 8 ) ] 1/2
increase, the outlet eventually begins to flow full, and
Q
~ [1 + 1.0 + 0.5 + ( 0 . 0 4 2 ) ( 6 0 ) ] l/2
pipe flow prevails as shown in Fig. 5.3. A s t a g e - d i s ­
charge curve is developed by plotting Q versus Η for
which simplifies to each of the t h r e e relationships (weir, orifice and pipe
flow). T h e minimum flow for a given h e a d is the actual
Q = 11.25(// + 1 3 . 8 ) . 1/2
discharge used. This process is d e m o n s t r a t e d in the
next problem.
Substituting a value of Η equal to 1 ft yields a Q equal to 43
cfs if pipe flow exists. To determine if pipe flow exists for a
drop inlet spillway, the discharge computed by pipe flow,
orifice flow, and weir flow would be compared. If the pipe Example Problem 5.4 Stage-discharge curve
flow discharge is the smaller of the three, pipe flow would
occur. An 18-in. diameter CMP with an 18-in. vertical riser is
being used as the principal spillway for a detention structure.
The pipe is 50 ft long and has one 90° bend. The top of the
inlet riser is 10 ft above the bottom of the outlet. Develop a
stage-discharge curve. Assume a free outfall.
Using Flow Control Structures as Spillways
A given spillway can have a variety of stage discharge Solution: Weir flow control will occur first. The governing
equation (using C = 3.0) is
relationships, depending on the head. A n example of
the impact of flow changes as a result of changing head
Q - 3.0 LH . 3/2

for a d r o p inlet is shown in Fig. 5.3. W h e n the water


level is just above the riser crest (very low head), the For an 18-in. pipe, L is the circumference of the pipe, which
riser crest acts like a weir, and flow is weir controlled. is 4.7 ft; therefore
W h e n the flow is weir controlled, the water level inside
t h e d r o p inlet is lowest n e a r the center of the inlet. As Q = 14.1// 3 / 2
. (a)
Hydraulics of Flow Control Devices 151
After a certain depth of flow occurs, the discharge may be These calculations can be done quite easily using a computer
orifice controlled or spreadsheet. The minimum flow for any head represents the
actual flow. A plot of the stage-discharge relationship is
Q = Ca(2gH) . 1/2
given in Fig. 5.4.

The area of an 18-in. pipe is 1.77 ft . Assuming a value of 0.6


2

for C",
Q = 8.51// 1 / 2
. (b)
Rockfill Outlets as Controls
For pipe flow, Eq. (5.7) is used, or
Rock is by far t h e most a b u n d a n t , and generally
available, building material on e a r t h and can often be
a(2gH') l/2

(c) obtained quite inexpensively. T h e major expenses asso­


(1 + K + K e b + K L) 1/2
C
'
ciated with rock are grading, transporting, and placing
t h e stone. T h e relative p e r m a n e n c e of rock was recog­
Use K = 1.0 and K = .5. K is determined from Table 5.1
nized ages ago a n d rock has often b e e n used to con­
c h c

(using an η of 0.025) as 0.07. The head is determined for this


case to be the distance from the surface to a point 0.6 struct structures t h a t a r e designed for long life. Rock is
diameters above the outlet as shown in Fig. 5.2. The head Η' used as a hydraulic control for many purposes such as
then is given in terms of the stage Η by t o construct protective c h a n n e l linings and breakwa­
ters, a d d stability to d a m s , a n d provide energy dissipa­
/ / ' = / / + 10 - (0.6)(1.5), tion zones for reservoir outlets. Recently rockfill has
which simplifies to b e e n used as a flow control structure. A s such, infor­
mation o n hydraulics of flow in rockfill is n e e d e d .
H' = H+ 9.1.
Therefore Analytical Procedures for Calculating
Flow Hydraulics
1.77[2g(/Y+ 9 . 1 ) ] 1 / 2

T h e study of flow through rockfill has generally


(1 + 1.0 + 0.5 + 0.07 X 5 0 ) 1 / 2
occurred at low Reynolds n u m b e r s in laboratory condi­
tions. Relatively few insitu m e a s u r e m e n t s have oc­
Q = 5.80(// + 9 . 1 ) . 1/2
(d)
c u r r e d because of t h e difficulties in describing the
A stage-discharge curve is developed by plotting Q versus Η stone shapes and controlling t h e flow so that t h e d e ­
for* each of the relationships given in Eqs. (a), (b), and (d). sired flow conditions occur. M u c h of t h e research in
152 5. Hydraulics of Structures

this area has dealt with solution of groundwater and t h e hydraulics of flow through a rockfill of varying
well problems. A n example of this work is that of gradation using sediment-free water. They used the
Stephenson (1979) who sought to correlate research of standard deviation of the particle diameter, σ, as the
flow/head loss relationships for porous granular media m e a s u r e of gradation instead of intrinsic permeability
over a wide range of Reynolds numbers ( 1 0 ~ to 1 0 ) . 4 4
or simple porosity as h a d previous w o r k e r s . A set of
1

His work, using uniform rock sizes, showed that head equations to predict t h e h e a d loss was developed based
loss is proportional to the flow velocity squared. Based u p o n a friction factor—Reynolds n u m b e r relationship.
on this effort, the head loss equation can be solved Six different models were evaluated using 16 rockfill
analytically for many cases in which the flow is fully structures, with 96 tests conducted using three replica­
developed turbulent flow. tions. T h e equations were found to predict the actual
T h e analysis of flow in rock media with uniform h e a d loss data to an average error of 8%.
diameters utilizes a variation of the D a r c y - W e i s b a c h T h e standard deviation was found to b e b e t t e r than
equation porosity alone as a predictor for describing the hy­
draulics of flow. In the original equations proposed by
dh 1 V}
(5.8) H e r r e r a (1989), porosity was included as a p a r a m e t e r .
dl d 2g In a later p a p e r by H e r r e r a and Felton (1991), porosity
was deleted, since it was approximately constant at a
where dh /dl is the gradient of head through t h e rock value of 0.46 throughout all of the tests. In order to
fill, / is the D a r c y - W e i s b a c h friction factor, d is the allow the relationships to b e used for a wide range of
average diameter of the rock, V is the velocity in the
p
conditions, porosity has b e e n included in the equations
pores, and g is acceleration of gravity. This is a varia­ described in this chapter. T h e best set of working
tion of Eq. (4.21). In the form used by Stephenson equations for describing the hydraulics of flow and for
(1979) for rock fill, the constant 2 is left out of the design, considering the H e r r e r a and Felton data, are
equation and a macro-velocity is substituted for p o r e proposed to b e t h e following [this is Model 3 in Her­
velocity. T h e macro-velocity is the velocity o n e would rera (1989)]
have if the flow through the pores were distributed
uniformly over the entire cross section (see discussion 1. Reynolds n u m b e r equation given by
in Chapter 11) and is related to the pore velocity by the
(d - a)V
porosity, or (5.12)

Κ = ν/ξ, (5.9)
2. Friction factor given by
where V is the macro-velocity and ξ is the porosity.
Folding the constant 2 in Eq. (5.8) into the / term, the gdi 2
dh
fk = (5.13)
modified D a r c y - W e i s b a c h equation becomes dl

dh 1 Κ 2 3. T h e friction f a c t o r - R e y n o l d s n u m b e r relationship
(5.10a) given as
~dl =fk
dp Τ 9

1600
+ 3.83 (5.14)
where f is the modified friction factor. Stephenson
k

(1979) proposed that the friction factor could b e


given by
const T h e average diameter, d, and the standard deviation, σ , can be
!

(5.10b) obtained after a sieve analysis using the relationships


fk =
~kT +f
"
d =
where / , is the friction factor for fully turbulent flow
and R is Reynolds n u m b e r given by
e

and
Vd 1/2
(5.11)

Stephenson proposed that f equals 1, 2, or 4 for


t
where all summations are from 1 to η, η is the number of sieve
smooth polished stone, semirounded stone, or angular
groupings from a sieve analysis, d is the average diameter of the
stone, respectively.
i

rock obtained by averaging the sieve opening of two succeeding


R e c e n t work performed by H e r r e r a (1989) and sum­ sieves, and tv is the percentage
{ by weight of the rocks assigned a
marized in H e r r e r a and Felton (1991) sought to define diameter d,.
Hydraulics of Flow Control Devices 153
ι ROCKFILL Hence
HYORAULIC PROFILE
q 1.0
V = - — = — = 0.217 m / s e c .
/lave 4.6

ο ο 2. Calculate Reynolds number. From the problem, σ =


0.001 m, d = 0.02 m, ν = 1.0 Χ 1 0 " m / s e c , and ξ = 0.4; 6 2

hence Eq. (5.12) becomes

(d - a)V (0.02 - 0.001)(0.217)


Ο R
< = νξ =
1.0ΧΐΟ-'(0.4) = 1 0
' m

Ο
3. Calculate friction factor using Eq. (5.14):

1600 1600
fk = + 3.83 = — — + 3.83 = 3.98.
J k
R e 10,308
Figure 5.5 Definition sketch rockfill equations.
4. Calculate head loss from Eq. (5.13):

gd{ 2
dh
4. T h e h 2 —h ave relationships (see Fig. 5.5) given by Α­ ν 2
dl
h x —h 0 + dh (5.15) or

and f V dl
k
2
_ (3.98)(0.217 )(1) 2

dh = = 5.98
h + h gd£ 2 =
(9.8)(0.02)(0.4 ) 2

x 2

h = (5.16)
5. Calculate upstream and average depth from Eqs.
(5.15) and (5.16):
w h e r e d is average diameter of rock fill ( m ) , ξ is
porosity, σ is standard deviation ( m ) , ν is kinematic h = h + dh = 0.78 + 5.98 = 6.76
x 2

viscosity ( m / s e c ) , h is the average water profile


aye h +h
2 x 0.78 + 6.76
inside the rockfill, h is exit d e p t h of the water in t h e
2
^ave =
= 3.77.
rockfill, h is the u p s t r e a m d e p t h , dh is static h e a d
x

drop as flow moves through the rockfill, dl is the flow 6. Additional trial calculations: Since 3.77 is less than the
length through the rockfill, f is t h e friction factor, R assumed value of 4.6, additional trials are necessary. One
k e
approach would be to use 3.77 as the new trial value. This,
is Reynolds number, and V is average bulk velocity
however, leads to oscillating values. An alternate approach is
equal to <?/A , w h e r e q is the discharge p e r unit
ave
to use a second trial that is an average of the original value
width of rockfill. H e r r e r a suggested that a value of 0.46 and the calculated value. Some oscillation occurs with this
be assigned to ξ in design calculations for a graded approach, but the calculated values converge. By combining
rockfill constructed by dumping. Eqs. (5.12)—(5.16) using the values given in the problem, dh
and h can be related by
Procedures for using the rockfill relationships a r e
aye

given in Example Problem 5.5. 1.08 122.1


(a)
r i
ave n
»vi»

and
Example Problem 5.5 Flow hydraulics
through a rockfill 1.08 122.1
hx = h + dh = 0.78 +
2 - — + — r — (b)
A rockfill dam is composed of rock having an average
diameter d of 0.02 m, porosity ξ equal to 0.4, standard Using these relationships, calculations can be quickly made.
deviation σ of 0.001 m, and length dl equal to 1.0 m. Water Assuming that the new trial value, h , is the average of ayenew

with a kinematic viscosity ν of 1 X 1 0 " m / s e c is flowing 6


the previous trial value and the previous calculated value,
through the rock at a rate q of 1.0 cms/m width. Down­
stream conditions control the exit depth of the water h at avc,cal 4.6 + 3.77
2
= 4.18.
0.78 m. Find the upstream height h . x ave, new 2 2
Solution: From Eq. (b), using A a v e , ew
n = ·18 for A , hx = 8.02 and
4
ave

1. Assume a trial A equal to 4.6 m. The average bulk or


ave 8.02 + 0.78
macro-velocity equals q/h where q is flow per unit width.
ayef
= 4.40.
154 5. Hydraulics of Structures
The new value of 4.40 is still different from the previous From Eq. (a),
value of 4.18; hence additional trials are necessary as shown
f = (47,500)(0.064) = 3040.
below. k

From Eq. (5.15),


Trial ave.trial ave.new 1600
fk = + 3.83 = 4.36.
1 4.6 3.77
J k
3040
2 4.18 4.40 Additional trials are shown below using Eq. (a), (b), and (c).
3 4.29 4.22 Calculated f is used as trial f in the next step.
k k

4 4.26 4.27
Trial Calculated
OK
Trial No.
fk fk
1 3.83 4.36
Thus, h avc = 4.27. For this value, dh = 6.95 and h = 7.72. x

Restated, this means that a head of 7.72 m on the rockfill 2 4.10 4.37
with a thickness of 1 m will discharge 1 m / s e c / m of width if
3
3 4.24 4.38
the downstream depth is 0.78 m. 4 4.31 4.38
Note that this problem illustrates a limitation of rockfill 5 4.35 4.39
structures. A head of cover 7 m is required to discharge 1
OK
cms/m. Such a structure would not be feasible. Example
Problem 5.6 illustrates a more typical application of these
structures. Hence, for h = 0.5 m, f = 4.38, V = 0.06 m/sec, / i
x k = a v e

0.5/2 = 0.25 m, q can be solved from q = / i K = 0.0150 ave

m / s e c . For a width of 2 m,
2

Example Problem 5.6 Stage discharge equation Q = 2q = (2)(0.0150) = 0.030 m / s e c . 3

If the rock fill in Example Problem 5.5 is 2 m wide and is Other values are tabulated below.
used as the spillway from a sediment pond, determine the
stage discharge relationship up to an upstream depth of 2 m, η
ι V q Q
using depths of 0.5, 1.0, 1.5, and 2.0 m. Assume that the (m) (m/sec) (m) (m /sec)
2
(m ) 3

downstream slope is such that the downstream depth is


0.5 4.39 0.0600 0.25 0.0150 0.0300
negligible.
1.0 4.22 0.0862 0.50 0.0431 0.0862
Solution: This is also a trial and error computation process. 0.0800
1.5 4.15 0.1065 0.75 0.1600
For a given value of h , a value of dh can be estimated from
x
2.0 4.10 0.1234 1.00 0.1234 0.2468
Eq. (5.16), since h is assumed to be approximately zero.
2

Knowing dh, the unknowns in Eq. (5.14) are f and V. In k

making the calculations, a value of f is assumed, starting


k A plot of Q vs h would constitute a stage-discharge rela­
x

with a value of 3.83 (from Example Problem 5.5) and checked tionship for a negligible downstream depth.
by iteration. Under these conditions, Eqs. (5.12) to (5.14) can
be simplified to

(0.02 - 0.001)K T h e p r o c e d u r e shown in Example Problem 5.6 ap­


= 47,500K (a)
(1.0 X 10" )(0.4)
6
plies only to those conditions w h e n the downstream
d e p t h is negligible. F o r conditions w h e r e downstream
and flow occurs, t h e d e p t h , h would b e a function of flow.
2

(9.8)(0.02)(0.4 ) dh 2
dh In this case, it would b e necessary to calculate h for 2

2 V = 0.03136^. l (b) each discharge, using o n e of t h e flow equations in


C h a p t e r 4.
Since the downstream depth, h is assumed to be negligible,
2

Graphical Procedure for Rockfill Flow Hydraulics


dh = h (c)
v
T h e p r o c e d u r e s p r e s e n t e d above require detailed
For an /ij or dh of 0.5 m and an assumed f k of 3.83, Eq. (b) computations. C o m p u t e r s and spreadsheets facilitate
can be solved for V or these computations considerably. W h e n quick esti­
m a t e s are n e e d e d , graphical p r o c e d u r e s are helpful. By
1/2
(0.03136)(0.5) utilizing Eq. (5.12) through (5.16), t h e graphical rela­
V = = 0.064.
3^83 tionships shown in Fig. 5.6 w e r e developed for predict-
Hydraulics of Flow Control Devices 155
10 ' ' ' '' I l_ l_l I I Μ Also, since dl is 1 m,
— -_i.-4_i-i-i.t- 1
I I1—IUI4.4444
I III
—I1— I It-hI•-<Μ
•»• -t—-«--1--t--M-1-t
dh = 7.5q '
0 653

. I 1 1 1 1 1 111
I I
I TTTK 11 11 11 I1 I 1Μ1 11 1 or
\ ll \ l I ι I11It ι ι" ι~ι ι γι ι
ι i N j 11 1/0.653
v ^sJrs. II II II II II II II dh
ITSJ I III2ml i % J ! ! ! ! q =
I I ISJ 7.5
t ι ι i x i "•"^I J 11 1 1 1 1 11 1II 1
- - —4I—I4-I4-I4-IV _IP ^NJV_4-4-4.4444
1 1
II II II II IΜI ^SJ i^-L » ι ιι 1ι1
ι ι ι ι ι Since the downstream depth is assumed to be negligible, dh
II II II II II II | o | ι ι ι ι 11 is equal to h The results are tabulated below along with a
III III IΜ I I • τι - ^I >πο 11 II 11
m

11
v

I II II II 1 1 1 1 1 1 1 1 summary of the predicted values from Example Problem 5.6.


0.01 0.1 Also included is a tabulation of values calculated using the
d i a m e t e r d (m)
procedures in Example Problem 5.6, but with the assumption
of a porosity of 0.46 and a standard deviation of half the
average diameter, or 0.01 m. These were the standard condi­
0.7 I
11I IIΓΤΠΠΤ
II II II II II II 1 1
11 1 11 1
1 1—ΓΤΠ1Ί
1
11 11 111 tions used for developing Fig. 5.6.
11 1 1 1
1—(-4H-H4 1 1 1 1
1
1— 1 1
f--1- -
f1HH1
H -
fII 1
1 1 1 1 1 1 11 1 1 1 1 1 1 1
> -J ±UAlL
— m - H + H M-HH+tt Discharge q (m /sec)
2

r r
ι mm ι ι I Γτ 1 1 1 _^l 1 1 1
' ι ι i ι ι ι 11 Upstream
i--._rf__C44.444 1
1 1—U44444 Example New
0.65 1 11 11 1 1 111 11 II 1 depth Graphical Problem 5.6 computation
X mmr^T
1 1 (Mill ι ι ι μ ι 11 111 H11+H11HH11-f 11 11 11
1
h (m) procedure ζ=0.40, σ=0.001 ζ=0.46, o=d/2
X-—\ (--h-mn-i 1 1 l__.JJ_U x

I 1 1 1 1 1 111 11 1
J11 111 L4JJJJ 1 11 1 1 11 11 11
1 1 1
1 1
1 1
1 1
1 1
1 1
r—ι—ι-τ-1-ι-ΐΎ
1
1 1 1 1 1 1 11 0.5 0.0158 0.0150 0.0162
1 r-ι—r-it
ι ιιι ιI
U
1 11 U1 4-1444
1 1 1 11
i-1ι 1ι1 U 4-1-1-14
1 1 1 •1 1 1I - - + + 4 H - H 4 1.0 0.0457 0.0431 0.0474
0.6
0.01
— 0.1
1 1 1 1 1 1 11 1.5 0.0850 0.0800 0.0887

1 1 1 1 1 II 1 2.0 0.1321 0.1234 0.1379


d i a m e t e r d (m)

Figure 5.6 Constants for the rockfill head loss equations.


The results indicate that the graphical procedure gives good
agreement with the analytical procedure. It is important to
note again that conditions deviating greatly from the stan­
ing the average gradient through rockfill, dh/dl. T h e s e dard conditions utilized for Fig. 5.6 could cause the proce­
relationships can b e used to develop h e a d loss as a dure to be in error.
power function of flow, or

dh/dl = aq b
(5.17)
Single- and Multistage Risers
where a and b are constants d e p e n d e n t on rock size
and flow p a t h length, dl T h e discharge q is flow p e r S t a g e - d i s c h a r g e relations for outlet structures are
unit width of rockfill and has units of c m s / m . T h e b a s e d on t h e physical characteristics of the outlet
gradient, dh/dl, is dimensionless. A n example showing structure. O u t l e t structures c a n have many different
the use of this equation is given in Example P r o b ­ e n t r a n c e a n d exit conditions. O u t l e t structures may
lem 5.7. consist of a single weir or orifice, multiple weirs or
orifices, or weirs a n d orifices used in conjunction with
e a c h other. Traditionally, basins have b e e n designed to
control t h e runoff from a long r e t u r n event (i.e., 100-
Example Problem 5.7 Use of graphical method
year event). A basin sized for a long r e t u r n event tends
for rockfill outlets
t o overcontrol m o r e frequent events (i.e., 2-year events).
Utilize the graphical method in Fig. 5.6 to calculate the If t h e design is b a s e d o n frequent events, t h e structure
stage discharge relationship in Example Problem 5.6. Assume t e n d s to u n d e r c o n t r o l long r e t u r n events. T h e s e obser­
a porosity of 0.46. Compare the results to those obtained in vations have led t o regulations that require two or
Example Problem 5.6. m o r e stage risers. T h e development of s t a g e - d i s c h a r g e
curves for single-stage risers follows t h e procedures
Solution: The solution with the graphical method elimi­
nates the trial and error involved in the process. From Fig. that w e r e discussed previously. Two-stage risers can
5.6, a = 7.5 and b = 0.653; hence consist of any combination of orifices, weirs, and pipes.
M o s t c o m m o n is a riser consisting of an orifice for low
dh/dl - 7.5-7 0653
. flow ( m o r e frequent) events a n d a weir for high flow
156 5. Hydraulics of Structures
(less frequent) events. Orifices a r e often staggered (ft). T h e overtopping discharge coefficient C is d e ­ d

around t h e circumference of t h e riser. fined as C = k C , w h e r e C is t h e discharge coeffi­


d t r r

A multistage riser may b e created by having m o r e cient a n d K is t h e submergence factor. T h e coeffi­


t

than o n e row of orifices on a riser. T h e size or n u m b e r cients a r e given in Fig. 5.7. R o a d overtopping locations
of orifices may change from o n e level on t h e riser t o are characterized by a sagging vertical curve such that
another. Equations used for estimating t h e discharge the length a n d elevation of t h e roadway crest a r e
for weirs a n d orifices have b e e n presented previously in difficult to quantify. F e d e r a l Highway Administration
this chapter. (1985) suggested two m e t h o d s of characterizing t h e
sagging vertical curve.
For culvert designs t h a t accompany road overtop­
Broad-Crested Weirs
ping, it may b e a d e q u a t e t o represent t h e sagging
A broad-crested weir supports t h e flow in t h e longi­ vertical curve by a single horizontal line. T h e length of
tudinal direction (direction of flow) so that t h e n a p p e this line is t h e n taken to b e t h e length of t h e weir.
flowing across t h e weir does not spring free from its T h e second design m e t h o d involves breaking t h e
upstream face as shown in Fig. 5.1. Broad-crested weirs sagging vertical curve into a series of horizontal seg­
are usually calibrated in t h e field or by using a model. ments. Flow across each segment is then calculated for
They tend to b e structurally stronger than sharp-crested a specified h e a d w a t e r using E q . (5.20). Flows from
weirs, a n d they a r e particularly useful in locations each segment a r e t h e n accumulated t o obtain t h e total
where sharp-crested weirs suffer maintenance prob­ flow across t h e roadway. T h e elevation a n d consequent
lems. A n example is a forest area w h e r e large limbs or h e a d Η for each horizontal segment is that of t h e
logs may crash into a sharp-crested weir a n d cause segment, n o t a n average.
damage. Streeter (1971) described t h e discharge rela­ While calculation of t h e flow over t h e roadway is
tionships for broad-crested weirs and showed that relatively simple, t h e difficulty is that this only repre­
sents a portion of t h e design flow if t h e r e is a culvert
Q = 3.09L//
t
3 / 2
, (5.18) through t h e road fill a n d this culvert continues to carry
flow. T o d e t e r m i n e t h e h e a d for a given discharge, a
where Q is t h e theoretical discharge from a broad-
t
trial a n d error p r o c e d u r e is required to d e t e r m i n e t h e
crested weir of width L operating with a h e a d of H. flow passing through t h e culvert a n d t h e a m o u n t flow­
Streeter also noted that calibration using a broad- ing across t h e roadway. If t h e head is given a n d a
crested weir produced an equation of t h e form discharge is to b e calculated, t h e total discharge is
simply t h e sum of that across t h e roadway a n d that
Q = 3.03L// 3 / 2
, (5.19)
through t h e culvert.
x

where Q is t h e discharge based on experimental d a t a


x

for a broad-crested weir having a well-rounded u p ­


stream edge. T h e two equations agreed within 2 % . HYDRAULICS OF CULVERTS
Streeter also indicated that viscosity a n d surface ten­
sion have a minor impact on t h e discharge coefficients Culverts a r e conduits that a r e commonly used to
of weirs. Equations (5.18) a n d (5.19) a r e inadequate for pass drainage water through e m b a n k m e n t s . They a r e
design of emergency spillways or road overtopping employed b e n e a t h access or haul roads a n d perpendic­
since they d o not consider t h e effects of flow length ular to roadside ditches. T h e selection of a culvert size
across t h e crest of t h e spillway. to convey flow adequately is p r e s e n t e d in this section.
Roadway overtopping occurs when t h e headwater T h e design of ditch relief culverts is similar to that of
rises to t h e elevation of t h e roadway. Flows u n d e r trickle tubes a n d is discussed in a subsequent section.
these conditions a r e similar to that of a broad-crested R e c o m m e n d e d spacing of ditch relief culverts is a
weir with t h e roadway serving as t h e weir. T h e flow function of road gradient.
across t h e roadway is defined by
Culvert Classes
Q0 = C L(mV ) \
d r
3/
(5.20)
Chow (1959) divided culvert flow into six categories
w h e r e Q is t h e overtopping flow rate ( f t / s e c ) , C is
Q
3
d
as shown in Fig. 5.8. E a c h of these categories is d e ­
the overtopping discharge coefficient, L is t h e length scribed along with m e t h o d s for predicting flow for each
of t h e roadway crest (ft), a n d HW is t h e u p s t r e a mr
type. Chow indicated that t h e e n t r a n c e of a n ordinary
d e p t h (ft), measured from t h e roadway crest t o t h e culvert will not b e submerged if t h e outlet is not
water surface at a location upstream of weir drawdown submerged unless t h e h e a d w a t e r is greater than some
2.90 1.00
0.16 0.20 0.24 0.28 0.32
HW„ ft

A DISCHARGE COEFFICIENT FOR 0.90

HW / L > 0.15 \

\
r r

3.10 GRAVEL-^1
,.ΡΑΥΕΡ 0.80
3.00

2.90
0.70
C f 2.80

2.70
0.80
2.80

2.S0
1.0 2.0 3.0 4.0 0.50
IW , ft
r
0.» 0.7 Ο.β 0.9 1.0

h , / HW,
Β DISCHARGE COEFFICIENT FOR
C SUBMERGENCE FACTER
HW,/!-,^ 0.15

Figure 5.7 Discharge coefficients for roadway overtopping (Federal Highway Administration, 1985).

critical value, H*. T h e value of H* may vary from 1.2 corrugated pipes and concrete pipes, respectively. F r o m
to 1.5 times the culvert height as a result of e n t r a n c e Fig. 5.8, t h e h e a d w a t e r height Η must b e greater than
geometry, barrel characteristics, and approach condi­ t h e critical height / / * , a n d tailwater d e p t h should be
tions. A critical headwater height equal to 1.5 times t h e less t h a n t h e height of t h e culvert. Calculate discharge
culvert height (diameter for circular culverts) is reason­ from Eq. (5.5).
able for preliminary analysis.
Type 3—Outlet Not Submerged, Η > H* Pipe Not 9

Flowing Full This corresponds to t h e hydraulically short


Type 1—Outlet Submerged T h e pipe will flow full.
condition. Check Figs. 5.9 or 5.10 to see if the proper
Discharge is calculated using Eq. (5.5).
conditions are m e t for t h e pipe to b e classified as type
3 by locating the intersection of t h e bed slope and
Type 2—Outlet Not Submerged, Η > H* Pipe Flowing y
ratios in the a p p r o p r i a t e figure. F r o m Fig. 5.8, the
Full This corresponds to a hydraulically long condi­ h e a d w a t e r height Η must be greater than the critical
tion. Figures 5.9 or 5.10 can be checked to provide an height H*, and tailwater d e p t h should be less than the
estimate of whether p r o p e r conditions are met for height of the culvert. Discharge is inlet controlled and
158 5. Hydraulics of Structures

FLOW T H R O U G H NONPRISMATIC CHANNEL S E C T I O N S

Type Profile

(1) O U T L E T S U B M E R G E D
H>d
Y <d t

Full Flow

(2) O U T L E T U N S U B M E R G E D
Η > H*
Y <d t

Full Flow, long pipe

(3) O U T L E T U N S U B M E R G E D
Η > H*
Y <d t

Partly Full, short pipe

(4) O U T L E T U N S U B M E R G E D
Η > H*
Y
t> Y
c
Subcritical Flow

(5) O U T L E T U N S U B M E R G E D -p=r
Η > Η* H*
Y
t< Y
c
Subcritical Flow
Control at Outlet

(6) O U T L E T U N S U B M E R G E D -r=r
Η > Η* H

Y
t< Y
c
Supercritical Flow
Control at Entrance
Figure 5.8 Types of culvert flow (Chow, 1959).

0 5 10 20 25 30IS 35 0 5 10 15 20 25 30 35
_L_ JL.
D 0
Figure 5.9 Criteria for hydraulically short and long culverts rough corrugated pipes. Type 3 is short and Type
2 is long (Carter, 1957).
Hydraulics of Culverts 159

0.09

0.08 h
r/D,w/D-0.04 " CONTROL SECTION W.S
_ CO
0.07 I" °.
ο
I Flgure 5.12 Typical inlet control flow condition (after Federal

>
ο
0.06 Highway Administration, 1985). HW, headwater; TW, tailwater, W.S.,
ο water surface; d , critical depth.
c

CO
ul 0.05
ο.
3
(Λ o.04 a. Submerged

8m 0.03
WATER
SURFACE

CONTROL
0.02 SECTION
DOWNSTREAM

0.01
b. Unsubmerged

D
d c (CONTROL SECTION)
Figure 5.10 Criteria for hydraulically short and long concrete pipes
(Carter, 1957). Figure 5.13 Typical outlet control flow conditions (after Federal
Highway Administration, 1985). H W , headwater, TW, tailwater; W.S.,
water surface; H, losses through culvert.

can be determined by using dimensionless plots devel­


oped by Mavis (1942) as shown in Fig. 5.11.
s h a p e of t h e profile will d e p e n d o n t h e d e p t h of flow at
t h e outlet.
Types 4-6—Outlet Unsubmerged, Η < H* U n d e r T h e factors that influence energy and hydraulic grade
these conditions, the pipe flows as an o p e n channel. lines in culvert discharge are illustrated in the culvert
T h e discharge for a given head d e p e n d s on the pipe schematic shown as Figs. 5.12 and 5.13 (inlet and outlet
slope, entrance geometry, pipe roughness, and pipe control, respectively). T h e s e factors are used to deter­
size. T o accurately predict discharge, it is necessary to mine t h e type of control for t h e discharge. T h e flows
develop a flow profile through t h e pipes. T h e exact shown are either full flow, partly full flow, or free
surface flow.

Inlet Control
Inlet control occurs w h e n t h e section that controls
flow is located at or n e a r t h e entrance to the culvert.
Discharge is d e p e n d e n t only o n the geometry of the
inlet a n d t h e h e a d w a t e r d e p t h for any particular cul­
vert size and shape. T h e inlet will continue to control
flow as long as water flowing through t h e barrel of the
culvert does not impede flow. If control is at the inlet,
downstream hydraulic factors such as slope, length, or
surface roughness will not influence capacity. However,
roughness does influence t h e critical slope at which
inlet control occurs (American Concrete Pipe Associa­
, . 5/2
tion, 1985). Smooth culverts placed o n a very flat slope
can have inlet control, w h e r e a s rough culverts have to
Figure 5.11 Stage-discharge relationship for a circular pipe with b e installed o n a much higher slope to have inlet
the control of the inlet (adapted from Mavis, 1942). control. Several types of inlet control a r e shown in
160 5. Hydraulics of Structures

simple empirical m e t h o d s cannot account for all of


t h e factors that impact flow, b u t they can b e used t o
WATER SURFACE estimate flow through culverts for t h e conditions they
HW represent. C h a r t s in B u r e a u of Public R o a d s (1965a, b)
provide a systematic design m e t h o d that is based on
Outlet unsubmerged
determining the h e a d w a t e r d e p t h from the charts as­
suming both inlet and outlet control. T h e condition
yielding t h e higher h e a d w a t e r d e p t h controls. Profes­
sional groups such as t h e A m e r i c a n Concrete Pipe
HW
Association ( A C P A ) (1985) provide excellent refer­
"^JlDlllll lllr^ ences for maify of t h e figures and techniques that are
Outlet submerged
used in culvert selection. They can often supply d e ­
Inlet unsubmerged
tailed design specifications a n d design p r o c e d u r e s for
specialized products. Additionally, they are often able

.Ε 5
to supply design d a t a on loading.
WATER SURFACE
HW
Critical Depth in Culverts

Inlet submerged
3 ^ W h e n t h e sum of kinetic energy plus the potential
energy for a specified discharge is at its minimum,
-MEDIAN DRAIN critical flow occurs. A n o t h e r way to view this is that
during critical flow, maximum discharge through a pipe
occurs with any specified total energy h e a d . F o r a given
flow rate, the d e p t h of flow a n d slope associated with
WATER SURFACE critical flow define t h e critical d e p t h a n d critical slope.
If t h e culvert has an u n s u b m e r g e d outlet, t h e maxi­
Outlet submerged m u m capacity of t h e culvert is established w h e n critical
Figure 5.14 Types of inlet control (after Federal Highway Admin­ flow occurs (American C o n c r e t e Pipe Association,
istration, 1985).
1985).

Culvert Nomograph Procedure


T h e A m e r i c a n C o n c r e t e Pipe Association (1985) has
Fig. 5.14. Figures showing t h e critical d e p t h for various
suggested t h e following p r o c e d u r e for sizing culverts.
shapes, sizes, and materials are available in many hy­
T h e basic p r o c e d u r e for sizing a culvert for a given
draulic references. If a culvert is operating u n d e r inlet
flow r a t e involves selecting a trial size culvert, deter­
control, it will not flow full throughout the entire
mining w h e t h e r inlet or outlet control prevails, and
length of the pipe.
t h e n finding t h e h e a d w a t e r required for t h e controlling
Outlet Control condition. If t h e h e a d w a t e r is unacceptable (too high
or low), a n o t h e r trial culvert is selected and t h e process
Outlet control occurs when control originates at or
r e p e a t e d . T h e p r o c e d u r e is outlined below. T h e nota­
near the culvert outlet. In this case, discharge is d e p e n ­
tion at this point utilizes H W as t h e h e a d on t h e
dent not only on all of the hydraulic factors u p s t r e a m
culvert instead of Η used earlier, in keeping with
from the outlet including size, shape, slope, length,
standard terminology utilized by public agencies.
surface roughness, headwater depth, and inlet geome­
try, but on tailwater d e p t h as well.
Determine Design Data.

1. Design discharge Q (cfs) for t h e design storm.


Culvert Selection and Design 2. Length L (ft) of culvert.
Culvert selection techniques can range from solution 3. Slope of culvert.
of empirical formulas to comprehensive mathematical 4. Allowable h e a d w a t e r d e p t h H W (ft). This is t h e
al

analysis of specific situations (American Concrete Pipe vertical distance from t h e culvert invert (elevation of
Association, 1985). T h e many hydraulic considerations t h e inside b o t t o m of t h e culvert) at t h e inlet side of the
involved make a precise mathematical evaluation dif­ culvert t o the maximum w a t e r surface elevation per­
ficult and extremely time consuming. T h e relatively missible.
Hydraulics of Culverts 161

5. Flow velocities or tailwater d e p t h in the down­ Table 5.4 Entrance Loss Coefficients (after FHA, 1985)
stream channel.
6. Size, shape and entrance type for trial culvert. A Outlet control, full or partly full entrance head loss
suggested trial size is a diameter (or height for non-cir­
cular culverts) of H W divided by 2.
al
2g_

Find Maximum Headwater Depth for Trial Culvert Under


Inlet and Outlet Control Conditions Type of structure and design of entrance Coefficient K m

A. Inlet control Pipe, concrete


1. Given Q, D , and entrance type, select the Projecting from fill, socket end (groove end) 0.2
appropriate control nomograph to find headwater Projecting from fill, square cut end 0.5
depth required (Fig. 5B.1 or 5B.2).
Headwall or headwall and wingwalls
a. Connect the given culvert diameter D and
Socket end of pipe (groove end) 0.2
discharge Q with a straight line. Continue the line
Square edge 0.5
to the first H W / D scale, indicated as (1).
Rounded (radius = iiD) 0.2
b. Find the H W / D scale that represents the
entrance type used. If necessary, extend the point Mitered to conform to fill slope 0.7

of intersection from the first line horizontally to End section conforming to fill slope* 0.5
scale (2) or (3). Beveled edges, 33.7° or 45° bevels 0.2
c. Multiply H W / D by D to calculate H W . Side- or slope-tapered inlet 0.2
2. If H W is greater or less than allowable, select
another trial size until the H W is within the desired
range.
Pipe, or pipe-arch, corrugated metal
B. Outlet control
Projecting from fill (no headwall) 0.9
1. Given Q, D , entrance type, and estimated tail-
Headwall or headwall and wingwalls square edge 0.5
water depth T W (feet) above the outlet invert for the
Mitered to conform to fill slope, paved or unpaved slope 0.7
design flow in the downstream channel.
a. Select the outlet control n o m o g r a p h for the End section conforming to fill slope* 0.5

desired culvert configuration (Fig. 5B.3 or 5B.4). Beveled edges 33.7° or 45° bevels 0.2
Find the entrance coefficient K from Table 5.4.
c
Side- or slope-tapered inlet 0.2
b. Find the K e on the length scale on the
nomograph.
c. Connect the K point on the length scale to
e

the size of the culvert using a straight line and Box, reinforced concrete
mark the point where the straight line crosses the Headwall parallel to embankment (no wingwalls)
"turning line." Square edged on three edges 0.5
d. F o r m a straight line with the point m a r k e d Rounded on three edges to radius of IT barrel
on the turning line and the design Q and project dimension, or beveled edges on three sides 0.2
to the head scale. R e a d Η on the h e a d scale. Wingwalls are 30° to 75° to barrel
2. If the tailwater, T W , elevation is lower than the Square edged at crown 0.4
top of the culvert outlet, use
Crown edge rounded to radius of h barrel
dc + D dimension, or beveled top edge 0.2
(5.21) Wingwall at 10° to 25° to barrel
Square edged at crown 0.5
or T W , whichever is greater, where d is the critical
c
Wingwalls parallel (extension of sides)
depth (feet) determined from the corresponding crit­
Square edged at crown 0.7
ical depth chart.
Side or slope-tapered inlet 0.2
3. If T W elevation is higher than or equal to the
top of the culvert outlet, set h equal to T W . Find
0
°Either metal or concrete sections commonly available from manu­
H W using facturers. From limited hydraulic test they are equivalent in operation to a
headwall in both inlet and outlet control. Some end sections, incorporat­
HW - Η + h r S L.
0 (5.22) ing a closed taper in their design, have a superior hydraulic performance.
162 5. Hydraulics of Structures
C. C o m p a r e the headwaters required from sections A point of intersection from the first line horizontally to
(inlet control) and Β (outlet control) to d e t e r m i n e scale (3) and obtain a value of H W / D = 1.37.
which is higher. T h e higher headwater controls a n d is c. Multiply H W / D by D = 5 ft to calculate HW =
the flow control existing u n d e r the design conditions 5(1.37) = 6.85 ft. This is the HW required for inlet
for the trial size u n d e r consideration. control.
2. Since HW < HW , OK. al

D . If H W is higher than acceptable and outlet condi­


B. Outlet control
tions control, select a larger culvert size and revise H W
1. Given Q = 200 cfs, D = 5 ft, grooved end projecting,
using section Β for outlet control. It is not necessary to and TW = 4 ft above the outlet invert for the design flow
recheck inlet control if the smaller size culvert proved in the downstream channel.
to be satisfactory in section A. It is also possible to a. The outlet control nomograph for the circular con­
select a n o t h e r shape or inlet condition if desired a n d crete culvert is Fig. 5B.3. The entrance coefficient K = e

repeat the procedure. 0.2 from Table 5.4.


b. Find the K on the length scale on the nomograph.
e

Determine Outflow Velocity c. Connect the K point on the length scale to the
e

size of the culvert using a straight line and mark the


A. If outlet control exists with tailwater, outflow point where the straight line crosses the "turning line."
velocity equals the Q/A. If the outlet is not sub­ d. Use the point marked on the turning line as the
merged, the flow area A is usually based on a flow pivot point and connect with the design Q. Read Η =
depth equal to t h e average of the vertical dimension 2.80 ft on the head scale.
(or diameter for circular pipes) and t h e critical d e p t h 2. Since the TW elevation (4 ft) is lower than the top of
calculated previously. the culvert outlet (5 ft), obtain the critical depth d from c

B. If inlet flow governs, outflow velocity is approxi­ the critical depth chart for 5-ft-diameter circular pipe with
m a t e d assuming o p e n channel flow and using Manning's Q = 200 cfs (Fig. 5C.1) as d = 4.10 ft. Substitute d and
c c

D in Eq. (5.21) for h :


equation as described in C h a p t e r 4. 0

d +D c 4.10 + 5
K = — — - 4 5 5 ft.

Example Problem 5.7 Culvert size with inlet control


Find HW under outlet control using Eq. (5.22)
(culvert nomograph method)
HW = Η + h - S L0 0 = 2.80 + 4.55 - 0.02(180) = 3.75.
Determine the diameter of a circular concrete culvert
(n = 0.012) that must carry a design flow of 200 cfs if the
culvert length is 180 ft and the culvert slope is 0.02 ft/ft. The C. Compare the headwaters required from sections A
allowable HW equals 10 ft. Assume that the culvert has a (inlet control) and Β (outlet control) to determine which is
projecting entrance, and TW equals 4 ft. higher. Since a higher HW is required for inlet control
(HW = 6.85 ft > HW = 3.75 ft), inlet conditions control un­
Solution: der the design conditions for the trial size under considera­
tion.
Determine design data.
1. Q = 200 cfs (given). D. Since HW = 6.85 ft is lower than the allowable HW =
2. L = 180 ft (given). 10.0, the culvert is adequate. If desired, a smaller culvert size
3. S = 0.02 ft/ft (given). might be selected and the process repeated to check if the
4. HW = 10 ft (given).
al smaller culvert might be satisfactory.
5. TW = 4 ft (given).
6. Trial size D = H W / 2 = 5 ft.
aI

Concrete culvert with η = 0.012 and projecting entrance. Culvert Capacity Chart Technique
Find maximum headwater depth for trial culvert under inlet
and outlet control conditions. T h e culvert capacity chart technique offers an alter­
native m e t h o d to d e t e r m i n e the required culvert size or
A. Inlet control h e a d w a t e r d e p t h (Portland C e m e n t Association, 1962;
1. Given Q, D, and entrance type, the appropriate inlet
A m e r i c a n C o n c r e t e Pipe Association, 1985). C h a r t s for
control nomograph to find required headwater depth is
n u m e r o u s shapes, including circular, square, rectangu­
Fig. 5B.1.
a. Connect the diameter D = 5 ft (60 in.), and dis­ lar, and oval pipes, have b e e n developed.
charge Q with a straight line. Continue the line to the F o r specific conditions, a simple p r o c e d u r e is to use
first H W / D scale, indicated as (1). culvert capacity charts such as shown in Fig. 5 A . 1 - 5 A . 8 ,
b. The H W / D scale, which represents the grooved- which are based on t h e d a t a in B u r e a u of Public R o a d s
end projecting entrance, is indicated as (3) so extend the (1965a). T h e culvert capacity charts enable a simple
Hydraulics of Culverts 163

technique to be used for selection of culverts. T h e case, t h e a p p r o p r i a t e outlet control nomograph de­
technique is somewhat limited because t h e pipe must scribing full pipe flow in t h e nomograph procedure
b e installed with the entrance and material as specified should b e used to estimate performance of the cul­
for the specific charts used. vert.
T h e Portland C e m e n t Association (1975) outlined
use of these figures to obtain a direct solution of Determine Outflow Velocity Outflow velocity is deter­
culvert size without the i n l e t / o u t l e t comparison re­ mined in the same m a n n e r as when the nomographs
quired in the nomograph p r o c e d u r e discussed previ­ are used. Culvert capacity charts permit quick estima­
ously. tion of h e a d w a t e r height given a p e a k flow rate and
culvert size if t h e specified design conditions are met.
Determine Design Data
T h e charts also can b e used to select culvert size
1. Design discharge Q (cfs) for the design storm. n e e d e d to m e e t a given h e a d w a t e r limitation such as to
2. Length L (ft) of culvert. prevent roadway overtopping.
3. Slope of culvert.
4. Allowable headwater d e p t h H W (ft). This is t h e
ai

vertical distance from t h e culvert invert at t h e inlet Example Problem 5.8 Culvert size with inlet control
side of the culvert to the maximum water surface (culvert capacity chart method)
elevation permissible.
5. Flow velocities in the downstream channel. Determine the diameter of a circular concrete culvert for
6. Size, shape, and e n t r a n c e type for trial culvert. A the conditions stated in Example Problem 5.7 using the
suggested trial size is a diameter (or height for culvert capacity chart.
noncircular culverts) of H W divided by 2.
al Solution:
Find Culvert Size Determine design data.
1. Locate the appropriate culvert capacity chart for 1. Q = 200 cfs (given).
the culvert size that is approximately half the allowable 2. L = 180 ft (given).
headwater depth. A typical culvert capacity chart is 3. S = 0.02 ft/ft (given).
shown in Fig. 5A.2. D e n o t e d on this chart are headwa­ 4. HW = 10 ft (given).
aI

ter values on the ordinate, discharge quantities on t h e 5. TW = 4 ft (given).


abscissa, solid curves designating inlet control, and a 6. Trial size D = H W / 2 = 5 ft.
al

dash curve indicating outlet control. Some figures con­ Concrete culvert with η = 0.012 and groove end projecting
entrance.
tained in Appendix 5 A do not contain both solid a n d
dash curves for all culvert diameters. This is an indica­ Find culvert size.
tion that the culvert would be unsuitable for the type 1. Select the appropriate culvert capacity chart for the
of control that is missing from t h e n o m o g r a p h . T h e culvert size £>, which is half the allowable headwater depth,
horizontal dotted line denotes that the accuracy of or D = H W / W = 10/2 = 5 ft.
chart values below this line are quite good, w h e r e a s 2. Estimate the index number, L/100 S , which accounts 0

for roughness and length effects relating to pipe flow charac­


prediction accuracy decreases above this line.
teristics:
2. Estimate the index n u m b e r by dividing L by 100
5 . This index accounts for roughness and length ef­
0
L 180
fects relating to pipe flow characteristics. 100 S 0
=
100 (0.02) = 9
°*
a. If the calculated index value for a specific p r o b ­
lem is less than or equal to t h e n u m b e r shown on the Since the calculated index value (90) is less than the number
solid curve for a specified culvert, the culvert will b e shown on the solid curve (700) as shown in Fig. 5A.2 for a
in inlet control and the solid curve will describe 5-ft circular concrete culvert, the culvert will be in inlet
culvert performance. control, and the solid curve will describe culvert perfor­
mance. Reading up from a Q = 200 cfs yields HW = 6.6 ft.
b. If t h e calculated index value is greater t h a n t h e
Since 6.6 ft is less than the allowable HW of 10 ft, the culvert
solid curve value but less than or equal to the dash
will carry the design flow under inlet control.
curve index value, outlet control dominates and t h e
point denoting culvert performance can be located
by interpolation. Example Problem 5.9 Headwater height required
c. A calculated index value greater than a particu­ for culvert with inlet control
lar dash curve index value is indicative of full pipe
flow, and full pipe flow conditions prevail and the During construction in a subdivision, a groove-edged en­
culvert capacity chart should not be used. In this trance, 42-in., 160-ft concrete culvert is placed on a 3 % slope
164 5. Hydraulics of Structures
beneath a temporary road. The design storm event yields a to the 100 cfs peak flow and extend it to read the head scale
peak runoff of 100 cfs. Find the headwater height required to value of 3.00 ft. To determine the headwater, solve
pass the flow. Eq. (5.22):
Solution: Select the appropriate capacity chart as Fig. HW = Η + h - 0 S L.
0

5A.2. The index is


Substituting
160
Index = = 53. HW = 3.0 + 3.5 - (0.0010)(160) - 6.34 ft.
100S 0 100(0.03)

Enter the capacity chart on the abscissa with the peak flow of
100 cfs. Drawing a vertical line at 100 cfs intersects the solid Example Problem 5.12 Culvert selection
line (inlet control) for the 42-in. culvert at a headwater height with headwater restriction
of 5.9 ft. The index value for inlet control given on the solid
line is 400 for the 42-in. culvert size. Since the calculated A corrugated metal pipe with a projecting entrance is to
index value (53) is less than the solid curve value (400), inlet be located under a road during construction of an airport.
control is indicated. Thus, the required headwater height The culvert length is 120 ft, and it has a slope of 0.003 ft/ft.
HW is 5.9 ft. Peak flow from the design storm is calculated to be 200 cfs.
The maximum allowable headwater is only 8 ft. Estimate the
culvert size that will be needed to satisfactorily convey the
Example Problem 5.10 Headwater height required design flow while not exceeding the allowable headwater
height.
for culvert with outlet control.
Solution: The index value is
Use the situation described in Example Problem 5.9, ex­ 120
cept assume a slope of 0.15%. Index = - 400.
1005 (100)(0.003)
Solution: The solution follows the same steps as in Exam­
ple Problem 5.9. The difference will be the impact of the The appropriate culvert capacity chart based on the pipe
index value. The index value is material, size, and index value is Fig. 5A.6a. The culvert size
160 that can convey the specified drainage without exceeding the
Index = = 1067. headwater constraint is selected by: (1) locating the 200 cfs
100S 0 100(0.0015) discharge on the abscissa, (2) extending a line upward to the
culvert size that yields a headwater value of 8 ft or greater,
The calculated index value, 1067, lies between the solid and
and (3) interpolating for the calculated index value of 400.
dash curves, 400 and 1200, for the 42-in culvert; thus outlet
The required culvert is a 72-in.-diameter circular corrugated
control is indicated. Read the required headwater height by
metal pipe having a projecting entrance.
interpolating for the index value of 1067 between 400 and
1200, and extend a horizontal line to the left axis. The
required headwater is 6.2 ft.

Trickle Tube Spillway


Trickle tube spillways are often considered to be just
Example Problem 5.11 Headwater required with full
a variation of the c o m m o n culvert that is placed in a
pipe flow
dam. In actuality, many culverts placed in roadways act
This problem is similar to the preceding example problems like a primary spillway for a reservoir. T h e major
except that the slope equals 0.10%. design difference is that trickle t u b e spillways are gen­
Solution: The index value for this case is erally considered to have steeper slopes t h a n typical
culverts. Flow at any given h e a d will b e d e p e n d e n t on
160 w h e t h e r the flow is controlled by t h e inlet acting like a
Index = = 1600.
100S 0 100(0.0010) weir or an orifice, by o p e n channel flow through the
pipe, or by full pipe discharge. Chow (1959) indicated
This value is above the dash line in Fig. 5A.2, indicating pipe that long pipes usually flow full w h e n the h e a d exceeds
full flow. Thus, the pipe full outlet control nomograph (Fig. some critical value H*. Chow r e c o m m e n d e d a value of
5B.3) must be used. Using Fig. 5B.3, assume tailwater is at
1.5 times the pipe d i a m e t e r for a reasonable value of
the crown (top) of the 42-in. culvert so that h = 3.5 ft.
H* in o r d e r to usually have full flow. Submergence of
0

Draw a line connecting the culvert diameter of 42 in. with


the culvert length of 160 ft along the groove-edged curve t h e outlet often leads to full flow. O t h e r hydraulic
(K = 0.2 in Fig. 5B.3). This located the pivot point on the characteristics also are similar to culverts. If the outlet
c

turn line. Now draw a second line connecting the pivot point of a trickle tube is not submerged, t h e t u b e may not
Hydraulics of Culverts 165

flow full even if the head exceeds the critical value H* 0.6D above the outlet invert as a reference,
and is considered to be hydraulically short. In other
cases, it is considered hydraulically long. Carter (1957) Η' = Η + 3.00 - 0.6 X {§ = Η + 2.10.
provided graphs (Figs. 5.9 and 5.10) that define whether
Hence,
a pipe is hydraulically short or long.
A detailed analysis of pipe flow control in a trickle Q = 4.957(if + 2 . 1 0 ) . 1/2

tube spillway would require the computation of back­


water curves in the pipe for a n u m b e r of discharges. If For example, with Η equal to 2 ft, Q equals 10.04 cfs. For
the trickle tube is on a mild slope ( y > y ) , flow would inlet control, data such as those shown in Fig. 5.11 can be
n c
used. The following table illustrates its use.
enter the pipe at approximately the normal d e p t h of
flow. In this case, the relationship between discharge
and head could b e obtained by computing t h e normal Calculation of discharge with inlet control for
depth of flow for a range of discharges and then Example Problem 5.13

computing the required head from QI-Jzd™


//(ft) HID (from Fig. 5.11) Q (cfs)
V2
V2

0.50 0.33 .10 1.56


1.50 1.00 .40 6.25
where K is the entrance loss coefficient and y is the
e n 2.00 1.33 .57 8.91
normal depth. This m e t h o d serves only as a very rough 3.00 2.00 .82 12.82
approximation and should be used with caution. Sensi­ 4.00 2.67 .96 15.01
tive projects should always have calculations m a d e with 5.00 3.33 1.09 17.04
detailed water surface profiles.
If the pipe is on a steep slope ( y < y ) , t h e n a 5
n c 2

flow profile would develop within t h e pipe and the To determine if open channel flow controls the discharge,
above relationship would only be approximate. It should values of y are assumed, Q, A, and V determined using
n

be noted that as the pipe slope increases, the range of open channel flow relationships given in Chapter 4 for a
entrance control also increases. Example Problem 5.13 circular conduit and calculating the required total head from
illustrates the development of a s t a g e - d i s c h a r g e rela­
tionship through a trickle tube.
H'=—(1 + K) + y.
e a

The results are shown in the following table.


Example Problem 5.13 Stage-discharge curve for a
trickle tube
Open channel flow calculations
for Example Problem 5.13
As part of a pond design, a stage-discharge curve for an
18-in. corrugated metal trickle tube must be found. The pipe Yn Q V //'
is 100 ft long, placed on a slope of 3 % with a free outfall and
a square entrance. 0.15 0.21 2.24 0.31
Solution: The discharges when pipe flow prevails can be 0.30 0.86 3.43 0.66
found using Eq. (5.5), or
0.45 1.93 4.33 1.03
0.60 3.32 5.03 1.38
0.75 4.93 5.58 1.72
yJ\+K +K
t b + KL C
0.90 6.62 5.98 2.01
For a corrugated metal pipe, η = 0.024 (Table 4.1). From 1.05 8.25 6.24 2.26
Table 5.1, K is 0.0621 for an 18-in. pipe with η = 0.024.
c 1.20 9.63 6.36 2.45
Table 5.1 also indicates that the area is 1.770 ft . Assuming 2

1.35 10.50 6.27 2.57


K = 1 and K = 0 results in
e b
1.50 9.86 5.58 2.47

1.770^2 X 32.2 Χ Η'


= 4.957νΪΓ.
y/l + 1.0 + (0.0621)(100) The discharge for a given head is the minimum value
obtained for any one of the different controls. Values are
For a 100-ft pipe on a 3 % slope, the change in elevation tabulated in the following table for pipe flow control, inlet
between the upper and lower invert is 3.0 ft. Using a point control, or open channel control.
166 5. Hydraulics of Structures
Comparison of pipeflow, inlet, or channel flow for Example Problem 5.13

Discharges (cfs)

Head Type pipe Pipe Inlet Open Design


(ft) 29n H/R
2 4/3
(Fig. 5.9) flow* control'' channel 1
discharge (cfs)

0.50 — — — 1.56 0.56 0.56


1.50 — — — 6.25 3.89 3.89
2.00 0.123 Long 10.04 8.91 6.62 6.62
3.00 0.185 Long 11.22 12.82 — 11.22
4.00 0.247 Long 12.27 15.01 — 12.27
5.00 0.309 Long 13.23 17.04 — 13.23

g = 4.957(// + 2 . 1 0 ) .
fl 1/2

^Discharge with inlet control calculations.


Interpolated from open channel flow calculations.

Plotting the design discharge as a function of head will grade a n d / o r road material. Typical spacings range
yield the stage-discharge curve. In this example, flow is from 100 ft for steep slopes ( > 10%) to 300 ft for mild
controlled by either open channel flow or pipe flow, depend­
slopes ( < 5%).
ing on the head. If the slope were substantially increased,
flow would have inlet control. Recall that if the trickle tube is
on a hydraulically steep slope, the open channel flow calcula­ Downdrain: Function, Use, Type, and Design
tions are only approximate. Consider the data in open chan­ Downdrains provide erosion protection while con­
nel flow calculations table and a depth of 0.75 ft. The Froude
veying concentrated runoff from o n e point on a slope
number can be determined from Eq. (4.11) as
to another. Riprap-lined channels are limited to gentle
slopes. S t e e p slopes require use of downdrains to route
F =
water collected by diversions and ditch relief culverts
safely downstream.
For y = 0.75 ft and
n D= 1.5 ft, Downdrains, such as those used on surface-mined
lands and construction roads, can consist of chutes,
A TTD /8
2
TTD
= 0.589 flumes, and half-round rigid and flexible pipes. They
Τ D 8 may also b e constructed of geotextile materials such as
F = 5.58 = 1.28. fiberglass, excelsior matting, j u t e mesh, and plastic
# 2 . 2 X 0.589" sheeting. T h e s e geotextile materials should not b e used
on slopes greater 1 5 % according to F e d e r a l Highway
The flow is supercritical, and thus the open channel calcula­
tions are only approximate. Similar calculations can be made Administration (1975).
for other flow depths. Two general types of downdrains can b e installed,
d e p e n d i n g on t h e flow requirements and availability of
the materials. Sectional downdrains are prefabricated
sectional conduits constructed of half- or third-round
pipe, corrugated metal, bituminized fiber, or other
Ditch Relief Culvert materials. Flexible downdrains are less p e r m a n e n t
In mining, construction, and similar land-disturbing structures that are constructed of heavy-duty fabric, or
activities, it is often desirable to divert flow away from o t h e r materials. Both sectional and flexible downdrains
a road ditch so that it will not attain sufficient volume, are supported by the surface profile and are con­
velocity, or depth to erode the ditch. Ditch relief structed to carry the design storm as o p e n channel flow
culverts perform this function by collecting runoff in a using the o p e n channel design techniques.
culvert constructed across the road at a downgrade Corrugated metal pipe and bituminized fiber pipe
angle (typically about 30°). T o further protect down­ are used as m o r e p e r m a n e n t structures and are placed
stream areas, flow exiting a ditch relief culvert should in cut a n d fill slopes. They are then covered much as a
be conveyed down the fill slope to a sediment t r a p or spillway in a d a m might be. T h e lead section of a
stabilized area by a downdrain instead of releasing it downdrain is normally prefabricated and placed in
on unprotected areas. compacted fill protected by rock riprap with a sand
T h e specific interval required between ditch relief bedding. T h e pipe slope drain is covered with approxi­
culverts is normally regulated as a function of road mately 1 ft of fill or spoil material. A n important
Hydraulics of Emergency Spillways 167

DIMEN­ SIZE FIGURE


SION A Β
Η l.5 #
2.0*
d 8" ιο·
L 5* 6'

ENERGY DISSIPATOR IS
CONCRETE BUILDING BLOCKS
| — L —»|0N EDGE ANCHORED TO LINING

RIPRAP IS LAYER OF
ROCKS OR RUBBLE

L
]*'MIN I
TOE OF SLOPE
P L A N VIEW
Figure 5.15 Example of a chute design (adapted from Soil Conservation Service, 1969).

component is the outlet protection consisting Of sized t h e downstream e d g e of t h e flat crest. This point
rock riprap or a stilling basin and an energy dissipator. locates t h e control section since the velocity and depth
Such a pipe downdrain system has design r e q u i r e m e n t s of flow are defined. If t h e exit channel slope is less
similar to those considered previously u n d e r trickle t h a n the critical slope, n o control section exists. Emer­
tubes and culverts. gency spillways can b e constructed b o t h with and with­
O t h e r downdrains function like a chute or flume to out control.
provide conveyance in a high-velocity, o p e n channel T h e hydraulics of broad-crested spillways with a
suitable for carrying water to a lower elevation without control section can b e analyzed with the aid of several
erosion. Chutes or flumes are typically constructed of n o m o g r a p h s a n d simple equations, which are pre­
concrete or comparable material using either formed sented in this section. T h e hydraulics of broad-crested
or freeformed methods. A schematic of a p e r m a n e n t spillways without control sections are not amenable to
chute structure is given in Fig. 5.15. Basic c o m p o n e n t s simple solution a n d will not b e covered.
are similar to those of a pipe d r o p downdrain. A chute
can b e designed using either t h e s t a n d a r d o p e n chan­ Evaluation of Head-Discharge Relations for an
nel procedures described previously, or it can be sized Emergency Spillway
as a function of drainage area.
T h e Soil Conservation Service (1968) developed a
m e t h o d of analysis for broad-crested spillways, such as
HYDRAULICS OF EMERGENCY SPILLWAYS emergency spillways. T h e analysis assumes that control
exists at t h e downstream edge of the crest as summa­
Emergency spillways are typically flat across the top rized by Barfield et al. (1981). Presented are analysis
with sloping inlet and exit sections as shown in Fig. p r o c e d u r e s that can b e used to:
5.16. If the exit channel slope is g r e a t e r than the (1) define the discharge in the spillway for a given
critical slope, then flow passes through critical d e p t h at value of h e a d , H , in the reservoir, or
p
168 5. Hydraulics of Structures

Figure 5.16 Illustration of a broad-crested spillway with a control section.

(2) define t h e required spillway dimension for a given for a rectangular cross section. T h u s Q is uniquely r

discharge. related t o H . T o define Q for a given value of H


cc r T p9

Relationship Between Discharge a n d H it is necessary t o relate H a n d H . ec p


p

T h e relationship between Q a n d H for a trape­ ec

W h e n flow occurs across t h e crest of a spillway with zoidal section is m o r e complex t h a n for a rectangular
a control, as shown in Fig. 5.16, critical d e p t h a n d channel. T o develop t h e relationship, it is necessary t o
critical velocity occur at t h e control section. T h e sum r e t u r n to t h e energy principles discussed in C h a p t e r 4.
of t h e velocity head, V /2g, 2
and d e p t h of flow, y , at c
T h e total energy h e a d for any flow is given by
this point is total critical head, H , or ec

// e = £ = K /2* + y 2
(5.29)
H ec =y + c V /2g 2
(5.23)

based on Eq. (4.9), where t h e subscript c refers to or in terms of discharge Q a n d cross-sectional area A,
critical flow and g is t h e gravitational constant. F o r a
rectangular cross section, t h e critical d e p t h is given in E = Q /2gA 2 2
+ y. (5.30)
terms of discharge as
Ε can b e differentiated with respect t o y t o obtain
.1/3
y ,r = (v /g)
2 (5.24)
c dE _ Q 2
dA
(5.31)
where q is discharge p e r unit width a n d t h e subscript r dy gA 3
dy
refers to a rectangular channel. F o r a channel of bot­
tom width b and discharge Q, y becomes ct A t critical d e p t h , dE/dy is equal to zero; hence,

1/3
Q 2
A c

y ,r = (5.25) (5.32)
c
bg
2
gA\ dA/dy'

Combining Eqs. (5.23) and (4.10), it can b e shown that w h e r e A is t h e cross-sectional area at critical d e p t h .
c

F r o m Fig. 4.25, dA/dy is equal to t h e t o p width; T;


" e c . r = bc.f (5-26) hence, at critical d e p t h ,
Equation (5.25) can be solved for Q in terms of T Htcx

and b using Eq. (5.26) to yield Q2

(5.33)
2gA 271
Qr = CiY s H X b
/2 l/2
e
2
t
(5.27)
where T is t h e t o p width at critical d e p t h . Relation­
c

= Q.5Mg Hl .b x/2 /2
(5.28) ships given in C h a p t e r 4 can b e used for A a n d 7^ t o c
Hydraulics of Emergency Spillways 169

yield necessary to calculate Q', the discharge in a 100-ft rectangu­


lar spillway with H = 5.0 ft. From Eq. (5.28),
ecr

g 2
Vc
2
(b+zy )y c<z CtZ

Q' = 0.544g /\H , Y b, l /2

2g(A )
Ct2
2
2g 2(b + 2zy , ) c z ' (
' > cc r

which after substituting values becomes


where the subscript ζ refers to the side slope. Defining
Hcc and Ε at critical depth, Eq. (5.34) can b e used in
z
Q' = (0.544)(32.2 f t / s e c ) 2 1 / 2
(5 ft) 3 / 2
(100 ft)
Eq. (5.30) to obtain
= 3451 f t / s e c .3

(3b + 5zy )y c c From Eq. (5.38),


#ec,z - (5.35)
2b + 4 z y c
1.5* +zH„ (1.5)(100) + (3)(5)
= (3451 cfs)
Equations (5.34) and (5.35) can b e solved simultane­ 150 150
ously to obtain a Q-H relationship for a trapezoidal
cc Q = 3796 cfs.
channel. Unfortunately, t h e solution is implicit.
The percentage error in the estimate of Q can be determined
A simple approximation can b e m a d e to t h e Q — H ec
from Fig. 5.17, or from Eq. (5.38). From Fig. 5.17,
problem based on the premise that the ratio of dis­
charges of two trapezoidal channels with equal critical % Error = 0.13%.
specific energy heads is equal to the ratio of their Using Eqs. (5.38), (5.39a), and (5.39b),
average widths (Soil Conservation Service, 1968), or
zH„ (3)(5ft)
Q b+zy G = = 0.15
c
b 100 ft
(5.36)
Q b' + z'y' c
L = 1.1 + [ l o g ( l / G ) ] [ 0 . 5 9 - 0 . 1 3 6 l o g ( l / G ) ]
10 10

W h e n the approximation is based on a rectangular = 1.1 + [ l o g ( l / 0 . 1 5 ) ] [ 0 . 5 9


10

channel ( ζ ' = 0) of width, b' = 100, and using y = c -0.136 l o g ( l / 0 . 1 5 ) ] = 1.49


10

2 / 3 / / , then Eq. (5.36) can be written


e c

% Error = 2 . 2 7 [ G ] = 2.27J0.15]L 14 9
- 0.13%.
Q_ l.5b +zHe c , ζ
t

(5.37) The corrected value of Q is


Q' 150
3796
(2 = = 3801 cfs.
T h e approximate error involved in Eq. (5.37) can be 1 - %e/100 1 - 0.13/100
determined from Fig. 5.17 or from the approximation
Although the first estimate is adequate, this will not always
be the case.
% E r r o r = 2.27 [ G ] , L
(5.38)
where
G = zH /b ccz (5.39a) T h e problem now remains to relate H to H in cc p

and o r d e r to develop a Q-H relationship. T o do so re­


p

quires t h e development of backwater profiles. By com­


L = 1.1 + [ l o g ( l / G ) ] [ 0 . 5 9 - 0.136 log ( 1 / G ) ] .
1 0 1 0 puting n u m e r o u s profiles, the Soil Conservation
(5.39b) Service (1968) developed nomographs of H versus p

Hec for a standard reference section of b = 100 ft,


Example Problem 5.14 illustrates the use of the proce­ ζ = 2, and η = 0.04 for the nine types of spillways
dure for calculating Q for a given value of H . ec
shown in Fig. 5.18. T h e H versus H curves are
p e c

shown in Figs. 5 D . 1 - 5 D . 9 located in the appendix to


this chapter.
Example Problem 5.14 Discharge for a T h e relationship between H p and H ec can be writ­
broad-crested spillway ten 2

Hp = H ec + h, { (5.40)
It has been determined that the total energy head at the
control section of a broad-crested trapezoidal spillway is 5.0 where h f is t h e friction h e a d loss. T h e inaccuracies in
ft. The side slopes are 3 : 1 , and the bottom width is 100 ft.
Deteremine the discharge in the spillway.
Hec is used here without superscripts r or ζ to denote a more
Solution: To use Eq. (5.37) to calculate Q, it is first general case.
170 5. Hydraulics of Structures

0.3 0.4 0.5 0.6 0.8 I 2 3 4 5 6 7 θ 9 I0 20


Critical Specific Energy Head, H 6 C f ft
Figure 5.17 Error in Qc where Qc is determined by the approximate relation Q /Q'
c c = (1.56 + z / / ) / 1 5 0 . (Soil
e c

Conservation Service, 1968)


Hydraulics of Emergency Spillways 171

S «0.00
0

/
Reservoir · Control Reservoir Control
Water Section Water
0
Section
Surface Surface

Control Reservoir
Control
Section *
W o t f

g Water
Surface Section 5

Reservoir
Control Reservoir
- Wote Section Water
' Surface Surface Control 3
Section

Reservoir -
*% Woter
57 Surface

I s 0 ·ο·οο I I I

Figure 5.18 Cases 1 - 9 used by the Soil Conservation Service w h e n developing curves of HD versus Hei
172 5. Hydraulics of Structures

Figs. 5 D . 1 - 5 D . 9 result from inaccuracies in h . T h e s e f t h e minimum slope that can b e used to g u a r a n t e e a


inaccuracies are given for selected cases in Figs. control section ( S > S ) for a discharge equal to β / 4 .
0 c

5 E . 1 - 5 E . 5 . Corrections in these figures are based on a W h e n designing an exit channel, it is important to


standard condition where b is 100 ft, ζ is 2, and η is select a slope steep enough to ensure a control section
0.04. Obviously, the only factor of major importance is over the flow range from β / 4 to β .
the channel roughness n. By using Figs. 5 D . 1 - 5 D . 9 T h e p r o c e d u r e for selecting channel dimensions is
and the corrections in Figs. 5 E . 1 - 5 E . 5 , the relationship best illustrated by example. Details are given in Exam­
between H and H is established as d e m o n s t r a t e d in
ec p ple Problem 5.16. T h e bottom width in Example Prob­
Example Problem 5.15. lem 5.16 is actually quite large. It could b e reduced by
increasing t h e permissible velocity in the exit channel
or by decreasing the exit slope.
Example Problem 5.15 Head in reservoir
with emergency spillway
Example Problem 5.16 Emergency spillway design
Determine the value for H in Example Problem 5.14 if
p for sediment pond
the total length of the spillway is 300 ft, the spillway is
grassed so n = 0.04, and the section is constructed like An emergency spillway is being designed for a construction
Case 1. site sediment pond. The design discharge is 600 cfs. The exit
Solution: Since b = 100, Fig. 5D.1 can be used without slope is to be grass lined with a Manning's η of 0.04 and a
making corrections for channel width. For H of 5.0 and permissible velocity of 6.0 ft/sec. Machinery limitations re­
ec

L = 300 ft, the value of H from Fig. 5D.1 is 6.3 ft. quire that 3 : 1 side slopes are the steepest that can be used.
p

The effects of side slope, z, can be determined from Fig. The rest of the spillway will be grassed with an η of 0.04. The
5E.4. From Fig. 5E.4 for a side slope of 4 : 1 , the ratio of Case 1 cross section in Fig. 5.18 will be used. Calculate the
head loss at ζ = 4 to that of ζ = 2 is value of H needed to convey the flow, and size the emer­
p

gency spillway. The spillway length, L, is 100 ft. The exit


slope, 5 , is 0.03 ft/ft.
tf, /A, -0.994;
iZ i2
0

Solution: From Fig. 5E.6, the maximum permissible value


hence, corrections for ζ are not required for a 4 : 1 side for H with S = 0.04 ft/ft and V - 6.0 ft/sec is
cc 0 p

slope. Since a 3 : 1 side slope is closer to a rectangular section


than a 4 : 1 side slope, the ratio of h to h would be even Hec max = 1.28 ft.
{ z { 2

closer to 1.0.
From Fig. 5D.1, H for H p ec of 1.28 ft, L of 100 ft, and a
Manning's η of 0.04 is

" , o . 0 4 = 1.75
P ft.
Spillway Dimensions for a Given Discharge
Since η is 0.04 on the spillway crest, correction is not neces­
W h e n an emergency spillway is being selected, it is
sary. The friction head loss is therefore
necessary to determine the design dimensions n e e d e d
to transmit some expected rare event. This requires Kom = 1-75 - 1.28 = 0.47 ft.
maximum values for H and H b e established based
p cc

on channel stability in the exit section of the spillway If η had not been equal to 0.04, the friction head loss would
(shown in Fig. 5.16). T h e maximum value of H is ec
be computed using Fig. 5E.1.
selected to ensure that the velocity in the exit section Corrections must also be considered for the side slopes
since ζ is equal to 3 instead of 2. By interpolating between
does not exceed the permissible velocity. T h e p a r a m e ­
lines for ζ = 2 and ζ = 4 on Fig. 5E.4,
ter to be used will be the limiting velocity. This re­
quires the development of backwater profiles just as in
the case of Q-H relationships. Again on the basis of
f,2
p h

numerous computed water surface profiles, the Soil


Conservation Service has developed n o m o g r a p h s that Therefore, the effects of ζ variations or h are not significant {

relate the maximum permissible H to permissible ec and


velocity and channel slope. T h e s e n o m o g r a p h s are
shown in the appendix to this chapter as Figs. 5E.6 and
5E.7 for Manning's η of 0.04 and 0.02, respectively. = 1.28 + 0.47
Also shown in Figs. 5E.6 and 5E.7 are curves that show = 1.75.
Hydraulics of Emergency Spillways 173

The bottom width must now be determined by using T h e solution p r o c e d u r e involves t h e following steps:
Eq. (5.38). Using Eq. (5.28), the discharge for a standard
100-ft rectangular channel with H of 1.28 ft would becc n (1) Calculate H for each value of H .
ec p

(2) Calculate Q from t h e Q-H ec relationship


Q' - 0S44g ' H*gb
l 2

[Eq. (5.28)] for rectangular sections.


(3) Correct for a trapezoidal section [Eq. (5.37)].
= (0.544)(32.2 f t / s e c ) ( 1 . 2 8 f t ) ( 1 0 0 ft)
1/2 3/2

(4) A d d t o t h e Q for t h e principal spillway.


= 447 cfs.
T h e s e calculations a r e illustrated in Example Prob­
From Eq. (5.38) lem 5.17.

1.56 + zH ec 2

* * 150

substituting values and solving for b Example Problem 5.17 Stage-discharge curve for
emergency spillways
1.5b + (3)(1.28)
600 = 447 - Develop a stage-discharge curve for the spillway in Exam­
L 1 5 0
ple Problem 5.16 if the side slopes are 3 : 1 and the η value is
b = 131 ft. 0.04 on the crest. Assuming that the spillway crest is 12 ft
above the principal spillway inlet and that the pipe used in
Use 130 ft. From Fig. 5.17, the error involved in using Example Problem 5.4 is the principal spillway, calculate the
Eq. (5.38) is approximately 0.01%; hence, the bottom width is total discharge from the reservoir.
130 ft. Solution: From Example Problem 5.16, the spillway is
trapezoidal with the following characteristics:

b = 130 ft
Stage-Discharge Curve for Emergency Spillways
L = 100 ft
T h e relationships p r e s e n t e d in this section can b e
ζ =3
used to develop a s t a g e - d i s c h a r g e curve for a broad-
crested emergency spillway. This can t h e n b e a d d e d to η = 0.04.
the discharge through t h e principal spillway to deter­
mine the total discharge. Calculations are summarized in the following table.

Head above crest Discharge in Discharge in


of emergency Total 100 ft rectangular trapezoidal Head above Discharge in
spillway 0
energy head^ spillway c
spillway** principal principal Total
"ec & Q spillway' spillway^ discharge*
(ft) (ft) (cfs) (cfs) (ft) (cfs) (cfs)

0 0 0 0 12.00 26.6 27
0.75 0.35 83 108 12.75 26.9 135
1.0 0.61 191 250 13.00 27.3 277
1.5 1.04 426 562 13.50 27.6 590
2.0 1.45 700 930 14.00 27.9 958

Assumed.
e

Figure5D.l.
fe

Equation (5.28).
c

^Equation (5.37).
H+
e
p 12 ft.
^From equations in Example Problem 5.4 (Q = 5 . 8 ( / / + 9 . 1 ) ) or read from stage-discharge curve.
1/2

^Columns 4 + 6. (Discharge in trapezoidal spillway plus discharge in primary spillway.)


174 5. Hydraulics of Structures

CULVERT OUTLET PROTECTION Table 5.5 Coefficients for Scour Prediction (FHA, 1975)

Coefficients
S c o u r Hole G e o m e t r y
hole dimension α β θ y
Erosion resulting from the discharge from a culvert
or chute o n t o an unprotected, erodible material will Depth (h ) s

form a hole or depression known as a scour hole. A n TW < 0.5D 0.82 0.375 0.10 1.0
estimation of scour hole size can best be d e t e r m i n e d
TW>0.5D 0.76 0.375 0.10 1.0
through inspection of a similar existing structure lo­
cated in a comparable soil environment. If this is not
Width (W )
possible, a p r o c e d u r e that allows a prediction of t h e s

TW<0.5D 0.55 0.915 0.15 1.0


depth, width, length, and volume of a scour hole as­
suming a sandbed culvert discharge area is p r e s e n t e d TW > 0.5D 0.39 0.915 0.15 1.0

in this section. T h e prediction of scour hole geometry


is indicative of t h e erosion hazard and n e e d for soil Length (L ) s

protection. T h e Corps of Engineers conducted studies TW < 0.5D 1.67 0.71 0.125 1.0
that indicated that scour hole geometry was related to TW>0.5D 2.85 0.71 0.125 1.0
tailwater conditions (Fletcher and G r a c e , 1972). They
defined two tailwater categories: Volume (v ) s

TW < 0.5D 0.29 2.0 0.375 3.0


(i) TW < \D, (5.41) TW > 0.5D 0.24 2.0 0.375 3.0

(2) TW > ±D, (5.42)

where T W is tailwater d e p t h and D is t h e culvert


diameter.
T h e prediction equation n e e d e d to d e t e r m i n e scour Example Problem 5.18 Scour hole geometry
hole geometry is ( F H A , 1975)
Determine the scour hole geometry caused by the dis­
charge of a long 36-in. concrete circular culvert on a 0.5%
Scour geometry = < £ ( D ) ( Q / D ) V .
r 2 5
(5.43)
slope discharging at 60 cfs.
Solution: Assume 20 min as the time duration t for the
Replacing the diameter, D , by an equivalent d e p t h
purposes of this example only. This t represents the time
duration of the maximum peak discharge from a design
y = (A/2) ,
e
05
(5.44) storm.
The next step normally will be to calculate normal depth
(TW) using the procedure presented in Example Problem
where A is an area with a width that is twice t h e d e p t h
4.4. For this example assume TW = 1.1 ft.
of flow, generalizes Eq. (5.43) to any culvert s h a p e so Determine brink depth, y , using Fig. 5.20 for circular
0

that culverts to calculate

Q_ 60 TW 1.1
a( y Y(Q/y ' )V,
2 5
Scour geometry = e (5.45) = 3.85 and — = — - 0.37.
D25
3^ D 3
where y a s equivalent d e p t h in feet, / is time of peak
e Then the brink depth can be found using Fig. 5.20,
flow duration in minutes, and α, γ , β, and θ are
coefficients that are d e p e n d e n t u p o n the desired pa­ ^ = 0.65 Λ y = 0.65(3) = 1.95 ft.
0

rameter: length, width, depth, or volume of scour, as


listed in Table 5.5. If the time of p e a k flow duration of
Determine area of flow using Eq. (4.24) as A = 4.86 ft . 2

the design storm is unknown, a maximum time of 30 Calculate the equivalent depth for nonrectangular culverts as
min should be used in the above equation ( F H A , 1975).
Dimensionless rating curves for outlets of rectangular y = (A/2)
e
1/2
= (4.86/2) 1 7 2
= 1.56 ft.
culverts (Fig. 5.19) or circular culverts (Fig. 5.20) sim­
plify the calculation of brink d e p t h as illustrated in the Noting that TW is 1.1 ft, which is less than 0.5D or 1.5 ft, the
following example. appropriate coefficients are selected from Table 5.5.
Culvert Outlet Protection 175

TW/D
Figure 5.19 Dimensionless rating curves for outlets of rectangular culverts on horizontal and mild slopes (after
Federal Highway Administration, 1975).

Calculate depth of scour using Eq. (5.45) and Table 5.5. Calculate volume of scour from

Scour depth - ay (G/y ' ) V


r
e
2 5

Scour volume = 0 . 2 9 ( 1 . 5 6 ) ( 6 0 / ( 1 . 5 6 ) ) ( 2 0 ) °
30 25 ZU

= 0.82(1.56) (60/(1.56) )° (20)


10 25 375 0 1 0

= 1319 ft . 3

= 5.3 ft.

Calculate width of scour using the equation


0.915 0.15
Scour width = 0 . 5 5 ( 1 . 5 6 ) ( 6 0 / ( 1 . 5 6 ) )
10 25
(20)' Energy dissipators — Rock Riprap Aprons
= 21 ft.
T h e design for rock riprap energy dissipators is based
Calculate length of scour from on research conducted at Colorado State University
( F H A , 1975). T h e conclusion drawn from this study is
Scour length = 1.67(1.56) ° ( 6 0 / ( 1 . 5 6 ) ) ° ( 2 0 ) °
1 25 71 1 2 5

that scour hole geometry is related to riprap size (d50),


= 32 ft. discharge (QX the flow depth at the discharge point
176 5. Hydraulics of Structures

TW/D
Figure 5.20 Dimensionless rating curves for outlets of circular culverts on horizontal and mild slopes (after Federal
Highway Administration, 1975).

known as the brink depth ( y ) , and tailwater depth


0
energy dissipator pool length is 10 A or 3 W , whichever
s 0

(TW). Also, it was found that a riprapped scour hole is greater, and the overall basin length is 15 A or 4 W , s 0

functioned very efficiently as an energy dissipator if the whichever is greater. W is equal to either the diame­
Q

ratio of tailwater to brink depth was less than 0.75. For ter for a pipe culvert or the barrel width for a box
greater ratios the high velocity discharge core passed culvert.
through the basin creating a shallower but longer scour A n alternative design p r o c e d u r e for a rock riprap
hole. Thus, the downstream channel required rock pipe outlet is to use an a p r o n as shown in Figs. 5.22
riprap lining. and 5.23. T h e design p r o c e d u r e relates riprap size and
Rock riprap basin geometry is shown in Fig. 5.21. apron dimensions to culvert discharge, pipe diameter,
T h e basin is excavated and lined with riprap. T h e and tailwater conditions (Environmental Protection
surface of the riprapped floor is constructed at a d e p t h Agency, 1976). For tailwater d e p t h less or greater t h a n
h below the culvert exit. T h e ratio of scour hole d e p t h
s
the culvert centerline, Figs. 5.24 or 5.25, respectively,
to d 50should be greater than 2 and less than 4. T h e should b e used. T h e design curves are based on circu-
Culvert Outlet Protection 177

N01C A- WMU» tASM TO CONFORM TO NATURAL STREAM CHANNEL TOP Of ftP*AJ> Ν


FLOOR OF IASM SHOULD Κ AT THE SAME ELEVATION OX LONER THAN NATURAL
CHANNEL tOTTOM AT SEC A-A.

TOR QT NATURAL

. EXCAVATE TO THtS UN€.


BACKFIL VRTH RIPRAP

8ERM IS REQUIRED
TO SUPPORT WPRAP

2d.fi OR 1.M MAX.

" SEC. 0 - 0

Figure 5.21 Schematic of riprapped culvert energy basin (adapted from Federal Highway Administration, 1975).

lar conduits flowing full. T h e design p r o c e d u r e is illus­ Therefore, the energy dissipator pool length is 20.3 ft. The
trated in the following examples. overall basin length = max[15/i ,W ] = max[ 15(2.03), 4(3)]
s 0

= 30.5 ft.

Example Problem 5.19 Rock riprap selection


for control of scour
Example Problem 5.20 Rock riprap apron
below culvert
Determine the rock riprap requirements for the conditions
given in Example Problem 5.18. A circular 36-in. concrete conduit flowing full discharges
60 cfs. The tailwater, calculated by Manning's equation, is 1.1
Solution: The tailwater and brink depth from Example ft. Determine the median rock diameter by weight, d , and 5Q

Problem 5.18 are 1.1 and 1.95 ft, respectively. Thus T W / the appropriate length and widths of the rock riprap apron.
y < 0.75. The equivalent depth, y is 1.56 ft and flow area,
0 e Solution: Calculate the apron width using either
A, is 4.86 ft . Assume rock with a d equal to 0.55 ft is
2
50

available. W =D + L
a a (TW < 0.5D)
Compute V - Q/A = 60/4.86 - 12.3 fps. Compute
Q
\¥ = D + 0.4L (TW > 0.5D).
Froude Number, F: Λ a

Use the low tailwater equation since


12.3
F = = 1.74. TW = 1.1 < 0.5D = 1.5.
V^T ^32.2 X 1.56
Hence, Fig. 5.24 is applicable. From Fig. 5.24, read both
median rock diameter and minimum apron length by drawing
Since d = 0.55 ft, d /y
50 = 0.35. From Fig. 5.23, / i / y =
50 e s e a vertical line from the abscissa at Q = 60 cfs to the pipe
1.3 so h = (1.3X1.56) = 2.03. Check that 2 < h /d
s < 4; s 50 diameter, D = 36 in. Read both the median rock diameter
just barely OK! Depth of scour, h = 2.03 ft. Length of s and minimum apron length from the appropriate ordinate
energy dissipation pool is the maximum of 10h or 3W or s 0 axes. Thus d = 0.6 ft and L = 21 ft. Therefore the apron
50 a

width using the low TW equation is


max[10h ,3W ] s 0 = max[(10)(2.03),(3)(3)]. \Υ = D + L = 3 + 21 = 24 ft
Λ a
178 5. Hydraulics of Structures

TAILWATER FLOW
JZ

SLOPE=0

SECTION
Figure 5.22 Rock riprap apron (adapted from Environmental Protection Agency, 1976).

and will always occur prior to equilibrium conditions. Com­


putation of tailwater is difficult because of changes in
d mM = 1.5 X d*« = 1.5 X 0.6 = 0.90 ft. channel cross section d u e to flood events, seasonal
vegetal changes, and the inherent difficulty of measur­
ing channel properties in poorly defined e p h e m e r a l
streams. F o r these reasons, calculated designs should
consider the worst possible tailwater condition, i.e., a
In sizing rock riprap energy dissipators used at con­ low tailwater d e p t h .
duit outlets, the relative tailwater d e p t h is a significant M e d i a n rock diameter, d , is predicted in t h e exam­
50

factor in design. Maximum scour d e p t h is associated ple problems. A standard practice is to p r e p a r e a


with low tailwater conditions, whereas maximum scour well-graded mixture down to the 1-in. size particle such
hole length is associated with high tailwater depths. that 5 0 % of the mixture by weight shall be larger than
For ephemeral flashy streams, tailwater d e p t h will lag the d size ( E P A , 1976). R i p r a p size distribution crite­
50

significantly during stream rise. H e n c e , low tailwater ria are presented in C h a p t e r 4.


Culvert Outlet Protection 179

FROUDE NUMBER VG23K*)


Figure 5.23 Relative depth of scour hole versus Froude number at brink of culvert with relative size of
riprap as a third variable (Federal Highway Administration, 1975).

Problems Estimate the pipe discharge if pipe flow controls and


t h e w a t e r level is 4 ft above the top of the riser.
(5.1) A 36-in. circular C M P riser serves as a vertical (5.4) For the spillway indicated in problems (5.2)
inlet for the principal spillway of a stormwater deten­ and (5.3), plot a s t a g e - d i s c h a r g e curve. Indicate on the
tion structure. Estimate the discharge if the head on curve t h e stages at which each type of control (weir,
the riser is 1.75 ft. orifice, or pipe) occurs. D o all t h r e e control types
(5.2) A primary spillway for a detention pond con­ occur?
sists of a 30-in. corrugated metal pipe connected to a (5.5) A large stormwater detention structure is to be
48-in. vertical riser. Calculate the head required to constructed using a 12-ft-diameter C M P for the riser
pass a discharge of 80 cfs if weir flow controls. W h a t if and an 8-ft C M P for the barrel. T h e difference in
orifice flow controls? elevation b e t w e e n t h e riser opening and the bottom of
(5.3) T h e spillway in the previous problem is 80 ft the downstream e n d of the pipe is 28 ft. T h e pipe is
long and has o n e 90° bend where the top of the 48-in. approximately 160 ft long, (a) W h a t flow occurs through
riser joins the 30-in. pipe; the riser is 12 ft above the the spillway for water levels that are 1, 2, 4, 6, and 8 ft
bottom of the outlet. Assume that there is n o tailwater. above the riser? (b) If four 1-ft-diameter holes are cut
180 5. Hydraulics of Structures

DISCHARGE. CFS
Figure 5.24 Design of outlet protection—minimum tailwater condition, T w < 0.5D (Environmental Protection
Agency, 1976).

2 3 4 6 8 10 20 30 40 60 80 100 200 300 400 600 1000


DISCHARGE, CFS
Figure 5.25 Design of outlet protection—maximum tailwater condition, T w ^ 0.5D (Environmental Protection
Agency, 1976).

into the riser 3 ft below its top, what discharge will pass (5.6) A gravel roadway is constructed in a low-lying
through the four holes with the water level at 1, 2, 4, area such that the roadway is frequently overtopped as
and 8 ft above the riser? (c) W h a t is the total discharge a result of severe storms. T h e roadway is 40 ft wide,
through the pipe? (d) How might the orifices be sized and its elevation is 36 ft. (a) If t h e water level upstream
to provide better stormwater control? (e) Explain of the roadway is 2 ft above the crest of the roadway,
whether you would expect two rows (each consisting of what is the discharge across t h e roadway? (b) If the
four holes) of 8-in.-diameter holes to provide better roadway is paved, what u p s t r e a m d e p t h would be re­
results? Assume that one row is 2 ft below the riser quired to carry the same flow? (c) Would paving re­
invert and the other row is 4 ft below the riser invert. duce flooding problems?
Culvert Outlet Protection 181

(5.7) Compute the head loss in a 200-ft section of U.S. Geological Survey, Circular 376. U.S. Government Printing
2 X 2-ft concrete pipe flowing full and discharging Office, Washington, D C .
Chow, V. T. (1959). "Open Channel Hydraulics." McGraw-Hill,
40 cfs. Assume a roughness coefficient of 0.015.
N e w York.
(5.8) Calculate the discharge in a 300 ft, 24-in.-diam-
Environmental Protection Agency (1976). "Erosion and Sediment
eter pipe with η equal to 0.014. T h e pipe is flowing Control Surface Mining in the Eastern U.S.," Vol. 2, "Design."
full, and head on the pipe is 7 ft. EPA-625/3-76-006.
(5.9) Plot a stage-discharge curve for a trickle tube Federal Highway Administration (1975). Hydraulic design of energy
spillway (CMP) that has 24 in. diameter. T h e pipe is 50 dissipators for culverts and channels, Hydraulic engineering circu­
lar N o . 14. Federal Highway Administration, Washington, D C .
ft long and is located on a 2 % slope. It has a free
Federal Highway Administration (1985). Hydraulic design of highway
outfall.
culverts, Hydraulic design series No. 5, Report N o . FHWA-IP-85-
(5.10) Determine the critical d e p t h and critical head 15. Federal Highway Administration, Washington, D C .
for a 5000 cfs discharge across an emergency spillway. Fletcher, B. P., and Grace, J. L., Jr. (1972). Practical guidance for
T h e spillway is rectangular with a bottom width of estimating and controlling erosion at culvert outlets, Corps of
100 ft. Engineers Research Report H-72-5. Waterways Experiment Sta­
tion, Vicksburg, MS.
(5.11) T h e emergency spillway in the previous prob­
French, R. H. (1985). "Open-Channel Hydraulics." McGraw-Hill,
lem is to be compared with a broad-crested, trape­ N e w York.
zoidal spillway having 4 : 1 sides slopes and bottom Grant, D . A. (1978). "Open Channel Flow Measurement Handbook,"
width of 150 ft. If total energy head at the control 1st ed. Instrument Specialties Co., Lincoln, N E .
section is 4 ft, what is the discharge in the spillway? Herrera, Ν. M. (1989). Defining the hydraulics of flow through a
(5.12) D e t e r m i n e the total energy head H in prob­
p
rockfill of varying gradation using sediment-free water, Unpub­
lished Ph.D. dissertation. University of Kentucky, Lexington, KY.
lem (5.10) if total length of the spillway is 240 ft and
Herrera, Ν . M., and Felton, G. K. (1991). Hydraulics of flow through
constructed as in Case 2. T h e spillway is grassed with
a rockfill dam using sediment-free water. Trans. Am. Soc. Agric.
tall fescue and has 3 : 1 side slopes. Eng. 34(3): 8 7 1 - 8 7 5 .
(5.13) Develop a stage-discharge curve for a deten­ Hoffman, C. J. (1974). Outlet works in "Design of Small Dams,"
tion pond having a principal spillway as in problem Chap. X. U.S. Bureau of Reclamation, U.S. Government Printing
(5.3) and an emergency spillway as in problem (5.12). Office, Washington, D C .

Note that the discharges must b e a d d e d to get total Kao, Τ. Y. (1975). Hydraulic design of storm water detention struc­
tures. In "Proceedings, National Symposium on Urban Hydrol­
discharge. Assume that the spillway crest is 9 ft above
ogy and Sediment Control, U K Y B U 109, College of Engineer­
the principal spillway inlet. ing, University of Kentucky, Lexington, KY."
(5.14) A 48-in.-diameter circular concrete culvert Mavis, F. T. (1942). T h e hydraulics of culverts, Bulletin 56, The
draining a subdivision will have a p e a k discharge of Pennsylvania State College, Engineering Experiment Station, PA.
75 cfs for 20 min as a result of a 10-year storm. McCuen, R. H. (1989). "Hydrologic Analysis and Design."
Determine the scour hole geometry and rock riprap Prentice-Hall, Englewood Cliffs, NJ.
Portland Cement Association (1962). "Culvert Design Aids: A n
requirements if the downstream channel has a 0.6%
Application of U.S. Bureau of Public Roads Culvert Capacity
slope. Charts." Portland Cement Association, Chicago, IL.
Portland Cement Association (1975). "Concrete Culvers and Con­
duits." Portland Cement Association, Skokie, IL.
Soil Conservation Service (1951). "Engineering Handbook," Hy­
References draulics Section 5. U.S. Department of Agriculture, Washington,
DC.
American Concrete Pipe Association (1985), "Concrete Pipe Design
Soil Conservation Service (1968). "Hydraulics of Broad-Crested
Manual." American Concrete Pipe Association, Vienna, V A .
Spillways." Technical Release N o . 39, Engineering Division, Soil
Barfield, B. J., Warner, R. C , and Haan, C. T. (1981). "Applied Conservation Service, U.S. Department of Agriculture, Washing­
Hydrology and Sedimentology of Disturbed Areas." Oklahoma ton, D C
Technical Press, Stillwater, OK.
Soil Conservation Service (1969). "Entrance H e a d Losses in Drop
Brater, E. R , and King, H. W. (1976). "Handbook of Hydraulics." Inlet Spillways," Design Note N o . 8. Engineering Division, D e ­
McGraw-Hill, New York. sign Branch, Soil Conservation Service, U.S. Department of Agri­
Bureau of Public Roads (1965a). Capacity charts for the hydraulic culture, Washington, D C .
design of highway culverts, Hydraulic engineering circular No. 10. Soil Conservation Service (1984). "Engineering Field Manual." U.S.
U.S. Government Printing Office, Washington, D C . Department of Agriculture, Washington, D C .
Bureau of Public Roads (1965b). Hydraulic charts for the selection of Stephenson, D . (1979). "Rockfill in Hydraulic Engineering."
highway culverts, Hydraulic engineering circular N o . 5. U.S. Gov­ Elsevier, N e w York.
ernment Printing Office, Washington, D C . Streeter, V. L. (1971). "Fluid Mechanics." 5th ed. McGraw-Hill,
Carter, R. W. (1957). Computation of peak discharge at culverts, NY.
ό
Channel Flow Routing
and Reservoir Hydraulios

In Chapter 4, steady flow in channels was considered. For a steep, prismatic, smooth channel with little
Steady flow refers to situations where t h e r e is no storage and n o intermediate inflows, the hydrograph at
change in flow characteristics at a point in a channel Β would be very much like the hydrograph at A. T h e
with respect to time. Similarly in C h a p t e r 5, steady flow flow would simply pass through the channel reach with
through hydraulic structures such as spillways, culverts, little alteration in the hydrograph except a time lag
drop structures, etc., was covered. C h a p t e r 3 was de­ equal to the travel time between A and B. If the
voted to estimation of runoff hydrographs from rainfall channel between A and Β was flat, irregular in shape,
events. In that chapter it was shown that runoff hydro- hydraulically rough, and had a lot of storage capacity,
graphs are time-varying descriptions of flow and thus the hydrograph at Β would be considerably different
natural runoff events produce unsteady flows. than the original hydrograph at A. In the absence of
This chapter is devoted to the analysis of unsteady inflows between points A and B, the hydrograph at Β
flow through channels and reservoirs. Storage in chan­ for this latter case would likely have a longer time to
nels and reservoirs has a major impact on flow hydro- peak, a longer time base, and a lower p e a k flow.
graphs. Generally, peak flows are reduced and base
times of hydrographs are prolonged. These characteris­
tics can b e used to great advantage in the design and FLOW ROUTING
operation of systems for controlling runoff events. Con­
sider two points, A and B, separated by several hun­ Flow routing consists of analytical techniques for
dred feet along a stream. T h e flow hydrograph at point determining the outflow hydrograph from a stream
A may b e known and the hydrograph at point B, reach or reservoir from a known inflow hydrograph.
downstream from A, desired. Obviously factors such as T h r o u g h flow routing, the impacts of channel a n d / o r
channel steepness, channel roughness, channel shape reservoir characteristics on hydrographs can be deter­
and available storage between points A and 5 , as well mined. F u r t h e r the impact of channel modifications or
as any additional flows into the channel between points changes in reservoir spillway characteristics on outflow
A and B, will impact the shape of the hydrograph hydrographs can be determined. Flow routing is essen­
at B. tial in any storm water runoff study. Flow routing is

182
Flow Routing 183

central to designing structures to control storm w a t e r


runoff events and to mitigate the impact of flood flows. In looking at t h e individual t e r m s a n d approximating
Flow routing is also an integral c o m p o n e n t of sediment partial derivatives by numerical derivatives, we have
pond design as it provides a m e a n s of estimating t h e
detention time and thus the settling opportunity time S 0 = 5 Χ 10" 3

for sediment particles.


Flow routing is logically a topic in hydraulics, yet it is 1 dv 1 0.5 f t / s e c / h r
Χ -ΓΤΤΤ- — = 4.3 Χ 10" 6

central to an understanding of hydrology. T h u s flow ~g~dt 32.2ft/sec 2


" 3600sec/hr
routing is treated both in books on hydraulics, espe­
cially o p e n channel hydraulics, and " e n g i n e e r i n g " hy­ V dv 5 ft/sec 0.5 f t / s e c / m i
= 1.5 Χ Μ Γ 5

drology. Flow routing procedures range in complexity g dx 32.2 f t / s e c 2


5280 f t / m i
from simple storage routing p r o c e d u r e s to relatively
complex procedures based on simultaneous solutions dy_ 1 ft/mi
= 1.9 X 10 - 4
to the hydrodynamic equations dealing with t h e conser­ dx 5280ft/mi
vation of m o m e n t u m and mass.
Storage routing is sometimes known as hydrologic
T h u s from Eq. (6.3),
routing since the basic equation is the continuity equa­
tion. Several methods of hydrologic routing are p r e ­
sented h e r e , including basic storage routing, Musk­ S =5X
; 10' 3
- 0.0043 Χ 1 0 " 3

ingum routing, convex routing, and kinematic routing.


Hydraulic flow routing is different from hydrologic + 0.015 Χ 1 0 " - 0.19 Χ 1 0 " = 4.69 Χ 1 0 "
3 3 3

routing in that it is based on both t h e m o m e n t u m a n d


continuity equations. Hydraulic routing is usually ac­ or S is 9 6 % of 5 .
f 0

complished by a numerical solution of t h e governing A s a second example, consider a stream where


equations or by t h e m e t h o d of characteristics. In C h a p ­ changes occur m o r e rapidly. In this case t h e slope
ter 3, the basic hydraulic equations for flow routing might be 0.02 f t / f t , t h e change in d e p t h 1 f t / 2 0 0 ft, a
were developed as Eqs. (3.36) and (3.37). T h e continu­ velocity of 10 fps, a velocity gradient of 1 f p s / 5 0 0 ft, a
ity equation was given as r a t e of change of velocity of 1 f p s / 3 min, and n o
lateral inflow. T h e individual terms would t h e n b e
dA dv dA
——V A-—h v~r~ = q(x, t) (6.1)
dt dx dx S n = 2 Χ 10" 2

and the m o m e n t u m equation was given as 1 dv 1 1 ft/sec


g^t 32.2 f t / s e c 2
3 min X 60 s e c / m i n
dv dv dy qv

* ^ ^ 7 - * ° - ' °· · 1.7 Χ 1 0 " 4


+ + + (5 5 ) = (6 2)

V dv 10 f t / s e c 1 ft/sec
= 6.2 Χ 1 0 " 4

T h e relative importance of the various terms in this g dx 32.2 f t / s e c 2


500 ft
last equation determines to a large extent the degree of
simplification warranted. This in t u r n d e t e r m i n e s t h e dy_ lft
flow routing method(s) that is a p p r o p r i a t e . Of course, = 5 Χ 10" . 3

dx 200 ft
the full m o m e n t u m equation can always be used but
may require more detailed information and calculation
T h u s from Eq. (6.3),
than necessary.
For example, consider a stream with a slope of 0.005
S = 0.02 - 0.017 - 0.062 - 0.500 = 0.01421
ft/ft, a change in d e p t h of 1 f t / m i l e , a velocity of 5 fps, f

a velocity gradient of 0.5 f p s / m i l e , a r a t e of change in


velocity of 0.5 f p s / h r , and n o lateral inflow. E q u a t i o n or S = 7 1 % of S .
f 0

(6.2) can b e written as In case two t h e t e r m s are m o r e nearly of t h e same


o r d e r of m a g n i t u d e t h a n they are in case o n e . In case
1 dv ν dv dy one, a simpler routing p r o c e d u r e might b e adopted
S = S - - — (6.3)
since S is nearly equal to S . In case two, such a
f 0
J
g dt g dx dx 0 f

simplification may a d d significant error. W h a t follows


184 6. Channel Flow Routing and Reservoir Hydraulics
is a discussion of routing techniques ranging from significant changes in the inflow hydrograph may occur
storage routing relying on the continuity equation to within the selected At and not be reflected in the
the use of the full hydrodynamic equations. outflow hydrograph, since they may pass completely
through the reach during the time interval.
F o r small channels, the nonuniformity of the channel
CHANNEL ROUTING may necessitate shorter reach lengths. As a general
guide, the average flow velocity may b e computed at
Storage Routing bankful discharge or at a flow equal to 7 5 % of the
hydrograph p e a k flow.
T h e simplest form of routing is based on the continu­ A common m e t h o d for solving Eq. (6.5) is to plot
ity equation and is known as storage routing. In C h a p ­ S + Ο At/2 and S - Ο At/2 versus the discharge or
ter 3, Eq. (6.1) was shown to come from the continuity d e p t h . T h e s e curves are called the characteristic curves.
equation [Eq. (3.34)]. If this equation is written over a Example Problem 6.1 illustrates the construction of
finite interval of time, Δ / , the result is these curves and their use in routing.

L + I 2 Oi + 0 2

- ^ γ 2
Δ / - - y - ^ At = S - S
L
2 v (6.4) Muskingum Method
A modification of storage routing that considers a
This equation may b e rearranged to linear change in d e p t h along the reach is known as the
Muskingum m e t h o d . In t h e previous discussion of stor­
0 O L +1 age routing, the storage in a channel reach was related
2 x 2

S +-±At
2 = S x - ^ A t + -±-^At. (6.5) to d e p t h of flow in the reach based o n t h e outflow rate
in t h e reach. In reality t h e flow d e p t h would not be
constant along the reach because the inflow rate would
In these equations, the subscripts 1 and 2 refer to not b e t h e same as t h e outflow rate. If the flow was in a
conditions at the beginning and end of a time interval, rising stage, t h e inflow into t h e reach would exceed the
respectively. / represents the inflow, Ο the outflow, outflow. T h u s t h e d e p t h of flow at the upstream end of
and S the storage in a channel reach. T h e time interval the reach would exceed that of t h e downstream e n d of
is represented by At. the reach. A " w e d g e " of storage above a uniform flow
T h e storage in a channel reach d e p e n d s on the rate corresponding to t h e downstream flow depth would
channel geometry and depth of flow. T h e flow rate may exist. T o partially overcome this nonuniformity, the
be related to t h e depth of flow, assuming steady, -uni­ Muskingum m e t h o d of streamflow routing makes the
form flow using Manning's equation. A simple method storage in t h e reach a linear function of both the inflow
for computing the storage is to base it on the average and outflow rate
cross-sectional area of the reach for a given flow rate.
T h u s S would be determined by developing a flow
S = k[xl + ( 1 -χ)θ], (6.6)
r a t e - a r e a relationship based on Manning's equation
for each cross section. T h e length of the channel
section multiplied by the average cross-sectional area where k and χ must b e determined from channel
of the channel at a given flow rate would give the characteristics. T h e coefficient k is known as the stor­
storage in the reach at that flow rate. age constant a n d is approximately equal to the travel
Tributary inflows, overland flow, and ground water time through the reach. F o r best results, both k and χ
contributions to flow can be added to the inflow or should b e based on observed hydrographs. A n χ of
outflow of the reach as appropriate. Normally, channel zero corresponds to reservoir storage routing. A value
routing requires that the channel be divided into sev­ of χ = } makes the storage a function of the average
eral reaches. T h e outflow from one reach becomes the flow rate in the reach based on t h e inflow and outflow.
inflow to the next reach downstream. T h e channel In the absence of streamflow records on / and 0 , k
should be divided into reaches having relatively uni­ may b e estimated as t h e flow travel time in the reach,
form hydraulic properties. T h e routing interval At and χ may b e taken as about 0.25. T h r o u g h manipula­
should not exceed one-fifth to one-third of the time to tion of Eq. (6.5) and (6.6), o n e can obtain a linear
peak of the hydrograph being routed. T h e routing expression for outflow in the form
interval should not exceed the travel time through the
reach. If these guidelines on At are not followed, 0 2 = C /
0 2 + CJ X + CO 2 l9 (6.7)
Channel Routing 185

where at the p e a k flow rate and S is the slope of the


0

channel. A s previously discussed, the routing interval


-kx + 0.5Δί
should not exceed i / 5 .
c =
0 k(l - * ) + 0.5 Δ ί
(6.8) p

kx + 0.5 At
(6.9)
Convex Routing
* ( 1 -x) + 0.5 Δ ί
T h e Soil Conservation Service (U.S. D e p a r t m e n t of
* ( 1 -x) - 0.5 Δ / Agriculture, 1971) presents a channel routing proce­
c =
2 A:(l -x) + 0.5 Δ ί
(6.10) d u r e similar to the Muskingum m e t h o d that is known
as the Convex m e t h o d . T h e Convex m e t h o d of routing
Curves relating C , C and C to t h e outflow may
0 v 2 involves only inflow-outflow hydrograph relationships.
be constructed. If k and χ are assumed constant, t h e T h e continuity equation is not directly involved, neces­
routing is greatly simplified since C , C and C t h e n
0 v 2 sitating close a d h e r e n c e to procedures recommended
become constants. by the Soil Conservation Service. In the Convex
If streamflow data on inflow and outflow to the m e t h o d , the routing equation is
reach of interest are available, values of k and χ may
be determined from these data. Equation (6.6) shows 0 ^ ( l - C ) O
2 l + CIu
(6.14)
that if S is plotted against xi + (1 - x)0, a straight
line with a slope k should result. Since the inflow and w h e r e C is a p a r a m e t e r such that 0 ^ C <. 1.0. T h e
outflow hydrographs are known, Eq. (6.4) can b e solved p a r a m e t e r C can b e estimated from
for S . T h e initial value for S is not important since
2 l
C = ϋ/(1.7 + υ), (6.15)
the slope of the S vs xi + (1 — x)0 relationship and
not the intercept is being sought. With S known, w h e r e υ is the average flow velocity of the reach, υ
values of χ are assumed and S plotted against xi + may b e c o m p u t e d at bankful discharge, at a flow equal
(1 - x)0. Several values of χ are tried. T h e value that to 7 5 % of t h e p e a k flow or at some other appropriate
gives the smallest loop in the plotted relationship is value. T h e C value may also b e approximated from the
taken as the desired χ value and the slope of the χ in the Muskingum m e t h o d as C « 2 χ if an appropri­
plotted relationship is taken as the k value. ate χ is available.
Generally for small catchments, measured inflow T h e p r o p e r routing interval to use with the Convex
and outflow hydrographs are not available, and k and m e t h o d is
χ must b e approximated. Several attempts to derive
procedures for estimating k and χ that are better t h a n Δί = CK, (6.16)
simply setting k equal to reach travel time and χ equal
w h e r e Κ is a p a r a m e t e r similar to k of the Muskingum
to some assumed constant have b e e n m a d e . A proce­
m e t h o d s and may be estimated as the travel time
dure developed by Cunge (1967) has received wide­
through the reach. Generally, computing At from Eq.
spread acceptance. In the M u s k i n g u m - C u n g e method,
(6.16) results in an inconvenient time interval. T h e C
k is determined from
value of C * for a m o r e convenient time interval can be
k = Ax/c, (6.11) calculated from

(Δί*/Δί)
where Ax is the reach length and c represents a flood C* = 1 - ( 1 - C) (6.17)
wave celerity determined from
w h e r e Δ ί is from Eq. (6.16) a n d Δ ί * is the desired
c = mv. (6.12) time interval. T h e ratio At*/At should be kept as n e a r
unity as possible. E q u a t i o n (6.14) is now the routing
T h e coefficient m comes from the uniform flow equa­
relationship with C replaced by C*. N o t e that t h e ratio
tion and can be taken as 5 / 3 (Viessman et al, 1989).
of At*/At can b e varied by changing the reach length
T h e velocity, υ, may b e taken as the average velocity at
or the routing interval. Again Δ ί should b e limited to
bankful discharge. T h e value of χ is then determined
from
In 1983, the Soil Conservation Service (SCS) re­
placed t h e Convex m e t h o d in their T R - 2 0 program
(6.13)
X
2\ S cAx)'
0
with a m e t h o d known as the att-kin m e t h o d because it
combines elements from the kinematic method with an
where q is the flow per unit width generally calculated
0 A t t e n u a t i o n p r o c e d u r e based on storage routing. T h e
186 6. Channel Flow Routing and Reservoir Hydraulics

storage routing assumes

Q = kS m
(6.18)

and t h e kinematic routing uses

±
Q = bA m
(6.19)

w h e r e Q is flow, 5 is t h e storage, A is t h e flow area,


and b, k, and m are coefficients. T h e factor m relates
At
velocity and wave celerity

c = mu. (6.20) "T


Based on Manning's equation, the SCS uses an m of
5 / 3 . T h e equations and repeated routings are used t o
assure that t h e peak flow resulting from t h e kinematic
Upstream - ΔΧ -Downstream
routing equals that from the storage routing a n d that
t h e p e a k outflow from t h e reach occurs at the time of
maximum storage. Figure 6.1 Finite difference grid.

Kinematic Routing
If all of the terms of Eq. (6.2) are neglected except Substituting these equations into Eq. (6.21) results in
the last o n e , the result is
Q4-Q2 A 4 - A 3 + A 2 - A 1 _
S = S + — = q
f 0 Ax 2 At

or a condition of uniform flow. Kinematic routing is or


based on t h e continuity equation and a uniform flow
equation. In theory, kinematic waves will not acceler­ A A A l + A 3 At
ate significantly and flow downstream without a p p r e ­
ciable change in shape or loss in peak flow. Kinematic
routing solves the continuity equation in t h e form which may b e written

dQ dA A 4

— + — = 4 (6.21) XQ + γ = a + β, (6.26)
dx dt
4

and a flow equation or rating function where

Q-f(A), (6.22) Δί
λ = (6.27a)
Αχ
where Q is t h e flow rate in cfs, A is t h e flow a r e a in
square feet, q is any lateral inflow ( + ) or outflow ( - ) A,+A
« = —γ—
3

(6.27b)
along the channel in cfs p e r foot, χ is t h e distance
along t h e channel in feet, and t is t h e time in seconds.
β = λβ 2 + Atq-A /2. 2 (6.27c)
Several numerical solution techniques are available.
T h e o n e that follows is based on Brakensiek (1966). E q u a t i o n (6.26) may b e solved by noting that Q =
Using t h e grid system shown in Fig. 6.1, t h e following f(A) a n d rewriting t h e equation as
approximations are m a d e :
0 = α + β - kf{A ) 4 -AJ2. (6.28)
dQ = Q ~Q4 2
(6.23)
dx Ax A value of A that satisfies E q . (6.28) is sought, and
A

dA A 4 —A 3 +A 2 — Ax
t h e n Q is calculated from Eq. (6.22). q represents
4

(6.24) lateral inflow and inflow from very small tributaries.


~dt ~ 2 At
Inflow from large tributaries would b e a d d e d to
(6.25) Q4 prior t o routing down t h e next channel reach.
M a n n i n g ' s equation can b e used to define Eq. (6.22).
Hydraulic Flow Routing 187

HYDRAULIC FLOW ROUTING

Chow (1959), H e n d e r s o n (1966), Chow et al (1988),


and others discuss hydraulic flow routing based on
equations of spatially varied (nonuniform), unsteady
flow (Saint-Venant equations). T h e s e equations involve
continuity and m o m e n t u m as given in Eq. (6.1) a n d
(6.2). T h e s e equations can be written as Figure 6.2 Channel cross section for Example Problem 6.1.

dv dA dA
A— + ν— + — = q (6.29)
dx dx dt v
'
equation using N e w t o n - R a p h s o n iterative techniques
and to perform channel routings.
F r e a d (1985) h a s c o m b i n e d D E W O P E R a n d
D A M B R K in a p r o g r a m known as F L D W A V adding
dv dv g dyA vq
features not found in either of t h e two-component
programs. F L D W A V is a large, generalized, one-di­
mensional routing p r o g r a m with many features. This
w h e r e A is area, υ is velocity, χ is distance, q is inflow versatility and power comes at the expense of com­
per unit length of channel, g is gravity, S is the p u t e r size a n d speed a n d input d a t a requirements.
channel slope, and S is the friction slope. A consistent
f

set of units must be used in these equations.


Equations (6.29) and (6.30) may be solved using
finite differences or t h e m e t h o d of characteristics. Ei­ Example Problem 6.1 Storage routing
ther treatment is beyond the scope of coverage p r e ­
sented here. T h e r e a d e r can consult o n e of t h e above A channel is 2500 ft long, has a slope of 0.09%, and is
references for more detailed information. clean with straight banks and no rifts or deep pools. The
appropriate Manning's η is 0.030. A typical cross section is
Very few computer models use the full m o m e n t u m
shown in Fig. 6.2. Along the length of the channel there is no
equation for routing. Situations involving backwater lateral inflow. The inflow hydrograph to the reach is triangu­
effects, tidal flows, surges, and flow junctions w h e r e lar in shape with a base time of 3 hr, a time to peak of 1 hr,
large tributaries enter the main channel are examples and a peak flow rate of 360 cfs. Route the hydrograph
of when the full m o m e n t u m equation should b e used. through the channel reach using the storage routing proce­
T h e National W e a t h e r Service D A M B R K model dure.
(Fread, 1977) uses the full equation to route the flow
Solution: Table 6.1 contains the channel hydraulic proper­
surge resulting from a d a m failure as it passes down
ties required to compute the quantities needed for the rout­
the channel below the dam. T h e National W e a t h e r ing using Eq. (6.5). The calculations required for the entries
Service model D E W O P E R (Dynamic Wave O p e r a ­ in this table are illustrated for a depth of 3.0 ft. The area at
tional Model) (Fread, 1978) solves the full m o m e n t u m y = 3 is equal to the area at y = 2 plus the area between

Table 6.1 Channel Properties for Example Problem 6.1

y A wp R V Q S S-OAt/2 S+OAt/2
(ft) (ft )2
(ft) (ft) (fps) (cfs) (ft )3
(ft )3
(ft )3

0 0 0 0 0 0 0 0 0
1 8 10.46 0.76 1.25 9.97 20,000 17,010 22,990
2 20 14.94 1.34 1.81 36.2 50,000 39,140 60,860
3 40 27.1 1.48 1.93 77.27 100,000 76,818 123,182
4 72 39.29 1.83 2.23 160.68 180,000 131,795 228,205
5 116 51.44 2.26 2.56 297.3 290,000 200,809 379,191
6 172 63.61 2.7 2.89 497.58 430,000 280,727 579,273
188 6. Channel Flow Routing and Reservoir Hydraulics

y = 2 and y = 3: computations. One step in the routing is illustrated for the


time interval from 40 to 50 min.
A3 = A + ^4 -3
2 2

L +I 240 + 300
2

= 20 + Μ + z < / = 20 + 14 X 1 + 6 X l = 40 ft
2 2 2
„ Δί X 600 = 162,000 ft 3

2 2
wp = wp +
3 2 wp _ 2 3 y = y from the previous step = 3.84 ft.
x 2

2 Λ. 12 2 ^ / 2 ^ ^
S - Ο At/2 is read from the routing curve corresponding to
= 14.94 + 2 ( 6 + l ) 2 1 4 9 4 + 1 2 > 1 7 = 2 7 J ft

y = 3.84 ft as 121,500:
40
= 1.48 ft
H>/? 3
27Λ h + /
2
S + Ο At/2 = At + S - Ο At/2
1 49 1 49
V = —Λ /3 1/2 2
5 . _ ( 1 4 8 )2/3 ( 0 0 0 0 9 ) l/2 = Χ 93 f p s
= 162,000 + 121,500 = 283,500.
y is read from the routing curve corresponding to S +
β = vA = 1.93 X 40 = 77.27. 2

OAt/2 = 283,500 as 4.40 ft.


The discharge is determined from Manning's equation as
Storage in the reach, 5, is computed as the product of the
Q = vA where υ and A are computed for the channel
reach length and the cross-sectional area: f

corresponding to a depth of 4.40 ft. The result is Q = 208


cfs. Similar calculations are carried out for all of the time
S = LA = 2500 X 40 = 100,000 ft . 3

intervals. Figure 6.4 shows a plot of the resulting hydrograph.


The travel time through the reach is estimated at bankfull
discharge as
Example Problem 6.2 Muskingum-Cunge routing
t = L/υ = 2500/2.9 = 862 sec = 14.4 min
t

t /5
p = 6 0 / 5 = 12 min. Work Example Problem 6.1 using the Muskingum-Cunge
routing procedure.
Based on these two estimates, a routing interval of Δ* = 10 Solution: The Muskingum-Cunge procedure relies on Eqs.
min or 600 sec is chosen: (6.7) and (6.11) to (6.13). The bankful velocity is 2.89 fps;
hence, from Eq. (6.12) using m = f,
S - OAt/2 = 100,000 - 77.27(600)/2 = 76,818 ft 3

c =v
m = f 2.89 = 4.82
S + OAt/2 = 100,000 + 77.27(600)/2 = 123,182 ft . 3

Ax 2500
k = = 518.7
A plot of S ± Ο At/2 vs y yields the storage characteristic 4.82
curves shown in Fig. 6.3. These curves along with equation
6.5 are used to do the routing. Table 6.2 contains the actual - i ( i - - i U .
2\ SqcAx)

600 q is the flow per unit width and is approximated as a flow


0

rate divided by the top width:

.2 5 0 0 μ 500
« ο - β - - 8 . 0 6
o
4 0 0 \- _ 1/ 8.06
b
= 0.131
ο
X =
2\ ~ 0.0009(4.82)2500
ο
Γ- 3 0 0 * ( 1 -x) + 0.5 At = 518.7(1 - 0.131) + 0.5(600) = 750.75
-kx + 0.5 At -518.7(0.131) + 300
0.309
"Ο 2 0 0 C
° 750.75 750.75
Ο
kx + 0.5 Δί 518.7(0.131) + 300
I = 0.490
100 750.75 750.75
+
k{\ - x) - 0.5 Δί 518.7(1 - 0.131) - 300
= 0.201
750.75 750.75
0 1 2 3 4 5
Depth o f Flow ( f t ) o 2 = C / + CI + CO
0 2 X X 2 x

Figure 6.3 Storage characteristic curves. = 0.309/ + 2 0.490Λ + 0.201OP


Hydraulic Flow Routing 189
Table 6.2 Storage Routing Computations

t / (7,+/ )Δ//2 y\ S-OAt/2 5+ΟΔ//2 0


2
yi
(min) (cfs) (ft ) 3
(ft) (ft ) 3
(ft )
3
(ft) (cfs)

0 0 0 0
10 60 18,000 0 0 18,000 0.82 7
20 120 54,000 0.82 13,800 67,800 2.15 40
30 180 90,000 2.15 43,000 133,000 3.12 85
40 240 126,000 3.12 82,000 208,000 3.84 144
50 300 162,000 3.84 121,500 283,500 4.4 208
60 360 198,000 4.4 158,000 356,000 4.87 275
70 330 207,000 4.87 191,000 398,000 5.11 316
80 300 189,000 5.11 208,500 397,500 5.11 316
90 270 171,000 5.11 208,500 379,500 5 297
100 240 153,000 5 200,800 353,800 4.85 272
110 210 135,000 4.85 190,000 325,000 4.68 246
120 180 117,000 4.68 178,000 295,000 4.49 220
130 150 99,000 4.49 163,000 262,000 4.26 188
140 120 81,000 4.26 148,000 229,000 4 161
150 90 63,000 4 132,000 195,000 3.73 133
160 60 45,000 3.73 114,000 159,000 3.4 105
170 30 27,000 3.4 97,000 124,000 3.01 78
180 0 9,000 3.01 77,000 86,000 2.49 52
190 0 0 2.49 55,000 55,000 1.87 31
200 0 0 1.87 36,000 36,000 1.39 18
210 0 0 1.39 25,000 25,000 1.07 11
220 0 0 1.07 18,500 18,500 0.85 7.5
230 0 0 0.85 14,000 14,000 0.68 5
240 0 0 0.68 11,000 11,000 0.56 3.5
250 0 0 0.56 9,000 9,000 0.47 2.5
260 0 0 0.47 7,000 7,000 0.37 2
270 0 0 0.37 6,000 6,000 0.33 1.5

Sum 3240 Sum 3225

The routings based on this relationship are shown in Table Solution: The solution is based on Eqs. (6.14)-(6.17) using
6.3. As an example calculation, the outflow at 60 min is an average bankful flow velocity of 2.8 fps:
calculated as
0 2 = (1 - C)0 1 + CI X

0 = 0.309(360) + 0.490(300) + 0.201(248) = 308.


2
υ 2.8
C = = = 0.622
The results are plotted in Fig. 6.4. 1.7+ u 1.7 + 2.8
Δ ί = CK
Κ = travel time « L/υ = 2500/2.8 = 892.9

Example Problem 6.3 Convex routing Δ ί = 0.622(892.9) = 555 sec

Do the routing of Example Problem 6.1 using the Convex C * « 1 - (1 - Ο ' * δ / Δ /


= 1 - (1 - 0.622) 6007555
- 0.65
routing method. 0 = (1 - 0.65)0! + 0.65/! = O.350J + 0.65/!.
2
190 6. Channel Flow Routing and Reservoir Hydraulics

Table 6.4 contains the calculations, and Fig. 6.4 shows the Table 6.3 Routing Calculations for Example
results of this routing. Problem 6.2

Time Inflow Outflow


(min) (cfs) (cfs)
Example Problem 6.4 Kinematic routing
0 0 0

Do the routing of Example Problem 6.1 using the kine­ 10 60 19


matic routing approach. 20 120 70
30 180 129
Solution: From Fig. 6.1, it can be seen that the points in
the grid and the subscripts on the symbols in the kinematic 40 240 188
routing relations have the following meaning: 50 300 248
60 360 308
Subscript Meaning 70 330 340
80 300 323
1 Upstream end of reach at start of time interval
90 270 295
2 Upstream end of reach at end of time interval
100 240 266
3 Downstream end of reach at start of time interval
110 210 236
4 Downstream end of reach at end of time interval
120 180 206
130 150 176
Δί is taken as 10 min. Ax is 2500 ft. The parameter λ is 140 120 146
given by Eq. (6.27):
150 90 116
At 10 min X 60 sec/min 160 60 86
λ = = 0.24 sec/ft.
Ax 2500 ft 170 30 56
180 0 26
Equation (6.28) can be written
190 0 5
α + β =\f(A ) 4 +A /2 4 200 0 1
210 0 0
or
220 0 0
0 . 2 4 / ( Λ ) + 0.5 A = α
4 4 +β.
230 0 0
Figure 6.5 is a plot of this function. Points on the line of 240 0 0
a + β or 0.24 f(A ) -I- 0.5 A are calculated by assuming a
4 4
250 0 0
value of y and calculating A and f(A ) or Q . The routing 4 4 4
260 0 0
is done by calculating a + β at each time step from Eq.
(6.27) and then determining A and Q from the curves. 270 0 0
4 4

Table 6.5 contains the routing computations for the kine­ 280 0 0
matic method. The computations are illustrated by consider­ Sum 3240 3240
ing the time interval from 50 to 60 min. Recall that condi­
tions at points 1, 2, and 3 are known and conditions at point
4 are being sought. Also recall that for a prismatic channel
under the assumption of uniform flow, the relationship be­
tween flow area and flow depth is the same all along the routing. It can be seen that the results are very similar to
channel. From Table 6.5, A A > and A are 133.5,126, and l9 2 3 those of the other routing methods. It could be expected that
111.2 ft , respectively:
2
the differences would increase if the routing involved several
channel reaches.
a - (A x +A )/2 3 - (133.5 + 111.2)/2 = 122.35
β - XQ - A /2 2 2 = 0.24(330) - 126/2 = 16.2
α+β = 122.35 + 16.2 = 138.55.

From Fig. 6.5, A and Q are found to be 123.5 ft and 2 RESERVOIR ROUTING
4 4

321 cfs, respectively. For the next time increment, Q and A x x

are Q and A from, the previous time, Q and A are Q Reservoir routing is generally d o n e by using storage
2 2 3 3 4

and A from the previous time, Q is the inflow hydrograph


4 2
routing similar to t h e storage routing p r o c e d u r e ex­
value at the end of the time increment, and A is the flow 2
plained earlier for channel routing. T h e continuity
area corresponding to Q . Figure 6.4 contains the results of
2 equation in t h e form of Eq. (6.4) a n d (6.5) is used.
Reservoir Routing 191
400 400

300 300

υ
200 h 200
ο

100 h 100 h

0 50 100 150 200 250 0 50 100 150 200 250


Time ( m i n ) Time (min)

400 400

300 h 300 h
0)
υ
200 \- 200 h
ο

100 h 100 h

50 100 150 200 250 0 50 100 150 200 250


Time (min) Time (min)
Figure 6.4 Comparison of routing results.

Outflow from the reservoir is controlled by the princi­ course, the point of intersection of the two hydro-
pal spillway at a rate depending on the height of water graphs.
above the inlet. A t the beginning of inflow, the d e p t h T h e storage volume required in the reservoir is the
of water above t h e inlet is small, and the inflow to t h e area b e t w e e n t h e inflow a n d outflow hydrographs prior
reservoir will exceed the outflow. A s a result, the d e p t h to the p e a k outflow as shown in Fig. 6.7. A n additional
of water in the reservoir will increase, thereby increas­ feature of the hydrographs is that the areas u n d e r the
ing the outflow rate. This process will continue until inflow and outflow hydrographs must b e the same if
the outflow rate equals the inflow rate. This point will the initial water level is at the spillway crest.
occur sometime after the p e a k inflow r a t e has oc­
curred. From this point, the outflow will exceed the
inflow and the depth of water or storage in the reser­ Graphical Routing: Puis Method
voir will decrease. Figure 6.6 illustrates this process. Puis m e t h o d is a p r o c e d u r e for graphically solving
T h e relationship of storage relative to inflow and t h e continuity equation using storage characteristic
outflow hydrographs is shown in Fig. 6.7. T h e p e a k curves as was d o n e in t h e case of stream routing. T h e
discharge of the outflow hydrograph corresponds to the s t a g e - s t o r a g e curve is developed from topographic in­
point where it intersects the inflow hydrograph. W h e n formation relative to the reservoir site, and the
the outflow is a maximum, the storage will also b e a s t a g e - d i s c h a r g e curve is based on the hydraulics of the
maximum, since discharge and storage are both in­ reservoir outlet as shown in C h a p t e r 5. T h e routing
creasing functions of stage. T h e maximum storage and time interval should b e 10 to 2 5 % of the time to peak
outflow occurs when inflow equals outflow, which is, of of the inflow hydrograph to ensure that the numerical
192 6. Channel Flow Routing and Reservoir Hydraulics
Table 6.4 C o n v e x Routing Results 250 700
650
Time Inflow Outflow
- 600
(min) (cfs) (cfs)
flow-
200 -j 5 5 0
0 0 0 -ΐ 5 0 0
10 60 0
£ l 5 0 area
i 4 5 0

20 120 39
123.5 -j 4 0 0 £
ο
30 180 92 υ 321 -I 3 5 0 ^
40 240 149 I 300 £
50 300 208 250
< 100 - m ι
60 360 268
m ι 200
150
70 330 328
50 - 100
80 300 329
50
90 270 310
0
100 240 284 0 50 100 150 200 250 300
α + β
110 210 255
Figure 6.5 Channel properties for kinematic routing.
120 180 226
130 150 196
140 120 166
150 90 136 in Table 6.6 and Fig. 6.8 using a routing time interval of 10
min.
160 60 106
170 30 76 Solution: The results of the calculations are shown in
180 0 46 Table 6.6. columns 1 and 3 contain topographic information
relative to the site. Column 2 contains the reservoir stage
190 0 16
assuming a datum of 0 at elevation 400.0 ft. Column 6
200 0 6
contains the discharge for the reservoir spillway correspond­
210 0 2 ing to the stage of column 2. Column 4 is the storage volume
220 0 1 contained between the indicated stage and the immediately
230 0 0 preceding stage. The incremental storage is calculated from
240 0 0
AS-{A +A )(z -z )/2,
1 2 2 x
250 0 0
260 0 0
where ζ refers to elevation. Thus for the increment from
270 0 0 401.5 to 402.0 ft,
280 0 0
Sum 3240 3240 AS = (6.0 - 4.5)(402.0 - 401.5)/2 = 2.6 acre-ft.

The total storage in column 5 is the sum of the incremental


storages. The storage at 402.0 ft is
averaging does not diminish the impact of the p e a k
S = 4.4 + 2.6 = 7.0 acre-ft.
flow. For reservoirs that store a large part of the flow,
the routing interval may be increased after the peak The last two columns are based on a time increment of 10
outflow rate has occurred. If the routing time interval min. The entry in column 7 for 402.0 ft is determined as
is changed, a new set of storage characteristic curves
will be required because they d e p e n d on At. Extreme 7.0 - 120.3(3600)(10)/60/43560/2 = 6.2.
care must be used in any graphical routing as the
errors tend to be cumulative rather than random. The results of the calculations are plotted in Fig. 6.9.

Example Problem 6.5 Storage characteristic curve Example Problem 6.6 Puis routing

Develop the storage characteristic curves for a reservoir Route the hydrograph of Example Problem 6.1 through
whose stage-area and stage-discharge relationships are shown the reservoir of Example Problem 6.5 using the Puis method.
Reservoir Routing 193

Table 6.5 Routing Computations for Example Problem 6.4

fii G 3 l
A
^3 α Q2 A
l
β α+β \ 04
(min) (cfs) (cfs) (ft )2
(ft )2
(ft )2
(cfs) (ft )2
(ft ) 2
(ft )2
(ft ) 2
(cfs)

0 0 0 0 0 0 60 32.7 -1.95 -1.95 0 0


10 60 0 32.7 0 16.35 120 57.5 0.05 16.4 17.9 31
20 120 31 57.5 17.9 37.7 180 79 3.7 41.4 42.66 84
30 180 84 79 42.66 60.83 240 98 8.6 69.43 67.5 148
40 240 148 98 67.5 82.75 300 117 13.5 96.25 90.2 215
50 300 215 117 90.2 103.6 360 133.5 19.65 123.25 111.2 281
60 360 281 133.5 111.2 122.35 330 126 16.2 138.55 123.5 321
70 330 321 126 123.5 124.75 300 117 13.5 138.25 123.5 321
80 300 321 117 123.5 120.25 270 108 10.8 131.05 118 304
90 270 304 108 118 113 240 98 8.6 121.6 110 277
100 240 277 98 110 104 210 89 5.9 109.9 100 245
110 210 245 89 100 94.5 180 79 3.7 98.2 91.5 217
120 180 217 79 91.5 85.25 150 68 2 87.25 83 192
130 150 192 68 83 75.5 120 57.5 0.05 75.55 72.5 163
140 120 163 57.5 72.5 65 90 45 -0.9 64.1 63.1 135
150 90 135 45 63.1 54.05 60 32.7 -1.95 52.1 52.4 106
160 60 106 32.7 52.4 42.55 30 17.5 -1.55 41 42.5 83
170 30 83 17.5 42.5 30 0 0 0 30 32 59
180 0 59 0 32 16 0 0 0 16 17.5 31
190 0 31 0 17.5 8.75 0 0 0 8.75 10.3 15
200 0 15 0 10.3 5.15 0 0 0 5.15 7 8
210 0 8 0 7 3.5 0 0 0 3.5 5 6
220 0 6 0 5 2.5 0 0 0 2.5 3.5 3
230 0 3 0 3.5 1.75 0 0 0 1.75 2.5 2
240 0 2 0 2.5 1.25 0 0 0 1.25 2 1
250 0 1 0 2 1 0 0 0 1 1.5 1
260 0 1 0 1.5 0.75 0 0 0 0.75 1.25 1
270 0 1 0 1.25 0.625 0 0 0 0.625 1 1
280 0 0

Sum 3240 3249

Solution: The routing calculations based on Eq. (6.5) are From Eq. (6.5),
contained in Table 6.7. The time increment from 50 to 60 0 L +1 Ο
2

min. is used to illustrate the steps involved: S + — Δί = — - — Δί + S - — Δί - 4.55 + 5.12 = 9.67.

From Fig. 6.9, y is found to be 2.34 ft, and from Fig. 6.8,
2
I = 360 cfs,
x I = 360 cfs
2
the outflow is found to be 130.2 cfs. The resulting hydrograph
is shown in Fig. 6.10.
300 + 360 10(60)
Δί = X ' / = 4.55 acre-ft.
2 43560
Numerical Routing
y is the stage at the end of the previous time interval or 1.89
x

ft. S = ( 0 / 2 ) Δί is read from Fig. 6.9 corresponding to a O n e numerical reservoir routing procedure is to
stage of 1.89 feet as 5.12 acre-ft. solve Eq. (6.4). T h e unknowns in this equation are S 2
194 6. Channel Flow Routing and Reservoir Hydraulics

and 0 . Both 5 and Ο are functions of t h e stage y.


2

Therefore, S and Ο are functionally related allowing a


unique solution for 0 . T h e solution is not explicit and
2

thus requires an iterative solution. Fortunately, the


solution converges very rapidly, and only 2 or 3 itera­
tions a r e required in most instances. O n e solution
p r o c e d u r e is as follows:

1. A s s u m e 0 = O 2 v

2. Calculate AS from Eq. (6.4).


3. Calculate S = S + AS. 2 x

4. D e t e r m i n e y for S and the s t a g e - s t o r a g e curve.


2 2

5. D e t e r m i n e 0 for y from the s t a g e - d i s c h a r g e


2 2

curve.
6. R e p e a t steps 2 - 5 until 0 remains unchanged. 2

This p r o c e d u r e is illustrated in Example Problem


Figure 6.6 Illustration of the continuity relationship in reservoir 6.7.
routing.

Example Problem 6.7 Numerical routing

Route the inflow hydrograph of Example Problem 6.1


through a reservoir having stage-storage and stage-dis­
charge relationships defined by

S = l.ly 1.727
and
Ο = 7y 0 4 9 8
,

Figure 6.7 Features of inflow and outflow hydrographs. where storage, 5, is in acre-feet; stage, y, is in feet; and

Table 6.6 Storage Characteristic Curves for Example Problem 6.5

Elevation Stage Area Incremental Storage Flow S-OM/2 S+OAt/2


(ft) (ft) (acres) volume (a-f) (a-f) (cfs) (a-f) (a-f)

399.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0


399.5 0.0 0.5 0.1 0.1 0.0 0.1 0.1
400.0 0.0 1.0 0.4 0.5 0.0 0.5 0.5

400.5 0.5 2.0 0.8 1.3 49.9 0.9 1.6


401.0 1.0 3.0 1.3 2.5 85.1 1.9 3.1
401.5 1.5 4.5 1.9 4.4 104.2 3.7 5.1
402.0 2.0 6.0 2.6 7.0 120.3 6.2 7.8

402.5 2.5 7.5 3.4 10.4 134.6 9.4 11.3


403.0 3.0 9.0 4.1 14.5 147.4 13.5 15.5
403.5 3.5 12.0 5.3 19.8 153.0 18.7 20.8
404.0 4.0 15.0 6.8 26.5 158.4 25.4 27.6

404.5 4.5 17.5 8.1 34.6 163.6 33.5 35.8

405.0 5.0 20.0 9.4 44.0 168.7 42.8 45.2

405.5 5.5 24.0 11.0 55.0 173.6 53.8 56.2

406.0 6.0 28.0 13.0 68.0 178.4 66.8 69.2


Reservoir Routing 195

Stage (ft) Stage (ft)


Figure 6.8 Stage-discharge and stage-storage relationship for Ex- Fj g u r e 6 < 9 storage characteristic curves for Example Problem 6.6.
ample Problem 6.6.

Table 6.7 Routing Calculations for Example Problem 6.6

Time Inflow / Δί y\ S-OM/2 5+0Δ//2 ?2 Outflow


(min) (cfs) ave
(a-f) (ft) (a-f) (a-f) (cfs)
(ft)

0 0 0.00 0.0
10 60 0.41 0.00 0.0 0.41 0.25 17.9
20 120 1.24 0.25 0.3 1.54 0.61 66.4
30 180 2.07 0.61 0.6 2.69 1.00 85.1
40 240 2.89 1.00 1.5 4.39 1.44 102.1
50 300 3.72 1.44 3.0 6.72 1.89 117.0
60 360 4.55 1.89 5.1 9.65 2.34 130.2
70 330 4.75 2.34 7.9 12.65 2.72 140.3
80 300 4.34 2.72 10.8 15.10 3.00 147.4
90 270 3.93 3.00 13.1 16.10 3.10 148.5
100 240 3.51 3.10 14.0 17.10 3.20 149.7
110 210 3.10 3.20 15.0 18.20 3.31 150.9
120 180 2.69 3.31 16.1 18.79 3.36 151.4
130 150 2.27 3.36 16.6 18.87 3.37 151.6
140 120 1.86 3.37 16.8 18.66 3.35 151.3
150 90 1.45 3.35 16.6 18.10 3.30 150.8
160 60 1.03 3.30 16.0 17.03 3.19 149.6
170 30 0.62 3.19 14.9 15.50 3.04 147.8
180 0 0.21 3.04 13.9 14.10 2.89 144.6
196 6. Channel Flow Routing and Reservoir Hydraulics

discharge, 0 , is in acre-feet per hour. Use the numerical


routing method, a time increment of 10 min, and an initial
stage of zero.
Solution: The solution is shown in Table 6.8. Note that
after 3 time steps, the solution converges in two iterations
and after about 10 time steps convergence has been achieved
after the first iteration. The last column contains the outflow
hydrograph.
The calculations are illustrated by considering the line
corresponding to a time of 90 min. The heading numbers in
parentheses refer to column numbers in Table 6.8.

(1 and 2) From inflow hydrograph

(3)
Ι + I
γ 2 300 + 270 ft 3

60 90 120 150 180


2 2 sec
Time (min)
X10 min X 60 sec 1 acre-ft = 3.93.
min 43560 ft 3
Figure 6.10 Hydrographs for Example Problem 6.6.

(4 and 5) y S and O = y , S , and 0


l9 l9 x 2 2 2 from previous
time step (columns 15, 16, and 17).
/ S2 γ/ ·
1 7268
ι 15.81 \ν 0.5791
Λ + h °\ + o 3.218.
2

(6) Δ5 = - 2 Τ ^ Δ / - - ^ Τ ^ Δ /
(17) 0 = 7(3.218) ' 0 498
= 12.528.
/ 12.08 + 12.08 \ 10 2

= 3.93 - — = 1.91. acre-ft 1 hr


\ 2 /60 (18) 0 = 12.528 — Χ
(7) S = S + AS = 13.93 + 1.91 = 15.84.
2
hr 3600 sec
2 x

43560 ft 3

I S 2 ^Z 1 7 2 6 8
/ 15.84 \ 0 5 7 9 1

= 151.6 cfs.
acre-ft

(9) 0 2 = 7(y ) 2
0 4 9 8
= 7(3.22) · 0 498
= 12.54.
/ι +' I 2 0 ι + 0 2
O t h e r numerical techniques can b e used to solve the
(10) AS = At Δί continuity equation as it applies to reservoir routing.
v
2 2 A s in all hydrologic problems, care must b e taken to
/ 12.08 + 12.54 \ 10
= 3.93 - — = 1.878. e n s u r e that a consistent set of units is used. C o m p u t e r
\ 2 /60 programs to carry out t h e routing by numerical m e a n s
(11) Si = S + AS = 13.93 + 1.878 = 15.808.
1
are widely available. T h e s e p r o g r a m s generally require
inputs having specific units and are structured to inter­
/ S 2 \ 1 / 1 7 2 6 8
/ 15.81 \ 0 5 7 9 1

polate between points on user-supplied s t a g e - s t o r a g e


and s t a g e - d i s c h a r g e relationships.

Note at this point that the y calculated in column (12) 2

agrees with the y in column (8) and the solution could stop.
2
USES OF RESERVOIR ROUTING
We carry it one additional step.
Reservoirs a r e typically used in flood control so that
(13) 0 2 = 7(y ) 2
0 4 9 8
= 7(3.218) · 0 498
= 12.528. some p r e d e t e r m i n e d p e a k outflow r a t e is not exceeded,
to delay t h e flow so that high runoff rates from t h e
L +I 0 i + 0
2 2

controlled catchment d o not coincide with high flows


(14) Δ 5 = - ± - ^ Δ ί - - ^ Δί
from a second catchment, to detain water so t h a t
/ 12.08 + 12.53 \ 10 suspended particles may settle prior to the flow enter­
= 3 .93 - ( ) - = 1.879. ing a stream, or for a water-use function such as
recreation or water supply. Regardless of the use of
(15) 5 = $! + AS = 13.93 + 1.879 = 15.81.
2
t h e reservoir, routing is used to d e t e r m i n e the impact
ο on «ν
Κ RO
On —« 3 Κ NO CO Ο

^ T f v O O N r r O r - < N
CN 5 ρ -Η -Η ο ON γ~* co ON
Si co co co' co cn cn cn -Η"
Ο

00 On cn 00 On «ο
Ν α no θ\ cn co Tf «η «η
Ο ΓΟ Γ- cs cn en co co co co CO CO CO CO
Ο -Η' ~

73
η oo oo
Ε
^ c£ «ο
ο —• —• CN —' —

CS 5 S S 2 £ 8 3
C N C N C N C O C O C O C O C N C N C N - H
D -I
ο 3

CO —* On oo On CN 00
-H m on co r- «η «η >ο ON
Ο © © —« -ί CN cs cs CO CO CO CO CO CO CO CO
co cn

3
CO Ό CO — Ο
R ^ CO
ο -ί iri co'

73
—< οoo CO
ο
<? 7 Τ 7

Γ- ο VO Γ-
—;
oo
^
On
«η
Ο
Tt
-Η'
S 2 2 8 3 g NO
CO
—>
On
Ν (S Ν CN* —<'

Ο —· -Η

8 ? S α
Ν©

no
\5
CO
r»;
Q
TF
«-Η

νο
ο ο ο CO 00 -Η
OO OO 00 OO

73
CO CO CO Ο «Ο

Ε —Η CO —' Γ"* CO Γ** —' ΙΟ 00


CN — -Η
^
Ο

©
CN
-Η'
Γ"*
-Η'
«? <? 7 7 7

8 8 2 τ*
"<t oo
p m
>η *t ^Ο 2 2 S 3
oo
-Η CN* CN CN*CO
Ο Ο co

8 CN
CN
ON
νο
OO
«η


ο ο no «ri
§
Ω
8 8 cs η- on

I
3
Ο Ο
Ν
Ο
on
©
co r-»
~ co co co co

Γ**· NO Ό CO CN -Η
ο CN 0O Ο Ν© CN
CN CN CO CO CO CN CN ~ -J — © Ο

ε
° S § 00 S 8 8 CO 8 g S CN2 —· — S 8
00 —« CN CO CO CO

Φ
3
Ό ε
8 8 2
198 6. Channel Flow Routing and Reservoir Hydraulics

of t h e reservoir on hydrographs; d e t e r m i n e reservoir bypassed by t h e flow. In designing a sediment basin, it


design requirements, maximum water levels, and outlet is desireable to minimize d e a d storage so that the
hydraulic requirements; and in some cases to provide entire volume of the basin is effective in retarding the
guidance in the operation of spillway gates and valves flow and allowing sediment to settle from suspension.
to m e e t downstream flow requirements for flood con­ T h e theoretical detention time in a rectangular basin
trol or low flow augmentation. with a constant flow rate is simply the average flow
through time given by
Flood Peak Reduction
O n e of the most common and important uses of (6.31)
" Q/A'
reservoirs is to reduce flood peaks. Based on t h e inflow
hydrograph and the allowable downstream p e a k flow,
w h e r e L is the flow length, Q is the constant dis­
the storage requirements and outlet hydraulic charac­
charge, and A is the cross-sectional flow area.
teristics can be determined. For small reservoirs with
F o r flows that vary in time, t h e definition of deten­
noncontrolled outlets, an exact design may not b e
tion time is m o r e complex. O n e definition is based on
achieved because of site characteristics and available
t h e plug flow or first in-first out concept. T o use the
pipe sizes for spillways. SCS (1986) T R 5 5 contains
plug flow concept, the inflow hydrograph is divided
some general guidelines. T h e general steps in the de­
into plugs of equal volume as shown in Fig. 6.11. T h e s e
sign p r o c e d u r e are:
plugs a r e r o u t e d through t h e reservoir, and t h e deten­
1. D e t e r m i n e the inflow hydrograph for t h e design tion time for each plug is d e t e r m i n e d . T h e theoretical
event. detention time is the average detention time of the
2. D e t e r m i n e t h e allowable peak outflow from the plugs. Obviously, plugs during the initial part of the
reservoir based on downstream conditions or governing flow have a d e t e n t i o n time shorter than the average
regulations. and t h e later plugs have a detention time greater than
3. Estimate t h e storage requirement by t h e average. Calculations based on the plug flow con­
a. plotting t h e allowable peak outflow on t h e cept are illustrated in c h a p t e r 9.
falling limb of t h e inflow hydrograph, A n alternative definition of detention time is based
b. approximating the outflow hydrograph u p to on t h e centroid concept. Theoretical detention time is
t h e peak outflow as either a straight line or an defined as the time difference between the centroids of
estimated curved line, and t h e inflow and outflow hydrographs:
c. estimating the storage required as t h e volume
of flow between t h e inflow and outflow hydrographs TA — T MRT T M (6.32)
from the initiation of flow to the p e a k outflow rate.
4. Estimate t h e maximum stage from t h e s t a g e - s t o r ­ w h e r e T is t h e detention time and T and r
D are the
MI mo

age relationship. time to t h e centroids of the inflow and outflow hydro-


5. D e t e r m i n e t h e spillway hydraulic configuration
required to produce t h e peak flow with the available
storage.
6. R o u t e the inflow through the reservoir based on
this spillway.
7. Adjust the spillway characteristics and r e p e a t step
6 until a satisfactory design is achieved.
- Inflow hydrograph

Detention Storage
- Detention time
Detention storage time is a measure of the average
residence time of water in a reservoir. It is a useful
m e a s u r e of t h e average opportunity time for sediment Plug volume (shaded area)
to settle out of flow. T h e actual detention time and the
Outflow hydrograph
theoretical detention time may differ because of d e a d
storage in the reservoir. D e a d storage represents the
volume of water in the reservoir that is not displaced 12 16
during t h e passage of flow through the reservoir. D e a d Time ( H R )
storage is found in areas of the reservoir that are Figure 6.11 Plug flow concept.
Uses of Reservoir Routing 199

Table 6.9 Computations for Example Problem 6.8

(b) q · q,(tk-t>/t. t / 0 tl tO
(min) (cfs) (cfs) (cfs-min) (cfs-min)

0 0 0 0 0
10 48 6 480 60
20 88 21 1760 420
30 100 37 3000 1110
40 92 50 3680 2000
50 76 58 3800 2900
60 59 60 3540 3600
70 43 58 3010 4060
80 31 54 2480 4320
90 21 48 1890 4320
100 14 41 1400 4100
110 10 34 1100 3740
120 6 28 720 3360
130 4 23 520 2990
140 3 18 420 2520
150 2 14 300 2100
160 1 11 160 1760
170 1 9 170 1530
180 0 7 0 1260
190 0 5 0 950
200 0 4 0 800
Figure 6.12 Triangular hydrograph approximations.
210 0 3 0 630
220 0 2 0 440
graphs, respectively. T m can b e calculated from 230 0 2 0 460
240 0 1 0 240
jtq(t)dt
250 0 1 0 250
Τ = (6.33)
jq(t)dt
Totals 599 595 28,430 49,920

where t is the time from t h e beginning of t h e hydro-


graph and q(t) is the flow rate at time t. A numerical T = 48 84 Γ =
α
36
m min min min
approximation for T is m

Τ = (6.34)
x
m Σ 9,
where t and q represent hydrograph time a n d flow
t {

values. T h e summations must b e carried until q is


Example Problem 6.8
i

negligible. Detention time


Designing a reservoir to have a specified detention
The inflow and outflow hydrographs shown in Table 6.9
time can be quite tedious. Trial outlet structures a n d a
are the result of flow from a 100-acre catchment. Compute
routing computating for each trial would have to b e the detention time for the reservoir.
examined until an outlet that would provide t h e re­
quired detention time could b e found. Using a triangu­ Solution: Equation (6.34) is used to calculate the time to
lar hydrograph approximation, it is possible to obtain the centroid of the inflow and outflow hydrographs. The
computations are shown in Table 6.9. Using Eq. (6.34),
very good estimates of the required design p a r a m e t e r s ,
which can then be refined if necessary based on d e ­
T = 28430/599 = 48 min
tailed routings. T h e required relationships are shown mi

in Fig. 6.12. Tmo = 49920/595 = 84 min.


200 6. Channel Flow Routing and Reservoir Hydraulics
From Eq. (6.32), 120

^ - ^ - 7 ^ - 84 - 48 - 36 min.

Example Problem 6.9 Approximate detention time

Estimate the detention time for the situation of Example


Problem 6.8 using the triangular approximations.
Solution (refer to Fig. 6.12):

q pi = 100 cfs * 1 iph


= 60 cfs « 0.6 iph.

The volume of flow is found from Eq. (3.46).

Δ/Σ<?« 10(599) 50 100 150 200 250


V = = 0.99 in. Time (min)
60.5/1 60.5(100)
Figure 6.13 Hydrographs for Example Problem 6.9.
The base times, f , for the triangular hydrographs are found
b

from Equation (c)

2V 2(.99)
= 1.98 hr required.
Ϊ
Solution: Use the triangular approximation. From Eq. (f) of
2V 2(.99) Fig. 6 . 1 2 ,
'bo« = 3.30 hr.
0.6 'mo = 'd + 'mi = 3 . 0 0 + 0 . 8 3 = 3 . 8 3 hr.
Equation (d) gives From Eq. (e),

'b + 'p 1.98 + 0 . 5


= 0.83 hr

for the inflow hydrograph. Equation (h) gives / 2 - 0.5 \ / 3 X 3.83 - 2 \

("ππ"Κ (
— Ϊ
— Κ" + 2= 0
4po 0.6
'PO - t - / - £ = 1.98 - 1 . 4 8 — = 1.09 hr.
bi r i

4 i P i.o 1.5^ + 9 . 5 ^ - 2 = 0.

Equation (d) gives This quadratic equation has as a solution q^ = 0 . 2 0 4 iph


or 2 0 cfs. The required storage is calculated from Eq. (i) as
'b + 'p 3.30 + 1.09
= 1.46 hr.
mo 3 3
S - t (q
bi pi - q )/2
po = 2 ( 1 - 0 . 2 0 4 ) / 2 = 0 . 8 in.

Thus the detention time from Eq. (6.32) is The required storage is about 0 . 8 in. or 6 . 6 7 acre-ft.
Based on the stage-storage curve for the reservoir, the
'd = 'mo - 'mi = 1 . 4 6 - 0 . 8 3 = 0 . 6 3 hr = 38 min,
stage corresponding to 0 . 8 in. of storage can be determined.
which agrees very well with the result obtained using the A principle spillway that will discharge 2 0 cfs at this stage
actual hydrographs. can be selected. A detailed routing of the inflow hydrograph
Figure 6 . 1 3 shows the actual hydrographs and the triangu­ can be done, and the actual detention time calculated. If the
lar approximations. The figure illustrates why the triangular detention time is not close enough to 3 hr, the design can be
estimate is a reasonably good one. modified as needed. Generally the triangular approximations
will produce very good first estimates of the required final
design.

Example Problem 6.10 Detention storage design

Assuming that a detention time of 3 8 min is inadequate,


Storm Water Management
estimate the storage required and the peak outflow rate for
the situation of Example Problem 6.8 if a detention time of D e t e n t i o n basins are often used in storm water m a n ­
3 hr is required. Obviously, a new outflow hydrograph is agement systems to reduce p e a k flows. Often, regula-
Uses of Reservoir Routing 201

in t h e undeveloped state, in t h e developed state with­


out detention storage, and in the developed state with
d e t e n t i o n storage. Also shown is the hydrograph from
catchment B. N o t e that t h e p e a k of the hydrograph
from catchment Β occurs after t h e p e a k from catch­
m e n t A. Figure 6.15d shows the combined hydrograph
from catchments A a n d B . Both the situation where
t h e hydrograph from catchment A is routed through a
d e t e n t i o n basin a n d w h e r e a d e t e n t i o n basis is not
p r e s e n t are shown.

I A significant finding is t h a t by not controlling the


flow from catchment A , t h e flow from that catchment
is allowed t o pass from t h e system prior t o the arrival
of t h e high flows from c a t c h m e n t B. W i t h detention
storage, the flow from catchment A is delayed and
caused to coincide with t h e flows from catchment B.
T h e n e t result is t h a t without detention storage in
point C " sf
c a t c h m e n t A , t h e total flow at point D is n o greater
point D t h a n it is with d e t e n t i o n storage. T h u s the expense and
Figure 6.14 Catchments A and Β with catchment A set to undergo legal liability associated with t h e detention basin is
a major land-use change. largely wasted as far as flow at point D is concerned.
T h e p u r p o s e of this discussion is to point out that
piecemeal installation of d e t e n t i o n basins without con­
tory authorities require that any land-use changes must sideration of how they fit into an overall storm water
be accompanied by measures that ensure that the peak m a n a g e m e n t system may not contribute to p e a k flow
discharge from a design storm after land-use changes reduction at a d o w n s t r e a m location even though they
will not exceed the peak discharge for the same event r e d u c e p e a k flows immediately below and on the same
prior to the changes. Often this results in t h e require­ stream as t h e d e t e n t i o n basin. Storm water manage­
ment of detention storage being included in the devel­ m e n t systems must b e considered o n a regional basis
opment plans. For discussion purposes, consider that a n d t h e placement of d e t e n t i o n storage optimized from
watershed A in Fig. 6.14 is going to change from a a regional standpoint to obtain maximum storm water
wooded area to an intensive urban development. Based control p e r dollar of expenditure.
o n considerations in C h a p t e r 3, it would b e expected A further complication in the design of detention
that the runoff hydrograph after the development would storage arises if t h e d e t e n t i o n basin does not empty to
have a greater p e a k flow, a shorter time to peak, and a the spillway level b e t w e e n major runoff events. If this is
larger volume than the hydrograph for the same event the case, a continuous simulation hydrologic model
prior to development. Figure 6.15a shows t h e two may b e required to c o m p u t e a continuous water bal­
hydrographs of concern. Obviously the development ance o n t h e basin so that t h e p r o p e r initial stage can b e
has increased the peak flow from catchment A. T o used in routing a storm water runoff event through the
control this increase in runoff, a storm water detention basin. Hydrologic models a r e discussed in more detail
basin can be installed n e a r the outlet of the watershed. in C h a p t e r 13. A p r o p e r w a t e r balance on the deten­
If properly designed, it is possible to have an outflow tion basin would have t o consider evaporation and
hydrograph from the basin with a peak discharge equal seepage from the basin.
to the peak flow from the catchment prior to its devel­
opment. Figure 6.15b shows the predevelopment and
routed post development hydrographs with t h e same Problems
peak discharge. T h e runoff volume is greater after the (6.1) A storm water drainage channel has b e e n con­
development and is unaffected by the detention basin. structed through an u r b a n area. T h e channel is trape­
If flow as it leaves catchment A is the primary zoidal in shape with a 15-ft bottom width, 1:1 side
concern, the situation depicted in Fig. 6.15b could slopes, a slope of 0 . 3 % a n d a Manning's η of 0.02. A t
represent a satisfactory final design. If, however, flow the u p s t r e a m e n d of t h e channel, t h e inflow hydro-
at point D , the combined outlet of catchments A and Β graph has a p e a k flow rate of 750 cfs and a time to
is of concern, further analysis is required. Figure 6.15c p e a k of 1 hr. T h e hydrograph may be described using
shows the runoff hydrograph produced by catchment A Eq. (3.59). T h e total volume of runoff is 1.8 in. from
202 6. Channel Flow Routing and Reservoir Hydraulics

the 500-acre drainage area. R o u t e t h e hydrograph (6.4) Using t h e triangular approximations of Fig.
through a 2000-ft segment of the channel. 6.12, estimate the detention time of flow of problem
(a) U s e storage routing (6.3).
(b) U s e M u s k i n g u m - C u n g e routing (6.5) C o m p u t e t h e detention time of the flow of
(c) U s e convex rooting problem (6.3) from E q s . (6.32) and (6.34). C o m p a r e the
(d) U s e kinematic routing. results with t h e results of problem (6.4).
(6.2) R o u t e the hydrograph of problem (6.1) through (6.6) Estimate a in t h e outflow relationship q =
2000 ft of a trapezoidal channel having a 10-ft bot­ ahl/1
to give a detention time of 2.5 hr for t h e situa­
tom width, 2 : 1 side slopes, a slope of 0.2%, and a tion described in problem (6.3). W h a t is the required
Manning's η of 0.040. storage? U s e t h e triangular approximation.
(6.3) A reservoir has a stage-discharge relationship (6.7) U s e a detailed routing to check your solution
given by the equation q = 200Λ and a s t a g e - s t o r a g e
2
to problem (6.6). W h a t is t h e actual detention time as
curve given by S = 20Λ , where q is cfs, h is in feet,
2
defined by Eqs. (6.32) a n d (6.34)?
and 5 is in acre-feet. T h e inflow hydrograph may (6.8) A 0.5-mile catchment has t h e following runoff
2

b e described by Eq. (3.59) with a p e a k flow r a t e of characteristics from a design storm prior to and follow­
1500 cfs from t h e 2 m i l e drainage area and a time to
2
ing development. Estimate t h e storage required for a
p e a k of 75 min. T h e volume of runoff is 2.4 in. Esti­ flood control reservoir to limit t h e postdevelopment
m a t e the outflow hydrograph. p e a k outflow to t h e predevelopment.
Uses of Reservoir Routing 203

Predevelopment Postdevelopment References


Brakensiek, D . L. (1966). Storage flood routing without coefficients,
Runoff volume 3.0 in. 4.0 in.
U.S. Department of Agriculture Publication A R S 41-122.
Peak discharge 400 cfs 800 cfs Chow, V. T. (1959). " O p e n Channel Hydraulics." McGraw-Hill,
Time to peak 45 min 30 min N e w York.
Chow, V. T., Maidments, D . R., and Mays, L. W. (1988). "Applied
Hydrology." M c G r a w - H i l l , N e w York.
(6.9) A potential reservoir site has b e e n surveyed Cunge, J. A . (1967). O n the subject of flood propagation method.
with the results shown below. Calculate and plot the /. Hydraul. Res. IAHR 7(2):205-230.
Fread, D . L. (1977). The development and testing of a dam-break
s t a g e - s t o r a g e curve using feet and acre-feet. A s s u m e
flood forecasting model. In "Dam-Break Flood Modeling Work­
that the storage below 410 ft is zero. shop, U.S. Water Resources Council, Washington, D C . "
Fread, D . L. (1978). National Weather Service Operational Dynamic
Wave Model: Verification of mathematical and physical models
in hydraulic engineering. In "Proceedings 26th Annual Hy­
Elevation Area (acres)
draulics Division Special Conference, A S C E , College Park, M D . "
Fread, D . L. (1985). Channel routing. In "Hydrological Forecasting"
410 0
(M. G. Anderson and T. P. Brent, eds.). Wiley, N e w York.
411 1 Henderson, F. M. (1966). " O p e n Channel Flow." MacMillan, New
412 4 York.
Soil Conservation Service (1986). Urban hydrology for small water­
413 8
sheds. Technical Release N o . 55. Soil Conservation Service, U.S.
414 10 Department of Agriculture, Washington, D C .
415 13 U.S. Department of Agriculture (1971). Hydrology. In "SCS Na­
tional Engineering Handbook," Sect. 4. Soil Conservation Ser­
416 15
vice, Washington, D C .
417 20 Viessmann, W., Lewis, G. L., and Knapp, J. W. (1989). "Introduction
to Hydrology," 3rd ed. Harper & Row, N e w York.
Sediment Properties
and Transport

INTRODUCTION BASIC PRINCIPLES OF SEDIMENTATION

Sediment is composed of many materials, including Types of Settling


individual primary particles, aggregates, organic m a t e ­ Particle settling has b e e n divided into four types
rials, and associated chemicals. Sediment properties based on concentration and the tendency of particles
impact how each individual or aggregate particle be­ to hinder each other:
haves in flowing water. Size, shape, and density affect
the settling velocity, which in turn affects sediment Discrete particle settling (Type 1) refers to settling in
transport rates and at what points particles deposit. low-concentration solutions in which particles tend to
Sediment is considered to be fully characterized w h e n fall independently of each other, i.e., as discrete parti­
its shape, size, density, constituent texture, mineralogy, cles.
and stability are known (Martin et al., 1955). Flocculent settling (Type 2) refers to settling in which
Chemical properties of a particle d e t e r m i n e if pri­ dilute solutions of particles coalesce and form particles
mary particles will aggregate and form larger particles of larger mass and higher settling velocities.
with a different shape and lighter densities. Densities Hindered or zone settling (Type 3) refers to settling
of the aggregates are typically significantly less than in which particles are so concentrated that forces b e ­
that of primary particles ( R h o t o n et al., 1983). Chemi­ tween particles interfere (or hinder) the settling of
cal properties of a s e d i m e n t - w a t e r mixture may either surrounding particles. This results in uniform settling
encourage flocculation or cause dispersion of the ag­ of all particles a n d is characterized by a distinct
gregates. Organic matter, iron oxides, and carbonates liquid-solid interface at the top of the solid mass. T h e
can contribute to bind aggregates (Martin et al., 1955). relatively sharp definition at the interface has led to
Some of the properties discussed above are beyond t h e title zone settling.
the scope of this text and are appropriate for texts on Compression settling (Type 4) refers to settling in
soil chemistry and mineralogy. In this chapter, e m p h a ­ which the particles are so concentrated that interparti-
sis is placed on properties related to particle settling cle bridging has formed a stable structure requiring
and sediment transport. compression for further settling. T h e force causing

204
Basic Principles of Sedimentation 205

τ — π — π — π — π ~

CL
Ixl
a
\ SPHERES \J
10 J U L_L J U Ϋ ,I V
10 3
10 2
10 1
1 10 10 2
10 3
10* 1 0 5
10 6

Re

Figure 7.2 Drag coefficients versus Reynold's number for spheres


and disks from experimental data (adapted from Rouse, 1950).

TIME / \
CYLINDER

Figure 7.1 Four types of settling (after Tapp et al. 1981). y


w h e r e ν is kinematic viscosity. E q u a t i o n (7.3) is gener­
ally considered to be valid u p to a value of R equal to e

0.5.
F o r t h e Stokes' range (i.e., R < 0.5), Eqs. (7.1) and
compression results from the weight of additional par­
e

(7.2) can b e simplified to yield


ticles settling on t o p of t h e settling mass.
1 dg 2

T a p p et al (1981) indicate that all four of these V = — — (SG-1) (7.4)


settling types can occur simultaneously in a sediment
s
18
pond. T h e zones (Fig. 7.1) do not occur instantly, but w h e r e S G is t h e specific gravity of t h e particles.
begin to form at varying times. Since each of these F o r nonspherical particles a n d for fall velocities out­
types of settling can occur in a sediment pond, a side t h e Stokes' range, experimental data must be
detailed discussion of each type is given. relied on. A c o m m o n m e t h o d is to use a curve such as
that shown in Fig. 7.2 for C to calculate fall veloci­
Discrete Particle Settling (Type 1)
D

ties. T h e results define fall velocities of various diame­


A particle falling in turbulent free water moves in ters settling in water at specified t e m p e r a t u r e s as shown
response to the difference in the submerged weight of in Fig. 7.3. E q u a t i o n (7.4) simplifies for S G = 2.65 and
the particle and t h e drag of the fluid on the particle. At
steady state, the forces are in equilibrium as described
by

ird2

pVs 2
πά 3

(p -p)g,
s (7.1)

where C is the drag coefficient, which is normally a


D

function of Reynold's number, d is particle diameter,


p is particle density, ρ is fluid density, V is settling
s s

velocity of t h e particle, a n d g is acceleration of gravity.


Stokes showed that the coefficient of drag for spheres
depends on Reynold's number, or

C D = 24//? , e (7.2)
.005 .01

where i ? is the Reynold's n u m b e r given by


e V.(fps)

Figure 7.3 Sedimentation diameter versus fall velocity in water


* e = Kd/v, (7.3) assuming a specific gravity of 2.65 (adapted from Rouse, 1950).
206 7. Sediment Properties and Transport
quiescent water at 68° F to become acceptable m e t h o d for accounting for turbulence is
presently available. Graf (1971) presents further discus­
K = 2.8W ,
s
2
(7.5) sion of these concepts.
where V is given in feet per second and d is in
s

millimeters. Flocculation. Flocculation results in formation of ag­


T h e settling velocity of large particles can b e esti­ gregates through t h e satisfaction of surface charges on
m a t e d by a Lagrangian polynomial that is fitted through the sediment. T h e effect is very similar to that of
t h r e e points of a sedimentation curve such as that in aggregation in that particles formed have a specific
Fig. 7.3. Wilson et al. (1982) presented on analysis that gravity less t h a n that of individual primary particles.
is similar to the approach used to d e t e r m i n e Stokes' Flocculation is discussed further in a subsequent sec­
settling velocity. This analysis evaluated the drag coef­ tion.
ficient, C , using experimental d a t a for Reynold's
D Particle settling velocities for type 1 settling are
n u m b e r s larger that 0.5. T h e results were typically d e t e r m i n e d b o t h experimentally a n d theoreti­
cally. T h e settling velocities of sand-sized and larger
log ίο Κ = - 0 . 3 4 2 4 6 ( l o g 10 d) 2
+ 0.98912 log 1 0 d particles are usually d e t e r m i n e d by using the size of
+ 1.14613 [cm/sec] (7.6a) sieve openings and calculating a corresponding settling
velocity using Stokes' law (size distribution classifica­
or tions are p r e s e n t e d in a later section). For silt and clay
size particles, t h e size distribution must be taken by a
log V = -0.34246(log d)2
+ 0.98912 l o g d
10 s 10 10
settling analysis using a hydrometer or pipette or by
- 0.33801 [ft/sec], (7.6b) m o r e recently available electronic instrumentation. If
only t h e primary particle size distribution is to be
where d is the diameter of the particle in millimeters obtained, a hydrometer is a d e q u a t e since particles are
and V is the particle settling velocity in centimeters
s
dispersed a n d t h e specific gravity is known with a
per second or feet p e r second, respectively. Factors relatively high d e g r e e of confidence. If an analysis of
that affect fall velocity of discrete particles include t h e the particle size distribution of a mixture of aggregates
following. a n d primary particles is being m a d e , it is necessary to
use t h e pipette m e t h o d and calculate equivalent fall
Particle shape. Particle shape affects t h e drag coef­ velocities since t h e specific gravity is not known. Since
ficient, which in turn affects the fall velocity. This is e r o d e d sediment is typically composed of a combina­
normally accounted for by classifying particles accord­ tion of primary a n d aggregated particles, a pipette
ing to an equivalent fall diameter. F u r t h e r information analysis is preferred. Obviously, a dispersing agent is
is contained in Graf (1971) and Simons and Senturk not used in an aggregate analysis since this would
(1977, 1992). result in dispersion of aggregates in the sample. It
should b e recognized that t h e use of a dispersing agent
Aggregation. Aggregation effectively alters t h e spe­ for primary particles is t h e major difference in tech­
cific gravity of soil particles. Since an aggregate con­ nique that separates determination of aggregate parti­
tains p o r e space, its specific gravity is less than that of cle size distributions from determination of primary
the primary particles from which it was formed. Al­ particle size distributions. Details on hydrometer and
though aggregation decreases the specific gravity, the p i p e t t e analyses can be found in standard texts dealing
diameter of an aggregated particle is so much greater with soil physics and soil mechanics. T h e U.S. D e p a r t ­
than its primary particle components that an aggre­ m e n t of Agriculture ( U S D A , 1979) presents substantial
gated particle typically has a higher settling velocity discussion of sieve, pipette, and hydrometer analyses.
than its component primary particles. Site-specific fall
velocity data should b e collected if possible.

Turbulence. T h e effects of turbulence on settling Example Problem 7.1. Settling velocity


velocity are somewhat mixed. Turbulence reduces of discrete particles
the drag coefficient C by changing the boundary
D

layer around the particle from laminar to turbulent. Compare the settling velocities using Eqs. (7.4) and (7.6)
For smaller particles, this is offset by the tendency of for 0.0002-, 0.002-, 0.02-, and 2.0-mm particles if SG = 2.65
turbulence to diffuse the particles from a zone of for all particles and settling occurs in water at 68° F.
higher concentration n e a r the bottom to a zone of Solution: Solving for the middle size (0.02 mm) first, values
lower concentration n e a r the surface. N o completely for particle diameter (0.002 cm), gravitational constant (980
Basic Principles of Sedimentation 207
c m / s e c ) , kinematic viscosity at 68° F (0.01003 cm /sec)
2 2
sedimentation basin, velocity gradients in t h e basin,
(from the General Appendices), and particle specific gravity concentration of particles, particle size, and interparti-
are substituted into Eq. (7.4): cle forces that cause particles t o repel each other.
Studies associated with surface mine sedimentation
1 dg 2

V = — — (SG-1) indicate that flocculation can b e a naturally occurring


18
s

p h e n o m e n a , d e p e n d i n g on the chemistry of the water


1 0.002 cm X 980 c m / s e c
2 2 2
a n d the mineralogy of t h e clay particles ( T a p p et al.,
(2.65 - 1) 1981; Evangelou et al., 1981). Evangelou et al. also
Is 0.01003 c m / s e c 2

indicate that dispersion is a possibility u n d e r certain


= 0.0358 cm/sec. conditions. Dispersion may b r e a k t h e aggregated parti­
cles down into primary particles, thus decreasing their
Similarly, for a particle diameter of 0.02 cm, V is 3.58 s

cm/sec, and for a diameter of 0.2 cm, V is 358 cm/sec. The settling velocity. T h e flocculation p h e n o m e n a are dis­
s

settling velocities for 0.02- and 0.2-cm particles would be cussed further in this section, a n d t h e dispersion p h e ­
outside the Stokes' range, since Reynold's numbers would be n o m e n a are delayed to a subsequent section of this
chapter.
(3.58)(0.02)
= 7.138 T o begin a discussion of the flocculation process,
0.01003 consider a stabilized colloidal suspension. This suspen­
for the 0.02-cm particle and 7138 for the 0.2-cm particle, sion consists of dispersed or individual colloidal parti­
using the calculated velocities. cles with a surface charge s e p a r a t e d at considerable
Using the Wilson et al. (1982) relationship [Eq. (7.6)] and a distances in an a q u e o u s solution. F o r clay particles, the
particle diameter of 0.02 mm, surface is negatively charged. Dispersed particles are
said to be stabilized since they repel each other. For
logK = - 0 . 3 4 2 5 ( l o g < / ) + 0.98912 l o g d + 1.14613
s l0 p
2
l0 p flocculation t o occur, particles must b e destabilized
and brought together to allow contact and flocculation.
logK = - 0 . 3 4 2 5 ( l o g 0 . 0 2 ) + 0.98912(log
s 10
2
10 0.02)
T h r e e theories are often used to describe t h e p r o ­
+ 1.14613
cess of destabilization—the double electric layer the­
logK = -1.522
s
ory, chemical bridging, and p H . T h e double electric
z p c

K = 0.03 cm/sec.
s layer theory considers only electrostatic attraction and
diffusion. In this theory, t h e largest concentration of
The final estimated settling velocities for 0.00002-cm
c a t i o n s is n e a r t h e surface of the colloidal particle
1

(0.0002-mm), 0.0002-cm (0.002-mm), 0.002-cm (0.02-mm),


0.02-cm (0.2-mm), and 0.2-cm (2.0-mm)-diameter particles, because of t h e electrical attraction of the negative
respectively, are: charge on the colloid. T h e concentration of cations
decreases with distance away from t h e colloidal surface
Vs = 3.58 X 1 0 " cm/sec,6
using Stokes' law [Eq. (7.5)] as shown schematically in Fig. 7.4. T h e resulting con­
V = 3.58 x 1 0 " cm/sec,4
using Stokes' law
s
centration gradient of cations causes a gradient of
Vs = 3.58 X 1 0 " cm/sec,2
using Stokes' law
forces t h a t repel o t h e r clay particles that are sur­
Ks = 1.94 cm/sec, using Eq. (7.6a)
V = 27.79 cm/sec, using Eq. (7.6a). r o u n d e d by positively charged ions as shown in Fig. 7.5.
s

T h e layer in which the concentration of counterions is


Values predicted from Eq. (7.6) agree favorably with val­ increased is called t h e diffuse layer (See Fig. 7.4). As
ues from Fig. 7.3 and differ substantially from Stokes' law. the concentration of ions in the a q u e o u s phase in­
This difference demonstrates the influence of the drag coef­
creases, the thickness of the diffuse layer decreases,
ficient on the settling velocity at Reynold's number greater
causing t h e m a g n i t u d e of the repulsive forces to de­
than 0.5. Equation (7.6) would be appropriate for the larger
particles. When using Stokes' law, a check to assure that the crease m o r e rapidly at intermediate distances from t h e
Reynold's number is less than 0.5 should always be made. surface of the colloidal particle. T h u s the forces of
repulsion d e p e n d on b o t h t h e solid a n d aqueous phases.
T h e forces of attraction between colloidal particles
result from van d e r Waals energy, shown schematically
Flocculent Settling (Type 2)
in Fig. 7.5, which d e p e n d s on the mass of the particle.
Colloidal particles in a dilute suspension will some­ T h e sum of t h e energy levels of repulsion and attrac­
times settle as discrete particles a n d sometimes coa­ tion is the resultant energy curve. If the concentration
lesce and form "floes" of n u m e r o u s particles, which of counterions in the a q u e o u s solution is low, the
have sufficient mass to settle rapidly. Metcalf and Eddy
(1979) reported that the degree of flocculation d e p e n d s
on the opportunity for particle contact, d e p t h of the Positively charged ions, sometimes called counterions.
208 7. Sediment Properties and Transport

PARTICLE SOLUTION BULK REACTION 1


INFTAL ADSORPTION (OPTIMUM POLYMER DOSAGE)
Q (FAST)
DESTAB1UZED
POLYMER COLL0DIAL PARTICLE PARTICLE
^ B REACTION 2A ^

V^T^N-. rffs RATE DEPENDENT ^ ^n^W


V \U UPON MASS J£S
DESTABILIZED PARTICLE 1 1 W < S P 0 R r
FLOC
REACTION 2 B >\%/

^
SECONDARY
REACTION ADSORPTION
3 LjJ
IONS WITH SAME CHARGE vu ι ivy η %j *
AS COLLOIDAL PRAT1CLE INITIAL ADSORPTION
.__ <RESTABIUZED
Μ^Τ*^
(EXCESS
XSS POLYMER DOSAGE) Λ Λ Λ
PARTICLE

EXCESS POLYMER OOUJOOIAL PARTICLE STABLE PARTICLE


REACTION
mON
RUPTURE 4A 1^ •
DISTANCE FROM PARTICLE S U R F A C E
RE OF FLOC J j + Jr
Figure 7.4 Schematic of electric double layer theory (after Tapp
et al., 1981). FLOC Η SHEAR Ji FLOC FRAGMENTS
(^J
REACTION 4 B
^JT ADSORPTION
SECONDARY AD

FLOC RESTABIUZED PRATtCLE

double layer thickness is high, and a potential energy 1) CHARGE NEUTRALIZATION 2) BRIDGING 3) COMBINATIONS
barrier exists, as shown in Fig. 7.5. A s the concentra­ Figure 7.6 Processes involved in chemical bridging (after Tapp
tion of counterions increases, the diffuse layer becomes et al, 1981).
thinner, reducing the forces of repulsion at intermedi­
ate distances. This ultimately eliminates the potential
energy barrier at sufficiently high concentrations. At
this point, the particle is destabilized and flocculation
can occur. Charge reversal can sometimes occur as a
result of overdosing when using a trivalent ion (three
positive charges) such as aluminum.
In chemical bridging theory, t h e second flocculation
mechanism, it is proposed that a polyelectrolyte (a
long-chain polymer with ionizable groups) can attach
REPULSIVE POTENTIAL ENERGY
CURVE DUE TO COUNTER IONS
itself to the surface of a colloidal particle at o n e or
m o r e absorption sites. Its length causes the polyelec­
trolyte to extend beyond the diffuse layer into t h e
RESULTANT INTERACTION solution. This extension can thus b e attached to vacant
ENERGY CURVE absorption sites on other colloidal particles causing
OT flocculation. A schematic of the processes involved is
given in Fig. 7.6. F u r t h e r discussion can b e found in
Metcalf and Eddy (1979).
< DISTANCE FROM Procedures for predicting changes in particle size

ϋ
SURFACE OF CLAY
PARTICLE d u e to flocculation are limited primarily to experimen­
ο
Q_ tal type relationships. T a p p et al. (1981) summarizes
the equations developed by various researchers. T h e
VAN DER WAALS ATTRACTIVE most promising p r o c e d u r e is t h a t p r o p o s e d by
ENERGY CURVE A r g a m a n and Kaufman (1970). They postulate t h a t
floes grow d u e to collisions between u n r e a c t e d colloids
and floes that are simultaneously coming apart d u e to
Figure 7.5 Energy levels of attraction and repulsion near a col­
shear forces resulting from the turbulence of the flow.
loidal particle (after Tapp et al., 1981). If n represents the n u m b e r of colloidal particles avail-
x
Basic Principles of Sedimentation 209
able for flocculation, then PORTS FOR
SAMPLING
^ = 100%
dnx rate of flocculation
1Γ of primary particles

rate of b r e a k u p of
+ (7.7a)
floes to primary particles

or
dn x

= H lF + B .
XT (7.7b)
~~dt

A r g a m a n and Kaufman (1970) proposed that the rate


of flocculation could b e given by

H lF = -^K aR n n [u ] ,s
3
F x F
,2
a (7.7c)

where K is a p a r a m e t e r that relates the effectiveness


s

of t h e m e a n square turbulent velocity, [ w ' ] , in mix­ 2 2


a
20 30
ing, α is a p a r a m e t e r that defines the probability of a TIME ( m i n )
colloidal particle permanently sticking to the floe after Figure 7.7 Application of settling tube for determining suspended
a collision, R is the radius of a floe, n is the n u m b e r
F x solids removal by flocculation settling (Haan and Barfield, 1978).
concentration of colloidal particles, and n is the n u m ­ F

ber concentration of floes. A r g a m a n and Kaufman also


propose that the rate of b r e a k u p is given by
In t h e absence of applicable procedures for predict­
ing floe growth, most designs of settling chambers and
BXT = B^n [u> ]„ F
2
(7.8) t r e a t m e n t facilities in which chemical flocculants are to
b e a d d e d are b a s e d on settling tube analyses. A settling
t u b e can b e constructed of material having convenient
where Β is a floe b r e a k u p constant and R is t h e
diameter, typically 3 - 1 6 in. A n example is shown in
x

radius of the colloidal particle. Combining the two


Fig. 7.7. Height of t h e tube should be equal to the
expressions yields
height of the settling chamber. A sample containing
sediment and flocculant is a d d e d to the tube; the
= ^-4TrK aR n n s
3
F x F + S^n j[ ' ] . F W
2
a (7.9) solution is thoroughly mixed for a uniform distribution;
a n d quiescent settling begins. After settling is initiated,
samples are t a k e n at various d e p t h s and time intervals
A r g a m a n and Kaufman (1970) found good agreement and analyzed for total suspended solids (TSS). These
between predicted and observed values of η in a x are c o m p a r e d to t h e initial concentration of TSS and
settling tank using alum as the coagulant with a sus­ the percentage of particles removed calculated for each
pension of clay colloids. T a p p et al. (1981) found that d e p t h at a given time. Percentage removal values are
the prediction accuracy for effluent suspended solids t h e n plotted as shown in Fig. 7.7. T h e percentage
concentration from a pilot size sediment p o n d was removal for a n ideal clarifier with n o turbulence is then
improved when Eq. (7.9) was used in conjunction with the average percentage removal at a time in the set­
various models of pond performance. Unfortunately, tling t u b e equal to detention time of t h e clarifier.
values for K and Β are not generally available for
s

different systems, and procedures for predicting [u ] ,2


a

are rather limited. Additional details of the T a p p et al.


study are discussed further in a subsequent section. Example Problem 7.2. Suspended solids
determination using a settling tube

From turbulent flow theory, u' is the deviation of the instantaneous A sediment pond is being designed to treat drainage that
velocity from the time average velocity at a point. The mean square is pumped from a mine. The estimated flow rate through the
velocity is thus the time average of u' and is a measure of the 2
system and surface area are such that the pond has a deten­
kinetic energy contained in the turbulence of flow. tion storage time of 2 0 min. A settling tube test is conducted
210 7. Sediment Properties and Transport
on the drainage from the mine while adding a commercial
flocculant. The resulting removal curve is shown in Fig. 7.7.
1.2
The depth of the pond is 5 ft. Calculate the removal effi­
ciency of the pond.
1.0
Solution: The percentage removal is calculated by weight­
ing the removal efficiency over given depth intervals, or
9 ^
Ml/D)tfR(y)dy = (l/Z))/?,Ay, where /?(y)is the removal
efficiency as a function of y and R is the average removal in LU
efficiency over the interval Ay,-. Using values from Fig. 7.7, u%
x u:
1 100 + 80 80 + 70
2 (0.8) + — — (0.4) oz
£
= 5
Ϊ 9 4
60 + 50 ο 3 Λ

70 + 60 • Ο ^sL COMPRESSION
(0.9) + — — ( 2 . 9 ) — -^-SETTLING

.21
Ε = 64%.

Obviously, the trapping efficiency is low. To improve the


pond's performance, the detention time would have to be 30 60 90
increased by increasing the pond size or a better flocculating
TIME ( m i n )
agent would have to be found.
Figure 7.8 Sample settling test for hindered and compression set­
tling.

Hindered or Zone Settling (Type 3)


A s t h e concentration of sediment increases in den­
until it reaches what might b e t e r m e d a short-term
sity, a point will be reached such that t h e settling of
constant value. This " c o n s t a n t " value continues t o d e ­
particles is hindered. T h e settling of particles displaces
crease, b u t at a slow rate. T h e short-term constant
water that flows upward through the openings between
density will d e p e n d on t h e m a k e u p of the sediment,
the particles. A s a result, particles tend t o settle as a
as well as t h e conditions of deposition. Lara and
blanket, with all of t h e particles maintaining t h e same
P e m b e r t o n (1963) suggested that t h e short-term (or
relative velocity. This is known as hindered settling.
initial) weight density W of deposited sediment can b e
H i n d e r e d settling would b e expected n e a r t h e bottom
given by
of a sediment pond.
Variability encountered with hindered settling is so W=W PC C + WP m m + WP, s s (7.10)
great that settling tests are typically required for a
w h e r e W , W , a n d W a r e unit weights of clay, slit, a n d
c m s
design. T h e s e tests would b e necessary if a system were
sand, respectively, a n d P , P , a n d P are t h e fractions
c m s
being designed for pumping the deposited sediment
of clay, silt, a n d sand, respectively. Values for W , W , c m
(which corresponds t o sludge in water t r e a t m e n t termi­
and W a r e given in Table 7.1. Miller (1953) developed
s
nology) from a pond. Metcalf and Eddy (1979) give an
a p r o c e d u r e for calculating how W, as calculated from
excellent discussion of the topic.
Eq. (7.10), changes with time or
Compression Settling (Type 4)
W =W+ 0.4343K (ΙηΓ-1) , (7.11)
After hindered settling continues for a long period T
Τ - 1
of time, a layer of particles with a definite structure
begins to form at the bottom of t h e basin or pond. w h e r e W is t h e value given by Eq. (7.10), Τ is time in
F u r t h e r settling of these particles results only from years, a n d Κ is given by
compression of the layer, thus the term compression K = BP + BP + B P, (7.12)
c c m m s

settling.
Compression settling, like hindered settling, is highly where B , B c m9 and B are coefficients given in Table
s

variable, usually requiring settling tests for satisfactory 7.1.


design information. These settling tests typically are
observations of the depth of the liquid-solid interface
Settleable Solids
with time. A n example is shown in Fig. 7.8.
A s can b e seen in Fig. 7.8, the height of the sludge Conceptually, total suspended solids can b e divided
or compressed sediment decays exponentially with time into settleable solids and those that are so small that
Basic Principles of Sedimentation 211

Table 7.1 Values of Unit Weight of Reservoir Sediments 0

Values of coefficients
inEq. (7.10) and (7.12)

Type of reservoir operation Κ W


m
m

I. Sediment always submerged 26 16.0 70 5.7 97 0


II. Moderate-considerable drawdown 35 8.4 71 1.8 97 0
III. Reservoir normally empty 40 0.0 72 0.0 97 0
IV. Riverbed sediments 60 0.0 73 0.0 97 0

a
W values from Lara and Pemberton (1963). Β values from Miller (1953).

ter d with a settling velocity V will settle through a


t S£

distance
DIMENSIONS OF (7.13)
IMH0FF CONE
D = 15.57" T h u s all particles of size will have settled o u t of
V = 1 Liter
T
zone V, in Fig. 7.9. T h e fraction of particles settling
r = ? into t h e solids zone is
a =14.64°
(7.14)

w h e r e V is t h e total volume of the Imhoff cone. From


T

geometry
7 7
* ο θ

(7.15)

Hence,

tan (g/2)y?2

F = (7.16)
rD 2

Solids If t h e velocity of a particle that will settle through D in


Figure 7.9 Definition sketch for settleable solids equations (after Τ is given by V a n d y, is VJ from Eq. (7.13), then D
0

Barfield et al., 1981).


becomes V T. D is also r / t a n ( 0 / 2 ) . Therefore,
0
2 2

ν 3
VT V
3 3 5

ρ — ___ — s
'
= 7 SI
^· (7-17)
W

they remain in suspension d u e t o Brownian motion. ' D 3


VT 3 3

Actually, settleable solids a r e defined as t h e volume of


Let ΔΧ b e t h e fraction of particles r e p r e s e n t e d by the
particles that settle in t h e bottom of a n Imhoff cone in έ

particles of size d . T h e n t h e mass of sediment trapped


1 h r of quiescent settling. T h e dimensions of a n Imhoff t

in t h e size range d is
cone a r e shown in Fig. 7.9. Since t h e displacement of t

an Imhoff cone is 1 liter, t h e units of settleable solids mass, = F CV AX (7.18)


i T i

are typically milliliters p e r liter.


T o convert TSS to settleable solids, t h e size distribu­ or
tion of t h e particles a n d their weight density must b e
tan 0 2
lV si V 3

used. T h e following derivation, developed by Barfield mass. ντ 2 2


— SI
et al. (1981) a n d Wilson et al. (1982), assumes that t h e r 2 8 1
V yη
volume of settleable solids is small relative t o t h e total (7.19)
volume. A n o t h e r assumption is that t h e particles a r e
uniformly distributed at t h e start of t h e test. A t t h e w h e r e C is t h e TSS concentration a n d V T is t h e total
end of a test of duration Τ hours, a particle of diame­ volume of the Imhoff cone.
212 7. Sediment Properties and Transport

T h e total mass trapped over all size ranges is ter can be solved from Eq. (7.5), or
* 1 / 2 /
0.00036 Y\ 1 / 2

, (7.20) 2.81 = 0.011 mm (fine silt).


V 3

Thus, one would expect to trap all of the fine silt and larger
particles. The size distribution for the sediment is given in
where η is the n u m b e r of size fractions of particles
Fig. 7.10. For diameter of 0.011 mm, X is 0.33. The follow­ Q
with a settling velocity less t h a n V . T h e total settleable0
ing data are tabulated for computing settleable solids.
solids (SS) per unit original volume is
Diameter
SS = M /WV ,
T T (7-21) SI
Range* (mm) Average (mm) (ft/sec)
where W is the dry bulk density of the solids. Based on
0.0010 0.00075 0.04 1.58 χ ΙΟ" 6
3.40 x 10' 9

Τ = 1 hr and the values given in Fig. 7.9 a n d using


0.0010-0:0016 0.0014 0.04 5.51 χ ΙΟ" 6
1.42 x 1 0 7

V = D/T, k is defined
0

0.0016-0.0024 0.0021 0.04 1.24 χ ΙΟ" 5


1.62 x 10' 6

3600 sec3 3
0.0024-0.0034 0.0030 0.04 2.53 x Ι Ο 5
1.39 x 10* 5

0.0034-0.0043 0.0042 0.04 4.96 x 10" 5


1.05 x 10" 4

(15.57/12) J
ft 3

0.0043-0.0064 0.0056 0.04 8.81 x 10" 5


5.88 x ΙΟ" 4

= 2.135 Χ 1 0 1 0
sec /ft . 3 3
(7.22) 0.0064-0.0082 0.0070 0.04 1.38 x Ι Ο 4
2.26 x 10" 3

0.0082-0.011 0.0095 0.04 2.54 x ΙΟ" 4


1.76x Ι Ο 2

Then

C ΣΔΧ ί = 0.33 UcV 3


ΔΧ = 0.02
SS = — (1 -X ) + Σ*Κ?Δ*ι (7.23) SI I

W 0

"From Fig. 7.10.


*From Fig. 7.10.
T r o m Eq. (7.5) or Fig. 7.3.

Example Problem 7.3. Use of an Imhoff cone Using Eq. (7.23) and assuming W of 70 l b / f t (1120 mg/ml), 3

1000 mg/liter
If the peak effluent TSS from a pond is 1000 mg/liter and SS = [(1 - 0.33) + 0.02]
the effluent sediment size distribution is given by Fig. 7.10, 1120 mg/ml
calculate the settleable solids concentration in the effluent. SS = 0.616 ml/liter
Solution: Solving first for V , 0

D (15.57/12) ft T h e major problem involved in the estimation of SS


K
o = τ = ' ν = 0.00036 ft/sec.
n
Τ (3600 sec) from TSS is determining a value for W. Compression
settling in t h e bottom of an Imhoff cone causes bulk
Assuming a temperature of 68° F, then the equivalent diame-
density to change with time. Therefore w h e n measur­
ing SS, it is important to m e a s u r e the volume of
deposited material at 1 hr.
d -
1 .00075 mm
100 r
d ~
2 0.0014 mm
dj- 0.0021 mm
d4 - 0.0030 mm Bulk Density
d5 - 0.0042 mm
Particle density and bulk density are two basic ways
to express sediment weight. Particle density is defined
as t h e mass of a unit volume of sediment solids. A
simple example is that if 1 c m of solid material weighs
3

2.65 g, t h e particle density is 2.65 g / c m . Particle 3

densities generally fall between 2.60 and 2.75 g / c m 3

for mineral particles. Organic m a t t e r weighs much less


than an equal volume of mineral solids and often has a
particle density of 1.2 to 1.4 g / c m . 3

.0001 .001 .01 .1 Bulk density provides a second way of defining sedi­
PARTICLE DIAMETER (mm) m e n t weight. Bulk density is defined as t h e mass of a
Figure 7.10 Particle size distribution for Example Problem 7.2. unit volume of dry sediment, including both solids and
Particle Size Classifications 213

A similar relationship occurs in settleable material in


Fine soils the bottom of a pond. D e p e n d i n g on the size of the
material t r a p p e d in the pond, considerable range in the
1.8-^
bulk density can occur. Also since the material is often

SΊ.6
quite loose, it is susceptible to mixing and eventual
transport from the p o n d because of turbulence caused
by runoff entering the pond. This concept must be
recognized if sediment storage is to be properly ac­
Coarse soils counted for w h e n sizing a pond. Equations (7.10) and
3 (7.11) can be used to m a k e this evaluation.
CD

1.2H
PARTICLE SIZE CLASSIFICATIONS

Textural Classification
VERY LOOSE VERY T I G H T
COMPACTNESS OF SOIL Z O N E T h e size of individual grains (primary particles) is
Figure 7.11 Relationship between compaction and approximate the basis for traditional approaches to classifying soils
ranges of bulk densities for coarse- and fine-textured soils (after (and sediment). McKyes (1989) and the U S D A (1979)
Brady, 1974). discuss a n u m b e r of classification systems based on
individual grain sizes, indicating that t h e number of
divisions, as well as their breakpoints, varies from one
pores. Brady (1974) discussed the factors alfecting bulk system to another. Several of the m o r e common classi­
density, indicating that loose and porous materials fication systems are shown graphically in Fig. 7.12. A
have lower bulk densities than m o r e compact materi­ major difficulty in using the classifications shown in
als. Sandy soils, particularly those with low organic Fig. 7.12 is that most soils d o not fall into only one of
matter, generally fit tightly together. Soil particles in a the categories, but instead are combinations of several
fine-textured soil typically do not lie closely together, categories. O n e technique that is commonly used to
but instead tend to bridge between individual particles classify these mixed soils is to use a textural triangle as
so that large pores may exist. T h e bulk densities of shown in Fig. 7.13. T o use a textural triangle, the
clay, clay loam, and silt loam surface soils normally sample is analyzed using only material that is less than
range from 1.00 to about 1.60 g / c m , whereas sandy
3
2 m m in diameter (i.e., sand or smaller). Aggregates
loams may have bulk densities ranging from 1.20 to are pulverized or soaked to separate aggregates into
1.80 g / c m (Brady, 1974). Figure 7.11 shows a general
3
primary particles. T h e sand, or coarse material, can be
relationship between bulk density and compaction for sized mechanically using sieves. T h e material caught in
both coarse- and fine-textured soils. T h e bulk density each sieve is then weighed, and the fraction of the
for fine soils is more sensitive to compaction because of original sample is calculated. Since silt and clay parti­
large void spaces in the loose condition. Compaction cle fractions are too fine to b e s e p a r a t e d using sieves,
allows these void spaces to be filled with solids. they must be d e t e r m i n e d using either hydrometer or

International SAND
Society of CLAY SILT GRAVEL
Soil Science Fine j Coarse
0.002 0. JZ 0.2 I 0 mm

0.002 0.05 0.10 0.25 0.5 1.0 2.0 mm


United States Very Very
Department CLAY SILT Tine Fine Med. Coarse coarse GRAVEL
of Agriculture
SAND

United States SAND


Public Roads CLAY SILT GRAVEL
Administration Fine Coarse
m 005 025 10 mm
Figure 7.12 Common classification systems (after Brady, 1974).
214 7. Sediment Properties and Transport
Table 7.2 Grade Scale of Sediment Particle Sizes
[Adapted from USDA (1979)]

Large sizes

Metric unit English Unit


Soil textural class (mm) (in.)

Very large boulders 4096-2048 160-80


Large boulders 2048-1024 80-40
Medium boulders 1024-512 40-20
Small boulders 512-256 20-10
Large cobbles 256-128 10-5
Small cobbles 128-64 5-2.5

Very coarse gravel 64-32 2.5-1.3


Coarse gravel 32-16 1.3-0.6
Medium gravel 16-8 0.6-0.3
Fine gravel 8-4 0.3-0.16
Figure 7.13 Textural triangle for soil classification (after Soil Sur­
vey Staff, 1951). Very fine gravel 4-2 0.16-0.08

pipette analysis as described previously in this chapter.


Details concerning these procedures and the equip­ Table 7.3 Size Limits for Sand and Finer Particles (USDA, 1979)
ment required for analysis can b e found in A S T M
Soil textural class Range in diameter (mm)
(1985) and U S D A (1979). T h e general terms used to
describe soil texture in relation to the basic soil textu­ Very coarse sand 2.0-1.0
ral class names for several soils are presented in Tables Coarse sand 1.0-0.5
7.2 and 7.3.
Medium sand 0.5-O.25
Fine sand 0.25-0.10
Very fine sand 0.10-0.05
Example Problem 7.4. Textural classification Silt 0.05-0.002
Clay Below 0.002
Two surface soils are analyzed using a standard hydrome­
ter test to obtain the following percentages of sand, silt, and
clay particles.

Particles (%) A Β
(USCS) in classifying solids for engineering purposes.
T h e U S C S bases classification upon the soil's particle
Sand 65 30
size, gradation, plasticity index, and liquid limit. It
Silt 22 40
utilizes the material having less than 3-in. diameter.
Clay 13 30
T h e Soil Conservation Service (1984) lists several out­
standing features of U S C S :
What is the textural classification of each soil? 1. Test procedures are simple.
Solution: Locate the point in the textural triangle (Fig. 2. I m p o r t a n t physical characteristics are described.
7.13) where each of the three percentages intersect. In this 3. Results are realistic and consistent in properties.
case, the USDA soil classification for soil A is found to be
sandy loam and for soil Β is clay loam.
Evaluation of Size Distribution Data
Considering all of t h e various properties of sedi­
T h e Soil Conservation Service (1984) and many other ment, size generally has the greatest importance to the
agencies use the Unified Soil Classification System engineer working in t h e area of sedimentation. Size
Particle Size Classifications 215

can b e measured readily and also can be related to


other characteristics such as particle shape, specific
gravity, and settling velocity. Simons a n d Senturk (1977,
1992) provide a detailed discussion of t h e physical
properties associated with sediment. O n e of these
physical properties is particle size. It can b e defined o n
several bases: volume, diameter, weight, fall velocity,
and sieve size. All of these except volume are d e p e n ­
dent u p o n shape and density of t h e particle. Generally
particle diameter characteristics for a specific soil a r e
expressed as a particle size distribution, either as ag­
gregated particles or primary particles, that is p r e ­
sented graphically as a curve similar to t h e examples
shown in Fig. 7.14. Each curve shows t h e fraction (or
percentages) finer plotted versus particle size and rep­
resents a size frequency distribution. Some users plot
fraction greater (instead of fraction finer) as t h e d e p e n ­
PARTICLE DIAMETER
dent variable. T h e fractions (or percentages) a r e b a s e d
Figure 7.15 Size distribution curve showing D , D , D , and
on t h e weight of sample that is smaller or larger t h a n 3 5 5 0 6 5

other representative diameters.


the specified size of t h e particles. D e p e n d i n g upon t h e
soil texture and gradation, t h e shape of t h e particle
size distribution can vary greatly as shown by t h e t h r e e
curves in Fig 7.15. D — t h e maximum size for t h e smallest 5 0 % of the
5 0

Specific points in t h e particle size distribution a r e sample a n d corresponds t o t h e median diameter. This
identified by Simons a n d Senturk (1977, 1992) as hav­ value is most naturally assumed as t h e best single size
ing b e e n used by researchers to define various sedi­ to represent a soil or sediment mixture, but because of
m e n t properties. Some of t h e commonly used points various influences including t h e wide variation in gra­
from the distribution include t h e following: dation that may b e in t h e mixture, this size is not
necessarily t h e best size for representing a given sedi­
D —the
35 maximum size for t h e smallest 3 5 % of t h e m e n t ' s transport a n d settling properties.
sample. Einstein (1950) specified this size to define D —the
65 maximum size for t h e smallest 6 5 % of the
grain size for sediment mixtures w h e n developing his sample. R e s e a r c h e r s have often used this size t o indi­
method for partitioning fluid drag. Additional informa­ cate roughness of a sediment mixture. It is sometimes
tion concerning this application is located in a subse­ used with D in this m a n n e r .
3S

quent section.

Example Problem 7.5. D 3S9 D 5 0 , and for three


100 soil textures

1 80 Estimate D , D , and D
35 5Q 65 for the three soil textures
shown in Fig. 7.14.
Kt / SILT LOAM
Solution:
W Q
SANDY L0AM
Uj uj 40 D 3 5 (mm) D 5 0 (mm) D 6 5 (mm)

Clay 0.0012 0.0017 0.0036


Silt loam 0.006 0.014 0.031

s ο
" alLl Sandy loam 0.062 0.133 0.295
or X 0.002 0.005 0.02 0.05 0.2 0.5 2
PARTICLE DIAMETER ( m m , log s c a l e )
Figure 7.14 Example particle size distributions for three soils (after
Brady, 1974). These examples are not intended to be typical for a
given textural classification. Particle size distribution for a given soil Two o t h e r terms associated with particle size distri­
would vary widely from these examples. butions a r e m e a n diameter and geometric m e a n size
216 7. Sediment Properties and Transport
(Simons and Senturk, 1977, 1992). M e a n diameter is Primary particle size distributions are of concern
given by when one is interested in identifying soils, probable
particle shapes, chemical contituents attached to parti­
cles, or water chemistry. Aggregated particle size distri­
ΣΔ,Α butions are of interest primarily when dealing with
Dm = (7.24) sedimentation because the density a n d size will affect
100 '
both the soil erodibility in the field and the settling
velocity in a sediment basin. Also, almost all soils
where Δ, represents incremental percentages on the e r o d e as a mixture of primary and aggregated particle
y-axis of Fig. 7.15 or 7.16, and D represents the m e a n
t
sizes.
for the sizes defined by the u p p e r and lower values of
Several researchers have recently studied techniques
the interval M e a n diameter D is then computed
m
for estimating the e r o d e d size distribution based on
using Eq. (7.24). laboratory or field tests of soil samples. Soil aggregates
T h e geometric m e a n diameter may be approximated have b e e n particularly difficult to characterize in terms
from of p a r a m e t e r s that are useful for engineering design of
erosion control measures and sediment control struc­
A log Djf tures because of the complex n a t u r e of soil aggregates
D = antilog (7.25)
(Tollner and Hayes, 1986). O n e of the major problems
g m
100
associated with soil aggregates is that they are subject
to destabilizing forces both in n a t u r e and in procedures
T h e geometric m e a n diameter is also not typically the
that are used to characterize aggregates. R a i n d r o p
same as the median diameter, D . T h e geometric
impacts, runoff, and channel flow each contribute
5 0

mean and median diameter will match only if the size


erosive energy, which may result in changes to the
distribution graphs as an S-curve on a semilog scale
aggregate characteristics as compared to the aggregate
and is symmetric about 5 0 % .
conditions at the site prior to rainfall ( R h o t o n et al.,
Several of the laboratory procedures for determining
1982). T h e stability of an aggregate describes the re­
the particle size distribution have b e e n discussed previ­
sponse of the aggregate to four inherently active soil
ously in the sections dealing with settling. Detailed
forces (Hartge, 1978):
information on the laboratory tests for aggregated and
primary particles can be found in soil mechanics or soil 1. weight of the particle,
physics references or are described in A S T M Stan­ 2. forces transmitted by the solid phase,
d a r d s . It must be stressed that one must recognize the
3
3. forces transmitted by the liquid a n d / o r gas phase
difference in laboratory procedures used to obtain pri­ within the aggregate, and
mary versus aggregated particle size distributions in 4. forces originated by contacting interfaces.
order to properly apply the results. A field manual
Stability is especially crucial in soils with substantial
developed by U S D A (1979) presents a thorough expla­
clay content because of the dynamic n a t u r e of clays.
nation of many traditional methods for testing soils or
A m o n g researchers and field personnel, t h e r e is con­
sediment. Recently, additional instruments have b e e n
siderable variation in the analytical procedures utilized
developed. T h e wide diversity in detection techniques
to characterize eroded sediment aggregates. In addi­
for determining particle size was discussed by Burcham
tion, the results of a n u m b e r of studies show that
(1989). Techniques range from sieving (a method that
considerable differences in predicted sediment aggre­
is visible in early Egyptian drawings) to sophisticated
gate size distribution can be expected with classical and
electronic methods. Methods used to delineate particle
analytical procedures (Tollner and Hayes, 1986). T h e
size include sieving, incremental sedimentation, mi­
characterization of soil or sediment properties needs to
croscopy, centrifugal methods, photoelectric methods,
provide absolute results if they are to be useful for
turbidimetric methods, electrical sensing zone meth­
engineers. Many present m e t h o d s provide relative re­
ods, radiation scattering, adsorption, laser light extinc­
sults that are useful for comparison only. Basic ap­
tion, sound attenuation, and particle separators. N o
proaches for determining aggregate size have not
one method is completely adequate, typically because
changed much in years. Changes that have occurred
of either a limited sampling range or difficulties in
are in technology for sensing concentration and size of
obtaining a representative sample.
aggregates in solution and improved technology such as
microcomputers for analyzing the information. How­
ever, standard m e t h o d s are available for characterizing
A S T M Standards are available for a variety of tests and are avail­
able from the American Society for Testing Materials. the size and mineralogy of primary particles. A t t e m p t s
Particle Size Classifications 217

have also b e e n m a d e to develop m e t h o d s to predict 100

eroded size information based on primary particle size 90


1'
R [INITIAL CONC.|
information a n d / o r field samples. 80 I ι
F ,1
I
(3 70 / ι
/
1,1
UJ
T.I
^ 60 Ξ =70%
Size Distribution after Sedimentation CD
1 I
In the design of sediment control structures, it is I so
40
1
1]
[

necessary to know how size distribution changes as ι-


α: 1
ΖUJ 30 L
sediment moves to a watershed exit or after significant u_
/ •
•.E.=20*
ζ
deposition has occurred. Hayes (1979) and Hayes et al. # 20
/
(1982) describe an approach that considers the impact 10 2<
of trapping. T h e m e t h o d subtracts a portion of the I- I ά
particle size distribution curve based u p o n what has ί.01 0.1 1 10
been trapped. Hayes assumed that as the sediment DIAMETER (mm)

transport capacity decreases, the largest particles would Figure 7.17 Adjusting particle size distribution for trapping.
settle out first. If sediment transport capacity continues
to decrease, the size of the particles remaining in flow
also decreases. A simple example of this concept, as
tained using t h e p r o c e d u r e s described (Hayes et al.,
illustrated in Fig. 7.1, is that if 2 0 % of the original
1982). T h e assumptions that are m a d e should be recog­
sediment concentration is trapped, the largest 2 0 % of
nized. Sediment deposits often contain some fine­
the sediment is assumed to be t r a p p e d . T o obtain an
grained particles mixed with large particles, and large
estimate of the new particle size distribution for the
particles are occasionally found far downstream from
remaining sediment, the diameter originally corre­
w h e r e they would b e expected if the assumption were
sponding to D in Fig. 7.17 becomes D for the new
completely valid. This is likely due to the stochastic
8 0 l00

distribution through a normalization process. O t h e r


n a t u r e of turbulence and its effects on bedload, as well
points along the original size distribution curve are also
as the exponential n a t u r e of deposition of suspended
adjusted based o n trapping until a completely new
load. A completely a d e q u a t e explanation of this phe­
particle size distribution is obtained as shown as the
n o m e n o n is currently not available.
intermediate curve in Fig. 7.17. T h e process is re­
Barfield et al. (1979) also proposed a procedure
peated, as necessary, to obtain distributions after addi­
based on the assumption that the largest particles in
tional trapping. In Fig. 7.17, the 7 0 % curve represents
t h e flow settle out first w h e n deposition occurs. Using
an estimate of the size distribution remaining if 7 0 % of
this assumption, the size distribution of the material
the original sediment is trapped. Limited data have
exiting a subwatershed contained within a larger water­
been p r e s e n t e d but available data show acceptable
r
shed can b e estimated if the e r o d e d size distribution
aggreement between measured values and values ob-
a n d the delivery ratio for t h e subwatershed is known.
T h e size distribution of the percentage finer (PF) after
partial deposition has occurred is given by
100
PF 2 / = PF /D U l i ; PF 2 / < 100, (7.26)

w h e r e P F and P F are shown in Fig. 7.18, and D is


2 / l f u

UJ
the delivery ratio for subwatershed i. A n example
m e t h o d for calculating D is the Modified Universal u

Soil Loss equation, or M U S L E (Williams, 1977), of the


form
cc
UJ 0.56
ζ 95(g,-<? ) P<
(7.27)
R Ά :

w h e r e Q is runoff volume in acre-feet, q


i is peak pi

).01 0.1 1 10 discharge in cfs, A is the watershed area in acres, and


t

DIAMETER (mm) R is the rainfall factor for the storm (see further
Figure 7.16 Example Problem 7.6 figure showing size distribution discussion of the M U S L E in C h a p t e r 8). T h e size
curves. distribution of material from subwatershed i that
218 7. Sediment Properties and Transport
the delivery ratio for subwatershed if the delivery ratio for
subwatershed i is 0.4.
Solution:

Diameter 0

(mm) % %

0.020 0 0
PARTICLE SIZE 3
·'
0.035 5 12.5
Figure 7.18 Changing particle size distribution along a flow path
0.048 10 25
(after Barfield et al., 1979).
0.064 20 50
0.250 85 100
0.300 90 100
reaches the main watershed exist can be estimated
0.700 100 100
from

PF = PF /D ; PF < 100, (7.28) "Diameter selected from distribution A in


3 / 2 | 2 / 3 /
Fig. 7.16.
*PF| . read from points corresponding to
where D is the fraction of material exiting subwater-
2i
f

diameter in column 1.
shed i that reaches the main watershed outlet. O n e P F calculated from Eq. (7.26) with D
C
2 1 u

method for calculating D is the Williams routing 2i equal 0.4.


equation [see C h a p t e r 8, Eq. (8.106)]:
Note that values of P F must be less than 100. Partial
2|

deposition at additional downstream points could be calcu­


D 2i = ^ = - * - y (7.29) lated in a similar manner.

or
D2i = -BT /D^
e ti} 9 ( 7 3 0 )

DEVELOPING SIZE DISTRIBUTION DATA


where Y is the sediment from subbasin i that reaches
t

the main watershed exit, and Y is the yield at the 0i


Introduction
subwatershed exit. A composite size distribution curve Several radically different approaches have b e e n
for an entire watershed can be determined by combin­ considered for estimating e r o d e d particle size distribu­
ing the curves from each subwatershed. For example, if tions. O n e approach applies simulated rainfall o n t o a
the fraction of total yield from subwatershed i is small sample of a representative soil. T h e size distribu­
tion of t h e particles in collected runoff from the sample
P = y,/2Y„ (7.31)
t
is t h e n analyzed in the laboratory to obtain an estimate
then the composite size distribution curve weighted of the e r o d e d sizes. A second m e t h o d seeks to simulate
according to the fraction of total yield P is i
the impacts of erosion and transport using a series of
laboratory steps involving wetting a soil sample and
subjecting it to shaking. A third approach uses primary
PF, = Σ M P F * ] . (7.32) particle size data in empirical equations. Obviously,
1= 1
these m e t h o d s cannot directly consider any deposition
Additional discussion of equations for estimating that occurs in the field. However, they can provide a
sediment yield from subwatersheds, main watersheds, first estimate of e r o d e d size distribution until better
and delivery ratios is contained in C h a p t e r 8. m e t h o d s are developed. T h e m e t h o d s are described in
the following sections.

Example Problem 7.6. Percentage finer after Single-Nozzle Rainfall Simulator Method
partial deposition
A m e t h o d that was first proposed in Barfield et al.
Assume that distribution A shown in Fig. 7.16 represents (1979) utilizes a simple rainfall simulator technique to
the eroded size distribution for subwatershed i . Estimate the obtain a sample of the particle size distribution for
size distribution for material as it leaves the subwatershed if e r o d e d material. A sample of the soil of interest is
Developing Size Distribution Data 219

tained; and the size distribution of the fine silts and


clays is d e t e r m i n e d by a pipette analysis (without dis­
Spraying systems Pressure held persion). This modified wet sieve approach is simple to
V - J e t 8 0 1 5 0 nozzle at 5 PSI
perform, requires very little equipment, and appears
promising; but extensive d a t a have not b e e n presented
comparing the wet sieve approach to o t h e r techniques.
A n a p p a r a t u s for mechanizing the wet sieve process is
described in Barfield et al. (1983).

Container
>10 with expanded Laboratory Method
metal grid in
the bottom R h o t o n et al. (1982) proposed two methods that
differ in the technique involved in wetting the soil
Bucket with sieves
sample. O n e m e t h o d lets the sample soak in deaerated,
After wash sieving, distilled water. T h e second m e t h o d allows the sample
fine sizes are . Metal lip
ischarges to wet by placing the soil sample on filter paper, which
determined by \
pipetting or ^ onto sieves is placed on a saturated sponge in an enclosed
particles size tray containing distilled water. Samples equilibrate
analyzer
overnight and are t h e n transferred to a flask. Distilled
w a t e r is a d d e d . E a c h soil suspension is agitated on an
orbital shaker for varying lengths of time. Immediately
after agitation, size distributions are found using pro­
cedures identical to procedures used for field samples.
Figure 7.19 Schematic of soil erodibility test using single nozzle This process includes wet sieving through a stack of
rainfall simulator.
five sieves with openings of 1000, 500, 250, 125, and
63 μ. Material less t h a n 63 μ is transferred to cylinders
and analyzed by pipetting after dispersing with hexam-
placed in a small p a n with an expanded metal base as etaphosphates. Discrepancies between field measured
shown in Fig. 7.19. Simulated rain is applied by a and laboratory m e a s u r e d percentage finers increased
Spraying Systems Veejet 8 0 1 5 0 nozzle located 3 m
4 as sediment size decreased in R h o t o n et al.'s results.
(10 ft) above the sample. A s drops fall, larger drops Seventeen soils were tested, with 14 from delta and
gain speed and small drops slow so that each d r o p upland areas of Mississippi and 3 from Iowa. T h e
approaches its terminal velocity prior to reaching the wetting m e t h o d had n o significant effect on prediction
sample. This nozzle appears appropriate based on rain­ of size distribution. Agitation times significantly im­
fall simulator studies that have shown that it produces pacted the prediction of size distribution. Best fits of
drop sizes and kinetic energy approximately equal to the curves were obtained by varying the agitation times
that of a 26- to 5 1 - m m / h r rainfall (Meyer and from 5 min for M e m p h i s and Sharkey soils to 45 min
H a r m o n , 1979). Similarly, a Spraying Systems Veejet for Loring. Most soils had a best fit in the 10 to 20 min
80100 nozzle produces drops and kinetic energy ap­
4 range, and R h o t o n et al. concluded that an agitation
proximating those of a 2- to 1 3 - m m / h r rainfall. T h e time of 14 min is probably appropriate for most soils.
spray nozzle used is selected based on the intensity of This time would generally predict the eroded size
the design storm. T h e sample p a n is sloped so that the distribution within o n e standard deviation of the mea­
drainage flows onto sieves nested in an 18.9-liter (5-gal) sured values from field tests.
bucket. Simulated rainfall is applied until the desired
design storm's depth is reached in a rain gauge m o u n t e d
CREAMS Equation Method
adjacent to the pan. T h e sieves are gently washed t h r e e
times to grade the sand through coarse silt fractions, A third alternative in estimating the composition of
fractions removed from the bucket, and dried at 105° C e r o d e d materials was proposed by Foster et al. (1985)
for 24 hr (or until constant weight is obtained). After and forms the basis of particle size estimation in the
the last sieve is removed, the volume of water in the C R E A M S and W E P P models. This technique defines
bucket is measured; a concentration sample is ob- five particle classes (primary clay, primary silt, small
aggregate, large aggregate, and primary sand) based
u p o n the primary particle sizes of the original soil
4
Manufactured by Spraying Systems, Inc. (Wheaton, IL). matrix to describe the composition of the eroded mate-
220 7. Sediment Properties and Transport
Table 7.4a Representative Diameters by Classes Based on Soil Matrix Fractions
[after Foster et al (1985)]

Range limits
Representative diameter of clay in Specific
Class (mm) soil matrix gravity

Clay D c l = 0.002 2.65


Silt D s i = 0.010 2.65
Sand D s a = 0.200 2.65
Small Aggregate D sg = 0.030 Ocl < 0.25 1.80
D s g = 0.2 ( O - 0 . 2 5 ) + 0.030
c I 0.25 < O < 0.6 c l

D s g = 0.100 <9 >0.60


cI

Large Aggregate D , = 0.30 g O <0.15


c l 1.60
D, = 2 0cl g O >0.15
c l

Table 7.4b Fraction of Sediment by Class Based on Soil Matrix Particle


Size Distributions [after Foster et al (1985)]

Class Fraction of sediment in class Range limits of clay


(mm) in soil matrix

Clay F c l = 0.26O c l

Small Aggregate O <0.25


c I

F s g = 0 . 4 5 - 0.6 ( O - 0.25) c l 0.25 < O < 0.50 c l

^sg = °-6 0 O d > 0.50


c l

Silt ^si^si-^sg
Sand ^sa = 0 s a O - 0 c l ) 5

Large Aggregate F
lg = 1
-^cl-^i-^sg-^sa

rial. Tables 7.4a and 7.4b summarize the equations for Solution: Using the equations for each class shown in
each classification by size range where D , D , D , cl s i s a
Table 7.4b, the fraction of sediment by class is calculated
D , and D in Table 7.4 are the representative diam­
s g lg
based on soil matrix particle size distributions:
eters for clay, silt, sand, small aggregates, and large F = 0.26 O = 0.26 X 0.36 = 0.09
cl cl

aggregates, respectively. T h e fractions of clay, silt, and


sand in the soil matrix are denoted as O , O , 0 , F sg = 0.45 - 0.6(0.36 - 0.25) = 0.38
c l si s a

respectively. Table 7.4b contains the equations for frac­ F = O - F


si sl sg = 0.6 - 0.38 = 0.22
tion of sediment by classes where F , F , F , F , and
c l si s a s g

F are the fractions of sediment in clay, silt, sand, ^sa - " a . ) ' - 0.04(1 - 0.36) = 0 5

l g

small aggregates, and large aggregates, respectively. ^lg =


1~
F
c\ - si ~ Fig ~
F
^sa
= 1 - 0.09 - 0.22 - 0.38 - 0.0 = 0.31.
The fractions of sand-, silt-, and clay-sized particles are then
Example Problem 7.7. CREAMS equations used calculated by recognizing that large aggregates correspond to
to calculate eroded size distribution sand particles and small aggregates with silt so that
Fraction sand = F, + F g sa = 0.31
A soil has a dispersed soil matrix of 4% sand, 60% silt,
and 36% clay. Estimate the fractions sand, silt, and clay using Fraction silt = F sg + F = 0.38 + 0.22 = 0.60
si

the revised CREAMS equations. Fraction clay = 0.09.


Developing Size Distribution Data 221
The sum of the fractions totals 1.00 as it should. It should be Piedmont soil. Rainfall for the year of interest was
noted that combining these primary particles and aggregates averaged and a 0.7-in. average d e p t h was used for
into sand-, silt-, and clay-sized fractions does not mean that rainfall simulation with representative samples from
they settle as primary sands, silts, and clays. The aggregate
each of the nine plots. Runoff from the rainfall simula­
component has specific gravities in the range of 1.6 to 1.8 and
tion was wet sieved and pipetted as described previ­
would settle more slowly then primary particles in the same
ously. N o dispersing agent was used. A portion of the
size range.
surface soil sample was also subjected to a standard
primary particle textural analysis for percentage sand,
silt, and clay, which included dispersion. These data
Foster et α/.'s equations were evaluated using 28 w e r e used with Foster et α/.'s revised C R E A M S equa­
different soils. Values of specific gravity for each of t h e tions to account for aggregation in the eroded particle
five classes are also recommended. A specific gravity of size distribution. In five of the nine tests, the equations
2.65 was suggested for the primary particle size classes p r o d u c e d values of D closer to those measured than
5 0

(sand, silt, and clay), and values of 1.8 and 1.6 w e r e the simulation m e t h o d . T h e equation m e t h o d showed
suggested for small and large aggregates, respectively. little variation between tests, whereas the measured
Little quantitative information that evaluates the and rainfall simulation values covered a wide range.
accuracy of these methods is available. A limited com­ However, t h e rainfall simulator technique produced
parison of the eroded size distributions from the Foster e r o d e d size distributions that were closer to the mea­
et al. (1985) equation method and the Barfield et al. sured values than the equation for sizes less than
(1979) rainfall simulator m e t h o d m e a s u r e d values is 0.1 m m in seven of t h e nine tests. Neither of the two
contained in Holbrook et al. (1986) using data from m e t h o d s considers the influence of cover or other
three tillage treatments with three replications on a surface stabilization measures. T h e procedures are, at

Table 7.5 Time of Pipette Withdrawal for Given Temperature, Depth


of Withdrawal, and Diameter of Particles (after USDA, 1979)*

Diameter of
particle (mm): 0.062 0.031 0.016 0.008 0.004 0.002

Depth of
withdrawal (cm): 15 15 10 10 5 5

Time of withdrawal

Min: sec
Temperature
(°C) Sec Hr:min

20 44 2:52 7:40 30:40 61:19 4:5


21 42 2:48 7:29 29:58 59:50 4:0
22 41 2:45 7:18 29:13 58:22 3:54
23 40 2:41 7:8 28:34 57:5 3:48
24 39 2:38 6:58 27:52 55:41 3:43
25 38 2:34 6:48 27:14 54:25 3:38
26 37 2:30 6:39 26:38 52:2 3:33
27 36 2:27 6:31 26:2 52:2 3:28
28 36 2:23 6:22 25:28 50:52 3:24
29 35 2:19 6:13 24:53 49:42 3:19
30 34 2:16 6:6 24:22 48:42 3:15

"Values in this table are based on particles of assumed spherical shape with an
average specific gravity of 2.65, the constant of acceleration due to gravity = 980
cm/sec , and the viscosity varying from 0.010087 cm /sec at 20°C to 0.008004
2 2

cm /sec at 30°C.
2
222 7. Sediment Properties and Transport
best, a first estimate. Additional procedures are desir­ tion, occurs when the forces of flow are exactly
able and need to be developed so that sediment size balanced by t h e submerged weight of t h e particle.
distributions in runoff can be measured routinely. Variables that impact the point of incipient motion for
uniform sediment on a horizontal bed are critical shear
stress ( T ) , particle diameter (d), particle specific weight
Adjustment for Particle Density c

(y ), fluid specific weight ( γ ) , fluid density ( p ) , particle


s

In most situations, the specific gravity of sediment density ( p ) , and kinematic viscosity (v).
s

particles varies over a wide range. This is particularly Shields (1936) recognized that these factors could b e
evident in the case of mixtures that contain both aggre­ grouped into two dimensionless p a r a m e t e r s so that
gates and primary particles. Since the aggregates and
primary particles have different specific gravities, pro­ u d mc

(7.33)
cedures are needed to convert the data to an equiva­ (y. - y)d
lent size distribution curve based on equivalent settling
diameter. Foster et al. (1985) recognized this problem w h e r e u „ is the critical shear velocity defined as
c

and suggested separating the particles into the five


classes (sand, silt, clay, small aggregates, and large Λ/2

aggregates), which represent density, size, and compo­ (7.34)


ρ
sition differences. While this reduces the quantity of
computations necessary to describe a particle size dis­ which is t h e critical friction velocity. T h e left side of
tribution, it is often desirable to compute equivalent Eq. (7.33) represents the critical Shields' stress, r „ , c

diameters based on settling velocities and an assumed and the right side is the critical boundary Reynolds
specific gravity. This can be accomplished quite easily number, R . Experimental data collected in flumes
em

by rearranging Stokes' law so that the equation is with a flat b e d permitted graphical representation of
solved for the equivalent particle diameter, d, based the relationship. T h e original version of Shields' dia­
upon an assumed specific gravity. A spherical shape gram does not apply adequately to sediment having low
and specific gravity equal to 2.65 is typically assumed. specific gravity and small diameter. Modifications to
If a sample is pipetted without dispersing the sediment Shields' diagram have b e e n suggested by n u m e r o u s
and samples are withdrawn according to times and researchers according to Simons and Senturk (1977,
depths calculated using Stokes' law, equivalent diame­ 1992). Gessler (1971) r e a r r a n g e d dimensionless param­
ters are obtained that are based u p o n the specific eters and developed a graphical relationship that has
gravity used to calculate the withdrawal times. Table the d e p e n d e n t variable in only o n e of the dimension­
7.5 shows the withdrawal times and depths as a func­ less terms. T o overcome the difficulty with small parti­
tion of t e m p e r a t u r e for various equivalent diameters cles, M a n t z (1977) extended the range of Shields' dia­
( U S D A , 1979). gram to smaller particles. This extension is shown
along with the original curve in Fig. 7.20.

SEDIMENT TRANSPORT Example Problem 7.8 Use of Shields' diagram


to determine critical shear stress
O n c e a soil particle erodes, it becomes part of the
flow and may be transported a few millimeters or
Determine the critical shear stress on the bottom of an
hundreds of kilometers. T h e distance is d e p e n d e n t unlined, trapezoidal channel carrying clear water. The chan­
upon the sediment transport capacity of the flow. Fac­ nel has a mean flow velocity of 0.5 m/sec, and the bottom is
tors controlling sediment transport capacity can be composed of slightly rounded quartz having d = 40 mm. 50

grouped into three categories: fluid properties, sedi­ Assume that the water temperature is 68°F and the channel
ment characteristics, and hydraulic p a r a m e t e r s associ­ is wide and has a slope of 0.8%.
ated with the flow path. Chang (1988) described how Solution: Base the critical shear stress criterion on Shields'
the motion of a particle is controlled by two opposing diagram (Fig. 7.20) and assume that flow is in the rough zone
forces: the applied force and resisting force. T h e ap­ where R > 400. For this situation,
e m

plied force results from hydrodynamics of flow, and the


resisting force comes from the submerged weight of the

T )cT
. « 0.06
particle. T h e particle will move if the applied force d(7 & ~ Ύ)
exceeds the resisting force, as discussed in C h a p t e r 4.
T h e threshold point for movement, or critical condi- Rearranging to solve for critical shear stress, the bottom can
Sediment Transport 223

Suspended load. Sediment that stays in suspension for


1.0U
some extended period of time as a result of suspen­
x ^ ^ O r i g i n a l shields diagram-, sion by turbulence.
Washload. Sediment that is composed of particles
smaller than the channel bed material and originates
from the channel b a n k and upslope areas.

Bedload
1000
If the hydraulics in a channel are such that a small
difference in o n e of the hydraulic p a r a m e t e r s will cause
t h e critical condition of t h e bed to b e exceeded so that
Figure 7.20 Shield's diagram as modified to include the Mantz
loose particles move, bedload transport occurs. Graf
data (after Storm et al, 1990).
(1971) grouped n u m e r o u s bedload equations into three
classes that have slightly different approaches to bed-
load transport:
safely handle
1. T h e duBoys-type equations that utilize a shear stress
(r )
G c r = 0.06</(y s -γ) relationship.
kN 2. T h e Schoklitsch-type equations that utilize a dis­
(r )
0 c r = 0.06 X 0.040m X (2.65 - 1.0)(9.81)-3- charge relationship.
m
kN 3. T h e Einstein-type equations that are grounded in
= 0.039- 2 * statistical considerations of lift forces.
m
To determine if the assumption that R * > 400 is valid, the Simons and Senturk (1977, 1992) indicate that bed-
e

shear Reynolds number may be calculated as follows: load a m o u n t s to about 5 to 2 5 % of the suspended load.
They emphasize that although this amount is a rela­
/ r \\
cc
11 // 22
/ 0 ..003399kkNN/ /mm \ \ 2 2 1 / 2
tively small proportion, it controls the shape, stability,
\p") =
\ 1000kg/m 3
) and hydraulic characteristics of the channel. Concepts
p r e s e n t e d by Einstein (1942, 1950) represent the most
0.039kN - m 1000 kg - m / s e c 2
well-known approach to theoretically explain sediment

-(
X —
1000 kg kN transport. His work was summarized by Graf (1971),
who recognized that Einstein's work had at least two
= 0.198 m/sec.
significant differences from previous efforts.
(Note that 1 kN = 1000 kg - m / s e c . ) 2

1. T h e idea of defining a critical value at which sedi­


β _ u,* d __ 0.198 m/sec
ι X 0.040 m m e n t motion commences is demanding, and proba­
ν lx 10" m /sec6 2 bly futile. Einstein bypassed this concept.
2. Bedload transport is related to fluctuations in veloc­
= 7.92 Χ 10 . 3

ity instead of t h e average velocity. T h e commence­


Since 7.92 X 10 is greater than 400, the assumption is all
3

m e n t and termination of particle movement is ex­


right.
pressed statistically using probability theory to
relate instantaneous hydrodynamic lift to the parti­
cle's weight forces. Experimental evidence that indi­
Classification Based Transport cates a close association between t h e moving bed-
As a particle travels, it may be considered to be load and the bed was described (Graf, 1971).
either bedload, suspended load, or washload. It should
Simons and Senturk (1977,1992) also discussed some
be noted that a particle's classification can change as it
of the concepts that Einstein used to describe his
travels depending upon the fluid, sediment, and espe­
theories dealing with sediment transport in open chan­
cially hydraulic conditions. Simons and Senturk (1977,
nels. E a c h particle moving past a cross section of a
1992) presented extensive descriptions of each classifi­
stream at a point of interest must have eroded some­
cation.
w h e r e in the watershed upstream of the cross section,
Bedload. Sediment that moves by saltation (jumping), and it must be transported by the flow from the point
rolling, or sliding in the flow layer just above the bed. of d e t a c h m e n t to the cross section at the point of
This flow layer is generally considered to be only a interest. Each of these two requirements may control
few particle diameters thick. the sediment rate at the cross section, the availability
224 7. Sediment Properties and Transport

of sediment in the watershed, and the transport capac­ again. This continues until t h e particle comes t o a
ity of the stream. Typically the finer material, which is point where the lift is less than the submerged weight.
easily carried in large amounts by flow, h a s limited T h e actual distance a particle moves is thus random,
availability in t h e water shed. This material is also but t h e average distance that a particle moves was
generally considered as washload (see definition at the d e t e r m i n e d by Einstein t o be A d o r L

beginning of this section). T h e coarse material is much


more difficult to move by flows, so its rate of movement
\d
Ad =
L (7.35)
is limited by t h e transport capacity of the flow. This
material is designated as bed-material load a n d in­
cludes both bedload a n d suspended load. Although where ρ is the probability that the lift forces, L, a r e
there is n o sharp division between bed-material load greater t h a n the submerged weight, W, o r
and washload, washload is often considered t o b e silt ρ = P r o b ( L > W). (7.36)
and clay materials. This is not always appropriate be­
cause in some channels having extremely slow flows, U n d e r the conditions where the n u m b e r of particles of
the bed material consists of fine silts. T h u s Simons and size d being e r o d e d is equal t o t h e n u m b e r being
Senturk (1977, 1992) delineated a second criterion that deposited, Einstein showed that
used the D of the bed material as the dividing size
1 0

between bed-material load and washload. Washload is = A - φ, (7.37)


often assumed to travel through a system by streamflow 1 -P
with very little deposition. Hence, very little of this
w h e r e A is a constant, i , is the fraction of particles of
small material is deposited. s

size d moving in t h e bedload, i is t h e fraction of b


Total sediment discharge (also called total sediment
particles of size d in the bed material, and φ is given
load) consists of both bed-material load and washload.
by
Generally, total sediment load can only be estimated if
the washload is estimated by measurement, experi­ 1
ment, or upland sediment yield equations because most (7.38)
sediment transport methods can only determine bed- F
7s V Ws ~ y ι sd 3

material load. w h e r e g is sediment discharge p e r unit width of size


s

Einstein concluded that there is a close relationship d, y is specific weight of sediment, γ is specific weight
s

between bed-material load and bedload because there of water, and F is a factor for settling velocity given by
is a continuous exchange of particles between t h e bed
material and bedload. I n addition, particles are trans­ 36p 2
36v2

ported along the channel bed in a sequence of steps F=


3 gd (SG
3
- 1) gd (SG
3
- 1) '
with the average distance of each step proportional t o
the particle size. It was also noted that a particular (7.39)
particle is not always moving, but rather, deposits o n
where ν is fluid viscosity and S G is sediment specific
the b e d after a few steps. Einstein found that t h e
gravity. T h e probability that lift is greater than sub­
deposition rate p e r unit area d e p e n d s o n both t h e
merged weight is based o n fluid dynamics. T h e lift
transport rate past a particular point and the probabil­
force was assumed t o be defined by the velocity a t the
ity that t h e forces o n t h e particle a r e such that t h e
top of a laminar sublayer formed as a result of drag
particle will deposit.
d u e t o grain roughness (see chapter 10), o r
Einstein's Method

Bedload. Einstein (1942) developed a relationship


for sediment transport based o n probability theory.
= cJ^pK d\\\.6) gR'S, 2
(7.40)
Ignoring drag force effects o n a particle, Einstein as­
x

sumed that a particle would b e dislodged from t h e


w h e r e 5 is slope a n d u is the velocity a t the top of the
channel bed if the lift force exceeded the submerged b

b e d layer, given by (see chapter 10)


weight during a n exchange time. T h e particle would
move a distance of \d before striking t h e surface u = 11.61/: = 1 1 . 6 ν ^ 5 \ (7.41)
b

again. If the forces at the point of impact are such that


lift exceeds submerged weight, t h e particle will take where R' is the hydraulic radius d u e t o skin friction,
another hop of distance Ad before striking the surface Κ! is a coefficient relating the cross-sectional area of a
Sediment Transport 225
particle to the square of its diameter d and C is the L

coefficient of lift. M e t h o d s for determining R' are


given in Chapter 10. Based on the assumption that t h e
submerged weight is proportional to ( p - p)gd , s
3

Einstein proposed that

P «/(*), (7-42)
where φ is referred to as the shear intensity p a r a m e t e r
given by

P ~ Ρ d

Φ=
s
(7.43)
ρ R'S

and Β is a constant.
T h e combination of Eqs. (7.42) and (7.37) indicates
that
(7.44)

where the functional relationship must be defined ex­


5 4 3 1 10.6 . 0.2 0.1
perimentally, similarly to Fig. 7.24. 0.8 0.4
Based on data taken in a research flume, Einstein
proposed that

0.465φ = e ' ° 3 9 1 φ
. (7.45)

In order to use Eq. (7.45), it is necessary to define R' Figure 7.21 Hiding factor for use in Einstein's equation (after
values for a given water depth. Procedures for making Einstein, 1950).
these computations are given in C h a p t e r 10.

To simplify calculations, first determine the common part


of Eq. (7.39) as
Example Problem 7.9 Bedload transport using
v2
Einstein's equation 36 v 2
36 X (1 X 1 0 " m / s e c ) 6 2

gd (SG - 1)
3
9.81 m / m s e c X (0.013 m) (2.65 - 1)
2 3

A stream in South Carolina is estimated to have the


following hydraulic data: slope of 1.5%, average depth of = 1.01 X 1 0 " , 6

1.0 m, and average velocity equal to 1.4 m/sec. The channel which can be substituted under the radicals in Eq. (7.39) as
is wide and approximately rectangular with width of 7 m. A
F = ^ 2 / 3 + (1.01 Χ 1 0 " ) - Vl.01 6
Χ 1 0 " = 0.817.
6

bed material sediment sample has been collected and ana­


lyzed, from which it was determined that d equals 13 mm. 35
Substituting φ, F , and given values into Eq. (7.38),
What is the rate of bedload transport? 1 1 1
1.23 = j = 0.087g .
Solution: Use Einstein's equation. To simplify computa­ s

tions, it will be assumed R' equals R =A/P= 1.0. This 0.817 2650 y 1.65 . 9 8 ( 1 3 x 1 0 -3y
implies a wide channel and a smooth bed. From Eq. (7.44), Solving for g by rearranging yields
s

the dimensionless ratios are related such that 1.23


gs = = 14.1 kg/m-sec.
0.087
Finally, multiplying by the channel width provides an esti­
where ψ can be found by substituting into Eq. (7.43) as
mate of channel bedload transport as
t P s -p d (SG-l)rf G = Bg = 7 m X 14.1 kg/m-sec = 99 kg/sec.
s s

ρ R'S R'S Note: The assumption that R = R' was made to simplify
computation for the example. This would not, in general, be
(2.65 - 1) 13 Χ 10" m 3

= 1.43. true.
1 (0.015 m / m ) ( 1 . 0 m)

From Eq. (7.45), φ = 1.23, which can be substituted with


other known information into Eq. (7.38) to solve for g after s In a later modification of the procedures above,
calculating F using Eq. (7.39). Einstein (1950) utilized a distribution of velocities at
226 7. Sediment Properties and Transport

the t o p of the laminar sublayer and proposed that where ψ + is given by

K (p -p)gd 3

*·" Ws ' ·
2 s
( 7 5 1 )
ρ = Prob 1 > , (7.46)
C^pK^uKl - η) where β a n d β χ a r e given by
where K is a coefficient relating d to particle vol­
2
3
β = log 1 0 10.6 (7.52a)
ume, a n d η is a p a r a m e t e r that defines the probability and
distribution of u , the velocity at a point 0 . 3 5 * from
b
/ 10.6* \
the theoretical bed, where 0* = l o g l o ( - ^ - J , (7.52b)

X = 0.77Δ if Δ / δ > 1.80 ξ is a so-called hiding factor given by Fig. 7.21 as a


function of d /d , a n d Y is t h e so-called pressure
= 1.395 if Δ / δ ^ 1.80, (7.47) i x

correction factor given by Fig. 7.23 as a function of


d /S.
where δ is t h e thickness of t h e laminar sublayer given 65

by Einstein showed that a new transport p a r a m e t e r


could b e defined as
δ = n.6v/u: (7.48a)
(7.53)
and Δ is a characteristic length given by
Using this transport p a r a m e t e r a n d t h e assumption
k. dt
«65 that η is normally distributed, Einstein showed that
Δ = (7.48b)
X 1 /·β^. -1/η ( 0 _2
r .
Ρ = 1 Γ/! / e r
dt = ,
where k is t h e characteristic particle diameter given
%

by d 65 a n d χ is Einstein's correction factor for t h e (7.54)


velocity profile, given in Fig. 7.22. Using the log veloc­
w h e r e A., Β , a n d η a r e constants. Based o n experi­
φ 0
ity profile, Einstein showed that
mental data, Einstein showed that t h e constants are
given by 43.5, 0.143, a n d 0.5, respectively. Experimen­
u b = 5.75 i/ log [10.6A7A].
2 2 2
0 (7.49) tal data a n d a theoretical curve a r e shown in Fig. 7.24.
Wilson a n d Barfield (1986) showed that t h e lower limit
Using Eq. (7.49) for u , Einstein modified Eq. (7.46)
b
in E q . (7.54) should b e zero; however, t h e impact of
to obtain the inappropriate lower limit o n t h e constants was
negligible. P r o c e d u r e s for calculating bedload move­
1
m e n t with Einstein's equations a r e given in Example
Prob 1 > Φ. (7.50)
l-η Problem 7.10.
Sediment Transport 227

100
t

Φ.
CONCENTRATION
Figure 7.24 Einstein's bedload equation (after Einstein, 1950).

Suspended Load W h e n critical conditions at the b e d


are exceeded so that particle motion occurs, a portion
of the sediment is diffused into the main flow by the
turbulence and does not move along the bed. T h e
settling of particles tends to offset the upward diffu­
Flow
sion, ultimately reaching a steady-state condition. M u c h Velocity
m o r e suspended load is transported in areas n e a r t h e
bed than occurs in regions n e a r the surface. Figure
7.25 illustrates the general relationships between flow VELOCITY
velocity and sediment concentration as a function of Figure 7.25 General shape of concentration and velocity profiles
depth. U S D A (1979) presents a figure showing dis­ as a function of depth.
charge-weighted concentration of suspended sediment
for the Missouri River at Kansas City. Figure 7.26
illustrates how concentration varies appreciably with
water d e p t h for coarse materials but is relatively uni­
form for fine sediments. Although the data in Fig. 7.26
are for a much larger channel t h a n might b e of general
concern, the figure depicts o n e of the major problems
in obtaining representative samples of sediment load
from flowing water. A considerable lateral and vertical
concentration gradient can occur in a given stream
cross section. Graf (1971) also reports that the concen­ CONCENTRATION: 1 Space-100 ppm by Weight
tration of suspended load is higher n e a r the center of a Figure 7.26 Example showing discharge weighted concentration of
stream as compared to n e a r the banks. T h e general suspended sediment by particle size for Missouri River at Kansas
shape of the resulting sediment concentration curve is City ( U S D A , 1979).
given in Fig. 7.25. A t equilibrium, t h e relationship
between upward diffusion and downward settling is
given by
and the resulting concentration profile is given by
dC
V C = K—,
S (7.55) (7.56)
oz

where V is the settling velocity for a particular particle


s
w h e r e C is concentration at a depth ζ above the
diameter, Κ is the turbulent diffusivity for sediment, channel bed, C is a reference concentration measured
a

and C is sediment concentration. Based on an analogy at a distance a above the channel bed, D is the depth
E

between sediment diffusion and m o m e n t u m diffusion, of flow, and


the m o m e n t u m diffusivity can be used in Eq. (7.55), Z' = V /Ulk,
s (7.57)
228 7. Sediment Properties and Transport
where k is von Karman's constant (approximately 0.4 A~ 1 - ζ τ­
z l

for clear water). / , = 0.216 Λ Ε


αζ (7.61)
( I - ^ - ' ^ B L ζ
T h e total suspended load for a given particle size
can be determined by the integral ι
1 - ζ ζ­
I = 0.216
2 /
( l - A ) z l
' A In zdz,
G ss = f r CUdydx
=W D
= C Q,
m (7.58) E t

(7.62)
where w is t h e stream width; D is the,stream d e p t h ; C
w h e r e A = a /D. T h e integrals I a n d I have b e e n
E E x 2
and U are concentration and velocity, respectively, a t
evaluated numerically and a r e p r e s e n t e d in several
any point (x,y) above t h e stream bed; a is t h e E
texts such as Simons and Senturk (1977, 1992) and
reference point above the stream bed, usually a few
Graf (1971). Values are shown in Fig. 7.27.
times the m e a n size of the bed sediment; C is the m
T h e reference concentration C was assumed by a
m e a n suspended sediment concentration (discharge
Einstein t o b e t h e average concentration in the bed
weighted); and Q is the stream discharge.
layer, a layer that he defined to b e equal in thickness to
Using the logarithmic velocity profile of Einstein a n d two particle diameters. Assuming that the bedload is
Barbarossa (see chapter 10) for U and Eq. (7.58) for C, not uniformly distributed over this layer, t h e average
Eq. (7.58) can be written as concentration over this layer is divided by 11.6 t o get
t h e reference concentration, or
G s s = 11.6κ<: Ι/:<ι [/> /
β Ε Ε 1 +/ ],
2 (7.59)

where 1
(7.63)
11.6 I / : β .
30.2D
PE = 2.303 log 10 (7.60)
T h u s the s u s p e n d e d load can b e predicted in terms of

10
10

v a i i i m •hiiiiiiibbbiii βΒβΗΙΙΐ!" IBIIIIII H B I I I I I I

10
ΙΚϋίϋΙΙΚΙΙΙΙΙΙΙΒΙΙΙΙΙΙΙΙΒΙΙΙΙ 10 Slllllll uiiiiiiimiiiii

M^Hw<lh<l!
ΜΒΙΙΙΙΙΓΗΙΙΙΙΙΙΙ
3B£lt<li;
10'
£11 10" mmwm siiiiiiiibiiiiiii
ι rv" « • 1111 >
"ΗΕ3ΙΙ.:·ΙΙ
»£ί«:ΐι»ιιιι
h 10
l 2 t0 s s u a i m
=ΕΕ5:::::;;
:***>;*:··;^3ί?£<ιιιι
"•mi 2 * * :yLLiiiiiii
ιριιιιιι
:lAi::::: ;

• Ε Ε Ϊ laiin; ^ ^ i : ; n i
-1
10
=35; 10
.-ι iimiaiiiii! IIIIIIIIIBfiii!!!

10
-2 IBIIIIIIII ibiiiii •••IIIIIIHHIIIIII
-5 - 4 - 3 - 2 - 1 lllllllll llllllll llllllll • l l l l l l l
10
10 10 10 10 10 1

10~ 5
10~* 10~ 3
10~ 2
10~ 1

Figure 7.27 Values for the integrals / , and I 2 in Einstein's suspended load equation. Curve parameter is ζ (after
Einstein, 1950).
Sediment Transport 229

the bedload as estimating total load because it requires bed samples,


depth-integrated samples of total suspended matter,
= ^WE'I +'2]. (7.64) and stream flow m e a s u r e m e n t s . Because of these dif­
ficulties, work has continued to find simpler, more
where / is the fraction of t h e particles of size d
s s

accurate m e t h o d s for defining total load.


moving as suspended load.
Yalin Total Load Equation
The Total Bed-Material Load T h e total bed-material
load, G , is the sum of suspended load a n d b e d load O n e of t h e most commonly applied transport equa­
st

tions for small channels is t h e Yalin bedload equation.


corresponding to sizes existing in t h e channel bed. For
Yalin (1963) developed his bedload transport model
a given particle size fraction i , this is given by
s t
for uniform, cohesionless grains over a movable bed.
Ά = *Ό'*. + i g )
8 ss ss (7.65) T h e model was derived using dimensional analysis and
t h e average grain motion for uniform turbulent flow
or
with a laminar sublayer that does not exceed the bed
«Ά- m,g [P I
% E x + / + 1].
2 (7.66) roughness. Yalin (1963) first presented the model as
several equations that were later reduced by Alonso
Procedures for making t h e calculations required are et al. (1981) to obtain an equation for sediment concen­
given in Example Problem 7.10. tration as

S G dU.
C = 6.35 x 1 0 5
- s 1 l n ( l + as) , (7.67)
Example Problem 7.10 Application of Einstein's
vh as
method to Big Sand Creek [adapted from Einstein
(1950) and Graf (1971)1 w h e r e S G is sediment specific gravity, U is bed shear m

velocity in meters p e r second, ν is average velocity in


A test reach in Big Sand Creek that can be represented by m e t e r s p e r second, a n d h is flow d e p t h in meters. T h e
a trapezoidal cross section with 1:1 side slopes and a bottom o t h e r two terms were defined by Yalin as
width of 300 ft was selected. The channel slope was mea­
sured as 0.0007 ft/ft. A grain size distribution was obtained
a = 2A5Y / 1 2
SG" 0 4
, (7.68)
by averaging five samples taken to a depth of 2 ft. The sizes C

can be represented by four size fractions as shown in column


1 of Table 7.7, for which sediment transport is determined. w h e r e Y is t h e critical mobility factor found
cr from
These four size fractions represent 95.8% of the bed mate­ Shields' diagram and
rial. A flow-discharge versus sediment-discharge relation
was found for the test reach. The stage-discharge relation­
ship is shown in columns 14 and 17 of Table 7.6. (Procedures (7.69)
*cr
for making these calculations are illustrated in Chapter 10.)
Einstein estimated that the highest discharge of interest was
20,000 cfs. Water viscosity was given as 1.0 X 1 0 " ft /sec,
1 0 2 w h e r e Y is a mobility n u m b e r defined by
and the specific gravity of the sediment as 2.65. Hydraulic
radius, wetted perimeter, and cross-sectional area as a func­
tion of depth were determined on the basis of channel Y = (7.70)
geometry. 7s d

Solution: Transport of each grain fraction is calculated at T o find t h e critical mobility factor, Shield's diagram
each given flow depth. Einstein divided the computations (Fig. 7.20) is utilized to d e t e r m i n e t h e value for T C

into hydraulic computations (Table 7.6) and bed material corresponding to a given roughness Reynold's number,
load calculations (Table 7.7). Explanation and references to U d/p. This value for r is t h e n used to determine a
m c
figures and equations are included.
critical shear velocity of t / = ( r / p ) , which is then
+ c c
0 5

used in Eq. (7.70) to d e t e r m i n e Y . Foster (1982) cr

a d a p t e d t h e m e t h o d t o predict transport of individual


Einstein's bedload function was modified by Colby particle size classes to obtain a particle size distribu­
and Hubbell (1961) to c o m p u t e total load. Difficulties tion.
with Einstein's bedload m e t h o d were t h e complexity
and detail required to solve t h e equations for a practi­ Yang's Unit Stream Power Equation
cal problem (Graf, 1971). Graf indicated even m o r e A n o t h e r commonly used sediment transport rela­
difficulty with Colby and Hubbell's modified version for tionship for small channels is that of Yang (1972,
-* — —' — — CN

—• —· ο» cs cn
1 ο ~ ~ -: — —

VO 00 00 ΙΟ —«
cn ON ON On
—; ο ο ο

»η «ο
Ο ο Ο

Ό Ή ιη ^- α

8 8 8 8 8
Ο Ο Ο Ο Ο

s sa = 2 8
-Η cn «η r< ^ w*S

cn oo
8 S § 2 cn m oo

m cn

Ο «η
on
3 3
ON
cn r- —<
m m
Ν Ν Λ «η vd

Ο
00 CN vo »n m in
VO oo
CN CN CN
fN cn cn cn
Ο ο d d d d d
ο
ON
00 CN
Ρ CM CN
cn cs cn cn u-i vd
c
'3
00 CN
c r- Ρ cn CN
d d d d

1 vO ο
-α m ο
d d d d d d d

< cn vo q
in
«n
vd
ο

q q
CN 00 cn >n 00
in
ι*

ε r~ ττ cn in vo in oo
•^t cn r- «n TJ- cn
csi d d d d

Q t 5 ^ ^ »
\fiι ^; CK — CN CN
cn cn «n vd oo on d

χ
m »n cn r- ο cn
^ oo ο ο

d d d d d d d
ω
§
m in in on m cn
1 vo m m cn —<

υ
3
•η ο ο νο «η on ο
U q «η ^ «η on cn no
υ ·—··—< cn cn cn" m' cn

Χ 8 S δ S 8 g δ
d d d d d d d
νο ο CN q ο vo 00
O m —' Φ ο rn vo
b — —' cN cn cn m m
φ d d d d d d d
Ξ
ο «η © © © © q ©
d cn cn Tf in vd

230
3
«Λ
i 31 ϊ I ϊ § ϊ I ϊ ϊ IIS
3
c> fiK ο
<m
r-
m
r-.
no
ττ rn r» r«i
— <s »n σν

3
«Λ
ϊ! s 111 111111
3
S 3
s
3

iliiillllllilS! I ! 8 i S ! I ! I E 111
— — cm m >/-> ^ —• <s
_ Μ rn

£ 8 8 8 ° S - S S 8 8 8 8

1 © © ' © — - ^ ( S r i o o - ^ r i p n i o v d ©

1
ο
© d o d o © © © © © © © - - — ©

«ο

ON ο cn ~ —" _• _" m pi - _•_•_• _• ( s - ^ ^ ^ o o o — - * © ' © © © ©

.5
•S I Si ί ξ S Ρ Ϊ S ? S R ? Κ ? R ί
© © © © © © © © © ' © © d o

ω « — © ο ο © — © © © © © © ©

β-
^ ^ © © © © © r ^ o o m o o o o v o ^ - c s — —• (N fN m — — γν fN m
a
S § 2 ^ 8 8 Π S 2 S 2 ? 2 o o — • r s o v o o t N Q ' t o o r s i o i r ) ©

ο ^' 2 η ί S [: * S U S ξ 5 " -— - Oc sN rr t^ >s on «r~-n or o- ' © © r n « n v d o o o ^

Γ η φ > Λ © Γ Λ Γ ^ © © · - < — > < S f N m


o 3 § 2 ? ^ 8 δ S S S f ^ 8 O O - N N N O O O C O O O
© © © © © © © © © © © © © © ' ©
© © © © © © o d d © © © ©
8£§!=Τ£ £ 8 8 8 8 8 8 8 2 $
• o o o o o o r ^ i n © © © © ©
« 0 Ο Γ Λ « Λ 0 Ο © © · ^ « Λ © © © ©
T t d t ^ r ^ s d ^ d c N ^ o v r n o o r s
— r i « n e N « ^ r ^ — f n o o f n r ^ « ^ © — — <N CN CO — — <s <s po
« r > c s » 0 " * © v o © o o o o < N m O « n
~ © © © © © d c 4 © d © © ©

ci - d © © © ci © d © © d «n

2 8 8 8 8 8 8 8 = 8 8 8 8 © 3
* e - « r * i o o s o © « o » n g o — t s r o o v o m
© irj —· f"i « O f ; 8 " f i * < n i n * " n r^o — o o o r < - i « n r - o o r ^ r - r -
©_·_•_·_• — © © © o d d ©
"8

3
r ^
ssη
r n _ _ d d ©
?η ' w n < N
ft ?; s δ Ϊ ! ? s
— © © © © ci
mr-^ooornrNjvooospTi-r^m
O N O s o - ^ t m r o r - m v O ^ r n c M C N
- - © © © d o t s - - © © © © ©

m p o p p p o ^ p p p p p p w - i p p p p o p w - i p o o o p o
© ~ c 4 c * i ^ « n v o © — c i r o ^ - i n x o ©
"8 - N n ^ i n ^ d - ( N n ^ i n \ d

OO CN ©
^ Ο <N
0D

— ^ ro
IV

3 5> _ So

231
232 7. Sediment Properties and Transport
1973), a sediment transport equation based on unit Y a n g (1984) later developed a similar equation for
stream power. Using regression analysis, he concluded gravel transport in which Eq. (7.71) b e c a m e
that the product of average flow velocity, K, and the
slope of the energy grade line, 5, is the primary indica­ l o g C = 6.681 - 0.633 log l
10 t 10

tor of flow conditions. T h e equation described in Yang


(1972) relates total sediment concentration to unit
stream power so that
- 4.816logio(-^f)

l o g C = a + j B l o g ( K S - V S), (7.71) 2.784 - 0.305 log 1 0 ( ^ ) - 0.282 log 1 0 (^-)


1 0 t 10 CT

where C is total sediment concentration in ppm, VS is


t

unit stream power in meters per second, V S is the cr


°*"(ττχ)·
χι (
" 3)

critical unit stream power required for incipient mo­ Hirschi (1985) provided an approach that considered
tion, and a and β are parameters. nonuniform sizes.
T h e equation was later nondimensionalized but was
only applicable to one sediment size. This limitation
presents difficulties in establishing a representative Example Problem 7.11 Illustration of Einstein,
particle size (Yang, 1973). In nondimensional form, Eq. Yalin, and Yang sediment transport equations
(7.70) becomes
A channel has a channel width of 20 ft, slope of 0.1%, and
overtops at a depth of 3 ft. The material in the bottom of the
log 10 C = 5.435 - 0.286 log |
t 1 0 channel has a size distribution as shown below.

Size (mm) PF

- 0.457 log 1 0 | ^
2.0 100
1.0 95
0.5 90
1.799 - 0.409 l o g ( 10 —
0.2 75
0.1 60
0.05 40
- 0 . 3 1 4 log 10
0.03 20
0.01 0
VS K 5
r

xlog 1 0 — " (7.72) For the channel and size distribution given, determine the
sediment discharge utilizing the Einstein, Yalin, and Yang
equations.
where K is particle settling velocity in meters p e r
s
Solution: The size distribution must be divided into size
second, d is particle diameter in meters, ν is kinematic fractions. By first plotting the given distribution on graph
viscosity in square meters per second, and U„ is shear paper, it is possible to divide the distribution into five parts,
velocity in meters per second. Yang developed the each containing 20% of the distribution. If this is done, a
equation using flume and stream data with noncohe­ mean diameter and settling velocity for each fraction can be
sive sand. determined with a corresponding settling velocity. This infor­
mation follows.

Size range Mean diameter Fraction Vs Mean diameter


(mm) (mm) (%) (ft/sec) (ft)

2.0-0.26 0.7211 20 1.461 0.00236


0.26-0.1 0.1612 20 0.073 0.000528
0.1-0.05 0.0707 20 0.014 0.000231
0.05-O.03 0.0387 20 0.0042 0.000127
0.03-0.01 0.0173 20 0.0008 5.7 χ 10 5
Sediment Transport 233

These size fractions and representative diameters are used in


Size (mm) Percentage finer
each of the methods. The Einstein method calculations fol­
low the same procedure previously discussed in Example 0.0002 0
Problem 7.10 and shown in Table 7.7. Table 7.8 summarizes
0.001 10
the computations used for Einstein's method.
Solutions of the problem using Yalin's and Yang's equa­ 0.003 20

tions are illustrated in Tables 7.9 and 7.10, respectively. 0.01 40


Detailed descriptions of the steps in the procedure are not 0.05 50
shown. Further discussion of partitioning of the hydraulic 0.10 70
radius and other aspects are explained and demonstrated in
0.30 100
Chapter 10. As can be seen from comparison of Tables 7.8,
7.9, and 7.10, computation of sediment discharge for a practi­
cal problem requires solution of a series of intricate equa­
(7.6) Estimate the settleable solids concentration
tions. Solution of the equations using computer modeling is
from a p o n d if the sediment size distribution is given
practically a necessity because of the likelihood of calculation
errors. Einstein (1950), Yalin (1963), Yang (1972), Graf by problem (7.4) and peak effluent TSS is 780 m g / l i t e r .
(1971), and Simons and Senturk (1977, 1992) provide addi­ (7.7) Two surface soils are analyzed using a stan­
tional information. dard hydrometer test to obtain the following percent­
ages of sand, silt, and clay particles.

Particles (%) A Β

Channel Armoring Sand 12 40

All particles on a channel bed d o not have the same Silt 64 25

susceptibility for lift. As a result, equations developed Clay 24 35


for critical shear stresses and velocity equations predict
that smaller particles are m o r e readily e r o d e d than are
larger ones. This results in the surface layer of the W h a t is t h e textural classification of each soil?
channel bed consisting of a larger fraction of coarse (7.8) Estimate D , D , a n d D for the sediment
3 5 5 0 6 5

particles than did the original bed. This results in what size distribution shown in problem (7.5).
is referred to as channel armoring (or bed armoring). (7.9) D e t e r m i n e the £> and D
50 representative
m

Channel armoring refers to the protection provided by particle sizes for the initial size distribution shown in
this coarse top layer to the finer particles that lie Fig. 7.17. A r e the m e d i a n and m e a n diameters the
underneath. Graf (1971) presents several photographs same in each case?
that vividly illustrate the protection provided by armor­ (7.10) Assume that distribution A shown in Fig. 7.16
ing. Chang (1988) discusses approaches for modeling represents the e r o d e d size distribution for subwater­
channel armoring. shed i. Estimate the size distribution for material as it
leaves the subwatershed if the delivery ratio for subwa­
tershed ι is 0.3. R e p e a t with distribution B.
Problems (7.11) A soil has a dispersed soil matrix of 3 7 %
sand, 5 0 % silt, and 1 3 % clay. Estimate the fractions
(7.1) C o m p a r e the settling velocities using the sand, silt, and clay using the revised C R E A M S equa­
Stokes equation and the Wilson et al. equation for tions. Also estimate the fractions of small and large
0.0004-, 0.004-, 0.01-, 0.02-, 0.04-, 0.10-, 0.2-, 0.4-, a n d aggregates. Plot the data as fraction finer versus size.
4.0-mm particles if S G = 2.65 for all particles and
(7.12) W h a t characteristic diameter is associated
settling occurs in water at 68°F.
with each of the classes for t h e soil matrix shown in
(7.2) R e p e a t problem (7.1) if the S G = 1.8. U n d e r problem (7.11)?
what conditions would you expect t h e S G to approach (7.13) D e t e r m i n e the equivalent diameter for a par­
this value?
ticle having a specific gravity of 2.65 if the actual
(7.3) R e p e a t problem (7.1) if the water is at 32 a n d specific gravity is 2.1 and the actual particle diameter is
100°F. A r e the results significantly different from those
0.14 mm.
obtained in problem (7.1)?
(7.14) Find the critical shear stress on the bottom of
(7.4) If the peak effluent TSS from a pond is an unlined, trapezoidal channel carrying clear water.
600 m g / l i t e r and the sediment size distribution is given
T h e channel has a m e a n flow velocity of 0.35 m / s e c ,
by Fig. 7.10, calculate the settleable solids concentra­
and the b o t t o m is composed of angular quartz having
tion.
d = 25 mm. Assume that the water t e m p e r a t u r e is
50

(7.5) Plot the following sediment size distribution. 68°F and the channel is wide and has a slope of 0.07%.
f ί Μ
0 < N — m O O J — m O f N O O O O O O O O

ώ ? 3
m ο cn ο vmo gV g O O O C N O O O ©

3 8 S S 3
r i ii 2£
o o o o o o o o

cm © ©

* a «ί 8 §

ο ο ο ο Ο Ο Ο
§ I
8 8
Ο (Ν OO ο Ο Ο Ο
o d d CM Tf
Γ» CM 00 cn cm
3 £ r» ov «n on 06 Γ» ON 00 — νο
— — <n m cn oo νο «η cm Ο Ο
- : © © © © © © ©

© © © © © © © © © O O

cn Ο (Ν cm ~ m ©
S
cm
Ι

3ί δc n ©
8 I

ii ii ώ uj
* 12 p §
« cn Ο cm r» >η ν£> © ο ©

ώ (8
νο m
Ο — Ο ο Ο Ο (Ν — © Ο © Ο Ο ©
"8

•5
S 3
«Ν «Ο Ο Ο Ο Ο Ο
e
"S S S 8
G

m ^ Ον t— γπ — Ο co ON <- Ο ίο
cn rn ^ «η cn Ή
«η ©
cn ^

I cn wn
ω
νο
ω
©
ω

ω
ρ».

Ή.

χ t s c n m r ^ o o — — Β ©
d d d d d d d d

oo vq — in oo § 8 8
r-^ op cn r-^ d

I— OC

•a
υ

CO
is!
φ
3
ο ©

234
so On oo
—· SO en so <n
oo

d d d
cn »n oc Wl
CN «n OO On

d dd d
ο
oo

dodo©
og
οm
TJ-
—« —ι οο

d dd d
r-~
—« cn co
go
—'
r-
cn
r--
m
oo
oo
Ο
«η
^
—·
—ι —·
fN
t^-
co ~> CN
IN
—<
—i
CN CO
—«

6.
d d SO CN CN CN
On
s^
CO *—'

0.01 O.OF
dd
^ CN Ο

odd SO Ο On Γ-
£ 2 § CN
CO
OO
—«
On Os
CO CN 2 2 2 ^ 1 ^ 8
cn m

cn oo «n r~- 00 CO

S 3 S 8odd
s s =
Β 8 δ3 8 8ο 8
ddοd d d s s
s
<
S ο
odd ο
odd
d d d d d
Ά 2 8
CN CN CN
d «g ο so cn
« Ρ ?
odd
d d d d d d d d d d d
oo os
"8 Os Tfr »n

•s 88 83 8888 8 * 8 88Sd Sο ο —ί ~

I ο
d d d d d d d d d d
so c-» ο ο d d d d d
<5 <5 £*
w w χ CO CN CN
8 8 8 8 S 8 S 8
8 88 8 8 8 d d d d d d d d d
§
ddddddddd
t^- —< ro
08880
τ}- _ Ο
00 rr» —1 os 00 r-

co co ^ ^ d d d d d
—· _·
8 83I 8 8
I Ο Ο Ο Ο Ο Ο Ο
o d d d d d
ο οο
2
ο ο ο ο οd
2 Μ
ο -d
ζ> « d§ d§ !d dS d=
ο
8 VO SO CO so —1 00 so On

d ~a $3 *
Tt CO
>η *η ρ ^ 00 00 »n On Ο
Ο «η οο co co —<
cn en T t

ι§
d d d d

Ο ο R-
>n
IN R- Ο Tf ρ TJ-
§ s 5 i §

ο ο οο οο οο ο ο
Tf

ο00000
CN ^ CN CN CN ·— CN 1

υ
δ s ο s s s
ρ ρ Ο
«n os

©9dε © ©9 d dd d
υ cn
4 r- cn I ϊ h (N I r»
so r-
d d d d d
rs
d d d
3

235
236 7. Sediment Properties and Transport

Table 7.10 C a l c u l a t i o n s for E x a m p l e P r o b l e m 7 . 1 1 ( Y a n g ' s M e t h o d )

R b
us V CT 1
s vf β' it

0.000719 0.219 0.014 10.56 3.107 0.443 0.088 5.395 1.239 — 0 0

0.443 0.020 15.00 2.788 0.443 0.145 5.325 1.191 — 0 0

0.675 0.025 18.5 2.634 0.443 0.271 5.283 1.163 — 0 0

0.729 0.026 19.2 2.608 0.443 0.371 5.276 1.157 — 0 0

0.000161 0.219 0.014 2.362 7.560 0.022 0.088 5.358 1.629 — 0 0

0.443 0.020 3.357 5.484 0.022 0.145 5.289 1.581 0.000 2.39E-08 2.39E-08

0.675 0.025 4.141 4.752 0.022 0.271 5.247 1.552 0.024 5.67E-06 5.67E-06

0.729 0.026 4.304 4.641 0.022 0.371 5.239 1.547 0.070 2.47E-05 2.47E-05

7.04E-05 0.219 0.014 1.033 -215 0.00424 0.088 5.338 1.844 52.58 0.001082 0.001082

0.443 0.020 1.469 17.44 0.00424 0.145 5.268 1.796 0.146 1.09E-05 1.09E-05

0.675 0.025 1.812 10.92 0.00424 0.271 5.227 1.76 1.896 0.00439 0.000444

0.729 0.026 1.883 10.24 0.00424 0.371 5.219 1.762 4.3508 0.001518 0.001543

3.87E-05 0.219 0.014 0.568 -7.84 0.00128 0.088 5.324 1.999 4.3166 8.88E-05 0.001171

0.443 0.020 0.807 -19.1 0.00128 0.145 5.254 1.952 14.444 0.00107 0.001081

0.67 0.025 0.996 -86.7 0.00128 0.271 5.212 1.923 87.95 0.020344 0.20788

0.729 0.026 1.03 -230.0 0.00128 0.371 5.205 1.918 301.1 0.105114 0.106657

1.73E-05 0.219 0.014 0.253 -2.84 0.00025 0.088 5.304 2.209 124.5 0.002563 0.003734

0.443 0.020 3.360 -4.1 0.00025 0.145 5.234 2.164 373.4 0.027661 0.028742

0.67 0.025 0.444 -5.36 0.00025 0.271 5.193 2.133 1492.0 0.3453 0.366088

0.729 0.026 0.462 -5.66 0.00025 0.371 5.185 2.127 2973.0 1.037594 1.144251

"d, mm (grain size): Representative examples are taken from a size distribution analysis (given in this problem). R, m (hydraulic radius with respect to
h

grains): Values assumed in this problem. See Chapter 10 for additional information. U, (shear velocity): (/» = VyRS . R , (shear Reynolds number): R , =
C d
t c

U.dv. 'V /V (critical velocity over settling velocity): For 0 < R . < 70, V /V = (2.5/(log fl .-0.06)) + 0.66. fy (particle settling velocity): V =
cl i e cr s c s s

2.81 (</*1000) /3.28. *V, m/sec (average flow velocity): Typically estimated from channel geometry and cover; assumed for this example. a (parameter in Eq.
2 h

7.71): α = 5.435 - 0.286 log, (V,d/v) - 0.457 ( l o g (ί/./V,)). 'β (parameter in Eq. 7.71): β = 1.799 - 0.409 log (V d/v) - 0.314 l o g (UJV )JC (total sed­
0 |0 10 s |0 % t

iment concentration in ppm): Eq. 7.72. q (sediment discharge): ς = 981 VDC,/10 with flow depths D corresponding to those from Table 7.9. ' I ^ (cumula­
k
s Λ
6
s

tive sediment discharge): Add q + previous Lq for grain size d.


st &

(7.15) A stream is estimated t o have t h e following References


hydraulic data: slope of 0.8%, average d e p t h of 2.1 m,
A l o n s o , C. V., Neibling, W. H., and Foster, G. R. (1981). Estimating
and average velocity equal to 0.6 m / s e c . T h e channel
sediment transport capacity in watershed modeling. Trans. ASAE
is wide a n d approximately rectangular with a width of 24(5):1211-1226.
10 m. A sediment sample has b e e n collected a n d Argaman, Υ. Α., and Kaufman, W. J. (1970). Turbulence and floccu­
analyzed, from which it was determined that d equals 35
lation. J. Sanitary Eng. Div. ASCE 96:223.

9 m m . W h a t is t h e rate of bedload transport if t h e A S T M (1985). Annual book of A S T M standards. American Society


for Testing and Materials, Philadelphia, PA.
channel bottom is smooth?
Barfield, B. J., Moore, I. D . , and Williams, R. G. (1979). Sediment
(7.16) Collect a bedload sample from a stream at
yield in surface mined watersheds, in "Proceedings, Symposium
your location. D e t e r m i n e t h e particle size distribution o n Surface Mine Hydrology, Sedimentology and Reclamation,
and plot as percentage finer versus diameter. W h a t is University of Kentucky, Lexington, KY," pp. 6 1 - 6 5 .
dl d? d 1
35 50 65 dj Barfield, B. J. Warner, R. C , and H a a n , C. T. (1981). "Applied
(7.17) Estimate or measure t h e hydraulic data for Hydrology and Sedimentology for Disturbed Areas." Oklahoma
Technical Press, Stillwater, OK.
the stream in problem (7.16) (use a flow m e t e r if
Barfield, B. J., Barnhisel, R. I., Powell, J. C , Hirschi, M. C , and
available), a n d estimate t h e rate of bedload transport
Moore, I. D . (1983). Erodibility and eroded size distribution of
using Einstein's method a n d t h e distribution d a t a pre­
surface mining spoil and reconstructed topsoil, Report IMMR
viously obtained. 8 4 / 0 9 2 . Institute for Mining and Minerals Research, University
(7.18) C o m p a r e t h e transport obtained by Yalin's, of Kentucky, Lexington, KY.
Yang's, a n d Einstein's methods for a stream at your Brady, N . C. (1974). " T h e Nature and Properties of Soils." Macmil-

location. lan, N e w York.


Sediment Transport 237

Burcham, Τ. N. (1989). Vortical particle size distribution system, Norma, ed.), N o . 7. Academic Press, N e w York.
Unpublished Ph.D. dissertation. Clemson University, Clemson, McKyes, E. (1989). "Agricultural Engineering Soil Mechanics," D e ­
SC. velopments in Agricultural Engineering 10. Elsevier, New York.
Chang Η. H. (1988). "Fluvial Processes in River Engineering." Metcalf, and Eddy, Inc. (1979). "Wastewater Engineering, Treat­
Wiley, New York. ment, Disposal, and Reuse." McGraw-Hill, N e w York.
Colby, B. R., and Hubbell, D . W. (1961). Simplified method for Meyer, L. D., and Harmon, W. C. (1979). Multiple-intensity rainfall
computing total sediment discharge with the modified Einstein simulator for erosion research on row sideslopes. Trans. ASAE
procedure, Water Supply Paper 1593. U.S. Geol. Surv. 22(1):100-103.
Einstein, H. A. (1942). Formulas for the transportation of bed-load. Miller, C. R. (1953). "Determination of the Unit Weight of Sediment
Trans. Am. Soc. Civil Eng. 107. for U s e in Sediment Volume Computation." U.S. Bureau of
Einstein, H. A. (1950). The bed-load function for sediment trans­ Reclamation, Denver, CO.
portation in open channel flow, Tech. Bulletin 1026. U.S. Depart­ Rhoton, F. E., Meyer, L. D . , and Whisler, F. D . (1982). A laboratory
ment of Agriculture. method for predicting the size distribution of sediment eroded
Evangelou, V. P., Rawlings, F., Crutchfield, J. D., and Shannon, from surface soils. Soil Sci. Soc. Am. J. 46:1259-1263.
E. A. (1981). A simple chemo-mathematical model as a tool in Rhoton, F. E., Meyer, L. D., and Whisler, F. D . (1983). Response
managing surface mine sediment ponds. In "Proceedings, 1981 of aggregated sediment to runoff stresses. Trans. ASAE
Symposium on Surface Mine Hydrology, Sedimentology, and 26(5):1476-1478.
Reclamation," College of Engineering, University of Kentucky, Rouse, H., ed. (1950). "Engineering Hydraulics." Wiley, New York.
Lexington, KY. Shields, A. (1936). "Application of the Theory of Similarity and
Foster, G. R. (1982). Modeling the erosion process. In "Hydrologic Turbulence Research to the Bed Load Movement, Vol. 26, pp.
Modeling of Small Watersheds" (C. T. Haan, ed.) Monograph 5 - 2 4 . Mitt. Preuss. Vers. Wasser Schiff. [In German]
No. 5. American Society of Agricultural Engineers, St. Joseph, Simons, D . B., and Senturk, F. (1977). "Sediment Transport Tech­
MI. nology." Water Resources Publications, Fort Collins, CO.
Foster, G. R., Young, R. Α., and Neibling, W. H. (1985). Sediment Simons, D . B., and Senturk, F. (1992). "Sediment Transport Tech­
composition for nonpoint source pollution analyses. Trans. ASAE nology." Water Resources Publications, Fort Collins, CO.
28(1):133-146. Soil Conservation Service (1984). "Engineering Field Manual." Soil
Gessler, J. (1971). Beginning and ceasing of sediment motion. In Conservation Service, Washington, D C .
"River Mechanics" (H. W. Shen, ed.), Chap. 7 Water Resources Soil Survey Staff (1951). "Soil Survey Manual," U.S. Department of
Publications, Fort Collins, CO. Agriculture Handbook N o . 18. U S D A , Washington, D C .
Storm, D . E., Barfield, B. J., and Ormsbee, L. E. (1990). Hydrology
Graf, W. H. (1971). "Hydraulics of Sediment Transport."
and sedimentology of dynamic rill networks. I. Erosion model for
McGraw-Hill, New York.
dynamic rill networks, Research report N o . 178. Kentucky Water
Haan, C. T., and Barfield, B. J. (1978). "Hydrology and Sedimentol­
Resources Research Institute, University of Kentucky, Lexington,
ogy of Surface Mined Lands." University of Kentucky, Lexington,
KY.
KY.
Tapp, J. S., Barfield, B. J., and Griffin, M. L. (1981). Predicting
Hartge, Κ. H. (1978). In "Structural Stability as a Function of Some
suspended solids removal in pilot scale sediment ponds utilizing
Soil Properties in Modification of Soil Structure." (W. E.
chemical flocculation, Technical report of Institute for Mining
Emerson, R. D . Bond, and A . R. Dexter, eds.). Wiley, New York.
and Minerals Research. University of Kentucky, Lexington, KY.
Hayes, J. C. (1979). Evaluation of design procedures for vegetal
Tollner, E. W., and Hayes, J. C. (1986). Measuring soil aggregate
filtration of sediment from flowing water, Unpublished Ph.D.
characteristics for water erosion research and engineering: A
dissertation. University of Kentucky, Lexington, KY.
review. Trans. ASAE 29(6): 1582-1589.
Hayes, J. C , Barfield, B. J., and Barnhisel, R. I. (1982). T h e use of
U S D A (1979). "Field Manual for Research in Agricultural Hydrol­
grass filters for sediment control in strip mine drainage. III.
ogy," U S D A Agriculture Handbook N o . 224. U.S. Government
Empirical verification of procedures using real vegetation, Tech­
Printing Office, Washington, D C .
nical report I M M R 8 2 / 0 7 0 . Institute for Mining and Minerals
Williams, J. R. (1977). Sediment delivery ratios determined with
Research, University of Kentucky, Lexington, KY.
sediment and runoff models. In "Proceedings, Erosion and Solid
Hirschi, M. C. (1985). Modeling soil erosion with emphasis on steep Matter Transport in Inland Water Symposium, IAHS, No. 122,
slopes and the rilling process, Ph.D. dissertation. University of pp. 168-179.
Kentucky, Lexington, KY. Wilson, Β. N., and Barfield, B. J. (1986). A detachment model
Holbrook, K. F., Ligon, J. T., and Hayes, J. C. (1986). Comparison of for non-cohesive sediment. Trans. Am. Soc. Agric. Eng.
methods for determination of eroded size distribution in sedi­ 29(5): 1300-1306.
mentology modeling, Paper N o . SER-86-206 presented at the Wilson, Β. N., Barfield, B. J., and Moore, I. D . (1982). " A Hydrology
Southeast Region Meeting of A S A E , Orlando, Fl. and Sedimentology Watershed Model," Part I; "Modeling Tech­
Lara, J. M., and Pemberton, E. L. (1963). Initial unit weight of niques." Department of Agricultural Engineering, University of
deposited sediments. In "Proceedings, Federal Inter-Agency Sed­ Kentucky, Lexington, KY.
imentation Conference," U.S. Department of Agriculture Miscel­ Yalin, M. S. (1963). An expression for bed-load transportation. J.
laneous Publication 1970. U.S. Government Printing Office, Hydr. Diu., Proc. ASCE 89(HY3):221-250.
Washington, D C . Yang, C. T. (1972). Unit stream power and sediment transport. / .
Mantz, P. A. (1977). Incipient transport in fine grains and flakes by Hydr. Diu., Proc. ASCE 98(HY10): 1805-1826.
fluids—Extended Shields' diagram. / . Hydraulics Diu. ASCE Yang, C. T. (1973). Incipient motion and sediment transport. J.
103(HY6):601-615. Hydr. Diu., Proc. ASCE 98(HY10): 1679-1704.
Martin, J. P., Martin, W. P., Paye, J. B., Raney, W. Α., and D e M e n t , Yang, C. T. (1984). Unit stream power equation for gravel. / . Hydr.
J. D. (1955). Soil aggregation. In "Advances in Agronomy (A. G. Div., Proc. ASCE 110(HY12).
Erosion and Sediment Yield

exception of additional cleanup costs after construction


INTRODUCTION is complete. Expenditures for sediment control are
therefore an economic liability. Since there are n o
Philosophy of Erosion and Sediment Control intrinsic economic incentives to control sedimentation,
T h e landscape of the earth has b e e n greatly influ­ it has b e e n necessary to appeal to the altruistic inter­
enced by the processes of soil erosion and sedimenta­ ests of o p e r a t o r s to implement controls. W h e n this has
tion. W h e n uninfluenced by the activities of h u m a n s , failed, regulations have b e e n imposed.
these processes often produce picturesque landscapes Soil erosion results w h e n soil is exposed to the
such as the G r a n d Canyon, Niagara Falls, and other erosive powers of rainfall energy and flowing water. It
places of natural beauty. However, when the landscape is not possible to conduct the massive earth-moving
is subjected to h u m a n activities with n o consideration operations necessary for major land disturbances with­
for conservation, the result is all too often a r a t h e r out exposing soil to these erosive forces. It is possible,
grotesque picture of fields riddled with gullies and however, to plan t h e operation and reclamation activi­
muddy streams whose channels are filled with sedi­ ties so that sediment production is minimized. T h r o u g h
ment. the use of properly designed and constructed conserva­
Soil erosion and sedimentation from construction tion practices and sediment control structures, t h e
and mining activities have many similarities with those adverse effects of land disturbances can b e m a d e mini­
from agricultural land. This is extremely helpful since mal.
agricultural erosion has b e e n studied for many years Large-scale a b a n d o n e d land disturbances will even­
resulting in the development of prediction algorithms tually heal themselves by natural processes, including
and control procedures. Economic incentives for ero­ natural revegetation and natural armoring. Curtis
sion control with construction and mining, however, (1971) analyzed the sediment yield from Appalachian
are drastically different from those for agriculture. Soil watersheds during periods prior to strip mining, during
loss from agricultural land represents a loss in nutri­ active operations, and subsequent to mining. H e found
ents and production capability, resulting in at least that sediment concentrations went from n e a r zero prior
some economic incentive to the landowner to reduce to mining to over 46,000 m g / l i t e r in runoff from large
erosion. In the absence of regulatory penalties, soil loss storms during mining operations. In the postmining
from a construction site or a surface mine represents period, sediment load decreased, even u n d e r poor
n o economic loss to the operator, with the possible reclamation, as the land t e n d e d to heal itself. This

238
Introduction 239

healing did not occur as a result of revegetation, but primarily from shearing forces of the channelized flow.
was a result of natural armoring by surface rocks that Tillage after a rainfall event will obliterate rills, form­
tended to cover the surface w h e n the fines were washed ing a new and different microrelief. Subsequent
away. T h e s e processes, along with natural revegetation, c o n c e n t r a t e d flow channels will occur in different loca­
proceed at differing rates, d e p e n d i n g on t h e climate, tions; thus rill location is generally assumed to be
geology, and chemistry of the exposed surface. Typi­ random.
cally, the rates are slow without h u m a n m a d e e n h a n c e ­ A s flow progresses further downslope and away from
ments, and environmental d a m a g e can be extensive the watershed divide, the location of channelized flow
prior to complete reclamation. Of course, a p p r o p r i a t e areas ceases to b e controlled by the microrelief and
reclamation and conservation practices dramatically becomes controlled by the prevailing macrorelief.
enhance natural reclamation processes and can elimi­ Tillage or landforming operations may obliterate any
nate undesirable environmental impacts. incised channels in these macrorelief drainage ways;
however, channels will t e n d to form again in the same
location unless t h e macrorelief is drastically changed.
The Erosion-Sediment Yield - Deposition Process Soil d e t a c h m e n t in macrorelief channels occurs primar­
ily as a result of t h e shearing forces of the channelized
Description of the Process flow, from headwall failure and from channel wall
Soil erosion involves detachment, transport, and sub­ failure. If t h e channels form in fields t h a t are being
sequent deposition (Meyer and Wischmeier, 1969). Soil tilled such that the resulting incised channels are oblit­
is detached both by raindrop impact and t h e shearing e r a t e d by subsequent tillage, t h e channels are called
force of flowing water. Sediment is transported down- e p h e m e r a l gullies (Foster, 1986). If t h e channels are
slope primarily by flowing water, although t h e r e is a p e r m a n e n t features with a dynamic headwall, vertical
small amount of downslope transport by raindrop channel walls, and dynamic features, they are called
splash. Runoff and resulting downslope transport d o classic gullies (Harvey et ai, 1985). T h e relationship
not occur until rainfall intensity exceeds infiltration b e t w e e n rills a n d e p h e m e r a l gullies is shown in
rate. For this reason, soil erodibility decreases as the Fig. 8.1.
infiltration rate increases. O n c e runoff starts, the quan­ A s flow moves still further away from the watershed
tity and size of material transported increases with the divide, channels t e n d to have relatively stable banks
velocity of runoff water. A t some point downslope, a n d n o headwalls with erosion occurring as a result of
slopes may decrease, resulting in a decreased velocity bed degradation from shear forces and channel wall
and transport capacity. At this point, sediment will be failure. A discussion of erosion and sediment transport
deposited, starting with the larger particles and aggre­ processes in these channels is given in C h a p t e r 10.
gates. Smaller particles and aggregates will be carried
further downslope, resulting in what is known as en­ Detachment
richment of fines. For this reason, t h e size distribution Erosion in interrill areas occurs primarily as a result
of eroded aggregates and primary particles has a major of raindrop impact. Interrill erosion is thus relatively
impact on soil erosion-deposition processes. i n d e p e n d e n t of slope length, after sufficient distance
downslope to g e n e r a t e enough runoff to transport
Terminology
e r o d e d sediment. Because interrill transport capacity
A n important part of developing an understanding from downslope splash a n d from overland flow does
of erosion and sedimentation literature is terminology. increase slightly with slope, interrill erosion is a linear
Soil eroded from exposed upland areas n e a r the water­ function of slope steepness. Rill erosion, on the other
shed divide is from rill and interrill areas. Interrill hand, is a strong function of both slope steepness and
areas are those zones between small channelized flows slope length. Since rill d e t a c h m e n t results from shear­
known as rills. Interrill erosion occurs in those areas ing forces of concentrated flow, increasing runoff depth
where flow is shallow overland or sheet flow and de­ and velocity will cause an increase in rill erosion. Both
tachment forces are primarily from raindrop energy of these increase with distance downslope. Shear and
impacting exposed soil. Rill erosion occurs when flow is resulting rill erosion also increase with slope steepness.
concentrated in microrelief channels with sufficient T h e impact of slope length and steepness on rill and
depth and slope to cause channel incision. T h e mi­ interrill erosion is illustrated in Fig. 8.2.
crorelief causing these small concentrated flow chan­
nels is generated by tillage or landforming operations. Deposition
Since flow depth in rills is typically sufficient to absorb Deposition of sediment can h a p p e n anywhere
falling raindrop impact energy, soil d e t a c h m e n t occurs downslope of the point of erosion, occurring when
240 8. Erosion and Sediment Yield

Figure 8.1 Illustration of rills and ephemeral gullies.

transport capacity of the flow is less than soil available et al. (1993a). Development of early empirical soil
for transport. Since transport capacity is a monotonic erosion models started m o r e than 50 years ago with
function of flow velocity, anything that reduces velocity efforts by Cook (1936), Zingg (1940), and Smith (1941)
in a flow segment increases deposition. Vegetal filters, to predict the impact of slope length and steepness,
terrace channels and check dams are examples of sedi­ cover, and supporting practices on soil erosion. Subse­
ment control practices that reduce velocity and in­ quently, Smith and Whitt (1948) developed a relation­
crease deposition. F u r t h e r discussion of these practices ship for soil erosion in Missouri in which annual
is given in Chapter 9. erosion from a claypan soil was used as a standard with
Components of the soil erosion-deposition process corrections for slope, slope length, soil classification,
are interrelated as shown in Fig. 8.3. Soil available for and conservation practice. Similar relationships were
transport at any slope segment is the sum of that developed for the cornbelt region. Based on a need for
carried from upslope plus that detached in the slope relationships for o t h e r regions, a national workshop
increment. Soil carried downslope is the lesser of was held in O h i o in 1946, which resulted in addition of
transport capacity or material available for transport. a rainfall factor and reappraisal of all previously devel­
Detailed information on all of the processes in Fig. 8.3 oped factors. T h e resulting so-called Musgrave equa­
are given in varying parts of Chapters 7, 8, 9, and 10. tion (Musgrave, 1947) included factors for rainfall,
surface runoff characteristics, slope steepness and
length, soil characteristics, and vegetative cover.
Erosion and Sediment Yield Modeling:
Success of the Musgrave equation led to a desire for
A Historical Perspective a truly national relationship that had universal p a r a m e ­
T h e information presented in this section is a sum­ ters and also led to establishment of the National
mary from several sources but primarily that of R e n a r d Runoff and Soil Loss D a t a C e n t e r at P u r d u e Univer-
Introduction 241

intensity, soil erodibility, slope length, slope steepness,


soil cover, a n d conservation practices, making it a
valuable tool for conservation planners. Because of
frequent misuse, Wischmeier (1976) strongly recom­
m e n d e d that its use b e limited to conditions for which
it was originally intended.
Most of t h e basic U S L E information was published
in 1965 in H a n d b o o k 282 with additional improvements
following in H a n d b o o k 537 in 1978. Subsequent to
H a n d b o o k 537, a n u m b e r of additional improvements
have b e e n m a d e , specifically

5 10 20 30 40 • revisions in the slope and slope length factors result­


ing from an evaluation of the original data base,
Distance Downslope (ft) • use of a subfactor a p p r o a c h for cover and manage­
m e n t factors with major new data available to define
Β the relationships,
0 First 60 min. c\ • use of a time-varying relationship to account for
• Second 60 min. \ J

r\
1.6 - N o Mulch — f r e e z e - t h a w effects on erodibility,
0
• • use of factors to account for susceptibility to rilling.
1.2 •
T h e s e d a t a have b e e n summarized in a new handbook
0.22 T/A

s
and the revised model described as the Revised U S L E
%

0.8 Straw
or R U S L E ( R e n a r d et al., 1993a). T h e R U S L E , in
addition to containing new p a r a m e t e r s , is available in
CL,
0.4 ο m 0.89 T/A _ computerized format.
" Straw
Efforts to develop a m o r e physically based approach

a
Ο
CO
0 Finely Sieved lead to a conceptual relationship known as the Meyer
model (Meyer and Wischmeier, 1969) shown in Fig. 8.3
ο and subsequently to the F o s t e r - M e y e r - O n s t a d ( F M O )
1.2 • model (Foster et al., 1977a, b). T h e F M O equation
£ a — • does not predict processes in individual rills, but rather
ο No Mulch • has algorithms to estimate the sum of all rill erosion on
CD 0.8
Ο • a slope segment. Instead of a combining rill and inter­
u • rill erosion estimates into o n e lumped prediction, in­
W 0.4 terrill erosion is predicted as a function of rainfall
Coarse Aggregates energy, interrill erodibility, slope steepness, and inter­
0 I I I I I I rill cover factor. Rill erosion is predicted as a function
0 5 10 15 20 25 30 of runoff volume, p e a k discharge, rill erodibility, slope
steepness, slope length, and rill cover and practice
Slope S t e e p n e s s (%) factors. Although separate factors were proposed for
Figure 8.2 Influence of slope length and steepness on interrill and
rill and interrill erosion, n o data base was developed
rill erosion. Pan size in Β was 2.5 X 2.5 ft and rainfall intensity was for the factors. T h e F M O model was used as the basic
2.5 i n . / h r . ( A ) Plot length effect on interrill and rill erosion. (B) relationship for the C R E A M S model (Knisel, 1980). A
Slope steepness effect on interrill erosion as measured in 2.5 X 2.5-ft modification to the F M O known as SLOSS was utilized
pans (after Meyer et al., 1975).
by Wilson et al. (1982, 1986) in the S E D I M O T II
model. T h e SLOSS modification includes procedures
for utilizing the F M O to calculate d e t a c h m e n t and the
sity in 1954. F e d e r a l - S t a t e cooperative research yielded Y a n g equation (Yang, 1972) to calculate sediment
10,000 plot years of data, which were analyzed by the transport capacity.
data center. T h e result was the Universal Soil Loss M o r e recent research has b e e n oriented toward pro­
Equation ( U S L E ) described by Wischmeier and Smith cess-oriented models that consider d e t a c h m e n t pro­
(1965, 1978) in Handbooks 282 and 537. T h e U S L E cesses in individual rills. Hirschi and Barfield (1988a, b)
describes erosion as a function of rainfall energy and developed a research-oriented single-storm model
242 8. Erosion and Sediment Yield

Soil F r o m
Upelope

Detachment Detachment Transport Transport


by Rainfall b y Runoff C a p a c i t y of R a i n f a l l C a p a c i t y of R u n o f f

Compare
Total Total
Detached Transport Capacity

If D e t a c h m e n t < Transport If D e t a c h m e n t > Transport

Soil C a r r i e d
Downslope

Figure 8.3 Interrelationship of detachment, transport and deposition of sediment (after Meyer and Wischmeier, 1969).

known as K Y E R M O , which predicted sediment de­ Storm et al. (1990). T h e following discussion draws
tachment and transport in individual rills as a function heavily on that document.
of flow rates and shear distribution, u n d e r the assump­
tion that rills were uniformly spaced and oriented u p -
and downhill. A national research t e a m collected a
large data base and developed a computerized continu­ Erosion Principles
ous simulation model, known as W E P P , that also p r e ­ Sediment Continuity Equation
dicts erosion in uniformly spaced rills assuming a rill
T h e relationship that is basic for fundamental ero­
spacing of 1 m (Lane and Nearing, 1989). T h e W E P P
sion processes is continuity of mass. For overland flow,
data base was used to develop prediction equations for
the continuity equation is (Foster, 1982)
the basic erosion parameters. F u r t h e r testing o n the
application of the model to watershed conditions is
needed, particularly for construction sites. R e c e n t re­ dq& d(cy)
search (Storm et al., 1990; Lewis et al, 1990) extended
the fundamental erosion process approach to include aT *-ir- ' "
+ D +D (81)

evaluation of the impacts of rill density distributions on


w h e r e q is sediment load, χ is distance downslope, p
s s
erosion.
is mass density of sediment particles, c is sediment
concentration, y is flow depth, / is time, D is rill r

erosion or deposition rate, and D is sediment deliv­ {

FUNDAMENTAL EROSION MODELING e r e d to t h e rill from interrill areas. T h e erosion param­


eters q , D , and D are m e a s u r e d p e r unit width of
s T x

T h e following is intended for the reader interested the field. T h e dq /dx s term represents the change in
in principles and concepts. Those interested primarily sediment flow rate along the slope, and p d(cy)/dt s

in applications should read subsequent sections on rill represents the change in sediment storage over time.
and interrill erosion modeling. A n overview of the state F o r flows that are shallow and gradually varied, the
of the art in fundamental erosion modeling is given by p d(cy)/dt
s storage t e r m may b e neglected, resulting in
Fundamental Erosion Modeling 243

a widely used steady-state continuity equation


Topsoil
Stockpile
= D + D,.
r (8.2)
dx Drainaf β Channel

Interaction between Sediment Load


and Transport Capacity ®
T h e interaction between sediment load and trans­
Β
port capacity is an important consideration in modeling Sediment Transport Rate
Sediment Detachment Rate
D in Eqs. (8.1) and (8.2). Foster and Meyer (1972a,
T

1975) proposed that rill detachment and deposition are


s
«*im e n t A < l d l t l

»
fro
st.c , o n

proportional to the difference between transport ca­ t p U

pacity and sediment load, or

(8.3)

where C, is a first-order reaction coefficient ( L ) , _ 1

and T is the sediment transport capacity. If it is


c

assumed that the maximum d e t a c h m e n t capacity, D , tc

is proportional to transport capacity, or

D r c = CX, (8.4)

then a relationship defining the interaction between


sediment load and transport capacity can be devel­
oped, as (Foster and Meyer, 1972a)
Figure 8.4 Illustration of interactions between sediment load, rill
(8.5) detachment and deposition, and flow transport capacity. (A) Plan
view. (B) Transport rate and detachment rate changes along the
channel.
From Eq. (8.5) it can b e seen that rill d e t a c h m e n t
equals detachment rate capacity when the sediment
load equals zero. Conversely, when sediment load
equals transport capacity, D approaches zero.
r where V is settling velocity, and q is flow discharge
s

Foster and Meyer (1972a) utilized the stream power p e r unit width of plot. W h e n Eq. (8.3) is applied to
concept to lend credence to these concepts, claiming channelized flow, Foster (1982) proposed that deposi­
that the finite amount of energy in a specified flow may tion at a given flow rate would be greater than that for
be used for either detaching or transporting sediment overland flow and that C should b e given by
1

particles, but the same energy cannot be used for both.


T h e D /D
r rc term in Eq. (8.5) represents the relative Cx = V /q.
s
(8.7)
amount of energy expended on sediment detachment,
and the term q /T s c indicates the relative amount of T h e relationships proposed in Eqs. (8.3)-(8.6) are
energy expended on sediment transport. T h e sum of illustrated conceptually in Fig. 8.4. In this figure, sedi­
these two terms equal unity, the total relative available m e n t free flow is assumed to exit the parking lot and
energy. e n t e r a channel with an erodible surface. Soil is de­
These concepts may also be applied to deposition. tached at a rate that decreases exponentially with
W h e n applying Eq. (8.3) to deposition, it is clear that distance until sediment load equals transport capacity.
C should be varied with sediment size to account for
x
W h e n additional sediment is a d d e d to the flow at the
the differing settling velocities between particle sizes. topsoil stockpile, deposition must occur until the sedi­
T o date, however, there are no validated relationships m e n t load again equals transport capacity. This simple
describing this p h e n o m e n o n . O n the basis of results by illustration shows how d e t a c h m e n t changes with sedi­
Einstein (1968), Foster (1982) proposed that C for x
m e n t load and transport capacity. In summary, detach­
overland flow may be approximated by m e n t capacity or potential is based on localized flow
conditions at a point, but the actual detachment rate is
C, = V /2q, s (8.6)
limited by available excess transport capacity.
244 8. Erosion and Sediment Yield
Foster and Meyer Closed Form Erosion Equations capacities at χ = L , respectively, and C is an integra­
0 2

Solutions to Eqs. (8.1)-(8.7) can be developed nu­ tion constant. If uniform slope and constant rainfall
merically; however, Foster and Meyer (1972a, 1975) excess are assumed, the resulting relationships are
developed a closed-form erosion model by simultane­
ously solving these equations utilizing auxiliary rela­
^ - - - ( 1 - 0 ) ( 1 - e- *)a~ ax l
(8.17a)
tionships. They assumed steady-state rainfall, uniform f
c0
shallow flow, and a steady-state rainfall excess given by

^ = - / - ( ! - 0 ) 0 - e - * * ) (8.17b)
(8.8) i
c0 ^0

ft (1 - 0)(1 -e~ *) ax

where q is rainfall excess at χ = L . In addition,


0 0 (8.17c)
potential detachment and transport rates, D and T , ft c 0 a
c c

respectively, were defined as


T h e general closed-form erosion equation is applica­
ft = C r3/2 D (8.9) ble to both d e t a c h m e n t and deposition, as well as
and several nonuniform slope conditions (Foster and Meyer,
1975). T h e erosion equation has b e e n validated using
71 = C r /2T
3
(8.10) field data (Foster and Meyer, 1972a, 1975) producing
acceptable predictions. In addition, the closed-form
where τ is the average shear stress on the soil surface,
equation was used in C R E A M S (Foster et al., 1980a,
and C and C are constants. Using auxiliary relation­
D T
b) for identifying d e t a c h m e n t and deposition of over­
ships from the Chezy equation, ft and can be 0
land and channelized flow.
defined as the detachment potential and transport ca­
pacity at χ = L , leading to the relationship
0

ft Is Example Problem 8.1 Illustration of the


(8.11)
D c0
+
Tc0
g
* !
closed-form equation

where At the end of a 100-ft slope, the sediment detachment


ft potential is 0.01 lb/sec · ft and the transport capacity is 2.0
2

8* rp (8.12) lb/sec · ft. Estimate the sediment load at 2, 10, 50, 75, and
i
c0 ft c 0 100 ft downslope, assuming steady uniform flow and an
and interrill detachment rate of 0.005 lb/sec · ft . 2

Solution:
X
(8.13) 1. Calculating parameters.

Using Eq. (8.11) with (8.2), Foster and Meyer ob­ * * = — = — = 0.01*
"0 100
tained a general solution to their equation as
L D (100 ft)(0.01 lb/sec · f t ) 2

ft
0 c 0

e x p * dx* + C
ax a = = = 0.5
= e 2
T 2.0 lb/sec · ft
ft c 0 c0

(8.14) L Dj 0 (100 ft)(0.005 lb/sec · f t ) _ 2


q ^
2.0 lb/sec · ft
where

2. Setting up the equations. From Eq. (8.17a)


a = (8.15)
»c0
<?s=[** -(l-e)(l-e-« *)/a]T x

and
c0

q = [x* - (1 - 0.25)(1 - e - ° * * ) / 0 . 5 ] [ 2 . 0 ]
s
5

θ = (8.16)
£
c0
or
where χ is the distance downslope, L is the slope 0

length, fto and Τ are the detachment and transport


Λ
q = 2.0x* - 3 . 0 ( 1
s -e- *). 05x
Fundamental Erosion Modeling 245

3. Tabulating the results. information, Eq. (8.19) may be used to estimate soil
splash from areas with various soils, topographic, and
X a
s surface conditions.
(ft) x+=x/L 0 (lb/sec-ft) Surface ponding dissipates raindrop energy; thus it
has also b e e n shown to affect significantly splash ero­
2 0.02 0.010
sion (Palmer, 1965; Mutchler and Larson, 1971). In
10 0.10 0.054
general, these studies show that splash increases u p to
50 0.50 0.336
a p o n d e d d e p t h of £ to 1 d r o p diameter with a de­
75 0.75 0.562 crease at d e e p e r depths.
100 1.00 0.819
Raindrop Splash Transport T h e concentration of de­
tached soil particles in r a i n d r o p splash is proportional
to the kinetic energy of raindrops and should be equal
in all directions. T h u s on a level surface, the net soil
Comments on Closed-Form Equation transport should b e zero even though the concentra­
T h e Foster and Meyer (1975) closed-form erosion tion of particles in t h e splash may be high. However,
model was developed using equations that describe t h e on a sloping surface, t h e splash trajectory goes farther
underlying relationships between detachment, deposi­ downslope t h a n upslope, just by geometric considera­
tion, and transport capacity. T o predict erosion rates, tions. T r a n s p o r t by splash should, therefore, b e pro­
the user must combine the closed-form equations with portional to both kinetic energy and slope, prompting
models of detachment capacity and transport capacity. Q u a n s a h (1981) to propose that
In this combined form, the closed-form model has b e e n
used as the building block for other models such as Q = a* V,
t e (8.20)
C R E A M S and the m o r e recent W E P P model.
where Q is t h e net splash transport (kilogram per
t

square meter), k is the total applied kinetic energy


e

Interrill Erosion (joule p e r square meter), s is t h e soil slope, and a, b,


Splash Erosion and c are constants. In general, t h e transport slope
exponent has b e e n found to be much higher than the
Most soil detachment in interrill areas is a result of d e t a c h m e n t slope exponent, indicating that splash
the force of falling raindrops (Young and Wiersma, transport is much m o r e sensitive to slope than splash
1973). For erosion to occur, detached particles must be detachment.
transported from the site of detachment. T r a n s p o r t of
detached soil particles occurs by raindrop splash as
Overland Flow Transport and Detachment: Sheet Erosion
well as by overland flow. Each of these concepts is
Overland flow transport and detachment results from
discussed separately in the following sections.
shearing forces of t h e thin film of runoff water in
interrill areas. Average shear stress in this film is the
Raindrop Splash Detachment T h e quantity of de­ product of the surface slope and overland flow depth, a
tached soil in interrill areas is related to the kinetic value that will b e very small and typically less than a
energy and m o m e n t u m of raindrops impacting the soil critical tractive or shear force necessary to initiate
surface (Free, 1960; Q u a n s a h , 1981; Gilley and Finkner, d e t a c h m e n t (Foster a n d Meyer, 1975). Although the
1985; Rose, 1960). Bubenzer and Jones (1971) p r o ­ shear forces are low, which would indicate minimal
posed that splash erosion could be estimated from transport capacity, r a i n d r o p impact o n the surface in­
creases the turbulence level and enhances transport
5 = ai k%p*
b
(8.18)
S
capacity. T h u s , overland flow is responsible for most of
where s is the soil splash, i is rainfall intensity, k is the transport of soil to rills, while detachment is pri­
s e

the total applied rainfall kinetic energy, p is the soil marily from r a i n d r o p impact (Young and Wiersma,
c

percent clay, and a, b, c, and d are constants. Hirschi 1973).


and Barfield (1988a) and Q u a n s a h (1981) suggest in­ Sediment transport equations developed for stream-
cluding the slope s as a parameter, or flow conditions have b e e n applied to overland flow
conditions, with varying degrees of success. T h e Yalin
5 S = ai klp s ,
b d
c
e
(8.19) (1963) equation is t h e most widely used transport equa­
tion for overland flow (Hirschi and Barfield, 1988a, b;
where s is slope, and e is a constant. Given sufficient Dillaha and Beasley, 1983). T h e Yang (1972) unit
246 8. Erosion and Sediment Yield
stream power equation has also been applied to over­ times when rainfall rate exceeds infiltration rate. T h e
land flow (Moore and Burch, 1986); however, it yields baseline erodibility factor, K , may b e estimated from
{

better results u n d e r d e e p flow conditions. soil properties using relationships developed by Alberts
et al. (1989). Canopy and ground effects are discussed
Net Interrill Erosion in a subsequent section on the W E P P model.
T h e quantity of sediment actually delivered to a As can be seen from the equations presented, a high
concentrated flow network is net interrill erosion. degree of empiricism exists in t h e description of inter­
Computationally, it is the net of detached soil particles rill erosion in current models. Operational models in
minus deposited material. A conceptually sound the near future are likely to continue to contain such
method for estimating net interrill erosion would be to empiricism d u e to t h e complexity of the processes.
predict interrill detachment and transport capacity sep­
arately and use a computational framework like that in
Fig. 8.3 to quantify the amount of sediment actually Rill Erosion
reaching rills (Hirschi and Barfield, 1988a, b). A fre­ Rill Networks
quently used alternative approach is to lump the pro­
cesses together in a regression equation. Since both Development and evolution of rills and rill networks
detachment and transport are functions of kinetic en­ are important erosion processes. Although significant
ergy, which is in turn a function of rainfall intensity, progress has been m a d e in describing the growth and
Meyer (1981) proposed that net interrill erosion on development of individual rills, minimal progress has
b a r e soil could be predicted as a simple function of been m a d e in describing rill networks. F u t u r e progress
rainfall intensity and soil properties, or in erosion modeling will hinge on better description of
the network processes in what is known as dynamic
D = al ,
{
b
(8.21) erosion models.

where D is net interrill erosion, / is rainfall intensity,


{ Network Development Rill networks are initiated on
and a and b are constants. T h e b exponent increases b a r e soil as a result of the microrelief when flow
as the percentage clay decreases, or (Meyer, 1981) concentrates in the microrelief channels. Initially, the
channels tend to be parallel (Leopold et al., 1964), but
b = 2.1 - F : clay content 2 0 - 5 0 % (8.22a)
c l
overtopping of the ridges between rills as a result of
b = 2.0: clay content > 5 0 % , (8.22b) changing flows results in some channels flowing at
angles to t h e prevailing slopes and ultimate coales­
where F is the fraction of clay.
cl cence of rills into a dendritic network (Mosley, 1974).
Erodibility and crop factors can be a d d e d into Eq. In laboratory studies, Mosley (1974) found that equilib­
(8.21) as was done in the U S D A W a t e r Erosion Predic­ rium sediment discharge rate for constant rainfall was
tion Project ( W E P P ) (Nearing et al., 1989) resulting in related to total rill network length.

D = K I?C G (R /W),
{ { e e s (8.23)
Rill Density and Its Effects on Erosion Rill density is
where K is a baseline interrill erodibility for b a r e soil,
x
the n u m b e r of rills p e r unit width. Studies have shown
7 is effective rainfall intensity, C and G are dimen­
e e e
that density varies with a n u m b e r of factors such as
sionless canopy and ground cover effects on interrill slope steepness and length, runoff rate, soil texture,
erosion, respectively, R is rill spacing, and W is a
s
soil erodibility, and the presence or absence of rainfall
computed rill width. T h e effective rainfall intensity in (Meyer a n d M o n k e , 1965). O n highly erodible soils, rill
Eq. (8.23) is calculated only for periods of rainfall density has b e e n shown to be high and rills have t h e
excess since detached soil is not transported substan­ same size from point of origin to end, indicating trans­
tially during periods without surface runoff. Account­ port-limiting conditions (Ellison and Ellison, 1947). O n
ing only for the periods of rainfall excess, t h e effective less erodible soils, rill densities are less and the rills
rainfall intensity then becomes (Nearing et al., 1989) vary in width and d e p t h from beginning to end, indica­
tive of detachment-limiting conditions. Meyer and
M o n k e (1965) observed that short slope lengths have
/2 = i
e Γ / Λ, 2 2
(8.24) higher rill densities relative to longer lengths.
r
e
J
t x

Mathematical models of rill density are extremely


where / is rainfall intensity, t is time (seconds), t is e
limited. Li et al. (1980) developed a rill density model
the cumulative time during which rainfall exceeds in­ for laminar and turbulent flows, expressed on a unit
filtration rate, and t and t are the start and end
x 2
width basis. Their model assumed that all rills were the
Fundamental Erosion Modeling 247

same size for a specified distance downslope. Numer­ rate, τ is flow shear stress along the rill boundary, r is c

ous empirical constants are required, limiting its use. the critical shear stress necessary to detach soil parti­
Foster and Lane (1981) criticized the Li et al. model, cles, and a and b are constants. T h e constant b is
indicating that their choice of a representative particle close to 1.0 a n d is typically assumed to be 1.0.
size in t h e Shield's diagram caused critical tractive If d e t a c h m e n t is to be estimated in a channel, it is
force to be underestimated. necessary to know the shear distribution along the
T h e effects of rill density on erosion are somewhat boundary. T h e average b e d shear stress, r , for uni­
a

unclear. Using the K Y E R M O model, Hirschi a n d form flow is given as


Barfield (1988b) performed a sensitivity analysis on t h e
r = yRS, (8.26)
n u m b e r of rills across a plot and found that the maxi­
a

m u m sediment yield occurred at about six rills in 15 ft w h e r e y is specific weight of water, R is hydraulic
for their test conditions. They proposed that the d e ­ radius, and S is rill b e d slope. Since the distribution of
cline in sediment yield at higher n u m b e r s of rills was t h e shear a r o u n d t h e rill boundary is nonuniform, the
d u e to lower flow rates in each rill as the surface runoff use of average shear stress to estimate detachment
was distributed over more rills. They also showed that potential could result in significant errors in estimating
the effect of rill n u m b e r on sediment yield is governed channel shape.
by the form of the rill d e t a c h m e n t and boundary shear
stress equations. Rill Geometry T h e importance of rill shape to rill
T h e W E P P Erosion model represents a rill network growth and development is primarily a result of its
as a series of parallel rills; hence rill density is analo­ effect on shear distribution. Prior to reaching a
gous to rill spacing. Based on a sensitivity analysis of nonerodible layer, rill s h a p e may b e approximated as a
the W E P P model, Nearing et al. (1989) indicated that rectangle ( L a n e a n d Foster, 1980; Foster and Lane,
W E P P predictions were somewhat insensitive to rill 1983) with a width given as
density; hence a default rill spacing of 1 m is used in
W=aQ , b
(8.27)
the W E P P model. Such an assumption, however, ig­
nores the complex interactions of rill networks. Lewis w h e r e W is channel width, Q is discharge in the rill,
et al (1990) developed a model similar to t h e W E P P and a and b are constants. Although other shapes
model, but with the capability of utilizing a r a n d o m have b e e n suggested (Rohlf, 1981), it is likely that a
distribution of rill n u m b e r s and flow in rills. Results rectangular cross section is developed as a result of
from a sensitivity analysis from the Lewis model shows side sloughing once the rill e n c o u n t e r s a nonerodible
that ignoring the stochasticity of rill networks can m a k e layer.
a significant difference in predicted erosion w h e n a
nonerodible layer is encountered, as frequently occurs Critical Shear Stress A soil's resistance to the shear­
in tilled soils. T h e presence of a nonerodible layer is ing forces of c o n c e n t r a t e d flow is determined by the
not considered in W E P P . critical shear stress, sometimes referred to as critical
tractive force. F o r noncohesive soils, Shields diagram
Growth and Development of Individual Rills (Shields, 1936) is t h e m e t h o d most widely used to
Development and growth of an individual rill is describe critical tractive force of individual particles.
governed by rill d e t a c h m e n t potential, transport capac­ T h e original Shields diagram does not apply to sedi­
ity, sediment load, and their interactions. T h e follow­ m e n t particles of low specific gravity and small diame­
ing sections discuss rill d e t a c h m e n t potential, which ter; however, M a n t z (1977) extended the Shields dia­
includes detachment from rill incision, headwall cut­ gram for smaller particles as shown in C h a p t e r 7.
ting, and sidewall sloughing. For cohesive materials, critical shear stress has b e e n
related to a n u m b e r of soil properties including soil
Rill Incision Shear stresses along a concentrated flow shear strength, soil salinity, and moisture content
boundary will lead to incision of a channel if the shear (Alberts et al., 1989); percentage clay, m e a n particle
exceeds the critical tractive force. T h e rate of soil size, dispersion ratio, vane shear strength, organic mat­
detachment in rills due to rill incision is typically ter content, cation exchange capacity, and c a l c i u m -
assumed to vary with shear excess and may be ex­ sodium ratio (Lyle and Smerdon, 1965); and plasticity
pressed as (Foster, 1982) index (Smerdon a n d Beasley, 1961). Foster (1982) rec­
o m m e n d e d t h e equation of S m e r d o n and Beasley (1961)
D r c = a ( T - r ) \
c (8.25) based on the dispersion ratio; however, Hirschi and
Barfield (1988a) used the relationship from Smerdon
where D rc is the maximum or potential rill d e t a c h m e n t a n d Beasley (1961) based on percentage clay.
248 8. Erosion and Sediment Yield

Typical critical shear stresses range from 1 to 30 Pa. (1992) summarize the recent studies on headwall cut­
For agricultural soils, Foster and Meyer (1975) recom­ ting and channel scour and propose a model, denoted
m e n d e d an average value of 2.4 Pa. Alberts et al. as C H A N N E L , that predicts channel incision, headwall
(1989) developed regression equations using the exten­ development, propagation, and washout. T h e model
sive W E P P field data (including corrections from considers d e t a c h m e n t as a function of shear excess as
Flanagan, 1990) and found that critical shear stress of given in Eq. (8.25) and an interaction of transport
cropland soils may be predicted from very fine sand capacity and sediment load as given by Eqs. (8.3) and
fraction, calcium carbonate fraction, sodium absorption (8.5). Shear downslope of the headwall is calculated by
ratio, soil specific surface area (milligrams ethylene submerged jet theory or by impinging jet theory, as
glycol mono-ethyl ether adsorbed per gram soil), sand appropriate. Rohlf (1993) summarizes recent studies
fraction, water-dispersible clay fraction, and clay frac­ on sidewall stability. They proposed a dynamic model
tion. For cropland soils with clay fraction greater t h a n of channel wall failure that includes the effect of water
0.30, Alberts et al. (1989) found that critical tractive movement into and out of the channel wall on slope
force may be predicted from volumetric water content. stability. T h e analysis requires simultaneous solution of
O t h e r relationships are being developed from the s a t u r a t e d - u n s a t u r a t e d flow equation for groundwater
W E P P data set (i.e., Wilson 1993). N o doubt, the final movement (see C h a p t e r 11) and stress-strain dynamics
result will be considerably different from the original for channel wall stability.
relationships.
Rill Erosion Models
Sidewall Sloughing and Headwall Cutting Sidewall T o date, physically based rill erosion models have
sloughing and headwall cutting are significant mecha­ b e e n based almost exclusively on shear excess con­
nisms in the propagation of rills. Sloughing results cepts. Assuming a rectangular channel cross section
from gravitational forces, flow hydraulics, and their and erosion based on shear excess, L a n e and Foster
combined effects. A sidewalPs resistance to failure (1980) and Foster and Lane (1983) developed a deter­
varies with slope geometry and soil properties such as ministic channel erosion model that was incorporated
cohesion, bulk density, void ratio, moisture content, into C R E A M S for describing e p h e m e r a l gully growth
and others. For rills, sidewall failure typically results in a tilled agricultural field. It has also been applied to
from gravity forces acting on an overhang caused by rill erosion by Storm et al. (1990) and Lewis et al.
undercutting, and translational or rotational slips (1990).
caused by shear failure along an internal surface. In the Foster and Lane (1983) model, channel devel­
Translational slips occur along a plane typically paral­ o p m e n t is partitioned into two distinct stages for steady
lel to the surface slope, and rotational slips occur along flow rate. During t h e initial stage, t h e channel bottom
a circular arc. erodes uniformly downward at a width equal to their
Hirschi and Barfield (1988a) incorporated rill side- so-called equilibrium width. A second stage of develop­
wall stability into their erosion model based on a m e n t occurs w h e n the channel bottom reaches a
critical slope concept. O n c e the sidewall reaches a nonerodible layer, after which lateral expansion is as­
critical slope, it was assumed to slough off forming sumed to occur with vertical sidewall sloughing. A
a stable slope and depositing the detached soil mass detailed discussion on the development and implemen­
into the rill. Bradford et al. (1973) developed a two-di­ tation of the model is presented in a subsequent sec­
mensional rotational slip-type bank failure model. They tion on concentrated flow modeling.
found that the factors controlling sidewall stability are Models of rill erosion can also be developed by
water table height, soil cohesion, and seepage rate. calculating the shear distribution around a rill (Rohlf,
T h e upslope propagation of rills through headcutting 1981; Hirschi and Barfield, 1988a; Fogle et al., 1992).
may contribute significant amounts of sediment. T h e A n erosion model incorporating rill network develop­
headwall is an abrupt break in the longitudinal channel m e n t was presented by Mossaad and W u (1984). Their
profile (Schumm et al., 1984) and is the transition model combined a stochastic surface roughness model
between wide shallow channelized flow and narrow with a deterministic interrill and rill erosion model
d e e p e r flow. Individual headcuts migrate upstream, describing rill network development over time.
and several headcuts may exist along the same channel. A process-based soil erosion model was developed
Most of the available literature on headcut develop­ for W E P P (Nearing et al., 1989). T h e rill erosion
ment and propagation is based on drainage basin mor­ model component is based on shear excess, and rill
phology. d e t a c h m e n t d e p e n d s on rill erodibility, hydraulic shear
Recent developments have included models of chan­ stress, surface cover, below-ground residue, consolida­
nel erosion that include physically based models of tion, and the ratio of sediment load to transport capac­
headwall cutting and sidewall sloughing. Fogle et al. ity. Net deposition is assumed proportional to excess
Rill and Interrill Erosion Modeling: USLE / RUSLE Empirical Models 249

sediment load, and sediment routing is performed by Κ is t h e soil erodibility factor, which is the rate of
applying the steady-state continuity equation. soil loss p e r unit of R (erosion index units) for a given
Average shear stress for a rectangular rill is found soil u n d e r continuous fallow with u p and downhill
using a rill width calculated from rill flow discharge, cultivation on a slope of 9 % with a slope length of 72.6
average slope gradient, hydraulic radius, and the ratio ft (22.1 m).
of friction factors for the soil and the total cross L is the slope length factor, which is the ratio of soil
section in the rill. T r a n s p o r t capacity in rills is found loss from a defined slope length relative to that from a
using a simplified transport equation, calibrated with slope length of 72.6 ft (22.1 m).
the Yalin transport equation. F u r t h e r details are given 5 is the slope steepness factor, which is the ratio of
in a subsequent section on theoretical models. soil loss from a slope with a given steepness relative to
that from a 9 % slope.
C is t h e cover a n d m a n a g e m e n t factor, which is the
RILL AND INTERRILL EROSION MODELING: ratio of soil loss from an a r e a with a given cover and
USLE / RUSLE EMPIRICAL MODELS m a n a g e m e n t relative to that from an identical area in
continuous fallow.
Basic Relationships Ρ is the supporting conservation practice factor,
which is t h e ratio of soil loss from a field with a
Soil erosion by water is the soil lost from a given
conservation support practice such as contouring rela­
slope, usually predicted on a per unit a r e a basis. Sedi­
tive to that with straight row farming u p - and downhill.
ment yield, on the other hand, is the a m o u n t of sedi­
ment that passes a given point on a watershed. Some of
T h e R U S L E was designed to b e of identical form to
the sediment that leaves a given slope is deposited;
t h e U S L E so that the U S L E p a r a m e t e r s could be used
hence, sediment yield and soil erosion are not the same
w h e r e desirable. This interchange is particularly useful
and should not b e confused as such.
in t h e case of t h e C factor. T h e R U S L E approach to
T h e Universal Soil Loss equation, ( U S L E ) , as dis­
t h e C factor is sufficiently complicated to require a
cussed earlier in t h e historical perspective, was devel­
c o m p u t e r for solution of most practical problems. For
oped to predict soil erosion, not sediment yield
many estimates, t h e accuracy of the U S L E data base
(Wischmeier and Smith, 1965, 1978). In fact, t h e t e r m
for C is a d e q u a t e . In this text, t h e U S L E data base for
soil loss is somewhat of a misnomer. Much of the soil
C factors is included along with some of the data base
displaced in the erosion process is deposited subse­
for the R U S L E . F o r simplicity, further reference to the
quently in flatter a n d / o r vegetated areas where t h e
R U S L E / U S L E are simply to R U S L E ; however, where
transport capacity is lower. T h u s not all eroded soil is
distinction between t h e two a r e important, the differ­
"lost" from the field.
ences will b e pointed out.
T h e U S L E is a relationship that has b e e n widely
Several cautions about the use of R U S L E should be
used for planning purposes to predict the impact of
considered. Predictions from R U S L E represent soil
land use on soil erosion. Originally developed to p r e ­
loss averaged over many storms a n d years. It also
dict annual soil erosion averaged over long time peri­
r e p r e s e n t s averages over a total field or disturbed area.
ods, it has b e e n modified to estimate monthly a n d
A t points on a slope or field, the soil loss will almost
single-storm erosion. As would be expected, t h e stan­
always be less t h a n or greater than t h e average values.
dard error of prediction increases for short-term and
For example, on long slopes, the u p p e r part of the
single-storm predictions.
slope will have lower erosion rates t h a n the lower part
Improvements to the U S L E based on m o r e recent
of the slope, but t h e average over the entire slope over
data as well as a new evaluation of the original U S L E
a long period of time should be approximated by
data base have resulted in a modification known as the
R U S L E predictions. Also, t h e energy content of rain­
Revised U S L E or R U S L E ( R e n a r d et al, 1993a). T h e
fall with a given intensity, as predicted by the R factor,
R U S L E / U S L E are multiplicative relationships, or
r e p r e s e n t s a value averaged over a large n u m b e r of
A = RKLSCP. (8.28) storms over a long period of time. T h e value for a
given storm could be m u c h greater or less than that
A is the average soil loss per unit of area, expressed predicted by t h e R factor.
in units selected for Κ and the time period specified by T h e original U S L E has b e e n extended to forest
R. Normal English units are t o n s / a c r e / y e a r , b u t o t h e r conditions and to construction applications. T h e
units are used. R U S L E manual does not include information for these
R is the rainfall/runoff factor, which is the n u m b e r conditions, but refers the r e a d e r to Dissmeyer and
of rainfall units for rainfall energy and runoff, plus a Foster (1981, 1984). T h e Dissmeyer and Foster data
factor for runoff from snowmelt. are incorporated into t h e tables in this chapter. Also
250 8. Erosion and Sediment Yield
1400

0 1 2 3 4 5 6 7 8 9 10

Rainfall I n t e n s i t y (in/hr)
Figure 8.5 Relationship between rainfall intensity and rainfall energy. T h e prediction equations are
those given by Eqs. (8.29a) and (8.29b). Data points are from Hirschi et al. (1983).

incorporated are data from other sources for construc­ Tonsf refers to force of rainfall impact as opposed to
tion and mining. T h e R U S L E is available in computer­ tons mass of sediment being e r o d e d . O t h e r relation­
ized format from the Soil Conservation Society of ships have b e e n proposed for kinetic energy, but the
America. relationship given in Eq. (8.29a) is t h e most widely
used.
In t h e R U S L E , t h e relationship of Brown a n d Foster
Rainfall Energy Factor R
(1987) was used for t h e W e s t e r n U.S., o r
Selection of the R Factor e = 1099[1 - 0 . 7 2 e x p ( - 1 . 2 7 / ) ] . (8.29b)
After evaluation of correlations between soil erosion R e n a r d et al. (1993a, Appendix C) r e c o m m e n d t h e use
and a n u m b e r of rainfall parameters, t h e R factor of Eq. (8.29b) in all future calculations of rainfall
selected by Wischmeier and Smith (1958) was t h e prod­ energy for all sections of t h e U.S., since it seems to
uct of rainfall energy and maximum 30-min intensity b e t t e r fit t h e data at lower intensities. T o convert to
divided by 100 for numerical convenience, known as total energy in a storm, e is multiplied by t h e d e p t h of
the EI 30 index. O n an annual basis, t h e EI value is
30 rainfall, or
the sum of values over the storms in an individual year. Ε = eP, (8.30)
R factors for the U S L E and R U S L E are identical
w h e r e Ε is total energy in a storm in ft · t o n s f / a c r e
except for additional relationships in t h e R U S L E that
and Ρ is total storm d e p t h of rainfall in inches.
account for ponding.
D a t a on rainfall energy a r e plotted in Fig. 8.5 along
Calculations of rainfall energy require an algorithm
with a plot of Eq. (8.29). This illustrates that t h e energy
relating energy to some measurable p a r a m e t e r . U p to
content predicted by E q . (8.29) is a value averaged over
an intensity of 3 i n . / h r , rainfall energy increases with
a large n u m b e r of storms. F o r any individual storm, the
storm intensity as a result of the fact that t h e d r o p size
actual energy could b e much larger o r smaller than
and fall velocity increase with intensity. Above 3 i n . / h r ,
that predicted.
the drop size reaches its maximum size and energy
Using t h e EI index, the R factor in Eq. (8.28) is
remains constant. Based on an analysis of G u n n a n d 30

given by
Kinzer (1949) data, Wischmeier and Smith (1958) pro­
posed that rainfall energy is related to intensity by R = Σ£/ /100, 3 0 (8.31)

e = 916 + l o g ι i < 3 in./hr where I is t h e maximum 30-min intensity (iph) for


30
10

the storm a n d t h e 100 in t h e d e n o m i n a t o r is used for


e = 1074 ι > 3 in./hr, (8.29a) numerical convenience. Details on computation of
where e is the kinetic energy in ft · t o n s f / a c r e · in. rainfall energy computation are given in R e n a r d et al.
and i is the average intensity of t h e storm in i n . / h r . (1993a, Appendix C).
Rill and Interrill Erosion Modeling: USLE / RUSLE Empirical Models 251

MOO

L12 5

Figure 8.6 Isolines of annual R factor for the Eastern U n i t e d States (after Renard et al., 1993b). Maps for areas in and
west of the Rockies are in Appendix 8A. R factor is in ft · tonsf · i n . / a c r e · hr · year. T o convert to SI units,
MJ · m m / h a · h - y, multiply by 17.02.

Units on the annual R factor are h u n d r e d s of ft · Average Annual R Values


tonsf · i n . / a c r e · hr · year. T h e h u n d r e d s t e r m comes Using Eq. (8.31), Wischmeier and Smith (1965) com­
from the 100 divisor in Eq. (8.31). In further reference, p u t e d values for t h e R factor for stations east of the
the units on R will be given as ft · tonsf · i n . / a c r e · Rockies a n d extended t h e values to t h e W e s t e r n States
hr · year for simplicity and t h e h u n d r e d s will be u n d e r ­ utilizing a correlation b e t w e e n 2-year, 6-hr rainfall and
stood. T h e units on single storm erosivity would b e t h e R factor (Wischmeier and Smith, 1978). In the
ft · tonsf · i n . / a c r e · h r · storm. development of R U S L E , additional analyses were m a d e
T o convert the annual R to SI units, M J · m m / h a · on longer d a t a bases and on hourly rainfall for stations
hr · year, multiply the English Units by 17.02. A sum­ in a n d west of t h e Rockies. T h e d a t a are summarized
mary of other conversion factors is given in t h e A p ­ in Fig. 8.6 for the e a s t e r n U n i t e d States.
pendix. As a note ft · tonsf · i n . / a c r e · hr · year m e a n s
Corrections for Ponding
ft tonsf in. Studies have shown that p o n d e d water on flat slopes
acre hr year t e n d s to absorb rainfall energy and retard erosion
252 8. Erosion and Sediment Yield

(Mutchler and Larson, 1971). T h e R U S L E contains a


modification for R that is a function of the R factor
and slope, as shown in Fig. 8.7. T h e corrected R fac­
tor is developed by multiplying the uncorrected R
factor from Fig. 8.6 by the correction factor in Fig. 8.7.
T h e correction factors in Fig. 8.7 are for annual ero­
sion. Corrections for single storms have not b e e n de­
veloped; however, as a first estimate, o n e might use
Fig. 8.7.

Distributions by Months
T h e R factor is not uniformly distributed over the
year, but has a monthly distribution that varies widely
with location. R factor distributions have been devel­
o p e d for the various zones in the U.S. These zones are
shown in Fig. 8.8. D a t a on the cumulative percentage
of annual R by months is given in Table 8.1 for
selected zones with a full tabulation given in A p p e n ­
dix 8A.

Single Storm R Factors


Although t h e U S L E a n d R U S L E w e r e developed for
predicting long-term annual soil erosion, the relation­
ship has b e e n used to predict single-storm erosion. In

Figure 8.8 R factor distribution zones (after Renard et al., 1993b). Zones 1-120 are for the contiguous U.S. Zones
1 2 1 - 1 3 9 are for Hawaii. Consult the local conservation office in Hawaii for distribution zones. Z o n e 140 is for Pullman, W A
and Northwest dryland winter wheat.
Rill and Interrill Erosion Modeling: USLE / RUSLE Empirical Models 253

Table β. 1 R Factor Distributions for Selected Zones Shown in Table 8.2 Historic Single Storm R Factors for Selected Cities
Fig. 8.8 (after Renard et al, 1993b) fl
(after Wischmeier et al, 1978)*

Geographic area from Fig. 8.8. Index values normally exceeded


once in Ν years
Date 10 25 86 104 105 112 119
Location N= 1 2 5 10 20
1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1/15 0.3 9.8 0.0 2.0 1.0 0.0 2.0 Alabama: Montgomery 62 86 118 145 172

2/1 0.5 20.8 0.0 3.0 3.0 0.0 4.0 California: Red Bluff 13 21 36 49 65

2/15 0.9 30.2 0.0 5.0 6.0 1.0 6.0 Georgia: Columbus 61 81 108 131 152

3/1 2.0 37.6 0.0 7.0 9.0 2.0 8.0 Indiana: Indianapolis 29 41 60 75 90

3/15 4.3 45.8 0.0 10.0 12.0 3.0 12.0 Kentucky: Lexington 28 46 80 114 151

4/1 9.2 50.6 1.0 13.0 16.0 4.0 16.0 Kentucky: Middlesboro 28 38 52 63 73

4/15 13.2 54.4 2.0 16.0 21.0 5.0 20.0 Maine: Portland 16 27 48 66 88

5/1 18.0 56.0 3.0 19.0 26.0 7.0 25.0 New York: Buffalo 15 23 36 49 61

5/15 22.7 56.8 6.0 23.0 31.0 12.0 30.0 North Carolina: Raleigh 53 77 110 137 168

6/1 29.2 57.1 11.0 27.0 37.0 17.0 35.0 Oklahoma: Ardmore 46 71 107 141 179

6/15 39.5 57.1 23.0 34.0 43.0 24.0 41.0 Oregon: Portland 6 9 13 15 18

7/1 46.3 57.2 36.0 44.0 50.0 33.0 47.0 South Dakota: Rapid City 12 20 34 48 64

7/15 48.8 57.6 49.0 54.0 57.0 42.0 56.0 Tennessee: Memphis 43 55 70 82 91

8/1 51.1 58.5 63.0 63.0 64.0 55.0 67.0 Wyoming: Cheyenne 9 14 21 27 34

8/15 57.2 59.8 77.0 72.0 71.0 67.0 75.0


Adapted from Wischmeier and Smith (1978). R factor values are in
a

9/1 64.4 62.2 90.0 80.0 77.0 76.0 81.0


English units: hundreds of ft>tonsf»in/acre«hr«year. To convert to SI units,
9/15 67.7 65.3 95.0 85.0 81.0 83.0 85.0 MJ»mm/ha«h»year, multiply by 17.02. Additional values are given in
10/1 71.1 67.5 98.0 89.0 85.0 89.0 87.0 Appendix 8A. Isolines of 10-yr single-storm factors are given in Figs.
8A.5-8A.8.
10/15 77.2 68.2 99.0 91.0 88.0 92.0 89.0
11/1 85.1 69.4 100.0 93.0 91.0 94.0 91.0
11/15 92.5 74.8 100.0 95.0 93.0 96.0 93.0 w h e r e R is the synthetic storm R factor, Ρ is precipi­
st

12/1 96.5 86.6 100.0 96.0 95.0 98.0 95.0 tation in inches corresponding t o a duration D in
12/15 99.0 93.0 100.0 98.0 97.0 99.0 97.0
hours, a and b are constants, and
x x

Values are cumulative percentage of total annual R for the day indicat­
a
/(£>) = 2.119D 0 0 0 8 6
. (8.33)
ed. Tabulations generously provided by K. Renard of the U S D A - A R S
(Tucson, AZ). Additional values are given in Appendix 8A. Values for a and b x x are summarized below.

Type
this computation, R is estimated for an individual storm *1
storm. Wischmeier and Smith (1965, 1978) evaluated R
I 15.03 0.5780
factors for individual historic storms of all durations
IA 12.98 0.7488
and summarized the information by return period.
These values were not upgraded in the R U S L E , so II 17.90 0.4134

values from the original analysis are given in Table 8.2 IIA 21.51 0.2811
for selected locations. A more extensive listing is given
in Appendix 8A.
Values in Table 8.2 are for historic storms. It is also Ateshian (1974) developed similar power relationships
useful to have values for D D F or SCS synthetic storms relating R to Ρ a n d D ; however, f(D) was repre­
24

used for design of structures as discussed in C h a p t e r 3. sented by a constant. Cooley (1980) indicates that the
Synthetic storm R values have b e e n d e t e r m i n e d for Ateshian equations can over- or underestimate by as
SCS type I, type IA, type II, and type H A storms by much as 4 0 % for high-intensity short-duration storms.
Cooley (1980) or Cooley (1980) summarized data sets comparing en­
ergy content of historic storms to predictions from
Eq. (8.32). A s expected, t h e synthetic storm value can
a P* D)
{
b e m u c h greater or less t h a n historic storms. It is
important to r e m e m b e r that SCS storms are design
254 8. Erosion and Sediment Yield
storms and are not intended to replicate particular 5. Correcting R factors for ponding. The R factor for SW
historic events. Georgia needs to be corrected for ponding on flat slopes.
From Fig. 8.7, the correction factor for a slope of 0.1%
and an R factor of 450 is 0.70. Therefore, the corrected R
value is
Example Problem 8.2 Determining R factors
ROOT = (450)(0.70) = 315.
Determine the following R factors for extreme Southwest
Georgia and extreme Northwestern New York: Annual R ,
January, April, and July R , 10-year historic storm R , and
10-year, 24-hr synthetic storm R . What would the R factor Special R Factors for the Pacific Northwest
be for a specific location in Southwest Georgia if the slope at
In the Pacific Northwest (PNW), rain or snow falling
that location were 0.1%?
on cropland produces erosion in excess of that ex­
Solution: pected from the R factor based on rainfall energy and
1. Annual R factor. From Fig. 8.6, the average annual R in maximum 30-min intensity, EI . Based on experimen­
30

English units is tal data taken in the P N W , R e n a r d et al, (1993b)


SW Georgia, R = 450 propose that an equivalent R can b e calculated from
NW New York, R = 80. the annual precipitation, or

2. January, April, and July R . From Fig. 8.8, SW Georgia is REQ = 5.9P, (8.34)
in zone 119 and NW New York is in zone 112. Values in
Table 8.1 are cumulative percentage of total annual R . The where R EQ is the equivalent R factor for the cropland
incremental percentage change of annual R can be deter­ in ft · tonsf · i n . / a c r e · hr · year and Ρ is annual rain­
mined from Table 8.1 for each month and multiplied by the fall in inches.
annual R to get the monthly value. For April in SW Georgia
(zone 118), the incremental change in percentage R is from Special Considerations for Higher Elevation
16.0 to 25.0 or 9%. With an annual R of 450, the incremental and Winter Conditions
R during April is 0.09 X 450 or 40.5. Tabulations for all time
periods are shown below. Wischmeier and Smith (1978) proposed in H a n d ­
book 537 t h a t precipitation falling in the form of snow
could be multiplied by 1.5 and a d d e d to the R factor to
January 1/1-2/1 April 4/1-5/1 July 7/1-8/1
fraction//? value fraction//? value fraction//? value account for the impacts of snowfall on erosion. T h e
developers of the R U S L E found this unsatisfactory
SW Georgia 0.04/18.0 0.09/40.5 0.20/90.0 and d o not r e c o m m e n d it. O t h e r t h a n the above-men­
NW New York 0.00/0.00 0.03/2.4 0.22/17.6 tioned analysis for the Pacific Northwest, they recom­
m e n d e d that n o correction be m a d e for snowfall.
A t higher elevations in the West w h e r e heavy snow­
3. Historic return period values. From Table 8.2, the 10-year fall is observed, it is possible that t h e EI index might
3Q

return period single-storm R for the nearest cities are (En­ be too high as a result of high snowfalls that accumu­
glish units)
late to large depths and d o not generate heavy runoff
Columbus, GA, R = 131 L0
rates. Predictions of erosion at these high altitudes,
Buffalo, NY, R = 49.
L0 based o n t h e EI index, would b e correspondingly too
30

These values compare favorably to values in Fig. 8A.5. high. W h e n evaluating erosion u n d e r these conditions,
4. Synthetic storms. Both locations utilize type II storms; the R factor based on total precipitation should be
hence a and b for Eq. (8.32) are 17.90 and 0.4134. From
x x
reduced to eliminate the snow component.
Appendix 3A in Chapter 3, the 10-year 24-hr storm precip­
itation for SW Georgia is 7.5 in. and NW New York is 3.5
in. Thus, for a duration of 24 hr, the R values can be Erodibility Factor Κ
calculated as
Definition of Erodibility

f(D) = 2.119(24)° 0 0 8 6
= 2.18 Ideally, soil erodibility is a measure of a soil's resis­
tance to the erosive powers of rainfall energy and
17.90(7.5) 2 1 8
runoff. Practically, in the R U S L E , soil erodibility is an
SW Georgia: R = , n. - 389
integration of the impacts of rainfall and runoff on soil
ST 4 1 3 4

(24)
loss for a given soil. Experimentally, soil erodibility is
17.90(3.5)* 1 8
the soil loss per unit rainfall index on a standard
NW New York: R = ' = 73.8.
ST 1 3 4
erosion plot, i.e., a plot u n d e r fallow conditions on a
Rill a n d Interrill Erosion Modeling: USLE / RUSLE Empirical Models 255
slope of 9 % with a slope length of 72.6 ft with u p - and Wischmeier et al. (1971), which was developed from
downslope tillage. U n d e r these conditions, L , 5 , C, d a t a collected on 55 midwestern agricultural soils. Soil
and Ρ are all equal to 1.0, hence erodibility in t h e n o m o g r a p h is predicted as a function
of five soil a n d soil profile p a r a m e t e r s :
measured erosion
K = (8.35) • P e r c e n t a g e silt ( M S ; 0.002-0.05 mm).
XEL 30 • P e r c e n t a g e very fine sand (VFS; 0.05-0.1 mm).
• P e r c e n t a g e sand (SA; 0 . 1 - 2 mm).
Practically, o n e seldom encounters standard conditions • P e r c e n t a g e organic m a t t e r ( O M ) .
for a test; hence, d a t a are taken u n d e r n o n s t a n d a r d • Structure (S ). x

conditions, and corrections are m a d e based on ac­ • Permeability (P ). x

cepted relationships for L , 5 , C, and P. Inaccuracies in


any of these o t h e r p a r a m e t e r s would be reflected in the It is important to n o t e that t h e size ranges given here
estimated Κ values. are not standard for some particle classifications. Codes
for structure a n d permeability are given in U S D A soil
Estimating Κ Factors for Average Annual Erosion survey manuals (Soil Conservation Service, 1983) avail­
A n u m b e r of studies of soil erodibility have b e e n able for most counties in the U.S. and in some foreign
m a d e with the U S L E / R U S L E format as summarized countries. T h e n o m o g r a p h is shown in Fig. 8.9.
by R o m k e n s et al. (1993). In the U S L E , Κ is assumed A n analytical relationship for the nomograph in
to be constant throughout the year. Tables of Κ values Fig. 8.9 is (Wischmeier et al, 1971)
are available from local Soil Conservation Service of­ 2.1 X 1 0 " ( 1 2 -
4
OM)A# U 4
+ 3 . 2 5 ( 5 ! - 2) + 2.5{ΡΧ - 3)
fices for most soils in the U.S. Κ values are also Κ =
100
tabulated in the m o r e recent soil survey manuals. In (8.36)
the absence of published data, a widely used relation­
ship for predicting erodibility is a n o m o g r a p h by w h e r e Κ is soil erodibility in tons p e r acre p e r unit

-—.70

PROCEDURE: With appropriate data,


enter teal* at left with % SHt + v f t (.002 - 0.1mm) and
proceed to points representing the sod's % sand (0.10-2.0mm),
% organic matter, structure, and permeability. In that sequence.
Interpolate between plotted curves. The dotted line illustrates
procedure for a sod having: sN-vfs 65%» sand 5% OM 2.8%,

T$Zk$r?ha"MJ mm
e
hmy
4 - S 0 M , 0 n :
^hmtT^L*
Figure 8.9 Soil erodibility nomograph of Wischmeier et al. (1971). The axes for Κ are scaled in both English and SI units.
256 8. Erosion and Sediment Yield

rainfall index (tons · acre · h r / h u n d r e d s · acre · ft · or tons · acre · h r / h u n d r e d s · acre · ft · tonsf · in.,
tonsf · in.), O M is the percentage organic matter, P is x w h e r e tonsf m e a n s tons of force from the energy equa­
the permeability index, S is the structure index, and
x tion and tons are mass of sediment being eroded. T h e
Μ is a function of the primary particle size fractions accuracy of Eq. (8.36) was evaluated by R o m k e n s et al.
given by (1993) using data from a n u m b e r of sources. In general,
the n o m o g r a p h worked well for midwest soils, but did
Μ = (%MS + %VFS)(100 - % C L ) , (8.37) not work well on soils from Hawaii or for subsoils.

where % C L is percentage clay ( < 0.002 m m ) and other


terms are as defined above. Equation (8.36) is valid for Seasonal Variation in Κ Values
% M S + % V F S less than 70. A n u m b e r of researchers have shown that soil erodi­
Since the units on Κ (English) are tons per acre per bility varies with antecedent moisture and with freezing
unit rainfall index and since the R factor is the EI 30 and thawing (Mutchler and Carter, 1983). W h e n aver­
index as given by Eq. (8.31), the units on Κ are A/R aged over a n u m b e r of years, freezing and thawing and

CASE I tmax<tmin

Κmax

•P*: CO.

I
Κnom I I
I I

- *max
I I
Froat Free Periods A t
r
.I..L t|>t„
Τ ' ' 1 1
' 3 6 5 Julian Day
120 150 180 210 240 270 300 330 360

CASE II
Figure 8.10 R U S L E procedure for predicting seasonal changes in Κ factor (after Romkens et al., 1993). It is not
recommended that these procedures be used for locations west of the Rockies.
Rill and Interrill Erosion Modeling: USLE / RUSLE Empirical Models 257

0 50 100 150 200 250 300 350 400 450 500

Annual R (ft · tonsf - i n / a c - h r - j r )

300

270

240 t m » = 154-.44R

s 2 210
ed £ 180

150

120

90

60 i

il 30

-30
0
•<
-60
50 100 150 200 250 300 350 400 450 500

Annual R ( f t - t o n s f · i n / a c · h r · y r )
Figure 8.11 R U S L E relationship between K /K , K /K , and the time of maximum
m&x min mikX nom

erodibility f
maxwith the annual R factor (after Romkens et al., 1993). Julian day is time measured
from January 1. It is not recommended that these procedures be used for locations west of the
Rockies.

high antecedent moisture conditions tend t o occur o n a (Julian days) of m i n i m u m erodibility. Julian days a r e
predictable basis. Procedures w e r e developed in t h e days n u m b e r e d sequentially from 1 t o 365 starting with
R U S L E t o account for this variability, as shown in J a n u a r y 1.
Fig. 8.10. In t h e R U S L E , correlations were m a d e b e ­ T h e p r o c e d u r e s illustrated in Figs. 8.10 a n d 8.11 a r e
tween t h e annual R factor a n d K /K max minwell as a s
divided into two categories. Case I is for i m a x less than
the time of maximum erodibility i , as shown in Fig.
m a x i , a n d Case II is for t
m i n max greater t h a n t . Proce­
min

8.11. In these relationships, i is t h e time (Julian


m a x d u r e s for using t h e relationships a r e illustrated in
days) of maximum erodibility a n d i is t h e time
m i n Example Problem 8.3.
258 8. Erosion dnd Sediment Yield
For the time periods when t > t and in the following year
x mxtx

when t < r , two sets of values are needed. For the


x m a x

Example Problem 8.3 Estimating Κ values periods when the temperature is below 27° F, November 16
to March 16, Eq. (d) applies, or
Estimate the annual Κ value using the Wischmeier nomo­ *i =K = 0.078; fj < 75; t > 320. (d)
mxn x
graph for a soil in central Minnesota that has the following
characteristics For the periods when 7 > 27° F, March 16 to April 19
a v g

(Julian Day 75 to 110) and September 6 to November 16


%SA = 20% (0.1 - 2.0 mm) (Julian Day 250 to 320), Eqs. (b) and (c) from Fig. 8.10 are
%VFS = 5% (0.05 - 0.1 mm) used. For the period September 6 to November 16, Eq. (b)
%MS = 45% (0.002 - 0.05 mm) yields
%CL = 30% ( < 0.002 mm)
%OM = 2 % K = 0.078exp[0.009(t -t )]
x x min

S x= soil structure = medium = 3 = 0.078exp[0.009(ii - 250)]; 250 < t < 320.


x (b)
P = permeability = rapid = 1
x
For the period March 16 to April 19 (Julian day 75 to 110),
R = annual R factor = 1 0 0 (English units)
Eq. (c) yields
Correct the annual Κ factor, accounting for the seasonal K = 0.078exp[0.009(ii - r + 365)];
x m i n
variability in Κ and the variability in the EI index. The 30

average number of frost-free days is 140. The average tem­ subject to K < K . x m dX

perature drops below 27° F on November 16 (Julian day 320) = 0.078 exp[0.009(ii - 250 + 365)]
and rises above 27° F on March 16 (Julian day 75).
= 0.078exp[0.009(ii + 115)] < 0.525; 75 < t < 110.
x

Solution: (c)
1. Nomograph value for Κ For the nomograph, %MS +
%VFS = 45 + 5 = 50. Using this value plus given values for 3. Correcting for the monthly rainfall energy distribution.
%SA, %OM, S and P in the nomograph, the value for Κ From Fig. 8.8, the zone for central Minnesota is 86. Utilizing
lt x

read directly from the nomograph is Κ = 0.21 (English units). the distribution in Table 8.1 for zone 86 and the relationships
Comparing this to the solution from Eqs. (8.36) and (8.37) above for K the Κ factor can be weighted as shown in the
x>

following table. Using the sum %R K in the last column, the X X

Μ = (45 +• 5)(100 - 30) = 3500 average Κ factor becomes


2.1 X 1 0 " ( 1 2 - 2 ) ( 3 5 0 0 )
4 1 1 4
+ 3.25(3 - 2) + 2.5(1 - 3 ) XR.K: 18.63
K = = 0.186.
100 100 100
Κ = 0.21. In this case, K is only slightly lower than the nominal K.
avg

The distribution of K, however, can allow for improved


2. Corrections for seasonal variability. For the location in
calculation of erosion by correlating appropriate values of Κ
Minnesota, the maximum erodibility would occur in the spring
with cover and rainfall during periods when cover may leave
and the minimum in the summer or fall; thus t is less than max
the soil exposed. A plot of the predicted K is given in Fig.
t , which is a Case I situation for Fig. 8.10 (t < t
x
min max min
8.12.
means that the date of maximum erodibility occurs before
the date of minimum erodibility). Using an R factor of 100 in
Fig. 8.11, the ratio K /K is 6.7, the ratio K /K
max min is max nom

2.5, and i is Julian day 110. For this situation


m a x

K„
= 2.5; K = (2.5)(0.21) = 0.525 (English Units) G 66
max
80
. 1 L _
|15 Day Average T e m p ^ ^
0.525 Η
= 6.7; K = = 0.078. 0
K„ min
6.7
^ .6
60 2u
cd

α
The Julian day for t is r plus the frost-free period. For
n m a x
\ 40 S

v
this location ο 3
redicted Erodibility Η
wu
Λ
'min - 'max + ^ - 110 + 140 = 250 (Julian day). 20 Q
Φ

F o
'max < * { < 'min> Julian day 110 to 250, K can be pre­
r

9)
x

dicted from Eq. (a) in Fig. 8.10, or 6A


CO
U
9)
A
min / 0.078 \ (/i-110)/140 60 120 180 240 300 360
Si
K> = K = 0.525
\ 0.525 j
n
t j J u l i a n Day

= 0.525(0.149) '" (/ 1,0)7140


; 110 < t < 250. (a) Figure 8.12 Predicted values of K for Example Problem 8.3. t
x
Rill and Interrill Erosion Modeling: USLE / RUSLE Empirical Models 259
Κ Value Computations for Example Problem 8.3

% annual R Julian fa v g
c
equation Kf
Time period =%R a
day* (°F) (in Fig. 8.10) (calc) %{R)K {

01/01-01/15 0 8 10 d 0.078 0
01/16-01/31 0 23 10 d 0.078 0
02/1-02/15 0 39 15 d 0.078 0
02/16-02/28 0 53 15 d 0.078 0
03/1-03/15 0 67 22 d 0.078 0
03/16-03/31 1 82 31 c 0.459 0.459
04/1-04/15 1 98 40 c 0.525 0.525
04/16-04/30 1 113 40 a 0.504 0.504

05/1-05/15 3 128 57 a 0.411 1.233


05/16-05/31 5 143 57 a 0.335 1.675

06/1-06/15 12 159 66 a 0.270 3.240


06/16-06/30 13 174 66 a 0.220 2.860

07/01-07/15 13 189 72 a 0.179 2.327

07/16-07/31 14 204 72 a 0.146 2.044

08/01-08/15 14 220 71 a 0.118 1.652

08/16-08/31 13 235 71 a 0.096 1.248

09/01-09/15 5 251 60 b 0.078 0.390

09/16-09/30 3 266 60 0.090 0.270

10/01-10/15 1 281 50 b 0.103 0.103

10/16-10/31 1 296 50 b 0.118 0.118

11/01-11/15 0 312 34 b 0.136 0

11/16-11/30 0 327 26 d 0.078 0

12/01-12/15 0 342 17 d 0.078 0

12/16-12/31 0 357 17 d 0.078 0

Total 100 18.63

T a b l e 8.1, zone 8b.


^Midpoint of half month intervals in Julian days. For example, the midpoint Julian day for the interval
2/1-2/15 is 39.
Assumed average temperatures for central Minnesota.
c

V a l u e s are given in English units (tons acre»hr/hundreds»acre ft tonsf»in.).


e e e

T h e results in Fig. 8.12 indicate discontinuities w h e n and decreasing infiltration. Rocks on the surface ab­
the air t e m p e r a t u r e drops below 27° F and rises above sorb energy a n d decrease erosion, which must be ac­
27° F. This discontinuity is reasonable and occurs as a c o u n t e d for in the C factor using the fraction of
result of freezing and thawing. W h e n frozen, soil is not surface cover (see subsequent discussion of C factor).
very erodible. W h e n the t e m p e r a t u r e rises above freez­ Rocks b e n e a t h t h e surface decrease infiltration rates
ing, the soil will immediately r e t u r n to a m o r e erodible and increase erosion, which must be accounted for in
state. t h e Κ factor. T h e change in saturated hydraulic con­
T h e procedures illustrated above are for central and ductivity d u e to rocks can be estimated by
eastern United States and should not b e used for sites
in and west of the Rockies. T h e p r o c e d u r e s are devel­ kh = * ( l - J l ) = 2k
f w t
K
^ (8.38)
oped for annual erosion, not for single storms.

where k bis the saturated hydraulic conductivity in


Soils with Rock F r a g m e n t s rocky soils, k is t h e saturated hydraulic conductivity of
f

In the U.S., 15.6% of the soils have significant rock the soil fraction, R is t h e fraction by weight of rock
W

fragments in the soil and on t h e surface. T h e s e rock fragments g r e a t e r t h a n 2 mm, a n d R is t h e fraction by


Y

fragments impact erosion by absorbing impact energy volume of rock fragments greater than 2 mm.
260 8. Erosion and Sediment Yield
T h e impact of changes in saturated hydraulic con­ on the C factor must be based on percentage ground cover,
ductivity on the Κ factor must be accounted for by the as discussed in a subsequent section.
nomograph in Fig. 8.9. T o accomplish this correction
using Eq. (8.38), relationships between hydraulic con­
ductivity and permeability classes used in Fig. 8.9 must Rough Estimates of Κ from Texturai Information
be known. Rawls et al. (1982) proposed the relation­ and Experimental Values for Construction
ship shown in Table 8.3. and Mined Sites
T h e U S D A - S C S has developed estimates of Κ
based on textural classification for topsoil, subsoil, and
Example Problem 8.4. Effects of rock fragments residual materials as shown in Table 8.4. These values
on A are first estimates only and d o not include the influ­
ence of soil structure or infiltration characteristics.
A silty clay loam soil is classified as permeability class 5. A limited n u m b e r of data sets have been developed
Based on textural information, soil structure, and a perme­ for drastically disturbed lands and for reconstructed
ability class of 5, Κ is estimated as 0.21 in English units. soils. A summary of the data is given in Table 8.5 along
What would be the value for Κ as corrected for rock frag­
with a comparison to values from the Wischmeier et al.
ments if the percentage of rock fragments greater than 2 mm
(1971) n o m o g r a p h shown in Fig. 8.9. T h e comparison is
occupies 40% of the soil mass by weight?
Solution: sufficiently favorable to warrant the use of the nomo­
1. Impact of rock fragment on hydraulic conductivity. From graph for a first estimate of Κ o n disturbed topsoil or
Table 8.3, fc for a silty clay loam soil is between 0.04 and
f
Α-horizon material. T h e comparison is not favorable
0.08 in./hr. Assume a value of 0.06 in./hr. From Eq. (8.38) for subsoil materials.

k = k (l
b f - R„) - 0.06(1 - 0.40) = 0.036 i n . / h r . Length and Slope Factors L and S
2. Estimating the revised permeability class. From Table T h e effects of topography on soil erosion are deter­
8.3, the permeability class for k = 0.036 in./hr is 6.
b
mined by dimensionless L and S factors, which ac­
3. Estimating the new erodibility. Entering Fig. 8.9 with an count for both rill and interrill erosion impacts.
estimated Κ of 0.21 for a permeability class of 5, the Κ value
for a class 6 permeability is estimated as 0.22 (English units). Slope Steepness Factor 5
It is again important to note that this procedure corrects T h e slope steepness factor S is used to predict the
only for the effects of rock fragments on infiltration. Impacts effect of slope gradient on soil loss. For slope lengths

Table 8.3 Soil Water Data for the Major USDA Soil Textural Classes
(after Rawls etal., 1982)

Saturated hydraulic
conductivity Hydrologic
Permeability soil
Texture class a
in./hr mm/hr group 6

Silty clay, clay 6 <0.04 <1 D


Silty clay loam, 5 0.04-0.08 1-2 C-D
sandy clay
Sandy clay 4 0.08-0.20 2-5 C
loam, clay loam
Loam, silt loam 3 0.20-0.80 5-20 Β
Loamy sand, 2 0.8O-2.40 20-60 A
sandy loam
Sand 1 >2.40 >60 A+

S e e Soil Conservation Service National Soils Handbook (SCS, 1983).


a

See Soil Conservation Service National Engineering Handbook (SCS, 1972,


A

1984).
I^ote: Although the silt texture is missing from the NEH because of inadequate
data, it undoubtedly should be in permeability class 3.
Rill and Interrill Erosion Modeling: USLE / RUSLE Empirical Models 261
greater than 15 ft, the S factor from the U S L E was d a t a from disturbed lands with slopes u p to 8 4 % ,
modified significantly by McCool et al. (1987, 1993) Mclssac et al. (1987) developed an equation similar to
after extensive evaluation of the original U S L E d a t a (8.39) and (8.40) with exponents in t h e same range;
base. T h e modified version is thus McCool et al. (1993) r e c o m m e n d that Eqs. (8.39)
a n d (8.40) also b e used for disturbed lands.
5 = 10.8 sin θ + 0.03; sin 0 < 0.09 (8.39) For slope lengths less than 15 ft, the S factor is not
S = 16.8sin θ - 0.50; sin Θ > 0.09, (8.40) as strongly related to slope (slope exponent less than
1.0) since rilling would not have b e e n initiated. T h e
where Θ is the slope angle. Based on an evaluation of r e c o m m e n d e d factor is

5 = 3.O(sin0) 0 8
+ 0.56 (8.41)
Table 8.4 Κ Value Estimates based on Textural Information
(English Units) (Soil Conservation Service, 1978) U n d e r conditions w h e r e thawing of recently tilled
soils is occurring and surface runoff is the primary
Texture Estimated Κ value a

factor causing erosion (typical of the Pacific Northwest


Topsoil
in the spring), t h e S factor should b e (McCool et al.,
1987, 1993)
Clay, clay loam, loam, silty clay 0.32*
Fine sandy loam, loamy very fine sand, sandy loam 0.24 0.6
5 = 4.25(sin 0 ) sin 0 > 0.09. (8.42)
Loamy fine sand, loamy sand 0.17
Sand 0.15 F o r thawing soils with slopes less than 9%, Eq. (8.39)
Silt loam, silty clay loam, very fine sandy loam 0.37 should be used.
T h e S factor in the R U S L E is significantly modified
from t h e original U S L E as a result of an extensive
Subsoil and Residual Material
reevaluation of t h e original d a t a base, addition of the
Outwash Soils
factors for short slope lengths, and new values for
Sand 0.17
thawing soils (McCool et al., 1987). T h e original data
Loamy sand 0.24
base did not include values beyond 2 0 % . W h e n using
Sandy loam 0.43 t h e quadratic form of the equation for S developed for
Gravel, fine to moderate fine 0.24 t h e original U S L E , projections beyond 2 0 % yielded
Gravel, medium to moderate coarse 0.49 unreasonably high values for erosion. T h e R U S L E
Lacrustrine Soils equation with the linear function corrects this problem.
Silt loam and very fine sandy loam 0.37
Slope Length Factor
Silty clay loam 0.28
Clay and silty clay 0.28 T h e slope length factor was developed by McCool
et al. (1989, 1993) from t h e original U S L E data base
Glacial Till
a u g m e n t e d with theoretical considerations. T h e L fac­
Loam, fine to moderate fine subsoil 0.32
tor retains its original form
Loam, medium subsoil 0.37
Clay loam 0.32
Clay and silty clay 0.28 L = (8.43)
72.6
Loess 0.37
Residual w h e r e λ is t h e slope length in feet, 72.6 ft is the length
Sandstone 0.49 of a standard erosion plot, and m is a variable slope
Siltstone, nonchannery 0.43 length exponent. Slope length, A, is the horizontal
Siltstone, channery 0.32 projection of plot length, not the length measured
0.28
along t h e slope. T h e difference in horizontal projec­
Acid clay shale
tions and slope lengths becomes important on steeper
Calcareous clay shale or limestone residuum 0.24
slopes.
T h e s e values are typical based only on textural information. Values for T h e slope length exponent is related to the ratio of
an actual soil can be considerably different due to different structure and rill to interrill erosion, β (Foster et al., 1977b; McCool
infiltration. et al., 1989, 1993), by
*Units on Κ in this table are English units (tons*acre»hr/hundreds»
acre»ft»tonsf»in.). To convert to metric units (t«ha*h/ha MJ»mm), multiply
e β
m = . (8.44)
A'values by 0.1317. 1 +/3 v
'
262 8. Erosion and Sediment Yield

Table 8.5 Experimental Κ Value Estimates for Disturbed Lands (English Units)

Reclaimed soil or Location of Κ


residual material experimental site Exp /Nomo
a 6
Reference

Hosmer silt loam Indiana 0.387/0.485 c


Stein etal. (1983)
Alfred silt loam Southern Indiana 0.812/0.485
Ava silt loam Southern Indiana 0.842/0.478
Graded overburden Southern Indiana 0.197-0.835/
0.250-0.478
Clinton silt \ozm d
Western Illinois 0.370/0.360 Mitchell et al. (1983)
Tama silty clay loam** Western Illinois 0.210/0.310
Hosmer silt loam*' Southern Indiana 0.450-0.650/
0.470
Sadler silt loam (A horizon) Western Kentucky 0.415/0.385 Barfield etal. (1988)
Sadler silt loam (B horizon) Western Kentucky 0.380/0.640
Shale spoil material Western Kentucky 0.140/0.180

"Values measured experimentally with rainfall simulators.


^Values calculated from Wischmeier etal. (1971) nomograph shown in Fig. 8.9.
Values in English units of tons acre hr/hundreds»acre»ft»tonsf»in. To convert to metric units of
c e e

t«a»h/ha»MJ»mm, multiply by 0.1317.


*The dominant soil series. Some mixing occurred with other series.

For soils that are classed as being moderately sus­ interrill erosion ratio is proposed. Selection of the
ceptible to erosion, McCool et al. (1989) proposed that appropriate column to use in Table 8.6 requires profes­
sional judgement. T h e assistance of a soil scientist may
11.16 sin θ
(8.45) b e helpful.
β mod ~~ , . Λ Χ 0.8 . f\ c c

3.0(sm 0 ) + 0.56 Combined Length and Slope Factors


w h e r e θ is the slope angle. Thus, the slope exponent is Combined slope length and slope steepness factors
a function of the slope angle Θ. were calculated using the factors from Eqs. (8.39) to
Soils in the R U S L E are classed as having low, mod­ (8.43). T h e s e combination factors are given in Fig. 8.13
erate, or high susceptibility to rill erosion. Equation for all susceptibilities and for thawing soils.
(8.45) is for soils that are moderately susceptible to
erosion. Conversions for soils that have low or high Irregular and Segmented Slopes
susceptibility to erosion are given in Table 8.6. Values Soil loss is strongly impacted by slope shape (Foster
in Table 8.6 are based on the assumption that moder­ and Huggins, 1979). A convex shape will have greater
ately erodible soils have a β defined by Eq. (8.45), soils erosion than a uniform slope by as much as 3 0 % . A
highly susceptible to rilling have a β that is twice that concave slope will have less erosion t h a n a uniform
given by Eq. (8.45), and soils with low susceptibility to slope. Foster and Wischmeier (1974) developed a pro­
rilling have a β that is defined by half that given by cedure for evaluating the impact of irregular slopes by
Eq. (8.45). dividing the slope into segments. T h e soil loss per unit
For soils in the Pacific Northwest, or other soils that area from the ith segment is
are exposed to runoff during thawing without sufficient
rainfall energy to cause interrill erosion, the values in \m+ 1 - A'
m +1
1
Table 8.6 should not be used. Instead, McCool et al. A, = RK C P S
i i i i (8.46)
(1989) recommend that a slope length exponent of 0.5 (A, 1 )72.6"
be used for all slopes. W h e n runoff on thawing soils is
exposed to rainfall sufficient to cause significant inter­ where λ, and λ , , ! are the slope lengths at the start
rill erosion, the slope length exponent for the low rill and end of segment i, and K C P and S, are U S L E i9 i9 h

to interrill erosion ratio should be used (column 1 in factors for segment i. Equation (8.46) can be used for
Table 8.6). For rangeland soils, the use of a low rill to each segment i. T h e total erosion from each segment
Rill and Interrill Erosion Modeling: USLE / RUSLE Empirical Models 263
Table 8.6 Slope Length Exponent m in Eq. (8.43)
(after McCool et al., 1993)* Dividing by η times t h e soil loss from a uniform slope
of equal length a n d assuming constant values of K C t i

Rill/interrill ratio P along t h e slope, a slope adjustment factor can b e


t
Percentage
slope Low* Moderate c
High* developed for each segment, or

0.2 0.02 0.04 0.07 +1 m+ 1


A s ιm (i - 1 )
0.5 0.04 0.08 0.16 (8.49)
=
~A =
~
1.0 0.08 0.15 0.26
2.0 0.14 0.24 0.39
w h e r e η is t h e n u m b e r of segments a n d S A F is t h e
3.0 0.18 0.31 0.47
slope adjustment factor. T h e s u m of t h e SAF, for a
4.0 0.22 0.36 0.53
given slope is equal t o t h e n u m b e r of segments n; thus
5.0 0.25 0.40 0.57 t h e average erosion over t h e slope is
6.0 0.28 0.43 0.60
8.0 0.32 0.48 0.65 R
10.0 0.35 0.52 0.68 >l = r E ^ W X ( S A F ) ,
l l i
(8.50a)
i=l
12.0 0.37 0.55 0.71
14.0 0.40 0.57 0.72
w h e r e L , is t h e slope length factor calculated from
16.0 0.41 0.59 0.74
Eq. (8.43) using t h e m value corresponding to t h e
20.0 0.44 0.61 0.76 segment steepness. I n t h e development of a S A F rela­
25.0 0.47 0.64 0.78 tionship, R, K, C, a n d Ρ remain constant over all
30.0 0.49 0.66 0.79 segments; t h u s E q . (8.50a) c a n b e solved for an equiva­
40.0 0.52 0.68 0.81 lent LS factor
50.0 0.54 0.70 0.82
60.0 0.55 0.71 0.83
LS = - Σ S/L^SAF),. (8.50b)
i = \
a
Values in table are not applicable to thawing soils. See
text for explanation.
*β = 1/2 value from Eq. (8.45) in Eq. (8.44). Factors calculated from E q . (8.50b) a r e given in Table
= 1 χ value from Eq. (8.45) in Eq. (8.44). 8.7. A n example of its use is given in Example Prob­
<*β = 2 χ value from Eq. (8.45) in Eq. (8.44). lem 8.5.

would b e A (k - A , ^ ) , a n d t h e average erosion p e r


Example Problem 8.5. Estimating LS factors
i i;

unit area over t h e entire slope length would b e A soil that is very susceptible to rilling has a slope length
of 210 ft and an average slope of 15%. Estimate the LS
factor if:
[λ7 - A7_V +1

A = R Σ K&PiSi (8.47) (1) the slope is uniform


A,72.6 m

i= l (2) the slope is convex with slopes of 10, 15, and 20% on
segments 1, 2, and 3
where A^ is t h e total slope length. Equation (8.47) can (3) the slope is concave with slopes of 20, 15, and 10% on
segments 1, 2, and 3.
also b e used to evaluate t h e effects of variation in Κ,
C, a n d Ρ over t h e slope length. Assume that the soil is not freezing and thawing.
A n alternate method for evaluating irregular slopes Solution:
is t h e use of a slope length adjustment factor (SAF). If 1. Uniform slope. The slope angle is
the slope is divided into η increments of equal length
ΔΧ, then θ = t a n " 0.15 = 8.53°. 1

From Eq. (8.45) for soils moderately susceptible to rilling,


\(ikX) m + l
- ([/ - l ] A * ) m + 1
]
A=RΣ K&PiSi
η AX72.6 m
11.16sin8.53
i= l
β τι = 1-37.
(8.48) 3.0(sin8.53) + 0.56 08
264 8. Erosion and Sediment Yield

I l l l l l l l l ι—I I I I III l l l l l l l l ι—I I Ι I III 1 — Γ Τ Τ Ί


1 5 10 50 100 5 10 50 100 50C
Slope Length (ft) Slope Length (ft)

ι f i 111 in
50 100 500 ) 5 1'0 5'0 100
Slope Length (ft) Slope Length (ft)
Figure 8.13 Combined slope length and steepness factors, L S , for the R U S L E for varying soil susceptibilities to rilling
and for thawing soils.

For a soil highly susceptible to rilling, β is doubled before and (8.43)


estimating m\ thus β = 2.74. From Eq. (8.44) LS = (16.8 sin θ - 0.50)

m = = 0 73 / 210 \ ° 7 3

1 + 2.74 ' - I— I (16.8 sin 8.53 - 0.50)


which is equal to the value in Table 8.6. From Eqs. (8.40) = 4.33.
Rill and Interrill Erosion Modeling: USLE / RUSLE Empirical Models 265

Table 8.7 Adjustment Factors to LS to Estimate Soil Loss on Irregular Slopes (after McCool
etal, 1993)°

Number Sequential Slope length exponent


of number of
segments segments 0.05 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

,i= 2 i= 1 0.97 0.93 0.87 0.81 0.76 0.71 0.66 0.62 0.57
2 1.03 1.07 1.13 1.19 1.24 1.29 1.34 1.38 1.43

3 1 0.95 0.90 0.80 0.72 0.64 0.58 0.52 0.46 0.42


2 1.01 1.02 1.04 1.05 1.06 1.05 1.05 1.04 1.03
3 1.04 1.08 1.16 1.23 1.30 1.37 1.43 1.50 1.55

4 1 0.93 0.87 0.76 0.66 0.57 0.50 0.44 0.38 0.33


2 1.00 1.00 0.98 0.96 0.94 0.92 0.88 0.85 0.82
3 1.03 1.05 1.09 1.13 1.16 1.18 1.20 1.22 1.23
4 1.04 1.08 1.17 1.25 1.33 1.40 1.48 1.55 1.62

5 1 0.92 0.85 0.73 0.62 0.53 0.45 0.38 0.32 0.28


2 0.99 0.97 0.94 0.90 0.86 0.82 0.77 0.73 0.69
3 1.01 1.03 1.04 1.05 1.06 1.06 1.06 1.05 1.03
4 1.03 1.06 1.12 1.17 1.21 1.25 1.29 1.32 1.35

5 1.05 1.09 1.17 1.26 1.34 1.42 1.50 1.58 1.65

S o i l loss factors = [ i
a
- (M) ] / n , where i is the sequential number of segment, m is the slope length exponent,
I + m 1 + m m

and η is the number of segments. Values are forced to give a factor total equal to n.

2. Convex slope. Values are tabulated below.

Slope Uniform Segment LS


Slope Slope expon. θ slope LS SAF, factor (Column 5 χ
segment percentage Table 8.6 (°) factor 0
Table 8.7 Column 6)

1 10 0.68 5.71 2.41 0.47 1.13


2 15 0.73 8.53 4.33 1.04 4.50

3 20 0.76 11.31 6.26 1.53 9.58


Σ=15.21

"Equation (8.40) χ Eq. (8.43) with λ = 2 1 0 ft.

From Eq. (8.50b)


LS = X/n = 15.21/3 = 5.07.
3. Concave slope. Values are tabulated below.

Slope Uniform Segment LS


Slope Slope expon. θ slope LS SAF, factor (Column 5 x
segment percentage Table 8.6
η factor* Table 8.7 Column 6)

1 20 0.76 11.31 6.26 0.44 2.75


2 15 0.73 8.53 4.33 1.04 4.50
3 10 0.68 5.71 2.41 1.49 3.59
1=10.84

fl
Equation (8.40) χ Eq. (8.43) with λ = 2 1 0 ft.
266 8. Erosion and Sediment Yield

From Eq. (8.50b)

Estimating Slope Lengths for Watersheds


LS = Σ/η = 10.84/3 = 3.61.
Slope length estimates for fields or watersheds with
4. Summary. n o n p l a n a r surfaces require considerable professional
judgment. A large n u m b e r of slope lengths occur in
Concave LS = 3.61 any given real watershed. Erosion can be estimated for
Uniform LS = 4.33 each of these and area weighted to determine average
Convex LS = 5.07. erosion on a watershed.
Slope length is defined as the slope distance from
T h u s the convex shape has the highest LS factor. the point of origin of overland flow to the point of

Table 8.8 Selected USLE C Values for Construction, Mining,


and Forest Lands.

Condition C factor References Condition C factor References

1. Bare soil conditions Undisturbed forest

Undisturbed except scraped 0.66-1.30 100-75% canopy, 100-90% litter 0.0001-0.001


Compacted 35-20% canopy, 7 0 - 4 0 % litter 0.003-0.009
1.00-1.40 a,b Permanent pasture and brush cover
Smooth
Root raked 0.90-1.20 0% canopy, 80% ground cover

Disk tillage Grass 0.013


Fresh 1.00 Weeds 0.043
After one rain 0.89 50% Brush, 80% ground cover
2. Mulch Grass 0.012
Straw Mechanically prepared woodland sites
0.5 tons/ac 0.30 Burned, 10% cover at ground
0.18 a,d
l.Otons/ac Good soil 0.240
0.09 a,d
2.0 tons/ac Poor soil 0.360
0.02 a,d
4.0 tons/ac Burned, 0% cover at ground
Wood chips Good soil 0.260
0.90 a,d
0.5 tons/ac Poor soil 0.450
0.70 a.d
2.0 tons/ac Disked, 0% cover at ground
0.42 a,d
4.0 tons/ac Good soil 0.720
0.22 a,d
6.0 tons/ac Poor soil 0.940
3. Chemical binders
Asphalt emulsion, 605 gal/ac 0.14-0.52
Aquatan, Terra-tack 0.67
4. Seedings
N o prepared seedbed
New planting 0.64
After 60 days 0.54
Prepared seedbed
New planting 0.40
After 6 0 days 0.05

Note. Additional values are given in Appendix 8B.


Transportation Research Board (1980).
^Barfield a/. (1988).
Wischmeier and Smith (1978).
''Meyer et al. (1972). C factors for mulch vary depending on slope length and steepness. Slope length limits apply (see Appendix 8B, Table 8B.5).
Rill and Interrill Erosion Modeling: USLE / RUSLE Empirical Models 267

Table 8.9 Selected USLE C Values for Cropland


(after Wischmeier and Smith, 1978) a

Crop Corn Corn Corn Soybeans Small


stage fc HP C LP* NT' HP' Grain*

Fal 0.44 0.65 — 0.45


SB 0.65 0.78 0.08 0.69 0.29
1 0.53 0.65 0.08 0.57 0.24
2 0.38 0.45 0.08 0.38 0.19
3. 0.20 0.26 0.07 0.18 0.06
4 — — 0.19
Figure 8.14 Example slope lengths (after Dissmeyer and Foster,
1984). Slopes: ( A ) If undisturbed forest soil above does not yield F o r a complete listing, see Handbook 537, Wischmeier and Smith
a

surface runoff, the top of slope starts with e d g e of undisturbed forest (1978).
soil and extends down slope to windrow of brush if runoff is concen­ ^Period Fal (Rough fallow), inversion plowing to secondary tillage.
trated by windrow; (B) point of origin of runoff to windrow if runoff Period SB (seedbed), secondary tillage for seedbed preparation until 10%
is concentrated by windrow; (C) from windrow to flow concentration of canopy cover, Period 1 (establishment), 10 to 50% canopy cover.
point; ( D ) point of origin of runoff to road that concentrates runoff; Period 2 (development), 50 to 75% cover. Period 3 (maturing crop), 75%
(E) from road to flood plain where deposition would occur; (F) on cover to harvest. Period 4 (residue or stubble), harvest to plowing or
nose of hill, from point of origin of runoff to flood plain where seeding.
deposition would occur; (G) point of origin of runoff to slight Corn, fall tumplowing, high productivity.
c

depression where runoff would concentrate. Corn, fall tumplowing, low productivity.
rf

Corn, minimum tillage, plant in stubble,


e

•f Soybeans after corn, fall turnplow, high productivity.


*Small grain, after corn, average productivity, 4 0 % cover after
emergence.
concentrated flow or until deposition occurs. A field
evaluation is helpful to d e t e r m i n e points at which flows
become concentrated a n d points at which deposition
occurs. Example slope lengths are given in Fig. 8.14.
thus conversion of rainfall t o stemflow typically reduces
erosion. T h e quantity of stem flow d e p e n d s on the type
Cover Factor C
of vegetation, with grasses having a high ratio of stem
Processes Being Considered in Cover Factor to drip flow and b r o a d leaf plants such as soybean or
T h e cover factor accounts for the effects of cover o a k having a low ratio. D r i p flow, if coming from leaves
above the ground, ground cover, root mass, incorpo­ high on trees, can contain as much energy as undis­
rated residue, surface roughness, and soil moisture on t u r b e d rainfall. T h u s , conversion of rainfall to drip flow
soil erosion. In general, these factors can be divided does not always lead to reduced erosion. In forest
into three categories: watersheds, of course, t h e litter will absorb much of
t h e energy of drip flow and protect the soil surface.
• Above-ground effects.
• Surface effects, including surface cover and surface Surface Effects. T h e impact of surface cover on rain­
roughness. d r o p energy impact a n d shearing forces of runoff is
• Below-surface effects, including active root growth, included in surface effects. In addition, surface effects
incorporated residue from previous crops, and effects include t h e impact of surface conditions on transport
of tillage and land disturbances on structure and capacity of runoff. T h e s e processes are affected by
consolidation. residue on t h e surface, surface roughness, and to an
Each of these effects are discussed below. extent the residue incorporated in the soil.

Above-Ground Cover. Above-ground cover intercepts Subsurface Effects. Subsurface effects include t h e ef­
rainfall and absorbs the raindrop energy. I n t e r c e p t e d fects of live root mass, residual root mass, incorporated
water moves to the surface by either stem or drip flow. residue, consolidation, compaction, prior land distur­
Stemflow is that intercepted rainfall that moves down bance, and soil moisture on erosion. T h e s e effects are
the main stem of a plant to the ground while drip flow the most difficult to quantify d u e to significant changes
is intercepted rainfall that collects on leaves and drips during a growing season and interactions between the
to the ground. Stemflow causes no interrill erosion; components.
268 8. Erosion and Sediment Yield

Tabulated Values for C 4/1-6/1 Stripped of vegetation (no cover, poor soil,
burned)
T h e simplest approach to defining the C factor is to
6/1-8/1 Active mining
use tabulated values for C that lump together all of the 8/1-9/1 Regraded (no cover, disked, poor soil)
factors discussed above. A n extensive data base of 9/1-11/1 Permanent seeding (first 60 days)
lumped p a r a m e t e r s collected for the U S L E is available 1 1 / 1 - 1 2 / 3 1 Permanent seeding (remainder of the year).
for use. Selected values are included in Table 8.8. for
The soils have a high tendency to rill, and slopes are 50%
construction, mining, and forest lands and in Table 8.9
with slope lengths of 150 ft. Estimate the weighted annual C
for agricultural lands. Additional data are given in factor using the tabulated values in Table 8.8, and calculate
Appendix 8B. the average annual erosion. Assuming that a 10-year storm
C factors change with land-use cover, resulting in occurs during the most susceptible period, estimate the 10-
varying erosion rates over the course of a season, even year single-storm erosion. Compare these predictions to the
with invariant month-to-month rainfall energy. This erosion rate from a corn field in West Kentucky with the
variation can be accounted for by weighting according same soil on a slope of 8% and slope length of 150 ft if the
to the fraction of R factor received during a given following sequence of operations is followed:
month, following the weighting p r o c e d u r e used for the 4/15 Seedbed preparation and planting
Κ factor in Example Problem 8.3. This weighting pro­ 5/15 10% canopy
cedure is illustrated in Example Problem 8.6. 6/15 50% canopy
7/1 75% canopy
10/1 Harvest
10/15 Turnplow.
Example Problem 8.6. Estimating erosion Assume that the corn crop has a low yield. In both cases,
with tabulated C factors on disturbed lands assume that Ρ = 1.0
Solution:
A 40-acre field near Middlesboro, Kentucky, has a Κ East Kentucky Strip Mine
value of 0.35 (English units tons · acre · hr/hundreds · 1. Estimating a weighted C factor. Computations for the C
acre · ft · tonsf · in.). The field is subjected to the following factors and'weighting factors are given in the table below.
sequence of operations: Middlesboro, Kentucky, is in zone 104 in Fig. 8.8. Using the
sum of the values in Column 6, the weighted C factor is
1/1-4/1 Dense forest (75% effective canopy) C = Σ column 6 / Σ column 4 = 63.84/100 = 0.64.

Example Problem 8.5 Weighted C Factors—Eastern Kentucky Strip Mine

Cumulative Percentage annual R Cover


percentage during factor Weighted
Time period* 1
Activity" annual Rb
period c
cover factor 1

1/1-4/1 Dense forest 13 13 0.001^ 0.013


4/1-6/1 Bare soil, burned 27 14 0.45' 6.300
6/1-8/1 Active mining 63 36 1.0 f
36.000
8/1-9/1 Regraded (bare) 80 17 0.94* 15.980
9/1-11/1 Perm, seeding 93 13 0.40* 5.200
(first 60 days)
11/1-12/31 Perm, seeding 100 7 0.05 Λ
0.350
(remainder of year)
Σ= 100 63.840

"Specified in the problem.


*Value from Table 8.1 (Zone 104 in Fig. 8.8).
''Incremental value from column 3.
''Table 8.8, item 5.
Table 8.8, item 7.
f Construction value = 1.0.
T a b l e 8.8, item 7.
'Table 8.8, item 4 (seedbed was disked in regrading operation).
'Column 4 χ Column 5.
Rill and Interrill Erosion Modeling: USLE / RUSLE Empirical Models 269
2. Estimating LS factors. From Table 8.6, m = 0.82 for 2. Estimating LS factors. From Table 8.6, m = 0.65 for an
50% slope and high tendency to rill. For a 50% slope, 8% slope and a high tendency to rill. For an 8% slope,
θ = t a n " 0.5 = 26.6°. Using Eqs. (8.43) and (8.40)
1
θ = t a n " 0.08 = 4.6°. Using Eqs. (8.43) and (8.39)
1

LS = [16.8 sin θ - 0.5] LS = [10.8 sin θ + 0.03]


72.6 72.6
0.82 150
150 [10.8 sin (4.6) + 0.03] = 1.44.
[16.8 sin (26.6) - 0.5] = 12.73. 72J>
72L6
3. Calculating average annual erosion rates. From Fig. 8.6,
3. Calculating average annual erosion rates. From Fig. 8.6, the average annual R factor is 250. Using Eq. (8.28)
the average annual R factor is 175. Using Eq. (8.28)
A = RKLSCP = (250)(0.35)(1.44)(0.50)
A = RKLSCP = (175)(0.35)(12.73)(0.64) = 63.0 ton / acre · year
= 499 ton/acre · yr. 4. Calculating single-storm erosion rates. The most sensitive
period is during the seedbed period with a C factor of 0.78.
4. Calculating single storm erosion rates. The most sensitive The 10-year single-storm R factor for the city closest to west
period is with a C factor of 1.0. The 10-year single-storm R Kentucky given in Table 8.2 would be for Memphis, Ten­
factor for Middlesboro is given in Table 8.2 as 63 (a similar nessee or 82, which compares favorably with Fig. 8A.5 in
value is obtained from Appendix 8A). Using the LS factor Appendix 8A. Using the LS factor calculated above and the
calculated above and the C factor of 1.0 C factor of 0.78,
A = RKLSCP = (63)(0.35)(12.73)(1.00) A = RKLSCP = (82)(0.35)(1.44)(0.78)
= 280.7 ton/acre · storm. = 32.2 ton / acre · storm
Comparison of Predicted Erosion Values
Western Kentucky Cropland
1. Estimating a weighted C factor. Western Kentucky is in Eastern KY Western KY
zone 105 in Fig. 8.8. C factors and weighting factors are strip mine cropland
tabulated below. The weighted C factor is
Annual (ton/acre year) e
499 63
C = Σ column 6 / Σ column 4 = 50.08/100 = 0.50. 10-year storm (ton/acre»storm) 281 32

Example Problem 8.5 Weighted C Factors—Western Kentucky Corn

Cumulative Percentage annual Cover


percentage R during factor Weighted
Time period* Activity* annual R b
period c
cover factor

1/1-4/15 Rough fallow 21 21 0.65 d


13.65
4/15-5/15 Seedbed (0-10%) 31 10 0.78 7.8
5/15-6/15 Period 1 ( 1 0 - 5 0 % ) 43 12 0.65 7.8
6/15-7/1 Period 2 (50-75%) 50 7 0.45 3.15
7/1-10/1 Period 3 ( 7 5 % - h v ) 85 35 0.26 9.1
10/1-10/15 Period 4 residue 88 3 0.26' 0.78
10/15-12/31 Rough fallow 100 12 0.65 7.8

Σ= 100 50.08

^Specified in the problem.


*Value from Table 8.1 (zone 105 in Fig. 8.8).
incremental value from Column 3.
J
Table 8.9, Column 2.
'Assumed the same as period 3.
^Column 4 x Column 5.
270 8. Erosion and sediment yield

The strip mine obviously has a much larger erosion rate by height, surface roughness, below-ground root mass and
almost a factor of 10. This is due partially to the greater C residue, prior cropping, and time. For cropped areas,
factors, but primarily due to the higher LS factor resulting Y o d e r et al. (1993) recommend that the values be
from much steeper slopes prevalent in that area. estimated for 15-day periods of time.

Canopy Cover Subfactor


T h e effects of canopy cover and height on energy
Cover Factors: Subfactor Approach reduction of falling rain are given by the canopy sub-
for Agricultural Lands factor. R a i n d r o p s either fracture into smaller drops
T h e reader not interested in the refinement offered with less energy or drip from leaf edges. T h e canopy
by subfactors may wish to skip to the supporting prac­ cover subfactor is
tice factor section and read the section on tabulated Ρ
factors. C c c = 1 - F e~°c
l
", (8.52)

w h e r e F is t h e fraction of surface covered by canopy


c
Introduction to the Subfactor Approach
and Η is the average canopy height in feet. This is
A n alternative to the tabulated values is t h e subfac­ the original relationship proposed graphically by
tor approach proposed by Wischmeier (1975) and Wischmeier and Smith (1978) in which it was assumed
Mutchler et al. (1982), which allows for m o r e detailed that the fraction of rainfall intercepted is equal to the
evaluation of the interacting processes affecting cover fraction of canopy cover. It was also assumed that
than possible with the tabulated values in Tables 8.8 intercepted rainfall leaves the canopy at height Η with
and 8.9. Originally proposed as a tool to develop com­ a d r o p size of 0.1 in. Q u i n n and Laflen (1983) reported
posite C factors for the U S L E , the subfactor approach that the relationship gave satisfactory results for cover
is used totally in the R U S L E . T h e actual algorithm for although the assumptions were not exactly correct. T h e
the R U S L E follows that of Laflen et al. (1985). Subfac­ r e c o m m e n d e d values for Η and F are listed in Table
c

tors are used to account for prior land use, canopy, 8.10A for selected crops.
surface cover, surface roughness, and soil moisture. In
this text, procedures for estimating cover factors are Surface Cover Subfactor
presented for three land uses: agricultural and range- T h e impacts of surface cover include a reduction in
land, disturbed forest, and construction and mining. In soil exposed to rainfall energy, reduction in transport
each case, the cover factor is developed as a function capacity, and deposition in p o n d e d areas. Included in
of several subfactors to account for the p a r a m e t e r s surface cover is residue, rocks, and other material in
listed above as well as special p a r a m e t e r s for construc­ contact with the ground surface. T h e surface cover
tion and forest lands. For cropland, the subfactors for factor is
above- and below-ground effects would change rapidly,
but with rangeland, the subfactors would change slowly. 0.08

T h e subfactors presented in this section are those of C s c = exp -bR c (8.53)


the R U S L E , as presented in the draft documentation 6 + R c

available as this text went to press. T h e original proce­


dures, proposed in a 1991 draft are likely to change. where R c is the fraction of ground cover, R is a G

Reference should b e m a d e to the final document variable to account for the effects of surface roughness
( R e n a r d et al. 1993) for further information. on the effectiveness of mulch, and b is a constant.
Using the subfactor analogy, the C factor for agricul­ R e c o m m e n d e d values for b are given in Table 8.10B.
tural and rangelands is defined by Residue cover, R is estimated from
c

R = l - e (8.54)
^ ^ plu ^ c c ^ sc ^ s r ^ s m » (8.51) c

where R is the fraction of residue cover, R


c
w IS
where C is the prior land-use factor, C
p l u is the c c residue weight (pounds p e r acre) and a is the ratio of w

canopy cover subfactor, C is the surface cover subfac­


s c area covered to mass of residue (acres per pound).
tor, C is the surface roughness subfactor, and C
s r is s m Example values for a are given in Table 8.10.
w

the soil moisture subfactor. These subfactors d e p e n d T h e r a n d o m roughness p a r a m e t e r in Eq. (8.53), R , G

on cropping and management and can be expressed as accounts for the impact of rainfall and buried residue
a function of residue cover, canopy cover, canopy on r a n d o m roughness and its further impact on the
οο 2
ιο in ιλ
CN CN CN
<n co oo R» O m
DODO*--—*

Μ /Ι I—
Ο Ο Ο
* 1

— <Ν m γο
2 ο (Μ I s
^
I I <
"8 ·§ 2> Ζ

Γ* ΙΛ ΙΛ
§ 11 I I i
ο - £

>> ο
3
*8 £υ1
Ϊ !
Ο Η Ζ

ο S" > » * I 8 fi 8 ϊ SI a s
(Ν 00 Μ ρ>η coρ

S £ -° η η η
Ο Ο 0

*9 2
- t α
CO ε γ-
r-
ο
in ^
γ- νδ
ο ο ο
»η
ο
CM CO .i I ζζ
u "3
Ο Ο Ο Ο Ο Ο Ο Ν η Tt
8 § — co Tf rf Ο O D D
Ο Ο Ο Ο Ο Ο s ζ Ο Ο
CO ΟΟ
CN Q Ο —
ν Ό Λ
£
2
Ti­ — οο m
ro ι- rt ro Tt
Ο Ο O D D
sι e - IN ΟΟ
Ο Ο r- cn cn cn οο
ι ΐ f - (Ν (Ν Γ) -<
ο ο ο ~ ~" ~" '
8 8 8 8 8 8 8
ill C3 Ο Ο Ο Ο Ο Ο Ο
| O m in
> — CO Γ­

ι
Ο Ο © - - «ι I
•3 2 Ο ο

δ •a
cx
•e §
— m 00 - ι 1
ε
ε V) £3

&0 Ο Ν ΟΟ ΟΟ ΟΟ ,
< OQ U U CO CO
Ζ Η &I
00 60 — —
2 CO CO CO U U

CO | 5 S
S 2° $ ρ S S g I
•η m m

ρ
υ"
Τ5

1| S

8 8 8 8
s δδ
C/3 ε- S O D D
Ο
« «ι ·5 G
lit C
Ο Q.
06 «λ3
ο ο g § 3 3 3 ι "δ •a
a °^ Ο m m
ω !
rrv I

D
ο £ S £ 3 3 3 ι 3
^ —• cn IN M IN 1

ε
οο
ε ^ ο — _ λ
Φ •a
Δ
ο
6* .
co ω 11 <
3 <
272 8. Erosion and sediment yield

Table 8.11 Selected Field Operations and Associated Parameter Values for
RUSLE (after Yoder et ai, 1993)

Percentage Soil
random surface residue Tillage surface
roughness buried depth disturbed
Field operations (in.) (%) (in.) (%)

Chisel (2-in. shovels) 0.9 25 8 100


Cultivator, row 0.6 30 4 85
Disk, 1-way (18- to 24-in. disks) 1.1 40 4 100
Drill, conventional 0.4 10 2 80
Harrow (tine) 0.4 5 4 100
Moldboard (8 in. deep) 1.9 90 8 100
Planter, row 0.4 15 4 20

mulch effects. T h e factor is given by total root zone. Values for n and a are given in Table i t

8.10F for selected plant communities in the Western


RG = (25.4/? R - 6)(1 - e -o.roi5*s) -o.i4/> e T ;
U.S.
T h e total below-surface mass in the u p p e r 4 in., R , s

R G > 0.0 (8.55) is the sum of root mass and buried residue, or

where R is the total r a n d o m roughness (inches) after


R
/? e = i? c r + R (8.57)
a field operation, P is the total rainfall (inches) after
T

the last field operation, and R is the total root and s

buried residue after tillage in the top 4 in. of soil where R and R
sr are live root mass and buried
br

(pounds per acre). Selected values for the live root residue, respectively (pounds p e r acre). T h e mass of
mass component of R , R , are given in Table 8.10A.
s sr
buried residue at any time d e p e n d s on the initial mass
T h e buried residue component of R , R is discussed s br
of residue at harvest, the rate of decomposition of
u n d e r the prior land-use subfactor. Total r a n d o m residue, and tillage sequences after harvest. In cases
roughness is t h e standard deviation of land surface where the initial mass of residue is not known, it can
elevation after furrows and slopes are removed from be estimated from the grain yield and residue-to-grain
calculations. Example values for r a n d o m roughness are ratio in Table 8. IOC. Residue decomposition in draft
given in Table 8.10D for rangeland and in Table 8.11 documentation available for R U S L E at the writing of
for tillage operations. this text was estimated by relationships from Gregory
et al. (1985), or
Below-Ground Root Mass Example d a t a are given in
Table 8.10A and Ε for below-ground live root mass,
R . For those crops where data are not available,
sr 1 - 14.1Ζλ ( ^ - 3 0 ) - ^ (8.58)
below-ground live root mass can be estimated from the
above-ground root mass by
where P is the fraction of initial mass remaining
R i

^sr ~ B
AG i i> n a (8.56) after D days, U is a constant for the given crop
x

producing the residue, R is the average radius of a


where R is the root mass in the u p p e r 4 in. (pounds
sr
residue stem, T is air t e m p e r a t u r e in degrees Faren-
ai

per acre), B is the above-ground biomass (pounds


AG
heit during the D days, A is an antecedent moisture
x mi

per acre), a is the ratio of root mass (pounds p e r acre)


t
term, and C is the carbon to nitrogen ratio for the
N

to above-ground biomass (pounds per acre), and n is k


residue. For periods w h e r e T < 30° F little decompo­ ai

the ratio of root mass in the u p p e r 4 in. to that in the sition occurs; hence P = 1.0. Selected values for U/R
Ri
Rill and Interill Erosion Modeling: USLE / RUSLE Empirical Models 273
and C are given in Table 8.IOC. T h e effects of se­
N wilting point to a d e p t h of 6 ft, the value for C s mis 0.0.
lected tillage practices on buried residue are given in A conservative estimate is to assume a value of 1.0.
Table 8.11. Values for A are estimated on an annual
m In the W e s t e r n U.S., particularly the Pacific North­
basis for the Western U.S. by west, Κ values are not varied with season; hence, soil
moisture corrections a r e in order. Information is given
A m = 0.018/? , A N (8.59) in A p p e n d i x 8C, T a b l e 8C.3, on replenishment and
depletion rates for these lands. Moisture balance com­
putations are m a d e on 15-day increments and com­
where is the annual rainfall in inches. For areas
p a r e d to field capacity a n d wilting point values to
east of the Rockies, A is estimated for 15-day periods
m
d e t e r m i n e C . Y o d e r et al. (1993) r e c o m m e n d that a
by
s m

linear relationship b e used b e t w e e n 1.0 at field capacity


and 0.0 at t h e wilting point. Soil moisture factors are
^ m i = 0.22i? 1 5 ) / , (8.60)
not used for rangelands.

where R , is the rainfall in 15-day increments. T h e


i5
Below-Ground Effects: Prior Land Use (PLU) Factor
cumulative value of residue decomposition effects, P , RC
T h e subfactor for prior land use ( P L U ) is used to
at any time is the product of P for t h e 15-day Ri
estimate the impact of prior cropping, tillage practices,
increments after residue was produced, or
soil consolidation, time, and biological activity on ero­
sion. F o r example, when a change is m a d e from a
m e a d o w to a cropland, t h e r e is a residual beneficial
PRC= UP i, (8.61)
i= 1
R
effect of the m e a d o w on erosion for 2 years. W h e n the
change is m a d e from cropland to meadow, there is a
where N is the n u m b e r of 15-day increments evalu­ residual detrimental effect for 3 years. T h e P L U factor
p

ated. is generalized to consider both type effects through the


At the writing of this text, the authors of R U S L E relationship
( R e n a r d et al., 1993) were planning to change t h e form
of the residue decay relationship in Eq. (8.58). T h e C p l u = D e- >«>,
en
C
(8.63)
changes are not expected to drastically change the
results for most situations.
where D e nis a density variable related t o tillage prac­
Surface Roughness Subfactor tices, C is a constant (Table 8.10B), and R is the
x s

a m o u n t of live roots and buried residue in the u p p e r


T h e direct impact of surface roughness on erosion is
4 in. of soil ( p o u n d s p e r acre). Values for the buried
given by the surface roughness subfactor. T h e indirect
residue must be estimated on the basis of: (1) tillage
impacts of surface roughness on the effectiveness of
impacts on buried residue as given in Table 8.11 and
mulch and residue as a surface cover are included
(2) residue decay. P r o c e d u r e s for making the estimates
u n d e r the surface cover factor in Eq. (8.53). T h e sur­
are given in Example Problem 8.7.
face roughness subfactor is given by
The parameter D en is used to account for the effects
of changes in surface density that occur as a result of
C s r = - e
0 X ) 2 6 R
°, (8.62) tillage practices. Tillage breaks soil particle bonds and
reduces density, thus increasing the potential for ero­
where R is defined by Eq. (8.55). Rainfall decreases
G
sion. With the passage of time and natural reconsolida-
surface roughness and thus decreases its impact on tion, these bonds reform and erosion potential is re­
erosion. Rainfall impacts are included in computation duced. Y o d e r et al. (1993) refer t o t h e work of
of R in Eq. (8.55).
G
Dissmeyer and Foster (1981) to show that D should
e n

vary exponentially from 1.0 for freshly tilled soil to 0.45


Soil Moisture Subfactor after 7 years with no tillage. T h e Dissmeyer and Foster
T h e soil moisture subfactor accounts for the effects d a t a are p r e s e n t e d in Fig. 8.15 as an estimator of D . e n

of antecedent moisture on infiltration. In general, the


effects of antecedent moisture on annual soil erosion Use of RUSLE to Estimate Soil Erosion
are accounted for by the seasonal variation in the Κ T h e R U S L E p r o c e d u r e allows the systematic evalua­
factor. For single storms, a correction may be n e e d e d . tion of t h e interactions between soil and plant proper­
W h e n the soil is near field capacity, the soil moisture ties that effect soil erosion; however, t h e procedures
subfactor, C , is 1.0. W h e n soil moisture is n e a r the
s m are computationally intensive and require a computer
274 8. Erosion and sediment yield
u residue from previous corn must be estimated from residue
ο
-ρ decomposition and tillage incorporation information. Param­
ο 1.0
€0 \ ο eters for Eq. (8.58) for predicting decay of corn residue left
& > Ο at harvest are taken from Table 8.10C, U/R = 0.0017 and
.8
C = 62. To utilize Eq. (8.58), the value for P must be
S3
CO r N

estimated for each time period and used multiplicatively.


Ri

C Calculations are summarized below by 15-day increments.


Ο \

"«3 Monthly precipitation is divided equally among 15-day peri­


CO δ « ods.
Δ
•i-H
i-M
• i
Ο
CO Ο Zonesville OH Τ
Δ Tyler. TX *15,
Observed Data Time period ( °atF ) (in.) Ρ b
ο • Lafayette, IN mi
ο
0)
10/1-10/15 60 1.5 0.33 0.888
8 10/15-11/1 60 1.5 0.33 0.888
11/1-11/15 50 2.5 0.55 0.876
11/15-12/1 50 2.5 0.55 0.876
Year s i n c e l a s t t i m e soil was tilled
12/1-12/15 42 2.0 0.44 0.940
Figure 8.15 Effects of reconsolidation on erodibility as reported by 12/15-1/1 42 2.0 0.44 0.940
Dissmeyer and Foster (1981). These values are recommended as 1/1-1/15 28 2.0 0.44 1.00
estimates of D in the prior land-use factor.
e n
1/15-2/1 28 2.0 0.44 1.00
2/1-2/15 32 2.5 0.55 0.987
2/15-3/1 32 2.5 0.55 0.987
3/1-3/15 45 3.5 0.77 0.870
or a major spreadsheet for any extensive use. A com­ 3/15-4/1 45 3.5 0.77 0.870
puterized version is available from the Soil and W a t e r 4/1-4/15 50 5.0 1.10 0.761
Conservation Society of America.
^Equation (8.60) A m i = 0.22 R l 5 i .
^Equation (8.58), using 15-day increments for D , ^ = [ΐ-(14,1)
(15) (0.0017) ( 7 ^ - 3 0 ) (^)f= [1-0.0058 ( 7 ^ - 3 0 ) (AJ] . 2

Example Problem 8.7 Use of RUSLE subfactors


to estimate the cover factor for agricultural land
From Eq. (8.61)
Estimate the C factor on April 15 for the following condi­
tions:
^RC= Π^, = 0.30.
Soybeans following corn. Corn harvested on Octo­
ber 1 with a yield of 6200 lb/acre; soybeans planted
on March 1. Monthly average temperatures: Octo­ The total residue mass present at April 15 is thus 0.300 times
ber, 60° F; November, 50° F; December, 42° F; Jan­ that left at harvest on October 15. From Table 8.10C, the
uary, 28° F; February, 32° F; March, 45° F; and residue-to-grain ratio for corn is 1.0; thus the residue pro­
duced equals the weight of grain or 6200 lb/acre. The total
April, 50° F. Soil moisture is at field capacity.
residue present at April 15, accounting for decay by P , is
Moldboard plowing on November 1; row planter on RC

thus
March 1. Precipitation during the period from
October 1 to April 15 is: October, 3 in.; November,
5 in.; December, 4 in.; January, 4 in.; February, 5 Residue = (0.300)(6200) = 1860 lb/acre.
in.; March, 7 in.; April, 10 in.
Of this residue, moldboard plowing on October 1 buried 90%
Assume that the soil is such that erosion is a typical rill and to a depth of 8 in. (Table 8.11). This would leave 45% in the
interrill mixture. upper 4 in. and 45% below 4 in., assuming a uniform burial
Solution: distribution. On March 1, row planting placed 15% (Table
1. Prior land-use factor [Eq. (8.63)]. From Fig. 8.15, D e n
8.11) of the mulch remaining on the surface (10% of original)
for freshly tilled soil (45 days after tillage) is still approxi­ into the 0 to 4-in. zone. Thus the fraction of original surface
mately 1.0. From Table 8.10B, the prior land-use parameter residue that is in the 0- to 4-in. zone on April 15 is
C is 0.00088. The prior land-use factor C is for the effects
x plu

of crops grown in previous years. Hence, the C residuep I u fraction of


parameters are for corn residue; whereas active growing original surface = [0.45 + 0.15(1 - 0.90)]
roots will be soybeans. April 15 is 45 days after planting; residue in upper 0 - 4 in.
hence from Table 8.10A, the growing root mass, / ? = sr

92 lb/acre for soybeans. The below-ground and above-ground = 0.465.


Rill and Interill Erosion Modeling: USLE / RUSLE Empirical Models 275

On a mass basis, the buried residue in this zone is thus roughness of 0.4 in.,

tf 6r = (0.465)(/V)(6200) RQ = [25AR R - 6][1 - £-0.0015/^-0.UP T

= (0.465)(0.300)(6200) = 865 lb/acre. = [(25.4)(0.4) - 6][1 - e - 0 0015(957)] -0.14(12)


e

Since planting incorporates only 15% of the surface residue RG = 0.591.


(Table 8.11), 85% of the residue at planting remains on the
surface. Thus, the fraction remaining on the surface on April Calculating surface cover subfactor. Equation (8.53) is
15 is 85% of the 10% left initially after turnplowing or used with R of 0.058, R of 0.616, and b of 3.5 to obtain
C G

(0.85X0.1) = 0.085. The above ground residue, K , is there­ w

fore 0.085 times the residue left, after accounting for decay, / 6 \ 0 0 8

Q = exp ( - 3 . 5 ) ( 0 . 0 5 8 )
or c
\ 6 + 0.591 j
/? w = (0.085)(/> )(6200) RC = 0.818.
= (0.085)(0.300)(6200) 4. Surface roughness subfactor [Eq. (8.62)].
= 158 lb/acre.
Q = £-0026/? G . ^-0.026(0.591) = Q 935

The total mass in the upper 0-4 in. of soil on a per inch basis
is the sum of the incorporated residue after accounting for 5. Soil moisture subfactor. Soil moisture is at field capacity;
decomposition, R , and the below-ground root mass, R , or
BT ST
thus

R S =R ST +R BT = 92 + 865 = 957 lb/acre. c = i.o.


sm

Using R = 957, D
S e n = 1.0 and C = 0.00088 in Eq. (8.63)
x
6. Calculating C factor [Eq. (8.51)].
for C plu
C =
^pluQxQcQrQm
Cpiu - £>cne- *
ClR
- (1.0)e- 0 0 0 0 8 8
< 9 5 7 )

= (0.431)(0.715)(0.818)(0.985)(1.0)
= 0.431. = 0.248.

2. Canopy cover factor C . From Table 8.10A, the canopy cc

height and canopy cover for soybeans at 45 days is Η = 0.5 ft


and F = 0.3. The cover factor from Eq. (8.52) is thus
c

C = 1 - (0.3)e"° 1 ( 0 5 )
= 0.715.
cc
Cover Factors: Subfactor Approach for Construction
and Mined Lands
3. Surface cover factor C . Parameters needed for surface sc

cover factor predictions in Eq. (8.53) are the fraction of Introduction


ground cover, R , the roughness parameter, R , and the
Soils that are drastically disturbed a n d / o r recon­
C G

coefficient b. From Table 8.10B, b = 3.5 for typical cropland


erosion. structed have erosion characteristics that are different
Fraction of residue ground cover R . From earlier calcu­ from agricultural lands. With passage of time, the
C

lation in part 1 of this problem, the above-ground residue forces of weathering a n d reconsolidation t e n d to move
weight fl is 158 lb/acre of corn residue. From Table 8.10C,
w
these characteristics m o r e toward agricultural lands.
A for corn residue is 0.00038. From Eq. (8.54)
w N o published subfactor p r o c e d u r e has b e e n devel­
o p e d for construction and m i n e d lands, but a modifi­
R c = l - £-0.00038(158) = o.058. cation of Eq. (8.51) seems t o b e in order. In addition to
the factors that are included in Eq. (8.51), other impor­
Surface roughness factor. Equation (8.55) is used to cal­ tant factors include effects of compaction and timing
culate R , which is then further used in Eq. (8.53) to
G
after reconstruction. N o published comprehensive pro­
calculate the effects of random roughness on both the surface
c e d u r e s have b e e n developed for all the subfactors;
cover subfactor and a surface roughness subfactor. From
h e n c e p r o c e d u r e s p r o p o s e d below are somewhat spec­
Table 8.11, the random roughness is 0.4 in. after row planting
on March 1. (Note that row planting would obliterate rough­ ulative, especially for below-ground biomass. T h e C
ness elements remaining from earlier turnplowing.) The pre­ factor for construction a n d mined lands could be rep­
cipitation between March 1 and April 15 from the tabulation r e s e n t e d by
in part 1 is 3.5 + 3.5 + 5.0 or 12.0 in. Also from part 1,
R = 1057 lb/acre. Thus from Eq. (8.55) with a random
S
C — C-rC D ^ p l u Q c Q c Q r ^ s m » (8.64)
276 8. Erosion and sediment yield
2.8 3.0
Ο Topsoil Q
2.6 Δ Shale
x Subsoil X Subsoil
Η 2.4 u Ο Topsoil
u ο
2.2 +>
u ο 2.0
2.0 ed
3 «Μ
X
Ο
CB
1.8
9 1.6 CO O O ^ s*
CO ^ ^ Γ Δ Ο
Δ
V
1.4 1.0
β 1.2 cn
1.0 α
4)
0.8 Q
12
1.0 1.1 1.2

Time After Reconstruction


Relative D e n s i t y
With Stockpiled Soil (Months)
( P l o t D e n s i t y / A v g Tilled D e n s i t y )
Figure 8.16 Subfactor to account for the effects of time after
reconstruction on the C factor (after Barfield et al., 1988). These Figure 8.17 Subfactor for the effects of compacted density on the
results are based on studies of reconstructed surface mine soils. C factor (after Barfield et al., 1988). These results are based on
studies of reconstructed surface mine soils.

where C is the subfactor for time and C


x is the D value for D en in Eq. (8.63) should b e conservatively set
subfactor for density and other factors as previously at 1.0.
described for Eq. (8.51) for agricultural lands.
Other Subfactors, C , C , , , C , C
c c s r s m

Time Subfactor T h e prior land-use factor discussed above for con­


After soil reconstruction, there is an a p p a r e n t loss of struction and mined lands should b e different from
soil structure resulting in greatly enhanced erosion, that of agricultural lands. O t h e r factors are a function
even u n d e r compacted conditions (Barfield et al., 1988). of what occurs on the surface and above ground and
With the passage of time and weathering, soil structure should be estimated the same as for agricultural lands,
recovers and erosion rates decrease. T h e impacts of given the a p p r o p r i a t e crop p a r a m e t e r s . However, since
such changes have b e e n evaluated by Barfield et al. growth in the early years after reclamation can be
(1988) who proposed the relationship shown in Fig. considerably different from agricultural lands, the crop
8.16. p a r a m e t e r s given in Table 8.10 should be adjusted
accordingly based on experience in the area. A n expe­
Density Subfactor rienced agronomist or reclamation specialist could be
A s soil is compacted artificially, its inherent resis­ of assistance.
tance to erosion decreases. This is in contrast to the
effects of reconsolidation after tillage, w h e r e the in­
crease in density causes a decrease in erosion (Barfield Example Problem 8.8 Cover Factor
et al., 1988; Israelson et al., 1980). In t h e latter case, for construction sites
the increase in density occurs as soil aggregate bonds
are enhanced, resulting in increased erosion resistance. Estimate the C factor for a reconstructed surface-mined
Mechanical compaction apparently does not create such land 6.5 months after regrading and reseeding with rye and
resistance. A subfactor for the effects of density changes fescue on October 1. One ton per acre of straw mulch is
due to mechanical compaction is given in Fig. 8.17. applied at seeding. During regrading, the soil is compacted to
1.1 times its loose tilled density due to compaction effects of
Prior Land-Use Factor reclamation machinery. Assume no seedbed preparation and
adequate moisture. Assume the same temperature and mois­
W h e n mulch is applied and the surface is planted to ture conditions as in Example Problem 8.7.
vegetation, Eq. (8.63) can be used to estimate C . p l u

Since vegetation does not grow as rapidly on recon­ Solution:


1. Subfactors for density and timing. Density subfactor for
structed land, the root mass estimates from Tables 8.10
relative density of 1.1 (Fig. 8.17)
should be decreased significantly. In the absence of
long-term data on erosion on reconstructed lands, the C = 1.5.
D
Rill and Interill Erosion Modeling: USLE / RUSLE Empirical Models 277
Timing subfactor for time after reconstruction of 6.5 months 5. Calculating C factor. From Eq. (8.62),
(Fig. 8.16)
C = 1.36. C = CCC C C C C
O T plu cc sc ST sm
T

= (1.5)(1.36)(1.0)(0.559)(0.347)(1.0)(1.0)
2. Subfactor for surface cover. As a first approximation,
assume no mulch decomposition and utilize values from = 0.395.
Table 8.8, item 2,

C sc = 0.18.

This would be an acceptable value for short periods after Cover Factors: Subfactor Approach for Disturbed
application of mulch and for longer periods if a rough esti­ Forest and Woodlands
mate were acceptable. As an alternative, Eqs. (8.53) through
Factors for Undisturbed Forests and Grassed Forests
(8.60) can be utilized. First, estimate the fraction of ground
cover by mulch (residue) using Eq. (8.54). For this equation, As discussed previously, first estimates of C factors
a = 0.00060 from Table 8.10C (assuming wheat straw), and
w
for undisturbed forests a n d grassed forest lands are
b = 3.5 from Table 8.10B. Using a mulch rate of 1 ton/acre b e s t e s t i m a t e d w i t h t a b u l a t i o n s d e v e l o p e d by
or 2000 lb/acre and a residue decay P of 0.300 from R C Wischmeier and Smith (1978) as summarized in Table
Example Problem 8.7, 8.8 a n d A p p e n d i x 8B. In this section, cover factors are
addressed for forests a n d woodlands that are disturbed
fl = (0.300)(2000) = 600 lb/acre.
to some d e g r e e .
w

Use of P from Example Problem 8.7 is justified since the


Introduction to Forest and Woodland Subfactors
R C

same temperature and moisture conditions were assumed.


From Eq. (8.54), Cover factors for forest and woodlands with some
disturbance can be estimated using procedures devel­
fl - l _ -Mw = 1 - £-0.00060(600)
o p e d by Dissmeyer a n d Foster (1984). Using a subfac­
c 6

= 0.302. tor analogy, the C factor is estimated from a product


of factors to account for b a r e soil, fine root mat, soil
From Eq. (8.53), assuming zero random roughness (R = 0),
reconsolidation, canopy, steps, depression storage, and
G

k 0.081 contour tillage. Dissmeyer and Foster (1984) evaluated


/ 0
C = exp -bR t h e accuracy of t h e m e t h o d on four plots and 35
sc
6 + R
watersheds in the southeast. T h e correlation between
r r

predicted and observed values was 0.90 and the aver­


age prediction e r r o r was ± 7 1 % .
= exp -(3.5)(0.302)
U + 0/
Bare Soil and Fine Root Mat Subfactor
= 0.347.
U n d i s t u r b e d soils in forested areas tend to have
A C of 0.347 is the more appropriate value since it takes
x
infiltration rates greater t h a n rainfall rates. T h u s , bare
into account residue decay.
soil a r e a s b e c o m e t h e primary source of sediment and
3. Canopy subfactor. This computation requires an esti­
mate of canopy cover. At 6.5 months, the primary canopy will runoff. A 0 % cover in forest lands frequently results in
be rye, which is a small grain. Since planting was in the fall, zero sediment yield as a result of the root mass, whereas
the crop is classed as winter grain. From Table 8.10A, the zero cover for agricultural land will normally have
canopy cover and height for small grains on agricultural lands some sediment yield. T h u s t h e cover factors for forest
after 6.5 months would be 1.0 and 2.5 ft. For an unprepared land use are different from agricultural land uses.
seedbed, the values would be smaller. For a first approxima­ T h e b a r e soil and root mass subfactors are given in
tion half these values will be assumed, or F = 0.50 and c
Table 8.12 for forest lands. T h e reconsolidation factor
Η = 1.25 ft. In actual practice, reclamation specialists should
discussed in t h e following section is also incorporated
be consulted for information on crop stands under reclaimed
into these values.
conditions. Using these values and Eq. (8.52),
Soil Reconsolidation Subfactor
C cc = 1 - F e~ c
0AH
= 1 - 0.50^"° 1(1 2 5 )
= 0.559.
After a tillage operation, soil reconsolidates natu­
4. Other factors. Since the surface roughness is zero, and rally as a result of wetting and drying cycles. Unlike
moisture is at field capacity C and C = 1.0. Immediately sr s m
mechanical compaction, this reconsolidation improves
after reconstruction, there will be no crop residuals; hence soil particle bonding with a resultant increase in ero­
sion resistance and decrease in erosion rates. Plots
Cp -1.0. l u maintained in a fallow condition at Zanesville, Ohio,
278 8. Erosion and sediment yield

Table 8.12 Selected Subfactors for Disturbed Forest Lands (after Dissmeyer and Foster, 1984). Additional Data Are Available in
Appendix 8D

A. Effect of bare soil, fine root mat of tree roots, and soil 3. Tilled soil with poor initial fine root mat in topsoil. Subsoil has good
reconsolidation on C factor structure and permeability

Time (months) since tillage


1. Untilled soils Percentage
bare soil 0 12 thru 36 72+
Percentage of bare soil with
dense mat of fine roots in 0 0.0000 0.0000 0.0000
top 3 cm of soil 1 0.0021 0.0025 0.0018
Percentage
bare soil 100 60 0 10 0.027 0.033 0.023
40 0.122 0.144 0.104
0 0.0000 0.0000 0.0000
100 0.530 0.630 0.450
1 0.0004 0.0007 0.0018
10 0.005 0.009 0.0123 B. Step effect on soil erosion
40 0.023 0.042 0.104
Percentage of total slope in steps
100 0.099 0.180 0.450 Percentage
slope 0 30 50 100

5 1.00 0.98 0.96 0.92


10 1.00 0.81 0.68 0.36
15 1.00 0.75 0.59 0.18
2. Tilled soils with good initial fine root mat in topsoil. Subsoil has good
structure and permeability 30+ 1.00 0.72 0.53 0.06

Time (months) since tillaee


Percentage C. Contour tillage subfactors for forestlands
bare soil 0 12 and 72+ 24+ thru 60
On Degrees off contour
Percentage contour
0 0.0000 0.0000 0.0000
slope 0 15 45 90
1 0.0014 0.0018 0.0020
10 0.019 0.023 0.026 1.0 0.80 0.88 0.94 1.00
40 0.083 0.104 0.117 10.0 0.80 0.88 0.94 1.00
100 0.360 0.450 0.510 19+ 1.00 1.00 1.00 1.00

had a decreased erosion rate with consolidation after storage. T h e depression storage subfactor is estimated
tillage, decreasing to 0.45 of the original tilled value by selecting the appropriate factor from Fig. 8.18. For
after 7 years. Most of this decrease occurred prior to 3 example, if the surface is composed primarily of 6-in.
years. Subfactors for reconsolidation are given in Table clods, the factor would be 0.5. Likewise, if the surface
8.12A, in combination with bare soil and fine root is b a r e soil c h o p p e d with l-in.-deep slits along the
subfactors. W h e n considered alone, the soil reconsoli­ contour, the factor would b e 0.8.
dation subfactor proposed by Dissmeyer and Foster is
given in Fig. 8.15. Step Subfactor
In forest lands, debris washed by surface runoff will
Canopy Subfactor form small dams with subsequent ponding and deposi­
A canopy of forest vegetation will have an impact tion. T h e final result is a series of steps that have the
similar to that of agricultural crops if the canopy heights a p p e a r a n c e of mini-terraces, as shown in Fig. 8.19.
are similar. Thus, the above-ground canopy subfactor Based on observation from 100 steps throughout the
should be estimated from Eq. (8.52). southeast, Dissmeyer and Foster (1984) found that the
deposition behind the slopes occurred at a slope of
Depression Storage Subfactor 3 % , as shown in Fig. 8.19. Using this measurement, the
W a t e r stored in depressions cannot transport sedi­ Foster and Wischmeier (1974) relationship for irregu­
ment off site; thus erosion is reduced by depression lar slopes, and the assumption that steps were small
Rill and Interill Erosion Modeling: U S L E / R U S L E Empirical Models 279

Smooth-no Storage b e t w e e n soil c l o d s


storage
1.0

Storage c r e a t e d b y 0-8 0.8


tractor treads
0.8

Small a m o u n t of a n d d o w n s l o
P e 0 9

bare soil
0.2 t o 0
Figure 8.18 Subfactors for on-site depression storage in forests (after Dissmeyer and Foster, 1984).

and randomly distributed, Dissmeyer and Foster (1984) that ridges from disking are not as high as typical
developed the relationship for steps given in Table agricultural rows. T h e s e factors are given in Table
8.12B. In general, this relationship should be used 8.12C.
when forest lands are being disturbed for logging or
other operations.
Example Problem 8.9 Estimating forest lands cover
Contour Tillage Subfactor factors with subfactors
C o n t o u r disking and plowing in forests that have
Estimate the cover factor for a forest land that has been
been disturbed will generally reduce erosion. Such
subjected to logging operation. In the logging operation, 40%
operations are most effective w h e n conducted on t h e
of the surface area is disturbed. Tracked vehicles that gener­
contour. W h e n this has not b e e n accomplished, correc­ ate random slits 1 in. deep were used. The remaining vegeta­
tions are necessary. tion provides only 30% above-ground canopy cover at a
Dissmeyer and Foster (1984) modified the U S L E height of 20 ft. Ground slope is 10%. Estimate the C factor
factors for contour plowing to account for disking on assuming: (A) a fine root mat covers the entire area and (B)
the contour, deviations from the contours, and the fact logging removes the root mat.
280 8. Erosion and sediment yield

used to evaluate the effects of contour tillage,


stripcropping, terracing, subsurface drainage, and dry­
land farm surface roughening. T h e effects of sod-based
crop rotations, minimum tillage, residue management,
and humid area surface roughening are included in the
C factor, as discussed in the previous section. T h e Ρ
factor is typically used only for agricultural lands and
rangelands, but could be used with some caution on
Debris Dam
Original Ground Line construction and disturbed lands.
A first approximation to the Ρ factor can be devel­
Figure 8.19 Illustration of step formation in forest lands (after
o p e d by using tabulated values from the U S L E data
Dissmeyer and Foster, 1984).
base. This can be used by the practitioner interested in
first or rough estimates, particularly for planning pur­
poses. For m o r e refined estimates that allow for de­
Solution: tailed consideration of a variety of combined practices,
With Fine Root Mat a subfactor approach from the R U S L E would be pre­
1. Subfactor for reconsolidation, bare area, and fine root ferred. Both procedures are presented here.
mass. Using Table 8.12A, with 100% fine root mass, 40%
bare soil, untilled Tabulated Ρ Factors from the USLE
Subfactor 1 = 0.023. Tabulated total values for the Ρ factor are available
2. Canopy subfactor. From Eq. (8.52) with a canopy cover for contouring, strip cropping, and contour terracing,
of 30% at a height of 20 ft, as given in Table 8.13. A detailed discussion of the
values is beyond the scope of this text. A few com­
Subfactor 2 = C = 1- F e~ 0lH
cc c
ments, however, are in order. First, it should be noted
= 1 - 0.3e- ( ( U X 2 0 )
= 0.96. that some of the practices have limits within which they
3. Depression storage subfactor. From Fig. 8.18, slits up- are assumed to b e totally effective. T h e s e limits de­
and downslope would have a subfactor of 0.9 and across crease with increasing slope. For example, Ρ values
slope would be 0.8. Using an average increase with slope, and the maximum effective slope
Subfactor 3 = 0.85. length for contouring decreases with slope. Limits on
contouring result from the decreasing surface storage
4. Step subfactor. Assume that all disturbed areas (40%)
with steeper slopes and the greater tendency to form
are in steps. From Table 8.12B, with a ground slope of 10%,
rills at steeper slopes.
and 40% of the area in steps,
Subfactor 4 = 0.75. The RUSLE Subfactor Approach to the Ρ Factor
5. Contour tillage subfactor. Disturbed areas are not tilled; A n alternative to the tabulations in Table 8.13 for
therefore, contour tillage factor is 1.0. the Ρ factor is to use a subfactor analogy from the
6. Calculating C factor. R U S L E , or
C = (0.023)(0.96)(0.85)(0.75)(1.0) P = PP P ,
c st ter (8.65)
= 0.014.
w h e r e P is the contour subfactor, P is the strip
c st

Fine Root Mat Removed cropping subfactor, and P is the terracing subfactor.
t e r
The only change in the solution is to subfactor 1 above.
Use of this subfactor analogy allows a more detailed
With zero fine root mat in the disturbed area, subfactor 1 for
evaluation of factors affecting P, particularly when
the effect of fine root mat and soil reconsolidation from
Table 8.12A with 40% bare area would be 0.104. Thus the C considering a combination of practices. Also, it allows
factor becomes a correction for the impact of large storms on contour­
ing.
C = (0.104)(0.96)(0.85)(0.75)(1.0) = 0.064, At the writing of this text, the R U S L E was in draft
form. Revisions are likely to be m a d e in some of the
subfactors before final printing.
Conservation Support Practice Ρ Factor
Contour Support Factor P T h e contour support fac­
c

Introduction tor accounts for the impact of tillage on the contour on


T h e conservation practice factor, P, by definition is soil erosion. If the surface is tilled u p - and downslope
the ratio of soil loss from any conservation support or is relatively smooth, a drainage p a t t e r n that allows
practice to that with u p - and downslope tillage. It is e r o d e d sediment to be readily transported downslope
Rill and Interill Erosion Modeling: USLE / RUSLE Empirical Models 281

Table 8.13 Selected USLE Ρ Factors (after Wischmeier and Smith, 1978)

1. Ρ values and slope length limits for contouring


Land slope Maximum length*
percentage Ρ value (ft)

1 to 2 0.60 400
3 to 5 0.50 300
6 to 8 0.50 200
9 to 12 0.60 120
13 to 16 0.70 80
17 to 20 0.80 60
21 to 25 0.90 50

2. Ρ values, maximum strip widths, and slope length limits for contour
stripcropping
Ρ values^
Land slope Strip width c
Maximum length
Percentage A Β C (ft) (ft)

1 to 2 0.30 0.45 0.60 130 800


3 to 5 0.25 0.38 0.50 100 600
6 to 8 0.25 0.38 0.50 100 400
9 to 12 0.30 0.45 0.60 80 240
13 to 16 0.35 0.52 0.70 80 160
17 to 20 0.40 0.60 0.80 60 120
21 to 25 0.45 0.68 0.90 50 100

3. Ρ values for contour-farmed terraced fields

Farm planning Computing sediment yield

Land slope Contour Stripcrop Graded channels Steep backslope


Percentage factor factor sod outlets underground outlets

lto2 0.60 0.30 0.12 0.05


3 to 8 0.50 0.25 0.10 0.05
9 to 12 0.60 0.30 0.12 0.05
13 to 16 0.70 0.35 0.14 0.05
17 to 20 0.80 0.40 0.16 0.06
21 to 25 0.90 0.45 0.18 0.06

^ i m i t may be increased by 25% if residue cover after crop seedings will regularly
exceed 50%.
A for 4-year rotation of row crop, small grain with meadow seeding, and 2 years of
b
y

meadow. B, for 4-year rotation of 2 years row crop, winter grain with meadow seeding,
and 1-year meadow. C, for alternate strips of row crop and small grain.
Adjust strip-width limit, generally downward, to accomodate widths of farm equipment.
c

develops. If tillage is on the contour, flow collects in a m o u n t of runoff, as well as the p e a k intensity. W h e n
the furrows between tillage ridges, allowing significant contour tillage marks have a cross-contour component,
amounts of deposition. T h e effectiveness of contouring as they frequently do, t h e effectiveness of contouring is
d e p e n d s on the ability of the tillage marks to store also reduced. A s flows move downslope, the quantity of
runoff and is obviously impacted by the size or rough­ runoff increases, reducing t h e effectiveness of a given
ness of the tillage system, the slope of the system, the contour tillage system. T h u s a critical slope length is
Ο Ο Ο Ο Ο Ο 14
1i Ο Ο Ο Ο Ο Ο
Χ) £ δ Ε; 3; 9
Ο
Ο
Ο
Ο
Ο
Ο
00
Ο
90
Ο
Μ
Ο
If
5 1
Ο Ο Ο Ο Ο Ο

Ο Ο Ο Ο Ο Ο
Q νθ Ο τ* 00 Q 00 «-Η ^ · so * - ·
ON 00 00 Ρ- νθ Φ
Γ**· ON 00 00 ON
Ο' Ο Ο Ο Ο Ο" D D D D D
ΟΟ ΟΟ Γ - R- NO Φ
Ο Ο Ο Ο Ο Ο
Β-1
C»~> Σ\ •<*· Ο
Γ- νθ νθ ν©
Ο Ο Ο Ο

I
»η οο —> fN
οο r » r- νο
Ο Ο Ο Ο Ο Ο

Σ\ (Ν «Η R- § 9
&s
Γ"» Ρ* νθ «Ο
Ο Ο 4J Μ <Λ
D O D O
Ο 2 ,ω ,u
M »M
Ο R-
TJ-

Ο Ο' Ο" Ο na na

Ο ^ Η φ M Ο I 8.'
«2 ΟΟ R» ΝΟ CO «S
Ο Ο Ο Ο Ο Ο

(Ν m m rn <Ν Ο
Γ-· Ό «η m cm
D D D D D D
cs °t 00
Ο Ο Ο Ο Ο Ο Ο
Ν© ·Η vC Γ- ON O
^ M C* <S
4> Λ
D D D D D D

ON 8.
Ο
ϊ
Β

1
On
1 t \0 Ν Χ Φ

•A 1 Χ* © Ο Ο Ο Ο Ο

U
Π R» OO O N Ο
O D D - ! ο' a
8—
I 5
a I υ
CO
D
at δ
s
Μ m
Ο Ο Ο' Ο
(Ν (Ν Γ-
Ο
Ο Γ-
Ο Ο'
ON O
~
a

3
'•3
CO I M ν> νο Ό OO Ο
a. Ο Ο Ο Ο D ~
δ-
Ι 3 S

co
}ι 8 Τ 00 00 Ο
I S Ρ
Ο Ο —'

Φ
ο | | 1* « Ο Ο Ο Ο Ο Ο Ο Ο Ε ·2?
5

8 £1
Ο Ο
«Η Ο
©

(δ ·§
«*1
Rill and Interill Erosion Modeling: USLE / RUSLE Empirical Models 283
typically defined, beyond which contouring effective­
ness is decreased.
T h e size of the ridge-furrow storage system, slope, 0)
To
the degree of the cross-contour c o m p o n e n t of t h e >
tillage marks, the runoff amount, and p e a k runoff r a t e
impact the effectiveness of contour tillage. Using t h e
limited data base available plus simulations with the
I
00
SJ
C R E A M S model, Foster et al. (1993) developed Ρ
subfactors for the impacts of contouring, which include
these p a r a m e t e r s . Base values for contouring, P , are b
c
υ
given in Table 8.14A for slopes that d o not exceed t h e •
co
critical slope length. For slopes that exceed t h e critical
slope length, values for P are given in Table 8.14B.
b ο
+>
Corrections for cross contouring and storm intensity CO

were developed using the C R E A M S model and a r e


90 100 110 120 130 140 150 160
given in Figs. 8.20 and 8.21. In revisions being p l a n n e d
to the initial draft of R U S L E , equations will b e used in R Factor for 10 yr. Storm (ft · tonsf •in/ac»hr»yr)
lieu of Figs. 8.20 and 8.21. Figure 8.20 Effect of storm erosivity on the contouring subfactor
Foster et al determined critical slope lengths from a (after Foster et al., 1993).
simplification of a mulch stability analysis, or

10471rt|1.5
(8.66) A n alternative to rotating strips is to include perma­
1.667„
(EI ) S
lo
on

n e n t strips of grass on the slopes. If the grass strips are


located within t h e growing crops, t h e strips are re­
w h e r e A is t h e critical slope length, n is Manning's n,
ferred to as buffer strips. If they are located along the
c t

EI is the 10-year storm EI index in English units, S


b o t t o m of t h e slope, they are referred to as filter strips.
l0

is slope in feet per foot, and r is a dimensionless runoff


Buffer strips, being p e r m a n e n t l y located, d o not con­
reduction factor to account for infiltration.
tribute as heavily as strip cropping to soil conservation
as t h e soil is not filtered in a location w h e r e it is used
Strip Cropping Conservation Support Factor P Strip
for further production of the row crops. T h u s , if o n e is
x

cropping is the use of alternating strips of close grow­


considering conservation planning, less credit would be
ing vegetation such as grasses and legumes b e t w e e n
strips of clean-tilled or nearly clean-tilled row crops, all
planted on t h e contour. Typically, t h e strips are ro­
tated.
100

Ii \
80
0.8 1
OB
OB
CP
d c

60 X W
a Storm Severity ^
ο Adjustment Value ( P ) s

Η From Figure 8.20 Ν


M 40
d
•Η

3α 20 i

οο ι
ι i
I 1

0.2 0.4 0.6 0.8 0.2 0.4 0.6 0.8 1.0

Row Grade/Field Slope Contour Factor Value P„


Figure 8.21 Adjustment to contour subfactor to account for grades off the contour and storm severity (after Foster et al.,
1993).
284 8. Erosion and sediment yield

given for the filter strip than if one is considering


sediment yield and off-site effects. T h e filter strips, Example Problem 8.10 Estimating conservation
being located permanently below the cropland, would practice factor
not contribute at all to maintaining soil productivity;
hence they would be considered ineffective as a conser­ A planner is interested in the impact of a proposed con­
servation plan on agricultural production. The plan includes
vation practice, but would contribute to controlling or
contour strip cropping with conventional tillage sloped at 2%
reducing sediment yield.
to the contour, and tile outlet terraces, spaced at 100 ft.
As flow moves from the clean-tilled strip through the Terrace channel slope is 1.0% and general downhill slope is
grass strip, transport capacity is decreased a n d sedi­ 8%. The EI index (R factor) for a 10-year storm is 100
30

ment may b e deposited. T h e effectiveness of the strip (English units) and ridge heights from tillage are moderate
d e p e n d s on strip width, slope, and type of tillage. (4 in.). The slope has equal width strips of alternate row crop
Using the C R E A M S equation, Foster et al. (1993) and winter small grain that are rotated.
developed a computational procedure that is included
in the R U S L E computer program. Foster et al. also Solution:
published typical subfactor values for selected strip 1. Contouring subfactor. From Table 8.14A, the uncor­
crops and buffer and filter strips. Selected values are rected contouring subfactor for 4-in. ridge heights (moderate)
given in Table 8.14E. and 8% slope is P = 0.55. Also, from the same table, the
b

critical slope length is greater than 100, so the value of 0.55 is


Terracing Support Factor, P Terraces collect flows appropriate. From Fig. 8.20, the storm severity adjustment
t e r
factor P for a 10-year R of 100 and P of 0.55 is P = 0.55.
from slopes and divert them to a stabilized waterway or s b s

The ratio of tillage slope to downhill slope is 0.02/0.08 =


to closed outlets, thus preventing long slope lengths. In
0.25. From Fig. 8.21 with P of 0.55 and a grade-to-slope
s
addition, flat sloped terrace channels cause deposition. ratio of 0.25,
Foster and Highfill (1983) developed factors to account
P = 0.76.
for the effects of terraces on soil loss, also accounting c

for deposition in channels. T h e net soil loss is that lost 2. 'Strip-cropping subfactor. From Table 8.14E, the strip
from the slopes minus channel deposition. Values for cropping subfactor for RC-WSG rotation is
the conservation practice factor for terracing, P , are ter
Px = 0.86.
given in Table 8.14C. In addition to using P subfac­
t e r

tor, the slope length used in the R U S L E , should be the Since the interest is in conservation planning (effects on crop
terrace interval. production), this is the appropriate number.
Terraces reduce sediment yield in two ways, by de­ 3. Terrace subfactor. From Table 8.14C for terrace spacing
of 100 ft and closed outlets
creasing the slope length and by allowing deposition in
the terrace channels. These two factors are considered P ter = 0.5.
separately in the R U S L E . T h e impact of slope length
4. Calculating support practice factor. From Eq. (8.65),
change is included in the LS factor by reducing the
slope length in the calculation. T h e effects of deposi­ Ρ = P e s t e r = (0.76)(0.86)(0.5) = 0.33.
tion are included in the terracing factor in Table 8.14C.
T h e factor Pter accounts for deposition in t h e terrace
channel. T h e impact of terraces on the loss of soil from
the slope is reflected in the slope length factor calcula­ Prediction of Annual Erosion Using RUSLE
tion based on the terrace spacing. Estimating annual erosion with the R U S L E can be
Unlike the U S L E data base, the impact of contour­ quite tedious, d u e to a n e e d to consider so many
ing is not included in the Ρ subfactor for terracing. variables that change with time. T h e p r o c e d u r e , how­
Hence, to get a Ρ factor for both contouring and ever, is readily a d a p t e d to a spreadsheet or to a com­
terracing (the normal combination), the two R U S L E p u t e r algorithm. As discussed earlier, computerized
subfactors must be multiplied together as given in Eq. version of the R U S L E is available from the Soil and
(8.65). W a t e r Conservation Society of America.

Rangeland Support Practices Rangeland t r e a t m e n t


practices include surface roughening by tillage or pit­ RILL AND INTERILL EROSION MODELING:
ting, contouring, and terracing. T h e impact of terracing COMMENTS ON PROCESS-BASED MODELS
and contouring must be evaluated with the support
factors discussed above. T h e effectiveness of other R e c e n t developments have led to process-based
practices is given in Foster et al. (1993). models such as C R E A M S (Foster et al., 1980a, b) and
Calculating Concentrated Channel Flow Erosion 285
W E P P (Lane and Nearing, 1989). T h e s e models are bank angle, sediment deposition, and vegetation
sufficiently complex to require a c o m p u t e r for solution; establishment.
thus illustration of model applications is beyond the (5) T h e n a t u r e of the sediment e r o d e d and trans­
scope of this chapter. However, since these models are p o r t e d affects t h e morphology of the channel and
being widely used, a discussion of the basic concepts the n a t u r e of t h e channel adjustment.
seems appropriate a n d is given in a section at the e n d
Models of the channel headwall migration and channel
of this chapter.
b a n k failure are currently being developed, but are not
available at this writing.
In general, channel erosion estimates are highly em­
CALCULATING CONCENTRATED CHANNEL pirical, relying on field surveys. O n e exception would
FLOW EROSION be the use of the model of Foster and L a n e (1983) and
its derivatives for c o n c e n t r a t e d flow erosion in terraces
Background and e p h e m e r a l gullies. T h e s e models are discussed
below.
Concentrated flow erosion is classed as channel ero­
sion when the location of channels is controlled by t h e
macrorelief. As discussed earlier, concentrated flow Foster and Lane Model
erosion is classed as rill erosion when the location of
T h e Foster and L a n e (1983) model is composed of
channels is controlled by the tillage microrelief. Con­
four major elements:
centrated flow erosion, w h e n controlled by the
macrorelief may be classed as e p h e m e r a l gully erosion, • A n equilibrium channel width model
classical gully erosion, or channel erosion. T h e s e chan­ • A model for conveyance function
nels erode by three mechanisms: • C h a n n e l erosion model prior to reaching a nonerodi­
ble layer
• channel bed degradation due to shear • C h a n n e l erosion model after reaching a nonerodible
• channel wall failure layer.
• knickpoint (headwall) advance.
Foster a n d L a n e developed t h e model for steady-state
A characteristic of concentrated flow channels is that flow, but use has extended to varying flow rates. A n
there is no smooth transition from a zone of wide overview of t h e model is given below. Storm et al.
shallow flow to an incised channel with d e e p e r flow. (1990) and Foster (1982) discuss the model further.
R a t h e r , the transition occurs abruptly at a knickpoint
or headwall. In fact, several such transitions may occur Equilibrium Channel Geometry
in a concentrated flow channel. Eventually, t h e knick- T h e basic d e t a c h m e n t relationship is the shear ex­
points merge, making one incised channel. cess concept, or
Physically based m e t h o d s for locating knickpoints
are not well developed. Present analytical m e t h o d s are D rc = /C (r-r ), r c (8.67)
based primarily on geomorphological studies with lim­ where D rc is d e t a c h m e n t rate potential ( k g / m · sec),
ited transferability to other climatic regions. In general, K is rill erodibility ( m / s e c ) , and τ and T are actual
v c

these studies indicate that some threshold flow nor­ and critical tractive force respectively (Pa). T h e data
mally exists above which the channel is incised. base for the W E P P model ( L a n e and Nearing, 1989)
Harvey et al. (1985) summarized much of the work can b e used to estimate K and r . In development of
r c

on gully and channel erosion. Based on the existing their equilibrium channel width model, Foster and
literature, they concluded the following: L a n e (1983) assumed that a symmetrical distribution of
shear exists, given by
(1) A gully may develop in a short time d u e to exceed­
ing an intrinsic or extrinsic threshold.
τ* = - = 1.35[ΐ-(1-2*,) · ]; 2 9
X* < 0.5,
(2) T h e response of the system to gullying is complex;
secondary responses complicate the adjustment to (8.68)
change.
(3) Empirical data bases for a h o m o g e n e o u s region w h e r e τ * is a dimensionless shear, τ is actual shear, r a

can be used to estimate thresholds. is the average shear stress given by


(4) All incised channels follow the same evolutionary
r = yRS, (8.69)
trend: initiation, headwall migration, channel a

widening, channel slope reduction, reduction of a n d X* is a normalized distance along the wetted
286 8. Erosion and sediment yield
0.16

V
0.14

\
Ν
Ν
\
0.12 2
\

\
s \ 0.10 <2 S3
Ο
\ 51
\
\
\ 0.08
\
«υ

t
\
V 0.06
\ fl
cd
\
\
\ 0.04 -3
Normalized Width - W« \ g
Normalized Hydraulic Radius - R » \ ο ο
\ 0.02 2 υ
V

\> 0.00
0.00 0.10 0.20 0.30 0.40 0.50 0.00 0.10 0.20 0.30 0.40 0.50

Normalized Critical Distance, X* Normalized CriUcal Distance, X , c

Figure 8.22 Normalized equilibrium characteristics for an eroding Figure 8.23 Conveyance function of Foster and Lane (1983). Tab­
channel (adapted from Lane and Foster, 1980). Tabulations are ulations are given in Appendix F, Table 8F.1.
given in Appendix F, Table 8F.1.

etry by: (1) selecting a value for X* (X* when τ * = c

perimeter starting at the water surface. X* is given by T * ) , (2) dividing the boundary b e t w e e n X* = 0.5 and
c

X* = X* c into 50 segments of length ΔΑ"*, (3) deter­


= */WP, (8.70)
mining an average Χ*, τ * , a n d thus a for each
where W P is wetted perimeter. Equations (8.68) and segment, (4) determining the final coordinates of the
(8.69) can be used to show that the maximum shear on segment, knowing α, ΔΧ*, and the starting coordi­
the channel bottom is 1.35 yRS at X* = 0.5. T h e nates, a n d (5) starting with X* = 0.5, repeating (3)-(5)
distribution given by Eq. (8.68) is symmetrical about until X* = X* . At X* , t h e boundary is vertical to
c c

X* = 0.5. X* = 0.0. Knowing t h e coordinates of X* , the values c

Foster and Lane (1983) also defined normalized for a n d / ? * were d e t e r m i n e d . By varying X* , the c

channel hydraulic radius and channel width as relationship in Fig. 8.22 was developed.

R* =*/WP (8.71) Conveyance Function


and
can b e d e t e r m i n e d once X* is known. Rela­
c

(8.72) tionships a r e n e e d e d to estimate X* and equilibrium c

where W is the equilibrium channel width. A normal­


eq
width, W . T o accomplish this, Foster and L a n e (1983)
eq

ized p a r a m e t e r known as X* was also defined as X*,


c
developed a conveyance function to predict X* . Start­ c

where τ = T . Using procedures defined below, the


c
ing with M a n n i n g ' s equation and Eqs. (8.69)-(8.72),
relationship between X* and R* and
c w e r e de­ they showed that
veloped as shown in Fig. 8.22. 3/8

T o define the channel geometries in Fig. 8.22, t h e * ( * * c ) =


1
ys nQ
(8.74)
assumption was m a d e that the channel erodes verti­
cally, rapidly adjusting to an equilibrium width. T h e w h e r e g(X* ) is a function only of X* as defined by
c c

rate of vertical movement was assumed to b e constant Fig. 8.23. If g(X* ) is greater t h a n 35, a value of 35 is
c

across the entire cross section at a rate defined by the typically used. T h e development of Eq. (8.74) is illus­
maximum shear stress. T h e rate of movement normal trated in Example Problem 8.11.
to the channel boundary at any point was defined by
t h e actual shear stress. T h e angle between t h e normal
and vertical movement would be defined by Example Problem 8.11 Development of conveyance
τ - τ„ *c function for Foster and Lane model
cos a = . (8.73)
1.35T 1.35T* — τ * c
Starting with Manning's equation, show that the con­
Utilizing Eqs. (8.67)-(8.72), Foster and L a n e (1983) veyance function of Foster and Lane (1983) given by Eq.
developed relationships for normalized channel geom­ (8.74) is correct.
Calculating Concentrated Channel Flow Erosion 287
Solution: Manning's equation [Eq. (4.23)] can be written in quantities to channel geometry. By assuming a rectan­
SI units as gular geometry with an equilibrium width, Foster and
L a n e (1983) showed that
Q = - R 2
^ 2
. (a)
η
WP =
w 2

(8.75)
Using the definition R = Λ / W P and R* = /?/WP,
W eq - 2R '
R 2

A = (b) w h e r e W is the equilibrium width. Utilizing Manning's


R* eq

equation, it can b e shown that


Hence
S/3 l/2 3/8
nQ
R S

(c) WP = (8.76)
nR, V5
and
3/8
a n d from Eq. (8.72)
nQ
R = l/2 R/.3 g
(d)
S
3/8
nQ
W*R-^\ (8.77)
From the definition

(e) T h u s , knowing t h e conveyance function, values for


yRS'
X* ,
c R* and can be determined. Knowing R*
a definition for R can be written as and , a value can b e d e t e r m i n e d for W P and W . eq

Knowing R * and W P , R can b e determined from


R =
ySr*
(0
R = R*V/P. (8.78)

Equating Eqs. (d) and (f),


It is significant to point out that Eq. (8.77) would
1
3/8 predict that W is a power function of discharge for a
y_s_ nQ tq

=
(g) given slope, critical tractive force, and bulk density.
rR/
c
3 8
r c
yfS
This is in keeping with empirical relationships given in
C h a p t e r 10 relating channel width and flow rate.
Since both T and R* can be predicted as functions of
# c X* ,
c

then Eq. (g) can be written T h e relationships developed to this point are for
channel geometry. Procedures discussed in the next
3/8
1 yS «β sections are used to convert these geometrical relation­
(h) ships to erosion rates. T h e discussion is divided into
T
* c r v
* '<
erosion prior to reaching a nonerodible layer and ero­
which is Eq. (8.74). Using Fig. 8.22 to define the relationship sion after reaching a nonerodible layer.
between X* and R * and Eq. (8.68) to define the X* -r *
c c

relationship, the functional relationship between X* and c Stage 1: Channel Erosion Prior to Nonerodible Layer
g(X* ) given in Fig. 8.23 was developed
c
Prior to reaching t h e nonerodible layer, the channel is
assumed to e r o d e vertically at a width equal to W and eq

at a potential rate defined by t h e maximum tractive


T o determine the normalized p a r a m e t e r s R* and force (see Fig. 8.24A) or
, X* can be determined for given values of Q, 5 ,
c

y, and τ from Fig. 8.23 and Eq. (8.74). It should be ETC = D WTC eq = K (l35r -r )W
T a c eqi (8.79)
noted that X* is undefined for g(X* )
c values less c

than 1.8. This corresponds to a flow condition with w h e r e E is the potential rate of vertical erosion, K
rc T

insufficient shear force to cause erosion and channel is the soil erodibility, and r is given by yRS [Eq. a

incision. Also, over certain ranges of g(X* \ X* is c c (8.69)]. Using Manning's equation for R (Eq. (d), Ex­
double valued. T h e conservative p r o c e d u r e for estimat­ ample Problem 8.11)
ing erosion is to select the lower values.
Given X* , values for the normalized p a r a m e t e r s 3/8
c
nQR,
R* and are determined from Fig. 8.22. Relation­ r = yS
a
(8.80)
ships are then needed to convert these normalized yfS
288 8. Erosion and sediment yield
bottom a n d channel wall. T h e erosion rate is given by
dW
r ),
c (8.84)

w h e r e d is t h e d e p t h to nonerodible layer.
ne

T h e final channel width, W , will b e reached when {

X*c is such that t h e shear on the channel wall is


everywhere less than r . T h e value of A"* correspond­
c

ing to this condition will b e d e n o t e d as


Nonerodible Layer
(8.85)
Β * cf
WP'
where y is t h e depth of flow when W = W . By assum­
{ {

ing a rectangular channel shape, it can b e shown that


the final width can b e given by

= 1- 2X *cf (8.86)
WP
7 and
^^^^ ^y ^ 1
1

(8.87)
WP k
* Cf
Nonerodible Layer

Figure 8.24 Equilibrium geometry for ( A ) stage 1 and ( B ) stage 2 Using Eq. (8.87) in (8.86) yields
channel erosion.
(8.88)

and substituting E q . (8.88) into Eq. (8.74) for R*, o n e


can obtain
T h e maximum rate of downward movement will b e
defined by 3/8
1 nQ
r* c f [** c f (l -2*, c f )] 3 / 8
r c

(8.81) (8.89)

Equation (8.89) must b e solved implicitly for X* . cf

where p is soil bulk density.


b T a b u l a t e d values for g(X* ) versus X* are given in
ci cf

Actual d e t a c h m e n t will b e limited by the ratio of Appendix 8F, Table 8 F . 1 . Solution procedures are il­
sediment load to transport capacity, as given in E q . lustrated in Example Problem 8.12. A t first glance, it
(8.5), or may seem that Fig. 8.23 could b e used to solve Eq.
(8.89). Such is not t h e case. Figure 8.23 a n d t h e g(X* ) c

relationship in A p p e n d i x 8 F a r e based o n Eq. (8.74)


Μ = (8.82)
and cannot b e used to solve Eq. (8.89) as t h e two
equations a r e different.
F r o m E x a m p l e P r o b l e m 8.11, E q . (d), R =
Stage 2: Channel Erosion Alter Reaching a Nonerodible [nQ/S ] R / .
l/2 3/8 3 8
Using t h e known value for X* c{

Layer After reaching a nonerodible layer, t h e channel from Eq. (8.89), t h e final hydraulic radius is
is assumed to expand laterally at a rate defined by 3/8
shear at t h e intersection of the erodible channel wall nQ
R = (8.90)
and the nonerodible bed, or X* = y / W P (See Fig. b

8.24B). T h u s
Using Eq. (8.87), Eq. (8.90), a n d t h e fact that R* =
dW y ( r b - r c ) Λ / W P , t h e final width is given by
(8.83)
-,3/8
nQ (l-2*„ c f )
(8.91)
where r b is shear at t h e intersection of the channel yfS X ^
Calculating Concentrated Channel Flow Erosion 289
In the transition between initial width and final width, and L a n e model, regression equations known as the
Foster (1982) defined a dimensionless time and width E p h e m e r a l Gully model were developed for W , W , eq f

from a n d τ (Watson et al., 1986), or


t(dW/dt) h

(8.92) W = 2.66(^ 3 9 6
)(,I - 0 3 8 7
)(S- 0 1 6
)(T -°- 2 4
) (8.97)
W -
t W-m
eq c

and
and

W = (8.93)
W { = 179(^ 52)
5
( n 0.556 ) ( 5 0.119^ -0.476^ T ( g 9 g )

where (dW/dt) is the initial value for dW/dt and W


in in

is the initial channel width u p o n reaching a nonerodi­ w h e r e q is p e a k discharge ( m / s e c ) , η is Manning's


p
3

ble layer. For steady-state flow, W would be equal to m


roughness, 5 is slope ( m / m ) , and r is critical tractive c

W , the equilibrium width prior to reaching t h e


&q
force (Pa). W h e n estimating the erosion rate, the maxi­
nonerodible layer. m u m shear force on the channel bottom (1.35yftS) is
T h e rate of change of channel width in the Foster estimated by
and Lane model was expressed by
τ = 4867(^· 3 7 5
)(η°· 3 7 5
)(5 · 0 8 1 1
). (8.99)
dWi/dt* =
= eΡ~ *
1
(8.94)

or In application of the E p h e m e r a l Gully model,


(8.95) W a t s o n et al. (1986) r e c o m m e n d that p e a k discharge
b e used to predict channel equilibrium width and rate
Knowing W*, the actual width at any time is deter­ of growth of t h e gully. In using the E p h e m e r a l Gully
mined from model, it should b e r e m e m b e r e d that t h e regression
W=W*{W -W ) + W- . (8.96) equations w e r e developed from a d a t a base generated
i in m

by the Foster a n d L a n e model and that Fig. 8.23 or


T h e use of the Foster and L a n e model is illustrated in Table 8F.1 in A p p e n d i x 8 F would indicate an u n d e ­
Example Problem 8.12. fined value for X* w h e n g(X* ) c is less than 1.8. As c

stated earlier, this would correspond to insufficient


shear strength to cause channel erosion and an incised
The Ephemeral Gully Erosion Model channel.
T h e complexity of the Foster and L a n e model dis­ A comparison b e t w e e n t h e Foster and Lane predic­
cussed above led to a desire for an explicit prediction tions a n d the E p h e m e r a l Gully model predictions is
equation. Using a data base g e n e r a t e d from the Foster given in Table 8.15 for selected input p a r a m e t e r s .

Table 8.15 Comparison between Predictions from the Foster and Lane Concentrated Flow Model and the
Ephemeral Gully Model.

Equilibrium width Final width Maximum shear


Critical
tractive
Flow rate Slope Manning's force FLM EGM FLM EGM EGM
(m /sec)
3
(m/m) η x (Pa)
c g(X* )
c Eq.(8.77) Eq. (8.97) Eq. (8.91) Eq. (8.98) FLM* Eq. (8.99)

0.1950 0.02 0.03 1.35 43.98 0.734 0.624 4.186 5.623 39.45 29.65
0.1500 0.02 0.03 2.00 26.91 0.648 0.512 3.068 4.035 35.88 26.88
0.0050 0.02 0.03 2.00 7.52 0.170 0.133 0.434 0.617 10.09 7.51
0.0010 0.02 0.03 2.00 4.11 0.085 0.070 0.164 0.254 5.53 4.10
0.0005 0.02 0.03 2.00 3.17 0.062 0.053 0.106 0.173 4.27 3.17
0.0005 0.01 0.03 2.00 1.80 0.045 0.060 0.062 0.159 2.21 1.80

fl
FLM, Foster-Lane model; EGM, ephemeral gully model.
M . 3 5 y R S [see Eqs. (8.68) and (8.69)].
290 8. Erosion and sediment yield
0.140
Example Problem 8.12 Computing concentrated
W (0.735)(1300)
eqPh
flow erosion by Foster and Lane Model
= 1.46 X 1 0 " m/sec or 0.527 m / h r .
4

A construction area has been regraded to include a 20-m


long drainage channel on a slope of 2%. In channel construc­ 7. Time to reach nonerodible layer (d ne = 0.5 m).
tion, the soil is graded to a depth of 0.5 m over a rock layer.
d 0.5 m
Prior to stabilizing the channel, a storm occurs with a peak np

= 0.95 = 1 hr.
discharge of 0.195 m / s e c . If the soil in the unstabilized state
3
M rc 0.527 m / h r
has an erodibility of 0.005 sec/m, a critical tractive force of
1.35 Pa, and a bulk specific gravity of 1.3, estimate the From geometry for a rectangular channel, the depth of flow y
erosion potential per unit length of channel and the total is ^(WP - W ). Thus, when erosion has reached the
eq

erosion for the entire channel. The storm has a duration of nonerodible layer, the following calculations can be made for
6 hr and a runoff volume of 1728 m . 3
y, the dimensionless distance X* corresponding to y, r*
corresponding to y, and the actual shear at y, r , by using b

Solution: Eqs. (8.68), (8.69), and (8.70),


1. Conveyance factors [Eq. (8.74)]. Assume a smooth soil
WP - W 0.988 - 0.735
η = 0.03 using the values in Appendix 8E, Table 8E.2. A pr

= 0.1265 m
frequently used factor in the Foster and Lane equations is
0.1265
X *. — = 0.128
J i 2 S £ 2 > R . o.303. WP 0.988
/0.02 •
τ* =1.35[l-(l-2**) 2 9
]
Using γ = 9803 N / m , the conveyance factor given by Eq.
3

(8.74) is = 1.35(1 - [1 - ( 2 ) ( 0 . 1 2 8 ) ] ) = 0.777. 29

/ nQ \ 3 / 8
yS (9803
(9803)(0.02) From earlier calculations, r = 29.2 Pa, which does not a

= 44.0. change until the nonerodible layer is reached, at which point


,35 the flow depth will change. When the nonerodible layer is
reached, the nondimensional shear corresponding to y would
For g(X* ) > 35, use 35.
c
be τ * = r / r , hence
2. Determining X* , W* and R*. From Fig. 8.23 or
b a

c f

Appendix 8F, Table F.l, T = r * r = (0.777)(29.2) = 22.69 Pa.


b a

g(X* ) c = 35.0; X+ = 0.0; W* = 0.744; R* = 0.151.


c
From Eq. (8.83),
3. Wetted perimeter [Eq. (8.76)]. ldW\ r (0.005)(22.69 - 1.35)
c

3/8 (1300)
nQ P b

WP = R~ 5/8
= (0.303)(0.151)" 5/8
= 0.988 m.
= 8.21 x 1 0 " m/sec or 29.55 c m / h r . 5

4. Equilibrium width [Eq. (8.77)]. 8. Initial erosion rate after reaching nonerodible layer [Eq.
(8.84)].
W cq = WP W* = (0.988)(0.744) = 0.735 m.
tdW\
5. Equilibrium hydraulic radius [Eq. (8.78)].

R = fl^WP = (0.151)(0.988) = 0.149 m. = (1300 k g / m ) ( 8 . 2 1 Χ 1 0 " m/sec)(0.5 m)


3 5

6. Stage 1 erosion rate and maximum downward movement. = 0.053 k g / m · sec.


Using Eq. (8.79M8.81) with p of 1300 k g / m 3

9. Final width [Eqs. (8.89) (8.90), and (8.91)]. Calculating


b
y

/Q n \3/8 X* cf>

τ = ysi y = - I R/ 3 8
= (9803)(0.02)(0.303)(0.151) 3/8
3/8
ι
α
nQ
S(**cf) = 3/8
= 29.2 Pa
£ r c = A: (1.35r -r )^
r a c e q
This must be solved by iteration for X* . From item 1 c{

above, the RHS of the equation is 44.0. Thus, a value for


= (0.005)[(1.35)(29.2) - 1.35](0.735)
X* that makes g(X* ) equal to 44.0 must be found. Since
cf c{

= 0.140 kg/sec · m g(X* ) is solved analytically for X* , no cutoff of 35.0 is


c{ cf
Calculating Concentrated Channel Flow Erosion 291

used, as in the case with g(X*). For a first trial, let X*cf
Time after Time after Hourly
0.05; then from Eqs. (8.68) and (8.89),
start of reaching average
storm <*ne t
T = 0.355 and g(X* f) = 9 0
< 44
·°· (hr)
C
(hr) (sees) (kg/m sec) (kg/m»sec)
# c f e

Other trials are given below.


0.0 — — 0.140
0.5 — — 0.140 0.140
**cf g(X* ) cf
1.0 0.0 0 0.053

0.05 9.0 1.5 0.5 1800 0.051 0.051

0.02 29.21 2.0 1.0 3600 0.049

0.015 42.80 2.5 1.5 5400 0.047 0.047

0.014 46.94 3.0 2.0 7200 0.045

0.0147 43.98 OK 0.043


4.0 3.0 10,800 0.041
0.040
Using X* cf = 0.0147 in Eq. (8.91),
5.0 4.0 14,400 0.038
3/8 3/8
l-2*. " 0.037
c f
\ n Q
]
6.0 5.0 18,000 0.35
. ( * . c f ) V
\
0.358
3/8
1 - (2)(0.0147)
= 0.303 = 4.19 m.
(0.0147) 5/3
12. Total detachment potential (length = 20 m). Total ero­
sion is given by
10. Dimensionless time and width [Eqs. (8.92) and (8.93)].
£tot= LErcM;X20m.
t(dW/dt) (0(8.21 Χ 10" ) in
5

t = 1—25. = — = 2.376 X K T * 5

4.19 - 0.735 Using Δί of 1 hr,


w -w t x
W - 0.735 E = (0.358 kg/sec · m ) ( l hr)(3600 sec/hr)(20 m)
w-w, _=
tot

WL = n
0 . 2 8 9 ( ^ - 0.735).
= 25,776 kg.
W - W f x 4.19 - 0.735
11. Potential detachment rates after reaching nonerodible 13. Converting to sediment concentration. Conversion of
layer [Eq. (8.84)]. From the chain rule of differentiation, Eq. total erosion to an average concentration can be made by
(8.94) and the definitions given in item 10, noting that the average sediment concentration is simply the
mass of sediment divided by the runoff mass, which is runoff
dW dW dW* dt* _ tdW\ volume times density of water. In this case, runoff mass is
1728 m X 1000 k g / m ; hence the average concentration is
3 3

From item 10 above, t* = 2.376 X 1 0 " t, and from item 7 5 25,776 kg


C =
(dW/dt) = 8.21 X 1 0 " m/sec; hence
in
5
1728 m X 1000 k g / m
3 3

dW = 0.0149 kg/kg or 14,901 mg/liter.

= 8.21 X io-V 2376X10


' m/sec. These values represent what could happen if transport capac­
dt ity is not limiting. Corrections for the effects of transport
Finally, using Eq. (8.84), capacity are done in a following section.
dW , dW
E = —d rc Ph nc = (1300 k g / m ) — ( 0 . 5 m) 3

dW
= 6 5 0 — - k g / m · sec The DYRT Model for Concentrated Flow Erosion
dt
A recent modification of t h e Foster a n d L a n e (1983)
or model was developed by Storm et al. (1990). Known as
D Y R T , this model allows for variable flow rates and
E - (650)(8.21 X i - ) e - 5 2 3 7 x 1 0
" '
5

can b e modified for layers of varying density and trac­


rc 0

= 0.0533<?- ' 2 376x10


" ' k g / m · sec.
5
tive force. T o account for variable flow rates, proce­
d u r e s were developed to account for flow changes.
Values for E rc are tabulated below. C o m p u t a t i o n s of channel width a r e m a d e numerically
292 8. Erosion and sediment yield

rather than using the exponential form of Eq. (8.95), or and

dW M rrt = A*rc, ot(l ~ ψ)


P = (1.46 Χ Ι Ο " m / s e c ) ( 1 - 0.7)4

Wt+* = <+~^ W Δ
'' ( 8 1 0
°)
= 4.38 X 1 0 - 5
m/sec.
where W a r e channel widths at time t a n d
n The time to reach the nonerodible layer (d nc = 0.5 m) is thus
t + Δ ί , a n d Δ Ms t h e solution time interval. 0.5 m
T h e use of Eq. (8.100) with t h e Foster a n d L a n e t. =
4.38 x 10~ m/sec
n 5
'react
(1983) model for equilibrium channel width allows for
= 11,415 sec or 3.17 hr.
computation of rill erosion u n d e r varying flow rates.
3. Correcting the rate of widening for actual erosion. The
Procedures for t h e computation are given in Storm
equilibrium and final widths would not be affected by sedi­
et al (1990).
ment load, but the rate of change would be. Hence from
items 4 and 9 in Example Problem 8.12, W = 4.19 m and {

Potential versus Actual Channel Erosion W = 0.735 m. Also, from items 7 and 8, the potential rate
cq

of widening and erosion rate after reaching the nonerodible


T h e concentrated flow models presented in this sec­ layers are (dW/dt) = 8.21 X 10~ m/sec and £
inpot = 5
r c > p o t

tion predict detachment potential in a channel. Poten­ 0.053 k g / m · sec. Actual rates would be [using an analogy to
tial d e t a c h m e n t can b e translated to actual d e t a c h m e n t Eq. (8.5)]
by using a modification of Eq. (8.5) or
-'react C.pOt A
rp

D = r D (\-qJT ),
xc c (8.5) \ CJ 1

= 0.053(1 - 0.7)
where D is detachment potential a n d D is actual
r c r = 0.016 kg/sec · m
detachment. This can be combined with t h e continuity and
equation [Eq. (8.1)] t o predict sediment load at any
point on a channel. T h e procedure is illustrated in
I dt / , \ dt /in,pot\ TJ
Example Problem 8.13. in act C

= (8.21 Χ Π Γ m / s e c ) ( l - 0.7) 5

= 2.46 X 1 0 " m/sec. 5

Example Problem 8.13 Effects of transport capacity The actual erosion rate would be given by
on detachment
tdW\
Calculate the total channel erosion in Example Problem
8.12, assuming that the channel drains a regraded area and ( dW\
that the ratio of sediment load to transport capacity, q /T , is s c = (1300g/m )| — J 3
(0.5 m) kg/:m · sec.
0.7 in the runoff from the regraded area.
The dimensionless time, t* is corrected for actual erosion
Solution: rate by
1. Converting potential to actual detachment. From Eq.
(8.5), t(dW/dt) intact 2.46 Χ 1 0 " 5

t* =
W - W { x 4.19 - 0.735
D = r D (l-Q /T ).
rc & c

= 7.12 Χ 1 0 " / . 6

D represents actual and D potential erosion as calculated


t rc From item 11 of Example Problem 8.12,
in Example Problem 8.12.
dW\ ldW\
2. Correcting initial erosion rates and time to reach
nonerodible layer. Corrections are needed for the rate of , dt /act I dt / i n a

actual erosion to determine the time to reach the nonerodi­ Using the value calculated above for (dW/dt) in act

ble layer. From item 6 in Example Problem 8.12, E = rcpoi

0.140 k g / m · sec and A f = 1.46 X 10~ m/sec. Correct­ 4

( dt =2.46x 10-V "\


7 1 2 X 1 0
rcpot

ing for the effects of transport capacity, an analogy can be /act


made to equation 8.5, or Hence
Rc.act = (1300 k g / m ) ( 0 . 5 m)(2.46
3
X 1 0 ' m/sec)e"
5 7 1 2 x l 0 _ 6 /

χ
-'rc.act ^rc.pot | l - ^ - j = (0.140kg/m · sec)(l - 0.7)
= 0.016iT 7 1 2 x l 0
~ ' k g / m · sec.
5

= 0.042 k g / m · sec Values are tabulated below.


Estimating Sediment Yield 293

of deposition occurring, a sediment delivery ratio has


b e e n defined as
Example Problem 8.12
Y
4. Corrected erosion rates. D = : — , (8.101)
gross erosion X watershed area
Time after Time after Hourly
w h e r e Y is the sediment yield from a watershed and
start of reaching average
gross erosion would be the erosion per unit a r e a occur­
storm t
(hr) (hr) (sec) (kg/m«sec) (kg/m sec)
e ring on t h e watershed. Gross erosion is composed of
rill and interrill erosion, gully erosion, and stream
0 0.00 0 erosion. O n disturbed areas, sheet and rill erosion are
1.00 0.00 0 0.0420° 0.0420 the principal c o m p o n e n t s of gross erosion. In the previ­
2.00 0.00 0 0.0420° 0.0420 ous section, p r o c e d u r e s were presented for predicting
3.00 0.00 0 0.0420° 0.0420 rill and interrill erosion as well as e p h e m e r a l channel
3.16 0.00 0 0.0420° erosion. Procedures a r e now n e e d e d for determining
the delivery ratio in o r d e r to obtain sediment yield.
3.17 0.00 0 0.0160
In this section, empirical delivery ratio methods are
4.00 0.83 2988 0.0157 0.0225
discussed. It should b e pointed out that the degree of
5.00 1.83 6588 0.0152 0.0155
understanding of sediment delivery ratios is probably
6.00 2.83 10,188 0.0149 0.0151
less t h a n any o t h e r a r e a of sedimentation.
^-^rc.act =
0.1791 Application of a delivery ratio to estimated erosion
should b e d o n e with careful consideration of the man­
°Values determined by previous computation for erosion prior to
n e r in which t h e model considers deposition. For ex­
reaching the nonerodible layer. The last value at 3.16 hr is just prior to
reaching the nonerodible layer. ample, t h e U S L E / R U S L E does not include proce­
d u r e s for evaluating deposition, whereas t h e C R E A M S
Total detachment is a n d W E P P models (discussed in a subsequent section)
E toX = lOm^E^Ati consider deposition on hillslopes and in e p h e m e r a l
gullies. T h u s , t h e delivery ratio for the U S L E / R U S L E
= (20 m)(0.1791 k g / m · sec)(l hr)(3600 s e c / h r )
would n e e d to evaluate deposition that occurs in over­
= 12,895 kg.
land flow prior to reaching concentrated flow channels.
The ratio of actual to potential erosion, using the estimated
potential erosion of 25,776 kg in Example Problem 8.12, is Graphical Methods for Predicting Delivery Ratios
12,895/25,766 = 0.50. Thus the effect of having q /T equal
A n u m b e r of methodologies have b e e n proposed to
s c

to 0.7 was to reduce the erosion by 50%. Obviously, the effect


is not linear. predict t h e sediment delivery ratio. T h e s e include sim­
5. Calculating concentration. To calculate concentration, ple e s t i m a t e s by an areal relationship a n d a
the sediment load entering the ditch must be combined with relief-length ratio. Also, t h e accounting of many on-site
the erosion rate to get total sediment. Since only the ratio of factors such as w a t e r available for overland flow; tex­
qJT was given in the problem, sediment concentration at
c
t u r e of e r o d e d material; ground cover; slope shape,
the end of the channel cannot be calculated. gradient, and length; surface roughness; and additional
site-specific factors have b e e n r e c o m m e n d e d by the
Forest Service (1980). Additionally, delivery ratio con­
cepts based o n storm modeling techniques have b e e n
advanced by Williams (1977).
ESTIMATING SEDIMENT YIELD
Area-Delivery Ratio Relationship A first approxima­
Erosion - Sediment Delivery Ratio Method tion of t h e delivery ratio may be obtained by the area
effects illustrated in Fig. 8.25. Erosion estimates used
The Concept of a Sediment Delivery Ratio in developing delivery ratio curves were based on the
T h e soil loss equations described earlier are useful simple U S L E . Therefore the curves should only be
tools for predicting the a m o u n t of soil loss from a field, used with erosion estimates from the U S L E or R U S L E .
referred to as gross erosion. Some of the models, such Considerable scatter is present in the data; however,
as R U S L E , assume no deposition. Between the field t h e general t r e n d indicates a strong effect of area on
and point of final deposition, sediment will normally delivery ratio. Differences exist in the form of curves
have numerous opportunities to b e deposited, reducing for different areas. Therefore, it is desirable, where
t h e sediment yield accordingly. T o quantify the a m o u n t possible, to obtain a curve specific for a given area. As
294 8. Erosion and Sediment Yield

cο
3
ο
co 1.00
*4

Ο
3
cd
< 0.10

Red Hill* Physiographic Area-Texas k Oklahoma


Missouri Bosh Loess HNIs - Iowa k Nebraska
15
Q
Btackiand Palrles - Texas
Sond-AAY Hllls-Mlssiseippi
Piedmont Physiographic Area-
North Carolina, South Carolina k Georgia
.01
I I I I Mill I I I Mill T T I I I ΙΙΜΓ

A .01 .1 1.0 10 100

D r a i n a g e Area ( S q u a r e Miles)
Figure 8.25 Sediment delivery ratio versus drainage area size for use with
U S L E / R U S L E (after Boyce, 1975).

expected, a method as simple as this yields only a


rough approximation and has limited applicability in Ξ
determining the expected sediment impact of alterna­
tive surface treatment techniques. In particular, the —

method is not recommended for single storms.

The Effects of Channelization on Delivery Ratio The Ξ


degree of channelization affects how efficiently e r o d e d
Dertivery Ratio Curve Rolling —
sediment can b e transported through a watershed —
Red Plains Lane I Resource Area
channel system. A well-channelized watershed will
transport most eroded material out of the watershed, w
.01
whereas a poorly channelized watershed will transport 0 .01 .02 .03 .04
the sediment slowly, leaving many opportunities for
deposition. O n e measure of channelization is known as Relief-Length (R/L) Ratio
the relief-length ratio, calculated as Figure 8.26 Effect of relief-length ratio on delivery ratio for use
with the U S L E / R U S L E (from Renfro, 1975).

elevation difference between watershed divide


at the main stem and the watershed outlet
length of flow path along the main stem material, (5) surface runoff, (6) slope gradient, and
(7) surface roughness. T h e p a r a m e t e r s are shown on
A n example of the effects of relief-length ratio on the Fig. 8.27 and discussed below.
sediment delivery ratio is shown in Fig. 8.26 for the
Red Hills area of Oklahoma and Texas (Renfro, 1975). Runoff Factor T h e surface runoff factor, given by
It should be applied with caution to other areas, al­ p e a k discharge in cfs/ft, defines t h e quantity of water
though the shape of the curve should be similar for available to transport sediment for a storm. In the
most areas. absence of p e a k discharge information, t h e magnitude
of the runoff factor can b e estimated by
Forest Service Sediment Delivery Index Model
T h e Forest Service (1980) developed a methodology F = 2.31 Χ 1 0 " ( 7 L ,
5
(8.102)
for predicting sediment delivery ratio that can be used
for a single storm. T h e method uses a stiff diagram to where F is runoff rate p e r foot of slope width (cfs/ft),
predict delivery ratio as a function of (1) delivery σ is rainfall excess ( i n . / h r ) , and L is the length of the
distance from the slope to stream, (2) slope shape, disturbed area (ft). Values greater than 0.1 are as­
(3) percentage of ground cover, (4) texture of eroded sumed to b e 0.1.
Estimating Sediment Yield 295

T e x t u r e of
Percent Eroded Surface
Ground Material Runoff
Cover (% S i l t k S m a l l e r ) (cfs/ft)
100

-75

-50

X * - .25
v °
1 2 3 Tfc
Slope ι ι ι ι ι ι ι 3
Shape
.3

.2

.1

Delivery Surface Slope


Distance Roughness Gradient
(feet)
Figure 8.27 Stiff diagram for estimating sediment delivery with the
U S L E / R U S L E (after Forest Service, 1980).

Texture of Eroded Sediment Texture of e r o d e d sedi­ Surface Roughness Surface roughness is a subjective
m e n t is a p a r a m e t e r used t o define t h e impact of index of t h e impact of roughness on sediment delivery.
particle size on delivery. T h e p a r a m e t e r is t h e percent­ A value of 0 indicates a smooth surface a n d 4 is a very
age finer than 0.05 m m (silt size a n d finer). A value of rough surface.
100%, for example, m e a n s that all particles a r e silt size
and smaller.
Use of the Forest Service Model T h e forest service
m e t h o d should only be used with t h e U S L E / R U S L E
Ground Cover Factor T h e ground cover factor refers
estimates of gross erosion. T o use t h e m e t h o d , all of
to the percentage ground cover in t h e flow p a t h b e ­
t h e p a r a m e t e r s must b e estimated a n d plotted on t h e
tween the source area and t h e stream. G r o u n d cover,
a p p r o p r i a t e axis in Fig. 8.27. All of t h e points are then
in this case, is defined as cover such as litter that is in
connected, forming a polygon. T h e ratio of the area
contact with t h e surface. Z e r o indicates n o cover a n d
within t h e polygon to t h e total a r e a of t h e rectangle is
100 m e a n s complete cover.
used in Fig. 8.28 to predict t h e sediment delivery index
(delivery ratio). P r o c e d u r e s a r e illustrated in Example
Slope Shape Factor Slope s h a p e is t h e factor that
P r o b l e m 8.14.
accounts for t h e impact of concave or convex surfaces
on a sediment delivery. T h e s h a p e refers t o slope
shape between t h e source area and channel. A factor
of zero represents a convex shape a n d four a concave
shape.
Example Problem 8.14 Predicting delivery ratio
with Forest Service method

Slope Gradient Slope gradient is the average slope


A construction operation is disturbing a large area in a
between t h e source area and t h e receiving channel. forest watershed upslope of a sensitive stream. The regula­
tory authority requires that a riparian forest zone be left
Delivery Distance Delivery distance is t h e flow dis­ undisturbed for 200 ft on either side of the stream. During a
tance in feet between the source area and channel. storm, the sediment yield from the disturbed area to the edge
296 8. Erosion and Sediment Yield
of the forest is estimated from the RUSLE to be 30 tons/acre
with a size distribution that includes approximately 50% silt
plus clay. The disturbed area has a slope length of approxi­
mately 720 ft. The peak rainfall excess from the site is 4.50
in./hr. The riparian zone between the disturbed area and
the stream has the following characteristics:
Slope shape, moderately convex
Change in elevation, 50.0 ft
Average slope, 25%
Ground cover, 100%
Surface roughness, rough
Flow distance, 200 ft.
Estimate the sediment yield to the stream, both with and
without the riparian zone, if the disturbed area extends 1000
ft along both sides of the stream. If the runoff volume is 3.0
in., estimate the average sediment concentration of the flow
entering the stream. 0 10 20 30 40 50 60 70 80 90 100
Solution: Percent Area From Stiff Diagram
1. Area of polygon on stiff diagram. From the data above,
Figure 8.28 Diagram to convert percentage of area in Fig. 8.27 to
the texture index is percentage silt plus clay or 50; the slope
sediment delivery ratio (after Forest Service, 1980).
gradient is 25.0; surface roughness value is 3.0; delivery
distance is 200 ft, slope shape is 1, and percentage ground
cover is 100. Peak flow rate is Using these values in Fig. 8.27, a polygon is plotted as shown
in Fig. 8.29 and the relative area determined to be 0.10 or
10%.
q = 2.31 X 1 0 - V L = 2.31 X 10~ (4.50)(720)
5
p
2. Delivery ratio. Using 10% in Fig. 8.28,
= 0.075 cfs/ft. D = 0.02.

T e x t u r e of
Percent Eroded Surface
Ground Material Runoff
Cover (% Silt * S m a l l e r ) (cfs/ft)
100

.75

Χ «Ρ -
1 2 3x > \
Slope
Shape 1 L 1 J_ f 1 J 3i

.2

'of . 1

Delivery Surface Slope


Distance Roughness Gradient
(feet)
Figure 8.29 Application of Forest Service Stiff diagram in Example
Problem 8.14.
Estimating Sediment Yield 297
3. Sediment yield. period of observation. Otherwise, the design capacity
may be far too small or too large. T o select a compari­
Y = D X (gross erosion)
son reservoir, it is important to select a site that has

( 1 acre \ similar characteristics insofar as they affect sediment


yield. T h e m o r e important characteristics to compare
y = 9.91 tons, are relief ratio, area, vegetative cover, and fraction of
area disturbed.
4. Estimating concentration. The runoff volume is A problem often e n c o u n t e r e d with the reservoir
Q = runoff depth X area survey m e t h o d for small watersheds is that many sedi­
m e n t a t i o n surveys d o not adequately describe the char­
= ( £ f t ) ( 7 2 0 x 1000 ft )
2

acteristics of t h e drainage area. T h e sediment yield


= 180,000 ft = 4.13 acre · ft.
3 expected from a highly disturbed watershed cannot b e
adequately extrapolated from t h e yield measured dur­
Sediment concentration is the mass of sediment divided by
ing an existing nondisturbed land use. Simply adjusting
the mass of water or
factors in the U S L E / R U S L E , such as K, C, and Λ
(9.91 tons)(2000 lb/ton) will not be a d e q u a t e since the entire watershed
C = drainage p a t t e r n a n d flow characteristics of individual
(62.4 ib/ft )(180,000ft )
3 3

streams have b e e n drastically modified through in­


= 0.00176 lb/lb or 1760 mg/liter. creased p e a k flows and varied sediment loads as a
result of disturbance. T h u s , caution should be exer­
This is the concentration entering the stream.
5. Sediment yield without the riparian zone. Without the cised in the application of this m e t h o d .
riparian zone, the sediment yield would be

(30 tons/acre)(720 ft)(1000 ft) Estimating Sediment Yield with Modified Universal
γ = = 496 tons.
43,560 ft / a c r e
2
Soil Loss Equation (MUSLE)
The average concentration would be Research in recent years has g e n e r a t e d new models
of soil erosion and sediment yield with varying degrees
(496 tons)(2000 lb/ton) of applicability. Some of these models are application
C =
(62.4 lb/ft )(180,000ft )
3 3
oriented, while o t h e r s have not reached t h e point of
application. Some of t h e models have b e e n combined
= 0.088 lb/lb or 88,000 mg/liter.
and computerized along with hydrology models and
Obviously, riparian zones have a major impact on sediment reservoir models to yield a complete computational
yield. package of watershed hydrology and sedimentology.
A detailed discussion of all t h e models is beyond the
scope of this chapter. O n e of the less complex relation­
ships is the Modified U S L E .
Reservoir-Survey Method of Estimating Williams (1976) p r o p o s e d t h a t the rainfall energy
Sediment Yield term, EI 30 index, in t h e U S L E could be replaced with
O n e of the most common m e t h o d s of estimating the a runoff energy t e r m in t h e U S L E to predict sediment
volumes of sediment produced is by the Reservoir- yield directly. P r o c e d u r e s were developed for homoge­
Survey method. T h e Soil Conservation Service main­ neous watersheds using a lumped p a r a m e t e r approach
tains a program in which thousands of reservoirs are and for n o n h o m o g e n e o u s watersheds using sediment
surveyed annually to d e t e r m i n e the quantity of sedi­ routing procedures. A l u m p e d p a r a m e t e r approach is
ment deposited. Information is also obtained, if avail­ o n e in which t h e entire watershed is represented by
able, on watershed characteristics. W h e n a new reser­ o n e characteristic p a r a m e t e r .
voir site is proposed in an area, the SCS geologist
simply finds the record of the closest surveyed reservoir MUSLE Lumped Parameter
that has similar watershed characteristics. T h e mea­ Williams (1977) and Williams and Brendt (1972)
sured loss of storage capacity in acre-feet p e r year p e r developed the M U S L E using data from 778 storms on
acre watershed for the existing reservoir now becomes watersheds n e a r Reisel, Texas, and Hastings, N e ­
the design sediment storage volume. braska. T h e drainage areas ranged from 2.7 to 4380
Caution should be exercised in selecting record years acres and t h e average slope and slope lengths ranged
for which the survey was m a d e to assure that the from 0.9 to 5.9% and 258 to 570 ft. H e replaced the R
rainfall over the watershed was n e a r normal during the factor in the U S L E with various p a r a m e t e r s and used
298 8. Erosion and Sediment Yield

the resulting equation to predict the sediment yield (4) Calculate t h e a m o u n t of sediment from each
from the watersheds. T h e p a r a m e t e r that gave the best subwatershed that reaches t h e exit by assuming that
estimate was Qq , the product of runoff volume and
p t h e r a t e of deposition is proportional to t h e particle
peak discharge. Williams d e n o t e d this t e r m runoff en­ size, a m o u n t of sediment, a n d travel time, or
ergy. T h e resulting equation is
dY,
(8.105)
Υ = 9 5 ( β X q f {K}J,LS}J,CP}^
56
(8.103) dt
p

w h e r e Y is t h e yield at any point downstream of the


{

w h e r e Y is single-storm sediment yield in tons, Q is subwatershed exit, β is a routing coefficient, and D 50

runoff volume in acre-feet, q is peak flow in cfs, a n d


p is t h e average d i a m e t e r of sediment. E q u a t i o n (8.105)
{K} , {L5} , and {CP\ are area weighted average
a a can b e integrated to obtain
U S L E / R U S L E p a r a m e t e r s for the watershed. Obvi­
ously, from the definition of delivery ratio, t h e single- -BTti
(8.106)
storm delivery ratio is
w h e r e Y is the yield at the subwatershed exit, Y is
0i {

0.56
95(Q X qj t h e sediment from w a t e r s h e d i that reaches t h e main
p

D (8.104) w a t e r s h e d exit, a n d T is t h e travel time from t h e


tl
R X area
subwatershed to the main w a t e r s h e d exit.
(5) Calculate t h e total sediment yield for t h e entire
w h e r e area is the watershed area in acres, R is t h e
w a t e r s h e d by
single-storm value of the rainfall factor in English
units, and Q a n d q are acre-feet and cfs.
p

(8.107)
MUSLE Routing Procedures i= l

T h e use of the M U S L E as given by Eq. (8.103) (6) D e t e r m i n e t h e routing coefficient from


should b e limited to those watersheds that a r e rela­
tively homogeneous. W a t e r s h e d s that are in t h e pro­
cess of being disturbed are typically very heteroge­ (Q X < 7 £ P
5 6
= Σ (β,X l J - 5 6
e - B T
< ^ ,
1= 1
neous. Williams (1975, 1978) p r o p o s e d t h e following
(8.108)
p r o c e d u r e to account for watershed heterogeneity:

(1) Divide the watershed into η h o m o g e n e o u s wa­ w h e r e (Q X < ? ) is t h e runoff energy t e r m for the
p ws

tersheds as shown in Fig. 8.30 and d e t e r m i n e t h e travel entire watershed.


time T from each subwatershed to the watershed
tl
Williams p r e s e n t e d verification d a t a from five storms
outlet. on a 4380-acre watershed with an average slope of 2 %
(2) D e t e r m i n e the sediment yield Y for each ho­ 0i
in Reisel, Texas, a n d p r o p o s e d t h a t t h e relationship
mogeneous subwatershed using Eq. (8.103). h a d acceptable accuracy.
(3) D e t e r m i n e t h e D , average d i a m e t e r of sedi­
5 0
Williams' routing p r o c e d u r e is simple to use but
ment, exiting each subwatershed. does not predict a time distribution of sediment (sedi-
g r a p h ) o r a size distribution of sediment. N e i t h e r is
flood plain erosion or channel erosion considered.
Williams (1978) p r e s e n t e d a m o r e complex p r o c e d u r e
for routing that considered b o t h channel erosion and
deposition. T h e p r o c e d u r e s have not b e e n verified nor
© widely a d o p t e d .

Size Distribution of Sediment


T o use Eq. (8.108), it will be necessary to have a D 50

value for t h e sediment exiting t h e subwatershed.


Barfield et al. (1979) p r o p o s e d a simple p r o c e d u r e

© based o n t h e assumption that all of t h e larger particles


in t h e flow settle out first in a deposition process.
Based on this assumption, t h e size distribution of ma­
Figure 8.30 Example watershed division for M U S L E routing pro­ terial exiting t h e subwatershed c a n b e estimated from
cedures. t h e size distribution of t h e p e r c e n t a g e finer ( P F ) of
Process-Based Erosion Models: CREAMS Semi-Theoretical Rill and Interrill Model 299
e r o d e d p a r e n t material. T h e s e p r o c e d u r e s a r e dis­ PROCESS-BASED EROSION MODELS: CREAMS
cussed in C h a p t e r 7. SEMITHEORETICAL RILL AND INTERRILL MODEL
Foster et al. (1985) divided e r o d e d soil into clay, silt,
and sand fractions plus large a n d small aggregates. T h e Basic Equations
fraction of these c o m p o n e n t s in runoff is predicted by
Erosion Model
regression equations based o n a limited a m o u n t of
data. T h e i n d e p e n d e n t variables in t h e regression T h e C R E A M S (chemicals runoff erosion in agricul­
equation is t h e fraction of t h e c o m p o n e n t s in t h e tural m a n a g e m e n t systems) interrill a n d rill erosion
p a r e n t material. T h e s e procedures a r e also discussed model (Foster et al., 1980a) is based o n a semitheoret-
in C h a p t e r 7. ical relationship known as t h e F o s t e r - M e y e r - O n s t a d
( F M O ) equation (Foster et al., 1977a). Foster et al.
start with a partition of erosion into rill a n d interrill
PREDICTING THE TIME DISTRIBUTION components
OF SEDIMENT: A SEDIGRAPH G = G + G, r x (8.114)

In t h e sediment yield computational p r o c e d u r e s p r e ­ w h e r e G is t h e cumulative rill erosion ( m a s s / w i d t h ·


x

sented thus far, only t h e total storm sediment yield was time) and G xis c u m u l a t i v e interrill erosion
calculated. A p r o c e d u r e is n e e d e d t o translate this ( m a s s / w i d t h · time) at any distance X down t h e hill-
yield t o a sedigraph, a time distribution of sediment slope. Relationships for predicting rill a n d interrill
yield. T h e simplest p r o c e d u r e is t o m a k e t h e assump­ erosion in t h e F M O a r e based o n d e t a c h m e n t as a
tion that concentration is a power function of water function of surface shear for rill erosion a n d rainfall
discharge o r energy for interrill erosion. Starting with t h e assump­
tion that interrill erosion is proportional t o slope and
c = kq a
(8.109) rainfall intensity a n d that d e t a c h m e n t in rills is propor­
tional t o tractive force t o t h e § power, t h e F M O
w h e r e k a n d a a r e constants. Based o n a value for a, A:
equation fof cumulative erosion at any point X downs­
can b e calculated from runoff a n d sediment yield infor­
lope is given by
mation as discussed below.
Since t h e mass flow rate of sediment is given by G =X K a'(sme) F C P
2
r
e
t r T

(8.110) + XK (V sin e)I C P ,


X t x x (8.115)

then w h e r e K a n d K a r e rill a n d interrill erodibilities, θ is


T x

slope angle, C a n d C a r e rill a n d interrill cover factor,


r x

(8.111) P a n d P a r e rill a n d interrill conservation practice


T x

factors, F a n d I a r e rill a n d interrill erosivities, and


t t
T o obtain t h e total sediment yield, E q . (8.111) must b e
a', b', a n d e' a r e constants. Foster et al. (1977a) pro­
integrated over t h e entire storm, o r
posed t h a t

y= [ kq dt,
Dst a+l
(8.112) F = 1 5 β ^
t
3
(8.116)

h ~~ EI ,3Q
(8.117)
w h e r e Z) , is t h e storm duration. Since A: is a constant,
s
w h e r e Ε is rainfall energy, 7 is t h e maximum 30-min 30

it can b e d e t e r m i n e d from
intensity, Q is runoff volume in watershed feet, a n d q p

is p e a k discharge in volume p e r area p e r time (feet p e r


k = (8.113) second). By equating t h e F M O equation a n d t h e U S L E ,
Foster et al. (1977b) developed t h e following form for
t h e rill a n d interrill erosion c o m p o n e n t s

w h e r e t h e storm is divided into η increments for t h e IEI 30

( 5 + 0.014) K&Pi
purpose of evaluating t h e integral. T h u s given a runoff
hydrograph a n d total sediment yield, a sedigraph can
(8.118)
be calculated. Values for a a r e typically n e a r 0.5 t o 1.0.
O t h e r models for developing sedigraphs have b e e n m-l

presented (Williams, 1979; R e n d o n - H e r r e r o , 1974; D = 37983mK,<r


IC p
1/3
SKCP2
r r r
72.6
R e n a r d a n d Laursen, 1975; Bruce et al., 1975), b u t a r e
much more complex than t h e p r o c e d u r e s above. (8.119)
300 8. Erosion and Sediment Yield
where D and D { are denoted as interrill and rill
rc Specific equations defining these cases are given in the
erosion potential in l b / f t · see, S = sin θ, σ is the
2
ρ C R E A M S manual (Knisel, 1980).
peak flow rate in f t / s e c , V is runoff volume in feet, K
u r T r a n s p o r t capacity calculations with Yalin's equation
and K are rill and interrill erodibility in English units,
x requires a value of shear stress acting on the soil
and m is a slope length exponent. T h e use of the surface. In C R E A M S , shear stress on the soil is esti­
potential instead of actual erosion definition allows t h e m a t e d from
relationship to be used where deposition occurs, as is 0.9
discussed later. Equations (8.118) and (8.119) are the
basic erosion relationships for C R E A M S . In the usuage
T SO =m s s
(8.122)

developed for C R E A M S , the only time Ρ is used is for


where shear stress on the soil in p o u n d s p e r square
contouring. Since separate K, C, and Ρ values for rill
foot, γ is the weight density of water in p o u n d s p e r
and interrill erosion are not available using the U S L E
cubic foot, y is the d e p t h of flow relative to b a r e soil
modeling format, the interrill and rill values are as­ bs

(i.e., the flow d e p t h assuming a b a r e soil), S is the sine


sumed to be the same and equal to the U S L E values.
of t h e slope angle, n is M a n n i n g ' s η for b a r e soil
bs

Erosivity of Storm (assumed to be 0.01), and n is the total Manning's n.


coy

Values for the p a r a m e t e r s are given in Appendix 8E,


Storm erosivity is calculated from rainfall distribu­
Tables 8E.1 and 8E.2. D e p t h of flow relative to bare
tion information. For breakpoint rainfall (rainfall dis­
soil is estimated from Manning's equation
tributed over the duration of the storm) energy is
0.6
calculated from Eq. (8.31). If the rainfall distribution is <?"bs
not available, the storm EI index is estimated from (8.123)
30
1.55 / 1 2

daily rainfall, or
w h e r e q is water flow rate in cfs/ft.
EI 30
= 8.0Κ^· , 51
(8.120)
100 Application of the Model
where K is the 24-hr rainfall in inches.
R
T h e equations used in C R E A M S are steady state.
Slope Exponent m T o use t h e m on a storm basis, Foster et al. (1980a)
r e c o m m e n d t h a t calculations b e m a d e with p e a k dis­
T o limit excessive erosion estimates on long steep
charge as a flow r a t e . This should b e used to estimate a
slopes, the slope exponent is defined as
sediment concentration. T h e concentration is then used
5.011 with the runoff volume to estimate total sediment yield.
m = 1.0 + X > 150 ft C R E A M S is a model of erosion for a field-sized area
In* (8.121)
and is not intended to estimate watershed erosion,
m 2.0 X < 150. although it has b e e n used for that purpose. C o m p u t a ­
By combining rill and interrill components, Foster tion of soil erosion with the C R E A M S model is too
et al. (1980a) point out that the effective exponent is a complex t o b e m a d e without a computer; hence t h e
function of slope, slope length, and ratio EI /V a . 3Q u p
l/3
model is available in a c o m p u t e r format.

Transport Capacity, Particle Size Distribution, and


Erosion Interactions
PROCESS-BASED EROSION MODELS: WEPP
Transport capacity in the C R E A M S model is calcu­ THEORETICAL RILL AND INTERRILL MODEL
lated by the Yalin equation using the C R E A M S e r o d e d
size distributions, as discussed in C h a p t e r 7. T h e inter­ Background
action between d e t a c h m e n t and transport is defined by
A group of hydrologic and erosion scientists, primar­
the closed-form erosion solution, given in Eq. (8.17).
ily in the U S D A - A R S , conducted a massive data
T h e closed-form equation is applied to four different
collection a n d modeling effort with the objective of
cases
developing a process-oriented model that incorporates
Case I —Deposition over the entire segment m o d e r n erosion and hydrologic science (Lane and
Case II — D e t a c h m e n t by flow on the u p p e r end and Nearing, 1989; Foster a n d L a n e , 1987). T h e hydrologic
deposition in the lower end and sediment algorithms w e r e developed at several
Case III —Deposition on the u p p e r end and detach­ erosion laboratories and calibrated on data collected
ment by flow on the lower end on crop lands, rangelands, and forest lands research
Case IV — D e t a c h m e n t over the entire slope. sites dispersed throughout the 48 states. Results from
Process-Based Erosion Models: WEPP Theoretical Rill and Interrill Model 301

the studies included a data base on soil erodibility N e t deposition is defined by a modification of Eqs.
parameters, rill and interrill hydraulic p a r a m e t e r s , and (8.6) and (8.3), or
effects of buried residue on erosion p a r a m e t e r s .
T h e W E P P erosion model is included as a compo­ D =
x 0 ^(« -r ),
T s c (8.126)
nent of an overall continuous simulation model includ­
ing a rainfall generator, runoff predictor, plant growth w h e r e β is a p a r a m e t e r defining t h e impact of turbu­
τ

simulator, and sediment yield model. T h e final result is lence on settling. F o r W E P P , β = 0.5. Storm et al. τ

a process-oriented continuous simulation c o m p u t e r (1990) showed that Eq. (8.126) is a modification of the
model for predicting runoff and soil erosion on hill- overflow rate discussed in C h a p t e r 9. T h e formulation
slopes as affected by soils, climate, and m a n a g e m e n t . in E q . (8.126) is t h e laminar form of t h e overflow rate,
Based on the experimental results, prediction equa­ but t h e relationship a p p r o a c h e s that of a turbulent
tions for hydrologic and erosion p a r a m e t e r s have b e e n form because (1) t h e slope is discretized into segments
developed and evaluated. F u r t h e r evaluation of t h e and (2) t h e assumption is m a d e that flow is completely
p a r a m e t e r prediction relationships continues. All units mixed at the beginning of each segment. Storm et al.
in the model are SI. T h e model will be available as a (1990) showed that discretizing t h e slope into 20 seg­
personal computer program in a hillslope version and m e n t s will e n s u r e that the deposition model ap­
an upcoming so-called watershed version. T h e water­ p r o a c h e d that of a fully turbulent model.
shed version, like the C R E A M S model, is intended for
small upland watersheds and will include impound­ Equilibrium Channel Width
ment channel erosion elements. L a n e and Foster (1980) showed that concentrated
T h e following information covers erosion fundamen­ flow erosion tends to develop channels with vertical
tals, presented h e r e to show how theoretical concepts walls and with an equilibrium width that is propor­
are combined in W E P P to calculate sediment yield. tional to flow r a t e . In W E P P , a rectangular shape is
T h e W E P P model is still u n d e r development, hence assumed with a n equilibrium width given by
the alogrithms presented h e r e are tentative. Most algo­
Κ = c q* w
d
(8.127)
rithms have reached a final condition as of the writing
of this text with a final "freezing" of t h e algorithms w h e r e W is t h e equilibrium width (meters), q is flow
e

expected in early 1994. r a t e at t h e end of t h e slope (cubic m e t e r s p e r second),


and c and d
w are equilibrium p a r a m e t e r s that are
w

functions of soil and vegetation. Unlike t h e Foster and


Basic Governing Erosion Relationships L a n e (1983) model p r e s e n t e d in an earlier section, the
Continuity W E P P model does not provide for the presence of a
nonerodible layer. Presently, Gilley et al. (1990) recom­
T h e basic continuity relationship is the steady-state
m e n d that universal values of c = 1.13 and d = 0.303
spatially varied sediment continuity equation, given
w w

b e used. T h e Foster a n d L a n e (1983) concentrated flow


earlier by Eq. (8.2) or
model, discussed earlier, could b e used to correct c w

dqjdx = D +
{ D,
r and d w for soil properties. T h e flow rate for Eq.
(8.127) is a steady-state flow at the e n d of a slope,
where D and D are interrill and rill erosion rates (kg
{ T

assumed in W E P P to b e given by the p e a k runoff rate,


s e c " m ~ ) , q is sediment load ( k g / s e c · m), and χ is
1 2
s

or
distance downslope (m). In this formulation, D and D { T

are detachment rates per unit slope area, not p e r unit q=I XS ,
p a (8.128)
length of rill. Net soil detachment is related to detach­ w h e r e I is t h e p e a k r a t e of rainfall excess (meters p e r
p

ment potential and transport capacity by Eq. (8.5), or second), X is t h e slope length (meters), and 5 is the r s

D = O (l-<7 /r ), (8.124) rill spacing (meters). In t h e present formulation of


r r c s c
W E P P , 5 is assumed to b e 1 m.
r s

where D is detachment potential ( k g / s e c · m) and T


TC c
Flow d e p t h is calculated using the Darcy Weisbach
is transport capacity ( k g / s e c · m). D e t a c h m e n t poten­ friction factor [Eq. (4.21)], rill width, a n d average slope.
tial is defined by shear excess given in Eq. (8.25) or With t h e calculated d e p t h and width, hydraulic radius
D r c = tf (r-T ),
r C (8.125) is d e t e r m i n e d a n d shear stress o n the soil calculated
from
where rill erodibility K replaces a in Eq. (8.25), τ is
T

stress acting on the soil particles (Pascal), and T is r e = yS *(/ //,),


a s (8.129)
c

critical tractive force (Pascal). W h e n τ < r , detach­ c w h e r e r is shear stress on the soil at the end of an
e

ment is zero. average uniform slope (Pa), y is density of water


8. Erosion ond Sediment Yield
( N / m ) , S is the average channel slope, R is hydraulic
3 and
a

radius, / is friction factor for soil, and / is total 7^* _ t_


s t
(8.138)
friction factor. M e t h o d s for estimating / a n d / are t s *ce
given in L a n e and Nearing (1989).
T h u s , since T cl = fc r , tl
3/2
then

Sediment Transport Capacity Sediment transport is r* = k (ax%


c tr + fa,), (8.139)
calculated by a relationship developed by Finkner et al. where k tr is k /k .
t tl

(1989), or
Normalized Rill and Interrill Detachment
T = A: r 3 / 2
, (8.130)
c t
and Deposition Parameter
where T is transport capacity, τ is shear stress o n t h e
c
W E P P utilizes dimensionless d e t a c h m e n t a n d d e p o ­
soil surface, and k is a calibration coefficient calcu­
t
sition equations and p a r a m e t e r s in making calcula­
lated from T and τ estimated at the end of t h e slope.
c
tions. Using dimensionless p a r a m e t e r s , rill d e t a c h m e n t
Details on calculation of k are given in Finkner et al. t
is corrected for consolidation, f r e e z e - t h a w , and
(1989). below-ground root mass. T h e dimensionless rill p a r a m ­
eters are
Normalized Parameters Slopes are not necessarily
η = LK K K T /T (8.140)
uniform, but can be divided into segments. F o r each T TC TbT e C€

segment, a normalized slope is defined as and


T n = T T /T ,
C C C C E (8.141)
S* = 5 / 5 a = aX* + b, (8.131)
w h e r e 77 is a dimensionless rill d e t a c h m e n t p a r a m e t e r ,
where
Krc is a dimensionless subfactor for soil consolidation
X* =X/L. (8.132) a n d f r e e z e - t h a w effects, K is a dimensionless sub- Tbr

L is the slope length, S is the average slope m e a s u r e d


a
factor for root effects, r is dimensionless tractive c n

over L , and a and b are coefficients set to m a k e the force, a n d T is a dimensionless subfactor to account
c c

model fit the land slope. A n equivalent transport ca­ 1 for consolidation a n d f r e e z e - t h a w . T o apply t h e sub-
pacity T at the end of the uniform slope is defined as
ce
factor correction to Eq. 8.125, multiply k by k and r rc

a function of the shear on the uniform slope r , or e


krbr while T is multiplied by r ^ . T h e standard condi­
c

tion for W E P P , used to define K , is for b a r e soil r

T ce = kr/. tl
3 2
(8.133) immediately after tillage. A dimensional deposition
Shear stress versus distance downslope can b e derived p a r a m e t e r is given by
using the D a r c y - W e i s b a c h equation Φ=β (Κ/Ι ). (8.142)
τ ρ

2/3
w h e r e V is settling velocity ( m / s e c ) . A n o n d i m e n -
s

4 x s (8.134)
c sional interrill d e t a c h m e n t p a r a m e t e r is given by
e=LD /T , (8.143)
where C is t h e Chezy coefficient given by C = { ce

(8g// ) t is t h e friction factor for rills and X is t h e


0 5
where D x is interrill erosion r a t e given by
downslope distance. Likewise, r can b e defined from e
D^KJ&CcCiSn/Wt), (8.144)
slope length L and 5 . Defining τ * as τ / τ , a β

where S { is t h e interrill slope adjustment factor given


,2/3
X S by
τ* = (8.135)
5 = 1.05 - 0 . 8 5 e "
f
4 s i n i l
, (8.145)

or, since S/S a = aX % + b, where Ω is t h e interrill slope angle, K is t h e baseline x

interrill erodibility, / is t h e equivalent rainfall inten­ e

τ * = [aXl +bX*Y . /3
(8.136) sity ( m / s e c ) , and C is the canopy factor, C is the c gc

ground cover factor, 5 is rill spacing (meters), and W


Sediment load and transport capacity are also normal­ r s c

is equilibrium rill width (meters) defined by E q . (8.127).


ized to transport capacity at the end of the uniform
T h e canopy effect is given by a relationship similar to
slope, or
that in R U S L E , or
Qt = ^~ (8-137)
C = 1 - F e-° ,
c c
MH
(8.146)
w h e r e F is t h e fraction of ground cover a n d Η is crop
c

height (meters). T h e ground cover factor is given by


1
Given two points on a slope segment, a and b can be determined
for Eq. (8.131).
-2.5F ot (8.147)
Process-Based Erosion Models: WEPP Theoretical Rill and Interrill Model 303

where F is fraction of ground cover in the interrill


area.
gc

mt + b
(8.153)

Normalized Erosion Equation bqj (8.154)


Detachment and Deposition Equation
Utilizing Eqs. (8.2), (8.4), and (8.124) along with
W E P P p a r a m e t e r s , a fundamental erosion equation 4o* (8.155)
(detachment) can b e written as

dqt a n d q is t h e inflow of water (meters per second) at the


dx*
= - Ο 1 ~ (8.148) 0

top of the h o m o g e n e o u s section. Nondimensional


transport capacity is given by
where η, r , and ε are t h e normalized p a r a m e t e r s
c n

defined earlier. T h e dimensionless p a r a m e t e r s τ * , q*,


and τ * are normalized functions of X* defined ear­ + Β χ< + C ) .
0 0 (8.156)
lier. T h e equation must b e solved numerically. Utiliz­
ing Eqs. (8.2) and (8.126), t h e normalized deposition T h e form of the d e t a c h m e n t equation, Eq. (8.148),
equation is must remain t h e same. T h e form of the deposition
dqt/dx* = (ΦΛ.)(Γ* - tf *) + ε,
s (8.149)
equation, Eq. (8.149), is changed by replacing
with (φ/χ* + 4o )· Equation (8.156) would b e used on
(Φ/χ*)

where φ and ε are the normalized erosion p a r a m e t e r s any slope segment with inflow at the t o p of the slope.
defined earlier. In application, φ can be calculated for
each particle class, if desired.

Sediment Yield
Use of the WEPP Model
A s indicated earlier, the W E P P erosion model pre­
Sediment load is calculated from normalized sedi­
sented in this c h a p t e r is but o n e c o m p o n e n t of a large
ment load by
hydrology and sedimentology continuous simulation
1s = <ltT [W /S ),
ce e ts (8.150) model that includes crop growth models, hydrology
models, and p a r a m e t e r generation relationships. Its
where q is sediment load in kilograms per second p e r
s
use obviously requires a computer.
unit width of hillslope. Equations (8.148)-(8.150) would
be used for the first segment of a slope w h e r e t h e
inflow is zero at the top of t h e slope. Equations for Problems
nonzero inflow are given in the following section.
(8.1) Discuss the difference b e t w e e n rill erosion, in­
terrill erosion, e p h e m e r a l gully erosion, and channel
Evaluating Downslope Variability with Inflow erosion. E n u m e r a t e the processes a n d forces involved
at the Top of the Slope in each type of erosion.
(8.2) O n a uniform slope of 150 ft, three different
Downslope variability in flow rate and erosion pa­
mulch rates are being considered. T h e resulting inter­
rameters can be evaluated with W E P P . In W E P P , t h e
rill d e t a c h m e n t rates would be 0.008, 0.005, a n d 0.002
flow paths are divided into sections with homogenous
l b / s e c · ft. If the d e t a c h m e n t capacity a n d transport
erosion properties. Those sections are elements that
capacity at the e n d of the slope is 0.02 l b / s e c · ft and
may have complex topography, but t h e erosion a n d
1.8 l b / s e c · ft, respectively, estimate t h e sediment load
cover p a r a m e t e r s would be constant over a section.
at 25, 50, 75, 100, 125, and 150 ft downslope for each of
Each section is treated separately with a lateral inflow
the mulch rates. A s s u m e steady uniform flow. Plot the
defined at the top of the section. With lateral inflow,
results, showing the difference between sediment loads.
the nondimensional shear stress becomes
(8.3) A researcher measures rainfall energy of 1300
ft · t o n s f / a c r e · in. in a storm with an intensity of 2.0
τ . = {A xl 0 + B x*0 + C ) 0
2 / 3
, (8.151)
i n . / h r . Is this a reasonable value? Justify your answer.
where A Qi B , and C
0 0 are coefficients defined by C o m p a r e this value to predictions from equations of
Wischmeier and Smith and Brown and Foster.
(8.152) (8.4) Estimate the energy content (ft · t o n s f / a c r e ·
in.) of a storm with t h e following rainfall intensities
304 8. Erosion and Sediment Yield

Time
(8.10) Estimate t h e LS factor for a 250-ft slope
Time
period Intensity period Intensity length with an average slope of 1 5 % u n d e r t h e follow­
(min) (in./hr) (min) (in./hr) ing assumptions: (1) Uniform slope; (2) convex slope
divided into 5 equal segments with slopes of 5, 10, 15,
0-15 0.6 30-45 4.0 20, and 2 5 % moving from top t o bottom of t h e slope;
15-20 1.4 45-60 0.8 (3) concave slope divided into 5 equal segments of 25,
20-30 3.0 20, 15, 10, a n d 5 % moving from top to bottom; and (4)
5-shaped slope with slopes of 10, 15, 25, 15, and 10%
moving from t o p to bottom. A s s u m e a m o d e r a t e ratio
W h a t is t h e EI index? If this storm occurred in
30
of rill to interrill erosion.
Memphis, Tennessee, what would b e t h e return period,
(8.11) A 180-ft-long slope has t h e following charac­
based o n t h e EI index? W h a t would b e t h e r e t u r n
teristics.
30

period if t h e storm occurred in your area?


(8.5) D e t e r m i n e t h e following R factors for Los Section Parameter Value
Angeles, California and Rapid City, South Dakota:
Average annual R factor, 10-year, 24-hr historic storm 1 Length 60 ft
R factor, a n d 10-year, 24-hr synthetic storm R factor. Steepness 25%
If t h e slope being considered has a slope steepness of Erodibility 0.2
0 . 1 % , m a k e a correction for ponding. Cover factor 0.6
(8.6) If you a r e scheduling construction operations, Practice factor 1.0
how would t h e difference in distribution of R factor
affect your scheduling decision for Los Angeles and
2 Length 60 ft
South Georgia?
Steepness 15%
(8.7) Estimate t h e average annual Κ using t h e
Wischmeier nomograph if t h e soil has t h e following Erodibility 0.3

characteristics. Cover factor 0.2


Practice factor 1.0

Property
3 Length 60 ft
Value
Steepness 5%

%SA 0.1-2.0 mm 25% Erodibility 0.2

% VFS 0.05-0.1 mm 4% Cover factor 0.1

%MS 0.002-0.05 mm 40% Practice factor 1.0

%CL <0.002 mm 31%


%OM 1% Estimate t h e erosion from a single storm with an R
Structure Medium
factor of 100 for this site.
(8.12) Estimate t h e average annual C factor for corn
Permeability Moderate
grown in upstate N e w York u n d e r t h e following as­
Annual R factor English units 80
sumptions: (1) high production conventional tillage; fall
turnplowing o n Oct 15, seedbed preparation on May 1,
Correct t h e annual Κ factor accounting for seasonable planting o n May 1 5 , 1 0 % cover o n J u n 5, 5 0 % cover on
variations in Κ a n d t h e variability in t h e EI index.
30 July 1, 7 5 % cover o n July 15, harvest o n October 1
T h e average n u m b e r of frost-free days is 120, a n d t h e (assume that t h e C factor for period 4 is t h e same as
site is located in central New York. T h e average tem­ period 3); (2) minimum tillage with t h e same planting
p e r a t u r e drops below 27° F on D e c e m b e r 1 and rises and harvest dates a n d same dates for percentage of
above 27° F on March 1. M a k e a plot of t h e resulting Κ cover.
factor versus time of year. Indicate t h e nominal Κ (8.13) If t h e Κ in Problem (8.12) is 0.3, slope lengths
value on t h e plot. are 150 ft, a n d slope steepness a r e 8 % , estimate t h e
(8.8) If the soil in Problem (8.7) has rock fragments average annual erosion assuming a m o d e r a t e ratio of
greater than 2.0 m m that occupy 3 0 % of t h e soil by rill to interrill erosion a n d a n average R factor of 175.
weight, what would be t h e resulting Κ value? (8.14) If t h e field in Problem (8.12) is t h e same as
(8.9) Estimate the LS factor for t h e R U S L E for a that in Problem (8.7), estimate t h e average annual
slope of 10% u n d e r t h e conditions of low, m o d e r a t e , erosion, accounting for seasonal variation in both Κ
and high ratio of rill to interrill erosion a n d a slope and C. A s s u m e a slope length of 150 ft a n d a slope
length of 100 ft. steepness of 8 % .
Process-Based Erosion Models: WEPP Theoretical Rill and Interrill Model 305

(8.15) If the bulk density of the soil in Problem 15-day period. This m e a n s that the effects of each
(8.14) is 1.2, what d e p t h of soil is being e r o d e d annu­ tillage operation on the distribution of above- and
ally by the crop. below-ground residue must be calculated. Develop an
(8.16) For a typical soil and field in your location, algorithm to evaluate that distribution of residue, using
estimate the average annual erosion for a crop that a spreadsheet.
would routinely be grown. Correct for a n n u a l variation (8.20) Estimate t h e C factor for a reconstructed
in C and K. Based on information from t h e soil survey surface mine soil on March 15, assuming that the soil is
on allowable annual erosion, develop a cropping m a n ­ r e g a r d e d on O c t o b e r 1, reseeded to winter small grain,
agement plan to k e e p the erosion rate below the allow­ and mulched with 2 t o n s / a c r e of wheat straw mulch.
able value. D u r i n g reconstruction, the soil is compacted to 1.2
(8.17) A strip mine in Southwestern Wyoming has times its loose density. Assume t h e same t e m p e r a t u r e
the following sequence of operations. and moisture as in Problem (8.18).
(8.21) If you live in an a r e a w h e r e timber harvesting
Time period Description occurs, take a trip to an active operation. Estimate the
p e r c e n t a g e of b a r e soil, t h e percentage of soil with root
1/1-3/1 Rangeland, 25% cover mat, the canopy height a n d cover, a n d the percentage
3/1-6/1 Bare soil, stripped of vegetation of total slope in steps. Using this information, estimate
(no cover, poor soil) the C factor. Collect information on t h e soil, slope
6/1-9/1 Active mining length, steepness, and R factor, and estimate the aver­
9/1-10/1 Regraded age a n n u a l erosion.
10/1-12/31 Seeded to range grass (8.22) T a k e a trip to a farming operation with and
without conservation practices. Estimate t h e annual C
and Ρ factors using t h e U S L E and R U S L E approach.
(1) Estimate the C value for the site. (2) If the soil has
Using these factors a n d t h e R K L S factors, estimate the
an erodibility of 0.4, a slope steepness of 12%, a slope
average annual erosion.
length of 250 ft, and a low ratio of rill to interrill
(8.23) T a k e a trip t o a farm with a crop in t h e
erosion, estimate the average annual erosion rate.
development stage. F r o m visual observations and your
(8.18) Estimate the C factor o n March 15 for the
best estimates, develop the subfactors for the R U S L E
following conditions: soybeans following corn, corn
C factor.
harvested October 15 with a yield of 5500 l b / a c r e , soil
(8.24) Starting with M a n n i n g ' s equation, show that
moisture at field capacity, moldboard plowing on
Eq. (8.76) is correct.
November 1, row planting on M a r c h 1, and a m o d e r a t e
(8.25) A n e p h e m e r a l gully contains a flow rate of
mixture of rill to interrill erosion. T h e following
0.005 m / s e c a slope of 0.015, η of 0.025, bulk density
3

monthly average t e m p e r a t u r e s and precipitations are


of 1.4, and a r of 1.8 Pa. Estimate the equilibrium
applicable
e

width, final width, initial d e t a c h m e n t rate, and detach­


m e n t rate at reaching a nonerodible layer located at a
d e p t h of 0.3 m. Use both the Foster and L a n e model
a n d t h e E p h e m e r a l Gully model a n d c o m p a r e t h e
Monthly average Monthly average
results.
temperature precepitation (8.26) If the flow for Problem (8.25) is from a dis­
Month (°F) (in.) turbed area with a sediment load equal to half t h e
transport capacity, what would b e t h e d e t a c h m e n t rates
Oct 55 2.5
prior to reaching t h e nonerodible layer.
Nov 50 4.0
(8.27) A 20-acre watershed is being disturbed to
Dec 40 3.0 develop a major shopping center. T h e planning com­
Jan 25 3.0 mission requires that t h e average sediment concentra­
Feb 30 4.0 tion in t h e drainage shall be n o greater than 1000
Mar 40 6.0 m g / l i t e r in a 10-year, 24-hr type II storm. T h e devel­
o p e r p r o p o s e d that a 100-ft-wide riparian zone of natu­
ral vegetation would solve the problem. Is the devel­
o p e r correct? T h e riparian zone is wooded with 8 0 %
grass at t h e surface. T h e slope shape is concave, the
(8.19) If the R U S L E is to be used to estimate ero­ slope gradient is 12.5%, and t h e surface is moderately
sion, an estimate of the quantity of residue on the rough. T h e average watershed slope length for the
surface and below the surface must be m a d e for every contributing area is 200 ft with a 6 % slope and a Κ of
306 8. Erosion and Sediment Yield
0.32. T h e sediment in the runoff from the contributing Ateshian, J. Κ. H. (1974). Estimation of rainfall erosion index. Proc.
area is 5 0 % silt and clay. For the site being considered, Am. Soc. Civil Eng., 100UR3): 2 9 3 - 3 0 7 .
Barfield, B. J., Moore, I. D., and Williams, R. G. (1979). Sediment
the 10-year, 24-hr rainfall is 5.0 in., and the curve
yield in surface mined watersheds. In "Proceedings, Symposium
number is 80 during construction. T h e surface will be on Surface Mine Hydrology, Sedimentology and Reclamation,"
bare during construction. T h e peak runoff rate for the University of Kentucky, Lexington, KY," pp. 8 3 - 9 2 .
watershed is 800 c f s / m i l e · in. runoff. (Hint, utilize
2
Barfield, B. J., Barnhisel, R. I., Hirschi, M. C , and Moore, I. D .
Fig. 8.27.) (1988). Compaction effects on erosion of mining spoil and recon­
(8.28) For the postmining hydrograph in Fig. 3.42, structed topsoil. Trans. Am. Soc. Agric. Eng. 3(2): 4 4 7 - 4 5 2 .

determine a corresponding sediment graph if the con­ Boyce, R. C. (1975). Sediment routing with sediment delivery ratios.
In "Present and Prospective Technology for Predicting Sediment
stant a in Eq. (8.109) is 1.0 and the sediment yield is
Yields and Sources," Publication A R S - S 4 0 , pp. 6 1 - 6 5 . U S D A -
2000 tons. Express the sediment graph in terms of Agricultural Research Service.
concentration (milligrams per liter). Bradford, J. M., Farrell, D . Α., and Larson, W. E. (1973). Mathemat­
(8.29) A watershed is divided into t h r e e homoge­ ical evaluation of factors affecting gully stability. Soil Sci. Soc.
neous subareas as shown below. Estimate the total Am. Proc. 37(1): 103-107.
yield from each subwatershed and the combined water­ Brown, L. C , and Foster, G. R. (1987). Storm erosivity using ideal­
ized intensity distributions. Trans. Am. Soc. Agric. Eng.
sheds using the M U S L E procedures.
3<K2):293-307.

Average
Sub Hydraulic USLE particle
watershed parameters parameters diam(mm)

1 β (acre-ft) = 5.1 Κ = 0.3 0.11


q (cfs) = 81.2
p Length 175 ft
r ( h r ) = 0.5
t Stpns 4.3%
C = 0.45
P= 1.0

2 Q (acre-ft) = 6.3 AC = 0 . 4 6 0.09


q (cfs) = 1 0 2 . 0
p Length 220 ft
r ( h r ) = 0.32
t Stpns 8.2%
C = 0.84
Ρ = 0.1

3 Q (acre-ft) = 4.1 Κ = 0.26 0.32


q (cfs) = 66.2
p Length 160 ft
r ( h r ) = 0.25
t Stpns 10.0%
C=0.02
P= 1.0

(8.30) For watershed 1 in Problem (8.29), estimate Bruce, R. R., Harper, L. Α., Leonard, R. Α., Snyder, W. M., and
Thomas, A. W. (1975). A model for runoff of pesticides from
the interrill and rill detachment potential using the
small uplane watersheds. / . Environ. Quality 4(4):541-548.
C R E A M S model, Eqs. (8.118M8.124).
Bubenzer, G. D., and Jones, B. A. (1971). D r o p size and impact
velocity effects on the detachment of soils under simulated rain­
fall. Trans. Am. Soc. Agric. Eng. 14(4):625-628.
Cook, H. L. (1936). The nature and controlling variables of the water
erosion process. Soil Sci. Soc. Am. Proc. 1:60-64.
References Cooley, K. R. (1980). Erosivity "R" for individual design storms. In
Alberts, Ε. E., Laflen, J. M., Rawls, W. J., Simanton, J. R., and " C R E A M S — A Field Scale Model for Chemicals, Runoff and
Nearing, M. A. (1989). Soil component. In " U S D A Water Ero­ Erosion from Agricultural Management Systems," Vol. Ill, Chap.
sion Prediction Project: Hillslope Profile Model Documentation," 2, U S D A - S E A Conservation Report N o . 26, pp. 3 8 6 - 3 9 7 .
Chap. 6, N S E R L Report No. 2. National Soil Erosion Labora­ Curtis, W. R. (1971). Strip mining, erosion and sedimentation. Trans.
tory, U S D A - A R S , W. Lafayette, IN. Am. Soc. Agric. Eng. 14(3):434-436.
Process-Based Erosion Models: WEPP Theoretical Rill and Interrill Model 307

Dillaha, Τ. Α., and Beasley, D . B. (1983). Distributed parameter Foster, G. R., Meyer, L. D . , and Onstad, C. A. (1977b). A runoff
modeling of sediment movement and particle size distributions. erosivity factor and variable slope length exponents for soil loss
Trans. Am. Soc. Agric. Eng. 26(6):1766-1777. estimates. Trans. Am. Soc. Agric. Eng. 20(4):683-687.
Dissmeyer, G. E., and Foster, G. R. (1981). Estimating the cover- Foster, G. R., Lane, L. J., Nowlin, J. D . , Laflen, L. M., and Young,
management factor (C) in the universal soil loss equation for R. A. (1980a). A model to estimate the sediment yield from
forest conditions. / . Soil Water Conserv. 36:235-240. field-sized areas: Development of model. In " C R E A M S — A Field
Dissmeyer, G. E., and Foster, G. R. (1984). " A Guide for Predicting Scale Model for Chemicals, Runoff, and Erosion from Agricul­
Sheet and Rill Erosion on Forest Land," Forest Service Techni­ tural Management Systems," Vol. I, "Model Documentation,"
cal Publication RA-TP6. United States Department of Agricul­ Chap. 3, U S D A - S E A Conservation Report No. 26, p. 3 6 - 6 4 .
ture. Foster, G. R., Lane, L. J., and Nowlin, J. D . (1980b). A model to
Einstein, H. A. (1968). Deposition of suspended particles in a gravel estimate the sediment yield from field-sized areas: selection of
bed. Proc. Am. Soc. Civil Eng. 94(HY5): 1197-1205. parameter values. In " C R E A M S — A Field Scale Model for
Ellison, W. D., and Ellison, Ο. T. (1947). Soil erosion studies. IV. Chemicals, Runoff, and Erosion from Agricultural Management
Soil detachment by surface flow. Agric. Eng. 28(9):402-408. Systems," Vol. II, U S D A - S E A Conservation Report No. 26, pp.
Finkner, S. C , Nearing, Μ. Α., Foster, G. R., and Gilley, J. E. 193-281.
(1989). A simplified equation for modeling sediment transport Foster, G. R., Young, R. Α., and Neibling, W. H. (1985). Sediment
capacity. Trans. Am. Soc. Agric. Eng. 32(5): 1545-1550. composition for nonpoint source pollution analysis. Trans. Am.
Flanagan, D . C. (1990). WEPP Second Edition, N S E R L Report N o . Soc. Agric. Eng. 28(1): 133-146.
4. National Soil Erosion Lab, United States Department of
Foster, G. R., Weesies, G. Α., Renard, K. G., Yoder, D. C , and
Agriculture, Agricultural Research Service, Lafayette, IN.
Porter, J. P. (1991). Conservation practice factor. In "Estimating
Fogle, A. W., Barfield, B. J. (1992). C H A N N E L , a model of channel Soil Erosion by W a t e r — A Guide to Conservation Planning with
erosion by shear, scour, and channel headwall advance. Part I. the Revised Universal Soil Loss Equation ( R U S L E ) (Renard et
Model Development. Research Report 186 Water Resources al., eds.), Chap. 6, A R S publication, U.S. Department of Agricul­
Research Institute, Univ. of Kentucky, Lexington. ture, in press.
Forest Service (1980). "An Approach to Water Resources Evaluation
Free, G. R. (1960). Erosion characteristics of rainfall. Agric. Eng.
of Non-point Silvicultural Sources," a procedural handbook.
41(7):447-455.
EPA-600/8-80-012, Environmental Protection Agency, Washing­
Gilley, J. E., and Finkner, S. C. (1985). Estimating soil detachment
ton, D C .
caused by raindrop impact. Trans. Am. Soc. Agric. Eng.
Foster, G. R. (1982). Modeling the erosion process. In "Hydrology
28(1):140-146.
Modeling of Small Watersheds" (Haan, Johnson, and Brakensiek,
eds.), Monograph N o . 5. American Society of Agricultural Engi­ Gilley, J. E., Woolhiser, D . Α., and McWhorter, D . B. (1985a).
neers, St. Joseph, MI. Interrill soil erosion. I. Development of model equations. Trans.
Foster, G. R. (1986). Understanding ephemeral gully erosion. In Am. Soc. Agric. Eng. 28(1):147-159.
"Soil Conservation, Assessing the National Resources Inventory," Gilley, J. E., Woolhiser, D . Α., and McWhorter, D . B. (1985b).
Vol. 2. National Academy Press, Washington, D C . Interrill soil erosion. II. Testing and use of model equations.
Foster, G. R., and Highfill, R. E. (1983). Effect of terraces on soil Trans. Am. Soc. Agric. Eng. 28(1): 154-159.
loss: U S L E Ρ factors for terraces. / . Soil Water Conser. Gilley, J. E., Kottwitz, E. R., and Simanton, J. R. (1990). Hydraulic
38(1):48-51. characteristics of rills. Trans. Am. Soc. Agric. Eng.,
Foster, G. R., and Lane, L. J. (1981). Modeling rill density-Discus­ 33(6):1900-1906.
sion. Proc. Am. Soc. Civil Eng. 107(IR1):109-112. Gregory, J. M., McCarty, T. R., Ghidey, F., and Alberts, Ε. E.
Foster, G. R., and Lane, L. J. (1983). Erosion by concentrated flow in (1985). Derivation and evaluation of a residue decay equation.
farm fields. In "Proceedings, D . B. Simons Symposium on Ero­ Trans. Am. Soc. Agric. Eng. 28(1):98-105.
sion and Sedimentation," Colorado State University, Ft. Collins, Gunn, R., and Kinzer, G. D . (1949). Terminal velocity of fall for
CO, pp. 9.65-9.82.
water droplets in stagnant air. / . Meteorol. 6(4):243-248.
Foster, G. R., and Lane, L. J. (1987). User requirements, U S D A -
Harvey, M. D., Watson, C. C , and Schumm, S. A. (1985). Gully
water erosion prediction project (WEPP), N S E R L Report N o . 1.
erosion, Technical note 366. Bureau of Land Management, U. S.
National Soil Erosion Research Laboratory, United States D e ­
Department of Interior, Denver Service Center, Denver, CO.
partment of Agriculture, Agricultural Research Service, West
Hirschi, M. C , and Barfield, B. J. (1988a). K Y E R M O — A physically
Lafayette, IN.
based research erosion model. I. Model development. Trans. Am.
Foster, G. R., and Meyer, L. D . (1972a). A closed-form soil erosion
Soc. Agric. Eng. 31(3):804-813.
equation for upland areas. In "Sedimentation: Symposium to
Honor Prof. H. A. Einstein" (Shen, ed.), Chap. 12, pp. 12.1-12.19. Hirschi, M. C , and Barfield, B. J. (1988b). K Y E R M O — A physically
based research erosion model. II. Model sensitivity analysis and
Colorado State University, Ft. Collins, CO.
Foster, G. R., and Meyer, L. D . (1975). Mathematical simulation of testing. Trans. Am. Soc. Agric. Eng. 31(3):814-820.
upland erosion by fundamental erosion mechanics. In "Present Hirschi, M. C , Barfield, B. J., and Moore, I. D . (1983). Modeling
and Prospective Technology for Predicting Sediment Yields and erosion on long steep slopes with emphasis on the rilling process,
Sources," ARS-S-40, pp. 190-207. USDA-Agricultural Research Research report N o . 148. Water Resources Research Institute,
Services. University of Kentucky.
Foster, G. R., and Wischmeier, W. H. (1974). Evaluating irregular Horton, R. E. (1945). Erosional development of streams and their
slopes for soil loss prediction. Trans. Am. Soc. Agric. Eng. drainage basins, hydrological approach to quantitative morphol­
17(2):305-309. ogy. Geol. Soc. Am. Bull. 56:275-370.
Foster, G. R., Meyer, L. D . , and Onstad, C. A. (1977a). A n erosion Huang, C , Bradford, J. M., and Cushman, J. H. (1982). A numerical
equation derived from basic erosion principles. Trans. Am. Soc. study of raindrop impact phenomena: The rigid case. Soil Sci.
Agric. Eng. 20(4):678-682. Soc. Am. I. 46(1):14-19.
308 8. Erosion and Sediment Yield
Huang, C , Bradford, J. M., and Cushman, J. H. (1983). A numerical McCool, D . K., Brown, L. C , Foster, G. R., Mutchler, C. K., and
study of raindrop impact phenomena: the elastic deformation Meyer, L. D . (1989). Revised slope length factor for the Universal
case. Soil Sci. Soc. Am. J. 47:855-861. Soil Loss Equation. Trans. Am. Soc. Agric. Eng. 32(5): 1571-1576.
Israelson, C. E., Clyde, C. G., Fletcher, J. E., Israelson, Ε. K., Haws, McCool, D . K., Foster, G. R., and Weesies, G. A. (1993). Slope
F. W., Parker, P. E., and Farmer, Ε. E. (1980). Erosion control length and steepness factor. In "Predicting Soil Erosion by Water
during highway construction: Manual on principles and practices, — A Guide to Conservation Planning with the Revised Universal
Report 221. Transportation Research Board, National Research Soil Loss Equation ( R U S L E ) " (Renard et al., eds.), Chap. 4,
Council, Washington, D C . U S D A - A R S Special Publication, in press.
Kelly, W. E., and Gularte, R. C. (1981). Erosion resistance of Mclssac, G. F., Mitchell, J. F., and Hirschi, M. C. (1987). Slope
cohesive soils. Proc. Am. Soc. Civil Eng. 107(HY10):1211-1224. steepness effects on soil loss from disturbed lands. Trans. Am.
Kilinc, M., and Richardson, Ε. V. (1973). Mechanics of soil erosion Soc. Agric. Eng. 30(4):1005-1013.
from overland flow generated by simulated rainfall, Hydrology Meyer, L. D . (1981). How rain intensity affects interrill erosion.
paper N o . 63. Colorado State University, Ft. Collins, CO. Trans. Am. Soc. Agric. Eng. 24(6): 1472-1475.
Knisel, W. G., ed. (1980). " C R E A M S — A Field-Scale Model for Meyer, L. D., and Monke, E. J. (1965). Mechanics of soil erosion by
Chemicals, Runoff, and Erosion from Agricultural Management rainfall and overland flow. Trans. Am. Soc. Agric. Eng.
Systems," Conservation Research Report N o . 26. U.S. Depart­ 8(4):572-580.
ment of Agriculture. Meyer, L. D . , and Wischmeier, W. H. (1969). Mathematical simula­
Laflen, J. M., Foster, G. R., and Onstad, C. A. (1985). Simulation of tion of the process of soil erosion by water. Trans. Am. Soc.
individual storm soil loss for modeling the impact of soil erosion Agric. Eng. 12(6):754-758, 762.e
on productivity. In "Soil Erosion and Conservation" (El-Swaify Meyer, L. D., Johnson, C. B., and Foster, G. R. (1972). Stone and
et al., eds.), pp. 2 8 5 - 2 9 5 . SCS of Am., Ankeny, IA. wood chip mulches for erosion control o n construction sites. / .
Lane, E. W. (1953). Progress report on studies on the design of Soil. Water Conserv. 27(6):264-269.
stable channels of the Bureau of Reclamation. In "Proceedings, Meyer, L. D . , Foster, G. R., and Romkens, M. J. M. (1975). Source
American Society of Civil Engineers, Irrigation and Drainage of soil eroded by water from upland slopes. In "Present and
Division," Separate No. 280. Prospective Technology for Predicting Sediment Yields and
Lane, E. W. (1955). Design of stable channels. Am. Soc. Civil Eng. Sources," ARS-S-40, pp. 1 7 7 - 1 7 9 . USDA-Agricultural Research
Trans. 120:1234-1279. Service.
Lane, L. J., and Foster, G. R. (1980). Concentrated flow relation­ Mitchell, J. K., Moldenhauer, W. C , and Gustavson, D . G. (1983).
ships. In " C R E A M S — A Field Scale Model for Chemicals, Erodibility of selected reclaimed surface mine spoil. Trans. Am.
Runoff, and Erosion from Agricultural Management Systems," Soc. Agric. Eng. 26(5):1413-1417, 1421.
Vol. Ill, Supporting Documentation, Chap. 11, U S D A - S E A Con­ Moore, I. D . , and Burch, G. J. (1986). Sediment transport capacity of
servation Report N o . 26, pp. 4 7 4 - 4 8 5 . sheet and rill flow: Application of unit stream power theory.
Lane, L. J., and Nearing, Μ. Α., ed. (1989). Water erosion prediction Water Resources Res. 22(8): 1350-1360.
project: Hillslope profile model documentation, N S E R L Report Mosely, M. P. (1974). Experimental study of rill erosion. Trans. Am.
No. 2. U S D A - A R S National Soil Erosion Research Laboratory, Soc. Agric. Eng. 17(5):909-913, 916.
West Lafayette, IN. Mossaad, Μ. E., and Wu, Τ. H. (1984). A stochastic model of soil
Larson, W. E., Holt, R. F., and Carlson, C. W. (1978). Residue for erosion. Int. J. Numer. Analyt. Methods Geomech. 8:201-224.
soil composition. In "Crop Residue Management Systems" Musgrave, G. W. (1947). The quantitative evaluation of the factors in
(Oshwald et al., eds.), pp. 1-15. American Society of Agronomy, water erosion, a first approximation. J. Soil Water Conserv.
Madison, WI. 2(3): 133-138.
Leopold, L. B., Wolman, M. G., and Miller, J. P. (1964). "Fluvial Mutchler, C. K., and Carter, C. E. (1983). Soil erodibility variation
Processes in Geomorphology." Freeman, San Francisco. during the year. Trans. Am. Soc. Agric. Eng. 26(4):1102-1104,
Lewis, S. M. (1990). P R O R I L — A probabilistic physically based 1108.
erosion model, Unpublished master thesis. Department of Agri­ Mutchler, C. K., and Larson, C. L. (1971). Splash amounts from
cultural Engineering, University of Kentucky, Lexington, KY. waterdrop impact on a smooth surface. Water Resources Res.
Lewis, S. M., Barfield, B. J., and Storm, D . (1990). A n erosion model 7(l):195-200.
using probability distributions for rill flow and density, Paper N o . Mutchler, C. K., Murphree, C. E., and McGregor, K. C. (1982).
90-2623. American Society of Agricultural Engineers, St. Joseph, Subfactor method for computing C factors for continuous cotton.
MI. [In press for 1994 Transaction A S A E ] Trans. A m . Soc. Agric. Eng. 25(2):327-332.
Li, R., Ponce, V. M., and Simons, D . B. (1980). Modeling rill density. Nearing, Μ. Α., Foster, G. R., Lane, L. J., and Finkner, S. C. (1989).
Proc. Am. Soc. Civil Eng. 106(1 R l ) : 6 3 - 6 7 . A process-based soil erosion model for U S D A - w a t e r erosion
Lundgren, H., and Jonsson, I. G. (1964). Shear and velocity distribu­ prediction project technology. Trans. Am. Soc. Agric. Eng.
tion in shallow channels. Proc. Am. Soc. Civil Eng. 8(3):419-422. 32(5): 1587-1593.
Lyle, W. M., and Smerdon, Ε. T. (1965). Relation of compaction and Palmer, R. S. (1965). Waterdrop impact forces. Trans. Am. Soc.
other soil properties to erosion resistance of soils. Trans. Am. Agric. Eng. 8(l):69-70.
Soc. Agric. Eng. 8(3):419-421. Quansah, C. (1981). T h e effect of soil type, slope, rain intensity, and
Mantz, P. A. (1977). Incipient transport of fine grains and flakes of their interactions on splash detachment and transport. J. Soil Sci.
fluid—Extended shields diagram. Proc. Am. Soc. Civil Eng. 32:215-224.
103(HY6):601-615. Quinn, M. W., and Laflen, J. M. (1983). Characteristics of raindrop
McCool, D . K., Brown, L. C , Foster, G. R., Mutchler, C. K., and throughfall under corn canopy. Trans. A m . Soc. Agric. Eng.
Meyer, L. D . (1987). Revised slope steepness factor for the 26(5): 1445-1450.
Universal Soil Loss Equation. Trans. Am. Soc. Agric. Eng. Rawls, W. J., Brakensiek, D . L., and Saxton, Κ. E. (1982). Estimation
3<K5):1387-1396. of soil water properties. Trans. A m . Soc. Agric. Eng.
Process-Based Erosion Models: WEPP Theoretical Rill and Interrill Model 309

25(5):1316-1320. Sedimentology and Reclamation, College of Engineering, Univer­


Renard, K. G., and Laursen, Ε. M. (1975). Dynamic behavior sity of Kentucky, Lexington, KY."
model of ephemeral stream. Proc. Am. Soc. Civil Eng. Storm, D . E., Barfield, B. J., and Ormsbee, L. E. (1990). Hydrology
101(HY5):511-528. and sedimentology of dynamic rill networks. I. Erosion model for
Renard, K. G., Foster, G. R., Weesies, G. Α., McCool, D. K., and dynamic rill networks, Research report N o . 178. Kentucky Water
Yoder, D . C. (1993a). "Predicting Soil Erosion by W a t e r — A Resources Research Institute, University of Kentucky, Lexington,
Guide to Conservation Planning with the Revised Universal Soil KY.
Loss Equation R U S L E , " Publication A R S . U.S. Department of Transportation Research Board (1980). Erosion control during high­
Agriculture, Washington, D C , in press. way construction. Manual on principles and practices. Report
Renard, K. G., McCool, D . K., Cooley, K. R., Mutchler, C. K., and 221. National Cooperative Highway Research Program. Trans­
Foster, G. R. (1993b). Rainfall-runoff erosivity factor (R). Chap­ portation Research Board, National Research Council, Washing­
ter 2 In Renard, Foster and Weesies (eds) "Predicting Soil ton, D C .
Erosion by W a t e r — A Guide to Conservation Planning with the Watson, D . Α., Laflen, J. M., and Franti, T. G. (1986). Estimating
Revised Universal Soil Loss Equation R U S L E , " Publication A R S . ephemeral gully erosion, Paper No. 86-2020. American Society of
U.S. Department of Agriculture, Washington, DC, in press. Agricultural Engineers, St. Joseph, MI.
Rendon-Herrero, O. (1974). Estimation of washload produced on Williams, J. R. (1975). Sediment routing for agricultural watersheds.
certain small w a t e r s h e d s . Proc. A m . Soc. Civil E n g . Water Resources Bull. l l ( 5 ) : 9 6 5 - 9 7 4 .
100(HY7):835-848. Williams, J. R. (1976). Sediment yield prediction with Universal
Renfro, G. W. (1975). Use of erosion equations and sediment deliv­ Equation using runoff energy factor. In "Present and Prospective
ery ratios for predicting sediment yield. In "Present and Prospec­ Technology for Predicting Sediment Yields and Sources," Publi­
tive Technology for Predicting Sediment Yields and Sources," cation ARS-S-40. Agricultural Research Service, U.S. Depart­
Agricultural Research Service Publication ARS-S40, pp. 2 3 - 4 5 . ment of Agriculture, Washington, D C .
U.S. Department of Agriculture, Washington, D C . Williams, J. R. (1977). Sediment delivery ratios determined with
Rohlf, R. A. (1981). Development of a deterministic mathematical sediment and runoff models. In "Proceedings, Erosion and Solid
model for interrill and rill runoff and erosion, M.S. thesis, Uni­ Matter Transport in Inland Water Symposium," IAHS No. 122,
versity of Kentucky, Lexington, KY. pp. 168-179.
Rohlf, R. A. (1993). Groundwater flow and elastoplastic stress-strain Williams, J. R. (1978). A sediment yield routing model. In "Proceed­
modeling of channel bank failure. P h D dissertation, Department ings, American Society of Civil Engineers Conference," Verifica­
of Civil Engineering, University of Kentucky, Lexington. tion of mathematical and physical models in hydraulic engineer­
Romkens, M. J., Young, R. Α., Poesen, J. W., McCool, D. K., ing, pp. 6 6 2 - 6 7 0 . American Society of Civil Engineers, New
York.
ElSwaify, S. Α., and Bradford, J. M. (1993). Soil erodibility factor
Williams, J. R. (1978). A sediment graph model based on an instan­
K. "Predicting Soil Erosion by W a t e r — A Guide to Conservation
taneous sediment graph. Water Resources Res. 14(4):659-664.
Planning with the Revised Universal Soil Loss Equation R U S L E "
(Renard, Foster, and Weesies, eds.), Chap. 3, Publication A R S . Williams, J. R., and Brendt, A. D . (1972). Sediment yield computed
with U n i v e r s a l E q u a t i o n . Proc. A m . S o c . Civil Eng.
U.S. Department of Agriculture, Washington, D C , in press.
98(HY12):2087-2098.
Rose, C. W. (1960). Soil detachment caused by rainfall. Soil Sci.
89(0:28-35. Wilson, Β. N., and Haan, C. T. (1993). Development of a fundamen­
tally based detachment model. Trans. Am. Soc. Agr. Eng.
Schumm, S. Α., Harvey, M. D., and Watson, C. C. (1984). "Incised
36(4):1105-1114.
Channels: Morphology, Dynamics and Control." Water Re­
Wilson, Β. N., Barfield, B. J., and Moore, I. D . (1982). A hydrology
sources Publication, Littleton, CO.
and sedimentology watershed model. I. Modeling technique. D e ­
Shields, A. (1936). Application of the theory of similarity and turbu­
partment of Agricultural Engineering, University of Kentucky,
lence research research to the bed load movement. In "Anwen-
Lexington, KY.
dung der Aehnlichkeitsmechanik und der turbulenzforschung auf
Wilson, Β. N., Barfield, B. J., and Warner, R. C. (1986). Simple
die geschiebebewengung," Vol. 26, pp. 5 - 2 4 . Wasserbau der
models to evaluate non-point sources and controls. In "Agricul­
Preussischen Versuchsanstalt fur Wasserbau und Schiffbau, Heft,
tural Nonpoint Source Pollution: Model Selection and Applica­
Berlin. [A translation of this paper by Ott and van Uchelen is on
tion" (Giorgini and Zingales, eds.), pp 2 3 1 - 2 6 3 . Elsevier, New
file with the Engineering Societies Library]
York.
Smerdon, Ε. T., and Beasley, R. P. (1961). Critical tractive forces in
Wischmeier, W. H. (1975). Estimating the soil loss equations cover
cohesive soils. Agric. Eng. 42(l):26-29.
and management factor for undisturbed lands. In "Present and
Smith, D . D. (1941). Interpretation of soil conservation data for field
Prospective Technology for Predicting Sediment Yields and
use. Agric. Eng. 22:173-175.
Sources," U S D A - A R S Publication ARS-S-40, pp. 118-124. U.S.
Smith, D . D., and Whitt, D. M. (1948). Estimating soil losses from Department of Agriculture, Washington, D C .
field areas. Agric. Eng. 29:394-396. Wischmeier, W. H. (1976). U s e and misuse of the universal soil loss
Soil Conservation Service (1978). "Water Management and Sedi­ equation. J. Soil Water Conserv. 31:5-9.
ment Control for Urbanizing Areas." U.S. Department of Agri­ Wischmeier, W. H., and Smith, D . D . (1958). Rainfall energy and its
culture, Soil Conservation Service, Columbus, OH. relationship to soil loss. Trans. A m . Soc. Am. Geophys. Union
Soil Conservation Service (1983). "National Soils Handbook." Soil 39(2):285-291.
Conservation Service, U.S. Department of Agriculture, Washing­ Wischmeier, W. H., and Smith, D . D . (1965). Predicting rainfall-ero­
ton, D C . sion losses from cropland east of the Rocky Mountains—A guide
Stein, O. R., Roth, C. B., Moldenhauer, W. C , and Hahn, D . T. for selection of practices for soil and water conservation, Agricul­
(1983). Erodibility of selected Indiana reclaimed strip mine soils. tural Handbook N o . 282. U.S. Department of Agriculture, Wash­
In "Proceedings, 1983 Symposium on Surface Mining, Hydrology, ington, D C .
310 8. Erosion and Sediment Yield
Wischmeier, W. H., and Smith, D. D . (1978). Predicting rainfall Yoder, D . C , Porter, J. P., Loften, J. M , Simenton, J. R., Renard,
erosion l o s s e s — A guide to conservation planning, Agricultural K. G., McCool, D . K., and Foster, G. R. (1993). Cover manage­
Handbook No. 537. U.S. Department of Agriculture, Washing­ ment factor (c). In "Predicting Soil Erosion by W a t e r — A Guide
ton, D C . to Conservation Planning with the REvised Universal Soil Loss
Wischmeier, W. H., Johnson, C. B., and Cross, Β. V. (1971). A soil Equation ( R U S L E ) Publication A R S " (Renard, Foster, and
erodibility nomograph for farmland and construction sites. J. Soil Weesises, eds.), Chapter 5. U.S. Department of Agriculture,
Water Conserv. 26:189-193. Washington, D C .
Yalin, M. S. (1963). An expression for bed-load transportation. Proc. Young, R. Α., and Wiersma, J. L. (1973). T h e role of rainfall impact
Am. Soc. Civil Eng. 89(HY3):221-250. in soil detachment and transport. Water Resources Res.
Yang, C. T. (1972). Unit stream power and sediment transport. Proc. 9(6): 1629-1636.
Am. Soc. Civil Eng. 98(HY10): 1805-1826. Zingg, A. W. (1940). Degree and length of land slope as it affects soil
loss in runoff. Agric. Eng. 21:59-64.
Sediment Control Structures

Table 9.1 E r o s i o n and S e d i m e n t Controls


INTRODUCTION
On-site overland now controls Channel erosion controls
From an environmental standpoint, erosion preven­ Chiseling Vegetative controls
tion and on-site control are the most desirable strate­ Surface roughing Riprap
gies for controlling sediment pollution. Such strategies Vegetation Gabions
Mulching Energy dissipator
keep soil at the source, thus maintaining its value as a
Straw Drawdown tubes
natural resource and eliminating expensive cleanup Hay Culverts
requirements for sediment materials deposited off-site. Cellulose
Thus, on-site control is typically considered to b e the Fiberglass Small structures
first line of defense in any sediment control plan. Sawdust Filter fence
Hydromulch Straw bales
After exhausting on-site control as an option, the
Wood cellulose Level spreader
next line of defense is off-site controls. During the Wood chips Swirl Concentrator
massive land disturbances resulting from construction, Vegetative filter
surface mining, and some types of farming operation, Contouring Sediment traps
unprotected soil will invariably be exposed to the ero­ Strip crops Riprap outlets
Crop rotation Vegetated outlets
sive powers of rainfall and surface runoff in spite of
Conservation tillage Filter fabric
best control technologies. Some type of off-site control
technique is typically n e e d e d as an adjunct to on-site
Runoff diversions (swales) Sediment basins
controls; thus a well-designed sediment control plan
Vegetated Dewatered
should include both on- and off-site technologies. A Bare soil Permanent pond
summary of common control techniques is given in Riprap First flush filter
Table 9.1. Combination
Any evaluation of the effectiveness of sediment con­ Temporary lining Wetlands
trol techniques should be m a d e on a systems basis

311
312 9. Sediment Control Structures

rather than by an evaluation of individual components, Table 9.2 Summary of Measures of Performance of Sediment
because of the nonlinearity of the system. For example, Control Techniques
a sediment pond controlling runoff from a b a r e soil
1. Trapping efficiency. The easiest parameter to predict. Effective only
may be designed to t r a p 9 0 % of the sediment. After
in determining the amount of sediment trapped. Not a good measure
revegetation, however, the vegetation will tend to p r e ­ of the impact of a control structure on the environment.
vent erosion of the larger particles resulting in a finer
particle size distribution. T h e trapping efficiency may
2. Effluent concentration. Can be either peak or average storm effluent
drop drastically as a result of the change in inflow
concentration. Most difficult to predict. A small error in predicted
particle size caused by revegetation. As shown in a trapping efficiency can have a large impact on predicted effluent
later section, the reduction in trapping efficiency d u e concentration. Considers all particle sizes. Good measure of a struc­
to a finer particle size distribution can be demonstrated ture's effect on total turbidity.
theoretically. T h e incoming sediment load, however,
would also be reduced as a result of vegetation. T h e r e ­ 3. Settleable solids. Considers only those particles that would settle out
fore, the pond sediment effluent would be reduced, of an Imhoff cone in 1 hr; thus it generally is only those particles
even though the trapping efficiency is decreased by 0.01 mm and larger. Good measure of a structure's impact on those
particles likely to settle in downstream conveyance systems and
revegetation. This is illustrated in Example Problem
small reservoirs.
9.1. A conclusion from the above discussion is that
pond effectiveness cannot be evaluated in isolation
from a hydrologic and sedimentation analysis of the
entire watershed.

SEDIMENT DETENTION BASINS


Example Problem 9.1 Illustration of problems with
Introductory Comments
trapping efficiency
T h e most commonly used off-site control is a sedi­
Sediment yield in a given storm has been estimated to be m e n t detention basin (pond). Any size pond or reser­
200 tons for bare soil and 5 tons after vegetation is estab­ voir can serve as a sediment detention basin; however,
lished. The trapping efficiencies for the same pond are esti­ sediment detention basins are typically small struc­
mated to change from 90% for the bare soil to 70% for the tures. In general, basins that are designed primarily for
vegetated case as a result of size distribution changes. Esti­ sediment trapping have storage volumes that are equal
mate the sediment discharged from the pond in both cases.
to or less than the runoff volume in a design storm. In
Solution: Sediment discharge rate would be the product of that situation, thermal stratification and vertical mixing
incoming load times (1 - E), where Ε is trapping efficiency are not likely to be major problems. For larger reser­
(fraction). Results are tabulated below. voirs, such is not the case. T h e procedures presented in
this chapter apply primarily to t h e smaller reservoirs.
Incoming Trapping Sediment
Some authorities have issued regulations requiring
Watershed load efficiency outflow the sizing of sediment detention structures by a frac­
status (tons) (%) (tons) tion of watershed area disturbed, detention storage
time, size of storm to b e stored, or an effluent stan­
Bare soil 200 90 20.0 dard. T h e material p r e s e n t e d in this chapter is not
Revegetated 5 70 1.5 oriented toward any o n e design requirement. Instead,
procedures are p r e s e n t e d for evaluating the perfor­
m a n c e of a sediment control system regardless of the
design methodology. In addition, a discussion of the
Although the pond had a lower trapping efficiency when the
watershed was revegetated, the sediment discharge was less. limitations of some of the design methodologies is
This points out the difficulty of using trapping efficiency given.
alone as a design criteria. T h e major factors that control sediment transport
through a detention basin include:

• Physical characteristics of the sediment.


O n e can conclude from Example Problem 9.1 that • Hydraulic characteristics of the basin.
trapping efficiency alone is not an a d e q u a t e m e a s u r e of • Inflow sedimentgraph.
a sediment control system's performance. A summary • Inflow hydrograph.
of other descriptors of performance is given in • Basin geometry.
Table 9.2. • Chemistry of the water and sediment.
Sediment Detention Basins 313

A discussion of these p a r a m e t e r s is given in this sec­ Volume For Flood

tion, along with a discussion of t h e models m o r e com­


monly used for evaluating sediment trapping in ponds.
Finally, special design considerations are discussed. N o
attempt is m a d e to present information on structural
and geotechnical design.
Sediment detention basins can b e of t h e p e r m a n e n t
Storage
pool or self-dewatering type as shown in Fig. 9.1. In
(A)
any detention basin, a volume, ¥ is included for sedi­
s

ment storage. A storage volume, ¥ , is also provided


D

for detention storage to detain the design storm long


enough to allow a d e q u a t e sedimentation. In some cases,
additional p e r m a n e n t pool volume, V , is a d d e d above
p
Volume For Flood

the sediment storage volume as illustrated in Fig. 9.1C;


this additional volume protects stored sediment from
resuspension and provides for additional detention time
for first flush flow. A final flood storage volume, ¥ , is f

added for storage of sufficient water to prevent over­


topping and resulting dam failure during rare events.
Storage
Potential downstream d a m a g e will dictate the r e t u r n
(B)
period for which the flood storage volume is designed.
Procedures for determining ¥ , ¥ , and ¥ are intro­
s D f

duced in subsequent sections.

Factors Affecting Pond Performance Volume For Flood


Prevention
Emergency
T h e background information in this section is in­ Spillway
cluded to assist in understanding the process involved
in reservoir sedimentation and to assist in developing Volume To G i v e ^ .
Required D e t e n t i o n Impact
inputs to models of p o n d performance. T h e r e a d e r Time Basin
interested primarily in mathematical models of p o n d
Additional
performance might wish to read the section on model­ Permanent
Pool S t o r a g e Sediment
ing pond performance prior to reading this section.
Storage

Particle Size Distribution and Pond Performance


(c)
Particle size distribution effects on trapping effi­
ciency of a reservoir are reflected primarily by particle
fall velocities or settling velocities. T h e importance of
settling velocity to reservoir trap efficiency can be illus­
trated by a simple model. If steady-state flow is as­ Volume For Flood
Prevention
sumed in a rectangular channel, the distance required
Rock Fill
for settling of 1 ft can be calculated for varying particle Structure
sizes as shown in Fig. 9.2. Obviously, t h e smaller t h e
Volume To Give
particle, the longer the flow p a t h required through the Required D e t e n t i o n
Time
reservoir to trap the sediment. T h e importance of
particle size and particle size distribution is illustrated Sediment
Storage
in Example Problem 9.2 for a simple rectangular reser­
voir with constant flow rate. (D)

Figure 9.1 Illustration of different types of sediment detention


structures. ( A ) Self-dewatering-type sediment detention basin with
Example Problem 9.2. Simple illustration inlet to principal spillway located at top of sediment storage. A
of impact of particle size on pond trapping trickle tube will display similar hydraulic characteristics. (B) Self-de­
watering-type sediment detention structure with slotted principal
spillway. (C) Permanent pool-type sediment detention structure illus­
Water is being pumped into a sediment detention reser­ trating the additional storage volume. ( D ) Rock fill outlet type
voir at a constant rate resulting in an average flow through sediment detention structure.
314 9. Sediment Control Structures

m e n t is a major factor determining t r a p efficiency.


Thus, any sediment p o n d model or design procedure
should consider particle sizes of sediment reaching the
reservoir, and model users should emphasize accurate
estimates of inflow particle size distributions.
A n o t h e r p a r a m e t e r of importance to trapping effi­
ciency is flow velocity. By changing the flow velocity to
0.1 f t / s e c , the distance changes in Example Problem
9.2 from 200,000 ft to 20,000 ft and 2000 ft to 200 ft.
T h e effect of flow velocity is reflected in detention
FC
1X10 " Β
1X10" 5
1X10" 4
1Χ10" 3
1Χ10" 2
1X10" 1

storage time and overflow rate, two concepts discussed


in a subsequent section.
S e t t l i n g Velocity (ft/sec) In addition to particle size and flow velocity, particle
Figure 9.2 Pond length for quartz particles to settle 1 ft at various shape, particle density, turbulence levels, and sediment
forward velocities, using Stokes' law and assuming SG = 2.65 (after concentration have influences on settling velocities. A
Bondurant et al., 1975).
discussion of the effect of these p a r a m e t e r s on settling
velocity is given in C h a p t e r 7.

velocity of 1.0 ft/sec. The inlet to the outflow riser is situated Pond Hydraulic Response and Reactor Models
so that the average settling depth is 2 ft. Estimate the flow Hydraulic characteristics of a p o n d are represented
distance required to trap 90% of the two sediments shown by a variety of p a r a m e t e r s including detention storage
below.
time, d e a d storage, and short-circuiting. Typically, hy­
draulic effects are defined by either a hydrodynamic
Sediment A Sediment Β model or a reactor model that desribes mixing pro­
Particle size Particle size cesses. Hydrodynamic models utilize conservation of
(mm) % Finer (mm) % Finer mass or m o m e n t u m principles to derive partial differ­
ential equations that are usually difficult to solve. R e ­
0.002 10 0.02 10 actor theory models divide the p o n d into conceptual
0.02 50 0.2 50 chambers or reactors in which complete mixing, plug
0.2 100 2.0 100 flow, or a combination of mixing processes is assumed.
Alternately, some of the material may b e assumed to
bypass t h e p o n d entirely. Equations defining the hy­
Solution: Figure 9.2 can be entered either with fall velocity draulic response of t h e p o n d are derived by conducting
or equivalent settling diameter. We assume a SG of 2.65 and a mass balance on each reactor to account for the time
spherical particles and enter the figure with particle size. As variation of concentration of any tracer. Because of
a rough approximation, it is assumed that 90% or more of their relative simplicity, only reactor theory models are
the sediment will be trapped when the particle size corre­ discussed in this chapter.
sponding to 10% finer settles from the surface to the bottom.
In the following discussions, models defining the
To settle 2 ft, the flow path will be twice that for 1 ft.
hydraulic mixing of a p o n d are developed on the basis
For sediment A, the D is 0.002 mm (2 μτή). This is the
of the assumption of a tracer with n o settling velocity.
l0

upper limit of the clay size particles. The flow path to settle 1
ft is 100,000 ft, so the flow path to settle 2 ft is 200,000 ft. T h e principles d e v e l o p e d a r e t h e n applied to
For sediment B, the D is 0.02 mm (20 μηι). This the
10
sediment-laden flows in a subsequent section. Reactor
upper limit of the silt size particles. The flow path to settle 1 model discussion is p r e s e n t e d for two reasons:
ft is 1000 ft, so the flow path to settle 2 ft is 2000 ft.
It should be noted that the rough estimate of trapping • to develop background information for p o n d models,
efficiency used here is for steady flow rates and only accounts and
for trapping of those particles that could settle all the way • to illustrate numerical p r o c e d u r e s for analyzing tracer
from the surface to the bottom of the reservoir in the flow studies of p o n d mixing processes.
through time. A portion of particles that settle only a fraction
of that distance will also be trapped. This effect on trapping R e a c t o r models illustrated in this section include:
efficiency is considered in a later section. • single continuous stirred tank reactor
• continuous stirred tank reactors in series (CSTRS)
• plug flow reactors
It is obvious that size distribution of inflowing sedi­ • combinations of reactors (hybrid).
Sediment Detention Basins 315

or
dC
Rate of Change of
t
i' Q
A
(9.4)
Mass Rate In Mass Within the Reactor Mass Rate Out = I — dt.
QC0 0
h V

For an initial concentration of zero and a constant


inflow concentration C starting at time zero, Eq. (9.4)
0

can b e integrated t o

C0 C1 Q
In —t (9.5)
V
or
Cl
Q
Co In 1 - (9.6)

T h e equation can be simplified by the concept of a


theoretical d e t e n t i o n time, defined as
J
td
i d = V/β, (9.7)
Figure 9.3 A single continuous stirred reactor. ( A ) Single reactor
mass balance, (B) Outflow concentration for a single reactor with w h e r e t is t h e time required for a flow of Q to
d

continuous influent concentration.


completely displace a reactor volume ¥ assuming no
mixing. Using this definition of theoretical detention
time, Eq. (9.6) becomes
A detailed derivation of the single continuous stirred
reactor is given. Final equations are p r e s e n t e d for (9.8)
other reactors. Steady-state flows are assumed in devel­
o p m e n t of the reactor models.
w h e r e C /C
x 0 is defined as F(t) to d e n o t e that C /C { Q

Single Continuous Stirred Tank Reactor Using steady- is a function of t. It should b e understood that F is
state assumptions, a mass balance can b e conducted on dimensionless a n d is a function of the dimensionless
the completely mixed continuous stirred reactor shown time ratio t/t , but is written as F(t) for simplicity.
d

in Fig. 9.3. Assuming a tracer of concentration C 0


T h e effluent concentration curve shown in Fig. 9.3
introduced continuously into the reactor with a con­ corresponds to E q . (9.8). It should be noted that an
stant inflow and outflow volumetric flow rate of β , the assumption of idealized flow, i.e., n o d e a d storage or
mass balance can b e written as short-circulating, was m a d e in t h e development. These
concepts are discussed subsequently.
Mass Mass R a t e of change A n alternate m e t h o d to continuous injection for
rate rate = of mass in (9.1)
- defining mixing is the use of an instantaneous slug
in out the reactor injection of dye at time zero. T h e resulting effluent
or concentration curve defines what is known as a resi­
QC 0 - QC, = d¥C /dt, x (9.2) d e n c e time distribution ( R T D ) , which is essentially a
probability distribution for residence time in a reactor.
where C is concentration of tracer in the reactor a n d
x
Since inflow concentration is zero for all times greater
¥ is reactor volume. In a continuous stirred reactor,
t h a n zero for an instantaneous slug injection, t h e mass
the assumption that the reactor volume is completely
balance given by E q . (9.1) will become
mixed is typically m a d e . This m e a n s that t h e outflow
concentration is equal to the concentration in t h e d*C,
reactor, as shown in Eq. (9.2). -QC { = (9.9)
dt
Using the assumption of steady state, i.e., Q and V
are constant, Eq. (9.2) becomes or
Q dC x

(9.3) (9.10)
? <c„-c,)-^i
316 9. Sediment Control Structures

where the initial concentration, C is given by


x
A special relationship exists between E(t) and F(t).
F r o m Eqs. (9.8) and (9.14), it can be seen that
Cj = M / ¥ (9.11)

and Μ is the mass of tracer injected in the slug. T h u s , F(t)= f'E(t)dt. (9.15)
Eq. (9.10) becomes

(9.12) T h u s , E(t) is the probability density function and F(t)


is t h e cumulative distribution function for residence
T o be a probability distribution, the relationship must time. Specifically, the probability distribution for a
integrate to 1.0; thus, Eq. (9.12) must be modified to completely mixed reactor is an exponential distribution
with p a r a m e t e r l / / . A further discussion of t h e R T D
d

concept is given in Levenspiel (1972).


-'/'ο (9.13)
Series of Continuous Stirred Tank Reactors (CSTRS) In
or the development of Eqs. (9.8) and (9.14) for the single
continuous stirred reactor, it was assumed that the
1 tracer was instantly mixed throughout the reactor. O b ­
£(0 = = -e-'/'a, (9.14)
viously, a time lag is n e e d e d to allow the introduced
M/Q td

tracer to reach the outlet. This can b e accounted for by


where Μ is the mass in the injector slug and C /M/Q l
introducing t h e concept of a series of continuous stirred
is defined as E(t). Thus, a plot of E(t) versus t/t is a d
reactors as shown in Fig. 9.4. T h e inflow concentration
probability distribution known as the R T D , which d e ­ for a given reactor is t h e effluent from the previous
fines the probability distribution of residence times for reactor. By conducting a mass balance in t h e reactors,
any fluid particle entering a reactor. it can be shown (Levenspiel, 1972) that the Fit) and

1st 2nd I th - - - - n th

• > 1 .

I
• Qi • Q _ 2 Qi-1 „ Qn-1.. Qn
.c . Cn-1 C n

·• i 2

/ / / / J/ / 7777//;
Sediment Detention Basins 317

E(t) distribution can be given by

C /
F ( 0 - t t = l-exp - -
«<\'
n

Σ
nt i-l i Plug
Flow
Completely Mixed

•Q'
O-P1X

(9.16)
P,V
and
i-l Figure 9.6 A plug flow and CSTR in series (after Wilson et al.,
n\ I nt nt
£ ( 0 = = I — exp - — 1982). P,, Portion of the total pond volume that is plug flow.
M/Q t )(n-l)\\t
d d

(9.17)

where η is the n u m b e r of reactors, and zero factorial is flow to displace the reactor volume. This time, defined
defined as one. A comparison of the R T D s for varying earlier as t h e theoretical detention time, is given by
numbers of reactors is given in Fig. 9.4. As can be seen, t = ¥ / Q . T h e process is illustrated in Fig. 9.5. M a t h e ­
a

the C S T R S model can be used to predict a reservoir's matically, the effluent concentration for a plug flow
mixing responses, ranging from a completely mixed reactor would b e defined by
reactor to a nonmixed plug flow system discussed be­ nO = Ci/c 0 = o, t<t d

low. T h e n u m b e r of reactors that should be used is


F ( f ) - Q / C o - 1.0, t >t . d

discussed in a subsequent section on estimating reactor


parameters. F r o m this discussion, it can be seen that an obvious
advantage of the C S T R S model is that it can be used
Plug Flow Reactor As the n u m b e r of reactors in a to evaluate mixing as well as time delay effects of plug
C S T R S system approaches infinity, the volume of each flow.
reactor (Vj = V / a z ) approaches zero. In this case, the
inflow tracer is not mixed with any of t h e reactor Hybrid Reactors A n alternate to the C S T R S model
volume, thus yielding a tracer concentration curve as is the use of hybrid reactor models, in which pure
shown in Fig. 9.4 for η of infinity. Therefore, if an mixing reactors and plug flow reactors are combined in
inflow tracer is added to a plug flow reactor, the parallel or put in series as shown in Fig. 9.6. O t h e r
outflow concentration will equal the inflow concentra­ possible combinations exist. For the system shown in
tion after a time delay required for the steady-state Fig. 9.6, t h e resulting steady-state effluent concentra­
tion curves would b e given by
F(f) - C , =0, 0<t<(l-P )t
x d

Inflow Concentration Water Level


C\
ι r ^ - r -'Λ
Permanent
0 (9.19)
Co fi

Co ι5 CoZ Cq 3
Pool = ΛΛ
// Outlet
= 1 - exp
-[t-(l- P l )t ]d

?
I
1

where P is the reactor fraction attributed to the


{

completely mixed reactor and (1 - P ) is the fraction }

assigned to a plug flow reactor. T h e value for P would x


Β
n e e d to be developed from tracer studies or experi­
Influent Effluent
Concentration Concentration ence, as discussed subsequently. In Eq. (9.19), the dead
storage space is assumed to b e zero.

Dead Storage and Short-Circuiting in Reactor Models


Co, Co 4
Cos Co,
Co, D e a d storage and short-circuiting are two concepts
used to explain nonideal behavior of a reactor. Appli­
cation of these concepts is illustrated with a single
Time Time continuous stirred reactor as illustrated in Fig. 9.7.
Figure 9.5 Plug flow reactor (after Wilson et al., 1982). ( A ) Con­
Equations for o t h e r reactors are summarized.
ceptualized view of a plug flow reactor. (B) Influent concentration Conceptually, short-circuiting is viewed as that por­
and resulting effluent concentration for a plug flow reactor. tion of t h e flow that bypasses the reactor volume
318 9. Sediment Control Structures

entirely, thus short-circuiting directly to t h e outlet. T h e fractions f and f would n e e d t o b e estimated


x 2

D e a d storage is conceptualized as that portion of t h e from dye tracer studies or experience. Equations for
reactor that does not mix with the inflow. Using t h e o t h e r reactor models a r e given in Table 9 . 3 .
definitions given in Fig. 9 . 7 , f is the fraction of t h e
x

flow that moves through the reactor while ( 1 - f ) x

short-circuits a n d f is t h e fraction of t h e reactor that


2
Estimating Parameters for Reactor Models Parameter
is active while ( 1 - f ) is dead storage. Using these
estimation must be based on experience, previous anal­
2

definitions a n d a mass balance; t h e effluent from t h e


ysis, or dye tracer studies. Dye tracer studies a r e con­
active portion of the reactor becomes
ducted by injecting a tracer, such as a fluorescent dye
of known concentration, into t h e inlet a n d observing its
concentration at t h e outlet. F r o m t h e observed concen­
£-*--(-££)· (9 20)
· tration curve, p a r a m e t e r s a r e estimated by obtaining a
best fit of t h e data to t h e particular model selected.
and t h e total concentration C from t h e combined Obviously, t h e model selected will have an effect o n
T

short-circuiting and active flow is given by t h e value of t h e best fit p a r a m e t e r . T h u s , if p a r a m e t e r s


are being d e t e r m i n e d for use in a p o n d sedimentation
analysis, it is important that t h e reactor model used in
p a r a m e t e r estimation b e t h e same as that which forms
-'"*(-££)· <9
· 21)
the basis for t h e sediment or water quality model.

Table 9.3 Equations for Reactor Models with Dead Storage and Short Circuiting*

Single continuous stirred


reactor

Continuous stirred reactors


in series g"-*«p("U)
Co ~ " ™ V " h u) J l r
(9.23)

iti C ' - D ! V fi tj

Plug How reactor Qr = ( 1 _ / ( ) , K A u <· )


9 24

Co

&=1, t < h u ( 9 2 5
>
Co

Plug now-continuous CT = ( , _ F ) F O R , < , l . 4 />, , d

stirred reactor Co f\ (9.26)

g - " 4 f e a ^ T ^ ) ] " > " (-»

Diffusion plug flow 1 Γ - (1 - NF ) 1 2

reactor ( l )
" (nN N )
2 F D
05
^ L WFND) J <> 9
2
8

NF=QL ± = (9.29)
V t d

N =£-D (9.30)
UL

^Definition of terms: C ; total concentration of short circuited flow plus reactor flow; C , inflow
T 0

concentration; t, time; t , theoretical detention time = VIQ\ V, reactor volume; f , fraction of flow
d x

going to reactor; f , fraction of reactor that is active volume; l - / , fraction of reactor that is dead
2 2

storage; Ρ , fraction of hybrid reactor that is plug flow; D , turbulent diffusivity; U , mean flow
χ

velocity; L, length of the reactor.


Sediment Detention Basins 319

Time a
E(t) =
(min) (ppm) W = C,/C 0
dFldt l-C./Co

0.0 0.000 0.00 0.00 0.0029 1.00


3.0 0.006 0.16 0.01 0.0392 0.99
6.0 0.160 0.31 0.24 0.0686 0.76
9.0 0.286 0.47 0.42 0.0485 0.58
12.0 0.358 0.62 0.53 0.0304 0.47
15.0 0.410 0.78 0.60 0.0189 0.40
18.0 0.435 0.93 0.64 0.0208 0.36
21.0 0.495 1.09 0.73 0.0297 0.27
24.0 0.556 1.24 0.82 0.0238 0.18
27.0 0.592 1.40 0.87 0.0088 0.13
30.0 0.592 1.55 0.87 0.0039 0.13
33.0 0.608 1.71 0.89 0.0108 0.11
36.0 0.636 1.87 0.94 0.0108 0.06
39.0 0.652 2.02 0.96 0.0059 0.04
42.0 0.660 2.18 0.97 0.0039 0.03

T o r all except first and last values, E = (F


i ( + l - F _ ,)/2Δ/ = {F , - F j)/6
i i + t min.

T h e plug flow model requires somewhat different was 0.68 ppm, the theoretical detention time was 19.3 min,
estimating procedures. Assuming no short-circuiting, and the steady-state flow rate was 38.1 liters/min. Estimate
Griffin et al. (1985) proposed that dead storage for the the best-fit parameters for the CSTRS model, plug flow
plug flow model could be estimated from model, and plug flow diffusion model.
Solution: Note: The solutions to this problem were devel­
D S = 1 - t /t g d (9.31) oped with a spreadsheet, and results are rounded for inclu­
sion in tables. Values developed with a calculator will not
where t is the time to center of mass (centroid) of the
g
agree precisely.
E(t) curve. M e t h o d s for estimating t are given in g
1. CSTRS model. The model to be used is Eq. (9.23) in
Example Problem 9.3. Table 9.3. A direct solution is available only for an assump­
A n additional relationship for evaluating nonideal tion of one reactor. The problem here is to find both the
behavior is known as the plug-flow-diffusion model, dead storage and the number of reactors. To do this, Eq.
which combines axial diffusion with plug flow concepts. (9.23) is solved for all values of time in the data set and the
T h e E(t) R T D for the plug flow diffusion model is difference between observed and predicted values calculated.
given by Eqs. (9.28)-(9.30) in Table 9.3. Assuming n o The sums of squared differences defined by
dead storage or short-circuiting, Levenspiel and Smith m

(1957) showed that the term Ν in Eq. (9.30) could be


Ό D E V = Ε 0og O -\ogP f
x x

estimated from
are calculated where O and P are observed and predicted
x x

Ν Ό = £ ( γ / 8 σ + 1 - 1),
2
(9.32) values of (1 - C / C ) for a given value of t/t and m is the
T 0 d

number of observations available. Following Wilson et al.


where σ is the standard deviation of the E(t) curve. (1982), observed and predicted values are log transformed to
Procedures for calculation of p a r a m e t e r s from dye account for the exponential nature of Eq. (9.23). After mak­
tracer tests are illustrated in Example Problem 9.3. ing the computations, the values of DEV are tabulated in the
Suggestions for parameters in the absence of such data table below for varying values and dead storage fraction,
are given in a subsequent section. (1 - / ) , and number of reactors, n. The minimum least-
2

squares value occurs at


η = 2
1 - / = 15%.
Example Problem 9.3. Calibration of reactor
2

models and estimating dead storage These would be the best-fit values for the parameters. Obvi­
ously, if the log transformation had not been made, the
Results from a continuous injection dye test on a labora­ results would have been different. Values in the table below
tory pond are given below. The constant inflow concentration were calculated with base 10 logarithms.
320 9. Sediment Control Structures

Dead
s
P a c e
Number of reactors (n)

(%) 1 2 3 4 5

0.0 0.985 0.529 0.271 0.223 0.376


5.0 0.788 0.317 0.139 0.250 0.631
10.0 0.599 0.149 0.117 0.473 1.193
15.0 0.425 0.052 0.254 0.979 2.195
20.0 0.276 0.060 0.624 1.896 3.829
25.0 0.168 0.224 1.337 3.406 6.375
30.0 0.124 0.621 2.553 5.785 10.250
35.0 0.176 1.366 4.513 9.446 16.082

2. Plug flow model. Dead storage for the plug flow model and
will be calculated from Eq. (9.31) or

t = t*Q'dF
g = t t' .
d g

DS = 1 - t /t ,
g d

From Fig. 9.8, it can be seen that t' can be estimated from g

where t is the centroid of the area beneath the EU) curve.


g

By definition, the centroid of the HO is given by


f.-ft'dF- f[\-F(t')\df

CtE(t)dt Ct(dF/dt)dt
h = Λ)
= CtdF. = EIi-WlAi'-Ai'EIi-Wl.
*g -OO in
/ E(t)dt
For this case

As illustrated in Fig. 9.8, let t' be given by


Μ f 3 min
Δί' = Δ | - | = _ _ _ = 0.155
19.3 min
t' = t/t .
d

Hence and

Short Circuiting = ( 1 - f , ) Q Q

Co

P o n d Inlet
Qo
•|
fiQo V,
' Completely Mixed

2
'
I
Y, f , Q o Qo m
P o n d Outlet
Co Co ^ c T

1/
Dead Space =

/977777777777Z77777ZV//.
(1-f )"Y
2

1
Figure 9.7 A single CSTR with short-circuiting and dead space (after Wilson et al., 1982). / „ fraction
of discharge entering CSTR. / , fraction of completely mixed reactor that is active volume.
2
Sediment Detention Basins 321

Values for F(t') and [1 - F(t')] are tabulated below. Using Eq. (9.31),
Computation of F{V) and 1 -F(f)
for Example Problem 9.3 t a 16.48 min
DS = 1 - — = 1 —
F(0 t d 19.3 min
DS = 0.146.
0.00 0.00 1.00
0.16 0.01 0.99 In this case, dead storage is 0.15 for the plug flow model as
0.31 0.24 0.76 well as for the CSTRS model. This will not always be true.
3. Plug flow-diffusion model. The task here is to estimate
0.47 0.42 0.58
the axial dispersion numbers, N and N . Estimating N D F D
0.62 0.53 0.47
from Eq. (9.32),
0.78 0.60 0.40
0.93 0.64 0.36 N D = | ( ) / 8 σ + 1 - 1),
2

1.09 0.73 0.27


1.24 0.82 0.18 where σ 2
is the variance of the E(t) curve, given by
1.40 0.87 0.13
1.55 0.87 0.13
1.71 0.89 0.11
m
1.87 0.94 0.06
= E£,[','-'U] A<\ 2

2.02 0.96 0.04 ι = 1

2.18 0.97 0.03


where t' = t/t and t'^ is the value of t/t
d at which
% d

Σ = 5.51 F(t/t ) = 0.5. Using the data set, t'^ can be found by
d

interpolation, or

*'g = ' ' Σ t - HO]


Δ 1
= (0.155)(5.51) = 0.854
i = /;xi
g d = 0.854 X 19.3 min = 16.48 min.
Calculations are summarized below.

Average over time interval

£(',') =
AF/At' a

c T
F{t;)
*/ '7
0.0000 0.000 0.0 0.057 0.407 0.078 0.1002
0.1554 0.006 0.01 0.757 1.041 0.233 0.1246
0.3109 0.160 0.24 1.324 1.130 0.389 0.0408
0.4666 0.286 0.42 0.937 0.762 0.544 0.0009
0.6218 0.358 0.53 0.587 0.475 0.699 0.0068
0.7772 0.410 0.60 0.364 0.383 0.855 0.0292
0.9326 0.435 0.64 0.402 0.487 1.010 0.0905
1.0880 0.495 0.73 0.572 0.516 1.166 0.1778
1.2435 0.556 0.82 0.459 0.315 1.321 0.1734
1.3990 0.592 0.87 0.170 0.123 1.477 0.0992
1.5544 0.592 0.87 0.076 0.142 1.632 0.1575
1.7098 0.608 0.89 0.208 0.208 1.788 0.3040
1.8652 0.636 0.94 0.208 0.161 1.943 0.2995
2.0207 0.652 0.96 0.114 0.095 2.098 0.2192

2.1762 0.660 0.97 0.076 Σ = 6.244 Σ = 1.826

'Based on Eq. (9.15), £(/.') = [F(t\ ,) - F{t\ _ , ) ] / 2 Δ Λ Calculations


+

were made with a spreadsheet with F(t\) = C / C and not from rounded
T 0

values in the third column.


Ej = (E
b
i + , + E )l2 = average over time interval.
t

'>',. = (f, + 1 + /',.)/2-


322 9. Sediment Control Structures

Plug flow DS = 0.14; f 2 = 0.86,


F(t) = 0, t/td < 0.86
F(t) = 1.0, t/t d > 0.86.

fa Plug flow diffusion N D = 0.101; N = 0.052/,


F

Ο
Ηυϋ Ο
1 -(1-^VF) 2

£ ( 0 = exp
2 /TTW W
; f d 4N N
F O

- ( 1 - 0.052/)
t/t =f exp
d
2^(0.0520(0.101) 4(0.101)(0.052O
Figure 9.8 Illustration of procedures for calculating t' r

- ( 1 - 0.052O 2

= 3.89r exp 05

0.0210i

Values are plotted in Fig. 9.9. Based on visual observation


Check on E distribution
of this data set, the CSTRS model provided the best fit to the
t

data.
Σ Ε, Δί' = (6.244) (^j)= 0.97 s 1.0.

Calculating σ , 2
Circulation Patterns and Reactor Models. In the reac­
tor models p r e s e n t e d in Table 9.3, d e a d storage is
σ 2
= Σ(£,)(<;-<;ν ) Δ'' 8
2 visualized as p o n d volume that is bypassed entirely by
incoming flow. Such is not t h e case. A n example of a
( 3 min \
.
19.3 min /
= 0.284 circulation p a t t e r n in a p o n d with large dead storage is
given in Fig. 9.10. Studies by Griffin et al. (1985) and
σ - 0.532 by N o e and Barfield (1990) show that a significant

= έ()/8σ 2
+ 1 - 1) = £(^8(0.284) + 1 - l )
portion of t h e p o n d flow bypasses the outlet and forms
a recirculation pattern, moving into the area typically
N D = 0.101. considered to b e d e a d storage. This volume in the
recirculation p a t t e r n is not as effective in the sedimen­
Estimating N F from Eq. (9.29), tation process because it it initially bypassed by the
inflow. A t any point in time, t h e flow being discharged
t t
from a p o n d is a combination of flow that moves
directly across the p o n d and that which has recircu­
lated.
= 0.052*.
19.3 Detention Storage Time and Pond Performance

4. Summary. Using the data supplied, three reactor mod­ D e t e n t i o n storage time, which is a measure of flow
els have been calibrated to the RTD. For all models the through time in a structure, was shown in Example
assumption of no short-circuiting, f — 1.0, was made. Putting x Problem 9.2 t o have a significant impact o n trapping
the calibrated parameters into the reactor models, the follow­ efficiency. F o r steady-state systems, detention storage
ing equations were obtained: time was defined by Eq. (9.7) as

CSTRS: η = 2; (1 - f ) 2 = 0.15 or f2 = 0.85 = V/G. (9.7)


Physically, it is the time required for a given flow, Q, to
ι fx * \ A 1 displace t h e stored water ¥ .
f 2 t d ) ^ o - ^ \ n
f 2 ^ For nonsteady flows typical of stormwater, the defi­
1 nition of detention storage time is not quite so simple.
=ι 1.0 exp ( - 2 )
(0.85) (19.3)
D e t e n t i o n storage time for nonsteady flows, as dis­
cussed in C h a p t e r 6, is t h e average time that a given
flow resides in a p o n d . T h e concept of detention time
i( J__L) 2
0
+ if JL_L)2
1

for nonsteady flows using the plug flow concept is


0! \ 0.85 19.3 / 1! \ 0.85 19.3 / illustrated in Fig. 9.11 for a reservoir with and without
d e a d storage. Example computations of plug flow are
= 1 - [ e x p ( - 0 . 1 2 2 i ) ] [ l + 0.122r].
Sediment Detention Basins 323

t/t d =f

Figure 9.9 Predicted and observed values for F(t') for three reactor models in Example Problem 9.3. Values of
F for the diffusion plug flow model were determined by numerical integration of the £"(/) equation.

given later in Example Problem 9.7. Alternatively, de­


tention storage time for reservoirs that are completely
dewatered can be estimated by the time difference
between centers of mass of t h e inflow and outflow
Inlet hydrographs as illustrated in C h a p t e r 6. For reservoirs
with a p e r m a n e n t pool, such estimates from centers of
mass can be misleading. W a t e r stored in the perma­
n e n t pool will have a detention time equal to the time
since the last storm. Much of this water will be dis­
placed by the incoming flow, giving a detention time
greater than that estimated from hydrograph centroids
or from t h e average detention of a plug, as illustrated
in Fig. 9.1 I B .
O n c e a detention time is assigned to the p e r m a n e n t
pool volume, the total storm detention time, using the
centroid concept, becomes

r [(V -
d p V )r + (V-V + V )(r - 7 )]/¥
ds dp p ds mo mi

(9.33)

where ¥ is the volume in the p e r m a n e n t pool, ¥


p is d s

d e a d storage, T is the detention time assigned to the


dp

p e r m a n e n t pool, V is the storm runoff volume, and


T mo - T is the time between the center of mass of
mi

the inflow and outflow hydrographs.


Figure 9.10 Example circulation pattern for a model pond with a
T h e problem with using the centroid concept for
large dead storage estimate from dye tracer tests (after Griffin et al.
t
computing detention time for ponds with p e r m a n e n t
1985). pools is that of assigning an acceptable detention time
324 9. Sediment Control Structures
between the center of mass of the inflow and outflow hydro-
graphs. Assume V = 0.25V . d s p

Solution: Calculating permanent pool and dead storage


volumes.
V p = (0.25)(2.4) = 0.6 in.
V d s = (0.25)(0.6) = 0.15 in.

Solving for T m o - T . From Eq. (9.33), using 7


mi d p = 2.5
(Γ - Τ )
v
* mo 1
mi

YT d = (2.4)(24) = (0.6 - 0.15)2.5(7™ - T)mi

+ (2.4-0.6 + 0.15)(7 m o -7 m i ).

Hence
Β
Tmo ~T = 1 8 . 7 hr. mi

Using the procedures discussed in Chapter 6, a pond could


be designed to give a required detention time of 18.7 hr.

Ι \ ι—Permanent Pool
/ \ / Dead Storage
Although detention storage time is frequently pro­
posed as a design criteria, it should be pointed out that
it has severe limitations as a design p a r a m e t e r . Particle
size distribution, as shown by Example Problem 9.2,
A 2

has a major effect on trapping efficiency. Even when


evaluating relative trapping efficiency within a given
|——Tdp2— size distribution, McBurnie et al. (1990) showed that
trapping efficiency varied widely with surface area for a
Figure 9.11 Illustration of plug flow detention time concept in a given detention time, as shown in Fig. 9.12.
reservoir both with ( A ) and without (B) permanent-pool storage.
Reservoir Shape and Dead Storage
Reservoir s h a p e has a major influence on how effec­
to the water displaced from the p e r m a n e n t pool. Stud­ tively t h e p o n d volume is utilized in sedimentation. It is
ies using a computer model of reservoir sedimentology frequently assumed that some areas of a pond, referred
known as D E P O S I T S (Ward et al, 1977, 1979) indi­ to as d e a d storage, are bypassed and are therefore
cate that a conservative value is 2.5 to 2.7 times T — totally ineffective in the settling process. Classic exam­
mo

T . This value is actually the detention time of t h e last ples of reservoir shapes that have large dead storage
mi

plug of outflow in a storm. volumes are shown in Fig. 9.13.


T o circumvent the problem of stored permanent-pool T o minimize d e a d storage, the Environmental Pro­
water versus stormwater the Environmental Protection tection Agency (1976) r e c o m m e n d s that the ratio of
Agency (Driscoll et al., 1986) has proposed a model average length of flow p a t h to t h e effective width of the
that evaluates trapping separately for stormwater and reservoir ( L : W ratio = L/W ) be greater than 2.0. W
e e

permanent-pool water. T h e E P A model is discussed in is an effective width calculated from


a later section. (9.34)
K=A/L,
w h e r e A is t h e surface area of t h e reservoir. T h e
effective length of the reservoir can be increased by
Example Problem 9.4. Plug flow computation
installing baffles in the structure as shown in Fig. 9.14.
of theoretical detention time accounting
Estimation of d e a d storage volume is difficult at best.
for permanent pool
A s discussed earlier, dye tracer studies are typically
A reservoir is being designed for a detention time of 24 hr. used in t h e analysis. Since d e a d storage zones can
The detention time assigned to the permanent pool is occur in the vertical as well as horizontal plane, three-
2.5 ( T - T ). The runoff volume for a 10-year, 24-hr storm
mo mi
dimensional evaluations are n e e d e d to m a k e visual
is 2.4 watershed in. A permanent pool is to be constructed to observations of d e a d storage. Visual observations of
contain 25% of the runoff volume. What is the required time t h e tracers in two dimensions can be m a d e , but the
Sediment Detention Basins 325

Line sfc Area (ac)

ο ^
CO ^ 10000 L

•S 1
Pond Volume - .967 a c - f t

S ν

3 w

1000L
Η a
•&
I
100

8 10 12 14 16

Detention Time (hr)


Figure 9.12 Effects of detention time on effluent TSS for a hypothetical reservoir in Maryland's Coastal Plains.
Predicted values are from simulations using S E D I M O T II with ponds of varying surface area but the same volume
(after McBurnie et al., 1990).

three-dimensional picture of flow is not available with While there may be n o movement of sediment or
any reasonable expenditure of effort. Because of this tracer into dead storage d u e to the m e a n flows, there is
problem, dead storage estimates are typically m a d e a diffusion of sediment or tracer into and out of dead
from measured dye concentrations at the pond outlet, storage zones d u e to turbulent eddies and recirculation
resulting either from a slug of dye or continuous injec­ flows that may occur. T h e s e nonideal flow p h e n o m e n a
tion of dye placed at the pond inlet. If an appropriate tend to m a k e t h e dead storage volume part of the
reactor model is used, as discussed earlier, these dye active volume of the reservoir. Diffusion and recirculat­
concentrations can be translated into a d e a d storage ing flow may or may not be accounted for, depending
volume. on the model a d o p t e d to interpret the tracer study.
Dye injection studies in a model sediment pond were
used by Griffin et al. (1985) to estimate dead storage Reservoir Type and Pond Performance
using plug flow and C S T R S models. T h e results are Reservoirs are typically classified as permanent-pool
shown in Fig. 9.15. reservoirs and reservoirs without p e r m a n e n t pool. A
T h e results shown in Fig. 9.15 clearly show that d e a d p e r m a n e n t - p o o l reservoir, as shown schematically in
storage depends on the length-to-width ratios and not Fig. 9.1, is o n e in which a p e r m a n e n t pool of water is
inflow momentum. In general, short ponds with length- kept below the crest of the principal spillway. Theoreti­
to-width ratios of less than 2.0 have a dead storage of cally, the p e r m a n e n t pool has two major functions, to
2 5 % . Longer ponds with length-to-width ratios greater shield the deposited sediment to prevent resuspension
than 2.0 have a dead storage of 1 5 % . from large storms and to provide a body of water that
D e a d storage values for the C S T R S model vary with will be clarified by settling over the period between
the number of reactors assumed. Using the C S T R S runoff events. In the plug flow concepts used in some
reactor model, Griffin et al. (1985) showed that the models of p o n d performance, it is assumed that the
optimum number of reactors for a best fit to the dye inflow storm water displaces the clearer permanent-
tracer data was 2.0. pool water in a first-in first-out concept. Thus, the
It is important to note that dead storage calculations p e r m a n e n t - p o o l water is the first flow discharged, re­
from tracer studies will vary widely depending on the sulting in a higher quality effluent for the first flush of
model used. This is due partly to the fact that the term stormwater.
dead storage is somewhat misleading. In Fig. 9.13, it is T h e relative effectiveness of permanent-pool volume
implied that no flow enters the dead storage volume. d e p e n d s on reservoir shape. A s shown earlier, ponds
326 9. Sediment Control Structures

Normal Pool

(B)

Inflow

Inflow
Figure 9.13 Examples of basin shapes. Flow paths are what would be expected if recirculation is ignored.
Shaded areas are conceptualized dead storage areas (after Environmental Protection Agency, 1976).

with a low length-to-width ratio have higher dead stor­ Flocculation or dispersion processes can be ex­
age volume (Griffin et al., 1985). plained by the double layer theory or by particle bridg­
ing. As shown in C h a p t e r 7, the presence of a potential
Water Chemistry a n d Pond Performance energy barrier ( P E B ) will prevent two particles from
Some ponds are relatively clear within a few days flocculating. If the ionic strength of cations (concentra­
after a runoff event and others will not be clear when tions of cations) is increased, the double layer thick­
observed months later. Studies indicate that runoff ness will decrease and the magnitude of the P E B will
water chemistry is one of the major factors causing this decrease. A t some point, the ionic strength will reach a
difference ( T a p p et al., 1981; T a p p and Barfield, 1986; value such that the barrier is eliminated and particles
Evangelou et al., 1981), primarily due to its influence will flocculate. Thus, ionic strength has an important
on flocculation or dispersion. effect on flocculation.
Sediment Detention Basins 327

Outlet Riser

Inflow

•Sheets Of 4'x8'x1/2" Exterior


Plywood Or Equivalent
1 Riser Crest Elevation

P o s t s - m i n . Size 4" Square


Or 5" Round. Set At Least
3' Into The Ground.

Elevation
Figure 9.14 Sediment basin baffles for reducing dead storage (after Environmental Protection Agency, 1976).
328 9. Sediment Control Structures
50
1 1 1 1 ι I I
Plug Row Model
0 L / W - 1:1
x L / W β 1:2
— L / W » 3:1
A

α L / W - 2:1

ο
« y—Regression Line For
ο 0 χ / L / W < 2:1
3 * * *
CO Χ ΧΒ
A
«β A
0
Q A y—Regression Line For
0
α
A / L/W £ 2 : 1
• A • Π Α D
α • •—-
— α _

1 1 ι ι ι 1 1
1 2 3 4 5 6

Momentum Factor (x 10 ) s

Β
50
τ 1 —
CSTRS Model
A L / W α 3:1
ο L / W - 2:1
40 X L / W - 1:1
ο L / W - 1:2

30 -Regression Line For

I
L / W < 2:1
°x 0

ο «
20
Regression Line For
L / W i 2:1

10 L

1 6

Momentum Factor (x 1 0 ) 5

Figure 9.15 Relationship between dead storage and momentum factor for ponds of varying
length-to-width ratio. Computations were made with the plug flow and CSTRS models (after
Griffin et al., 1985). The momentum factor is the inflow momentum divided by the weight of fluid in
the pond. (A) Optimum dead storage value for plug flow model. (B) Optimum dead storage value
for CSTR model with two reactors.

In addition to ionic strength, the type of absorbed drated (surrounded by water molecules) such that it is
cation is also important. Calcium ( C a ) and Magne­
2 +
weakly absorbed. T h e charge density for N a does not +

sium ( M g ) are both divalent ions with a strong


2 +
change as rapidly from negative to positive as in the
electric charge. Thus, when absorbed on the colloidal case of C a , resulting in a P E B until very high con­
2 +

surface, the charge density goes from negative to posi­ centrations of sodium are reached. Therefore, the rela­
tive, resulting in a thick double layer and probable tive magnitude of N a to C a + M g is also impor­
+ + 2 +

flocculation. Sodium ( N a ) on the other hand is a


+
tant to the determination of flocculation potential. Two
monovalent ion with a weak charge and is highly hy- measures are used to describe the relative magnitude
Sediment Detention Basins 329

of these ions. O n e measure known as t h e exchangeable practice, t h e actual velocity would b e designed to be
sodium percentage (ESP) is the fraction of absorption less than V to prevent scour.
H

sites on the colloidal surface (exchange phase) filled by Relationships m o r e complex than Eq. (9.36) are
N a . T h e other measure, known as t h e sodium absorp­
+
available, b u t require complex computer models for
tion ratio (SAR), is the fraction of ions in solution that their execution. Wilson and Barfield (1985, 1986a) de­
are sodium. For soils that are heavy clays, with high scribe such a model, known as BASIN, which evaluates
numbers of exchange sites (high cation exchange capac­ sediment resuspension using t h e Einstein equation.
ity), the E S P and S A R are highly related. In soils that B A S I N is discussed further in a subsequent section.
are composed of kaolonitic colloids, t h e n u m b e r of
exchange sites with a double negative charge are low as
compared with heavy clays. As a result; t h e monovalent Modeling Pond Performance:
N a ion is more easily absorbed on t h e kaolonite; thus
+ Theoretically Based Predictors
the S A R and E S P values a r e typically not equal. Reservoir a n d p o n d sedimentation is not a new field.
In an effort to predict t h e occurrence of flocculation T h e first theories on reservoir trapping efficiency were
or dispersion, Evangelou et al. (1981) proposed that probably developed by Seddon (1989). Since that time,
the thickness of the double layer could b e used as a many models have b e e n developed. In t h e following
prediction parameter. Using electrical conductivity as section a brief overview of t h e most commonly used
a measure of ionic strength, Evangelou et al. analyzed models is given, starting first with steady-state and
a kaolonitic soil with a S A R of less t h a n 5.0 a n d a moving t h e n t o non-steady-state models.
solution phase composed primarily of C a + M g . 2 + 2 +

They proposed that a m e a s u r e of t h e double layer Steady-State Overflow Rate Models—Quiescent Flow
thickness could b e estimated from A n analysis of t h e trapping efficiency of rectangular
basins with steady-state inflows and outflows can be
1 m a d e using settling velocities, flow rates, and surface
DLT = 1/2
(9.35) areas. Based o n t h e trajectory of particles through a
[(EC)(0.014)]
settling basin, a critical settling velocity, V , that will c

just allow a particle t o settle to t h e bottom in its


where D L T is thickness in arbitrary units and E C is t h e trajectory through t h e basin can b e selected as shown
electroconductivity of t h e water in millimhos p e r cen­ in Fig. 9.16. T h e critical settling velocity can be given
timeter. For D L T values less than 15, Evangelou et al. by
indicated that flocculation would occur. D a t a for o t h e r
clay types and other dissolved chemicals were not V =
c D/T, (9.37)
evaluated, limiting the use of Eq. (9.35) for o t h e r
situations. w h e r e D is t h e d e p t h of t h e basin a n d Τ is t h e flow
through time. If it is assumed that flow is quiescent
Sediment Scour in Ponds
(turbulent free) a n d that n o resuspension occurs, then
W h e n water flows over deposited sediment, t h e lift all particles with a settling velocity greater than V will c

and drag forces will attempt to move particles o u t of be t r a p p e d . Based o n geometry, it can be shown that
the bed. T h e critical velocity that will cause motion in a the fraction of particles trapped with a settling velocity
rectangular settling tank with constant flow rate was
analyzed by C a m p (1946) using Shield's p a r a m e t e r s , or
Completely Mixed Completely Mixed
1/2 -Inlet Zone Outlet Z o n e — ν
8*(SG - l)gd
(9.36)
7 V s = Settling Velocity
Of Particle In
Question

where V is the critical velocity causing scour of a


H

particle of diameter d, g is acceleration of gravity, S G


is particle specific gravity, / is the D a r c y - W e i s b a c h
V = Design Settling
friction factor, and k is Shield's p a r a m e t e r discussed in
c

Velocity Of Particle
Just Removed
Chapter 7. Typical values for / are 0.02 to 0.03 for
settling chambers and k is 0.4 for granular materials Sludge Zone

and 0.06 for particles that have cohesive properties. Figure 9.16 Illustration of sediment flow trajectories in an ideal
Units on g and d will dictate t h e units on V . In H rectangular sediment pond.
330 9. Sediment Control Structures

Vs less than V is given by


c should again b e noted that the following assumptions
were m a d e :
(9.38)
• quiescent flow
• n o resuspension of sediment
Equations (9.37) and (9.38) can be converted to t h e • rectangular shaped reservoir
familiar overflow rate equation by noting that the flow • steady flow.
through time, Γ, is given by effective length of the
basin L divided by the flow through velocity K, or It was also implicit that inlet and outlet zones were
completely mixed and that settling was discrete particle
settling as described in C h a p t e r 7.
F = (9.39a) If X represents the fraction of particles finer than a
D/L/V DV/L
given size and dX represents a differential element of
X, t h e n the total trapping efficiency for a basin inte­
Further modification can be m a d e to yield grated over all sizes is

Ε = f°°FdX.
(9.39b)
WDV/(LW) '
If the fraction of particles with settling velocity less
where W is the width of the chamber. T h e t e r m WDV t h a n V is given by X , the total trapping efficiency of a
c c

is simply the discharge Q and LW is the surface area basin can b e calculated from
A. H e n c e
E = (l-X )+
c ( <-^dX
X

F = = 1 7 ^ 10, (9.40) (9.41)


Q/A

1 = 1 K
c
which is the familiar overflow rate equation. T h e quan­
tity Q/A is equal to V and is known as the overflow
c T h e use of Eq. (9.41) is illustrated in Example Problem
rate. Equation (9.40) illustrates that the trapping frac­ 9.5.
tion for a given particle is i n d e p e n d e n t of the d e p t h of T h e effluent size distribution can also b e estimated
the settling basin u n d e r steady-state conditions. It with the overflow r a t e concept. Since the fraction

100

0^=.00075mm
d 2=.0014mm
d =.0021mm
3

d =.0030mm
4

d =.0042mm
5

u .0056mm
Φ
α

3
ο
φ
α.

.0001

Particle Diameter (mm)


Figure 9.17 Size distributions used for Example Problem 9.5.
Sediment Detention Basins 331

trapped of a given size is V /V , t h e fraction dis­ sl c Assuming a temperature of 68° F, the corresponding equiva­
charged is 1 - V /V . If the total mass in t h e storm is
si c lent diameter sphere corresponding to V can be determined c

M , t h e mass discharged for a given size, i, is (1 —


s from Eqs. (7.5) (or (7.6) for larger particles) or
K / K ) A A M , a n d t h e fraction of total discharged
si c
r
l s

mass of size / is
/ V \ 1 / 2
/ 1.15 X 1 0 " \ 4 1 / 2

J =
c
(1-K /K )A*,M
s i C S mm. 00064
ΔΓΓ •= - ' ( i n ) - ( — 2 5 Γ -
°·' Ul-Vn/VJLXiM,
(9.42)
2. Calculating trapping efficiency. From Fig. 9.17, the frac­
T h e fraction finer than size / , F F , is simply t h e s u m tion of particles smaller than d is 0.24. The evaluation of
c
Q y

of A F F , for all smaller particles. the integral in Eq. (9.41) is given in columns 1-5 in the table
Q
below. From Eq. (9.41),
This approach has t h e inherent limitations of t h e
overflow rate concept. Computations a r e illustrated in
Example Problem 9.5.
ι = 1
V
c
Example Problem 9.5. Calculating trapping = (1 - 0.24) + 0.0634 = 0.8234.
efficiency and effluent size distribution
with the overflow rate concept Hence the trapping efficiency is approximately 82%.
3. Calculating effluent size distribution. The fraction of
A rectangular reservoir has a steady inflow and outflow effluent smaller than particle size /, AFF , is calculated Q y

rate of 5 cfs and a surface area of 1.0 acre. Calculate the from Eq. (9.42) and the fraction finer than size / is
overflow rate. If sediment inflow has a size distribution as
given in Fig. 9.17, calculate the trapping efficiency of the
structure and estimate the effluent size distribution.
FF , =
0 t EAFF ,,, 0

Solution:
1. Calculating the overflow rate. The overflow rate is given
by where ; is the number of particle sizes smaller than L The
5.0ft /sec 3
computations are summarized in columns 1-3 and 7-9 in the
c
A (1.0acre)(43,560ft /acre) 2 table below and the computed effluent size distribution plot­
ted on Fig. 9.17. The size corresponding to FF · is the larger Q

= 1.15 Χ 1 0 " ft/sec. 4


value of the size range in column 1.

1.0

.9

.8

.7

V ^ \
.6

.5
h
V
I
ο
.4 V V V
ο k
CO .3

.2

.1

0
.01 .05 0.1 0.5 10 50 100 500 1000

Overflow Rate - Q/A m / h r


Figure 9.18 Sedimentation as a function of overflow rate for various particle sizes (1 μ π ι = 1 0 " mm). 3
332 9. Sediment Control Structures
Tabulations for Example Problem 9.5

Particle
size Diameter FF Particle
range* v.b
Column 6/Σ Σ size for
SI
(mm) (mm) ΔΧ 2

(ft/sec) (VJV )AXf


c
(l-V /V )AX/<
si c (Column 6) e
(Column ΊΫ (Column 8)*

0.0001-0.00095 0.00075 0.04 1.58 x 10" 6


0.0005 0.039 0.222 0.222 0.00095

0.00095-0.0016 0.0014 0.04 5.51 x ΙΟ" 6


0.0019 0.038 0.216 0.438 0.0016

0.0016-0.0025 0.0020 0.04 1.24 χ 10" 5


0.0043 0.036 0.205 0.643 0.0025

0.0025-0.0036 0.0030 0.04 2.53 x 10" 5


0.0088 0.031 0.176 0.819 0.0036

0.0036-0.0045 0.0042 0.04 4.96 x 10" 5


0.0173 0.023 0.131 0.950 0.0045

0.0045-0.0064 0.0056 0.04 8.81 x 10~ 5


0.0306 0.009 0.050 1.000 0.0064

Σ = 0.24 1 = 0.0634 1 = 0.176 Σ = 1.001

'Taken from Fig. 9.17 for each diameter.


*Min of V = 2.81 d and Eq. (7.6).
si
2

c
V A X / 1 . 1 5 x 10-Λ
si

J
( l - V /1.15 x l O " ) ^ .
si
4

MFF 0 I = (Column 6)/Σ(1 - V /1.15 χ Ι Ο ^ Δ Χ ^ (Column 6/Σ Column 6) = Column 6/0.176.


s

supper limit of size range in column 1.

From Example Problem 9.5, it is clear that particle size and ε is t h e turbulent diffusivity for sediment of size d.
overflow rate are important parameters affecting the perfor­ Simplifying
mance of sediment ponds. This is further illustrated in Fig.
9.18. Since the horizontal axis (overflow rate) in Fig. 9.18 dC
decreases as area increases, an increase in surface area, for a ε — + V C = 0. S (9.43)
given particle size and discharge, will increase trapping. Thus, dz
for steady-state quiescent flow, the optimum rectangular
reservoir is one that maximizes surface area subject to the T o solve Eq. (9.43), it is necessary to have an estimate
constraint that particles not be suspended. This will typically of the eddy diffusivity ε. By assuming that the eddy
minimize depth. Also, from Example Problem 9.5 and Fig. diffusivity for m o m e n t u m and sediment are equal and
9.18, it is obvious that discrete particles 4 μτη and smaller are by using the logarithmic velocity profile to describe
very difficult to trap with any reasonably sized basin. As is velocity, the eddy diffusivity becomes
shown in following sections, the same conclusion can be
drawn about turbulent flow and non-steady-state flow.
e-*£/.(l-^)z, (9.44)

Steady-State Overflow Rate Models: Turbulent Flow w h e r e U* is the so-called shear velocity, D is depth of
flow, and k is von Karmon's constant given by 0.4 in
A theoretical analysis of turbulent flow is much more
clear water. For uniform flow in o p e n channels, shear
complex than that required for quiescent settling as a
velocity is defined as
result of upward diffusion of sediment by turbulence.
A n analysis of turbulent diffusion of sediment should
start with the equilibrium case in which the upward (9.45)
turbulent diffusion of sediment of a given size d just
equals the rate at which particles are settling (see Fig. w h e r e τ is shear on the channel b o t t o m and S is the
0

9.19A), or energy gradient.


Using Eq. (9.44), Eq. (9.43) can be integrated to
dC yield
~dz
C Ό/ζ -1 (9.46)
where ρ is density of water, V is settling velocity of a
s
D/a - 1
particle diameter of size d, C is the concentration
(mass s e d i m e n t / m a s s w a t e r ) of particles of size d, and where C a is the concentration of d-sized particles
Sediment Detention Basins 333

Downward Flux Due To Settling


p V s C A X !

—Z- . Plane Of Unit


^yr Width

Upward Flux Due To Turbulence = -ρ ε -lk-ΔΧ · I

Vertical Flux In = ρ θ ν Δ χ - ρ ε
δ

Horizontal Flux In = Horizontal Flux Out =


ρουΔζ-ρε -48-ΔΖ ρουΔΖ-ρε -|^-ΔΖ+
"fc [pCUAZ^e -§|ΔΖ]ΔΧ

Vertical Flux Out =

ρον Δχ-ρε
3 -|£ΔΧ+

[ρον Δχ-ρε - | £ Δ Χ ] Δ Ζ
3

Figure 9.19 Schematics for deriving sediment trapping efficiency equation for turbulent flow. ( A ) Schematic
showing equilibrium sediment movement. The minus sign on the upward flux due to turbulence accounts for
the fact that sediment diffuses from high to low concentration. (B) Control elements for deriving
non-steady-state sediment transport equation.

measured at a distance a above the channel bed and balance using the terms in Fig. 9.19B results in

z* = VJkU, (9.47) dpCAxbz d d dC


= - — p C i / Δ χ Δ ζ + — ρε — Δ χ Δ ζ
at οχ dx dx
Equation (9.46) describes the case where the rate of
d 3 dC
resuspension of sediment at the channel bed equals the - — pCV Δ χ Δ ζ + — ρε — Αχ Δ ζ .
s

rate of settling; hence there are no changes of concen­ dz dz dz


tration in the horizontal flow direction. In any small (9.48)
sediment pond this will not be the case; therefore an
equation that considers changes in the mean flow di­ After dividing by ρ Δ χ Δ ζ and assuming that U and V s

rection, must be derived. Such a derivation will require are constant,


a mass balance on a fluid element.
dC
A schematic of a mass balance of a fluid element is
shown in Fig. 9.19B. In such a mass balance, the rate of
— + U
dt
dC
dx
— + ΚdzdC— s

inflow of mass into the element minus ihe rate of


outflow of mass from the element must equal the rate d dC d dC
= —ε— + — ε—. (9.49)
of change of mass in the element. Performing the mass dx dx dz dz
334 9. Sediment Control Structures

T h e general solution to Eq. (9.49) will require t h e use


of numerical procedures. Dobbins (1944) developed an
analytical solution for the special case in which ε is a
constant by using what is known as a product series
solution. C a m p (1946) manipulated Dobbins solution
II
for the special case in which there is n o scour to obtain
ο
V

F = 1 85fe
2 5 , ) ( β ι + alT
2
>
ο
(9.50)

where
a
PS

(9.51)

B 2 = (9.52)
2B V X C

and a a , a ,...,a
v 2 3 are the real positive roots of t h e
n

transcendental equation Figure 9.20 Sediment trapping efficiency of a rectangular reservoir


with steady-state turbulent flow (after Camp, 1946).

a B x

2 Cot a = - (9.53)
a a n d (9.55), t h e horizontal axis of Fig. 9.20 becomes
KD 10V D 1/6

In Eq. (9.50), H is + 1 when a is in the first a n d S


n
(9.56)
second quadrants and - 1 when a is in t h e third and 2ε
fourth quadrants. Dobbins verified his equations exper­
Wilson a n d Barfield (1986b) summarize o t h e r meth­
imentally using carefully controlled laboratory studies.
ods for estimating turbulent diffusivities in sediment
Numerical computations using Eq. (9.50) require the
ponds. Details a r e given in Appendix 9A.
use of u p to 40 terms of t h e infinite series. T o simplify
T h e effect of turbulence o n basin t r a p efficiency is
the analysis procedures, C a m p (1946) modified Eq.
a p p a r e n t from Fig. 9.20. W h e n turbulence levels are
(9.50) to predict trapping efficiency of a rectangular
low, t h e value of V^D/Ιε is large a n d trapping effi­
reservoir and developed the graphical solution shown
ciency a p p r o a c h e s that of quiescent settling given in
in Fig. 9.20. T o use Fig. 9.20, it is necessary to have an
Eq. (9.40) as V /V . W h e n turbulence levels are high, ε
s c
estimate of ε. C a m p (1946) proposed that ε could b e
is large, V D/28
s is small, a n d trapping efficiency is
estimated from
reduced. F r o m Fig. 9.20 it can also b e seen that t r a p ­
ping efficiency for small values of VJV Cis relatively
ε = 0.75t/*D, (9.54)
unaffected by turbulence.
Since trapping efficiency in t h e turbulent flow condi­
where U* is shear velocity and D is d e p t h of t h e
tion is d e p e n d e n t o n K D / 2 c , it is not i n d e p e n d e n t of
rectangular reservoir. In this case, ε is t h e average
s

flow d e p t h as in t h e case of quiescent settling. How­


diffusivity over t h e entire flow cross section as opposed
ever, it is interesting t o n o t e how a doubling of d e p t h
to the point value given earlier in Eq. (9.44). Brown
would affect trapping. Since Eq. (9.56) indicates
(1950) proposed that t / be estimated from a modifi­
K D / 2 e is proportional t o D , t h e n changing t h e
# 1 / 6

cation of Manning's equation, or s

d e p t h by a factor of 10 would cause a change in


V D/28
s of 1 0 o r a factor of 1.5 (a 5 0 % change).
1 / 6

nVyfg F r o m Fig. 9.20, a 5 0 % change in V^D/Ιε would change


(9.55)
1.5D ' '
1 6
F by less t h a n 0.05 in all cases. This weak d e p e n d e n c e
of F on D p r o m p t e d C a m p (1946) to infer that t h e
where V is the average horizontal velocity in the reser­ o p t i m u m settling tank design is o n e that maximizes
voir and η is Manning's roughness. Using Eq. (9.54) surface area a n d minimizes depth, without causing
Sediment Detention Basins 335
1.0 Solution: From Example Problem 9.5,
υ
a
a.
.8
.6
f V c = Q/A = 1.15 Χ 1 0 " ft/sec 4

dc = 0.0064 mm
α
• Q = 5.0 cfs
Q 4. —I/ > 0 A = 1.0 acre.
%
A
ο
fa. First, the turbulence parameter of Dobbins must be calcu­
II λ/ lated. In the discussion of reservoir scour in this chapter
fal following Eq. (9.36), it was stated that the friction factor for a
reservoir is 0.02 to 0.03. Using a value of 0.03, it can be
.5 1.0 5.0 10.0 shown that Manning's η can be given by (see Eqs.
(4.21M4.23))

η = 1.49
Figure 9.21 Trap efficiency versus ratio of settling velocity to over­
flow rate for a high-turbulence model (after Chen, 1975).
where R is the hydraulic radius. Since R for a rectangular
section is

scour. This is t h e same conclusion reached earlier for WD (10)(5) 50


quiescent flow. = 2.5,
2D + W 2 X 5 + 10 20
Chen (1975) defined a high turbulence flow condi­
then
tion as o n e in which V D/2e is equal to 0.01 a n d
s

determined values of trapping efficiency versus VJV C 1.49(2.5) 1/6

for t h e condition. These values a r e plotted in Fig. 9.21 = 0.019.


n
~ ^ ( 8 X 32.2)/0.03
along with those predicted for quiescent settling. C h e n
utilized Vetter's (1940) equation a n d proposed that t h e The forward flow velocity is
turbulence values can b e approximated by
Q_ = 5
= 0.1 fps,
F=l-exp(-K /K ). s c (9.57) v=
A cs (5 X 10)

where A is the cross-sectional area perpendicular to flow.


cs
Total trapping efficiency for quiescent settling was Dobbin's parameter is, therefore,
found earlier by integrating E q . (9.40) over all size
ranges to yield KD 10KD 1 / 6
10(5) K 1/6
S

= 1.213 X 1 0 K . 3

2E nVyfe O.O^O.HOv ^ 7

dx.
Calculations for predicting trapping efficiency by Dobbin's
and Camp's procedures as well as those of Chen are summa­
T h e total trapping efficiency for fully turbulent flow rized in the Table 9.4.
Using the results from Example Problem 9.5 for quiescent
can b e found by integrating E q . (9.40) over all size
settling along with calculations from Table 9.4, the following
ranges to yield
comparison can be made:

. E - l - / l
e J - £ U . (9.58) Trapping
Procedure efficiency

These models a r e illustrated in Example Problem 9.6. Quiescent settling Eq. (9.41) 0.823
Dobbin-Camp (Fig. 9.20) 0.794
High-turbulence model (Eq. (9.58) 0.789
Example Problem 9 . 6 . Comparison of the turbulent
and quiescent flow models Comments: The differences in predicted trapping efficien­
cies are relatively small. As expected, the quiescent value is
Estimate the trapping efficiency of the reservoir in Exam­ the highest, the fully turbulent value (Chen) the lowest, and
ple Problem 9.5 using the turbulent flow procedures of Chen the Dobbin-Camp model intermediate. An explanation for
and those of Dobbin and Camp. Compare the results to the this lack of a major difference can be developed from Fig.
quiescent flow model (overflow rate) of Camp. Assume that 9.21.The maximum deviation between quiescent settling and
the reservoir is 5 ft deep and 10 ft wide. fully turbulent flow occurs at V /V = 1.0. For this problem,
s c
336 9. Sediment Control Structures

Table 9.4 Tabulations for Example Problem 9.6

Fully turbulent model Dobbin and Camp model


[Eq. (9.57)] (Fig. 9.20)
Average
Particle size diameter 1 =
range* VP 1-exp \0V D
si
U6

SI
(mm) (mm) IsX? (ft/sec) nVg d m
Ff

0.110-0.500 0.170 0.10 8.1 x 10" 2


706 1.0 0.100 98.5 1.0 0.10
0.072-0.110 0.088 0.10 2.17 χ 10" 2
189 1.0 0.100 26.0 1.0 0.10
0.048-0.072 0.058 0.10 9.45 χ 10" 3
82 1.0 0.100 11.5 1.0 0.10
0.035-0.048 0.040 0.10 4.50 χ ΙΟ" 3
39.1 1.0 0.100 5.5 1.0 0.10
0.026-0.035 0.030 0.10 2.53 χ 10" 3
22.0 1.0 0.100 3.1 1.0 0.10
0.17-0.026 0.022 0.10 1.36 x l 0 ~ 3
11.8 1.0 0.100 1.6 1.0 0.10
0.009-0.017 0.012 0.10 4.05 χ 10" 4
3.52 0.97 0.097 0.49 1.0 0.10
0.004-0.009 0.0065 0.10 1.19x 10" 4
1.03 0.64 0.064 0.14 0.65 0.065
0.002-0.004 0.0034 0.10 3.25 x 10~ 5
0.28 0.24 0.024 0.04 0.25 0.025
<0.002 0.0012 0.10 4.05 x W 6
0.04 0.04 0.004 0.00 0.04 0.004

Σ = 0.789 Σ = 0.794

"From Fig. 9.17.


b
V = 2.81 d?. This overestimates V for larger particles, but refined estimates would also yield F = 1.0 for these
si s

particles.
c
V / 1 . 1 5 x 10" .
jf
4

1.23xl0 V .
rf 3
5/

'From Fig. 9.20.

only 10% of the particles fell in that range. If 30 to 40% had m e n t of a better design m e t h o d for sediment structures
been in that range, the difference between the quiescent and ( W a r d et al, 1977, 1979). D E P O S I T S is o n e of the
fully turbulent model would have been 10 to 15% which is a pond options in S E D I M O T II (Wilson et al, 1982).
significant difference.
T o m a k e the model sufficiently general to be appli­
cable to most sediment basins, the flow within the
basin is idealized by the plug flow concept (see Fig.
9.11). Plug flow assumes delivery of the flow on a
first-in, first-out basis and allows n o mixing between
Variable Flow Rate—Plug Flow Model
plugs. Although this concept does not account for
Overflow rate models have found widespread accep­ short-circuiting or turbulent flow, provision for a cor­
tance in the design of settling tanks for water and rection factor to account for these p h e n o m e n a has
sewage treatment systems. They can also be applied to b e e n incorporated into the model.
the analysis of sediment ponds that have constant Computational procedures for D E P O S I T S are sum­
inflows as might well be the case with pumping from marized in W a r d et al (1977) and are illustrated in
quarries or d e e p mines. T h e majority of the sediment Example Problem 9.7. A description of the model
ponds, however, must handle surface runoff a n d inflow computational procedures follows, but can probably
rates that vary over several orders of magnitude during best b e understood after working through the example.
a runoff event. In D E P O S I T S , an outflow hydrograph is generated
O n e variable flow rate procedure that has found first by continuity routing through t h e reservoir and a
widespread acceptance in the analysis and design of total cumulated flow hydrograph determined for both
sediment ponds is a computer model known as D E ­ inflow and outflow. Cumulative flow is calculated by
P O S I T S . T h e D E P O S I T S Model ( d e t e n t i o n perfor­ summing the area u n d e r the hydrograph. T h e inflow
mance of sediments in / r a p structures) was developed and outflow hydrographs are then divided into plugs of
to study the sedimentation process in small reservoirs equal volume by equally dividing t h e cumulative inflow
in the hopes of providing an insight into the develop­ and outflow hydrograph into equal volumes. O n c e the
Sediment Detention Basins 337

hydrograph is divided into plugs, corresponding plugs


are identified on t h e inflow a n d outflow hydrograph,
the detention time for each plug is calculated, a n d an
Total Runoff Volume « 2.61 a c - f t
average depth of flow a n d surface area for t h e plug
residence time is determined. T h e volume of each plug
is divided into four layers for sedimentation computa­
tions. Using settling velocities calculated from Stokes
Law, t h e amount of sediment in each layer is calcu­
lated at t h e e n d of t h e plug detention time. Particles
are considered to be trapped as soon as they reach t h e
basin bed. Withdrawal can b e specified as uniform, or
distributed among t h e four layers at t h e user's option.
T h e outflow concentration for each plug of flow is
calculated from t h e mass of sediment a n d t h e mass of
water in t h e plug as it exits t h e pond.
Calculations of effluent size distributions follow simi­
lar procedures to those of t h e overflow rate m e t h o d 4 6 8 10
N
14
illustrated in Example Problem 9.5, but must b e com­ Time (hr)
puted for each layer. Thus, procedures a r e m o r e com­
plex. T h e procedure is developed in Example Problem
9.9. Further details a r e given in W a r d et al. (1977) a n d Β
100
Wilson et al. (1982). 15
Y/_ _ 14 ^

80 L _ /_ 13^f__

Example Problem 9.7 Illustration of D E P O S I T S


model computations ο
60
F Z' Q
Using the plug flow concept, divide the hydrograph in Fig. 7Jl^L
ι T
D.10/ i£ 8~
9.22A into 15 plugs. For plug 10, determine the average flow « J I.C 7
ο 40
depth and plug detention time. Divide the average depth into u Γ V " - -Ύ "V ~ 6
V
four layers, and calculate the particle size that will settle out rJL iι „ _ - /
1 1
\\\ ~ 5
of plug 10 during the plug detention time. Finally, assuming / j o /
7 ι t
j ο
4
20 p _
uniform withdrawal, calculate the effluent concentration for 1 _ 4»- / 3
/ 1
/
plug 10 assuming that the inflow concentration for plug 10 is / "Xr— ι 2
19,270 mg/liter and the inflow particle size distribution is
given in Fig. 9.23B.
2 4 6 8 10 12
Solution: I-
I Τ Time (hr)
1. Dividing hydrographs into plugs. Subdivision into plugs is
T
D.P

done from a cumulative inflow and outflow hydrograph where Figure 9.22 Hydrographs for Example Problems 9.7 and 9.8. ( A )
Inflow and outflow hydrographs, ( B ) Cumulative flow graphs.
jjl(t)dt ΣΓ /(<,)Δ/,.
=1

ic(0 =
/£/(/) Λ QTR

3. Average depth for plug 10. From Fig. 9.22B, the inflow
where IC(r) is relative cumulative inflow at time t, I(t) is
and outflow times for the tenth plug are
inflow rate at any time r, and Q is the total inflow volume.
TR

Using this approach, the inflow and outflow hydrographs in


Fig. 9.22A are transformed to the cumulative hydrographs in '/,ιο = 4.00 hr
Fig. 9.22B. To divide the hydrograph into 15 plugs, the i o l 0 = 7.62 hr.
vertical axis is simply divided into 15 equal increments. The
tenth plug is emphasized in Fig. 9.22B. Using the average reservoir depth versus time curve in Fig.
2. Plug detention time. Detention time for the tenth plug is 9.23 and /, and / given above, the average depth can be
1 0 D 1 0
shown in Fig. 9.22B as the time interval between plug inflow determined by numerically integrating the depth-time curve
and plug outflow, or and dividing by 3.62 hr to be
7 , = 7.62 4.0 hr = 3.62 hr or 217.2 min.
D 1 0
D i n = 1.51 ft.
338 9. Sediment Control Structures

A 5. Calculating effluent concentration. Once d , the particle c

2.0 diameter that will just settle out of a layer, is determined,


this can be transformed into a percentage finer by the follow­
ing assumptions:
• All particles larger than d are out of the layer and all
c

particles smaller than d are in the layer. c

• Inflow sediment is uniformly mixed.


Using d calculated above and the inflow size distribution
c

given in Fig. 9.23B, the following fractions finer than d were c

determined.

Layer Fraction finer,


2 4 6 *10 12 14
j (mm) FFW Ct> )
Time ( h r )
1 0.0032 0.26
Β
2 0.0045 0.30
3 0.0056 0.33
4 0.0064 0.34

0.60

The outflow concentration C OJ for layer / is given by

C ^ . - C j F F K , , ) ,

where C is the inflow concentration. Given the inflow con­


x

0.20 centration of 19,270 mg/liter, the effluent concentration for


.0032.
.0045
.oose
.006

each layer is

0.0
.0001 .001 .01 .10 Layer; C . (mg/liter)
c y

Particle Diameter (mm)


1 5010
Figure 9.23 Average depth and size distributions for Example
2 5781
Problems 9.7 and 9.8. (A) D e p t h - t i m e relationship. (B) Particle size
distribution. 3 6359
4 6552

4. Calculating particles that will settle from each layer. The The average outflow concentration is
pond depth, D, is divided into four layers. To settle out of
layer one, a particle would settle through a depth of D/4, to r _ Q i + C t 0 t 2 + C 0 t 3 + C 0 t 4

settle out of layer two, a particle must settle through a depth


of 2 D / 4 , layer three, 3 D / 4 , and layer four, 4 D / 4 .
The required settling velocities and particle diameters are 5010 + 5781 + 6359 + 6552
C = 0 = 5926 mg/liter
summarized below.

Required Diameter d · c

Settling depth, Settling velocity from Eq. (7.5)


7 , 10
D SD SD/r D 1 0 and Fig. 7.3
Layer (min) (ft) (ft/sec) (mm)

1 217.2 0 . 3 7 8 = D/4 0.0000290 0.0032


2 217.2 0.755 = D/2 0.0000579 0.0045
3 217.2 1.133 = 3 0 / 4 0.0000869 0.0056
4 217.2 1.510 = 0 0.0001158 0.0064
Sediment Detention Basins 339

T h e computations in Example Problem 9.7 are for a Computational details for the above table follow those of
totally dewatered reservoir, i.e., no p e r m a n e n t pool. Example Problem 9.7 and are left to the reader.
Procedures for considering the effects of a p e r m a n e n t
pool are given in Example Problem 9.8.
Effluent concentrations for plug 10 in Example Prob­
lems 9.7 and 9.8 are similar for t h e totally dewatered
p o n d as compared to the pond with a p e r m a n e n t pool,
Example Problem 9.8 Effects of permanent pool although the detention time is increased from 127 to
on DEPOSITS prediction 385 min. T h e increase in settling d e p t h offsets the
increase in the detention time. T h e p e r m a n e n t pool
Assume that the reservoir in Example Problem 9.7 has a does allow effluent early in the storm to be n e a r zero
permanent pool volume of 0.91 acre · ft and a dead storage as the stored permanent-pool water is discharged prior
of 25%. To develop this permanent pool, the crest of the to any stormwater being discharged. Calculation with a
riser is left at the same elevation, and a 1-ft-deep pool is computerized D E P O S I T S model would yield a trap­
excavated. How will this impact the effluent concentration of ping efficiency for Example Problem 9.7 (no p e r m a n e n t
plug 10 in Example Problem 9.7?
pool) of 63.2% and for Example Problem 9.8 (with a
Solution: Adding permanent pool is typically accomplished p e r m a n e n t pool) the trapping efficiency would be
by raising the elevation of the spillway. This modifies the 68.82%. T h e plug flow model, however, does not al­
outflow hydrograph and the average depth time relationship. ways predict that p e r m a n e n t pool will increase trap­
To accurately evaluate such a situation, the analysis given in ping efficiency.
Example Problem 9.7 must be redone. In this problem, it is
Computational details are quite tedious for using
assumed that the permanent pool is developed by excavation
this model. Software for making these computations is
in order to simplify computation. For a first approximation
with this small permanent pool, it will be further assumed available from the authors.
that the outflow hydrograph is not appreciably affected.
The permanent-pool water displaced is equal to the per­
manent pool volume minus the dead storage, or Example Problem 9.9 Size distribution equations
Volume displaced = V - V p ds for plug flow model
= 0.91 - (0.25)(0.91) = 0.68 acre · ft.
Develop a routine to predict the particle size distribution
From Fig. 9.22A, the total volume of runoff is 2.61 acre · ft.
for sediment in the discharge from a plug of flow, using the
Thus the runoff displaced represents 0.68/2.61 or 26% of the
DEPOSITS model approach. Assume uniform withdrawal
total volume. From Fig. 9.22B, the time required to displace
and a completely mixed inflow.
26% of the total flow is shown as Τ = 4.8 - 2.0 hr = Όρ

2.8 hr. (Note that the starting time for Fig. 9.22B is Solution: Let the inflow particle size distribution be de­
2 hr). Thus the permanent pool effectively has the impact of fined by FF W ), where d is particle diameter. For a layer j ,
f k k

adding 2.8 hr to the detention time of each plug. From the particle size that just settles out of the layer is d . All c}

example Problem 9.7, T = 3.62 hr; thus the effective


D 1 0
particles in the effluent from layer j are equal or smaller than
detention time of plug 10 with a permanent pool becomes d j . If FF j(d )
c Q represents the size distribution of the
k

r Dil0 = 3.62 + 2.80 = 6.42 hr. effluent from layer j , then FF j(d ) is 1.00. For diameters Q c y

smaller than d ·,
The value for average outflow time, t , remains the samet 10
c

as in Example Problem 9.7, or 4.0 hr. The value for average


outflow time, t , thus becomes 4.0 + 6.42 or 10.42 hr.
0 10
p
Fi(^k)
Other computations change also. From Fig. 9.23A, the aver­
age depth above the outlet riser for the period from 4.0 to F F M =
W O * 1 0
- ( A )

10.42 hr is 1.27 ft. Including the 1.0-ft permanent pool, the


average depth is 2.27 ft. Using this depth and the detention This equation effectively normalizes the effluent size distribu­
time, the following concentrations are calculated. tion from layer to the range of 0 to 1.0. The mass corre­
sponding to particles finer than the size d would be k

Average C Q

Layer (mg/liter) (mg/liter) M (d )


oJ k = FF (d )C oJ k Q (B)

1 0.23 4432
where C is the effluent concentration from layer / and q
Q y }
2 0.29 5588
is the water discharge rate from layer /. For uniform with­
3 0.31 5974 5588 drawal,
4 0.33 6359

QJ = 4 / 4 , (C)
340 9. Sediment Control Structures

where q is the total discharge rate for the plug. Hence t h e settling velocity corresponding to l.lq^/A. This
effectively m a k e s V equal to l.lq^/A.
c T h e calculation
KM) = ^C F¥ (d ). ( D )
OJ OJ k p r o c e d u r e s for using t h e E P A model a r e

1. T h e particle size distribution is divided into inter­


Summing over all layers to get the total mass finer than
d, M (d \ vals, a n d t h e m e d i u m d i a m e t e r in each interval is
k Q k

used to c o m p u t e t h e settling velocity, V , for that si

interval.
K(d ) - \ Σ C FF (d ). (E)
k OJ OJ k
2. T h e basin overflow rate, V , is calculated from t h e
c

routed p e a k outflow and t h e effective p o n d surface


The total mass discharged is area at t h e p e r m a n e n t pool using V = Llq^/A.
c

T h e effective a r e a is the area at t h e p e r m a n e n t pool


M T = f EC . 0e> (F)
minus t h e d e a d storage fraction.
3. T h e removal fraction, F, is calculated for each size
interval a n d t h e trapping efficiency, E, d e t e r m i n e d
The fraction finer than d summed over all four layers is
k
using p r o c e d u r e s shown in Example Problem 9.5.
therefore

F F (/I \ ^C 0 ) > FF 0 t y (^) Tapp Method 1 T h e T a p p M e t h o d 1 ( T a p p et al,


FFo(^k) · (G) 1981) is a modification of t h e steady-state overflow r a t e
equation. T h e m e t h o d is based on t h e discharge a n d
Equations (A) and (G) make up the size distribution algo­ effective surface a r e a at a given time. T h e basin over­
rithm for the plug flow model. flow rate for each time interval is calculated as

CQRGO
(9.60)
Variable Flow Rate—Modified Overflow Rate Models CA A t

Because of t h e no-mixing assumption m a d e in t h e


overflow rate derivation given earlier, these type rela­ w h e r e (K ), is t h e basin overflow rate at a given time,
c

tionships could b e classed as plug flow models. T o C O R is a constant, A is t h e basin surface area at a
k

apply t h e overflow rate concept to variable flow mod­ given time, C is t h e fraction of surface area that does
A

els, a flow rate and surface area must b e selected to not contribute to settling, and Q is t h e basin outflow
Q

define t h e overflow r a t e for each time increment. Sev­ rate. T h e calculation p r o c e d u r e is summarized after
eral different approaches have b e e n utilized as de­ t h e following section.
scribed below.
Tapp Method 2 T h e previous two m e t h o d s were
Early EPA Model T h e early E P A Model is essentially based on E q . (9.41), which uses surface a r e a in calcu­
an application of the overflow rate equation to t h e lating t h e basin overflow r a t e . T h e overflow rate m e t h o d
non-steady-state system (Hill 1976). In t h e E P A modi­ was originally developed assuming a basin with vertical
fication of t h e overflow rate method, surface area, A is sidewalls. A m o r e theoretically palatable equation for
set equal to the basin surface area at t h e top of t h e an irregular geometry like that found in a sediment
outlet riser, and Q is set equal to t h e p e a k discharge pond, would b e to consider settling d e p t h and volume
from the reservoir; hence in calculating t h e basin overflow. T h e equation used in
this m e t h o d is
F = (9.59)
1.2 q
po C' Q D
0R 0
(9.61)
where q^ is the p e a k outflow and t h e factor 1.2 is ¥-C V v

added to account for nonideal settling. T h e E P A


m e t h o d provides a poor indication of the effects of w h e r e (K ), is t h e basin overflow r a t e at a given time,
c

basin geometry, inflow hydrograph, and sedigraph C'OR is a constant, Q is t h e outflow rate, D is the
Q

shape on sediment trapping; however, because of its settling d e p t h , ¥ is t h e volume in t h e basin at a given
simplicity, it has b e e n used with some regularity in t h e time, a n d C is t h e fraction of d e a d space volume
v

analysis of surface-mined sediment ponds. T o use t h e accounting for d e a d storage. T h e quantity (V-C V) V

E P A procedures, the argument V /V in Eq. (9.41) is s c would b e t h e effective volume. T h e calculation proce­
replaced with V^A/l.lq^, and X is calculated from c d u r e would be t h e same as t h e T a p p M e t h o d 1, except
Sediment Detention Basins 341
the removal ratio would be calculated using the above Using this diameter in Fig. 9.23B, the value for X is 0.34. c

equation for basin overflow rate. Following procedures given in Example Problem 9.5, the
All three of the modified overflow rate m e t h o d s trapping efficiency for plug 10 is
described may be used with quiescent settling or turbu­
lent settling computations.
U - ( l - * c ) +

Dividing the particle sizes below 0.0064 mm into intervals of


Example Problem 9.10 Illustration of modified ΔΛ' = 0.04, and following procedures from Example Problem
overflow rate models for variable flow 9.5, the summation in the RHS is 0.047. Therefore trapping
efficiency for plug 10 is
Estimate the trapping efficiency of the reservoir in Exam­
ple Problem 9.7 for the time interval given by plug 10 using £ 1 0 = (1 - 0.34) + 0.047 = 0.707.
Tapp method 1. A stage-area-average depth curve is given
in Fig. 9.24 for the reservoir. The effluent concentration for plug 10 would thus be (assum­
ing the same C, as Example Problem 9.7)
1 0
Solution:
1. Determining average surface area for plug 10. From Fig. io = (1 - E )C l0 itl0 = (1 - 0.707)(19,270)
9.23, the average depth for plug 10 is 1.51 ft. Using Fig. 9.24,
the average surface area is given as 0.90 acres or 39,204 ft . 2 ίο = 5646 mg/liter,
2. Calculating overflow rate. From Fig. 9.22, the average
which is similar to the plug flow model. This will not always
discharge rate for plug 10 is 4.5 cfs. Assuming a dead storage
be true.
volume of zero, the overflow rate becomes

ί ο , 10 4.5
(K)io- = 0.000115 ft/sec. Variable Flow Rate—CSTRS Model
A l0 39,204
Plug flow concepts are useful for defining simple
3. Calculating trapping efficiency. Using Stokes' law [Eq. models, but d o not describe the mixing known to exist
(7.4)], the diameter corresponding to the overflow rate is in sediment basins. As will b e seen in a subsequent
discussion of model accuracy, plug flow models predict
ΓνΓ 10.000115 trapping efficiencies with reasonable accuracy, but do
dc = \ = \ —- = 0.0064 mm. not accurately predict timing or magnitudes of sedi­
V V
c
2.81 2.81
m e n t concentrations. T o overcome this deficiency,
Wilson and Barfield (1984) modified the C S T R S con­
cept for sediment p o n d modeling. A schematic of a
reservoir conceptualized as C S T R S is given in Fig.
9.25.
Effluent concentrations are predicted by conducting
a mass balance on each reactor, or

Mass rate into Mass rate out


reactor i of reactor i

Deposition rate R a t e of change of


in reactor i mass in reactor i

For a continuous stirred reactor, effluent concentra­


tion is equal to concentration in the reactor. Hence,
the mass balance for reactor i can b e written as

(9.62)
dt
Stage (ft) w h e r e Q _ and C _ are the water inflow rate and the
T X T X

Figure 9.24 S t a g e - a r e a - a v e r a g e depth curves for Example Prob­ influent sediment concentration, respectively, Q and T

lem 9.10. Average depth is an area-weighted depth over the entire C are the water discharge and the effluent sediment
z

reservoir.
concentration, respectively, £>/?, is the deposition rate
342 9. Sediment Control Structures
Number Of Reactors = π

1st 2nd ith nth


Reactor Reactor Reactor Reactor

Figure 9.25 Pond divided into a series of CSTRS (after Wilson and Barfield, 1985). n, number of
reactors.

of sediment in the reactor, V, is the volume of the the outlet of the pond, or
reactor, and t is time.
{Gjavg - ( G o ) avg
Knowing the concentration in all the reactors at the { G o U + ( / - 1)
start of a time step (C,° for i = 1 to n) and the volume
of each reactor at the start of the time step (vf for (9.64)
/ = 1 to n), the concentration in the i t h reactor, C, at w h e r e ( G J a v g * the average inflow rate for the i t h
s

the end of the time step is determined by using a finite reactor, {Q l is the average flow rate at the p o n d ' s
0 dyg

difference approximation inlet, ( G A v g the average discharge at the p o n d ' s


I S

outlet, and η is the n u m b e r of reactors.


T h e reactor volumes in Eq. (9.64) (V° and V") are
M<a--i «--iU- { & U C P / 2 )
c
- D E P + C/V?
C" = defined as
V? + A r { G } a v / 2; g

(9.63) ¥ j ° = [ P V ( 0 - D E A D ] /η, (9.65)


w h e r e P V ( i ) is the p o n d volume a t time t, D E A D is
the volume of the p e r m a n e n t pool that is d e a d space,
where Δ ί is the time increment used t o route t h e
and η is t h e n u m b e r of reactors. P V ( i ) is d e t e r m i n e d
sediment-laden flow, { G z - i Q - J a v g the average in­
l s

and stored by a hydrograph routing p r o c e d u r e . E q u a ­


flow mass rate during At, { G / } v g is the average outflow
tion (9.65) divides the p o n d volume equally among
a

discharge rate during Δ ί , and V f are the reactor


each reactor.
volumes at the beginning and end of the time incre­
T o u n d e r s t a n d the methodology that is used to esti­
ment, respectively, Cf and C" are the effluent/reactor
m a t e the deposition term DEP, of Eq. (9.63), it is
concentrations at the beginning and the end of the
useful to discuss the differences between t h e sedimen­
time increment respectively, DEP, is t h e mass de­
tology of sediment particles of a plug flow model, such
posited within Δ ί , and the other terms are as previ­
as D E P O S I T S and that of a mixing model such as
ously defined. In Eq. (9.63), ( G i - i Q - i i a v g * defined s

C S T R S . In a plug flow model, each particle of a given


from the previous reactor or from the inflow sediment
inflow slug of sediment has t h e same detention time.
1

graph (for the first reactor) and Cf is a known value


calculated by the previous time step.
T h e average discharge, { G / } , is estimated by inter­
a v g A slug of sediment is defined in this context as the amount of
polating between the average discharge at the inlet and sediment inflow into the 1st reactor during the time step.
Sediment Detention Basins 343

1st 2nd 3rd


Chamber Chamber Chamber

Figure 9.26 Conceptualization of a reservoir represented by the B A S I N model (after Wilson and Barfield, 1985).

As such, a single time p a r a m e t e r can be used to In the B A S I N procedures, the inflow size distribu­
characterize the fall time of this sediment. For a C S T R S tion is divided into a maximum of nine particle size
model, inflow sediment mixes with sediment from pre­ classes. Using the assumption that each layer is com­
vious inflow slugs; therefore, when sediment is dis­ pletely mixed horizontally with vertical diffusion only, a
charged from the pond, it contains some mass from the differential equation for concentration, C, of any parti­
current inflow slug as well as previous inflow slugs. cle size class at any level ζ in reactor i can be written
Consequently, a single time p a r a m e t e r cannot be used as
to represent the residence time of sediment particles
contained in each inflow slug since some sediment
dC QiU (z)
{ r dC d ( dC\
particles remain in the pond for a long time, whereas
other particles are discharged almost immediately. T o
handle a variation in the residence times of individual (9.66)
particles, the C S T R S model maintains a mass balance
for each inflow slug of sediment.
where C is the concentration flowing into a layer from
p
Computational algorithms for the C S T R S model are
the previous reactor, V is the effective volume of
ei
described in detail in Wilson and Barfield (1984). T h e s e
reactor i (actual volume minus d e a d storage), V is s
procedures are too complex for h a n d calculations a n d
settling velocity for the particle class (taken as positive
must be solved with computer software. T h e C S T R S
in the negative ζ direction), ε is turbulent diffusivity,
model is a component of the S E D I M O T II hydrology
Q is inflow to reactor i, and i / ( z ) is the fraction of
t f
and sedimentology computer model (Wilson et al.,
flow through the reactor that moves through a given
1982, 1986) and S E D C A D ( W a r n e r and Schwab, 1992).
level of the reactor, or
Variable Flow R a t e — B a s i n Model
Use of the C S T R S concept allows the effects of i/ (z) =
f t/(z)/i/ avg) (9.67)
mixing to be evaluated, but cannot be used to evaluate
bed scour and resuspension. T o overcome that deficit, where t / ( z ) is velocity at level ζ and t / is average
avg

a model of reservoir sedimentation, known as B A S I N velocity for the section. U (z) is a user input to the
f

was developed (Wilson and Barfield, 1985), in which model that can vary between reactors. It can be esti­
scour is predicted using a modification of the entrain- m a t e d from a known velocity distribution at the inlet to
ment equation from Einstein (1950) and resuspension the reservoir. For the special case where the reservoir
of scoured material predicted by a coupling of reactor water surface is rising slowly, Wilson and Barfield
and diffusion theory. Reactor theory is used to predict (1985) r e c o m m e n d that U (z) in each reactor / be
f

residence times of particles, and diffusion theory is related to that in the previous reactor by
used to model settling characteristics due to gravity
and vertical diffusion due to turbulence. A conceptual­
ization of a reservoir represented by BASIN is given in
Fig. 9.26. iMO = ι - ^ [ ι - t W O ] . (· )
9 68
344 9. Sediment Control Structures

where subscripts / and i - 1 refer to values in reactors necessarily b e equal to upward diffusion and the
i and i - 1. and Q^ can b e estimated from Eq.
x boundary condition is defined by
(9.64). T o start the computation, U can be estimated { 0

from the log velocity profile of Einstein (1950) given by dc


Eq. (10.15), or (9.75)
z= 0

D
(9.69) where 5 is the rate of scour due to turbulence. S is
C c

evaluated by Einstein's (1950) scour function. Coinci-


dentally, the net of scour and settling, F , is n

and
dc
U* l Ζ •VJC-.- (9.76)
U(z) = —log 30.27-* (9.70) z= 0

w h e r e , again, V is positive in the negative ζ direction.


s

where i/* is shear velocity given by (gDS) , g is


gravity, D is reservoir depth, S is channel slope, k is Variable Flow Rate Models—Evaluation of Accuracy
s

channel roughness given by the sediment diameter d , 65 Evaluations of the accuracy of variable flow rate
χ is Einstein's log velocity constant from Fig. 10.10, models have b e e n m a d e with laboratory and field data.
and k is von Karmon's constant ( = 0.4). H e n c e T h e D E P O S I T S model has b e e n evaluated the most
U(z)/Uavg can b e approximated by extensively on both laboratory and field data. In gen­
eral, estimates of trapping efficiency were reasonably
U{z) ln(30.2 z/k ) accurate for both field and laboratory data, but timing
%

ff.o(*) = (9.71) of effluent concentration and magnitudes of effluent


U.av In( 12.27 D/k,)' concentration were poorly predicted. Based on labora­
tory model studies, the modified overflow rate equa­
where D is the depth in the inflow channel. tions were n o more accurate than the D E P O S I T S
Boundary conditions are needed for the solution of model in predicting trapping efficiencies. W h e n com­
Eq. (9.66) at the inlet to the reservoir, at the water p a r e d to the D E P O S I T S model utilizing laboratory
surface, and at the bed of the reservoir. At the inlet to data, both the C S T R S and B A S I N models were more
the reservoir, sediment concentration can be given by accurate in predicting the shape of the effluent sedi-
m e n t g r a p h and peak concentration. Nevertheless, nei­
ther C S T R S nor B A S I N models are more accurate
C(z) =fc exp(-K z/e), (9.72)
than D E P O S I T S in predicting trapping efficiencies. In
1 s

general, B A S I N was n o m o r e accurate than the C S T R S


where k is a constant of integration based on mass
x model in predicting effluent concentration. T h e B A S I N
continuity or model, however, can be utilized to predict resuspen­
sion, whereas the C S T R S model cannot. Detailed com­
DC avg
parisons of model predictions are given in Wilson and
(9.73) Barfield (1985).
f (-K(z/e))U (z)dz'
Q
D
{

ε is the turbulent diffusivity at the inlet, C is the a v g

average inflow concentration at a given time at the Example Problem 9.11. Comparison of CSTRS and
inlet to the reservoir, and D is flow depth at the inlet. DEPOSITS model predictions
T h e boundary condition at the water surface is de­
fined by the statement of zero flux, i.e., upward diffu­ For the storm and reservoir in Example Problem 9 . 8 ,
sion equals downward settling, or compute the trapping efficiency and peak effluent concentra­
tion for the storm using the plug flow model (DEPOSITS)
and CSTRS models. Compare the models. Assume a dead
dC storage volume of 2 0 % .
Kd (9.74)
z =D
8
dz z =D Solution: The solution to the problem requires the use of
computer models due to the complexity of computations.
At the reservoir bed, the rate of settling would not Input data required by the models are summarized below
Sediment Detention Basins 345

1. Storage area discharge data (required of all models). T h e E P A methodology extends t h e single-storm con­
cept to long-term trapping in a reservoir. Stochastic
Discharge concepts are used for developing stormwater and sedi­
Stage Area (acre) (cfs) m e n t discharges into a p o n d along with a reservoir
sedimentation model to predict trapping u n d e r quies­
0.00 0.91 0
cent conditions. T h e overall concept is to utilize the
1.00 0.91 0
single-storm model along with rainfall and runoff
3.00 0.91 5.84 statistics to predict trapping during stormflow and to
6.00 1.50 6.22 use a quiescent model plus statistics on the interarrival
11.00 1.80 6.82 times between storms for trapping after storm flow
16.00 2.20 7.36 ceases. P e r h a p s the weakest link in t h e approach is a
lack of physical basis for the models used for trapping.
T h e single-storm reservoir model used for predicting
2. Special input. trapping of a given particle u n d e r dynamic (storm­
water) conditions is an empirical relationship

No. Inflow Withdrawal


1 Κ Λ-β
Model reactors distribution distribution F= 1 - (9.77)
1 +
βΚ
DEPOSITS 0* Mixed Uniform
w h e r e K is settling velocity of the particle, V is
s c
CSTRS 2 Mixed Uniform
overflow rate defined as Q/A in Eq. (9.40), and β is a
a
Zero reactors actually correspond to an infinite number
turbulence or short-circuiting p a r a m e t e r reflecting the
of reactors using the CSTRS model (see Fig. 9.4). nonideal performance of t h e pond. R e c o m m e n d e d val­
ues for β are
3. Computed values. Trapping efficiencies and peak ef­
β = 1, very poor performance
fluent concentrations were predicted with computerized algo­
β = 2, average performance
rithms. Values are summarized below.
β = 3, good performance
β > 5, very good performance.
Trapping Peak effluent
efficiency concentration F o r β = 1, Eq. (9.77) reduces to
Model (%) (mg/liter)

Plug flow (DEPOSITS)


CSTRS
68.8
72.4
52,527
17,492
κ [κ
F o r small values of VJV , this becomes equal to the
C

quiescent overflow rate equation, but deviates for larger


values. For β = oo, t h e equation becomes

F = ι - - *'\ e
v

Modeling Pond Performance: 1986 EPA which is the fully turbulent model given earlier by Eq.
Urban Methodology (9.57). A comparison of the models is given in Fig. 9.27,
showing that t h e E P A model does not predict trapping
Model Description efficiencies g r e a t e r t h a n the fully turbulent model of
T h e E P A methodology (Driscoll et al., 1986) at­ Eq. (9.57), even u n d e r their so-called " i d e a l " condi­
tempts to evaluate long-term sediment trapping in a tions.
reservoir. T h e m e t h o d accounts for trapping u n d e r T o estimate total sediment removed for a single
storm flow (dynamic) conditions and subsequent set­ storm, Eq. (9.77) would b e applied to all particle sizes
tling u n d e r the quiescent conditions that occur in p o n d s in a p r o c e d u r e like that applied to the overflow rate
after most stormwater is discharged. T o m a k e such an p r o c e d u r e (Example Problems 9.5 and 9.6).
evaluation, the model must account for variations in T o predict long-term trapping, t h e E P A model com­
stormflow and variations in the duration of quiescent bines Eq. (9.77) with stochastically generated flows.
conditions. T o predict settling during quiescent peri­ T h e analysis is divided into dynamic (stormwater) flows
ods, the model must account for the portion of and quiescent flows between stormwater events. Dy­
sediment already deposited a n d that remaining in sus­ namic flows a r e assumed to b e g a m m a distributed and
pension at the time all stormwater is discharged. characterized by a m e a n flow and a coefficient of
346 9. Sediment Control Structures

1.0
< [Coefficient of Variation
of Runoff Voiume CVQ
0.8

0.6

Pi
Pi 0.4
a
u
Η
d 0.2
ο
sο 0.0
0.0 0.2 0.4 0.6 0.8
cd
u
Removal at Mean Runoff Flow
Maximum Removal at Very Low Flow ( E m A ) F

Figure 9.28 Long-term sediment trapping in sediment ponds un­


der storm water flow (after Driscoll et al., 1986).

V is m e a n runoff volume, and Q is the removal rate


R R

defined by Eq. (9.79). Values for T are given inIA

Figure 9.27 Comparison of the E P A and overflow rate models. T a b l e 9B.1 in A p p e n d i x 9B u n d e r t h e column labeled
"Interval."
T h e storage volume u n d e r quiescent conditions in
t h e E P A model is not assumed to b e a fixed quantity,
variation for flow, C V . Using the g a m m a distribution,
Q
but varies b e t w e e n storms. This is accounted for in the
Driscoll et al. (1986) proposed that the total removal
E P A model in Fig. 9.29 by relating t h e ratio of effective
efficiency could be calculated by
p o n d storage volume to m e a n runoff volume, V / ¥ , E R

to t h e ratio of storage volume (empty) to m e a n runoff


D R = L F
cvj (9.78) volume, ¥ / ¥ , a n d t h e removal ratio from Eq. (9.80).
B R

l/CV^-ln(£ /L ) m F T h e effective storage volume ratio from Fig. 9.29, V / E

V is t h e n used in Fig. 9.30 along with the coefficient


R

where D is long-term dynamic removal fraction for


R
of variation of runoff volume to estimate the percent­
stormwater, L is the removal ratio (fraction) for very
F
age removal u n d e r quiescent conditions. Computa­
low flow rates, E is m e a n storm removal fraction
m
tional p r o c e d u r e s are illustrated in Example Problem
[typically estimated from Eq. (9.77)], and C V is t h e Q
9.12.
coefficient of variation of flows. Equation (9.78) is
illustrated graphically in Fig. 9.28. Values for L and f

E must be determined for single storms, as is dis­


m

cussed subsequently.
Driscoll et al. (1986) r e c o m m e n d e d that removal Ε «
«Ε
efficiencies for quiescent conditions between stormwa­
ter flows be estimated from
> %

(9.79) 0.75
QR ~" K^Q>

where Q is quiescent removal rate, V is settling


R S

velocity, and A is surface area u n d e r quiescent condi­


Q

tions. For an average condition, the model defines a α


>\>
removal ratio as II

RR =
T Q
LA R

(9.80)
RY
: =
[jStorage Volume (Empty)
Mean Runoff Volume

Figure 9.29 Ratio of effective storage volume to mean runoff


where T LA is the average time interval between storms, volume (after Driscoll et al., 1986).
Sediment Detention Basins 347

A relationship for c o m b i n e d conditions, i.e., 1.0


stormwater trapping plus trapping during quiescent Ό 0.9
§> to
intervals between storms, is n e e d e d in o r d e r to esti­
m a t e the total long-term fraction removed. A s
11 0.8

stormwater moves through the pond, part of the sedi­


δ !ο 0.7
0.6
C
Q> Ο
ment is retained at the end of the storm as a conse­ E t 0.5
quence of settling, and part is discharged. Of that 0.4
remaining, some is in suspension in the p e r m a n e n t </) 8
0.3
pool and some has already settled to the bottom. Of ° σ
C ΐ- 0.2
that remaining in suspension, some will continue to 0)
Ο Ό

Ο C 0.1
settle out during the quiescent period between storms. Ρ
0.0
Any sediment remaining in suspension at the start of a 0.5 1.0 5.0 10.0
subsequent storm will be discharged. T o accurately
account for this combination of dynamic and quiescent Effective Basin Volume
Ratio:
settling, a complex computer model would be required. Mean Runoff Volume
T h e E P A model recommends a simple alternative Figure 9.30 Long-term removal ratio for sediment ponds under
quiescent conditions between storms (after Driscoll et al., 1986).
E T = 1 (1-£ D )(1-£ Q ), (9.81)

where E is total removal efficiency, Ε


T is total dy­ Ό

namic removal efficiency, and E is total quiescent


Q

removal efficiency. settling velocities and percentages in the table below are
T o use the E P A model, special rainfall statistics are applicable. Calculate the total fraction trapped for both
needed. Values r e c o m m e n d e d by Driscoll et al. (1986) stormflow and quiescent flow. Assume that the reservoirs
are given in Appendix 9B. Computational procedures have good hydraulic performance.
for the model are illustrated in Example Problem 9.12.

Settling velocity,
Particle class Percentage
Example Problem 9.12 Illustration of the /* finer (ft/hr)
EPA model
1 10 0.02

A reservoir is being constructed with the following charac­ 2 30 0.20


teristics: 3 60 1.00

4 80 8.00
Parameter Value 5 100 50.00

Average surface area (principal spillway


to emergency spillway) ft 2
10,000

Storage volume ft 3
40,000 Solution:
Average depth ft 4.0
1. Calculating mean runoff parameters. From the curve
number method in Fig. 3.21, the runoff volume for a rainfall
of 1.14 in. and curve number of 90 is 0.412 in. The peak
The reservoir is located in an area that has the following discharge for the average strom using the rational method
rainfall characteristics (see Appendix 9B) [Eq. (3.70)] is

Parameter Mean Coef. var.


q = C I 4 = (0.8)(0.3 in./hr)(80 acre)
p W

Average rainfall volume (in.) 1.14 1.3


Intensity (in./hr) 0.3 1.3 = 19.2 cfs or 69,120 f t / h r .
3

Duration (hr) 4.0 1.2


It will be assumed that the coefficient of variation of runoff is
Interval between storms T (hr) 100 1.0
the same as rainfall; thus
lA

The watershed has an area of 80 acres, a curve number of 90, C V = 1.3


Q

and a Rational Equation C factor of 0.8. The particle size


distribution for sediment in an average storm is such that the C V = 1.3.
R
348 9. Sediment Control Structures
2. Calculating removal under dynamic conditions. The over­ 3. Fraction trapped under quiescent conditions. The mean
flow rate for the mean storm will be conservatively estimated runoff volume is runoff depth (0.412 in.) times watershed
using the peak flow, or area or
/ 1 ft \ / ft 2
\
¥ = 0.412 in. X — —
R S80 acre X 43,560
69,120 f t / h r3 I 12in.]\ acre
= 6.91 ft/hr.
10,000 ft 2
- 119,645 ft . 3

The reservoir volume is given as 40,000 ft ; hence the ratio of 3

Since the pond will have good performance, β = 3. For


reservoir volume to runoff volume is
particle class 1, K = 0.02 ft/hr. From Eq. (9.77),
s

¥ B 40,000 ft 3

= 0.334.
1 0.02 VR~ =
119,645 ft 3

F = I - = 0.003.
1 +
36^T
The removal ratio from Eq. (9.80) is

Values for the other particle classes are given in the table
_ TQ _ T VA
below. Average trapping over all storms is given by Eq.
n n lA R lA s Q

(9.78). It will be assumed conservatively that the average


trapping is given by Eq. (9.77), or E = F. From Eq. (9.78) m

using C V = 1.3 and assuming that low flow trapping effi­


Q
Assuming that A is the average surface area, and knowing
Q

ciency, L , is 1.0,
F
that T is given by the problem statement as 100 hr,
lA

(100 hr) (10,000 f t ) V 2

1 + 1/1.3 RR = - , = 8.36 V ,
2

1/1.3 2

(119,645 f t ) 3
s

D = 1.0 = 0.023.
1/1.3 - ln(0.003/1.0)
R 2

where V is feet per hour. For a given K , RR can be


s s

calculated and used along with V / ¥ in Fig. 9.29 to B R

Trapping by each particle class averaged over all dynamic determine the ratio of effective storage volume to runoff
conditions is estimated in the table below. volume. This is then used in Fig. 9.30 along with C V to R

Fraction
Settling Single-storm removed
velocity storm over all
Particle Percentage in V a
trapping storms Column 5 χ
SI
class a
class* (ft/hr) Fb
Column 2

1 10 0.02 0.003 0.023 0.23

2 20 0.20 0.028 0.045 0.90

3 30 1.00 0.132 0.094 2.82

4 20 8.00 0.624 0.393 7.86

5 20 50.0 0.975 0.936 18.72

Total 100 E = 30.53


D

Given in problem statement.


fl
v

^Equation (9.77), using β = 3 for good performance, F = 1 - [1


+5

^Equation (9.78), assuming that L = 1.0 and C V F Q = 1.3, D R = 1.0 ^* _ Q


x p )

The fraction trapped under dynamic conditions is thus determine the fraction removed under quiescent conditions.
Computations are summarized in the following, assuming
30.53 that the size distribution does not change between storm flow
E d
- loo" - °· · 305 and quiescent flow.
Sediment Detention Basins 349

Fraction
Effective removed
Settling Removal ratio, volume under
Particle Percentage in velocity RR a
ratio, quiescent Column 6 x
class class V (ft/hr)
s (ft /hr)
3
conditions 0
column 2

1 10 0.02 0.17 0.100 0.09 0.09


2 20 0.20 0.167 0.334 0.25 5.0
3 30 1.00 0.836 0.334 0.23 7.5
4 20 8.00 66.86 0.334 0.23 5.0
5 20 50.0 418.00 0.334 0.23 5.0
Total 100 Σ =23.4

fl
RR = 8.36 V . s

^Figure 9.29 with V /V = 0.334.


B R

Tugure 9.30 with C V = 1.3. R

Thus of the procedures were developed from reservoir sur­


vey data and typically do not reflect many of the factors
23.4
that are known to affect p o n d performance. O n e of the
procedures is based on data generated by a simulation
model.
4. Total fraction trapped under combined conditions.

£ =1-(1-£ )(1-£ )
T D Q
Brune's Method
= 1 - (1 - 0.305)(1 - 0.234) Figure 9.31 shows the empirical curves developed by
= 0.468. B r u n e (1953) with data collected from 44 large reser­
voirs. B r u n e relates t r a p efficiency to the capacity-
inflow ( C / 7 ) ratio described by H a z e n (1904). Consid­
erable scatter exists in the data, and there appears to
Comments on EPA Model b e n o correlation with estimates for semidry reservoirs
that are typical of sediment structures.
T h e rainfall data provided by the E P A contain in­
T h e C/I ratio used as an independent variable in
formation for small storms where runoff is unlikely;
Brune's m e t h o d does not account for sediment charac­
therefore, designers are urged to make conservative
teristics, inflow sedimentgraph, outlet discharge curves,
decisions and consider doubling flow rates and volumes
and outlet discharge distribution variation with depth.
to be conservative. A n o t h e r option is to evaluate the
T h e capacity of a basin is a very poor indicator of the
rainfall data base for a given locale, eliminating all
geometry and anticipated flow conditions within the
small rainfall events not likely to produce runoff, and
basin.
develop a modified set of values for rainfall duration,
intensity, etc.
T h e graphical relationships and equations presented
by the E P A represent, at best, a rough estimate of
long-term trapping efficiency in a pond. It does, at
least, represent a useful framework with which to ap­
proach the problem. T o accurately predict long-term
trapping, a detailed model of reservoir performance
such as the C S T R S model (Wilson and Barfield, 1984)
should be combined with a continuous stormwater and
sediment runoff model to develop estimates.

Other Models of Pond Performance


Several researchers have proposed empirical proce­ Capacity-Inflow Ratio, (ac«ft Capacity)/(ac · ft Annual Inflow)

dures for evaluating reservoir trapping efficiency. Some Figure 9.31 Brune's trap efficiency curves (after Brune, 1953).
350 9. Sediment Control Structures
Churchill's Method
Churchill (1948) developed a m e t h o d based on re­
sults obtained from several T V A reservoirs in which a
sedimentation index is related to trap efficiency as
shown in Fig. 9.32. T h e sedimentation index is the
detention time divided by the mean velocity of the flow
through the basin.
While the method does not account for varying
10<
sediment sizes, many of the other factors affecting 10 10 10 10 10
5 e 7 8 9

sediment transport are directly incorporated in the Period of Retention


Sedimentation Index of Reservoir
Mean Velocity
sedimentation index. T h e small amount of scatter indi­
cates a high correlation with trap efficiency for the Figure 9.32 Churchill's trap efficiency curves (adapted from Brune,
large reservoirs analyzed. 1953).

Based on a theoretical analysis using the overflow


rate concept, C h e n (1975) showed that both Churchill's
and Brune's method overpredicted trapping efficiency HA/M V H R

for small particles and underpredicted trapping effi­


ciency for large particles.
100
The Maryland Model
A simulation analysis of sediment trapping in reser­
voirs used in sediment control for construction sites
Ο
was conducted by McBurnie et al. (1990) on several UJ
watersheds typical of conditions across Maryland. Us­ Ο

ing the S E D I M O T II model, they predicted sediment


trapping in reservoirs with varying pond volumes and Ο
2
surface areas. Inflows were simulated for disturbed CL
0.
watersheds with varying soil types. T h e results indi­ <

cated that the primary variables controlling trapping


efficiency were soil classification and the surface area
to peak discharge ratio. T h e results of the study, shown
in Fig. 9.33, can be used for a very rough estimate of
trapping in sediment detention ponds. T h e surface AC/CFS
area in the horizontal axis is the area at the crest of the SURFACE AREA/PEAK DISCHARGE
principal spillway.
Figure 9.33 Maryland model of trapping efficiency in a pond (after
McBurnie et al., 1990)
Predicting Pond Performance with Chemical
Flocculation
It has b e e n demonstrated clearly that chemical floc- and varying dosages of a given flocculant added. T h e
culants can be used to obtain sediment trapping effi­ mixture is typically stirred at a high speed of 100 rpm
ciencies near 100% if the pond is properly sized and for 2 min to properly distribute the flocculant, and then
adequate mixing is obtained (Tapp et al., 1981; T a p p at 60 rpm for 10 min followed by 40 rpm for 5 min. T h e
and Barfield, 1986; Janiak, 1979). Prediction of p o n d purpose of the lower speeds is to give a high probabil­
performance when using flocculants is a difficult prob­ ity of particle contact and resulting flocculation without
lem since floe sizes are changing due to flocculation in creating such large shear forces that floes break apart.
the reservoir. Procedures developed for predicting sed­ T h e jars are then tested for clay size particles remain­
iment removal with flocculation in steady-state flow ing in suspension. T h e dosage that gives the minimum
will not work with variable-flow rates. concentration of clay-sized particles is the optimum
T h e performance of a given chemical used to en­ dosage. F o r the o p t i m u m dosage, a particle size distri­
hance flocculation is typically determined by a proce­ bution can be developed by withdrawing samples at
dure known as "jar tests." A jar test apparatus consists varying d e p t h s and times as shown in T a b l e 9.5. This
of a stirring platform on which solution in beakers, size distribution information can then be used as input
typically 1000 ml, is stirred with variable-speed paddle to any of the p o n d models described earlier for a first
mixers. S e d i m e n t - w a t e r mixtures are placed in the jars estimate of trapping efficiency. As indicated earlier,
351
Sediment Detention Basins
Table 9.5 Sampling Time and Depths and Corresponding Overflow
Rates and Particle Diameters

Particle Equivalent
settling particle
Time velocity Overflow rate a
diameter''
(min-sec) Depth(cm) (ft/hr) (gal/dayft )
2
(mm)

2-21 10 8.37 1500 0.029


3-31 10 5.59 1000 0.024
5-18 7.5 2.79 500 0.017
8-49 5 1.11 200 0.011
17-40 5 0.56 100 0.007
35-20 5 0.28 50 0.005

a
\ gal/dayft = 40.7 liters/daym .
2 2

^Equivalent particle diameter is the diameter of spherical particle that,


according to Stake's law, settles the indicated depth in the given time at 20°C
and at a particle specific gravity of 2.65.

this would not account for the change in particle size It can be concluded that the use of chemical floccu-
distribution that occurs with flocculation in the pond. lants will improve t h e performance of sediment ponds,
T a p p et al. (1981) showed that the use of the size but procedures for predicting effluent concentrations
distribution determined from these jar test procedures w h e n using flocculants are not highly accurate.
in any of the variable flow models overpredicted the
trapping efficiency of a pilot-scale sediment pond. T h e
probable reason for the overprediction was the lack of Sediment Pond Design Procedures
p r o p e r mixing in the pilot-scale apparatus. T h e jar tests T h e following discussion is an overview of proce­
optimize the mixing process; hence maximum size floes dures for designing a sediment pond. Specific attention
would be obtained. In the pilot-scale a p p a r a t u s , mixing is paid to designing each component of the ponds
was conducted using techniques that would be avail­ shown in Fig. 9.1. A typical reservoir site plan is shown
able at a sediment pond located at a remote site in Fig. 9.34.
w i t h o u t power. This consisted of r u n n i n g t h e T h e procedures p r e s e n t e d are not specific to any one
w a t e r - s e d i m e n t - f l o c c u l a n t mixture through a baffle set of regulations, but are general enough to be adapted
system for the low-energy mixing with a total contact to specific design requirements.
time of 1 min. Such mixing would not be optimum;
hence the floes likely did not reach the maximum size. Designing for Sediment Storage Volume, ¥ s

It should be pointed out that although the floes in the


T h e first decision to be m a d e is the selection of an
pilot-scale apparatus did not reach optimize size, the
appropriate sediment storage volume. This volume is
effluent-suspended solids concentration for the floccu­
simply the storage occupied by the sediment deposited
lant tests were at least o n e o r d e r of magnitude lower
over the given design period. T h e design periods may
than those from identical tests without flocculant.
be the life of the reservoir or the time between clean
Because of the growth of floes in the pond, the outs. Using a c o m p u t e d sediment yield and a bulk
settling characteristics of the sediment change with density of deposited sediment, the sediment storage
time. T a p p and Barfield (1986) modified Eq. (7.7b) to volume is
account for floe growth. Since m e a n square turbulent
velocity is required as an input to Eq. (7.7b), T a p p used
the equations of Li et al. (1980) to predict turbulence. YD
T h e use of flocculation routines improved the accuracy ¥ =
s - , (9.82)
s
Wx 43,560 v
'
of the predictions in ponds; however, the procedures of
Li et al. are not applicable to nonrectangular reser­ w h e r e ¥ is sediment storage volume in acre · feet, Υ
s Ό

voirs. Wilson and Barfield (1986b) c o m p a r e d m e t h o d s is sediment deposited over the design period in pounds,
for predicting turbulence in sediment ponds, as sum­ and W is weight density of deposited sediment in
marized in Appendix A. p o u n d s p e r cubic foot. W can be determined from
352 9. Sediment Control Structures

Designing for Detention Storage Time T o design for


detention storage time, T , a reservoir storage volume
d

SCALE: Γ - 1 0 0 * a n d p e a k outflow discharge that gives t h e specified


detention storage time must b e selected. Procedures
for making these calculations were discussed earlier.
First estimates can be obtained from t h e triangular
approximation given in Fig. 6.12, but should b e checked
by a detailed routing. If the detention storage time is
too short, t h e outflow structure must b e reduced in size
a n d a new routing calculated to d e t e r m i n e a new
outflow hydrograph.
Credit for p e r m a n e n t pool volume can b e calculated
Top Of from Eq. (9.33). Caution should be exercised in select­
Sediment
Storage ing p e r m a n e n t pool volumes that approach the design
storm runoff volume w h e r e t h e r e is n o detention vol­
Emergency u m e ; i.e., t h e design storm does not appreciably in­
Spillway
crease t h e elevation of w a t e r in t h e p o n d . In this case,
t h e principal spillway must b e sized large enough to
Principal
Spillway
pass t h e p e a k of t h e design storm without an apprecia­
ble increase in water surface elevation. Short-circuiting
could be a major problem with t h e large pipe sizes that
result from such a design.
If a series of reservoirs in t a n d e m are being used, n o
acceptable definition of detention storage time has
b e e n established. In this case, it will b e necessary to
Φ Principal inflow Point During 10YR. 24Hr Storm
Principal Row Path During 10YR. 24HR Storm select a series of reservoirs that give an equivalent
Figure 9.34 Example of a sediment pond site plan. trapping t o o n e large reservoir with a given detention
storage time. O n e can show from t h e overflow r a t e
m e t h o d that it is m o r e effective to have o n e large
reservoir t h a n two smaller ones with a combined vol­
Eq. (7.10). Υ is the product of the sediment yield over
Ό u m e equal to that of t h e larger o n e .
t h e design period and t h e trapping efficiency. This
volume is assumed to be stored b e n e a t h t h e crest of Designing for Outflow Concentration T o design for a
the principal spillway as shown in Fig. 9.1. maximum outflow concentration, a design storm must
b e specified. In addition, a hydrograph, sedimentgraph,
Determining Peak Outflow Rate and Storage Volume
a n d particle size distribution must b e determined. If
T h e key problem in sizing a sediment detention these are used as input to o n e of t h e non-steady-state
structure is determining the p e a k outflow r a t e a n d sediment p o n d models, predictions can b e m a d e di­
storage volume to meet a given standard. A n u m b e r of rectly of s u s p e n d e d solids concentration in t h e effluent.
design criteria are utilized in various regions of t h e If t h e D E P O S I T S or C S T R S model that is in S E D I -
world. T h r e e criteria are discussed in this section. M O T II (Wilson et al., 1982) is used, t h e model can
also b e used to predict settleable solids.
1. Sizing the structure and outlet to m e e t a required
T h e above models require c o m p u t e r capability. In
detention time.
t h e absence of a computer, an average effluent concen­
2. Sizing t h e structure a n d outlet to m e e t a required
tration can b e c o m p u t e d from
effluent concentration, either suspended solids or
settleable solids.
3. Sizing the structure and outlet so that t h e p e a k C = (9.83)
outflow during o r after disturbance does not exceed
the predisturbance peak.
w h e r e C is average effluent concentration, Y is sedi­
Each of the criteria will be discussed subsequently. m e n t yield during t h e storm, Ε is c o m p u t e d t r a p
T h e design should b e based on t h e maximum storage efficiency of t h e reservoir (must b e calculated from o n e
and minimum outflow required by the standard as set of the previous procedures), γ is weight density of
by t h e appropriate regulatory authority. water a n d ¥ is t h e volume of runoff. T h e concentration
Sediment Detention Basins 353

predicted will b e t h e average concentration in a storm. d u r e s have b e e n developed by t h e Soil Conservation


O n e major problem with E q . (9.83) is t h e effect of Service (1975, 1986), in what is known as T R 5 5 , for
small errors in predicting t r a p efficiency. F o r example, determining storage volumes a n d p e a k outflow dis­
with a n inflow concentration of 100,000 m g / l i t e r (not charges. Details for making exact computations with
unusual for construction) a 1 % error in Ε c a n make a routing a r e given in C h a p t e r 6. C o m p u t e r models such
1000 m g / l i t e r error in t h e predicted concentration. as T R 2 0 (Soil Conservation Service, 1983) a n d S E D I -
This inherent limitation in t h e m e t h o d must b e recog­ M O T II (Wilson et aL, 1982) can also b e used.
nized.
After t h e effluent concentration is computed, this Sizing Emergency Spillways
can b e converted t o a settleable solids concentration, if T h e return period of a design storm for sediment
the effluent size distribution is known. Procedures for control will typically b e between 2 a n d 10 years. A
calculating t h e effluent size distribution a r e given in reservoir that h a s only sufficient volume t o safely pass
Example Problem 9.5 for t h e overflow rate a n d in this storm will overtop if a n event with a much longer
Example Problem 9.9 for t h e plug flow model. Proce­ r e t u r n period occurs during t h e life of the system. Such
d u r e s for converting a n effluent size distribution into overtopping could result in a d a m breach with subse­
settleable solids a r e given in Example Problem 7.2. q u e n t downstream flooding. T o prevent overtopping,
If t h e predicted effluent concentration is t o o high, a n emergency spillway is designed t o pass t h e rare
the storage volume will n e e d t o b e increased a n d t h e event through a channel constructed in stabilized soil.
outlet pipe size decreased. A n o t h e r option is t o modify This stabilized channel should n o t b e constructed
the size of p e r m a n e n t pool. A s shown in Example through t h e reservoir e m b a n k m e n t .
Problem 9.8, a n increase in p e r m a n e n t pool does n o t T h e r e t u r n period required for t h e emergency spill­
always increase t h e overall trapping efficiency; how­ way design storm d e p e n d s o n t h e hazard classification
ever, p e r m a n e n t pool does improve t h e water quality of t h e reservoir. Example hazard classifications used by
of t h e effluent from t h e first flush of flow. t h e Soil Conservation Service a r e given in Table 9.6
Selecting t h e correct combination of p e r m a n e n t pool along with typical associated design storm return peri­
and temporary detention is a trial a n d error process ods. Classifications a n d return periods a r e set by regu­
that should become easier with experience. N o opti­ latory authorities in each state.
mization algorithms a r e available. O n c e a r e t u r n period precipitation is selected for t h e
emergency spillway, it is translated into an inflow hy­
Design for Peak Discharge Reduction D e s i g n for p e a k drograph. If t h e reservoir is small, t h e emergency spill­
discharge reduction was discussed in C h a p t e r 6 u n d e r way is usually sized t o safely pass t h e p e a k inflow rate.
the title Flood Peak Reduction. Approximate proce­ If t h e reservoir is large, t h e flood should b e routed

Table 9.6 Dam Hazard Classification and Return Period Storm for Emergency Spillway Design*
Hazard Height-storage volume Typical design storm
classification* Description relationship0
for emergency spillways'*

Dams in rural and agricultural areas where Sx//<3,000


failure may damage farm buildings, 3,000 < S xH< 30,000 /> 100 + 0.12(PMP-/> ) 100

agricultural lands, townships, or country roads 30,000<5x// P I 0 0 + 0.26 (PMP - / > )1 0 0

Dams in predominantly rural and agricultural P + 0.40 (PMP - P )


l0O 1 0 0

lands where failure may damage isolated


houses, main highways, or minor railroads, or
cause interruption of relatively important public
utilities.
Dams located where failure may cause loss of PMP
life, serious damage to homes, industrial and
commercial buildings, important main
highways or railroads.

Adapted from Earth Dams and Reservoirs, TR 70, Soil Conservation Service.
fl

^Hazard classifications and return periods are specified by regulatory agencies. Description and values given here are typical.
5 , storage in acre feet. / / , dam height in feet.
c

= 25 year, 24-hr precipitation. P m = 100 year, 24-hr precipitation. PMP, probable maximum precipitation.
354 9. Sediment Control Structures
through the reservoir to take advantage of the reservoir Row Separator U Remaining Runoff - e d DetenUon
storage. T h e size of the emergency spillway for the
routed storm will be less than that required to pass the Firet 1 / 2 Inch Runoff
peak flow. Procedures illustrated in C h a p t e r 5 can be
used to size the spillway.
SedimenteUon FUtraUon
Geotechnical Considerations Figure 9.35 Flow chart for a first-flush filtration system.
In the design of a reservoir, attention must be given
to seepage through and under the dam and slope
stability. T o prevent this seepage, cutoff trenches are
sedimentation basin is filled with the required volume
constructed below the dam and antiseep collars are
for the first flush. In design calculations, t h e elevation
located along the principal spillway. In addition, the
corresponding to this first-flush volume would b e set as
dam must be properly compacted from appropriate
the elevation of the crest of the diversion weir in Fig.
materials. Details are given in appropriate soil mechan­
9.37. A t this point, water overtops the diversion weir
ics texts.
and most of the subsequent storm flow diverts directly
to the storm water detension basin.
Design of Ponds for Combination Sediment and W a t e r in the sedimentation basin is detained long
Stormwater Control e n o u g h to allow coarse particles to settle out. T h e
In some situations, ponds are used for both sediment remaining finer particles are discharged along with the
and storm water management. In this case, the final first-flush volume o n t o a filtration basin with a sand
design must accomplish both objectives. Typically, sedi­ b e d filtration system and an u n d e r d r a i n piping system
ment control standards are set for a more frequent as shown in Fig. 9.36. T h e volume of the sedimentation
return period, i.e., 2- to 10-year storm and storm water basin is set by the first-flush runoff to be treated,
control for a more rare event, i.e, a 10- to 100-year typically ^ to 1 in. T h e surface area of the sedimenta­
storm. As shown in C h a p t e r 10, the channel forming tion basin is set by overflow rate and particle size
event is a 2-year storm; thus some regulatory authori­ trapping requirements. If it is assumed that 100% of
ties require basin design for storm water control for particles with settling velocity V are required to be sr

2-, 10-, 25-, and 100-year events. This m e a n s that peak t r a p p e d in the sedimentation basin, then the overflow
storm water discharges after land disturbances must rate given by Eq. (9.40) can be used to d e t e r m i n e the
match those of predisturbed conditions for each of the required surface area, or
return periods. Design procedures for multiple return
period storms, therefore, must assure that the peak K_
= 1.0
discharges from the basin for each of the return period Q/A
events do not exceed those prior to disturbance. In
addition, if the basin is to be used for sediment control, and
the design must be checked for sediment effluent.
(9.84)
Designing for First Flush Filtration
After a disturbed area has been developed and stabi­ T h e flow rate can be approximated by a steady^state
lized, the highest concentration of sediment and o t h e r rate averaged over the flow time t , or f

pollutants frequently occurs during the first flush of


runoff when dry deposition materials are being washed D F F /4 W

A = (9.85)
from the unsaturated surface. A technique for control­
ling sediment in postdevelopment runoff is to use
first-flush filtration. In this process, the first \ to 1 in. where D F F is the d e p t h of runoff to be diverted, t is f

of runoff is diverted to a sedimentation/filtration basin the drawdown time, and A w is the watershed area.
and the remainder of the runoff goes directly into a A p p r o p r i a t e units should b e used. A value for K sr

storm water detention basin as shown conceptually in should be selected on t h e basis of a minimum particle
Figs. 9.35 and 9.36. A n example of a first-flush diverter size to b e trapped. T o prevent clogging of the filter
is shown in Fig. 9.37. As storm water flows into the beds, o n e would anticipate trapping most of the coarse
inlet channel in Fig. 9.36, the first-flush flows through sediment in the sedimentation basin and the remainder
the slots on the isolation baffle into the sedimentation of the fines in the filtration basis. T h u s a particle
basin shown in Fig. 9.36. This flow continues until the diameter that represents a large fraction of the sedi-
Sediment Detention Basins 355

Flood
Control Stormwater
Isolation Weir Outlet Flood Control
and Diversion Detention/Retention
Baffle Basin

Inflow

Sedimentation
Basin Emergency
(Full Water Quality Spillway
or F i r s t F l u s h Volume)

Subsurface
Drain
Perforated
Pipe S y s t e m

Filtered Outflow

T r a n s p o r t into Underdrain Piping


Sedimentation Basin System

Figure 9.36 Example of a complete first-flush filtration system. ( A ) Conceptual plan view. (B) Elevation A - A'.

ment in a postdevelopment situation should be se­ small enough to regulate the flows would likely plug
lected. T h e city of Austin, Texas, recommends a value from debris.
of 20 μ m (Austin, 1988) based on data from the T h e first-flush volume should be the runoff that
Environmental Protection Agency (Driscol et al., 1986). contains the first flush of material washed from the
Using a diameter of 20 μ,ιη or other site-specific mea­ surface. T h e first flush is often called water quality
surements, a corresponding settling velocity V can be sr volume and should include runoff from all impervious
calculated from t h e minimum value of Eqs. (7.5) and surfaces such as roadways, rooftops, and parking lots,
(7.6). T h e drawdown time, t , should be selected on the
{ plus all runoff from previous areas that drain onto
basis of interarrival time between storms as well as impervious areas. Typically this is assumed to be \ in.
safety and site characteristics. Values between 24 and of runoff (Austin, 1988).
60 hr are currently being used (Austin, 1988). Flow T h e surface area of the filter chamber is controlled
controls for the drawdown should not be based on pipe by the permeability and d e p t h of the sand bed. A
size, but should be manually adjustable. Pipe sizes minimum d e p t h of 18 in. is typically recommended
356 9. Sediment Control Structures
Top of Isolation Baffle Must be Greater
Than Maximum Water Surface Elevation
Over Diversion Weir For 25 Year Storm

Figure 9.37 Details of a system to divert the first flush of storm water. T h e elevation of the crest of the
diversion weir is set to allow overflow as soon as the first-flush volume is satisfied.

\
(Austin, 1988). Using Darcy's law for flow, Eq. (11.5),
the velocity of drainage flow through t h e filter is given
by Elevation at t = 0

AH
* A L ' · Unit
Ho
K P C R = ( 9 8 6 )

Area
Η at t
where V^ is flow through or percolation velocity (feet
x

p e r second), Κ is hydraulic conductivity (feet p e r sec­


ond), AH is head loss in flow through the filter, and
A L is the distance over which the h e a d is dissipated
(see Fig. 9.38). If the d e p t h of ponding on t h e filter at Filter fc

Media
any time is given by Η and the thickness of t h e filter is
given by L , then AH is given by Η + L and A L is
f f

given by L . T h e flow through velocity then becomes


f

Outflow Velocity = V p e r

Figure 9.38 Definition sketch for continuity relationship as ap­


K ^ - t f — — ( 9 . 8 7 ) plied to filtration chamber.

T h e surface area is set t o allow the p o n d e d w a t e r to water, or


drain in a specified time t . T o d e t e r m i n e t h e d e p t h
dr

that can b e removed, o n e must utilize the continuity dH/dt = -K (9.88)


equation. For the purposes of developing equations to
compute drawdown times, assume a column of water U s i n g D a f C y , s ^ E q m n for y

per'
initially at height H as shown in Fig. 9.38, with n o
Q
p

additional inflow. For this no-inflow situation, t h e rate ^ Η + L


of outflow p e r unit area of the water column, — = -K— . (9.89)
must equal the change in height of the column of at L {
Sediment Detention Basins 357

Separating variables a n d integrating yields Solution:


1. Sedimentation basin design. The runoff volume to be
Η + L = (H f Q + L )e~ <.
{
Kt/L
(9.90) diverted is V = D > 1 , where A^ is the watershed area,
d i v F F W

or
T h e total volume of water, C , p e r unit a r e a that V t P c r

has passed through t h e filter a t any time t is r


div = ( ^ f t ) (20 acre X 43,560 — ) = 36,300 ft . 3

\ 12 / \ acre /
C , = fV^dt (9.91)
V p e r
Assume that a 20 μπι (0.02 mm) particle is to be trapped in
the basin, the required settling velocity is determined from
or using Darcy's law for K , per
Eq. (7.6) in centimeters per second as

logioKr - - 0 . 3 4 2 ( l o g < 0 + 0.989 log </ + 1.15 1 0


2
10

r.per = f*K-—;—— dt. (9.92)


= - 0 . 3 4 2 ( l o g 0 . 0 2 ) + 0.989 log (0.02) + 1.15
10
2
10

= -1.518
Substituting E q . (9.90) for Η + L a n d simplifying f

y sr = io~ 1 5 1 8
= 0.030 cm/sec or 3.54 ft/hr.
Cy.pe r = (tf 0 + L )(l f ( ·
9 9 3
)
Using Eq. (9.85) to define surface area for the required
If C is set equal t o t h e total volume above t h e drawdown time and settling velocity
V t P e r

filter, H , t h e n E q . (9.93) becomes


Q
D y4 36,300 ft 3
F F w

A =
H -L (e " <-l).
Q t
K L
(9.94) tV f ST - (24 hr)(3.54 ft/hr)

Equation (9.94) defines t h e d e p t h of water that can b e A = 427 f t or 0.01 acre. 2

drained through a filter of thickness L with hydraulic f


Thus, the settling basin should have a surface area of at least
conductivity, Κ in time t. This can b e used t o define 427 ft . The required depth would be
2

t h e filter area required. If is t h e d e p t h of runoff


36,300 ft 3

from watershed area A , which is diverted, t h e n t h e w D = ^ - - 85 ft > 6.0 ft.


filter surface area A is given by 427 ft 2

Use D — 6.0 ft based on the safety requirements set by the


A — D A /H , (9.95)
{ FF W 0
problem statement. The surface area becomes
where H would b e defined by t h e required drainage
Q
36,300 ft 3

time. A maximum constraint for H would also typi­ - 6050 ft or 0.14 acre. 2

6.0 ft
Q

cally b e defined based o n safety a n d site requirements.


Computational procedures for first-flush filtration d e ­ Since the surface area is much grater than that required to
sign a r e given in Example Problem 9.13. trap a 20-μπι particle, the discharge time can be modified.
Using Eq. (9.85), the minimum discharge time now becomes

ϋ Α^
ΨΈ 36,300 ft 3

Example Problem 9.13 Illustration of first-flush


AV (6050 f t ) · (3.54 f t / h r ) 2
= 1.69 hr.
ST

filtration
An outlet must be sized to allow the sedimentation basin to
Runoff from a 20-acre watershed is to be controlled by a completely discharge in 1.69 hr or more. This would typically
storm water detention basin with first-flush filtration capabili­ be a slotted riser with an orifice on the outlet barrel. The
ties, including a sedimentation basin and a filtration basin. riser would be wrapped with filter cloth to prevent coarse
Regulatory requirements dictate that the first 0.50 in. of material from entering.
runoff be diverted. The filter is to be 18 in. thick with 2. Filtration basin. The maximum head on the basin to
sufficient underdrain capacity to remove peak flows. A sand allow discharge in 36 hr is given by Eq. (9.94) as
that has a hydraulic conductivity of 3.5 ft/day is to be used.
The sedimentation basin is to be drawn down in 24 hr or less. H - L {e " *
Q f
K L
- 1)
The filtration basin is to be drawn down in 36 hr. Water
Kt /I 3.5 ft/day
ft/day \
depth in the filtration and settling basins must not exceed (36 h r ) / 1 . 5 f t = 3.5
6.0 ft. Determine the area of the sedimentation basin and the T
L {f
=
\\ 24
24 hr/day
hr/day )
filtration basin. Following the recommendations of the City
H - 1.5 ft (e 3 5
- 1) = 48.2 > 6 ft.
of Austin, Texas, assume that 20-μπι and larger particles Q

should be trapped in the sedimentation basin. The maximum allowable head was stated to be 6.0 ft; thus
358 9. Sediment Control Structures
use H of 6.0 ft and reduce the drawdown time. The new
Q Table 9.7 Range of Measured Long-Term
drawdown time (which does not need to be controlled) is Pollutant Removal for Sediment Detention Basins
solved from Eq. (9.94), or (after Stahre and Urbonos, 1990)

6 ft = 1 . 5 f t ( e * , / L
' - 1). Item Removal percentage

Solving for Kt/L ,{


Total suspended solids (TSS) 50-70
Total phosphorous (TP) 10-20
Kt/L = f l n | ^ + l j = 1.609 Nitrogen 10-20
Organic matter 20-40
(1.609)(1.5 ft)
Lead 75-90
' " 3.5 ft/day
Zinc 30-60
= 0.690 days or 16.6 hr. Hydrocarbons 50-70

The required surface area is Bacteria 50-90

D^A„
'FF^w 36,300 ft 3

A = = 6050 ft or 0.14 acre.


2

6.0 ft
(3) T h e full volume of the basin should be drained in
The total surface area for the sedimentation basin and filtra­ 40 hr or less.
tion basin is 0.14 + 0.14 or 0.28 acre.
Stahre and U r b o n o s (1990) summarize the data from
Grizzard et al. (1986), Occoquan W a t e r s h e d Monitor­
ing Laboratory (1986), and Whipple and H u n t e r (1981)
N o routing time delays are accounted for in the
and propose the range of trapping percentages for
procedures above. T h e time delay for flow from the
extended detention basins as given in Table 9.7. T h e
sedimentation basin along with the discharge time from
values given are long-term averages. Considerable vari­
the filtration basin during inflow periods would m a k e
ation exists between storms. In some cases, a negative
the above analysis conservative. A more detailed analy­
trapping efficiency will exist for certain stroms, result­
sis using routing techniques could be used to refine the
ing from resuspension of sediment a n d / o r diffusion of
design and reduce the surface areas.
chemicals from previously deposited sediment. O t h e r
Provisions would need to be m a d e to clean the filters
data are given in Randall (1982) and R e e d (1978).
or provide for replacing the filter material.

Effect of Sediment Basins on Water Quality CONSTRUCTED WETLANDS


A portion of the sediments deposited in a reservoir Native Wetlands
will be in the clay fraction and have surfaces with a
charge density. This charge density, which varies with Native W e t l a n d s have b e e n defined by H a m m e r and
clay mineralogy, is known as the exchange p h a s e of a Bastian (1989) as
soil and attracts ions from the solution. A substantial
an ecotone, an edge habitat, a transition space
fraction of the total dissolved solids load ( T D S ) in a
between dry land and d e e p water, an environ­
flow will b e tied u p on the exchange phase and hence
m e n t that is neither clearly terrestrial nor
will b e t r a p p e d in a basin along with sediment. Present
clearly aquatic. (p. 5)
technology is not sufficiently developed to theoretically
predict the amount of dissolved chemicals t r a p p e d in a T h e U.S. Fish and Wildlife Service (FWS) also rec­
reservoir, although Evangelou et al. (1987) propose ognizes wetlands as a transition zone with a water table
procedures based on equilibrium chemistry for a lim­ that is periodically at or n e a r the surface or covered by
ited n u m b e r of metals. H e n c e , one must resort to shallow water, meeting o n e or m o r e of the following
empirical studies to define the effects of sedimentation criteria (Cowardin et al., 1979):
basins on water quality.
(1) A r e a s that primarily support hydrophyles.
Grizzard et al. (1986) recommend the following cri­
(2) A r e a s with predominantly u n d r a i n e d hydric soils
teria for dry basins designed for water quality:
(wet e n o u g h for long enough to produce anaerobic
(1) Detain the runoff volume from an average annual conditions that limit the type plants).
storm for 24 hr. (3) A r e a s with nonsoil substrate (such as sand or gravel)
(2) T h e runoff volume detained should be larger than that are saturated or covered with shallow water at
the volume of runoff from an average rainstorm. some time during the growing season.
Vegetative Filter Strips and Riparian Vegetation 359

This description by the F W S encompasses areas that i n s e c t / p e s t outbreaks, plant species should be mix­
are traditionally called swamps, marshes, and bogs. t u r e s instead of monocultures.
Swamps are wetlands that contain primarily woody T h e plant types present have the ability to transport
species, marshes are areas that contain soft-stemmed e n o u g h oxygen to their roots to survive anaerobic con­
plants, a n d bogs contain p r e d o m i n a n t l y mosses ditions. T h e film aerobic region, called the rhizosphere,
( H a m m e r and Bastian, 1989). that is p r o d u c e d a r o u n d t h e roots is essential to the
T h e natural dynamics of wetlands lead to alternative microbial population that modifies nutrients, metallic
wetting and drying as the area receive, hold, a n d recy­ ions, and o t h e r c o m p o u n d s . T h e microbial population
cle nutrients and runoff from u p l a n d areas. T h e wet­ is essential to t h e operation of wetlands and is, fortu­
ting and drying t e n d s to e n h a n c e native productivity. nately, ubiquitous, naturally occurring in most waters.
Continually wet areas have a different species t h a n T h e y are fast growing and genetically plastic and thus
alternatively wet and dry areas. can b e rapidly a d a p t e d to a variety of incoming pollu­
According to H a m m e r a n d Bastian (1989), t h e most tants. Because of this plasticity, wetlands have b e e n
important and least u n d e r s t o o d function of wetlands is used to treat u r b a n wastewater, acid mine drainage,
in water quality improvement. A s a result of complex highway runoff with high lead concentration, man­
chemical and biological interactions, these areas can ganese, and runoff laden with organics and nutrients.
remove or convert large quantities of chemicals from a They provide a simple alternative to physical t r e a t m e n t
variety of point and nonpoint sources, including sedi­ (see Fig. 9.39).
ment, metals, nutrients, a n d organic matter. Some Constructed wetlands should not be used as a pri­
chemicals are completely immobilized and some a r e mary settling system for sediment since sediment parti­
broken down into simpler forms. cles will t e n d to cause deltas a n d clog flow systems. If
sediment loads a r e high, p r e t r e a t m e n t is necessary.
Constructed wetlands a r e best used for removing con­
Constructed Wetlands
t a m i n a n t s o t h e r t h a n sediment from t h e flow, and thus
Constructed wetlands are defined by H a m m e r a n d should b e used downstream of a sediment control
Bastian (1989) as structure. H a m m e r (1989) edited a book that contains
an excellent summary of t h e state of t h e art in wetland
design.
a designed and m a n - m a d e complex of satu­
rated substrates, e m e r g e n t and submergent
vegetation, animal life and water that simu­
lates natural wetlands for h u m a n use and
VEGETATIVE FILTER STRIPS AND RIPARIAN
benefits. (p. 12)
VEGETATION

O t h e r terms used for constructed wetlands include Description of Vegetative Filter Strips
artificial, m a n m a d e , and engineered. Vegetative filter strips (VFS) are zones of vegetation
T h e most promising type of wetland for construction t h r o u g h which sediment a n d pollutant-laden flow are
purposes is a marshland, d u e to its adaptability to directed before being discharged to a concentrated
fluctuating water levels and rapidly growing vegetation flow channel. This control technique has b e e n de­
( G e a r t h a r t et al., 1983). Vegetation that is likely to b e scribed by a n u m b e r of different t e r m s including vege­
successfully grown in m a n m a d e wetlands includes cat­ tative filter strips, grass filters, grass filter strips, buffer
tail, bulrush, rush, and giant reed, all of which tend to strips, riparian vegetation buffer strips, and constructed
be adapted to high pollutant concentrations as well as filter strips. F o r t h e p u r p o s e of this text, two main
fluctuating water levels (Small, 1976). classifications will b e used with subclassifications for
Elements of constructed wetlands include ( H a m m e r each:
and Bastian, 1989)
C o n s t r u c t e d filter strips: Filter strips that are con­
(1) subsurface flow zone with varying rates of hydraulic
structed a n d m a i n t a i n e d to allow for primarily over­
conductivity
land flow t h r o u g h t h e vegetation. Vegetation is
(2) plants adapted to anaerobic conditions
grass-like plants with density approaching that of tall
(3) a water column above t h e surface
lawn grass.
(4) invertebrates and vertebrates
N a t u r a l vegetative strips: Any natural vegetative area
(5) an aerobic and anaerobic population of microbes.
t h r o u g h which sediment-laden flow is directed, in­
T h e subsurface zone includes the p o r o u s zone plus cluding riparian vegetation a r o u n d drainage chan­
the zone of deposited litter and debris. T o survive nels. Flow is typically not b r o a d overland sheet flow,
360 9. Sediment Control Structures

Inlet

Emergency
Spillway

Β Trash Hood

Pool Anti-Seep
Collars
Figure 9.39 A n example of a constructed wetland integrated as a part of a sedimentation basin. ( A ) Plan view.
(B) Section view.

but occurs in small concentrated flow channels or


flow zones. These channels occur as a result of
channelization resulting from the natural topography
as well as a result of the deposition delta that
frequently forms at the leading edge of the vegeta­
tion. Vegetation can range from grass-like plants to
brush or trees with ground litter.

Riparian vegetative buffer strips are strips of veg­


etation that grow along stream and concentrated flow
channels. T h e vegetation may b e constructed or natu­
ral.
T o b e effective, the V F S will normally be located on
Β
the contour perpendicular to the general direction of Bedload Suspended Load
DepoaiUon Zone
flow. A schematic of a typical V F S is given in Fig. 9.40. Zone
Source
Area

Effectiveness of VFS in Trapping Sediment


Deposited
It has been established that V F S are effective as a Sediment
nonpoint pollution control device when properly uti­ Figure 9.40 Schematic of a typical vegetative filter strip. ( A ) Plan
lized. M e a s u r e m e n t s o n naturally occurring vegetative view. (B) Profile view.
Vegetative Filter Strips a n d Riparian Vegetation 361

strips have shown that they are very effective in remov­ techniques can also b e used for design, as discussed
ing sediments and some dissolved solids, but the subsequently. Differences b e t w e e n constructed and
degree of the effectiveness is not clearly established naturally occurring V F S are discussed in this section,
except for a few specific cases (Dillaha et al., 1986, showing why the m e a s u r e m e n t and prediction tech­
1988; Cooper and Gilliam, 1987). Operational models niques can be be projected from o n e to the other
to evaluate the effectiveness of naturally occurring without modification. In t h e absence of any specific
strips have not b e e n developed, although a research data, a discussion is given of how o n e might use experi­
oriented model is available for natural grassed areas ence and predictions from constructed V F S to give a
(Inamdar, 1993; Barfield et al., 1993). Conversely, mea­ first estimate of the effectiveness of natural V F S .
surements and prediction m e t h o d s are fairly well de­ W h e t h e r designed and constructed or occurring nat­
fined for constructed V F S (Hayes et al., 1984) a n d urally, V F S remove solids primarily by t h r e e mecha­
evaluations can be m a d e of their effectivness. T h e s e nisms:

Figure 9.41 Illustration of the trapping mechanisms in V F S . N o t e that C(t) is a short zone and its length is typically
ignored. (A) Overview of all mechanisms. (B) Trapping mechanisms in zone D(t).
362 9. Sediment Control Structures
(1) Deposition of bedload material and its attached T h e s e trapping mechanisms are shown schematically
chemicals as a result of decreased flow velocities and in Fig. 9.41. T h e degree to which a V F S traps sediment
transport capacity. Such deposition occurs in a deposi­ thus d e p e n d s on a n u m b e r of factors. A factor of
tion wedge, either at the leading edge of the V F S or in considerable importance is the degree of channeliza­
a p o n d e d area upslope of the VFS. tion that occurs. If runoff occurs as shallow overland
(2) Trapping of suspended solids in the litter col­ flow spread fairly uniformly across the width of the
lected at the surface. W h e n suspended solids settle to filter, m o r e sediment will be t r a p p e d than if the flow
the bed, they are t r a p p e d in the litter at the soil occurs in o n e or m o r e small concentrated flow chan­
surface instead of being resuspended as would occur in nels. In channelized flow, d e p t h of flow will be greater,
a concentrated flow channel. W h e n the litter becomes the transport capacity greater, and hence the a m o u n t
inundated with sediment, trapping no longer occurs as of sediment transported through the filter will b e much
a result of this mechanism. greater (trapping efficiency will be less) than for the
(3) Trapping of suspended material that moves into case of shallow overland flow. Also, if flow is con­
the soil matrix along with infiltrating water. This is the centrated, the area w h e r e infiltration is occurring is
primary mechanism by which dispersed colloidal parti­ reduced, which reduces the trapping of suspended sed­
cles are trapped. iment.

Table 9.8 Factors Affecting Trapping Efficiency for Vegetative


Filter Strips

Parameter Effect on trapping

Flow rate Trapping inversely related to flow rate.


Size distribution Trapping directly related to particle size. Larger parti­
cles are more easily trapped. Sand size and larger parti­
cles can be trapped with short filters on relatively steep
slopes. Clay size particle are trapped only by infiltration
that occurs on long relatively flat slopes and highly per­
meable soils.
Slope Trapping inversely related to slope.
Type of vegetation Vegetation that grows in clumps tends to be less effec­
tive than uniform growth.
Density of vegetation Trapping directly related to the density of vegetation.
Stiffness of vegetation Trapping related to stiffhes and height of vegetation. If
the vegetation is not sufficiently stiff to remain erect
under storm flow conditions, trapping will be greatly
reduced as vegetation lays flat.
Height of vegetation Theoretically, trapping increases with height, assuming
that vegetation remains erect. Actually, tall vegetation
does not naturally remain dense near the ground, result­
ing in a decreased trapping based on the number of veg­
etal elements per unit area at the ground surface. Also
taller vegetation tends to lie flat during storm flow,
which reduces trapping.
Infiltration rate Trapping of suspended sediment, particularly colloidal
and small silt particles, is directly proportional to infil­
tration rate.
Mass of litter The depth of surface litter is directly related to the
amount of sediment that can be trapped before the
filter looses its effectiveness. For tree or brush like veg­
etation, trapping of sediment typically occurs as a result
of the litter; thus trapping efficiency is directly related to
mass of litter.
Degree of channelization Trapping is inversely related to the degree to which flow
is channelized. Channelized flow is deeper with higher
velocities than overland flow, resulting in higher trans­
port capacities and reduced trapping efficiencies.
Vegetative Filter Strips and Riparian Vegetation 363
Based on studies conducted by Hayes et al. (1978) m a d e between flow in vegetation with a spacing of 5 S

and Tollner et al. (1976), it was found that the trans­ and flow in a rectangular channel with a flow depth of
port of sediment, and hence trapping of sediment, in d , or
{

VFS that are designed and constructed to minimize or


eliminate concentrated flow is a function of a n u m b e r
of variables as shown in Table 9.8. Equations that V= m j^)/*V3 i/2 5 ( f t / s ec) (9.96a)
predict sediment trapping as a function of these vari­
ables are presented in a later section. or
Examples of the measured trapping efficiencies of
constructed and natural V F S are given in Table 9.9,
showing that the ability of the V F S to t r a p sediment is
quite high. F r o m this information, it is obvious why V m = (i£) R / sy
2 3 2
(m/sec), (9.96b)
V F S are r e c o m m e n d e d practices, particularly riparian w h e r e xn is a calibrated value for Manning's rough­
VFS. ness, 5 is the slope of the channel, and R is a
C s

hydraulic radius based on average spacing of the media


Analysis of Flow in VFS elements and the flow d e p t h , or

Flow velocities can be predicted in V F S by using a


specially calibrated form of Manning's equation (Hayes (9.97)
et al., 1978; Tollner et al., 1976) in which an analogy is 2d f + 5 S

Table 9.9 Examples of Measured Trapping Efficiency for Vegetative Filter Strips
Study location Sediment
and type trapping Impact on
vegetation Description of site efficiency water quality Reference

University of Constructed filter field site, 87-99% Not measured Hayes et al. (1984)
Kentucky length approximately 30 m
Fescue ( K Y 3 1 ) slopes 3-20%
University of Constructed filter at strip 70-90% Not measured Barfield and
Kentucky, SE KY mine site, 70 m slope 20%, Albrecht(1982)
Fescue (KY 31) monitored natural rainfall
for 1 year
Mississippi Contructed filter for > 90% Not measured Hayes and
State University construction site, high clay Harriston(1983)
Fescue (KY 31) soil, monitored natural
rainfall for 1 year, slope 3%,
30 m long.
North Carolina Natural riparian forested > 50% 50% Total Ρ Cooper et al.
State University area in Middle Coastal (1987)
Wetland Plains of North Carolina
Vegetation
Virginia Nine constructed VFS 81-91 % 5 0 - 6 9 % Total Ρ Dillaha et al.
Polytechnic controlling feedlot runoff. 6 0 - 7 4 % Total Ν (1986, 1988)
University Lengths 4.6 to 9.1 m
Orchard Grass with 5 to 15% slopes
Same type plots, controlling 70-84% 6 1 - 7 9 % Total Ρ Dillaha et al.
runoff from cropland 5 4 - 7 3 % Total Ν (1989)
North Carolina 4.3- and 5.3-m plots 70% 50% Ortho Ρ Parsons et al.
State University controlling runoff from 26% Total Ρ (1991)
Crab Grass and croplands 50% Total Ν
Bermuda
Maryland 4.6- and 9.2-m filters 66% 27% Total Ρ Magette et al.
K Y 3 1 Fescue controlling drainage from 0% Total Ν (1989)
plots spread with poultry
manure and liquid nitrogen
364 9. Sediment Control Structures

Table 9.10 Values for Calibrated Manning's Roughness xn Vegetative Density and Vegetative
Stiffness for Various Vegetative Types

Density* Max. height Calibrated Stiffness'*


Retardance (stems/ft / 2
before Manning ΜΕΙ
class* spacing in.) mowing c
roughness, unmowed/mowed Type
Vegetation unmowed/mowed **/s c
(in.) xn N/m2
stand

Vegetation typically recommended for VFS


Yellow bluestem AID 250/0.76 NA NA 300/0.1 Good

Tall fescue B/D 360/0.63/ 15 0.056 20/0.1 Good


Blue grama B/D 350/0.64 10 0.056 20/0.1 Good
Ryegrass (perennial) B/D 360/0.67/ 7 0.056 20/0.1 Good
Weeping lovegrass B/D 350/0.64 12 NA 20/0.1 Good

Bermudagrass B&C/D 500/0.54 10 0.074 9/0.1 Good

Bahiagrass C/D NA 8 0.056 5/0.1 Good


Centipedegrass C/D 500/0.54 6 0.074 5/0.1 Good
Kentucky bluegrass C/D 350/0.64 8 0.056 5/0.1 Good
Grass mixture* C/D 200/0.85 7 0.05 5/0.1 Good

Buffalograss D/D 400/0.60 5 0.056 0.1/0.1 Good

Vegetation not typically recommended for VFS


Alfalfa B/E 100/1.20* 14 0.037 Λ
20/0.05 Good

Sericea lespedeza C/E 60/1.55* 16 0.037* 5/0.05 Good

Common lespedeza D/E 30/2.19* 5 0.037 Λ


0.1/0.05 Good
Sudangrass D/E 10/3.80* NA 0.037 Λ
0.1/0.05 Good

"Best estimate by authors based on values from Table 4.3.


*Taken from Temple et al. (1987) except as noted. To convert densities for good stand to other stands, Temple et al.
recommend multiplying the given densities by j , f, 1, j , and f for poor, fair, good, very good, and excellent cov­
ers, repectively. The first number in the column is density in stems/ft . The second number is stem spacing in inches.
2

M a x i m u m heights recommended before mowing. Based on a value that would typically keep the vegetation erect.
Checks should be made with shear velocity as shown in the text. M o w i n g should be routine to promote a dense
growth, even if the maximum height given here is not reached. Heights of VFS after mowing would be 3 to 5 in.,
based on use of a standard tractor drawn field rotary mower. Adjustments should be made for other mowing imple­
ments.
^Adapted from Kouwen et al. (1981). Values are based on retardance class rather that calibration values given by
Kouwen et al. For retardance class D, an average of the calibration value was used, as the retardance value appeared
unrepresentative. For stands other than good stands, the retardance class changes for a given height; hence, MEI
changes. Judgement will need to be used to estimate this change.
'Values could vary widely, depending on mixture. If a given grass type predominates, values for that species should
be used.
/Adapted from Table V.l, Hayes et al. (1982).
*Not recommended for VFS. Studies by the authors have shown that spacings greater than 1 in. tend to cause scour
and are hence not recommended. These numbers are one-fifth those in Temple et al. (1987). For VFS, the spacing of
interest is that close to the surface. Values given by Temple et al. are for grass-lined channels where one is interested in
an effective spacing that relates the retardance of the vegetation to shear force; hence Temple et al. multiplied the actu­
al spacings for legumes by five to get an effective spacing.
A
Based on grain sorghum, Table V.l, Hayes et al. (1982).
Vegetative Filter Strips and Riparian Vegetation 365

From continuity, the flow per unit width, # , is given w Solution:


as 1. Parameters for calculation. From Table 9.10, assuming a
good stand of tall fescue
(9.98) MEI = 2 0 unmowed, 0.1 mowed
xn = 0 . 0 5 6
Equations (9.96)-(9.98) characterize the flow in V F S . Η = 1 5 in. or 1.25 ft unmowed
Values for the calibrated xn are given in Table 9.10 for Η = 4 in. or 0 . 3 3 3 ft mowed (footnote C, Table 9.10)
some of the various vegetal classes shown for grass S = 0 . 6 3 in. or 0 . 0 5 2 5 ft.
s

waterways in Table 4.3. T h e user is cautioned that the 2. Calculating velocity and depth of flow. Using Eqs.
values for xn are not standard values for Manning's η (9.96M9.98)
for overland flow, but are specially calibrated values,
based on an analogy to flow in a rectangular channel
with a width equal to the spacing of grass blades.
Standard Manning's η values for overland flow should
not be used in this equation. R.=
2d + 5
Flow velocity equations presented above are valid { S

only if the vegetation remains erect. Kouwen et al. 4w = V d.


m {

(1981) evaluated the erectness of vegetation by compar­


ing the shear velocity of the flow to a critical shear Solving the equations simultaneously,
velocity based on the stiffness of the grass. H e devel­
oped two relationships to predict the critical shear
velocity, or
1 5 / 0 0525 \ 2 / 3

K S - 0.091 + 20.76 ( M E I ) 2
(9.99) 0.074 = — d V ^ — — — j (0.08)" 2

and Collecting d terms into a term denoted as


{ f(d ),
{

0.106
[/ * = 0.754 ( M E I )
c 2 (9.100) d/ 5 3

f(d ) { =
(2d f + 0.0525) 273

where i/ * is the critical shear velocity ( f t / s e c ) of grass


c

that is elastic, t/ * is the critical shear velocity ( f t / s e c ) (0.074)(0.056)


= 0.069.
c

v2/3
of grass that is stiff, and M E I is the stiffness of the (1.5)(0.08) (0.0525) 1/2

grass ( N m ) as given in Table 9.10. Kouwen et al.


2

(1981) selected the lesser of the two values to c o m p a r e The solution is by trial and error, or
to the actual shear velocity. Actual shear velocity, of
Trial 1: d = 0.15 ft, f(d ) - 0.085, high
course, is given by
{ f

Trial 2: d f = 0.10 ft, f(d )


{ = 0.054, low
Trial 3: d f = 0.12 ft, f(d )
{ = 0.066, low
(9.101) Trial 4: d { = 0.125 ft, f(d )
{ = 0.069, OK.
Calculating velocity, spacing hydraulic radius from d ft and
where g is the gravitational constant with other terms subsequently checking the flow rate,
as described previously. If U* is greater than the lessor
of i/ * and t/ *, then the vegetation will not remain
c c
(0.125)(0.0525)
R* - T-TTTT - 0.022 ft
erect. (2)(0.125) + 0.0525
1.5
(0.022) (0.08) 2/3 1/2
» 0.594 ft/sec
0Ό56
Example Problem 9.14 Calculation of flow depth
and velocity in vegetative filter strips <? = (0.594)(0.125) = 0.074 cfs/ft
w OK.

Estimate the flow depth and velocity for a constructed 3. Estimating the Reynold's Number. Assuming that ν =
VFS with overland flow if the vegetation is perennial KY 31 1.0 Χ 1 0 " ft /sec,
5 2

fescue on a slope of 8% and the flow rate is 0.074 cfs/ft


width. Determine whether or not the grass will remain erect. VRm s (0.594)(0.022)
Estimate the Reynold's number. Λβ = =
—iF^— = 1306
·
366 9. Sediment Control Structures
4. Checking to see if vegetation will remain erect. Critical (2) Calculate t h e size distribution entering zone D(t)
shear velocities [Eqs. (9.99) and (9.100)] Unmowed - MEI = (3) Calculate the fraction t r a p p e d in zone D(t).
20
T h e steps are explained in the following sections.
US = 0.091 + 20.76 ( M E I ) 2
Subsequently, a summary is given in Appendix 9C of
t h e relationships utilized to calculate t h e length of the
= 0.091 + 20.76(20) = 8304 ft/sec 2

deposition wedge.
US = 0.754 ( M E I )
2
0 1 0 6
Definitions of deposition and transport can best be
given by moving from t h e downstream zones back to
= 0.754(20) 0106
= 1.04 ft/sec.
t h e u p s t r e a m edge. This is the approach taken in t h e
Use US = 1.04 ft/sec. following discussion. T h e p a r a m e t e r s discussed in
Mowed - MEI = 0.1 the following sections are shown on Fig. 9.41. A l t h o u g h
t h e length of zone D(t) is t r e a t e d as a constant in this
US = 0.091 + 20.76 ( M E I ) 2
discussion, it is d e n o t e d L(t) for t h e r e a d e r who might
want to use t h e relationships in Appendix 9C and
= 0.091 + 20.76(0.1) = 0.299 ft/sec 2

calculate the variation in Lit) with the formation of a


i/ * = 0.754 ( M E I )
c
0 1 0 6
deposition wedge.
= 0.754(0.1)° 1 0 6
= 0.591 ft/sec. In t h e following section, subscripts c and d refer to
inflow to zone C(t) and D(t). W h e n average p a r a m e ­
Use U* = 0.591 ft/sec. ters are n e e d e d on a segment, they are given the
Actual shear velocity [Eq. (9.101)] additional subscript a. T h e subscript ο refers to the exit
from zone D(t). For example, q so is sediment load
U* = (gd S ) f c
l/2
= [(32.2)(0.125)(0.08)] 1/2

leaving zone D(t), q is sediment load entering zone


sd

= 0.567 ft/sec. D(t), a n d q saa is average sediment load on zone D(t).


In the discussions that follow, relationships for trans­
Therefore, the grass will remain erect in either the mowed or
port and deposition are given for a single particle size,
unmowed condition as
although computations are usually m a d e for a range of
u* < us. particle sizes. A discussion is given, after presentation
of transport and deposition equations, of procedures
for partitioning sediment load among particle classes
a n d for estimating representative particle diameters.
Sediment Trapping in VFS T h e discussion given below starts first with deposition
in zone D(t) and then moves to zone C(t).
Constructed VFS
A s described earlier, constructed V F S are designed Deposition in Zone D(t) Z o n e D(t) represents that
for broad overland flow without significant channeliza­ area within the filter in which the layer of litter on the
tion. Mechanisms for trapping were illustrated earlier b e d has not b e e n totally filled with sediment; thus
in Fig. 9.41. T h e representation of the zones as func­ bedload transport would b e zero. Tollner et al. (1976)
tions of time (t) is used to illustrate t h e fact that t h e proposed that t h e trapping efficiency in this layer was a
length and location of the zones change with time. function of the n u m b e r of times a particle could settle
Based on studies by Tollner et al. (1976, 1978) and t o the b e d a n d of the flow Reynold's number. They
Hayes et al. (1984), equations for predicting sediment experimentally d e t e r m i n e d that
deposited and transported through each zone have
been developed. Actual application to a design situa­ T =
s " = exp(-1.05 X 1 0 - ^ 3 8 2
N -
f
0 9 1
),
tion would normally require the use of a computer
program. (9.102)
A n alternative approach to the use of t h e complete
set of equations is to assume that the length of the where T is the trapping efficiency in zone D(t), R is
s e

deposition wedge is negligible and calculate the trap­ the flow Reynold's n u m b e r , a n d N is t h e fall n u m b e r .
f

ping that occurs in the suspended load zone, given as Values for R w e r e defined using t h e spacing hydraulic
e

zone D(t) in Fig. 9.41. Using this approach, t h e follow­ radius in E q . (9.97), or
ing steps would be taken:
* e = Knda*sdaA> (9-103)
(1) Calculate q , the equilibrium sediment load enter­
sd

ing zone D(t) w h e r e ν is kinematic viscosity. Addition of t h e sub-


Vegetative Filter Strips and Riparian Vegetation 367
script d a t o V , R , a n d d in E q s . (9.96)-(9.98)
m s { w h e r e I is a dimensionless term related t o t h e average
indicates that these values a r e estimated in zone DU) infiltration rate. T o simplify further equations, / is
at t h e midpoint. Values for t h e fall n u m b e r N a r e {
taken as half t h e infiltration rate t o offset t h e fact that
given by (<7 d + # s o ) i divided by two, o r
s n o t
S

N =K L(0/K
f s m d A, a ) (9.104)
(9.110)
where V is t h e settling velocity of t h e sediment a n d
s 2q. wda
Lit) is t h e total length of zone DU) in Fig. 9.41.
Values for settling velocities a r e d e p e n d e n t o n particle where g a n d <?
w d wo a r e flow rates entering a n d exiting
size, as defined in C h a p t e r 7; thus t h e trapping effi­ zone D(t) a n d q wda is t h e average flow rate in zone
ciency changes with particle size. In t h e approximation DU) given by
being utilized here, LU) is a constant. If o n e wished t o
4wd +
calculate t h e advance distance of t h e deposition wedge, (9.111)
#wda ~~
XUX utilizing equations in Appendix 9C, LU) could b e
represented as a function of time by
T h u s , t h e sediment load exiting zone DU) for a given
L(t) = L T - * ( / ) , (9.105) particle size, w h e n considering infiltration, becomes

where L is t h e total initial length of the filter prior t o


T 4so = ( < 7 s d - 4 s d i ) ( l - 0 - 7 ; ) s

deposition a n d X(t) is t h e distance t h e deposition


wedge in zone Bit) h a s advanced downstream. It is or
assumed in E q . (9.105) that t h e length of zone C(t) is
4so = k d - (<7sd + 4 s o ) / ] [ 1 . 0 - T ]. (9.112)
small, as shown by Hayes et al. (1982, 1984). cs

In t h e development of E q . (9.102), it was assumed Solving for q , so

that all particles reaching t h e b e d a r e t r a p p e d in t h e


litter. While this is a good assumption initially, it g s d (1.0-/)(1.0-T c s )
becomes questionable with time a n d deposition; thus,
*° 1.0 + 7 ( 1 . 0 - Γ ) * ( }

Wilson et al. (1982) proposed a correction function t o


α

the trapping efficiency, or Converting t o a total fraction t r a p p e d , / , becomes d

C = 0.5exp[-3D ] + 0.5exp[l5(0.2D - D 2
)],
e p e p e p
r + 2/(l-r )
cs c s

(9.106) (9.114)
i+/(i-r„)
where D is t h e average d e p t h of sediment deposited
e p

in zone DU). T h e corrected trapping efficiency now T h u s , sediment discharge rate from zone DU) for a
becomes given filter slope, length, a n d media spacing can be
calculated, if the inflow rate of water and sediment, t h e
71. = C T . (9.107) infiltration r a t e , a n d t h e sediment particle size a r e
and t h e outflow sediment load (without considering known. T h e following calculations must b e m a d e
infiltration losses) becomes
(1) Knowing t h e inflow rate of water, q , t h e effective wd

4so = « s d ( l - 7 ; s ) . (9.108) length of zone DUX a n d t h e infiltration r a t e , t h e


outflow rate of water is given by
Sediment load in zone D(t) is further reduced by
infiltration. Hayes et al. ( 1 9 8 4 ) evaluated the impact of 4 w o = <?wd ~ < (0>
L

infiltration on sediment load by assuming that:


w h e r e i is t h e infiltration rate.
( 1 ) t h e difference in flow rate between t h e inlet of
(2) T h e average flow rate, q , o n DU) is d e t e r m i n e d
wda
zone DUX <7 , a n d outlet, # , is only a result of
wd w o
from E q . (9.111) a n d t h e infiltration p a r a m e t e r
infiltration, a n d
from E q . (9.110).
(2) t h e mass of sediment contained in a given infiltra­
(3) Calculate t h e average flow depth, d , a n d veloc­ f d a
tion volume is either transported into t h e soil m a ­
ity, K , from #
m d a a n d simultaneous solution of
w d a
trix by infiltration o r is t r a p p e d o n t h e surface.
Eqs. (9.96), (9.97), a n d (9.98).
Based on these assumptions, t h e sediment deposited (4) Calculate t h e fall n u m b e r , N , t h e Reynold's num­ f

as a result of infiltration, q , is sdl


ber, R , from Eqs. (9.103) a n d (9.104).
e

(5) Calculate t h e trapping efficiency for particles set­


tling t o t h e b e d by E q . (9.102).
368 9. Sediment Control Structures
(6) If a cumulative deposition of D (inches) has oc­ ter a r e discussed in a following section.
curred, calculate a corrected trapping efficiency by T o d e t e r m i n e the sediment discharge r a t e from zone
Eq. (9.106). C ( i ) , and h e n c e inflow into zone D(t), for a given filter
(7) Calculate the total sediment discharge corrected slope, length, media spacing, inflow rate of water, and
for infiltration by Eq. (9.113). representative sediment size, t h e following calculations
must be m a d e :
U s e of these procedures is illustrated in a subse­
quent example. (1) Calculate the flow d e p t h , velocity, and spacing hy­
draulic radius as in zone D(t), using the flow rate
Equilibrium Transport in Zone C(t). Z o n e C(t) is t h e q wd instead of # . A s s u m e n o infiltration in zones
w d a

zone in the filter where t h e r e is sufficient sediment A(t) t h r o u g h C(f); h e n c e


deposition to allow bedload transport, but not suffi­
cient deposition to alter the bedslope. Based on obser­
vations by Hayes et al. (1982), the length of zone C(t)
is small, and changes in flow rate and sediment load in (2) Calculate t h e sediment discharge rate, q , for the
sd

the zone are small. T h u s the equilibrium transport in representative particle size from Eq. (9.118).
zone C(t) is t h e transport into zone D(t). T h e sub­
P r o c e d u r e s for determining a representative particle
script d, therefore, is used to represent values in zone
diameter a r e discussed in a subsequent section.
C(t).
Tollner et al. (1982) developed a calibrated version
Particle Size Distributions, Effective Particle Diameters,
of t h e Einstein bedload function to predict sediment
and Sediment Load Partitioning. Sediment transport and
transport in this zone, or
deposition for areas u p s t r e a m of t h e V F S a n d in zones
ψ= 1.08ψ" 0 2 8
(9.115) A(t) through D(t) are illustrated in Fig. 9.41. Z o n e
D(t) is t h e only o n e discussed h e r e . Z o n e s A(t) and
where ψ is Einstein's shear intensity given by B(t) a r e discussed in A p p e n d i x 9C. Discussions to this
point have focussed on a single particle size. T h e use of
ψ = ( S G - 1) (9.116) a single particle size, however, to calculate deposition
5
c^sd could indicate complete trapping of a mixture, al­
though it is unlikely that fines would be t r a p p e d , par­
and φ is Einstein's transport rate function given by ticularly in t h e deposition wedge. T o apply these equa­
4sd
tions to t h e various zones, it is necessary to m a k e
(9.117) calculations for a variation of particle sizes and adjust
y /(SG
s] - l)gdl d ' t h e particle size distribution at various points in the
VFS.
where S is slope, S G is particle specific gravity, y is
c s
T h e following analysis is that p r e s e n t e d by Hayes
particle weight density (pounds per cubic foot in t h e et al. (1982, 1984). Based on their observations, it was
fps system), and d is the representative particle size
pd
a p p a r e n t that particles smaller t h a n 0.037 m m were not
for bedload transport. F o r computational purposes, t r a p p e d in t h e deposition wedge and that particles
Eq. (9.115) can b e rearranged to smaller than 0.004 m m were not t r a p p e d by settling in
3.57 zone D(t). T h u s t h e particle size distribution was di­
K(tfsA) vided into t h r e e classes:
4sd = (9.118)
(a) d > 0.037 m m
pi

where Κ is a constant given by (b) 0.004 m m < d < 0.037 m m


pi

(c) 0.004 m m > d . pi

Κ = (1.08) 3 5 7
y S 1 / 2
SG(SG _ i ) " 3 0 7
. (9.119a)
w
F o r computational purposes, t h e particles greater than
0.037 m m can b e divided into multiple size ranges.
If q is in p o u n d s p e r second p e r foot width, R
sd is in sd
Using trapping efficiencies a n d t h e size ranges d e ­
feet, and d^ is particle diameter in millimeters, then
scribed above, t h e size distribution was modified at
Κ = 6.462 X 1 0 S G ( S G - l ) "
7 3 0 7
. (9.119b) t h r e e points in t h e V F S as shown in Fig. 9.42. For
transport in t h e deposition wedge, zones A(t) and
Because q is a function of particle diameter, calcula­
sd
B(t), observations indicated that t h e particles being
tions must be m a d e based on an average particle transported as bedload were primarily those greater
diameter. Procedures for estimating an average diame­ t h a n 0.037 m m (Hayes et al., 1984); therefore, it was
Vegetative Filter Strips and Riparian Vegetation 369

Figure 9.42 Illustration of particle size distribution computations.

assumed that all particles smaller than 0.037 m m w e r e logarithmic interpolation as necessary. For bedload
transported to zone D(t) a n d that t h e representative transport, t h e m e a n d i a m e t e r of particles at the inlet is
diameter for bedload transport (also d^), was t h e not t h e d of t h e entire size distribution, but the m e a n
5Q

average diameter of all particles coarser than 0.037 d i a m e t e r of particles coarser t h a n 0.037 m m . This is
mm. Using these definitions, t h e total sediment load t h e size corresponding t o a fraction finer given by
incoming to t h e u p s t r e a m edge of the V F S is divided
into two groups, or
« , - f + / , i . (9-123)
ei-e5 + ei , +m
(9.120)

where q^ is the total load, q% is the coarse material where / r


u
is the fraction of coarse material. Given a , w

transport rate, and ql+ is the fine material (silt a n d t h e value for d is d e t e r m i n e d from t h e input size
m
pi

clay) transport rate. Further, distribution, as shown in distribution A A in Fig. 9.42.


T h e next task is to calculate t h e size distribution at
^ i + m
- ^i/r! (9.121) location BB in Fig. 9.42. Using t h e value for d^ = d , pi

t h e transport capacity for bedload can b e calculated for


and z o n e C(t) from Eq. (9.118). T h e fraction of sediment
t r a p p e d in t h e u p s t r e a m deposition wedge can b e cal­
^i = ^i(l.0-/ !), r (9.122) culated using

where f \ is the fraction of particles smaller t h a n 0.037


r

mm, the lower limit of coarse material. Values for f \ T f^SkZjk_ (9.124)
are determined from the input size distribution, using
370 9. Sediment Control Structures

Using this value for / , t h e fraction of sediment deliv­


ered t o zones C(t) a n d D(t) is
Example Problem 9.15 Trapping efficiency
of vegetative filter strips
D rd = l-ff^. (9.125)

Using D , t h e size distribution at the inlet, F F can b e A 50-ft-long vegetated filter strip is being built to control
rd i5
sediment from a disturbed area with an average sediment
converted t o that at BB, F F , by
discharge of 0.5 lb/ft/sec and a size distribution as given in
d

Fig. 9.43. The average inflow rate is 0.074 cfs/ft, the vegeta­
FF = FFj/ D (9.126)
tion to be used is KY 31 fescue, and the slope is 8%.
d rd

Estimate the trapping efficiency of the filter at the start of


with t h e constraint
the storm. Assume steady-state flow conditions, and a spe­
FF d < 1.0. cific gravity of sediment of 2.65, and a steady infiltration rate
of 0.30 in./hr.
T h e resulting size distribution is shown in Fig. 9.42 at Solution:
section BB. A s is obvious from t h e figure, f} a n d f r
li

1. Input parameters for filter (same as Example Problem


at B B a r e different from that at t h e inlet. 9.14). Assume a good stand, unmowed conditions are
In zone £>(/), sediment is trapped by settling a n d by
xn = 0.056; S = 0.63 in. or 0.0525 ft; Η = 15 in. or 1.25 ft.
s
infiltration as predicted by Eq. (9.114). W h e n consider­
ing total sediment t r a p p e d in zone D(t), it is necessary Unmowed height is
to divide t h e particles into three size ranges given Η = 4 in. or 0.333 ft.
above and calculate / , Eq. (9.114), for each size range.
2. Estimating flow depth and velocities in zone D(t). Calcu­
d

T h e following sizes a r e used for calculation


lations same as Example Problem 9.14,
(a) Coarse particles: d > 0.037 m m . Calculate rf
p pda R sd = 0.022 ft; V md = 0.594 ft/sec; d {d = 0.125 ft.
using size distribution at B B .
(b) M e d i u m (silt) particles: 0.004 m m < d < 0.037 3. Estimating average size distribution for transport capacity
p
in zones C(t) and D(t). From Fig. 9.43, the fraction of
mm, rf = 0.012 m m .
particles entering the VFS are / / = 0.36 and / = 0.64.
pda n

(c) Small (clay) particles: d < 0.004 m m , d = 0.002


r
p pda
From Eq. (9.123), the fraction corresponding to the average
mm. diameter of coarse particles ( > 0.037 mm) is
T h e total fraction trapped for each size is calculated
from Eq. (9.114) a n d designated as / , / , a n d f for d
f
d
m
d * v = / r i + - y = 0.36 + — = 0.68.
clay, silt, a n d coarse particles, respectively. Using these
values, t h e total trapping efficiency for t h e entire filter Thus from Fig. 9.43, the average diameter becomes
is then d pd = 0.17 mm.

/ , o = [ / + ( / c
) ( i - / ) ] [ ( ! - Λ ! ) ]
4. Estimating bedload sediment transport capacity in zone
d

C(t). From Eq. (9.118), using the symbol q* for the coarse d

+ /d (/ri-/r?)+/d /ri,
m f
(9-127) fraction of q , sd

where / ? is t h e fraction of inflow sediment smaller


^sd ,2.07
than 0.004 m m , a n d t h e total fraction of sediment "pd
reaching t h e outlet of the V F S is
Κ = 6.462 Χ 10 SG(SG - 1 ) ~ 7 3 0 7

£>ro=l-/,o- (9-128) = 6.462 Χ 10 (2.65)(2.65 - 1 ) ~


7 3 0 7
= 3.68 Χ 10 . 7

T h e particle size distribution at section C C in Fig. From item 2 above, R sd = 0.022; hence
9.42, F F , is given by
Q
(3.68 X 1 0 ) [ ( 0 . 0 2 2 ) ( 0 . 0 8 ) ]
7 357

FF G = FFj/D r o (9.129)
=
(OTTp
= 0.211 lb/sec · ft.
with t h e constraint
5. Estimating fraction of bedload trapped in upstream depo­
FF G < 1.0. sition wedge. Using Eq. (9.124)

T h e s e procedures a r e illustrated in Example Problem Qt\ ~ ltd


J c
9.15.
Vegetative Filter Strips and Riparian Vegetation 371

The incoming sediment load that is coarse material ( > 0.037 Using fl with FF in Fig. 9.43, the average diameter for
vd d

mm) is coarse particles in zone DUX d , is 0.1 mm. p d a

7. Estimating total sediment trapped. Ignoring the effect of


Qn - Q J n - (0-50 lb/sec · ft)(0.64) the deposition wedge on the effective length of zone D(t),
= 0.32 lb/sec · ft
and L(t) =L T = 50 ft.
0.32 - 0.211
/= — = 0.341. The average diameters for clay and medium size particles in
J
0.32 the model are 0.002 and 0.012 mm, respectively. The average
6. Estimating size distribution at BB (see Fig. 9.42). From diameter for coarse particles, d^, was determined under
Eq. (9.125), the fraction delivered to zone DU) is item 6 to be 0.1 mm. From Chapter 7, settling velocities are

D ^ - l - / · / » - 1 -(0.341)(0.64)
K = 0.021 ft/sec (Fig. 7.3)
= 0.782. StC

The size distribution at BB is given by V -2.81dl


ttm [Eq.(7.5)]
FF, FF
FF„- < 1.0. = (2.81)(0.012) = 4.05 X 10~ ft/sec
2 4

0.782
K -2.81dl [Eq.(7.5)]
This distribution is shown in Fig. 9.43. From this distribution,
tf

/ = 0.46 and f£ « 1 - 0.46 = 0.54. Using Eq. (9.123), the


r d

new value for a is denoted as a and is = 2.81(0.002) = 1.124 χ 1 0 " ft/sec.


2 5

y v d

0.54
a = 0.46 + — = 0.73. Trapping, settling, and infiltration must be estimated at the
vd
average flow condition for zone D(t); thus the infiltration
372 9. Sediment Control Structures
volume must be calculated. Knowing the infiltration rate i From Eq. (9.114), the total sediment trapped in zone DU)
including infiltration is
<7wo = 4wd ~ i L
U)
/0.30 \ / lhr \ r cs + 2/(i - τ) α

= 0.00035 cfs/ft. Assuming no prior deposition, 7 CS = T . From the above


S

calculations, / = 0.00473; hence


Hence

4 = 0.074 - 0.00035 = 0.07365 cfs/ft. T CS + (2)(0.00473)(1 - 7 ) CS


w o

1 + (0.00473)(1 - r ) cs

Thus infiltration has no significant impact on flow. For pur­


poses of calculating average depths of flow, it is therefore Values are tabulated below.
assumed that q = q = 0.074 cfs/ft and the average
wo wd

depths of flow and velocities are the same as those at the


Particle Calculated
inlet to zone DUX or
class Τ Parameter /d
V**-V* = 0-594 ft/sec * cs
Coarse 0.967 0.967
rffda = d = 0.125 ft u Medium 0.292 /d m 0.298
Λ,Η = 0.022 ft. Fine 0.000 0.009
//
Using Eq. (9.103) for Reynold's number and ρ = 1.0 X 10" 5

ft /sec,
2

The total trapping efficiency [Eq. (9.127)]


V R (0.594)(0.022)
R = —
e
mda sdi

— = 1306.5
/.o = [/ + / (i -/)](ι -Λ!) +/d (A! -fS) +/d7?-
d
c m
r

Separating the incoming sediment into coarse (sand), medium


Using Eq. (9.104), the fall number is (silt), and clay-sized particles is done from Fig. 9.43, or

V L(t) s K(50) Coarse fV = 0.64


Knda^fda (0.594)(0.125) Medium + fine f \ = 0.36 r

= 673 F . S Clay / ? = 0.18.


Calculating individual values The effects of the length of the deposition wedge (zones A(t)
and BU)) are being ignored in calculations of trapping in
Coarse Nf = (673)(0.021) = 14.13
zone DU) as affected by the effective length of the zone.
Medium (silt) N f
m
= (673)(4.05 Χ 1 0 " ) = 0.2726 4
However, bedload is actually being trapped there and with a
fraction trapped, / , of 0.341, as calculated under item 5.
Clay JV = (673)(1.124 Χ 1 0 " ) = 0.00756.
f
f
5
Hence, from Eq. (9.127) cited above, the total trapping in the
filter will be
From Eqs. (9.110) and (9.111), the infiltration parameter is
/ = [0.341 + 0.967(1 - 0.341)](1 - 0.36)
I = gwd ~ ^iso ^wd ~ <?wo = = 0-00035
t o

+ 0.298(0.36 - 0.18) + 0.009(0.18)


4wda
2
^wd iwo + 0 0 7 4

= 0.00473 /to = 0.681.

From Eq. (9.102), the fraction trapped by settling, T is Thus the filter traps 68.1% of the incoming sediment.
S

8. Calculating outflow sediment load. By definition of the


T = exp[-1.05 X l O " ^
s
3 8 2
^- 0 9 1
] total fraction trapped,

= exp[-1.05 X 10- (1306) 3 082


iV - f
091
] tfso^siO "/to)
= 0.5(1 - 0.681)
= exp[-0.377AT -°- ]. f
91

= 0.160 lb/sec · ft.


Calculating values for individual particle classes,
9. Particle size of outgoing sediment. From Eq. (9.125), the
Coarse r s
c
= 0.967 total delivery ratio of sediment at the filter exit is
Medium 7 s
m
= 0.291 ro =
D 1
" /to = 1
- 0.681 = 0.319.
Clay Γ/ = 0.0 From Eq. (9.126), the particle size distribution at the filter
Vegetative Filter Strips and Riparian Vegetation 373

exit is rectangular channel with a width equal to the media


spacing and flow in a V F S . Conversely, flow in litter
t e n d s to b e m o r e analogous to turbulent flow in a
p o r o u s media. Equations for such transport have not
This particle size distribution is shown in Fig. 9.43. b e e n developed.
10. Summary.
q = 0.5 lb/ft · sec
si Design of VFS
<? = 0.160 lb/ft · sec
so
In the design of V F S , o n e must select a vegetative
Total trapping efficiency equals 68.1%. type, a groundslope, and a length of t h e filter (distance
parallel to t h e flow path). T h e groundslope should be
as close to the natural slope as possible. Slopes less
t h a n 1% and greater t h a n 1 0 % should b e avoided, if
Natural VFS possible. In both cases, m a i n t e n a n c e becomes crucial.
Trapping in natural V F S differs from that in con­ V F S should be located on t h e contour to minimize
structed V F S due to differences in types of vegetation channelization. Vegetation should b e selected to be
and to the effects of channelized flow. Corrections dense, lawn-like grass in o r d e r to minimize channeliza­
n e e d e d for each effect are discussed below. tion. Turf specialists in an area should b e contacted for
information on desirable species.
Effects of Channelized Flow Channelization causes an T h e length of the filter is an important design pa­
increase in flow rate per unit width and a decrease in r a m e t e r . Typically, t h e length should b e such that
trapping. Such channelization is normally randomly excessive deposition will not occur during the design
distributed across a V F S , making it difficult to quantify. life. In selecting the filter length, the following steps
A first estimate of the effects of channelization on can b e used.
trapping could be m a d e by a correction to the inflow
(1) Select the design life a n d a maximum allowable
rate and sediment load, or
deposition. T e n years a n d 15 cm are r e c o m m e n d e d by
*wi = 4 w i / C f
C (9.130) Hayes and Dillaha (1991).
( 2 ) Estimate t h e long-term sediment yield entering
ii-ei/Co (9-131) the filter (see C h a p t e r 8) and a design single-storm
where C is a correction factor for channelized flow. A yield. Hayes a n d Dillaha (1991) r e c o m m e n d a 10-year
c f

possible correction factor would be the fraction of the storm.


VFS through which flow occurs. For example, if, as a ( 3 ) Calculate t h e required trapping efficiency to meet
result of channelization, a given V F S has flow through a water quality s t a n d a r d . If a w a t e r quality standard is
only 4 0 % of its area, then a first estimate for C might c f
not required, m a k e a first estimate of trapping effi­
well be 0 . 4 . After making the correction to flow rate ciency.
and sediment load, sediment trapping could b e evalu­ (4) Estimate t h e length necessary to prevent deposi­
ated by Eqs. ( 9 . 9 6 ) - ( 9 . 1 2 9 ) . tion above that allowable over the life of t h e system.
(5) U s e the filter length to calculate the trapping
Effects of Vegetation Type Impacts of vegetation type efficiency for the design storm, ignoring the deposition
on trapping are related to growth p a t t e r n s and type of wedge (zones A(t) a n d Bit) in Fig. 9.41).
litter. Grass-like vegetation that grows in clumps tends ( 6 ) R e p e a t (4) and (5) until t h e lengths match.
to be less effective as a sediment t r a p than uniformly (7) Estimate, if desired, the trapping efficiency for
growing grass. Tall grass-like vegetation tends to be t h e design storm, considering deposition in zones
less dense at the surface than low-growing-type vegeta­ A(t)-C(t) using relationships in A p p e n d i x 9C.
tion and hence has a reduced trapping efficiency. T h e s e
effects can be accounted for in the spacing p a r a m e t e r
5 S in Eqs. (9.96)-(9.129).
Example Problem 9.16 Design of a vegetative
For tree- and brush-like V F S , sediment trapping filter strip
normally occurs in the litter at the ground surface,
since the spacing of trees and brush are typically too Estimate the length of a VFS needed in Western Kentucky
large to have a significant impact on shallow overland to reduce the long term sediment yield to less than 1
flow. T h e flow equations described earlier were devel­ t o n / a c r e / y r if the actual sediment yield to the VFS is 5
oped on the basis of the analogy between flow in a tons/acre/yr with the size distribution given by FF in Fig.t
374 9. Sediment Control Structures

9.43. Flows and vegetation are the same as in Example For the flow rate of 0.074 cfs/ft, the depth of flow is
Problem 9.14 and 9.15. The average flow length from the determined the same as in Example Problem 9.14. The
disturbed area is 200 ft. results are d = 0.125 ft, R = 0.022 ft/sec, 5 = 0.0525 ft,
{ s S

Solution: xn = 0.056, V = 0.594 ft/sec, and R = 1306. Hence, since


m c

1. Selecting a vegetation type. The vegetation assumed in L equals L, the fall number becomes
T

Example Problem 9.15 was KY 31 fescue, which is acceptable


for growing conditions in Western Kentucky. = 13.5 KLn
2. Select a design life. Based on a desired frequency of Vd m { (0.594)(0.125)
repair, use a design life of 10 years.
3. First estimate the required filter length, based on depth of For L T = 8 ft,
deposition. As recommended by Hayes and Dillaha (1991),
Coarse N = 2.27 c

assume that the maximum allowable deposition is 15 cm and f

a bulk density of 1.5. Also for the first approximation, assume Medium N {
m
= 0.44
that the width of the filter (distance parallel to the slope)
equals the width of the disturbed area. To reduce the sedi­ Fine N * = 0.001.f

ment load from 5 tons/acre to 1 ton/acre, the trapping


Trapping efficiency for each particle size
efficiency must be
f > I = 0.8.
to
/ = T = exp[-1.05 X
d s lO- R° N - ]
3
e
S2
f
091

The total sediment to be trapped, using the given length of = exp[-1.05 X 10- (1306) 3 082
iV - f
091
]
200 ft for the disturbed area, is
= e x p [ - 0.377 Nf° 91
].
(0.8)(5 tons/acre · yr)(10 years)(l acre/43,560 f t ) 2

x(2000 lb/ton)(200 ft) Total trapping efficiency


= 367 lb/ft filter width.
Coarse / d
c
= exp[(-0.377)(2.27)" 091
] = 0.836
The units here are pounds per foot measured along the
contour. The depth of sediment to be trapped per unit area Medium / d
m
= 0.002
is the total mass of sediment per unit length of filter divided Fine /J = 0.0.
by the bulk density. The bulk density is

y - SG X 62.4 = (1.5)(62.4) = 93.6 l b / f t . 3 Using Eq. (9.127), assuming / = 0,


sb

Hence the depth of deposition is /to = /d (l " / , } ) + / d ( / r ! " /r?) +


C m
flfS
367 lb/ft = 0.836(0.64) + 0.002(0.36 - 0.18) + (0.0)(0.18)
Depth =
= 0.535.
367 lb/ft 3.92 To reduce yield to 1 ton/acre, the trapping efficiency must
ft.
~ L (ft)(93.6 l b / f t )
T
3 =
T7 be 0.80; thus the length must be increased.
5. Calculating the length required based on trapping effi­
To keep this depth below 15 cm or 0.49 ft, ciency. The length must be greater than 8 ft, which will make
3.92 / approach 1.0, as shown by the above computations. Based
d

on the settling velocity, / = 0, for any reasonable length. d


L
^ ( M 9 - 8
° f t
- Therefore using Eq. (9.127), the trapping efficiency becomes
4. Calculate the trapping efficiency. To be conservative, all
deposition in the deposition wedge (zones A(t) and B(t)) /,o = (/d )(i-/r!)+/d
c m
(/r!-/r?)
will be ignored). Thus the size fraction and diameters are
0.80 = (1.0)(0.64) + / ( 0 . 1 8 ) m

given as FFj in Fig. 9.43. Using Fig. 9.43, the values in the
d

following table were determined in Examples Problem 9.15. •m


0.80 - 0.64
/d - = 0.889.
0.18
Average diameter, Fall Velocity,
Keeping / > 0.889 will ensure a total trapping efficiency of
d
m

Particle Fraction,
(ft/sec)
0.80. From item 4 above,
class FF 4
(mm)

Coarse (sand) 0.64 0.170 0.021 / d


m
= e x p [ - 0 . 3 7 7 ( N ) " ° ' ] = 0.889 f
m 91

Medium (silt) 0.18 0.012 4.05 χ 10" 4

η -1/0.91
ln(0.889)
Fine (clay) 0.18 0.002 1.124 χ 10" = 3.595.
5

-0.377
Porous Structures: Check Dams, Filter Fences, and Straw Bales 375

Also, from item 4 above, sedimentation and hydraulics. T h e s e principles are dis­
N f
m
= 13.5 V L
sm r = (13.5)(4.05 X U T ) L
4
T
cussed in the next section, followed by a description of
3.595 the types of porous structures.
L t =
(13.5)4.05X10- = 6 5 7

Thus a total length of 657 ft of KY 31 fescue should reduce


Predicting the Trapping Efficiency of Porous
the sediment load to the desired level and remain effective
Structures
for the design life of 10 years. This is obviously an excessive A limited n u m b e r of prediction equations have been
length. An alternative would be to utilize some cultural developed for defining the sediment trapping efficien­
practices to reduce on-site erosion. cies of porous structures (Hirschi, 1981; W a r n e r and
Hirschi, 1983), as discussed in t h e section.
Porous structures trap sediment by three mecha­
nisms.
Impact of VFS on Water Quality
1. R e d u c e d total load transport capacity resulting in
D a t a sets on the impacts of V F S on water quality are
deposition (primarily bedload) behind the dam.
limited for anything other than sediment. T h e trapping
2. Mechanical filtration inside t h e porous media.
will be highly related to the fraction of sediment in t h e
3. Trapping of fines as a result of infiltration into the
clay size range since chemicals are attached to the
exchange phase of the clay. As shown by t h e previous deposited sediment.
examples, colloidal particles not flocculated or aggre­ T h e relative importance of the t h r e e mechanisms
gated are not likely to b e t r a p p e d by settling, but could varies between types of porous structures, as discussed
b e t r a p p e d by infiltration with high infiltration rates. below.
Thus a knowledge of the fraction of colloidal sediment Porous structures trap sediment only if the flow does
tied up in aggregates is important. This is discussed in not overtop t h e structure. O n c e overtopping occurs,
Chapter 7. the trapping efficiency effectively becomes zero.
A summary of selected data sets available o n t h e
Modeling Sediment Trapping by Settling
impacts of V F S on water quality is given in Table 9.9.
T h e quantities of dissolved solids trapped varies widely. Hirschi (1981) and W a r n e r and Hirschi (1983) devel­
T h e complete mechanisms involved are not well under­ o p e d equations defining sediment trapping as a result
stood or modeled at t h e present time. of reduced transport capacity behind porous structures.
T h e trapping equations are based on the following
assumptions (see Fig. 9.44)
POROUS STRUCTURES: CHECK DAMS, (1) A quiescent settling zone occurs upstream of the
FILTER FENCES, AND STRAW BALES sediment structure with a length calculated from a
single-step backwater curve.
Introduction (2) T h e settling d e p t h is t h e average of the depth of
Porous structures are finding wide use as sediment 1.1 times t h e normal flow d e p t h and the depth at
control techniques. Examples include rock fill check the d a m .
dams in a stream, rock over sand check dams, filter (3) T h e flow through velocity is either
fabric fence (filter fence), filter fabric over brush, (a) given as a function of porosity, or
gabions, and brush barriers. Many of these structures (b) defined by slurry flow rate.
are simply built with the materials on hand.
T h e following derivations a r e given for both assump­
Experimental data on the trapping efficiency on
tions (3a) a n d (3b).
porous structures are extremely limited. In general, it
is expected that the trapping efficiency will be low, Trapping Efficiency—Velocity Defined by Porosity In this
particularly for rock-filled check dams. R e e d (1978)
derivation, equations are developed only for triangular
measured a trapping efficiency of less than 5 % for a
channels with side slopes of 2 : 1 . T h e assumption is
check d a m associated with highway construction. D e ­
m a d e that the discharge is given by
pending on particle size, theoretical considerations in­
Q = VA = V A e, (9.132)
dicate that the trapping efficiency of filter fence could n n d d

be relatively high, as indicated in a subsequent exam­ w h e r e V is the normal velocity, A is the area corre­
n n

ple. sponding to normal d e p t h , K is the velocity at the


d

Evaluation of the sediment trapping by porous struc­ d a m , A is t h e cross-sectional area at the dam, and e
d

tures can be m a d e by utilizing basic principles of is t h e porosity of the d a m . Using Manning's equation
376 9. Sediment Control Structures

Q=£V A d d

Figure 9.44 Backwater profile behind porous dam.

to define normal flow d e p t h for a triangular channel, tively. If it is assumed that S is given by Manning's {

the normal depth becomes (assuming that the hy­ equation using the velocity at the dam, then
draulic radius is half the depth)

3/8
y + V /2g
a d
2
- y - q V /2g
q
2

nQ Δ* = , (9.139)
y„ = 1.024 (9.133) S -[nV /lA9R / ]
c d
2 3 2

ZS™
w h e r e R is the hydraulic radius at the dam. F o r Ζ
where Q is discharge in cfs, y is normal depth in feet,
d
n
greater than 2.0,
η is Manning's roughness, Ζ is the channel side slope,
and S is channel slope in feet per foot. If the assump­ * =y /2;d d (9.140)
tion is m a d e that the zone of quiescent flow starts at a
depth of y = L I y , then
q n
hence

y = 1.126 (nQ/ZS ' ) .


q
0
c
5 3/S
(9.134) = y /e
n + Q'e -
2
l . l y „ - 0.683 Q
X
S-(2.ngn e^/y^)Q'
2
' ( 9
' 1 4 1 )

Using Eq. (9.132) for continuity, and Α ά = ZyJ,


where
y = [A /Z}
d d
1/2
= [Q/eV Z] . d
1/2
(9.135)
Q = Q /2gy Z .
2 4
n
2
(9.142)
If further assumption is m a d e that
Since t h e constant used for Manning's equation in Eq.
(9.136) (9.141) is 1.49, all inputs must b e in feet p e r second
then and feet. T h e gravitational constant must be 32.2
ft/sec .2

(9.137) Hirschi (1981) proposed that the trapping efficiency


for a given particle size could b e given by
Next, the length of quiescent settling zone can b e
estimated from a single-step backwater curve, using (9.143)
Eq. (4.54), or
w h e r e V is the particle settling velocity, t
s is the D

Ax = (9.138) average flow through or detention time, and y is the a v g

average d e p t h in the quiescent zone. T h e value of y a v g

can b e given by
where E and E refer to total energy at the d a m and
d q

at the start of the quiescent zone and S and S are the y + y


d n
t c
y avg (9.144)
slope of the energy grade line and channel, respec­
Porous Structures: Check Dams, Filter Fences, and Straw Bales 377
or using Eq. (9.137),
Example Problem 9.17. Porosity-defined trapping
(9.145) efficiency of porous structure
2 l β
T h e flow-through time is determined by using t h e Calculate the trapping efficiency for a 0.10-mm particle
behind a 60% porosity check dam located in a triangular
average of V and V as an average velocity; hence
n d
channel with 3 : 1 side slopes, a bottom slope of 1%, and a
2 ax Manning's roughness of 0.015. The flow rate is 5 cfs. What
(9.146) would be the required dam height?
ν + ν
η Λ

Solution:
Since V = cV
d n9
1. Calculate the normal depth from equation 9.133.
2Zy AX 2
n

(9.147) 3/8 3/8


«2(1 + 0 v„ = 1.024
nQ
= 1.024
(0.015)(5)
= 0.609 ft.
zs«5
3(0.01) 05

T h e trapping efficiency for a given particle size is


determined by simultaneously solving Eqs. (9.133),
(9.141), (9.143), (9.145), and (9.147). T h e use of these 2. Calculate the depth of water at the dam from Eq. (9.137).
equations is illustrated in Example Problem 9.17. Ex­
ample predictions for a variety of conditions is also y n 0.609
= 1.015 ft.
given in Fig. 9.45. e 0.6

100 4.0

3.0

σ>
ω *a>
Χ x
Ο υ I 2.0
c Ε c Ε
σ σ
α [ο α
«4—
Ixl Ε UJ Ε
Ε Ε
C c
2
-J 1.0

0.0
5 10 15 20 0 5 10 15 20

Flow (CFS) Flow (CFS)


Figure 9.45 Example prediction of porous dam trapping efficiency for two channel slopes and a porosity of 60%. ( A )
S = 0.01. (B) S = 0.02. The maximum dam height corresponds to dam overtopping for a given flow rate.
c c
378 9. Sediment Control Structures

3. Calculate the length of the quiescent zone, Eq. (9.141). Diameters (mm) Class %
From Eq. (9.142), the parameter Q' is
0.004 20
0.01 30
(5f
= 0.314. 0.03 20
2gy*Z 2
~ (2)(32.2)(0.609) (3) 4 2

0.05 20
0.10 5
From Eq. (9.141), Δχ is
0.30 5

1 χ _ y /e
n + Q'6 -l.\y - 2
n 0.683Q'

S- (2.27^ £ · Λ · )β' 2 3 3 3 3 3 Solution:


c η

These conditions fit those in Fig. 9.45A; thus T values can s

_ 0 . 6 0 9 / 0 . 6 + ( 0 . 3 1 4 ) ( 0 . 6 ) - (1.1)(0.609) - (0.683)(0.314) 2
be read directly. The total trapping efficiency summed over
all particle sizes is then
0.01 - ( 2 . 2 7 ) ( 3 2 . 2 ) ( 0 . 0 1 5 ) ( 0 . 6 ) 2 3 3 3 3
(0.314)/(0.609) 1 3 3 3

Γ* - E(7*)(AFF,).
= 29.78 ft; use 30 ft.

Calculations are tabulated below.


4. Calculate V . n

1 4 9
/y \ n
2 / 3
1.49 / 0.609 \ 2 / 3

/2 Size
"•"—(τ) - -o^(—) s /! ( 0 0 1 )
(mm) τ
si
a

0.004 0.20 0.00 0.000


= 4.50 ft/sec. 0.01 0.30 0.01 0.003
0.03 0.20 0.03 0.006
5. Calculate t from Eq. (9.147). 0.05 0.20 0.014
D
0.07
0.10 0.05 0.26 0.013
2Zy 2
ΔΧ (2)(3)(0.609) (30) 2

= 8.34 sec. 0.30 0.05 1.00 0.050


eo + o (5)(1 + 0.6)
1 = 0.086

6. Calculate the setting velocity. From Fig. 7.3, V is 0.022 s


fl
From Fig. 9 . 4 5 A . Note: If the flow
ft/sec for a 0.10-mm particle. c o n d i t i o n s do not match t h o s e in F i g .
7. Calculate average depth from Eq. (9.145). 9.46A, T would need to be calculated using
s

Eq. (9.143).

yi n l\ 0.609 / 1 \
^ = y( 1 +
7 ) ~ ( 1 +
o^) = °- 812ft
- Thus
T r = 0.086 or 8.6%.
8. Calculate T from Eq. (9.143). s

It can be seen from this example that the trapping efficiency


VJn (0.022)(8.34) for porous dams is typically small. This is a result of the small
T, = = = 0.226 or 22.6%. detention time.
s
y _ 0.812

This is approximately the same value one obtains from Fig. A n important point should be m a d e about the value
9.45A.
of Manning's η for the above equations. T h e rough­
ness of t h e channel will be d e p e n d e n t on t h e equiva­
lent roughness of t h e deposited material. Relationships
Example Problem 9.18 Effects of size distribution given in C h a p t e r 4 and C h a p t e r 10 can be used to
on porous structure trapping efficiency predict η for a given particle size.

Using the same flow conditions and channel as Example Trapping Efficiency—Defined Flow Through Velocity
Problem 9.17 estimate the fraction of sediment trapped by T h e porosity definition of velocity is difficult to apply to
the porous dam if the sediment has the following size distri­ flow through filter fence-type material. A s an alterna­
bution. tive, the flow through velocity, known as slurry flow
Porous Structures: Check Dams, Fitter Fences, and Straw Bales 379

Table 9.11 Slurry Flow Rates Through T h e average velocity, of course, is


Porous Filter Material

Slurry flow rate


"avg 2
(9.150)
Material gpm/ft 2
ft/sec Reference

Straw bale 5.6 0.0125 a T h e travel time is


a
Burlap (10 oz.) 2.4 0.0053
0.000674 a
Synthetic 0.3
Ό = y—> (9.151)
fabric r
avg

V i r g i n i a Soil and Water C o m m i s s i o n


a

Conservation (1980). a n d t h e trapping efficiency is


M a r y l a n d Water Resources Administration
a

(1983).
τ = (9.152)

τ avg
rate, is defined for t h e material. Example slurry flow
where
rates are given in Table 9.11, as r e c o m m e n d e d by state
regulatory agencies. T h e value for synthetic fabric is
y d + y n
r e c o m m e n d e d by several states; however, studies by y avg (9.153)
Fisher and Jarrett (1984) and Waynt (1980) indicate
that the slurry flow rate of 0.5 g p m / f t is very low. A 2

further discussion is given in the following section on


filter fences. Rectangular Channels F o r a wide rectangular chan­
W a r n e r , as r e p o r t e d in W a r n e r and Hirschi (1983) nels of width b, w h e r e R = y„, and A = by„, t h e
and Barfield et al. (1981), derived equations defining equations are
sediment trapping behind a porous structure based on
a defined slurry flow rate, V . C h a n n e l shapes consid­ 3/5
sl nQ
ered are triangular and wide rectangular. Procedures yn = (9.154)
l.49Sy b 2

followed are similar to those based on porosity. In


addition to defined velocities, the equations are based
on normal velocities and d e p t h s from M a n n i n g ' s equa­ (9.155)
bV.,
tion and a single-step backwater profile.
Q/V b A - y„ - β 7 2 « 6 2Χ2
ζ

Triangular Channel F o r a triangular-shaped channel, (9.156)


using a defined slurry flow rate of K (feet p e r second), sl
S c

calculations of normal d e p t h a n d d e p t h at t h e d a m are


Q/by + v
m a d e with Manning's equation and continuity, given n A

(9.157)
earlier as Eq. (9.133), or 2
-.3/8 y + y
nQ n a
(9.158)
1.49 ZS / 1 2

AX 2AXby
Given that the area at the d a m is A = Zy\ for d
n

in = (9.159)
triangular-shaped channel and A = Q/V , t h e n d s{

l0.5

(9.148) τ = £ 1.0.
ZK. (9.160)

Using a single-step backwater profile defined by Eq. r avg


(9.139) and ignoring terms with powers of V greater sl
A relationship b e t w e e n slurry flow rates and porosity
than one, for p o r o u s rock fill d a m s is given in C h a p t e r 5 along
with example computations.
(Q/ZV ) ' -y -Q*/2gZ y<
0 5 2

P r o c e d u r e s for using Eqs. (9.154)-(9.160) are illus­


s[ n n

Ax = (9.149)
t r a t e d in Example Problem 9.19.
380 9. Sediment Control Structures

Compare those results to trapping by a filter fence with slurry


Example Problem 9.19 Trapping efficiency defined flow rates of 5 to 10 gpm/ft . 2

by slurry flow rates Solution: From Table 9.12, the slurry flow rate for straw
bales is K , = 5.6 g p m / f t (0.012 ft/sec). Slurry flow rates for
s
2

Estimate the trapping efficiency due to settling behind a the filter fence are given in the problem statement. To solve
straw bale filter assuming that a flow of 5 cfs is spread over for trapping efficiencies, one must make calculations for y n

a width of 200 and 75 ft, the slope is 1%, and the channel has from Eq. (9.154), y from (9.155), Δ * from (9.156), K
d avg

a roughness of 0.015. Assume a rectangular channel. The size from (9.157), y from (9.158), r from (9.159) and T from
avg D s

distribution is the same as that in Example Problem 9.18. (9.160). Values are tabulated below.

Material V xl0s l
2
ΔΧ Vavg )avg Ό
(ft/sec) (ft) (ft) (ft/sec) (ft) (sec)

Width = 200 ft
Straw bales 1.25 2.00 0.0275 196 0.461 1.01 425 421 V s

V = 5.6 gpm/ft
sl
2

Filter fabric
V = 0.3 gpm/ft
sl
2
0.0674 37.09 0.0275 3705 0.455 18.56° 8143 439V S

V = 10 gpm/ft
sl
2
2.23 1.12 0.0275 108 0.466 0.57 232 407V S

Width = 75 ft
Straw bales
V = 5.6 gpm/ft
sI
2
1.25 5.33 0.0497 526 0.677 2.69 777 289V S

Filter fabric
V = 0.3 gpm/ft
sl
2
0.0674 98.91 0.0497 9893 Ό.671 49.48* 14743 298V S

V = 10 gpm/ft
s l
2
2.23 2.99 0.0497 291 0.682 1.52 427 280V S

fl
Unrealistic situation.

Trapping efficiencies now must be calculated for each particle size and summed over all particles classes as shown below.

Trapping Efficiency for a Particular Size, T s i


a

Straw bale Filter fabric

Size V
v
si
b
vV
s\'· 5.6 gpm/ft 2
0.3 gpm/ft 2
10 gpm/ft 2

(mm) AFFj (ft/sec) b: 200 75 200 c


75 c
200 75

0.004 0.20 4.5 χ 10" 5


0.02 0.01 0.02 0.01 0.02 0.01
0.01 0.30 2.8 x 10" 4
0.12 0.08 0.12 0.08 0.11 0.08
0.03 0.20 0.00253 1.00 0.73 1.00 0.73 1.00 0.71
0.05 0.20 0.0062 1.00 1.00 1.00 1.00 1.00 1.00
0.10 0.05 0.025 1.00 1.00 1.00 1.00 1.00 1.00
0.30 0.05 0.13 1.00 1.00 1.00 1.00 1.00 1.00

r
ts = 2 : T
si A F F
i = 0.54 0.47 0.54 0.47 0.54 0.47

a
7 f r o m Eq. (9.160.).
s i

^Minimum of Eq. (7.5) or Fig. 7.3.


c
Since the values for y for this case are unrealistic, i.e. (overtopping occurs), these trapping fractions are presented
d

for illustrative purposes only.


Porous Structures: Check Dams, Filter Fences, and Straw Bales 381

Table 9.12 Results of Tests by Fisher and Jarrett (1984) on Selected Filter Fabric

Measured mechanical
Slurry flow rates (gpm/ft ) for flows with
2
filtration trapping
^ '
U 1 V
sediment consisting of efficiency
opening
size Clear Sand Coarse Silt-clay Sand Coarse Silt-clay
Fabric (Sieve No.) water silt silt

Cerex 34 NA 131 27 4.5 99 100 82 2


Cerex 68 NA 94 22 4.5 3 100 91 22
Supac 139 80/120 111 21 10.5 75 100 72 5
Supac 407 70/100 111 29 40.5 110 100 48 2
Typar 64 4/70 37 12 33.0 44 95 6 2
Mirafi 100 30 15 5 16.5 5 90 9 1

From Example Problem 9.19, o n e can conclude that tion. H e did not test t h e same materials as Fisher and
trapping is very much a function of flow r a t e p e r unit J a r r e t t (1984).
width and only weakly a function of slurry flow rate. In Procedures for estimating trapping by mechanical
the prediction equations, the decreased travel time action have not b e e n developed. In general, one has to
resulting from higher slurry flow rates are offset by use empirical d a t a for a given size fraction, such as
decreased d e p t h of flow, requiring less settling time to those in Table 9.12.
reach the bottom. Again, it should b e pointed out that
the trapping predicted here results from settling b e ­
hind the dam. Mechanical filtration is a separate issue,
discussed subsequently. Infiltration Impacts
Also, it should b e pointed out that the d e p t h of flow Sediment trapping by infiltration through the previ­
behind the filter fence is excessive for slurry flow rates ously deposited sediment also accounts for some of the
of 0.3 g p m / f t , the value typically r e c o m m e n d e d for
2
trapping. As infiltration occurs, suspended sediment is
filter fence fabric. A value of 5 to 10 g p m / f t gives a
2
transported to t h e soil matrix. This would be the pri­
more reasonable flow depth. mary mechanism for trapping clay size particles. Since
infiltration rates are typically much less than the slurry
flow rate, this mechanism occurs primarily after the
Mechanical Filtration
filter fabric is clogged and sediment laden flow drains
Studies by Fisher and J a r r e t t (1984) indicate that through previously deposited sediment.
mechanical filtration is very much d e p e n d e n t on t h e
type fabric, equivalent opening size ( E O S ) and the
sediment size distribution. As sediment flows through Types of Porous Structures
the filter cloth, holes are plugged by sediment in t h e Equations for trapping efficiency were given in the
size range of the E O S , if such sizes are present. A s a previous section. T h e following description gives some
consequence, the fraction of sediment t r a p p e d by m e ­ general details on design and installation.
chanical filtration and t h e slurry flow rates are highly
related to the size distribution of the sediment as well Rock Fill Dam and Gabion
as the filter material. Fisher and Jarrett (1984) indicate A commonly used porous structure is the rock fill
that slurry flow rates and fraction of sediment t r a p p e d check d a m , simply constructed by e n d dumping rock in
are related to flow orientation, thickness of the fabric, a channel as shown in Fig. 9.46B. In some cases, the
and accumulation of sediment on the upstream side of rock may cover a sand filter core. Sometimes these
the fabric. Results for the fabrics tested by Fisher a n d structures are used for grade stabilization as well as
Jarrett are shown in Table 9.12. sediment trapping. W h e n used for grade stabilization,
Studies by Waynt (1980) on 15 materials show slurry t h e heights of the dams and t h e distance between dams
flow rates much smaller and m o r e in the range of 0.3 is set by t h e slope, as shown in Fig. 9.47
g p m / f t . His results also show greater mechanical fil­
2
Sediment trapping with rock fill check dams occurs
tration. Waynt's results are based on low h e a d filtra­ as a result of all the mechanisms listed in the introduc-
382 9. Sediment Control Structures

Figure 9.47 Height and spacing relationships for check dams used
for grade stabilization and sediment trapping.

Filter Fence
A filter fence performs in much the same m a n n e r as
a porous rock fill check dam or a straw bale check dam.
It is normally used to control overland flow. A filter
Figure 9.46 Schematic of rock fill check dam and a gabion check fabric, typically a reinforced geotextile, is placed in the
dam. (A) Gabion and (B) end-dumped riprap check dam.
flow path, and trapping is provided by the t h r e e mech­
anisms described in the introduction to this section.
Design of the filter fence should require calculation
tion, but it appears that reduced total transport capac­ to ensure that the fence does not overtop and that the
ity behind the d a m is the primary trapping mechanism. required trapping efficiencies are met. Calculation pro­
Trapping within the dam itself could also be occurring, cedures should follow those in Example Problem 9.19
but to a lesser extent. Such trapping would be expected with modification to account for mechanical trapping,
with small rocks and small pores. T h e degree of trap­ if empirical d a t a are available for t h e particular fabric.
ping inside the pores has not been evaluated experi­ In the installation of the filter fences, particular care
mentally. Thus, rock fill check dams would not b e should b e taken to assure that the toe of the fabric is
expected to trap particles other than the larger bed- buried, as shown in Fig. 9.49. A s much as possible, the
load materials. fence should b e installed on the contour to prevent
T h e design of rock fill structures is based on the excessive concentration of flow at any o n e point. A
intended use. If designed as a grade control structure, filter fence installation is shown in Fig. 9.49.
the relationship between dam height and spacing of
check dams is based on slope as shown in Fig. 9.47. Brush Barrier
W h e n designed for trapping efficiency, calculation of Organic litter and spoil material from site clearing
trapping efficiency is m a d e as shown in Example Prob­ can be used effectively on a site by pushing or dumping
lems 9.17 to 9.19. a mixture of limbs, small vegetation, and root mat into
windrows along the toe of any slope where accelerated
Straw Bale Check D a m s erosion a n d runoff are expected. Anchoring a filter
Straw bales are frequently used to control overland fabric over the brush enhances the filtration capacity of
flow from disturbed areas as shown in Fig. 9.48. P r o p ­ t h e vegetation. M a i n t e n a n c e requirements are small.
erly installed, they can be effective in removing sedi­ W h e n covered with filter cloth, the upstream edge of
ment. Proper installation includes placing the bales in the filter cloth should be buried just upstream of the
a shallow trench and securing t h e m in place with barrier brush. A brush barrier installation is illustrated
stakes. In some cases, straw bales form a core of a rock in Fig. 9.50.
fill check dam. T h e rock fill is placed over the straw T h e trapping efficiencies of brush barriers without
bales, resulting in improved anchoring. filter fabric has not b e e n evaluated experimentally, nor
Design calculations are limited to making checks to have theoretical relationships b e e n developed to pre­
assure that the flow does not overtop the straw bales dict their effectiveness. Thus, design cannot b e based
and that the desired trapping efficiency can b e met. on predicted trapping efficiences. For brush with filter
Multiple-layer straw bale dams are not recommended. cloth cover, the effectiveness can be determined, as a
Flow and sediment trapping calculations should follow first approximation, by the effectiveness of the filter
those in Example Problem 9.19. material.
Sediment Traps 383

In general, the recommended height of the barrier long enough to allow sedimentation. Such traps are, in
should be at least 3 ft at construction, and the base fact, small sediment ponds, but typically have ill-
should be at least 5 ft. Some consolidation will occur defined outlet structures. Outlets can be open chan­
with time and vegetative decay, reducing t h e effective nels, riprap-lined weirs, riprap porous dams, or any
height. T h e effective life of the brush barrier is short, o t h e r of a large variety of controls. T h e outlet is
limited by the vegetative decay. typically selected to control the volume of water de­
tained and not the flow rate. Because of the lack of
attention to the outlet design, a maximum drainage
SEDIMENT TRAPS area of 5 acres is r e c o m m e n d e d (Clar et al., 1981).
T h e trapping efficiency of sediment traps can be
A sediment trap is a small temporary excavated defined by one of the several sediment pond models
basin that intercepts sediment-laden flow and detains it described earlier, if the stage-discharge curve is de-

Figure 9.48 Schematic of straw bale installation for sediment control.


384 9. Sediment Control Structures

Design of a t r a p must consider both volume and


surface area. Surface area is selected to give a desired
trapping efficiency. Calculations of trapping efficiency
are m a d e using Eq. (9.57). Volume is based on the
need to store trapped sediment and should theoreti­
cally equal some fraction of the sediment expected to
be t r a p p e d over the life of the trap. If the total volume
of t r a p p e d sediment cannot be stored, clean out will be
required.

INERTIAL SEPARATION: THE SWIRL


CONCENTRATOR
Figure 9.49 Schematic of a filter fence installation.
T h e swirl concentrator is a device designed to use
the centrifugal force of the flow to separate sediment
from water in storm runoff. T h e effluent sludge is
fined. As an alternative, H a n s e n (1973) utilized the transmitted to a small basin while the treated flow is
fully turbulent form of the overflow rate equation, discharged directly to a stream. A schematic is shown
given by Vetter (1940) and Eq. (9.57) or in Fig. 9.51. Flow enters through an inlet pipe at the
o u t e r edge of the concentrator and discharges through
an overflow section at the center of the device. Sludge
is discharged from the lower section. T h e device is
where V is the overflow rate given earlier by
c
portable and can be reused.
K-QM. Sullivan et al, (1976) p r e s e n t e d curves, as shown in
Fig. 9.52, for predicting the capture efficiency of a swirl
In the above equations, F is the fraction trapped, V is s concentrator for the size unit shown in Fig. 9.51.
the settling velocity, Q is the flow rate, and A is the W a r n e r and Dysart (1983) used these curves in con­
surface area. junction with the discharge from a 100-year storm in

Filter Fabric

Overland Flow

Toe of Filter Fabric Buried in a


Brush, Root Material, etc.
Trench and Backfilled
Figure 9.50 Schematic of a brush barrier installation.
Systems Approach to Sediment Control 385

Pipe ( . 6 1 m ) 100

Four Flow
Spoilers
Inlet
(.61x.61m)

2.5 3.0 3.5 4.0 5.0 6.0 7.0 8.0 9.0

Prototype Discharge ( f t /sec)

Figure 9.52 Swirl concentrator device particle size capture effi­


ciency for a 10% draw off (after Sullivan et al., 1976).

Β
mulch, than if the upslope areas are b a r e . This de­
crease is a result of the fact that the large particles will
be t r a p p e d by the mulch. T h e combination of the two
will, however, b e m o r e effective than either of the
Foul Outlet
(.10m)
practices used alone.
To Discharge
Tank T h e effectiveness of a system of controls is highly
Foul Outlet d e p e n d e n t u p o n site specific p a r a m e t e r s , including the
Discharge Adjusting specific combination of control techniques. A n evalua­
Device
tion of a proposed system, d u e to the nonlinearity of
Figure 9.51 Schematic of swirl concentrator (after Sullivan et al.,
the combined systems, will need to be m a d e with some
1976). ( A ) Plan view. (B) Elevation, section A - A'.
type of watershed model such as S E D I M O T II (Wilson
et al., 1982, 1986). Such a model must be capable of
central Appalachia and predicted a trapping efficiency evaluating systems of controls used in various combina­
of 3 3 % for a surface mine site. They proposed that the tions.
trapping in more common events would be much A n example analysis was m a d e by Barfield and Wells
higher. Caruccio and Buxton (1984) present results for (1981) using S E D I M O T II and reported in part by
a limited series of field tests in which they experienced W a r n e r et al. (1982b), which illustrated t h e effective­
problems generating vortex action in the separator. ness of a variation of erosion control techniques on the
Field studies are n e e d e d to settle the controversy. same location. T h e predicted p a r a m e t e r in the study
was the single storm (10-year, 24-hr) sediment yield at
varying stages in the mining operation. A n example of
SYSTEMS APPROACH TO SEDIMENT CONTROL the results is shown in Fig. 9.53. From inspection of the
results, it is obvious that the effects of a combination of
Sediment control by any o n e of the technologies controls is nonlinear and d e p e n d s on the location.
described in this chapter or previous chapters will b e Although total suspended solids (TSS) were not sum­
less effective than a combination of these systems to­ marized in this study, it would have been possible to do
gether along with on-site controls. Example combina­ so. TSS concentrations would respond in the same
tions may include the use of diversions, mulching, m a n n e r as single-storm sediment yields.
check dams, and grass filters. T h e effectiveness of a
combination of systems will not b e as great as a linear
combination of each of the systems. For example, a
Problems
check dam in a flow channel will have a lower trapping (9.1) Describe the criteria used to evaluate the ef­
efficiency if the upslope areas are covered with straw fectiveness of sediment control structures. Which crite-
386 9. Sediment Control Structures

YEAR FROM START OF MINING


Figure 9.53 Example analysis of the effects of a system of sediment controls on sediment yield from
an area surface mine in Northern Appalachia. Computations were made with the S E D I M O T II model
(after Warner et al., 1982b).

ria would you recommend as a design standard for storage volume is 1.0 acre feet and the p u m p rate is 3.0
sediment control structures for surface-mined lands, cfs. Dye was p u m p e d into the pond such that t h e
construction sites, and storm water detention struc­ inflow concentration was 0.334 ppm. If the outflow
tures for urbanized areas following construction? concentration varies as shown below, estimate the best-
(9.2) A pond is being used to control sediment from fit p a r a m e t e r s for the C S T R S , plug flow, and plug flow
p u m p e d discharge from a d e e p mine operation. T h e diffusion models.
Systems Approach to Sediment Control 387

Time Concentration Time Concentration Parameter Mean Coef. var.


(min) (ppm) (min) (ppm)
Average rainfall volume (in.) 1.4 1.2

0 0 320 Intensity (in./hr) 0.5 1.5


0.278
40 360 Duration (hr) 5.0 1.2
0.003 0.278
80 0.086 400 0.304 Interval between storms Γ Ι Α (hr) 80 5.0

120 0.143 440 0.318


160 0.179 480 0.326
T h e watershed area is 80 acre with a curve n u m b e r of
200 0.205 520 0.330
85 and a Rational Equation C factor of 0.6. T h e
240 0.217 560 0.332
particle size distribution for sediment in an average
280 0.247 600 0.334 storm is shown below. Calculate the total fraction
t r a p p e d for both stormflow and quiescent flow. Assume
(9.3) A rectangular reservoir has a steady inflow and that the reservoirs have good hydraulic performance.
outflow rate of 3 cfs and a surface area of 1.3 ac.
Calculate the overflow rate. Settling velocity,
(9.4) If the reservoir in Problem (9.3) has a size Particle class
distribution given by the inflow distribution in Fig. I % finer (ft/hr)
9.17, estimate the trapping efficiency. If the inflow
1 15 0.01
concentration of sediment is 25,000 m g / l i t e r , estimate
the effluent concentration. 2 25 0.25

(9.5) Estimate the outflow size distribution for Prob­ 3 55 1.49


lem (9.4). 4 90 7.32
(9.6) T h e mine operator for the pond in Problems 5 100 22.00
(9.3) to (9.5) finds that he cannot physically locate the
pond as described. Therefore, he proposes to construct
two ponds in series that have a combined volume equal (9.12) If o n e wanted to extend the sampling time
to that of the originally proposed pond, but with a and d e p t h in Table 9.5 to a particle diameter of 0.001
surface area of 0.65 acres for each pond. W h a t will be m m , what would be the sampling time if a depth of 5
the impact o n trapping efficiency? Estimate the total cm is used?
trapping efficiency of the two ponds and the effluent (9.13) A first-flush filtration system is proposed for a
size distribution for each pond. If the inflow concentra­ 30-acre watershed with established a p a r t m e n t s and
tion is 5000 m g / l i t e r , what will b e the outflow concen­ houses. Regulatory requirements specify that the first
tration for Problems (9.3) to (9.5) and for this problem? 0.4 in. of runoff must be diverted and filtered. Filters
(9.7) If the reservoir in Problems (9.3) to (9.5) has a must be at least 18 in. thick with a hydraulic conductiv­
depth of 2.0 ft and a width of 10 ft, estimate the effects ity of 3.5 f t / d a y . T h e sediment basin and filter must be
of turbulence on trapping efficiency. W h a t would be drawn down in 24 hr or less and water depths must not
the outflow concentration if the inflow concentration is exceed 4 ft. D e t e r m i n e the surface area and depth of
2000 m g / l i t e r ? the sedimentation basin and filter basin. Assume that
(9.8) Work Example Problem 9.7 for plug 12. 18 μ π ι and larger particles are to be trapped in the
(9.9) Work Example Problem 9.8 for plug 12. sedimentation basin. M a k e a sketch for the system.
(9.10) Work Example Problem 9.10 for plug 12. (9.14) A developer proposes that sediment basins
(9.11) A reservoir is being constructed with the fol­ are improperly constructed and that the deepest por­
lowing characteristics: tion should be at the inlet and the shallow portion near
the outlet. W h a t is your response to that proposal?
Parameter Value (9.15) A vegetated filter strip (VFS) is constructed
Average surface area (principal spillway
on a slope of 5 % a n d seeded to perennial rye grass. If
to emergency spillway) ft 2
15,000 the flow rate is 0.02 cfs/ft, estimate the flow depth and
d e t e r m i n e w h e t h e r or not the vegetation will remain
Storage volume ft 3
5 2,500
erect.
Average depth ft 3.5 (9.16) If the V F S in Problem (9.15) has a 100-ft flow
length and is subjected to a sediment load of 1.0
The watershed is located in an area that has the l b / f t · sec with a size distribution as given in Fig. 9.43,
following rainfall characteristics: estimate the trapping efficiency and the outflow con-
388 9. Sediment Control Structures
centration. Assume a steady infiltration rate of 0.20 Engineering, University of Kentucky, Lexington, KY.
in./hr. Barfield, B. J., and Hayes, J. C. (1988). Design of Grass Waterways
for channel stabilization and sediment filtration. In "Handbook
(9.17) A VFS is being used to control sediment near
of Engineering in Agriculture," Vol II, "Soil and Water Engi­
Atlanta, Georgia. If the sediment yield must b e less neering." C R C Press, Boca Raton, FL.
than 1.5 t o n s / a c r e · year and the actual sediment yield Bondurant, J. Α., Brockway, C. E., and Brown, M. J. (1975). Some
to the filter is 6 t o n s / a c r e · year, estimate the length of aspects of sedimentation pond design. In "Proceedings, 1975
filter required. Assume that the flow and vegetation is National Symposium on Urban Hydrology and Sediment Control,"
the same as those in Problems (9.15) and (9.16) and U K B U 109, pp. 1 1 6 - 1 2 2 . College of Engineering, University of
Kentucky, Lexington, KY.
that the average flow length from t h e disturbed area is
Brown, C. Β (1950). Sediment transport. In "Engineering Hy­
150 ft. draulics" (Hunter Rouse, ed.). Wiley, N e w York.
(9.18) Calculate the trapping efficiency for a 0.15-mm Brune, G. M. (1953). Trap efficiency of reservoirs. Trans. Am.
particle behind a check d a m in a triangular channel Geophy. Union 34 (3):407-418.
with 3 : 1 side slopes, slope 1%, η = 0.015, Q = 3 cfs, Camp, T. R. (1946). Sedimentation and the design of settling tanks.
Trans. Am. Soc. Civil Eng. 111:895-958.
and porosity 8 0 % .
Caruccio, F. T., and Buxton, H. (1984). A n evaluation of the swirl
(9.19) W h a t would b e the overall trapping efficiency concentrator: Its effectiveness in clarifying sediment laden
for the check d a m in Problem (9.18) if the sediment drainage. In "Proceedings, 1984 Symposium on Surface Mining,
has the following size distribution: Hydrology, Sedimentology and Reclamation," U K B U 136,
pp. 169-178. College of Engineering, University of Kentucky,
Diameter Percentage Diameter Percentage Lexington, KY.
(mm) in class (mm) in class Chen, C. (1975). Design of sediment retention basins. In "Proceed­
ings, National Symposium on Urban Hydrology and Sediment
0.006 15 0.05 15 Control," U K B U 109, pp. 2 8 5 - 2 9 8 . College of Engineering,
University of Kentucky, Lexington, KY.
0.01 25 0.10 10
Churchill, M. A. (1948). Discussion of "Analysis and use of reservoir
0.03 30 0.40 5 sedimentation data" by L. C. Gottschalk. In "Proceedings, Fed­
eral Interagency Sedimentation Conference, Washington, D C ,
pp. 139-140.
(9.20) If the dam in Problem (9.19) is a synthetic Clar, M. L., Das, P., Ferrandino, J. J., and Barfield, B. J. (1981).
Handbook of erosion and sediment control measures for coal
fabric with a slurry flow rate of 5 g p m / f t , what would
2

mines, Unpublished final report on Contract N o . J5104049 to


be the trapping efficiency? U . S. Office of Surface Mining. Hittman Associates, Columbia,
(9.21) W h a t are some possible reasons for differ­ MD.
ences in the slurry flow rates in Tables 9.11 and 9.12? Cooper, J. R., and Gilliam, J. W. (1987). Phosphorous redistribution
(9.22) Estimate the trapping efficiency, effluent con­ from cultivated fields to riparian areas. Soil Sci. Soc. Am. J. 51
(6):1600-1604.
centration, and effluent size distribution if a swirl con­
Cooper, J. R., Gilliam, J. W., Daniels, R. B., and Robarge, W. P.
centrator of the size shown in Fig. 9.51 is used to
(1987). Riparian areas as filters for agricultural sediment. Soil Sci.
control the flows in Problems (9.18) and (9.19). Assume Soc. Am. J. 51 (2):416-420.
that the inflow concentration is 10,000 m g / l i t e r and Cowardin, L. M., Carter, V., Golet, F. C , and LaRoe, Ε. T. (1979).
that the trapping efficiency is zero for particle sizes less Classification of wetlands and d e e p water habitats of the United
than 30 microns. States. U.S. Department of the Interior Publication F W S / O B S -
79/31.
Dillaha, Τ. Α., Sherrard, J. H., Lee, D., Shanholtz, V. O., Mostaghimi,
References S., and Magette, W. L. (1986). U s e of vegetative filter strips to
minimize sediment and phosphorous losses from feed lots: Phase
Austin, TX, City of (1988). "Environmental Criteria Manual." D e ­ I experimental plot study, Bulletin 151. Virginia Water Resources
partment of Public Works, City of Austin, TX. Research Institute, Virginia Tech University, Blacksburg, VA.
Barfield, B. J., Warner, R. C. and Haan, C. T. (1981). Hydrology and Dillaha, Τ. Α., Sherrard, J. H., Lee, D . , Mostaghami, S., and Shan­
Sedimentology of Disturbed Lands. Oklahoma Technical Press, holtz, V. O. (1988). Evaluation of vegetative filter strips as a best
Stillwater, OK. management practice for feed lots. / . Water Pollut. Control Fed.
Barfield, B. J., and Wells, C. G., (1981). Unpublished data associated 60 (7):1231-1238.
with studies for Hittman, Inc. (Available from senior author). Dillaha, Τ. Α., Reneau, R. B., Mostaghami, S., and Lee, D . (1989).
Barfield, B. J., Fogle, A . W., Carey, D . I., Inamdar, S. P., Blevins, Vegetative filter strips for agricultural nonpoint source pollution
R. L., Madison, C. E., and Evangelou, V. P. (1992). Water quality control. Trans. Am. Soc. Agric. Eng. 32 (2):513-519.
impacts of natural riparian grasses. Research Report N o . 184. Dobbins. W. E. (1944). Effect of turbulence on sedimentation. Trans.
Kentucky Water Resources Research Institute, University of Am. Soc. Civil Eng. 109:629-678.
Kentucky, Lexington. Driscoll, E. D., DiToro, D., Gaboury, D., and Shelley, P. (1986).
Barfield, B. J., and Albrecht, S. (1982). U s e of a vegetative filter zone Methodology for analysis of detention basins for control of urban
to control fine grained sediment from surface mine. In "Proceed­ runoff quality, Report N o . E P A 4 4 0 / 5 - 8 7 - 0 1 (NTIS N o . PB87-
ings, 1982 Symposium on Surface Mine Hydrology, Sedimentol­ 116562). U.S. Environmental Protection Agency, Washington,
ogy and Reclamation," U K Y B U 129, pp. 4 8 1 - 4 9 0 . College of DC.
Systems Approach to Sediment Control 389
Einstein, H. A. (1950). The bed-load function for sediment trans­ Hazen, A. (1904). On sedimentation. Trans. Am. Soc. Civil Eng.
portation in open channel flows, Technical Bulletin N o . 1026. 980:45-87.
U.S. Department of Agriculture, Washington, D C . Hill, R. D . (1976). Sedimentation p o n d s — A critical review. In "Pro­
Environmental Protection Agency (1976). Erosion and sediment con­ ceedings, 6th Symposium on Coal Mine Drainage Research,
trol—Surface mining in the Eastern U.S., Vol. I and II, E P A - Louisville, KY."
615/2-76-006. U.S. Environmental Protection Agency, Washing­ Hirschi, M. C. (1981). Efficiency of small sediment controls. Unpub­
ton, D C . lished Agricultural Engineering file report. University of Ken­
Evangelou, V. P., Rawlings, F., Crutchfield, J. D., and Shannon, tucky, Lexington, KY.
E. A. (1981). A simple chemo-mathematical model as a tool in Inamdar, S. P. (1993). Modeling sediment trapping in riparian vege­
managing surface mine sediment ponds. In "Proceedings, 1981 tative filter strips. Masters Thesis, Agricultural Engineering D e ­
Symposium on Surface Mining Hydrology, Sedimentology, and partment, Univ. of Kentucky, Lexington.
Reclamation," U K Y B U 126, pp. 4 9 - 5 8 . College of Engineering, Janiak, H. (1979). Purification of waters discharged from Polish
University of Kentucky, Lexington, KY. lignite mines, E P A - 6 0 0 / 7 - 7 9 - 0 9 9 . U.S. Environmental Protection
Evangelou, V. P., Barnhisel, R. I., Barfield, B. J., and Garyotis, C. L. Agency Industrial Research Laboratory, Cincinnati, O H .
(1987). Modeling release of chemical constituents in surface mine Kathuria, D . V., Nawrocki, Μ. Α., and Becker, B. C. (1976). Effec­
runoff. Trans. Am. Soc. Agric. Eng. 3 0 ( 0 : 8 2 - 8 9 . tiveness of surface mine sedimentation ponds, Report N o . EPA-
Fisher, L. S., and Jarrett, A. R. (1984). Sediment retention efficiency 600/2-76-117. U.S. Environmental Protection Agency, Cincin­
of synthetic filter fabric. Trans. Am. Soc. Agric. Eng. nati, O H .
27(2):429-436. Kouwen, N., Li, R. M., and Simons, D . B. (1981). Flow resistance in
Gearheart, R. J., Wilbur, S., Williams, J., Hull, D . , Finney, B., and vegetated waterways. Trans. Am. Soc. Agric. Eng. 24(3):684-690.
Sunburg, S. (1983). Final report City of Areata Marsh pilot Levenspiel, O. (1972). "Chemical Reaction Engineering." Wiley,
project effluent quality results—System design and management, N e w York.
Project report C-06-2270. City of Areata, Department of Public Levenspiel, O., and Smith, W. K. (1957). N o t e s on the diffusion-type
Works, Areata, CA. model for the longitudinal mixing of fluids in flow. Chem. Eng.
Gottschalk, L. C. (1965). Trap-efficiency of small floodwater retard­ Sci. 6:227-233.
ing structures. In "Proceedings, A S C E Water Resources Engi­ Li, R. M., Schall, J. D . , and Simons, D . B. (1980). Turbulence
neering Conference, Mobile, AL." prediction in open channel flow. Proc. Am. Soc. Civil Eng.
Griffin, M. L., Barfield, B. J., and Warner, R. C. (1985). Laboratory 106(HY4):575-587.
studies of dead storage in sediment ponds. Trans. Am. Soc. Magette, W. L., Brinsfield, R. B., Palmer, R. E., and Wood, J. D .
Agric. Eng. 28(3):799-804. (1989). Nutrient and sediment removal by vegetated filter strips.
Grizzard, T. L., Randall, C. W., Weand, B. L., and Ellis, K. L. Trans. Am. Soc. Agric. Eng. 32(2):663-667.
(1986). Effectiveness of extended detention ponds. Urban Runoff Maryland Water Resources Administration (1983). Maryland stan­
Quality, American Society of Civil Engineers. dards and specifications for soil erosion and sediment control.
Hammer, D . Α., ed. (1989). "Constructed Wetlands for Wastewater Maryland Water Resources Administration, Annapolis, M D .
Treatment: Municipal, Industrial and Agricultural." Lewis, McBurnie. J. C , Barfield, B. J., Clar, M. L., and Shaver, E. (1990).
Chelsea, MI. Maryland sediment detention pond design criteria and perfor­
Hammer, D . Α., and Bastian, R. K. (1989). Wetlands ecosystems: mance. Appl. Eng. Agric. 6(2):167-173.
Natural water purifiers? In "Constructed Wetlands for Wastewa­ N o e , S., and Barfield, B. (1990). Unpublished results of dye tracer
ter Treatment: Municipal, Industrial and Agricultural" ( D . A. studies in a model sediment pond. Department of Agricultural
Hammer, ed.) pp. 5 - 2 0 . Lewis, Chelsea, MI. Engineering, University of Kentucky, Lexington, KY.
Hansen, E. A . (1973). In stream sedimentation basins-a possible tool Occoquan Watershed Monitoring Laboratory (1986). Final Contract
for trout habitat management. U. S. Department of Agriculture, Report: Washington Area N U R P Project, Prepared for the
Forest Service, North Central Forest Experimental Station, East Metropolitan Washington Council of Governments.
Lansing, MI. Parsons, J. E., Daniels, R. B., Gilliam, J. W., and Dillaha, T. A.
Hayes, J. C , and Dillaha, T. (1991). Procedure for the design of (1991). The effect of vegetation filter strips on sediment and
vegetative filter strips, Report prepared for the U S D A - S o i l Con­ nutrient removal from agricultural runoff. In "Proceedings, Envi­
servation Service, Washington, D C . [Available from J. C. Hayes, ronmentally Sound Agriculture Conference, April, Orlando, FL."
Department of Agricultural and Biological Engineering, Clemson Randall, C. W. (1982). Stormwater detention ponds for water quality
University, Clemson, SC] control, stormwater detention facilities—Planning design opera­
Hayes, J. C , and Harriston, J. (1983). Modeling the long term tion and maintenance. In "Proceedings, Engineering Foundation
effectiveness of vegetative filters as onsite sediment controls, Conference, A S C E , 1982."
Paper 83-2081. American Society of Agricultural Engineers, St. R e e d , L. A . (1978). Effectiveness of sediment control structures used
Joseph, MI. in highway construction in Central Pennsylvania, Water Supply
Hayes, J. C , Barfield, B. J., and Barnhisel, R. I. (1978). Evaluation of Paper 2054. U.S. Geological Survey.
grass characteristics related to sediment filtration, Paper 78-2513. Seddon (1889). Cleaning water by settlement (original not seen, cited
American Society of Agricultural Engineers, St. Joseph, MI. Hazen, Α., 1904). J. Am. Assoc. Eng. Soc.
Hayes, J. C , Barfield, B. J., and Barnhisel, R. I. (1982). The use of Small, Μ. M. (1976). Data report-marsh/pond systems, U S E R D A
grass filters for sediment control in strip mine drainage. III. Report B N L 50600.
Empirical verification of procedures using real vegetation. Report Soil Conservation Service (1975, 1986). Urban hydrology for small
No. I M M R 8 2 / 0 7 0 . Institute for Mining and Minerals Research, watersheds, Technical release N o . 55. Soil Conservation Service,
University of Kentucky, Lexington, KY. U.S. Department of Agriculture, Washington, D C .
Hayes, J. C , Barfield, B. J., and Barnhisel, R. I. (1984). Performance Soil Conservation Service (1983). Computer program for project
of grass filters under laboratory and field conditions. Trans. Am. formulation hydrology, Technical Release N o . 20. Soil Conserva­
Soc. Agric. Eng. 27(5):1321-1331. tion Service, U.S. Department of Agriculture, Washington, DC.
390 9. Sediment Control Structures
Stahre, P., and Urbonos, B. (1990). "Stormwater Detention for Warner, R. C , and Dysart, B. C. (1983). Potential use of the swirl
Drainage, Water Quality and CSO Management." Prentice-Hall, concentrator for sediment control on surface mined lands. In
Englewood Cliffs, NJ. "Proceedings, 1983 Symposium on Surface Mining Hydrology,
Sullivan, R. H., Cohn, Μ. M., Ure, J. E., Parkinson, F. E., and Sedimentology and Reclamation." University of Kentucky, Col­
Zielinski, P. E. (1976). The swirl concentrator for erosion runoff lege of Engineering, Lexington, KY.
treatment, E P A 6 0 0 / 2 - 7 6 - 2 7 1 . U.S. Environmental Protection Warner, R. C , and Hirschi, M. C. (1983). Modeling check dam trap
Agency, Washington, D C . efficiency, Paper N o . 83-2082. American Society of Agricultural
Tapp, J. S., and Barfield, B. J. (1986). Modeling the flocculation Engineers, St. Joseph, MI.
process in sediment ponds. Trans. Am. Soc. Agric. Eng. Warner, R. C , and Schwab, P. J. (1992). " S E D C A D + Version 3
29(3):741-747. training Manual." Civil Software Design, A m e s , IO.
Tapp, J. S., Barfield, B. J., and Griffin, M. L. (1981). Predicting Warner, R. C , Wilson, Β. N., Barfield, B. J., Logsdon, D . S., and
suspended solids removal in pilot size sediment ponds using Nebgen, P. J. (1982a). A hydrology and sedimentology watershed
chemical flocculation, Institute for Mining and Minerals Re­ model. II. Users' manual. Special Publication. Department of
search Report IMMR 8 1 / 0 6 3 . University of Kentucky, Lexing­ Agricultural Engineering, University of Kentucky, Lexington, KY.
ton, KY.
Warner, R. C , Wilson, Β. N., Barfield, B. J., and Wells, L. G.
Temple, D . M., Robinson, Κ. M., Ahring, R. M., and Davis, A. G.
(1982b). Evaluation of alternative on site sediment controls using
(1987). Stability design of grass-lined open channels, Agriculture
the S E D I M O T II Model. In "Proceedings, Modeling and Simu­
Handbook Number 667. U.S. Department of Agriculture, Agri­
lation Conference, University of Pittsburg, 1982."
cultural Research Service.
Waynt, D . C. (1980). Evaluation of filter fabric for use as silt fences.
Tollner, E. W., Barfield, B. J., Haan, C. T., and Kao, Τ. Y. (1976).
V A Highway and Transportation Research Council, Richmond,
Suspended sediment filtration capacity of simulated vegetation.
VA.
Trans. Am. Soc. Agric. Eng. 19(4):678-682.
Whipple, W., Jr., and Hunter, J. V. (1981). Settleability of urban
Tollner, E. W., Barfield, B. J., Vachirakornwatana, C , and Haan,
runoff pollution. / . Water Pollut. Control Fed. 53(12): 1726-1732.
C. T. (1977). Sediment deposition patterns in simulated grass
Wilson, Β. N., and Barfield, B. J. (1984). A sediment detention pond
filters. Trans. Am. Soc. Agric. Eng. 20(5):940-944.
model using CSTRS mixing theory. Trans. Am. Soc. Agric. Eng.
Tollner, E. W., Hayes, J. C , and Barfield, B. J. (1978). The use of
27(5): 1339-1344.
grass filters for sediment control in strip mine drainage. I. Theo­
retical studies on artificial media, Report No. IMMR 3 5 - R R R 2 - Wilson, Β. N., and Barfield, B. J. (1985). Modeling sediment deten­
78. Institute for Mining and Minerals Research, University of tion ponds using reactor theory and advection-diffusion concepts.
Kentucky, Lexington, KY. Water Resources Res. 21(4):523-532.
Tollner, E. W., Barfield, B. J., and Hayes, J. C. (1982). Sedimentol­ Wilson, Β. N., and Barfield, B. J. (1986a). A detachment model
ogy of erect vegetal filters. Proc. Am. Soc. Civil Eng. for non-cohesive sediment. Trans. Am. Soc. Agric. Eng.
108(HY12):1518-1531. 29(2):445-449.
Vetter, C. P. (1940). Technical aspects of the silt problem on the Wilson, Β. N., and Barfield, B. J. (1986b). Predicted and observed
Colorado River. Civil Eng. 10:698-701. turbulence in detention ponds. Trans. Am. Soc. Agric. Eng.
Virginia Soil and Water Conservation Commission (1980). "Virginia 29(5):1300-1306.
Erosion and Sediment Control Handbook." Richmond, V A . Wilson, Β. N., Barfield, B. J., and Moore, I. D . (1982). A simulation
Ward, A. D., Haan, C. T., and Barfield, B. J. (1977). Simulation of model of the hydrology and sedimentology of surface mined
the sedimentology of sediment detention basins, Research report lands. I. Modeling techniques, Special publication. University of
N o . 103. Water Resources Research Institute, University of Ken­ Kentucky Agricultural Engineering Department.
tucky, Lexington, KY. Wilson, Β. N., Barfield, B. J., and Warner, R. C. (1986). Simple
Ward, A. D., Haan, C. T., and Tapp, J. S. (1979). "The D E P O S I T S modles to evaluate nonpoint source and controls. In "Agricult­
Sedimentation Pond Design Manual," OISTL. Institute for Min­ ural Nonpoint Source Pollution: Model Selection and Applica­
ing and Minerals Research, University of Kentucky, Lexington, tion." (Giorgini and Singales, eds.), pp. 2 3 1 - 2 6 3 . Elsevier, New
KY. York.
Fluvial Geomorphology
Fluvial Channel Analysis and Design

INTRODUCTION Time Frames of Fluvial Systems

The Fluvial System A n important distinction needs to be m a d e about


steady-state concepts as applied to fluvial systems. Flu­
Definition of a Fluvial System vial channels t e n d to b e dynamic, as a result not only of
A fluvial system can be divided, for convenience changing flows but also of continued erosion and depo­
sake, into zones of production, transfer, and deposition sition, which occur with steady flows. Fluvial channels
of sediment, as shown in Fig. 10.1. Given that these are thus always in a state of change, requiring that the
three processes occur in all t h r e e zones, the division is concept of a stable channel be applied to conditions
rather artificial. However, it is possible to speak of averaged over a period of time and not to a particular
individual zones where o n e of the three processes instant of time. If a channel system, over a period of
predominate. years, has developed hydraulic properties such that the
T h e processes occurring in zone 1, the zone of sediment load entering the system is transmitted
production, are of interest to hydrologists and soil through the system, it is said to be graded or in regime
erosion scientists. In this zone, processes are very dy­ (Mackin, 1948). Such a condition is a delicate balance
namic, responding to short-term climatic conditions, a n d would be modified by changes in any of the up­
and channels tend to b e unstable with rapid changes. stream zones in Fig. 10.1. For example, a channel-
Hydrologic and sedimentation processes occurring in straightening operation would increase the channel
zone 1 have b e e n discussed in detail in C h a p t e r s 2, 7, slope, resulting in a short-term increase in sediment
and 8. load, u p s t r e a m erosion, a wider channel, and possible
T h e processes occurring in zones 2 and 3 are of changes in the channel classification. A further discus­
interest in this chapter. Processes in these zones tend sion of the impacts of such changes on channel geome­
to be the most stable with the best defined configura­ try and properties is given in a subsequent section.
tions. Channel systems in these two zones remain dy­ T h e choice of a time period for analysis determines
namic and occasionally experience abrupt changes; w h e t h e r a variable is i n d e p e n d e n t or d e p e n d e n t . For
however, the specific processes tend to be amenable to example, over a short time period, the rate of sediment
prediction. Some of the deposition processes have been transports t e n d s to be a function of channel flow rate
discussed in C h a p t e r 9. and the channel hydraulic properties of slope, width,

391
392 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

and bed material composition, as discussed in C h a p t e r s in n a t u r e . This led S c h u m m (1977) to p r o p o s e t h e


4 and 7. Over long periods of time, however, these concept of thresholds for change, indicating t h a t
channel hydraulic properties are d e t e r m i n e d by flow changes in flow do not immediately lead to a change in
rate and incoming sediment load to zone 2. In fact, channel properties until some threshold level is
these channel hydraulic properties must ultimately as­ reached. A t that point, a major change in channel
sume values that allow the long-term sediment inflow properties is observed.
to equal t h e outflow. Therefore, the channel hydraulic
properties b e c o m e d e p e n d e n t variables and sediment Rigid versus Fluvial Channels
load becomes the i n d e p e n d e n t variable. In an analysis A detailed discussion is given in C h a p t e r 4 of flow in
of the impact of either m a n m a d e or natural p h e n o m ­ rigid channels, w h e r e t h e channel banks a n d bed d o
ena on channel geomorphology, it is important to con­ not change their configuration with flow. With the
sider the time frame of interest. T h e absolute duration exception of vegetative channels, t h e roughness was
of the time is not important. W h a t is important is the also assumed to remain constant. This is in contrast to
concept that a variable can be i n d e p e n d e n t or d e p e n ­ channels in alluvial material w h e r e channel width,
dent, d e p e n d i n g on the time frame. A n illustration of d e p t h , slope, curvature, and roughness change with
the changes of p a r a m e t e r s from i n d e p e n d e n t to d e p e n ­ flow r a t e and sediment load. Adjustment in these chan­
dent with time is given in Table 10.1. nel properties may b e slow or may occur in an episodic
T h e rate of change of channel hydraulic properties is n a t u r e , as suggested by t h e threshold concept. F o r
also variable for fluvial systems, tending to be episodic p u r p o s e s of this chapter, t h e following definitions are

PRODUCTION
ZONE 1

Watershed Sediment Deposition


Source Area Area
Figure 10.1 A n idealized fluvial system (after Schumm, 1977).

Table 10.1 Independent and Dependent Variables for Different Time


Spans (Adapted from Schumm, 1971a,b)

Status of Variable 0

Steady Graded Geologic


(short (long (very long
Variable term) term) term)

Geology I I

Valley morphology (slope, width, and depth) I I D

Water discharge I I X

Sediment inflow rate I I X

Channel slope, width, and depth ι D X

Bed sediment size ι D X

Channel sinuosity I D X

Channel roughness D D X

Observed sediment load and flow rate D X X

a
I; independent variable; D ; dependent variable; X; indeterminate.
Channel Classification 393

used: ing. M e a n d e r s tend to occur across the entire valley.


M a t u r e stream channels, in their natural state, tend to
Rigid channel: A channel whose slope, width, depth,
b e graded. Old streams occur in flatter flood plains and
roughness and curvature are assumed to remain con­
have flatter gradients. Stream m e a n d e r s do not typi­
stant.
cally occupy the entire valley. Natural levees occur
Fluvial channel: A channel whose slope, width, depth,
n e a r the stream with swamp-like areas adjacent to the
roughness, and curvature are functions of flow rate
levees. Tributaries flow for long distances parallel to
and sediment load.
the main stream channel before entering the channel
In the discussion in this chapter, a special section is at a natural break in the levee. T h e channels of most
devoted to those alluvial channels whose bed is com­ interest in small catchments fall in the youthful to
posed of gravel and large roughness elements. Typi­ m a t u r e classification.
cally, the size of the roughness elements in these Schumm (1971a, b) classified channels by discharge
channels approaches that of the flow depth, making the and sediment load into stable, depositing, and eroding
flow relationships more complicated. In addition, the channels. T h e form of the sediment load (bedload
bed material is either fixed or only mobile in larger versus suspended load), along with flow rate, deter­
runoff events, as contrasted to the m o r e common mines how the stream will shape its channel.
moveable beds where bed material is in transport in
most runoff events.
Channel response in a fluvial channel is a complex
interaction of many factors, not always completely p r e ­
Planform
dictable, but amenable to some analytical procedures. Based on a plan view of a river network, rivers can
T h e major physical factors involved are water dis­ b e classed as braided, straight, or meandering, or some
charge, sediment load, channel slope and shape, soil combination of the three, as shown in Fig. 10.2
and geologic characteristics of channel banks and beds, (Leopold and Wolman, 1957). If the channel walls and
vegetative effects, and h u m a n activities. b e d are alluvial, straight channels are typically artificial
situations or, at most, unstable and short lived. Even
u n d e r laboratory conditions, artificially formed straight
Overview of the Chapter channels b e c o m e naturally sinuous with time. A n ex­
T h e purpose of the discussion in this chapter is to ception results from anthropological activity where me­
develop a rational system of channel classification, to andering and resulting sinuosity can b e prevented to
discuss information on factors controlling channel some extent by channel bank controls.
properties, and to discuss changes that occur in the Braided channels are typically wide with ill-defined
transfer and deposition zone as a result of natural or banks and multiple islands. Inherently unstable, these
human actions. Finally, a discussion of both simple and channels experience rapid changes in shape, resulting
complex analytical models for predicting such changes in a braided pattern, thus allowing the channel to
is given. Each of these topics could fill an entire text; establish a sort of equilibrium between flow rates and
therefore, the coverage in this chapter is limited to an sediment load. D u e to the inherent instability, it is
overview. difficult to predict the precise n a t u r e of changes in
channel morphology resulting from changes in inputs.
L a n e (1957) p r o p o s e d that the mechanisms causing
braiding are: (1) steep slopes that result in wide shal­
CHANNEL CLASSIFICATION
low channels w h e r e multiple islands can form and (2)
excessive sediment loads resulting in deposition is­
Geomorphic
lands. Simons, Li and Associates (1982) propose that
Davis (1899) proposed that streams could be classi­ easily e r o d e d banks can also b e a cause of braided
fied based on stream age, using three age categories: channels. In general, braided streams have steep slopes
youth, maturity, and old age. This system of classifica­ and excessive sediment loads composed primarily of
tion would allow a stream to go from youth in the bedload materials.
headwaters to old age in the delta. M e a n d e r i n g channels have an S-shaped appearance
A youthful stream is o n e in the initial state of resulting from multiple m e a n d e r s . Within the main
development. Such streams are typically V shaped and channel itself is a narrow d e e p section known as the
very irregular, such as headwater streams in mountain thalweg with a m o r e sinuous shape than the main
tributaries. M a t u r e streams occur in b r o a d e r valleys channel. T h e deepest part of the channel exists in the
where channel slopes are flatter and downward cutting pool at the apex of the m e a n d e r . T h e shallow crossing
has essentially ceased, being replaced by bank widen­ can occur in the straight section between meanders.
394 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

A*
e
V
B' D'

Figure 10.2 Illustration of planform patterns for stream channels (after Simons et al., 1975; Simons, Li and
Associates, 1982). ( A ) Braided, (B) straight, (C) meandering.

Brice (1983) classified channels into four planforms: planform. T h e interrelationship is known as hydraulic
sinuous canaliform, sinuous point bar, sinuous braided, geometry. N u m e r o u s empirical relationships have b e e n
and nonsinuous braided channels. In general, the slope developed to predict this interrelationship, with equa­
increases with a change from sinuous canaliform to tions generally divided into those that predict geometry
sinuous point bar to braided channels. T h e primary at a point and those that predict changes as o n e moves
difference between the sinuous canaliform and sinuous from o n e point to a n o t h e r on a stream.
point bar is in the frequency of point bars. T h e sinuous As discussed earlier, gravel bed streams and streams
canaliform river has broad uniformly wide channels with large roughness elements frequently have rough­
with high sinuosity and a lack of distinct crossings and ness elements of magnitudes equal to or greater than
pools. the flow d e p t h . Equations developed in this section d o
not apply to those streams. A special section is in­
cluded later in the c h a p t e r to deal with the morphology
Description of such streams a n d their resistance to flow.
Simons, Li and Associates (1982) credit an additional
classification to Culbertson et al. (1967), who use vege­ Hydraulic Geometry at a Point
tation, sinuosity, and bank characteristics to character­ For fluvial channels, surface width B, m e a n d e p t h
ize streams. Bank characteristics include presence of D , m e a n velocity U, and suspended sediment load
oxbow lakes, m e a n d e r scroll patterns, type of bank, Q , are all related to w a t e r discharge, Q. Leopold and
s

bank height, type of natural levee, type of floodplain, M a d d o c k (1953) proposed that each w h e r e related to
and type of vegetal patterns on the bank. T h e s e sub- water discharge by a power function, or
classifications assist in determining possible changes
that might occur with changes in input p a r a m e t e r s . Β = CQ
a
a
(10.1)

D = CQ
b
b
(10.2)
CHANNEL MORPHOLOGY
U = CQ
C
C
(10.3)
Hydraulic Geometry and Shape of Channels Gs = CQ,d
d
(10.4)
W a t e r discharge and sediment load combine to de­
termine stream channel width, depth, velocity, and where C , a C,
b C , C,
c d a, b, c, and d are constants,
Channel Morphology 395

given in C h a p t e r 7. O n e of these procedures should be


used to develop specific sediment discharge curves.
It can be concluded from the relationships presented
that channel width, depth, average velocity, and sedi­
m e n t load all increase with water discharge at a point.
A n example is given in Fig. 10.4A for Brandywine
C r e e k in Pennsylvania (Wolman, 1955).

Hydraulic Geometry along a Reach


Studies similar to those of hydraulic geometry at a
point have been developed for hydraulic geometry along
channel reaches. A n example is shown in Fig. 10.4B for
Brandywine Creek in Pennsylvania. Chang (1988) pre­
sents the results of studies from twenty investigators
worldwide as summarized by Ming (1983). His results
show that the slopes of t h e regression lines of B, U,
and D versus discharge along a channel reach are
remarkably similar. Values of the exponents are:

a = 0.39-0.60, b = 0.29-0.40, c = 0.09-0.28.


T h e s e exponents, again, indicated that width, depth,
a n d flow velocity increase with increasing discharge.
T h e coefficients C , C , and C in Eqs. (10.1)-(10.4)
a b c

I i i i i i ni| 1 1 ii n i | 1 ι ι I d e p e n d o n a n u m b e r of variables, including the size of


100 1,000 10.000 the bed material and the type of channel bank. W h e n
prediction equations that correlate hydraulic geometry
DISCHARGE (cfs) to flow and sediment load are developed, the resulting
Figure 10.3 Suspended sediment load versus discharge for equations are known as regime relationships. T h e range
Brandywine Creek at Wilmington, Delaware (after Wolman, 1955).
of their usefulness is limited since they do not include
many of the important p a r a m e t e r s affecting hydraulic
geometry. Examples of additional p a r a m e t e r s include
interrelated by flow. If the average flow is defined as b e d material size and type of b a n k material. Schumm
Q = BDU, then a logical solution for the constants is (1968) evaluated hydraulic geometry of sand channels
in the G r e a t Plains of t h e United States and the
C A C e - l (10.5) Riverine Plains of New South Wales, Australia. Con­
siderable variation in width, d e p t h , and velocity was
and present in streams with the same discharge. All chan­
nels studied were stable with no major adjustments in
β + 6 + c = 1. (10.6)
10 years. Schumm d e t e r m i n e d that the shape of the
Leopold and Maddock (1953) obtained values for a, b, channel, as determined by the width to depth ratio, F,
and c of 0.26, 0.40, and 0.34, respectively, for rivers in was related to the p e r c e n t a g e of silt and clay, M , in the
the G r e a t Plains and the Southwest. Considerable scat­ b a n k material. Silt and clay were measured as the
ter in the data exists among rivers. They also found p e r c e n t a g e of sediment smaller than 0.074 mm.
that values for d were between 2 and 3, indicating that Schumm found that
sediment load increases with discharge m o r e rapidly
F = 255 M ~ 1 0 8
. (10.7)
than B, D, and U. Values for d vary widely among
rivers, as a result of varying bed material and channel Studies by Schumm and Khan (Schumm, 1968; Khan,
characteristics. A n example of a relationship between 1971; Schumm and Khan, 1972) also show that F is
sediment load and discharge for a given channel and related to the composition of the sediment load as well
location is given in Fig. 10.3 As can be seen from the as the total load. In general, width increases and flow
scatter, there is considerable variation among events. becomes more shallow as sediment load increases. In
Procedures for estimating a relationship between sedi­ studies by Schumm (1960) on ephemeral channels,
ment load and water flow-rate for a given channel and flows with a high percentage of clays and silts tended
location where monitoring data are not available are to have vertical side walls and small width-to-depth
396 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

Δ Β

Discharge (cfs) Bankfull Discharge (cfs)


Figure 10.4 (A) Relationship between hydraulic parameters related to discharge at a point for Brandywine Creek at
Cornog, Pennsylvania. (B) Relationship between hydraulic parameters and discharge along a 25-mile stretch of Brandywine
Creek, Pennsylvania (after Wolman, 1955).

ratios. Channels with primarily sandy sediment load flow and resulting scour. T h u s , the bankfull discharge
tended to have shallow flow depths and large width- is assumed to b e t h e event that forms the channel.
to-depth ratios. Leopold et al. (1964) found that t h e bankfull discharge
Specific regime relationships used for design of flu­ h a d a return period of 1.5 years for 13 stations in t h e
vial channels are given in a subsequent section of this E a s t e r n U n i t e d States. Williams (1978) utilized 233
chapter. sets of d a t a and found that the recurrence interval was
not consistent. C h a n g (1979b) evaluated several d a t a
sets and developed the relationship in Fig. 10.5. Dury
Channel-Forming Discharge (1973) analyzed data for t h e U.S. and showed that t h e
r e t u r n period of t h e bankfull discharge is 1.58 years.
Bankfull discharge is typically considered to b e the
channel-forming event. Flows lower than bankfull dis­
charge transport less sediment and cause less scour.
Channel Gradient
Flows grater than bankfull discharge are absorbed by Channel gradients generally decrease with distance
the overflow areas, without major increases in d e p t h of downstream in association with increased discharge
Alluvial Channel BedForm 397

and decrease in stream bed particle size. In general, braided channel at high slopes. Similar results were
the decrease is exponentially related to distance down­ found for the Mississippi River, in the early 1900s
stream, or before channelization occurred, which cut off some of
the m e a n d e r s and increased slope gradients. These
S = S e'0
aX
(10.8) results are also shown in Fig. 10.6B. O n e would expect,
at any point on the channel, to observe wide variations
and
in the sinuosity as a result of cutoffs and the changes in
dp = d p 0 e^ ,x
(10.9) channel slope that occur. T h e scatter in the Mississippi
river data is likely a result of this changing sinuosity
where S and d are slope and particle diameter at any
p
and length of channels as cutoffs occur.
X\ S and d
0 p 0are slope and particle diameter at X T h e results in Fig. 10.6 point to a threshold value for
equal to zero; and a and β are constants. Regression slope at which a change occurs and the channel goes
equations for slope as a function of m e a n annual from m e a n d e r i n g to braided. T h e presence of such a
discharge, average particle diameter, and percentage of threshold has b e e n verified by L a n d (1957) and Leopold
silt and clay in the total sediment load have been and W o l m a n (1957). T h e analysis of Lane is shown in
developed, but the relationships are not universally Fig. 10.7 indicating that changing the slope of a chan­
applicable. nel can lead to a transition from a meandering pattern
to a braiding p a t t e r n , or vice versa.

Meandering and Channel Sinuosity


As previously stated, a straight channel is inherently
unstable and will tend to develop m e a n d e r s unless
ALLUVIAL CHANNEL BEDFORM
artificially confined. Laboratory studies by Schumm
Description of Channel Bedforms
and Khan (1972) as shown in Fig. 10.6A, showed that
the degree of sinuosity for a laboratory sand channel Channel bedforms have long been of interest be­
with steady flow of 0.15 cfs increased with increasing cause of their impact on flow and sediment transport.
slope u p to a point and then declined until becoming a As bedforms change, resistance, flow depth, and width
398 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design
0.01
1.4
4)

o
£ 00.01
0)
' " ' T ^ - . ^ J
ο.
J

Straight Meandering , \ Braided α

a α ο.οοοι
1.1
cd
CO Ο

00.000110. 10 10 10 10« 10 10
0.9
0.4 0.8 1.2 1.6 2.0 2 J 3 e

Slope (percent) Mean Annual Discharge (cfs)


Figure 10.7 Lane's (1957) analysis of meandering as related to
Β 2.5 annual discharge and slope. The regression lines developed by Lane
are taken from streams ranging from highly meandering to braided.

ο
slopes approximately at the angle of repose. D u n e s
may have ripples superimposed on their surface.
CO
Transition or plane bed: As stream power continues
to increase, the roughness elements start to wash out
and a plane bed again forms. This form is also called
washed out d u n e s or sand waves. T h e bedform is
Valley Slope (ft/mile) unstable, frequently alternating between dunes, plane
Figure 10.6 (A) Laboratory studies. Relationship between slope
bed, and antidunes.
and sinuosity for laboratory studies at a flow rate of 0.15 cfs. The Antidunes: With a continued increase in stream
relationship would vary, dependent on flow rate and channel proper­ power, a bedform known as antidunes appears. A n ­
ties (after Schumm and Khan, 1972). (B) Relationship between slope tidunes have an a p p a r e n t upstream movement, al­
and sinuosity for the Mississippi River between Cairo, Illinois, and
though the net movement of sediment is downstream.
Head of Passes, Louisiana. Data obtained between 1911 and 1915
surveys before cutoffs (after Simons, Li and Associates, 1982). T h e surface water profile is in phase with the an­
tidunes. T h e amplitude of the antidunes increases as
stream power increases and surface waves reach a
point of breaking, leading to the term antidunes with
also change. Bedforms, however, are induced by flow breaking waves.
rates and shear forces: hence, there is a nonlinear Chutes and Pools: With a further increase in stream
relationship between flow and channel geometry for power, the bedform becomes chutes and pools. Sedi­
fluvial channels. ment deposits in large mounds, connected by chutes
As flow rates and bed shear forces change, t h e r e is with supercritical flow and pools that may be either
resulting change in bedforms as shown in Fig. 10.8. A supercritical or subcritical.
description of the characteristics of each bedform fol­
Characteristics of each of the bedforms are given in
lows. T h e bedforms follow an increase of stream power,
Table 10.2. Of particular interest is the roughness,
T U, where r is shear on the channel bed and U is the
0 0
which increases with increasing stream power and bed-
mean velocity.
form, but decreases significantly as the flow transitions
Flat bed: A channel with no bedforms occurring at from lower regime to u p p e r regime.
low stream power. Alternately, it is called a plane bed
and a smooth bed.
Prediction of Bedforms
Ripples: As stream power increases, roughness ele­
ments known as ripples form that have no consistent A n u m b e r of studies of bedforms have b e e n con­
wavelength. Average wavelengths are typically less than ducted using flume data as well as that from canals and
1 ft long and heights are less than 0.1 ft. natural channels (Simons and Richardson, 1966;
Dunes: As stream power increases further, bedforms Athaullah, 1968; van Rijn, 1984). P e r h a p s the most
known as dunes occur that are out of phase with the widely used relationship is that of Simons and Richard­
water surface. D u n e s tend to b e triangular with u p - son (1966), shown in Fig. 10.9. T o predict t h e bedform,
Alluvial Channel BedForm 399

A) T y p i c a l R i p p l e P a t t e r n E) P l a n e Bed

B) D u n e * w i t h R i p p l e S u p e r p o a e d F) A n t i d u n e e , S t a n d i n g Waves

C) D u n e s G) A n t i d u n e e , B r e a k i n g Waves

D) Washed O u t D u n e s o r T r a n s i t i o n H) C h u t e s a n d Pools
Figure 10.8 Bedforms in alluvial channels (after Simons and Richardson, 1966).

Table 10.2 Bedform Characteristics (after Simons et al, 1965; Simons and Richardson, 1966)

Phase
relation
between
Bed material Mode of bed and
concentration sediment Type of Friction water
Regime Bedform (ppm) transport roughness factor surface

Lower flow Ripples 1O-200 Discrete Form 0.05-0.13 Out of


regime Dunes with steps roughness phase
ripples dominates
Dunes 100-2000 0.05-0.15
Transition Alternates 1000-4000 Variable 0.03-0.08 Variable
zone between
dunes, plane
bed, and
antidunes
Upper flow Plane beds, 1500-3000 Continuous Grain 0.02-0.03 In phase
regime antidunes, >5000 roughness 0.03-0.08
chutes, and dominates 0.07-0.10
pools
400 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

relates Manning's η to particle diameter, or

η = </ 1 / 6
/21.1, (10.10)

w h e r e d is diameter in meters. F o r d in feet, E q .


(10.10) becomes Eq. (4.32), or

η =</ 1 / 6
/25.7. (10.11)

Using either Eq. (10.10) or (10.11) in Manning's equa­


tion, a relationship known as the M a n n i n g - S t r i c k l e r
equation results [Eq. (4.33)], or

U/U* =6.74(ΛΛ0 1 / 6
, (10.12)

w h e r e U is m e a n velocity and U* is shear velocity


given by

£/* = yfgRS . (10.13)

W h e n t h e channel is lined with mixtures of particles,


0.001 U ι ι ι I I
0 0.2 0.4 0.6 0.8 1.0 the average diameter is obviously not the diameter that
controls roughness. Meyer-Peter a n d Muller (1948)
proposed that d^ be used and that
Median Fall D i a m e t e r (mm)
Figure 10.9 Bedforms as related to stream power and median η = dl£ /26, 2
(10.14)
particle diameter (after Simons and Richardson, 1961).
w h e r e d^ is given in meters.
Equations (10.10M10.14) are for fully turbulent flow
and a fully turbulent boundary layer. F o r conditions
one needs only to know the stream power and median
w h e r e t h e boundary layer is not turbulent, i.e., a lami­
particle diameter.
n a r sublayer exists, o t h e r relationships are n e e d e d .
Two observations can be m a d e from Fig. 10.9. First,
Einstein (1950) proposed that the logarithmic velocity
for diameters greater than 0.64 mm, the bedform goes
profile be applied for both turbulent and laminar sub­
directly from flat to dunes without the intermediate
layer conditions a n d proposed t h e following relation­
ripple form. Also, based on the slopes of the transition
ships:
line between bedforms, d u n e bedforms will disappear
as one gets to large gravel-sized particles.
^ - = 5.751og (30.2^-) 10 (10.15)

FLOW RESISTANCE and

U ι D\
Channel roughness is composed of two components, — = 5.751og |l2.27-j, 1 0 (10.16)
grain roughness and form roughness. Grain roughness
is related to the shear forces and form roughness is
w h e r e u is t h e velocity at some distance ζ above the
related to the changes in pressure that result from
channel bottom, U is average velocity, D is d e p t h of
interactions of bedforms with the flow. T o u n d e r s t a n d
flow, and Δ is an a p p a r e n t roughness. Δ is related to
roughness of moveable beds, it is first necessary to
t h e roughness of the channel boundary, k , by
evaluate roughness of fixed beds. s

A = k/ , s X (10.17a)
Fixed Bed Roughness
w h e r e k is channel roughness (d ) and χ is a factor
s 65

If channels are lined with unmoveable uniform sand given in Fig. 10.10 that is used to correct t h e logarith­
or gravel elements of size d, the resistance is a result of mic velocity profile as transition occurs from a smooth
grain roughness. O n e widely used relationship for p r e ­ to a rough boundary layer. T h e p a r a m e t e r χ in Fig.
dicting roughness is the Strickler (1923) formula, which 10.10 is given as a function of k /S, w h e r e δ is t h e
s
Flow Resistance 401
1.8 roughness, and TJ is shear d u e to form roughness. If
Eq. (10.21) is divided by p i / , then a friction factor
2

\
1.6
relationship results, or

\
1.4 Smooth/ k =d65
s

Wall /
Rough (10.22)
X 1.2

//
Wail
w h e r e / , / ' , and / " are friction factors for the channel,

/
1.0
grain roughness, and form roughness, respectively. In
0.8
the following discussion, m e t h o d s for partitioning the
Δ= k s

0.6 ι iimn 1 1 1 1 I I II X hydraulic radius and channel slope are given. Methods
0.1 1.0 10 1 1 1 I I 100
1II for partitioning friction factors are given in Chang
(1988) and o t h e r references.

Figure 10.10 Einstein's correction factor χ for the logarithmic Partitioning Hydraulic Radius - Einstein Barbarossa
velocity profile (After Einstein, 1950). Method
Einstein a n d Barbarossa (1952) divided total resis­
tance into friction a n d form drag by partitioning the
thickness of the laminar sublayer given by hydraulic radius. They assumed that velocity could be
predicted by using a modification of the logarithmic
8 = 11.6 v/U*, (10.17b) velocity profile given by Eq. (10.16). In this partition­
ing, the hydraulic radius for predicting m e a n velocity in
where ν is kinematic viscosity (L T~ ). 2 {

Eq. (10.16) was assumed to b e that for grain roughness,


For flow in gravel bed channels, the relationships
R', or
used here need to be specially calibrated. T h e F e d e r a l
Highway Administration (1975) evaluated several data U I R'
x

sets and developed an equation for rock riprap as — = 5.751og 10 12.27— (10.23)
U'
1/2
ng
D l / 6
= 0.225 Μ 1/6
(10.18) where
D) ' (10.24)
which yields Eq. (4.32). Each of the terms in Eq.
(10.18) must have consistent dimensions. a n d χ in Fig. 10.10 is given as a function of k /S', with s

k defined by d . In calculating δ', of course,


s 65 is
used and in t h e calculation of t/*, R! is used. For a
Fluvial Bed Roughness given R\ the m e a n velocity and hence discharge can be
Predictors of roughness and velocities for beds with calculated. T h e difficulty in using Eq. (10.23) comes in
form roughness are based on partitions of drag b e ­ determining t h e value for R" and hence R for a given
tween form and friction drag. O n e approach to parti­ R'. T o m a k e this possible, Einstein and Barbarossa
tioning drag includes partitioning the energy slope, or proposed that

U
5 = 5 ' + S", (10.19) (10.25)
where 5 is total slope, S' is the slope required to
overcome the friction drag, and 5" is the slope re­ where
quired to overcome the form drag. Partitioning can
( P s - P ) ^35 ^35
also be based on the hydraulic radius, or Φ35 = - - ( S G - . ) - , (.0.26)

R = R' + R\ (10.20)
w h e r e S G is t h e specific gravity. T h e functional rela­
where R' and R" are the hydraulic radius relative to tionship for E q . (10.25) is shown in Fig. 10.11. Proce­
friction and form drag respectively. If Eq. (10.20) is d u r e s for using t h e relationship are given in a subse­
multiplied by yS, then a shear stress relationship re­ q u e n t example.
sults, or
Partitioning C h a n n e l S h e a r — E n g e l u n d Method
Το = τΌ + τ 5 , (10.21)
A n o t h e r widely used resistance formula is that of
where r 0 is total shear, r' 0 is shear d u e to grain E n g e l u n d (1966), in which the slope is divided into
402 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

Ο Missouri River near Fort Randall. SD


0 Missouri River near Pierre, SD
Q Missouri River near Omaha, NB
> A Elkhorn River near Waterloo, Ν Β
50 < Ί ή- Big Sloue River near Akron, IA
i X Platte River near Ashland, NB
0 Niobrara River near Butte, NB
f
ψ Salinas River at San Lucas, CA
• Nacimlento River near Juctlon, CA
4 Salinas River near Pasco Robles,
7

• Oh Ο
c

10
•f
J
if
c
5

0.5 1.0 5 10 50

Ψ
r as -
-
35
Ρ R'S
Figure 10.11 Einstein and Barbarossa relationship for form resistance (after Einstein and
Barbarossa, 1952).

form and grain roughness components. By evaluating E q u a t i o n (10.27) can b e further modified by multiply­
the expansion head loss that results as flow moves over ing by yR/[(y - y)d ]
s and by assuming that yRS'
5Q

a bedform element, Engelund proposed that the head equals yR'S to obtain
loss d u e to flow over a form roughness element of
length λ and height A is caused by expansion loss, (10.29)
Δ / / " = α ( Δ ί / ) / 2 # , where Δ ί / is the flow change
2

where
over the roughness element. Since U is discharge per
unit width over d e p t h (q/D) and the d e p t h on the yRS
form roughness in the lee of the form roughness ele­ τ* = (10.30)
(? - Ύ)<*50
ment is D - A / 2 and D + A / 2 , then the slope could
s

be divided into grain roughness and form roughness yR'S


τ* = (10.31)
components by

Q
and
5 = 5' + (10.27)
2g\ D - A/2 D + A/2 yR"S
τ* =
where D is d e p t h of flow, 5 ' is slope due to grain (y -
s y)d 50

(10.32)

-(f)
roughness, α is a loss coefficient, A and λ are the yh
u
2
2

height and average wavelength, respectively, of the


(y - y)^so gR
form roughness elements, and U is the mean velocity 5

of the flow. T h e second term on the right-hand side is Shear due to grain roughness is based on the logarith­
Δ Η/λ. Equation (10.27) can be modified to mic formula (Engelund and H a n s e n , 1967)

aU2
U R'
(10.28) — = 6 + 2.5 In (10.33)
2^A 2.5d 65
Flow Resistance 403

(·) T * = 1.581 ( % ' - 06) 1 / 2

0.04. I I I I Mill I I,I I Mill L 0.04


(b) T * - 7 V D » 0.47mm
•1/Ι.β
(c) T.= [ΐ.425(Τ*)" · -42δ]- 1 β

I I — Γ Τ Τ 1 1—ΓΤ
0.03 J Ripples | Dunes L 0.03

*tu>
d

0.02. L 0.02

0.01. 0.01
I I I I Mill 1 I I I Mill Γ
0.1 1.0

TU 0 (ft-lb/sec/ft ) 2

Figure 10.13 Flume data showing a relationship between bed


70 S roughness, bedform, and stream power (after Chang, 1979b).

(r -r)d o
8 5

Figure 10.12 Engelund's relationship for total shear and grain


shear. Solution:
Einstein-Barbarossa Method
1. Assume a value for R' and calculate a trial velocity.
Assume that R! =«1.0 to start the computation. From Eqs.
(10.23), (10.24), and (10.17) using R' instead of R,

Using this model, Engelund and H a n s e n (1967) devel­ U'* = (gR'S) x/2
= (32.2 x 1.0 x 0.0005) 172
= 0.127 ft/sec
oped the relationship in Fig. 10.12. According to C h a n g
II.61/ (11.6)(10" f t / s e c ) 5 2

(1988), the relationship is generally satisfactory for δ' = — = ; ___ = 0.000913 ft


v

predicting stage-discharge relationships, but should not t/; 0.127 ft/sec


be used for channels with very coarse sand beds. Pro­ k^ 0.00229 ft
cedures for making the calculations are illustrated in a = 2.51.
δ' 0.000913 ft
subsequent problem.
From Fig. 10.10 with kj& = 2.51, χ = 1.25. Using Eq.
From Fig. 10.12, it can be seen that a discontinuity
(10.23) with A: = d = 0.00229 ft,
exists in the roughness curve at τ * of 1.0, resulting
s 65

from a transition from the lower regime to the u p p e r 12.27/?'*


regime. Such a discontinuity has b e e n supported by i/ = (7; 5.75 log,
flume data as shown in Fig. 10.13, but it is less common
/ 12.27 X 1.0 X 1.25 \
in natural channels d u e to t h e variation in flow regimes = (0.127)(5.75).og ( — 10 9 ) = 2.79 ft/sec.
that exist across a cross section of varying depth.
2. Calculating R". Calculate a value for ψ , assuming that 35

R' is 1.0 ft. Utilizing a value of 2.65 for specific gravity,


(SG - l)d (2.65 - 1)(0.000492)
Example Problem 10.1 Calculation of
35
φ <- —
Ί = = 1.62.
stage-discharge curves
Ψ 3 5
R'S (1.0)(0.0005)
From Fig. 10.11, U/U'i = 26; hence
A channel has a flow width of 20 ft and a maximum depth
of 4.0 ft before overtopping the banks. The channel slope is U 2.79 ft/sec
0.0005 ft/ft. Bed material has the following particle size Ul = — = ^0.107 ft/sec
information: d = 0.15 mm (0.000492 ft), d = 0.3 mm
35 50 7/2
(0.107)
u:
2

(0.000984 ft), d = 0.7 mm (0.00229 ft). Develop a stage-dis­


65
R" = = 0.711 ft.
charge curve for depths up to bankfull discharge using the gS (32.2)(0.0005)
Einstein—Barbarossa and the Engelund methods. Assume
that the kinematic viscosity is 1 0 " ft /sec and that the 5 2 3. Calculate R, A, and Q.
channel is rectangular. R = R' + R" = 1.0 + 0.711 = 1.711 ft.
404 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

For a rectangular channel D = BR/(B - 2R) = 20/?/(20


-2R),
(9) R" = uz = 62.11i/f
(20)(1.711) gS (32.2)(0.0005)
D = = 2.06 (10) R - R' + R"
20 - (2)(1.711)
BR 20R
A = BD = (20)(2.06) = 41.2 ft 2

(11) D - Β -2R 20-2/?


Q =AU = (41.2)(2.79) = 114.9 cfs. (12) Q = BDU
4. Tabulation for other values of R'. Other values are Engelund Method
tabulated and shown below. 1. Assume a value for R' and calculate τ'*. Assume R' = 1.0
ft. From Eq. (10.31),
(1) (2) (3) (4) (5) (6) (7)
yR'S
/?'
u: V* X U Ψ35 uius
(r - Ύ)<*5Ο
s
0.1 0.040 0.79 1.50 0.67 16.2 7.0
0.5 0.090 1.78 1.50 1.86 3.24 15.0 R'S (1.0)(0.0005)
= 0.307.
1.0 0.127 2.51 1.25 2.79 1.62 26.0 (SG - l)rf 50
=
(1.65)(0.000984)
1.5 0.155 3.06 1.15 3.53 1.08 37.0
2.0 0.179 3.53 1.12 4.20 0.81 51.0 2. Determining τ * . From Fig. 10.12, since τ'* < 0.55,
3.0
lower flow regime is assumed. Using Eq. (a) in Fig. 10.12 for
0.219 4.32 1.08 5.34 0.54 85.0
T'„ - 0.307,
4.0 0.254 5.01 1.02 6.34 0.41 150.0

τ * = 1.581(τ; - 0 . 0 6 ) 05
= 0.786.
(1) (8) (9) (10) (11) (12)
3. Calculate depth, velocity, and discharge. From definition
R'
(ft)
υ:
(ft/sec)
R"
(ft)
R
(ft)
D
(ft)
Q
(cfs) of r. [Eq. (10.30)],

0.1 0.096 0.572 0.672 0.72 9.6


Mr,-?)*** M S G - 1 ) 450
0.5 0.124 0.955 1.455 1.70 63.2 R =
yS S
1.0 0.107 0.711 1.711 2.06 114.9
1.5 0.095 0.560 2.060 2.59 182.9 (0.786)(1.65)(0.000984)
= 2.55.
2.0 0.082 0.418 2.418 3.19 268.0 0.0005
3.0 0.062 0.239 3.239 4.79 511.6
For a 20-ft-wide channel,
4.0 0.042 0.110 4.110 6.98 885.1

BR (20)(2.55)
D = = 3.42.
Column entries in this table are determined from Β -2R 20 - (2)(2.55)
(1) Assumed value for R'
From Eq. (10.33) for mean velocity,
(2) £/; = v^R\S = ^32.2/? (0.0005) ft/sec, ,
Eq. (10.24)
k R'
=19.74£/;, Eq. (10.17) 6 + 2.5 In
(3) ^ =
δ' U.6v/U
r
u= i/; 2.5^ 65 t

IR'\
Fct — I,, Fig
(4) χ = Fct\ Fig. 10.10 Ϊ / ; = V i f ^ = ^/(32.2)(l)(0.0005) = 0.127
\ ° I
d = 0.00229 (given).
R'X 65

(5) U = £/;5.75log 10 12.27— , Eq. (10.23)


Hence
(1.65)(0.000492ft) 1.62
(6) φ = (SG - 1) 35L

35
(R'S) (/?')(0.0005) R' ' 1.0
U = 0.127 6 + 2.5 In
Eq. (10.26) (2.5)(0.00229)
U
(7) — = ^ / ( ψ Χ 3 5 Fig. 10.11 = 2.40 ft/sec.

U From continuity for a rectangular channel,


(8) Ul =
u/ui Q = BDU - (20)(3.42)(2.40) = 164.2 cfs.
Channels in Regime 405

4. Tabulations for other R' values.


Q * Constant
d » Constant •Max
0)
7s =• Constant ι / Q a t Constant
(1) (2) (3) (4) (5) (6) (7) (8)
s
Or
1 ' Slope
R' R D
(ft)
u:
(ft/sec)
U
(ft/sec)
τ.' τ.
(ft) (ft)
Q
(cfs)
\
\
1

1
s
^
S
es vim ^W 1
* \ A \ i S
ο
-1
0.5 0.090 1.55 0.154 0.485 1.58 1.88 58.3 0)
\ > ^
1.0 0.127 2.40 0.307 0.786 2.55 3.42 165.1 Ο I^MIn. I \ζ a
CO

a
1.5 0.155 3.09 0.461 1.001 3.26 4.84 297.3
B 2 \
2.0 0.179 3.69 0.614 0.614 2.00 2.50 184.5 1
/ Slope at
1 1 \
/ Constant Q. CO
3.0 0.219 4.74 0.921 0.921 3.00 4.29 406.7 1 \

0.254 5.68 1.228 1.381 4.50 8.18 929.2


/ 1
4.0
' τ f

Column entries in this table determined from Width


Figure 10.14 Illustration of the minimum stream power and maxi­
(1) Assume a value for R' mum sediment efficiency as applied to stable width determination.
(2) £/; - yfeRS - ν (32.2)Λ (0.0005) / ,

For a constant β» t n e
minimum value for 5 also corresponds to the
R' minimum stream power, yQS> and the maximum sediment transport
(3) U = t / ; 16 + 2.5 In Eq. (10.33) (after Chang, 1988; copyright © 1988. Reprinted by permission of
2.5d 65 t
John Wiley & Sons, Inc.)
yR'S
(4) ; =T
= 0.307R', Eq. (10.31)
(r - y) so s
d

(5) τ jk from Fig. 10.12 ical relationships were developed by Lacey (1930) fol­
yRS RS
Eq. (10.30) lowed by Blench (1970) and Simons and Albertson
*
( 6 ) r
(r - y)d
=
s (SG - 1 ) ^
50
= 1
(1960). M o r e recently, C h a n g (1980b, 1985b) developed
/? = T , / 0 . 3 0 7
£Λ 20 what is described as a rational basis for a regime
(7)£> = concept based on minimization of stream power. Chang
(B - 2R) (20 - 2R)
(1979b, p . 311) proposed as his hypothesis for stream
(8) β = BDU = 20 Di/.
Note: The stage-discharge curve results from a plot of Q power minimization:
versus D.
For an alluvial channel, the necessary and
sufficient condition of equilibrium occurs when
the stream power p e r unit channel length yQS
CHANNELS IN REGIME is a minimum subject to given constraints.
H e n c e , an alluvial channel with water dis­
The Regime Concept charge Q and sediment load Q as indepen­ s

dent variables tends to establish its width,


A channel is assumed to be in regime if it has not
d e p t h and slope such that yQS is a minimum.
changed its characteristics over a long period of time.
Since β is a given p a r a m e t e r , minimum yQS
This would imply that the channel hydraulic geometry
also m e a n s minimum channel slope S.
was in a state of equilibrium with the incoming flow
rate and sediment load such that the flow was just
carrying the sediment load introduced into the stream. Application of this principle is illustrated in Fig.
Such a concept would be significantly different for a 10.14. A family of curves with varying Q could be s

channel with widely varying flows as c o m p a r e d to a plotted like that of slope versus width in Fig. 10.14. For
canal with a reasonably stable discharge. Both flow a constant width, Q would increase with slope. Thus,
s

situations are discussed. for a constant slope, Q versus width would display thes

shape known, illustrating that the width corresponding


to t h e m i n i m u m stream power is also the most efficient
Regime Equations for Canals and Channels with
section for sediment transport.
Steady Discharge
T h e m e t h o d s of Blench, Simons and Albertson, and
Regime theory originated with analyses on irrigation C h a n g are illustrated in this section. In addition to the
canals in India and Pakistan in an attempt to develop information p r e s e n t e d on these methodologies, Chang
stable designs for canals transporting steady discharges (1988) points out that experience on canals in India
of water with high sediment loads. Early strictly empir­ and Pakistan indicates that F r o u d e numbers greater
406 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

than 0.3 make maintenance of straight channels diffi­ Table 10.3 Coefficient Values for Simons and Albertson
cult, even when designed with regime equations. A t Regime Equations (after Chang, 1988)
higher F r o u d e numbers, a thalweg is formed with a
Channel type*
tendency to m e a n d e r .
Coefficient 1 2 3 4 5
Simons a n d Albertson's Method
*b 3.50 2.60 1.70 2.20 1.75
Using data collected in India and Pakistan as well as
the Western United States, Simons a n d Albertson *d 0.52 0.44 0.34 0.37 0.23
(1960) divided channels into five types: (1) sand beds *u 13.90 16.90 16.00 — 17.90
and banks, (2) sand beds a n d cohesive banks with n o *r 0.33 0.54 — 0.87 —
sediment load, (3) sand beds and cohesive banks with ξ 0.33 0.33 0.29 — 0.29
sediment loads between 2000 and 8000 m g / l i t e r , (4)
cohesive b e d and banks, a n d (5) coarse noncohesive ^Channel types are: (1) Sand beds and banks, (2) sand beds and
material. Simons a n d Albertson developed t h r e e groups cohesive banks with no sediment load, (3) sand beds and cohesive banks
with sediment loads between 2000 and 8000 mg/liter, (4) cohesive bed
of graphical relationships a n d equations to define sta­
and banks, and (5) coarse non-cohesive material.
ble width and wetted perimeter, stable d e p t h a n d hy­
draulic radius, a n d stable velocity.
Chang (1988) grouped t h e Simons and Albertson
(1960) equations into three categories a n d p r e s e n t e d T h e s e constants a r e given in T a b l e 10.3. Equation
equations for t h e graphical relationships. Equations (10.40) is only applicable to t h e channel types 1, 2, a n d
for stable w e t t e d p e r i m e t e r a n d stable w i d t h 4 as identified in Table 10.3.
( f o o t - p o u n d - s e c o n d s ) a r e of t h e form Calculation of discharge from E q s . (10.34)-(10.41) is
accomplished by continuity, or
Ρ = KQ b
l/2
(10.34)
B„ = 0 . 9 P (10.35) Q = UB D
a
(10.42)

and F o r classification purposes, sediment sizes, channel


£ s = 1.0871?, + 2.17, (10.36) banks, a n d b e d a r e assumed to b e sand bed channels if
t h e b e d a n d b a n k materials a r e m e d i u m to fine sands.
where Ρ is the wetted perimeter in feet, K is a b Cohesive b e n d materials a r e those finer than sand size
constant d e p e n d e n t on type of channel, Q is discharge and coarse b e d materials a r e those with m e d i u m parti­
in cfs, B is average channel width in feet, a n d B is
a s cle sizes between 20 a n d 82 m m .
top width in feet. Equations for stable flow d e p t h a n d
hydraulic radius are of the form Blench Regime Method
Blench (1970) developed a regime m e t h o d that uti­
R = KQ 036
(10.37)
d
lizes a b e d factor, a side factor, a n d a flow resistance
D=\2\R\ R<lit (10.38) equation to develop relationships for stable width,
d e p t h , a n d slope. His relationships a r e
D = 2 + 0.93Λ; R > 1 ft, (10.39)
1/2
where R is t h e hydraulic radius and Κ is a coefficient ό FQ
b

d e p e n d e n t on channel type. Equations for stable flow B = (10.43)


rates are of the form
w h e r e B is average width in feet, F a n d F a r e b e d
a b s

U = K (R S)*
U
2
(10.40) and side factors, a n d Q is discharge in cfs. Average
and d e p t h in feet is d e t e r m i n e d as
0.37
U 2
UB» 1/3
(10.41)
~gDS D = (10.44)

where U is average velocity in the channel, ν is kine­


and stable slope is
matic viscosity, and K , K , a n d ε a r e constants d e ­
u r

p e n d e n t on channel type. Values for t h e constants ^•5/6^1/12


were developed by Chang (1988) from graphical rela­ S = (10.45a)
tionships presented by Simons and Albertson (1960). K Q {\
x
x/b
+ C/2330) '
Channels in Regime 407

where discharges. Using sediment discharge relationships and


3.63g stage flow rate predictors, Chang developed a model
(10.45b) known as F L U V I A L that determines the width corre­
sponding to a minimum stream power. A flow diagram
ν is kinematic viscosity in square feet p e r second, g is of F L U V I A L is given in Fig. 10.15.
the gravitational constant in feet per second squared, Based on computations using F L U V I A L , Chang
and C is sediment concentration in p p m by weight. (1985b) developed the graphical relationship in Fig.
Values for the bed and side factors w e r e d e t e r m i n e d 10.16 for side slopes of 2 : 1 , but subsequently sug­
by Blench to be gested applying the results to 1.5:1 channels also
(Chang, 1988). Similar curves could have been obtained
(10.46) for o t h e r sides slopes.
T h e critical slope for bedload movement in Fig.
and
10.16 is given by
F = 0.1
s slightly cohesive banks (very friable)
= 0.2 moderately cohesive banks (silty clay loam) -0.51
= 0.00238β (10.48)
= 0.3 highly cohesive banks (tough clay) (10.47)

where d 50is median diameter of sediment in millime­


ters. Blench also proposed that F increases with sedi­ w h e r e d is average particle size in millimeters and Q
5Q
b

m e n t concentration by 1 2 % for every 1% increase in C. is flow rate in cfs. Empirical relationships for surface
width Β and d e p t h D are
Chang's Rational Method
Chang (1980b, 1985b) utilize the concept of mini­ , 0.05

m u m stream power, yQS, to develop stable channel Β = 4.17 β 0 5


(10.49)
width, depth, and slopes for channels with constant

Q=l,000cf8
E n t e r Q.Qs
d, z, vt e t c . d=0.3mm <
Q =0.05cfs - 1 0 CO
I
9

\ z=2.0 18
I A s s u m e C h a n n e l Width B | * - Ο \ X
\ Ιβ
Ο
I v s
I ^ | A s s u m e C h a n n e l D e p t h Ί)| -4
Τ τ
Min. Z-2 or
C o m p u t e Slope Using
Sediment Transport Formula CO

C o m p u t e Q Using 10. Min. 700


Flow R e s i s t a n c e F o r m u l a -600

- A — - -500

No -400 to
-300
0. 200

0.025. - 4
No
—___n_
- 3 •a
V
—2

50 100 150 200

Β (ft)
Figure 10.15 ( A ) Flow diagram for F L U V I A L . (B) Illustration of the change in stream power (yQS) and other
parameters with changes in channel width (after Chang, 1980b).
408 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

10 10 2
10 3
10 4

Q (cfs)
Figure 10.16 Stable alluvial canal (or steady flow channel) design chart for side slopes of 1 . 5 - 2 : 1 (after
Chang, 1985b). Also shown is the bed load and velocity relationship for the stable section.
Channels in Regime 409
and Simons and Albertson Methods
1. Coefficients for equations. Channel type is similar to sand
/ s s Γ
03
bed and banks (type 1). From Table 10.3,
D = 0.055 - = - -f= Q°\ (10.50)
\ V 50
d
V 50 /
rf K = 3.5,
b K = 0.52,
d K = 13.9,
u

where Β a n d D a r e in feet. A n empirical equation for K = 0.33,


r ε = 0.33.
bedload is
2. Stable R values from Eqs. (10.37) and (10.38).

= 0 - 4 3 3 ^ + -f=. (10.51) R =KQ d


036
= 0.52(50) 036
= 2.126.

Since R < 1 ft, use Eq. (10.38) for D, or


where Q is volumetric bedload flow rate in cfs.
s

D=1.21, R = 2.51.
3. Average channel width [Eqs. (10.34) and (10.35)].
Example Problem 10.2 Illustration
of regime equations B = 0.9 Ρ = 0.9 K Q '
a b
] 2
= 0.9(3.5)(50) 1/2
= 22.27.

4. Average velocity and slope. From continuity, the re­


A flow of 50 cfs and a bed sediment load of 3.12 lb/sec is quired velocity is
to be conveyed in a canal in alluvial material with a median
bed particle size of 0.40 mm. Determine the stable width, Q Q 50
depth, and slope by the Blench, Simons and Albertson, and u = — = = 0.874 ft/sec.
A BD a (22.27)(2.57)
Chang methods if the channel banks are slightly noncohesive
with 1.5 :1 side slopes. Assume a viscosity of 1 0 " ft /sec. If 5 2

From Eqs. (10.40) and (10.41) for stable slopes,


the actual groundslope is 0.0003 ft/ft, what would be recom­
mended? [U/K ] u [0.874/13.9]
l/e 17033

S = ^ — = τ = 0.000050387
Solution: R 2
(2.13)'
Blench Method
1. Bed and side factors. From Eq. (10.46) with d = 50

0.4 mm,
U 2

S =
F = (1.9)(0.4)
h
05
= 1.20. gDK {UBJvf
x

From Eq. (10.47) for slightly cohesive banks, (0.874) z

= 0.000132
F.-0.1. (32.2)(2.57)(0.33)((0.876)(22.2)/10- ) 5

2. Average width and depth. From Eq. (10.43) for width


The two slopes are based on Reynold's number-type relation­
0.5
(1·20)(50) ships and on tractive force relationships.
Β = = 24.5 ft.
0.1
Λ

Chang's Method
1. Calculating a stable slope. Chang's method for analyzing
From Eq. (10.44) for depth
canals does not lend itself to the calculations shown above. A
D = , (0.1)(50)
1/3 slope must be specified. Starting with sediment load, a slope
= 1.51 ft. that will just transport the sediment can be defined. From the
1.20 2

calculations above, C = 1000 ppm. From Fig. 10.16 with


3. Average stable slope. Concentration of sediment C = 1000 ppm and Q = 50 cfs,

Qs 3.12 S/Jd = 1.52 x 10" . 3

C = = 0.001 lb/lb = 1000 ppm.


yQ (62.4)(50)
2. Stable channel dimension. From Fig. 10.16 with
From Eq. (10.45) for stable slopes S/d = 1.52 Χ 1 0 " and Q = 50 cfs,
05 3

(3.63)(32.2) # = 2 1 f t

Κ = = 2079 a

(io- )
λ Λ/Λ
5 1/4
D = 1.3 ft
.1/12 U = Q/(B D) = 1.8 ft/sec
(1.2) (0.1) 5/6 a

5 =
( 2 0 7 9 ) ( 5 0 ) ( 1 + 1000/2330)
1/6
S = (1.52 X l O " ) ^ = (1.52 X l O " ) ^ = 0.00096.
3 3

= 0.0001685. Using Eqs. (10.48)-(10.51) would have yielded similar results.


410 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

Summary of Predictions complish this end, t h e channel geometry is self-


forming. C h a n n e l characteristics such as width, d e p t h ,
Simons and slope, sinuosity, b a n k slope, a n d transverse slope in
Blench Albertson Chang b e n d s a r e d e p e n d e n t variables that allow t h e channel
Avg. width (ft) 24.5 22.3 21.0 considerable degrees of freedom t o come into regime
with inflow rate a n d sediment load.
Avg. depth (ft) 1.5 2.57 1.3
Early studies of stable channel morphology, p r e ­
Stable slope (ft/ft) 0.00017 0.000132 0.00096 sented previously in this chapter, w e r e based on statis­
tical analyses of field data, yielding only marginal in­
It should not be surprising that the slopes are drastically sight into t h e mechanics of channel change. M o r e
different. Slopes are highly variable in flat terrain. What is recent developments in modeling channel resistance to
significant is the impact of sediment load. The Simons and flow, sediment transport, flow in channel b e n d s , a n d
Albertson relationship does not consider sediment loads, channel b a n k behavior have given rise to complete
whereas sediment load is considered in both the Blench and channel models that can give a reasonable approxima­
Chang relationships. The Chang relationship is based on tion to natural channel morphology.
theoretical relationships and has been experimentally veri­ C h a n g (1985a) combined a n u m b e r of basic relation­
fied. It would thus be preferred in the absence of field data.
ships into a n overall model of flow a n d sediment trans­
Recommendation for a Ground slope of 0.0003 ft/ft port t o develop a model of stable channel geometry for
1. Discussion of options. Comparing the ground slope to sand b e d rivers. T h e analysis is based on t h e use of
predicted stable slopes where sediment load is considered, it bankfull discharge as t h e stable flow event. In t h e
is obvious that Chang slopes are greater than the ground analysis, physically based equations w e r e used, includ­
slope. This indicates that the ground slope is insufficient to ing relationships for: (1) flow continuity, (2) sediment
transport the sediment load and provide stable channel di­ transport, (3) flow resistance as related to flow rate,
mensions. In the absence of constructed works, deposition
(4) m i n i m u m stream power for a cross section a n d
would be expected with the resulting increase in channel
m i n i m u m stream power for a channel reach, a n d (5)
slope. The most reasonable action would be to develop a
sediment removal facility to reduce sediment load to the relationships for transverse sediment movement a n d
transport capacity. flow in channel b e n d s . Utilizing these relationships,
2. Analysis of stable channel using Chang's relationship. C h a n g (1979a, 1985a, 1986) developed t h e curves shown
For 5 = 0.0003, in Fig. 10.17, which relate discharge, slope, average
diameter, average depth, a n d average b o t t o m width for
S 0.0003
= 0.000474. rivers. C h a n g (1985a) q u o t e d experimental d a t a indi­
7d cating general a g r e e m e n t with theoretical relation­
ships.
From Fig. 10.16 with S/d ' - 0.000474 and Q = 50 cfs, the
0 5

stable sediment load would be In C h a n g ' s m e t h o d t h e channels a r e classified into


several regions.
C = 50 ppm.
Region 1
Also from Fig. 10.16, the stable channel geometry would be
T h e s e channels tend t o have flat slopes, low sedi­
Β = 17 ft
m e n t loads, a n d low velocities. C h a n n e l d e p t h is sensi­
D = 2.5 ft. tive t o slope, b u t width is insensitive t o slope. Flow
3. Recommendations. Construct a sediment diversion sys­ resistance is such that t h e channels a r e in t h e lower
tem to reduce sediment load to 50 ppm and construct a flow regimes. C h a n n e l s would generally be classed as
channel with average width of 17 ft and average flow depth of sinuous canaliform with constant d e p t h s a n d widths
2.5 ft on the prevailing ground slope. t h r o u g h o u t t h e m e a n d e r s . Regime canals normally fall
in this region, indicating that it is possible to have long
straight channels. W h e n flows a r e such that t h e chan­
nel moves intermittently t o region 2, channel sinuosity
Rational Regime Relationship for Natural Channels could result.
Natural channels a r e significantly different from con­ Empirical equations for surface width a n d d e p t h in
stant flow canals in that their flow rates a n d sediment Region 1 a r e
loads change frequently. U n d e r this concept, it is im­
p o r t a n t t o think of a natural channel in regime as o n e k 0.02

that is stable over a long period of time, discharging a Β = 3.49 )0.47 (10.52)
sediment load that equals t h e incoming load. T o ac­
Channels in Regime 411

REGION 4. Steep Braided Streams


REGION 3. Braided Point-bar And
Wide-bend Point-bar
Streams
REGION 2. Straight Braided Streams
REGION 1 . Equlwldth Point-bar
Streams And Stable
Cannals

2000

3000

— — — D .Channel Depth In Feet


c

d5Q— Average Diameter In mm

No Bedload
Movement
—ι—I I I 11 ll| I I 1111| 1 I III
— I " I ' l ιΊιιΐ| " I "I" m T T l
10 10 2
10 3
10 10
4

10
5

Bankfull Discharge (cfs)


Figure 10.17 Stable (regime) channel geometry for sand bed rivers with bankfull discharge (after Chang, 1985a).

and width plots have minimums in both lower and u p p e r


regimes. If the lower regime minimum occurs at slopes
0.4 equal to the valley slope, t h e channel can be straight. If
D c = 0.51(? 0 4 7
exp -0.38 - 1 (10.53) t h e lower regime minimum occurs at slopes lower than
the valley slopes, this implies the possibility of both
riffles and pools to maintain the channel at a slope less
where D is the center d e p t h in feet, Β is surface
c than the valley slope (Chang, 1985a). T h e s e channels
width in feet, d is average particle diameter in mil­
50 tend to b e alternating point bar or sinuous braided.
limeters and Q is bankfull discharge in cfs. For Β and For discharges less t h a n bankfull, streams can become
D in meters and Q in cms, replace the constant 3.49 Region 1 or 2 channels with significantly different
with 5.68 and 0.51 with 0.83. channel forms.
Empirical equations for Β and D for Region 3 are
Region 2
c

This is a region that occurs over a r a t h e r narrow


range of flow conditions. A minimum stream power Β = 33.2β •ί—ϊ (10.54)
exists in the lower regime and in the u p p e r regime, but
with a more defined value at the lower regime. Width-
and
to-depth ratios are usually large, leading to frequent
braided channels. Width and d e p t h are sensitive to
slope. Few natural channels occur in this region. A. = 0.015 - 0.025 In Q - 0.049 In β · ,
0 4 5

N o empirical equations are available for Region 2.


(10.55a)
Region 3
Channels in this region have widths and d e p t h s that with Β a n d D in feet, d in millimeters and Q in cfs.
c 5Q

are very sensitive to slope. A s a result, changes in slope F o r Β a n d D in m e t e r s and Q in cms, replace the
c

can quickly lead to braiding. Stream p o w e r - c h a n n e l constant 33.2 with 278 in Eq. (10.54) a n d use the
412 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

following expression for D c 2. Stable channel width and center depth. From Fig. 10.17
with S/dlf = 0.00179 and Q = 600 cfs.
0.45 D = 3.5 ft
D = - 0 . 1 1 2 - 0.0379 In Q - 0.0743 In Q
Β = 65 ft.
(10.55b) Use of Eq. (10.52) and (10.53) yields similar values.
3. Channel form. From Fig. 10.17, the channel is on the
border of Regions 1 and 2. Region 1 is indicated; therefore,
Region 4
the channel is a sinuous point bar stream with the potential
T h e s e channels are similar to those in Region 3, but to be a braided point bar steam.
have steeper slopes with a greater tendency to braid­
ing. Width-to-depth ratios are typically greater t h a n
100; thus the braided channels tend to be straight.
Prediction equations for Β and D are not available. c
GRAVEL CHANNELS
Chang (1988) utilized a sediment transport equation
by Engelund and H a n s e n (1967) to develop the b e d Resistance to Flow
material transport rate for Regions 1, 3, and 4, or
Region 1 Gravel channels and channels with large roughness
elements typically d o not follow the same resistance
c
relationships that are used for channels with smaller
= 9.66 X 1 0 - C 5 0 5 8
Cr° 1 7
(10.56)
m e d i a n d i a m e t e r b e d materials d u e to t h e highly
three-dimensional flow characteristics a r o u n d rough­
Regions 3 and 4 ness elements. In addition, the major fluid drag with
large roughness elements results from grain roughness
and not t h e presence of mobile bedforms that exist
7.28 X 1 0 - C 6 0 8 7
, (10.57) with smaller particles. T h e empirical d a t a available
indicate that relationships available for more moveable
beds d o not apply to these coarser materials.
where C is the bed material load in p p m by weight, d
Because of its dimensionless form, Bathurst (1985)
is in millimeters, and Q is in cfs. If Q is in cms, then
utilized the D a r c y - W e i s b a c h friction factor to evaluate
the constant in equation (10.56) is 5.27 X 1 0 ~ . Since 5

existing data. F o r uniform flow, it can b e shown that


Q is not a p a r a m e t e r in Eq. (10.57), the constant is
the friction factor is given by
unchanged with the use of Q in cms. Relationships
were not given for Region 2. SgRS
(10.58)
U2

Example Problem 1 0 . 3 . Illustration of Chang's or since U* 2


= gRS,
regime channel geometry relationship
for natural channels
U
_ (10.59)
t/7 V7* =

A bankfull flow of 600 cfs and a bed sediment load


concentration of 1000 ppm is flowing in a channel with a Using data from 16 field sites in the U n i t e d Kingdom,
median size of 0.4 mm. Determine the stable width, depth, Bathurst (1985) evaluated U/U* for a n u m b e r of flows
and slope by Chang's relationships for regime sand bed and developed the relationships in Fig. 10.18. For
streams. slopes greater t h a n 0.004, t h e scatter in the data is
Solution: quite large for any relative submergence, D/d . H4

1. Stable channel slope. Assuming Region 1, Eq. (10.56) Bathurst suggests that n o accurate relationship exists
can be used for slope with Q in cfs, C in ppm, and d in for all situations, but proposed t h e following relation­
millimeters: ship

= 9.66 X 1 0 - C G-° £ = / I = 5 .5.621og


62,o (£)
5 0 5 8 1 7

V^50
8 l 0
10 ( —I +4.0 (10.60)
"84 ,
= 9.66 X 1 0 - ( 1 0 0 0 )
5 0 5 8
(600) - 0 1 7
= 0.00179 m m " / 1 2

as a compromise w h e r e D is the d e p t h of flow and d M

S = 0.00179\/0J = 0.00113. is the d i a m e t e r for which 8 4 % of the roughness ele-


Gravel Channels 413

51 and
Slope > 0.004
-0.410
5 = 0.223Q (10.64)

where D is average d e p t h . Parker used the assump­


a

tion that shear stress at bankfull discharge exceeded


a
ο / Region of the critical shear stress by 2 0 % in developing Eqs.
' Sediment
Transport (10.63) and (10.64).
υ Chang (1980a) utilized the F L U V I A L model with its
a concept of minimum stream power, as illustrated in
b, Fig. 10.14, to develop a rational regime diagram for
0) gravel bed streams in a m a n n e r similar to that used to
Ο
e develop Fig. 10.16. In this development, a modification
$ of the Einstein bed load equation, calibrated for gravel
0) channels, was used for sediment transport calculations.
CO T h e results are shown in Fig. 10.19.
a)
Empirical equations for Β and D in Fig. 10.19 are a

0.001065<4
e
15

Relative Submergence D/ds4 Β = 1.905 + 0.249 In 0.42


047
s>s
SQ c

Figure 10.18 Resistance function for large roughness elements


and steep slopes (after Bathurst, 1985). (10.65)

and

ments are smaller. Bathurst suggests that the expected 0.000442^ό 15

errors associated with using Eq. (10.60) are likely to be Ζλ = 0.2077 + 0.04181η β 0 4 2
, 5>5 ,
SQ 0.42
C

on the order of 25 to 3 5 % . For D/d greater than 4,


S4

no acceptable relationship is available. (10.66)

where Β is surface width in feet, D a is average depth


Regime Relationships for Stable Gravel Bed Streams in feet, and Q is flow in cfs.

Regime relationships for gravel bed streams are dif­


ferent from those presented earlier. Parker (1979) ana­
lyzed empirical data and proposed that Example Problem 10.4. Illustration of regime
relationships for gravel channels
Β 1/2
= 4.4(Q*) (10.61) A major development is being planned in the headwaters
*50 of a mountain stream with bed material of average diameter
of 200 mm. The annual storm flow downstream from the
where Β is the surface width normalized by d , the 50 development is expected to change from 100 to 200 cfs as a
median bed particle diameter, and β * is a dimension­ result of the development. The stable surface width and
less flow p a r a m e t e r given by slope prior to development are 23 ft and 0.024 ft/ft. What
would be the expected change in channel geometry as a
result of the increase in Q.
Q
Q* = 1/2
(10.62) Solution:
[(SG - l)gd ]
5
50
1. Determining expected dimensions prior to development.
For a d of 200 mm, the vertical axis in Fig. 10.19 is
50

where SG is specific gravity of a particle and Q is


bankfull flow rate. Consistent units are n e e d e d in order 50 / 50 \ l M

= 0.024 0048,
for B/d50 and Q* to be dimensionless. Parker (1979) '50
extended the relationships using theoretical calcula­
tions to obtain relationships for D and 5, or From Fig. 10.19 for Q = 100 cfs and S (50/d ) = 0.0048, 50
1 15

the channel is in a zone where no transport is expected. The


channel is close to the threshold where B = 19 ft and a

^ - = 0.253β°* 415
(10.63) D = 1.5 ft. Therefore a width of 23 ft would be expected to
a

«50 be stable.
414 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

' I I 1111 τ—! I I 1111| 1—I I I l l l l | 1—π


Surface width of channel, B, feet ;
Average depth of chonnel, TTo, feet;

d.

0.01
ο
to
T3

0.001

0.0002 Η-ΗΗ) 1 I I lllll| 1 I I Mill-


Anolyticoi Mean Velocity. U, fps

o.oi L
ο
in

CO
0.001 L

0.0002 I I I lllll| 1 I I
Bedload Discharge, Q .cfs
s

d_ — 50 mm
50

in
o.oi Ρ
ο
m

Χ
C/D
0.001

0.0002

Bankfull Discharge (cfs)


Figure 10.19 Rational regime relationship for geometry, mean velocity, and bed load for gravel bed
channels. d 50 is in millimeters (after Chang, 1980a; copyright © 1988. Reprinted by permission o f John
Wiley & Sons, Inc.)
Modeling Channel Response to Change 415

2. Expected change in dimensions with flow change. From tonically with Q d . This principle is illustrated graph­
a 5Q

Fig. 10.19 with Q = 200 cfs, ically in Fig. 10.20. A change in o n e variable requires a
Β = 25 ft corresponding change in a n o t h e r variable to maintain
D = 1.9 ft
a
equilibrium. F o r example, an increase in sediment load,
Q , resulting from deforestation with n o appreciable
Q = 0.005 cfs of sediment. s

change in flow or average particle size, would require


s

The sediment load is a volume flow rate, but is very small. an increase in slope t o maintain equilibrium. T h e in­
Therefore the new channel would be expected to have little creased slope would result from deposition in the
sediment transport. channel of t h e excess of sediment load over t h e trans­
3. Expected change. It would be reasonable to assume that port capacity. T h e deposition would continue until the
the original channel had come into regime with its flow. slope is great e n o u g h to transport the increased sedi­
Assuming that Fig. 10.19 could be used in this situation to
m e n t load.
predict relative changes, then the ratio of width after devel­
opment and prior to development is 25/19 = 1.316. The
surface width, after development, would be expected to be

Β = (1.316)(23) = 30 ft.
Example Problem 10.5. Illustration of Lane's
geomorphic relationship
Since the sediment load is very small, the slope would not be
expected to change; hence 5 = 0.024 is still a stable slope. A channel-straightening project is proposed to eliminate
stream resistance to flow and minimize flooding in a subdivi­
sion. The channel is in relatively deep alluvial material. What
would be the impact of the project on equilibrium channel
properties in the proposed project area and downstream
MODELING CHANNEL RESPONSE TO CHANGE areas.

Over a period of time, channels will tend to come Solution:


into regime with flows, sediment loads, a n d properties 1. Impact in the project area. We assume that flow, sedi­
ment load into the straightened reach, and particle diameter
of the geologic parent material through which t h e
are unchanged. As a result of the straightening activity,
channel flows. These characteristics can change as a
channel slope will increase. Thus, we can write Lane's rela­
result of n a t u r e through climatic shifts or tectonic tionship [Eq. (10.67)] as
activity, and by h u m a n interference through surface
mining, urbanization, deforestation, channel straight­
Q°S ~
+
Qtd%
ening, reservoir construction, etc. Any of these changes
will result in new equilibrium conditions a n d a change
where the superscript 0 implies no change, + implies an
in channel properties. Examples of such changes a r e increase, and - a decrease. Since Q and d are unchanged,
50

abundant. Q must increase to offset the increase in slope. The incom­


s

A s a part of any development project, it would b e ing sediment load to the reach is not increased; thus the
desirable to predict the impact of development on increase in Q in the reach must come from detachment in
s

channel properties. Such a prediction may b e a simple the stream bed. Therefore, it is anticipated that channel
qualitative projection of new equilibrium conditions or erosion will increase in the section.
may be specific detailed predictions of change with 2. Downstream changes. In the downstream section that
time. T h e needs of t h e project will dictate t h e type of has not been straightened, there will be no increase in 5, Q,
analysis conducted. T h e following discussion outlines or d as a result of human activity. However, the sediment
50

load coming into the section has increased from upstream


the methodology for each analysis and gives examples.
channel erosion. Therefore, slope must increase to offset the
increased Q . The increased slope will actually result from
s

Qualitative Predictors aggradation occurring downstream. As seen in the next sec­


tion, an increase in sediment load results in a decrease in
Lane's Geomorphic Relationship depth, an increase in channel width, and an increase in
Lane (1955) developed a well-known geomorphic meandering and sinuosity. From Fig. 10.7, there is a tendency
relationship to predict a balance between t h e sediment toward braiding as the slope is increased. Therefore, the net
result of channel straightening is detailed below.
load, particle diameter, discharge, and slope in an
alluvial channel, or Local change.

QS ~ Q d ,
s 50 (10.67) 1. Increased slope
2. Channel erosion and increased sediment load
where t h e symbol ~ means that QS changes m o n o - 3. Increased channel width
416 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

Figure 10.20 Graphical illustration of Lane's (1955) geomorphic model of channel adjustment.

4. Decreased depth of flow In some cases, it is not possible to predict t h e impact


5. Tendency toward braiding. of a change in Q or Q on hydraulic geometry without
s

resorting t o detailed models of channel response. T h e s e


Downstream change.
a r e indicated in Table 10.4 by a ± superscript. D e ­
1. Aggradation and decreased slope tailed models are addressed in the following section.
2. Decreased meandering
3. Increased channel width
4. Decreased depth of flow
5. Tendency toward braiding. Example Problem 10.6. Illustration of Schumm's
To prevent these unwanted features, it will be necessary to qualitative model
use channel stabilization in the straightened section such as
grade stabilization structures and channel bank stabilization. A reservoir proposed in an arid area is to be built on a
stream that receives constant flow from snow melt. Assuming
that the discharge after the reservoir fills is approximately the
Schumm's Qualitative Model same as that prior to filling and that the sediment discharge
is essentially zero, estimate the qualitative impact of the
Schumm (1969) combined Lane's (1955) geomorphic reservoir on the downstream channel.
response as given by Eq. (10.67) with other regime
Solution: From Schumm's relationship in Table 10.4, the
equations and developed a summary of qualitative rela­
following response
tionships, as summarized in Table 10.4, that include
not only t h e p a r a m e t e r s of β , β , d , and 5 , but
8 5Q

channel depth, ratio of channel width-to-depth, aver­ Q~~ B-D F\-SSt


+

age wavelength of m e a n d e r s , and sinuosity. T h e s e rela­


tionships along with Eq. (10.67) can b e used to provide is proposed, which implies that average width decreases,
a fairly complete qualitative evaluation of channel depth increases, width-to-depth ratio decreases, meander
metamorphosis as inputs change. wavelength decreases, slope decreases, and sinuosity in-
Modeling Channel Response to Change 417

Table 10.4 Schumm's Qualitative Model of Channel Metamorphosis (after


Schumm, 1969,1971a, b)

Increase in discharge alone Decrease in discharge alone


Q- ~ BrDrF-k-S*
Q - B+D+F+X+S-
+

Increase in bed material discharge alone Decrease in bed material discharge alone
(2ί ~ B+DrF+k+PS- β" ~ B-D+F-XSrS*

Discharge and bed material discharge Discharge and bed material discharge
increase together (example: during decrease together (example: discharge
urban construction) from reservoir)
(tQ£~B*D F k+S S-
t ¥ t

Discharge increase, bed material decrease Discharge decrease, bed material increase
(example: urbanization after construction) (example: diversion for irrigation with
Q*Qr ~ &D FfrSS+
¥
sediment laden return flow)
Q-Qt ~ BtD-F+ktS+S-

Q, water flow rate; Q , bed material flow rate; B, average channel width; D , average
a
s

depth; F ratio of channel width to depth; λ, average wavelength of meanders; 5, channel


f

slope; 5 , sinuosity.
U

creases. The most important impact is an increase in depth


and decrease in slope. Since the sediment load is decreased, Example Problem 10.7. Utilization of Chang's
slope must decrease to come into regime. This must result regime channel geometry relationship for predicting
from channel degradation immediately downstream of the equilibrium changes in natural channels
dam, resulting in higher banks. At water withdrawal points,
this translates into increased pumping charges. There is also In Example Problem 10.3, a bankfull flow of 600 cfs and a
an impact on tributaries, resulting from increased slope at bed sediment load of 1000 ppm flowing in a channel with a
the junction with the main channel. median size of 0.40 mm yielded a channel with a bed slope of
0.00113 ft/ft, a width of 65 ft, and an average flow depth of
3.5 ft. If the channel is straightened and the slope increases
to 0.002 ft/ft, what would be the resulting channel width,
Chang's Quantitative Model flow depth, and sediment load? How does this compare
quantitatively to the Lane and Schumm relationships? Ex­
A s described earlier, Chang developed predictors of plain any difference.
channel morphology in sand bed streams by combining
Solution:
physically based equations for: (1) flow continuity, (2)
Chang's Analysis
sediment transport, (3) flow resistance as related to
1. Change in width and depth of flow.
flow rate, (4) minimum stream power for a cross sec­
tion and minimum stream power for a channel reach, 0.002
= 0.00316.
and (5) relationships for transverse sediment move­ ToX
ment and flow in channel bends. Utilizing these rela­
tionships, Chang developed the curves shown earlier in 2. Width and depth. From Fig. 10.17 with Q = 600 cfs and
Fig. 10.17, which relate discharge, slope, average diam­ S/d 05
= 0.00316
eter, average depth, and average bottom width. Using £ = 100 ft
a

the E n g e l u n d - H a n s e n (1967) sediment transport equa­


D = 2.5 ft.
tion, he also developed equations for sediment con­
centration for the equilibrium shapes. With these 3. Sediment load. After the modification, flow is in Region
graphical relationships and associated equations, it is 3 (braided channel); therefore, Eq. (10.57) is used
possible to develop first estimates of changes in chan­ [/0.87 1/0.87
nel regime characteristics corresponding to changes in 0.00316
C =
input. This is illustrated in the following example. 7.28 Χ 10" 6
7.28 Χ 10" 6
418 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

or Requirements of Dynamic Models


C = 1076 ppm.
Water Routing
4. Conclusions. Due to changes in slope the following
W a t e r routing must account for variations in resis­
changes are anticipated
tance resulting from bedform changes, for transverse as
well as longitudinal flow, and for upstream and down­
Before After
stream boundary conditions. W a t e r routing is discussed
Width Β Λ 65 ft 100 ft in C h a p t e r 6 and is typically accomplished using nu­
Depth 3.5 ft 2.5 ft merical solutions to the continuity and m o m e n t u m
Sediment concentration 1000 ppm
equations, or
1076 ppm
Classification Sinuous Braided
dA dQ
^ +
^ - « - =° ( · >
10 68

5. Comparison to Lane and Schumm analysis. Quantita­


tively, the local changes with straightening are the same as and
those indicated from Lane qualitative model in Eq. (10.67),
or I dQ dZ 1 d IQ \
2
Q
g°s ~
+
Q-d° .
x A ^ + g
T ^ A Y x U ) ^
X
g S
^ ^ q
^ ^
Since there is an increase in S with Q and d remaining
50
(10.69)
constant, there must be an increase in Q to balance the
s

relationship. From Table 10.4, it is seen that an increase in w h e r e Q is water discharge, A is t h e cross-sectional
sediment load alone leads to an increase in width, decrease area, χ is t h e streamwise coordinate, q is lateral x

in depth, and a decrease in sinuosity. These changes are inflow of water p e r unit channel length, g is gravita­
consistent with predictions from Chang's relationship. tional constant, 5 is slope of the energy gradient, and
e

Ζ is water surface elevation. U p s t r e a m boundary con­


ditions would b e the inflow hydrograph, and down­
stream b o u n d a r y conditions would be a water surface
DYNAMIC MODELS OF CHANNEL CHANGE level or a s t a g e - d i s c h a r g e relationship. Numerical so­
lution possibilities a r e included in a n u m b e r of s t a n :

A s has already b e e n pointed out in this chapter, d a r d references (i.e., U.S. Army Corps of Engineers,
channels in alluvial material are self-adjusting as they 1982).
change characteristics in response to changes in the A resistance relationship such as the E n g e l u n d curves
environment. Adjustments in width, flow depth, slope, in Fig. 10.12 is n e e d e d to relate channel roughness to
m e a n d e r length, sediment load, and channel form are flow r a t e . Such relationships must account for the
possible to allow a channel to adjust so that it can changes that occur in form roughness with changing
balance the sediment load. flows.
Early studies of channel change w e r e limited to Total energy gradient must account for both lateral
physical model evaluations, which are often expensive flow a n d transverse flow. Transverse flow must account
and time consuming. M o r e recently, improvements in for changes in secondary currents and transverse slopes
our understanding of fluvial processes has led to devel­ that occur in channel b e n d s . Relationships for making
o p m e n t of dynamic models of fluvial channels. Dawdy these computations are too complex to b e included
and Vanoni (1986) review the existing models, includ­ h e r e , but can be reviewed in C h a n g (1988).
ing the U.S. Army Corps of Engineers (1977), H E C - 6
model F L U V I A L - 1 1 by Chang and Hill (1976), and Sediment Load Calculations
Chang (1982). A summary of the requirements of a d e ­ A relationship must be available for predicting sedi­
quate models is given in the following section. m e n t transport capacity over a range of particle sizes.
T h e r e is a difference between dynamic models of A n u m b e r of relationships are available, as summa­
channel behavior and the regime models discussed rized in C h a p t e r 7. T h e relationship used can then be
earlier. T h e dynamic models respond to short-term combined with a b e d armoring function and a sediment
fluctuations, whereas t h e regime models are for long- continuity equation to predict actual sediment dis­
term equilibrium conditions. Thus, the models have charge rates. A sediment armoring function should
different purposes. T h e dynamic models, however, can allow for interactions of flow with the bed to d e t a c h
be used with long-term simulations to develop regime smaller particles and leave behind larger particles.
relationships. T h e s e larger particles will form an a r m o r over smaller
Dynamic Models of Channel Change 419

particles in the bed, preventing d e t a c h m e n t as flow Changes Due to Curvature Effect


rates decrease. A n example of an armoring relation­
A final correction must be m a d e for transverse sedi­
ship is that of B o r a h et al (1982).
m e n t transport, which creates a transverse channel
Sediment calculations should also account for the
slope. Available relationships correlate transverse sedi­
diffusion p h e n o m e n o n that moves suspended load into m e n t load to streamwise sediment load as a function of
the flow and prevents suspended load from coming to the bottom current in the transverse direction (related
rapid equilibrium with flow changes. to the curvature) and the transverse slope of the bed.
T h e basic sediment continuity relationship is Sediment continuity is also followed in the transverse
direction, or
dQ s

dZ 1 1 d
-(rtf)«0, (10.72)
~dt \ - P r d r
where Ρ is porosity of bed material, A is b e d surface
cb
w h e r e r is the radial coordinate at any streamwise
area, Q is b e d material discharge, a n d q is lateral
s s
position JC, a n d q' is t h e transverse sediment load.
s

inflow of sediment including that d e t a c h e d from the


bed. Equation (10.70) must b e solved numerically along
with Eq. (10.69). T h e o u t p u t n e e d e d for further calcu­ Application of the Models
lations is the change in bed area, AA over a routing
cb
T h e models that result from the relationships dis­
interval Δ ί . This can be used to d e t e r m i n e t h e change cussed above are very complex. Input p a r a m e t e r s are
in channel width and depth. not known with a high degree of certainty; particularly
bed roughness p a r a m e t e r s , b a n k and bed erodibility
Changes in Channel Width and Depth
p a r a m e t e r s , and sediment transport relationships. Suc­
Because channel width and d e p t h are not con­ cessful application of the model to a specific situation
strained, it is necessary t o select rules for computations requires, at least some field d a t a for calibration runs
based on additional principles. For example, it is possi­ specific t o t h e location. Examples of successful applica­
ble to have simultaneous erosion or deposition in banks tions of such models are included in Simons, Li, and
and stream beds, or erosion in one and deposition in Associates (1982) and C h a n g (1988).
the other. In the F L U V I A L model (Chang, 1982), for
example, the direction of width adjustment is such that
the channel is moving toward uniform stream power Problems
along the reach. If a channel reach is s t e e p e r t h a n (10.1) In an attempt to improve navigation in a very
adjacent sections, its width is decreased to reduce t h e sinuous channel, the public works d e p a r t m e n t in O u t e r
gradient, and vice versa. T h e rate of widening in Slobovia proposes t o eliminate a series of m e a n d e r s . If
F L U V I A L is controlled by t h e banks. If a channel is t h e stream channel is erodible, what will be the local
widening, the rate is controlled by t h e erodibility of the impacts and downstream impacts? How might these
banks. If it is decreasing in width, it decreases at a rate impacts be r e d u c e d ?
controlled by deposition along the banks. (10.2) A large forested watershed is to be devel­
O n c e the rate of widening is d e t e r m i n e d , a n d sub­ o p e d into a urban and industrial complex over a period
tracted from AA , cb the r e m a i n d e r can b e used to of 10 years, leaving m o r e t h a n 50 acres disturbed in any
determine bed profile adjustments. For example, in t h e 1 year. W h a t would b e t h e probable impact on the
F L U V I A L model, C h a n g (1982) adjusts t h e b e d eleva­ main channel draining t h e watershed (a) during con­
tion by using the shear excess concept, or struction and (b) after construction is complete.
(10.3) List the differences between a rigid bound­
ary and moveable boundary (fluvial) channel. For each
Λ 7 ( T
0 ~ T
c ) A

(10.71) type channel, indicate how t h e following p a r a m e t e r s


L(T -T ) 0 c Ay would change with increasing flow rate: depth, width,,
slope, width to d e p t h ratio, and sinuosity.
(10.4) For your local geographic region, select and
where Δ Ζ is a bed channel elevation correction, r is 0 describe in detail natural channels that are sinuous
the local tractive force, r is critical tractive force, m is
c canaliform, sinuous point bar, sinuous braided, and
an exponent, y is the horizontal coordinate, and t h e nonsinuous braided. Include in your description such
summation over Β indicates that the d e n o m i n a t o r is things as b o u n d a r y vegetation, channel slope, valley
summed over the channel width. slope, sinuosity, etc.
420 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design

(10.5) Explain to the best of your ability why flow information on the sediment load is available. During
d e p t h decreases and width increases as sediment load development, it is estimated that the average bed ma­
increases. terial concentration will be 10,000 p p m and after con­
(10.6) Describe the changes one experiences in struction the incoming b e d material load will b e n e a r
channel morphology as slope increases. W h a t causes zero. W h a t will be t h e expected channel characteristics
the p h e n o m e n o n of braiding? (width, d e p t h , slope) during and after construction?
(10.7) Describe the changes in bedform that are Explain how this will h a p p e n , giving a sequence of
observed as stream power increases from low to high events.
values. It is often observed in the transition zone that (10.16) T o allow for navigation on the Arkansas
the bedform oscillates between dunes, plane bed, and River, a series of locks and d a m s were built u p to
antidunes. Explain why this might occur. Tulsa, O k l a h o m a . W h a t changes in channel character­
(10.8) A riprap-lined channel is being p r o p o s e d istics would you expect in areas w h e r e the channel bed
with an average diameter of 0.25 ft and a of 0.5 ft. is erodible? (Hint: Utilize the slope of the energy
If the slope is 1.0% and the flow d e p t h is 2.0 ft, what is gradient instead of the slope of the channel in your
the Manning's η according to the Strickler equation analysis.)
and t h e F e d e r a l Highway Administration equation? (10.17) As a result of a series of events in the 1930s,
W h a t would be the average velocity a n d discharge per most of t h e cotton fields in the E a s t e r n Piedmont areas
unit width (assume a wide channel)? of South Carolina were converted from agricultural
(10.9) T h e flow in a 10-ft-wide rectangular channel lands to forest with a major decrease in sediment yield
with a slope of 0 . 1 % is 50 cfs. If the average roughness and runoff rates. W h a t impact would that have on
of the channel, k , is 0.01 ft, what is t h e d e p t h of flow
s
stream channel characteristics in t h e a r e a ? Justify your
according to Einstein's equation? (Hint: Replace D in answer.
Eq. (10.16) with hydraulic radius.) (10.18) A gravel channel has a d^ of 0.3 ft, a slope
(10.10) A 40-ft wide channel has a maximum d e p t h of 0.01 -ft/ft, and a width of 30 ft. Estimate the
of 5.0 ft before overtopping t h e banks. T h e channel s t a g e - d i s c h a r g e curve u p to a relative submergence of
slope is 0.0004 f t / f t and the bed material has the 5 using Bathurst's relationship. Use d e p t h s of 0.3, 0.5,
following characteristics: d = 0.12 mm, d = 0.25
35 50
0.75, 1.0 ft, etc.
mm, d = 0.65 mm. Develop a s t a g e - d i s c h a r g e curve
65 (10.19) A recreational development in Appalachia is
using (a) the E i n s t e i n - B a r b a r o s s a m e t h o d and (b) t h e expected to change t h e annual flow in a stream chan­
Englelund method. Assume that t h e water t e m p e r a t u r e nel from 150 to 400 cfs. Prior to development, the
is 20°C and that the channel is approximately rectangu­ surface width in the annual storm is estimated to be
lar. 30 ft with a slope of 0.03 f t / f t . W h a t would be the
(10.11) For Problem (10.10), what will b e the bed- expected change in channel characteristics as a result
form at a d e p t h of 3.0 ft? of t h e development? T h e average d i a m e t e r of the
(10.12) Explain the difference between a regime re­ channel material is 300 mm.
lationship for a steady flow channel and a channel
conveying naturally varying runoff.
References
(10.13) A flow of 75 cfs is being diverted from a river
with an average suspended load of 10,000 p p m of Athaullah, M. (1968). Prediction of bed forms in erodible channels,
sediment. Define a stable channel width, d e p t h , and Ph.D. thesis. Department of Civil Engineering, Colorado State
University, Ft. Collins, C O .
slope by the Blench, Simons and Albertson, and C h a n g
Bathurst, J. C. (1985). Flow resistance estimation in mountain rivers.
methods. T h e actual groundslope is 0.001 f t / f t . W h a t J. Hydraul. Eng. 111(4): 6 2 5 - 6 4 3 .
would you r e c o m m e n d ? Blench, T. (1970). Regime theory design of canals with sand beds.
(10.14) Why would the bankfull discharge be classed Proc. Am. Soc. Civil Eng. 96(IR2): 2 0 5 - 2 1 3 .
as the channel-forming event? Why would the a n n u a l Borah, D . K., Alonso, C. V., and Prasad, S. N. (1982). Routing
graded sediments in streams: Formations. Proc. Am. Soc. Civil
storm (2-year event) b e approximately equal to the
Eng. 102(HY12): 1486-1503.
bankfull discharge? Brice, J. C. (1983). Planform properties of meandering rivers. River
(10.15) A s a result of urbanization, the 1.5-year Meandering. In "Proceedings, 1983 Rivers Conference, Ameri­
storm is expected to change from 400 to 800 cfs in t h e can Society of Civil Engineers, N e w York."
drainage from a watershed. During construction, the Chang, Η. H. (1979a). Geometry of rivers in regime. Proc. Am. Soc.
Civil Eng. 105(HY6): 6 9 1 - 7 0 6 .
1.5-year storm is expected to be 600 cfs. Prior to
Chang, Η. H. (1979b). Minimum stream power and river channel
development, the channel draining the watershed has a patterns. / . Hydrol. 41: 3 0 3 - 3 2 7 .
bed material with an average diameter of 0.4 mm, a Chang, Η. H. (1980a). Geometry of gravel streams. Proc Am. Soc.
slope of 0.0002 ft/ft, and a surface width of 45 ft. N o Civil Eng. 106(HY9): 1443-1456.
Dynamic Models of Channel Change 421

Chang, Η. H. (1980b). Stable alluvial canal design. Proc. Am. Soc. Processes in Geomorphology." Freeman, San Francisco, CA.
Civil Eng. 106(HY5): 8 7 3 - 8 9 1 . Mackin, J. H. (1948). Concept of the graded river. Geol. Soc. Am.
Chang, Η. H. (1982). Mathematical model for erodible channels. Bull. 59: 4 6 3 - 5 1 2 .
Proc. Am. Soc. Civil Eng. 108(HY5): 6 7 8 - 6 8 9 . Meyer-Peter, E., and Muller, R. (1948). Formulas for bed-load
Chang, Η. H. (1984). Modeling of river channel changes. Proc. Am. transport. In "Proceedings, 2nd Meeting International Associa­
Soc. Civil Eng. 110(HY2): 157-172. tion Hydraulic Research," Paper N o . 2; pp. 3 9 - 6 4 .
Chang, Η. H. (1985a). River morphology and thresholds. / . Hydraul. Ming, Z. F. (1983). Hydraulic geometry of alluvial channels.
Eng. 11KHY3): 5 0 3 - 5 1 9 . /. Sediment Res. 4: 7 5 - 8 4 .
Chang, Η. H. (1985b). Design of stable alluvial canals in a system. / . Parker, G. (1979). Hydraulic geometry of active gravel rivers. Proc.
Irrigat. Drain. Eng. 111(1): 3 6 - 4 3 . Am. Soc. Civil Eng. 105(HY9): 1185-1201.
Chang, Η. H. (1986). River channel changes: Adjustments of equilib­ Richards, K. (1982). "Rivers, From and Process in Alluvial Channels."
rium. / . Hydraul. Eng. 112(1): 4 3 - 5 5 . Methuen, New York.
Chang, Η. H. (1988). "Fluvial Processes in River Engineering." Schumm, S. A. (1960). The effect of sediment type on the shape and
Wiley-Interscience, New York. stratification of some modern fluvial deposits. Am. J. Sci. 258:
Chang, Η. H., and Hill, J. C. (1976). Computer modeling of erodible 177-184.
flood channels and deltas. Proc. Am. Soc. Civil Eng. 102(HY10): Schumm, S. A. (1968). River adjustment to cultured hydrologic
1461-1477. regimes, Murrumbridgee River and Palco Channels, Australia,
Culbertson, D . M., Young, L. E., and Brice, J. C. (1967). Scour and U S G S professional paper N o . 598.
fill in alluvial channels, U.S. Geological Survey, Open File Re­ Schumm, S. A. (1969). River metamorphosis. Proc. Am. Soc. Civil
port. [Quoted by Simons, Li, and Associates, 1982; original not Eng. 95(HY1): 2 5 5 - 2 7 3 .
seen]. Schumm, S. A. (1971a). Fluvial geomorphology, historical perspec­
Davis, W. M. (1899). The geographical cycle. Geogr. J. 14: 4 8 1 - 5 0 4 . tive, In "River Mechanics" (Η. H. Shen, ed.). P.O. Box 606, Ft.
Collins, CO.
Dawdy, D . R., and Vanoni, V. A. (1986). Modeling alluvial channels.
Schumm, S. A. (1971b). Channel adjustment and river metamorpho­
Water Resources Res. 22(9): 7 1 S - 8 1 S .
sis, In "River Mechanics" (Η. H. Shen, ed.). P.O. Box 606, Ft.
Dury, G. H. (1973). Magnitude-frequency analysis and channel mor­
Collins, CO.
phology. In "Fluvial Geomorphology" (Morisawa, ed.), Publica­
Schumm, S. A. (1977). "The Fluvial System." Wiley, New York.
tions in Geomorphology, pp. 9 1 - 1 2 1 . S U N Y , N e w York.
Schumm, S. Α., and Khan, H. R. (1972). Experimental study of
Einstein, H. A. (1950). The bed load function for sediment trans­
channel patterns. Geol. Soc. Am. Bull. 83: 1755-1770.
portation in open channels, Technical bulletin 1026. Soil Conser­
Simons, D . B., and Albertson, M. L. (1960). Uniform water con­
vation Service, U.S. Department of Agriculture, Washington, D C .
veyance channels in alluvial material. Proc. Am. Soc. Civil Eng.
Einstein, Η. Α., and Barbarossa, N. (1952). River channel roughness.
86(HY5): 3 3 - 7 1 .
Trans. Am. Soc. Civil Eng. 117: 1121-1146.
Simons, D . B., and Richardson, Ε. V. (1961). Forms of bed rough­
Engelund, F. (1966). Hydraulic resistance of alluvial streams. Proc.
ness in alluvial channels. / . Hydraul. Div. Am. Soc. Civil Eng.
Am. Soc. Civil Eng. 92(HY2): 3 1 5 - 3 2 6 .
88(HY3): 8 7 - 1 0 5 .
Engelund, F., and Hansen, E. (1967). "A Monograph on Sediment
Simons, D . B., and Richardson, Ε. V. (1966). Resistance to flow in
Transport in Alluvial Channels." Teknisk Vorlag, Copenhagen,
alluvial channels, U S G S professional paper 4-22-J.
Denmark.
Simons, D . B., Richardson, Ε. V., and Nordin, C. F. (1965). Sedi­
Federal Highway Administration (1975). Design of stable channels
mentary structures generated by flow in alluvial channels, Special
with flexible liners, Hydraulic engineering circular N o . 15. U.S.
publication N o . 12. American Association of Petroleum Geolo­
Department of Transportation, Washington, D C . gists.
Foster, G. R., and Lane, L. J. (1983). Erosion by concentrated flow in
Simons, D . B., Lagasse, P. F., Chen, Υ. H., and Schumm, S. A.
farm fields. In "Proceedings, D . B. Simons Symposium on Ero­ (1975). The river environment—A reference document. U.S. D e ­
sion and Sedimentation, Colorado State University, Ft. Collins, partment of Interior, Fish and Wildlife Service, Twin Cities, MN.
CO," pp. 9.65-9.82. Simons, Li and Associates (1982). "Engineering Analysis of Fluvial
Khan, H. R. (1971). Laboratory studies of alluvial river morphology, Systems." Simons, Li and Associates, Ft. Collins, CO.
Ph.D. dissertation. Civil Engineering Department, Colorado State Strickler, A. (1923). Some contributions to the problem of the
University, Ft. Collins, CO. velocity formula and roughness factors for rivers, canals and
Lacey, G. (1930). Stable channels in alluvium. Proc. Inst. Civil Eng. closed conduits, N o . 16. Mitteilungen des eidgenossicschen Amtes
229: 2 5 9 - 3 8 4 . fur Wasserwirtschaft, Bern, Switzerland.
Lane, E. W. (1955). The importance of fluvial geomorphology in U.S. Army Corps of Engineers (1977). HEC-6, scour and deposition
hydraulic engineering. Proc. Am. Soc. Civil Eng. 81: 1-17. in rivers and reservoirs, users manual. Hydrologic Engineering
Lane, E. W. (1957). A study of shape of channels formed by natural Center, Davis, CA.
streams flowing in erodible material, M. R. D . sediment series U.S. Army Corps of Engineers (1982). HEC-2, water surface profiles,
No. 9. U.S. Army Engineering Division, Missouri River, U S Army users manual. Hydrologic Engineering Center, Davis, CA.
Corps of Engineers, Omaha, N E . van Rijn, L.C. (1984). Sediment transport. III. Bed forms and alluvial
Leopold, L. B., and Maddock, T. (1953). The hydraulic geometry of roughness. / . Hydraul. Eng. Am. Soc. Civil Eng. 110(12):
stream channels and some physiographic implications, U S G S 1733-1754.
professional paper N o . 252. Williams, G. P. (1978). Bankfull discharge of rivers. Water Resources
Leopold, L. B., and Wolman, M. G. (1957). River channel patterns: Res. 14(6): 1141-1154.
Braided, meandering and straight, U S G S professional paper Wolman, M. G. (1955). The natural channel of Brandywine Creek,
282-B; pp. 4 5 - 6 2 . Pennsylvania, U . S . Geological Survey professional paper 271.
Leopold, L. B., Wolman, M. G., and Miller, J. P. (1964). "Fluvial U.S. Geological Survey, Reston, V A .
11
Ground Water

INTRODUCTION water by volume, but the water may b e held so tightly


by the clay that it cannot b e removed effectively by
Most of the fresh water in the continental United pumping.
States is ground water. G r o u n d water refers to all of T h e exact d e p t h at which the saturated zone occurs
the water b e n e a t h the earth's surface. G r o u n d water varies by location (spatially) and through time (tempor­
may be found in zones that are either saturated or ally). T h e d e p t h to the saturated zone can vary from
unsaturated. Traditionally, the water of interest to zero at the soil surface to h u n d r e d s , even thousands, of
hydrologists is contained in saturated layers similar to feet. In some areas, it is almost impossible to locate the
those shown in Figs. 11.1 and 11.2. T h e unsaturated saturated zone because of rock formations or the ex­
zone is also known as the vadose zone or the phreatic tremely large d e p t h at which ground water is located.
zone. G r o u n d water in the saturated zone is consid­ Typically, t h e r e is a region above t h e ground water that
ered to b e water contained in underground formations consists of soil and other unconsolidated materials
in a saturated or near-saturated condition and u n d e r a containing some water, but the larger pore spaces are
pressure greater than atmospheric. W a t e r stored u n d e r not filled with water. T h e water in this zone is typically
pressures less than atmospheric such as water in the not available for pumping because it is generally held
vadose zone is sometimes thought of as u n d e r g r o u n d u n d e r tension. However, if sufficient water is available,
water, not as ground water. it will move through this u n s a t u r a t e d zone by gravity
G r o u n d water in saturated zones is located in drainage and b e c o m e part of the saturated zone (Fig.
aquifers. Aquifers are water-bearing formations that 11.2). T h u s , the u n s a t u r a t e d zone is important because
are saturated (pore spaces are filled with water). They it is a zone of transmittal to the saturated zone. As
are generally considered to be those soil and rock such, it becomes a source or sink for any chemicals
formations that have enough hydraulic conductivity to associated with it as it moves toward the saturated
allow water to be p u m p e d out through wells at useable zone.
rates. T h e Soil Conservation Service (1984) indicated Although ground water is conventionally classified as
that although there is water under most of the earth's in either the saturated zone or vadose zone, this divi­
surface, it may not be practical to utilize because it is sion of waters is somewhat artificial. T h e r e is no dif­
too difficult to p u m p or is held too tightly. For exam­ ference in physical or chemical properties of the water
ple, a heavy clay may contain a high percentage of or in the basic physical laws governing the movement

422
Introduction 423

Figure 11.1 Ground water's role in the hydrologic cycle (after Soil Conservation Service,
1984).

of water in these two zones. W a t e r is transferred freely e m e r g e as baseflow to b e c o m e surface water has a
between the zones. U n d e r unconfined (water table) major impact on t h e quality of that water. A n indica­
conditions, a saturated zone occupied by ground w a t e r tion of t h e impact of geologic factors on quality of
may become a part of the vadose zone as t h e w a t e r g r o u n d w a t e r is provided in T a b l e 11.1.
table drops. A t a later time, as the w a t e r table rises, T h e interchange b e t w e e n g r o u n d water, vadose wa­
the zone may again b e c o m e a p a r t of t h e s a t u r a t e d ter, and surface w a t e r points to t h e n e e d to consider
water zone. Virtually t h e same water may b e in t h e the entire hydrologic system in assessing t h e impact of
saturated zone at some time, vadose zone at a later alterations within a catchment on any aspect of the
time, and in the saturated zone once again at a still hydrology of that catchment.
later time. G r o u n d water is of interest to hydrologists and engi­
G r o u n d water becomes surface water w h e n it n e e r s for several reasons. T h e first arises because much
emerges as a spring or seep or w h e n it directly e n t e r s a of t h e drinking w a t e r used by h u m a n s and domesti­
stream, pond, or lake. G r o u n d water discharge t o cated animals comes from ground water. T h u s , t h e r e is
streams, springs, and seeps generally forms t h e base considerable interest in any activity that has potential
flow for small streams b e t w e e n major runoff-producing for impacting either t h e quantity or quality of ground
events. T h e pathways taken by water as it infiltrates water. Particular concern about the possibility of chem­
and percolates to become ground water and then to icals entering g r o u n d water because of the extreme
difficulty of removing contaminants from ground water
zones has b e e n expressed. Recently, considerable at­
tention has b e e n directed toward determining potential
for certain activities to cause ground w a t e r contamina­
tion ( C A S T , 1985; National R e s e a r c h Council, 1984).
A second interest, primarily to engineers, occurs when
construction, mining, or o t h e r activities involve ground
w a t e r aquifers or recharge areas or w h e r e ground
w a t e r restricts these activities (National Research
Council, 1990a, b). Additionally, ground water provides
a large portion of t h e flow to streams. G r o u n d water,
like almost all of t h e waters of t h e earth, is part of the
hydrologic cycle as described in C h a p t e r 3 a n d illus­
t r a t e d in Fig. 11.1.
Figure 11.2 Saturated and unsaturated zones in an unconfined T h e speed with which g r o u n d water moves from the
aquifer (after Heath, 1982). surface to o t h e r aspects of t h e hydrologic cycle can
424 11. Groundwater
Table 11.1 Naturally Occurring Inorganic Chemicals that Pollute
Ground Water (Adapted from Heath, 1982)

Substance Major natural sources

Bicarbonate and carbonate Solution of carbonate rocks,


limestone, and dolomite by water
Calcium and magnesium Rocks containing limestone,
dolomite, or gypsum
Chloride Either geologic seawater or
seawater in contact with aquifers
Fluoride Sedimentary or igneous rocks
Iron and manganese Iron present in most soils or
rocks, manganese less widespread
Sodium Similar to chloride
Sulfate Gypsum, pyrite, or other rocks
containing sulfur

vary greatly. In o n e location, water may move rapidly


through the hydrologic cycle as it goes from precipita­
tion to percolation into a rock fracture or a highly
structured soil. T h e n it may quickly flow to a spring or
stream and into an ocean. A t a n o t h e r location, water
may infiltrate and percolate into a slowly p e r m e a b l e
formation and require many years before it emerges
and reenters the more active parts of the cycle through
surface water flow or evapotranspiration.
Surface mining, construction, and similar land-dis­
turbing activities may alter many of the hydrologic
processes related to ground water (National Research Figure 11.3 Geographic delineation of ground water provinces
Council, 1990a). Infiltration, overland flow, surface (after Soil Conservation Service, 1984).
runoff, surface storage and detention, interception,
evapotranspiration, percolation, ground water flow, and
stream flow can all be impacted. As a m a t t e r of fact,
the only part of the cycle not generally considered to
phabetically. T h e Soil Conservation Service (1984) p r e ­
be potentially impacted by h u m a n activities is precipi­
sents a brief description of each province as follows.
tation, and even that assumption is currently u n d e r
serious challenge.
T h e flow and storage processes shown in Fig. 11.1 A. Atlantic and Gulf Coastal.Plain Province
are u n s t e a d y — t h a t is, the flow rate and volume of W a t e r is derived in r a t h e r large quantities from
water in any particular form of storage are constantly sands and gravels interbedded with clay. Large supplies
changing with time. T h e time rate of change of some are obtained from alluvial gravels in the Mississippi
processes such as precipitation and surface runoff may Valley and adjacent areas. T h e province includes ex­
b e rapid, while the rate of change of ground water tensive areas of artesian flow. In mineral content, the
storage and discharge may be very slow. ground w a t e r varies from low to high.

LOCATION OF GROUND WATER PROVINCES B. Northeastern Drift Province


G r o u n d water comes principally from glacial drift.
T h e U.S. Geological Survey identified significant T h e till yields small supplies to many springs and
ground water provinces in the United States based on shallow wells (less than 60 m); the outwash gravels
the area of coverage of water-bearing formations. Fig­ yield large supplies. Many drilled rock wells receive
ure 11.3 delineates these ground water provinces al- small supplies, chiefly from joints in crystalline rocks or
Location of Ground Water Provinces 425

in Triassic sandstone. G r o u n d water is generally soft only meager supplies, and as a rule t h e r e are no
and low in mineral content. intervening Paleozoic, Mesozoic, or Tertiary forma­
tions thick e n o u g h to yield much water. T h e drift and
rock waters range from soft waters in Wisconsin to
C. Piedmont Province highly mineralized waters in t h e western and north­
W a t e r generally low in mineral content is supplied in western parts of t h e province.
small quantities by the crystalline rocks and locally by
Triassic sandstone. Many shallow dug wells are sup­
plied from surface deposits or from the u p p e r decom­ I. Dakota Drift-Cretaceous Province
posed part of the bedrock. Many moderately d e e p T h e two important sources of ground water are the
(60-300 m) drilled wells are supplied from joints in the glacial drift and t h e D a k o t a sandstone. T h e drift sup­
crystalline rocks. Some wells in Triassic sandstone yield plies n u m e r o u s wells with h a r d b u t otherwise good
large supplies. water. T h e D a k o t a sandstone has extensive areas of
artesian flow t h a t supply many strong flowing wells, a
D. Blue Ridge-Appalachian Valley Province considerable n u m b e r of which are m o r e than 300 m
d e e p . T h e D a k o t a sandstone waters are highly mineral­
This is a region of rugged topography with n u m e r o u s ized but are used for domestic supplies.
springs that generally yield good-quality water from
Paleozoic strata, pre-Cambrian crystalline rocks, or
post-Cambrian intrusive rocks. T h e water is derived J. Black Hills Cretaceous Province
chiefly from springs, spring-fed streams, and shallow
wells. T h e conditions in this province are, on the whole,
unfavorable for shallow water supplies, because most
of the province is underlain by the Pierre shale or by
E. Southcentral Paleozoic Province shales of the White River group. T h e principal aquifer
T h e principal water sources are the Paleozoic sand­ is the D a k o t a sandstone, which underlies the entire
stones and limestones. In many of the valleys, large region except the Black Hills. This sandstone will prob­
supplies are obtained from alluvial sands and gravels. ably yield water wherever it occurs, and over consider­
able parts of the province it will give rise to flowing
wells; however, throughout much of the province it is
F. Northcentral Drift-Paleozoic Province far below the surface. In the Black Hills, water is
Most water is derived from glacial drift, w h e r e it is obtained from a variety of sources, ranging from pre-
generally hard but otherwise good. N u m e r o u s drilled Cambrian crystalline rocks to Cretaceous or Tertiary
wells produce large supplies from glacial outwash or sedimentary rocks.
from gravel interbedded with till. Many drilled wells
end in Paleozoic sandstone or limestone and receive
ample water. K. Great Plains Pliocene-Cretaceous Province
T h e principal aquifers of this province are the late
Tertiary sands and gravels (Ogahalla and related for­
G. Wisconsin Paleozoic Province
mations) and the D a k o t a sandstone. T h e Tertiary de­
Most of the water is from wells of m o d e r a t e d e p t h posits underlying t h e extensive smooth and u n e r o d e d
drilled into Cambrian or Ordovician sandstone or lime­ plains supply large quantities of water to shallow wells.
stone. These wells as a rule yield ample supplies of T h e D a k o t a sandstone underlies nearly the entire
hard but otherwise good water. In many of the valleys, province and gives rise to various areas of artesian
artesian flows are obtained from the Paleozoic aquifers. flow. T h r o u g h o u t much of the province, however, it lies
T h e region has no water-bearing drift except in the too far below the surface to be a practical source of
valleys, where there are water-bearing outwash gravels. water. W h e r e the Tertiary beds are absent or badly
e r o d e d and the D a k o t a sandstone is buried b e n e a t h
thick beds of shale, as in parts of eastern Colorado,
H. Superior Drift-Crystalline Province
developing even small water supplies may b e difficult.
In most parts of this province, satisfactory water Many of the valleys contain Q u a t e r n a r y gravels, how­
supplies are obtained from glacial drift. W h e r e the ever, which supply large quantities of good water.
drift is thin, water is generally scarce, because the Considerable Tertiary and Q u a t e r n a r y sections can
pre-Cambrian crystalline rocks in most places yield yield a supply suitable for irrigation.
426 11. Groundwater

L. Great Plains Pliocene-Paleozoic Province yields only meager a m o u n t s of generally poor-quality


T h e principal aquifers of this province are t h e late water or n o n e at all. H e n c e , locally, w h e r e the Fort
Tertiary and Quaternary sands and gravels, which give Union and Lance are absent or d o not yield enough,
the same favorable conditions as those in the G r e a t satisfactory water supplies are very difficult to obtain.
Plains P l i o c e n e - C r e t a c e o u s province. T h e Tertiary de­ T h e n o r t h e r n part of the province has a little water­
posits are underlain practically throughout the province bearing glacial drift.
by Permian or Triassic deposits, which in most places
yield little or only highly mineralized water. W h e r e the
P. Southern Rocky Mountain Province
Tertiary deposits are thin or absent, or w h e r e they
have b e e n eroded, the ground water conditions are In this m o u n t a i n province, underlain mostly by crys­
generally unfavorable. talline rocks, water is obtained chiefly from springs,
from streams fed by springs and melted snow, or from
very shallow wells near streams.
M. Trans-Pecos Paleozoic Province
T h e bedrock consists of Carboniferous, Permian, Q. Montana-Arizona Plateau Province
and Triassic strata, including limestone, gypsum, red
This large a r e a is mostly an arid to semiarid plateau
beds of shale and shaly sandstone, and some less shaly
underlain by sedimentary formations ranging in age
sandstone. In most of the province these rocks yield
from Paleozoic to Tertiary. T h e formations are not
only meager supplies of highly mineralized waters to
violently deformed, but they are warped and broken
d e e p wells. In t h e Pecos Valley, however, Carbonifer­
enough that the presence of ground water is closely
ous limestones and sandstones yield large supplies to
related to rock structure, and conditions vary over
n u m e r o u s flowing wells; the water is very h a r d but
short distances. O n the whole, water is neither plenti­
good enough for irrigation, domestic, and livestock
ful nor of very satisfactory quality. W h e r e thick forma­
purposes. Locally the bedrock is overlain by Q u a t e r ­
tions of nearly impervious material are at the surface,
nary water-bearing gravels.
or w h e r e the plateau is greatly dissected, as in the
G r a n d Canyon region, water is scarce. Locally, how­
ever, sandstone aquifers can b e developed and may
N. Northwestern Drift-Eocene-Cretaceous Province
yield very satisfactory supplies—in some places giving
G r o u n d water is obtained from glacial drift and from rise to flowing wells. T h e r e are also local deposits of
underlying E o c e n e and U p p e r Cretaceous formations. water-bearing Q u a t e r n a r y gravels.
W h e r e the drift is absent or not water bearing, wells
are sunk into the underlying formations, with variable
success. T h e E o c e n e and latest Cretaceous, which un­ R. Northern Rocky Mountain Province
derlie most of the eastern part of the province, gener­ This is a relatively cold region, chiefly mountainous
ally include water-yielding strata or lenses of sand, b u t with extensive valleys and plains. It is underlain by
gravel, or coal. T h e Cretaceous formations in the west­ a wide variety of rocks with complicated and diverse
ern part consist chiefly of alternating beds of shale and structure. A s in o t h e r m o u n t a i n regions, water is ob­
sandstone. T h e sandstones generally yield water, but tained largely from m o u n t a i n springs and streams.
the shales are unproductive, and w h e r e a thick shale Considerable w a t e r is available in places m a d e u p of
formation immediately underlies t h e drift or is at the ordinary alluvial sand and gravel and the outwash
surface, successful wells may b e difficult to obtain. In deposits of m o u n t a i n glaciers. W a t e r is also obtained
certain localities, upland gravels yield water to shallow from wells drilled into various pre-Cambrian a n d Ter­
wells. tiary rock formations.

O. Montana Eocene-Cretaceous Province S. Columbia Plateau Lava Province


E n o u g h fairly good water for domestic and livestock T h e principal aquifers of this province are the
supplies and even for small municipal supplies is ob­ widespread Tertiary a n d Q u a t e r n a r y lava beds and
tained from strata and lenses of sand, gravel, and coal interbedded or associated Tertiary sand and gravel. In
in the Fort Union (Eocene) and Lance (late Creta­ general, t h e lava yields a b u n d a n t supplies of good
ceous or E o c e n e ) formations that underlie most of this water. It gives rise to many large springs, especially
province. T h e s e formations usually rest on t h e Pierre along t h e Snake River in I d a h o . Locally, the lava or the
shale, a thick, dense U p p e r Cretaceous shale that i n t e r b e d d e d sand a n d gravel give rise to flowing wells.
Basic Concepts of Ground Water Hydraulics 427

However, much of the lava is so p e r m e a b l e a n d t h e w a t e r have delayed investigation a n d exploitation of


relief of t h e region is so great that in many places the t h e extensive ground water resources of t h e area.
water table can be reached only by d e e p wells. In
certain parts of the province, glacial outwash a n d ordi­ X. N o r t h e r n Coast Range Province
nary valley fill are also important water sources. G r o u n d water is found in the alluvial fill of the
valleys draining to the Pacific O c e a n . A small area in
t h e s o u t h e r n p a r t of t h e province contains h e a t e d
T. Southwestern Bolson Province
g r o u n d water, hot springs, a n d geysers. Because sur­
T h e principal source of water in this arid province is face w a t e r is a b u n d a n t and t h e province is relatively
the alluvial sand and gravel of the valley fill underlying undeveloped, little detailed information o n ground wa­
the n u m e r o u s intermountain valleys. In the elevated ter conditions is available.
marginal parts of the valleys, the water table may b e
far below the surface or ground water may be absent;
in the lowest parts, underlain by clayey a n d alkaline BASIC CONCEPTS OF GROUND WATER
beds, ground water may b e scarce and of poor quality; HYDRAULICS
at intermediate levels, however, large supplies of
good-quality water are generally found. Most of t h e Conservation of Mass
water in the valleys of this province is recovered by
means of pumping wells, but t h e r e are many springs T h e principle referred to as conservation of mass
and areas of artesian flow. In m o u n t a i n areas of t h e indicates that over some time interval, the difference in
province, many springs, small streams, a n d shallow t h e volume of w a t e r entering and leaving a control
wells furnish valuable supplies. e l e m e n t must equal t h e change in volume of water
stored in t h e element. Mathematically this is expressed
as
U. Central Valley of California Province
Good-quality ground water is found chiefly in allu­ /Δί - ΟΔί = Δ5, (11.1)
vial cones formed by streams emerging from t h e Sierra
Nevada, although water can b e obtained t h r o u g h o u t w h e r e / Δ ί is the inflow volume in time Δ ί , Ο Δ ί is the
the valley. T h e yield of cones flanking the Coast R a n g e outflow volume in time Δ ί , and Δ 5 is t h e change in
is small. Poor-quality water generally comes from t h e storage. A s a r a t e relationship, t h e continuity of mass
south and central sections, and somewhat b e t t e r qual­ is given by
ity from the north. Underlying p i e d m o n t deposits con­
I - 0 = dS/dt, (11.2)
sist of marine, lacustrine, a n d alluvial formations.
Highly mineralized water is found in d e e p strata
w h e r e / and Ο a r e t h e inflow a n d outflow rates (volume
throughout the valley and n e a r the ground surface in
p e r unit or time) a n d dS/dt is t h e time r a t e of change
the center of the valley. Extensive irrigation in the
in storage within t h e control element.
valley d e p e n d s on ground water p u m p e d from wells.
Over short intervals of time (up to a few years), the
change in inflow, outflow, or the volume of water
V. Coastal Ranges of C e n t r a l a n d Southern stored in t h e control element may b e substantial. In an
California Province undisturbed control element, in t h e absence of some
T h e principal ground water bodies are in t h e m o u n ­ natural, cataclysmic event, and for long time intervals
tain valley and piedmont plains draining to t h e Pacific (several years), t h e change in t h e average value of
Ocean. Aquifers consist of valley fill and alluvial sand these quantities is relatively small, a n d t h e inflow and
and gravel deposits. Locally, good water supplies are outflow volumes are nearly equal. U n d e r these condi­
developed from underlying younger Tertiary sand­ tions, a state of dynamic equilibrium exists.
stones. Heavy development of ground water along t h e T h e relationship between t h e inflow, outflow, and
coast for municipal and irrigation n e e d s has caused sea c h a n g e in storage for any control element defines the
water to e n t e r and contaminate aquifers in several hydrologic balance of t h e element. Any factors that
valley mouths. alter inflows, outflows, or storage characteristics poten­
tially alter t h e hydrologic balance and thus ground
water. Such factors include impacts of construction,
W. Willamette Valley-Puget Sound Province
mining, a n d o t h e r land-disturbing activities.
A large body of alluvium fills the structural trough T h e size of t h e control e l e m e n t must also b e consid­
forming this province. A b u n d a n t supplies of surface e r e d in evaluating changes to t h e hydrologic balance.
428 11. Groundwater
A small control element (a few acres in areal extent) by chemical reactions to form solid masses. They gen­
may have its hydrologic balance altered by minor dis­ erally consist of sedimentary and igneous rocks. Rocks
turbances within the, element. M o r e substantial alter­ of importance to ground water are limestone, dolomite,
ations would be required to alter the hydrologic bal­ shale, siltstone, sandstone, conglomerate, granite, and
ance on large control elements that contain many acres. basalt.
Hydrologic impacts of changes within a control ele­ Void spaces may consist of primary voids that were
m e n t may be transient. A disturbance upsets t h e state formed at the same time as the rock or secondary voids
of dynamic equilibrium existing within the control ele­ formed by fracturing of consolidated materials and
ment. A substantial time period, often several years, chemical action-forming solution channels.
may have to pass for a new state of dynamic equilib­ O n e characteristic of an aquifer is porosity, n, which
rium to come into existence. During this transient is defined as the fraction (or percentage if preferred) of
period, the hydrologic balance is certainly altered. Af­ volume that is occupied by voids divided by the total
ter the transient period, the new state of dynamic volume of the media so that
equilibrium must be compared to the original state to
determine if p e r m a n e n t changes to the hydrologic bal­
ance have occurred.
A control element that encompasses only water in
the saturated zone may be defined. T h e inflow to this w h e r e V , K , and K are the total volume, solid volume,
t s v

element could be flow from outside the element, perco­ and void volume, respectively. Materials with high
lation, and infiltration. Outflow from the element could porosity contain considerable water w h e n saturated.
be saturated zone flow, flow becoming surface water Porosity provides an u p p e r limit that represents the
via springs, seeps or direct discharge to surface water volume of water contained in a material when satu­
bodies, unsaturated flow to the vadose zone generally rated. All of the water contained in a formation will
from a water table, and transpiration via plants draw­ not drain solely d u e to gravity either because the pores
ing water directly from the saturated zone. are n o t connected or because some water is held too
tightly to t h e individual particles. T h e a m o u n t of water
Occurrence and Movement of Ground Water that will drain from a saturated material d u e to gravity
is referred to as t h e specific yield, while the a m o u n t of
T h e material that constitutes the earth's o u t e r m a n ­ water retained is known as the specific retention. F o r a
tle is composed of solid material and void spaces. T h e confined aquifer, water yield is defined as the volume
solid material may be in the form of individual parti­ of water an aquifer takes in or discharges p e r unit
cles or more massive rock formations. T h e void spaces surface area p e r unit change in head normal to the
are occupied by air or water. In the saturated zone, of surface. This yield is t e r m e d the storage coefficient.
course, the voids are filled with water (some e n t r a p p e d T h e sum of the specific yield and the specific reten­
air may b e present). tion is the porosity of t h e material as illustrated m a t h e ­
Water-bearing formations may b e either consoli­ matically by
dated or unconsolidated. Except for rock outcroppings,
the earth's surface is covered by a layer of unconsoli­ * = S + S ,
y r (11.4)
dated material that may range in thickness from a few
centimeters to several thousand meters. Consolidated w h e r e 5 is specific yield and S is specific retention.
y r

material always underlies the unconsolidated material T a b l e 11.2 contains typical values for porosity, specific
at some depth. Alternating strata of consolidated and yield, and specific retention ( H e a t h , 1982) for various
unconsolidated material may exist above t h e final con­ materials. A s shown, fine-grained material tends to
solidated rock. have higher porosity. Although clay has high porosity
Unconsolidated material consists of individual min­ because of the small size of individual particles, it has a
eral particles derived from the breakdown of consoli­ very low specific yield. This property is undesirable for
dated rock. Individual particles may range from clay- an aquifer used as a water supply. Coarse materials,
sized particles measuring fractions of a millimeter in sands, and gravels form most of the highly productive
diameter to rocks and boulders measuring several m e ­ aquifers and can yield u p to 8 0 % of their water as
ters across. Coarser materials, such as sand and gravel shown from the porosity a n d specific yield percentages
deposits, often m a k e excellent aquifers because they shown in Table 11.2.
allow water to move with relative ease. Estimates of ground water flow velocities for several
Consolidated material consists of mineral particles aquifer classes and selected hydraulic gradients are
that have b e e n fused together by heat and pressure or provided in Table 11.3. ( D u n n e and Leopold, 1978).
Basic Concepts of Ground Water Hydraulics 429

PERCHED WATER
Table 11.2 Selected Values of Porosity, Specific Yield, and
Specific Retention (Heath, 1982) a

Specific Specific
Material Porosity yield retention

Soil 55 40 15
Clay 50 2 48
Sand 25 22 3
Gravel 20 19 1
Limestone 20 18 2 Woter in
chemicol
Sandstone 11 6 5 combination
(semiconsolidated) with rock

Granite 0.1 0.09 0.01 Figure 11.4 Schematic showing relationships between influent and
Basalt (young) 11 8 3 effluent streams and ground water zones (after Vesilind et al., 1988).

^Values in percentage by volume.

sure are also known as artesian aquifers. A well pierc­


ing an artesian aquifer will flow without pumping if the
Table 11.3 A p p r o x i m a t e V e l o c i t i e s , V, for Different S o i l piezometric surface is above the ground surface. A
C l a s s e s and Hydraulic Gradients, ι ( A d a p t e d from D u n n e
p e r c h e d aquifer (Fig. 11.4) is an unconfined aquifer of
and L e o p o l d , 1 9 7 8 )
limited areal extent that retains w a t e r against gravity
k i V because of an underlying restricting layer such as clay.
Soil class (ft/day) (ft/ft) (ft/year) P e r c h e d aquifers may b e seasonal o r p e r m a n e n t .

Clean gravel 2.4 χ 1 0 4


0.001 8,860
0.05 443,000
Darcy's Law
Clean sand 2.4 χ 1 0 2
0.001 89 W a t e r moves within an aquifer d u e to hydraulic
0.005 4,430 gradients or water potential gradients created by grav­
Very fine sand 2.4 χ ΙΟ" 2
0.001 0.0089 ity. G r o u n d water flow may b e in any direction, even
0.05 0.044 upward. T h e actual flow p a t t e r n s may be difficult to
Unweathered 2.4 χ 10" 5
0.001 8.9 x 1 0 - 6 ascertain because of difficulties in defining under­
clay 0.05 4.4 χ 1 0 ^ ground formations. Although gravity is the dominant
force in ground water flow, surface topography often
does not match t h e water table of an unconfined aquifer
or the piezometric surface of a confined aquifer. T h e
hydraulic gradient, and not surface topography, con­
Figures 11.1 and 11.4 illustrate three types of trols ground water movement. A n example situation
aquifers—unconfined, confined, and perched. A n un- that is indicative of the complexity of ground water
confined aquifer, also known as a water table aquifer, flow problems is shown in Fig. 11.5. T h e figure shows a
has the free water surface as its u p p e r boundary. T h e p u m p i n g well and stream located in an unconfined
free water surface is also called the water table. With aquifer. T h e equipotential lines delineate points that
an unconfined aquifer, the u p p e r surface of the satu­ have t h e specified potential. T h e equipotential lines
rated zone is subject to rise or fall. A confined aquifer show drawdown occurring n e a r t h e pumping well a n d
(Fig. 11.1) has as an u p p e r boundary a relatively imper­ n e a r the stream. If the water level in the stream was
vious layer that restricts the height to which water can greater t h a n t h e water level in the aquifer, the stream
rise. T h e pressure potential of t h e water in contact would recharge the aquifer. It should also b e noted
with this confining layer is greater than atmospheric that the direction of ground water flow is perpendicu­
and defines the piezometric surface. T h e confining lar to the equipotential lines in isotropic aquifers. In
layer may be impermeable or only slightly permeable. some cases, it is desirable to draw equipotential lines
A buildup of pressure results so that if a well is drilled so that t h e same d r o p in head occurs across adjacent
into the aquifer, water will rise in a pipe to the height equipotential lines. T h e n flow lines are drawn so that
of the piezometric surface above the u p p e r boundary they form curvilinear squares with the equipotential
of the aquifer. Confined aquifers that are u n d e r pres- lines, and the flow is equal between adjacent flow lines.
430 11. Groundwater

IGNEOUS AND MET AMORPHIC ROCKS

BASALT
Unfractured Fractured
SANDSTONE
Sernlconsolldated

Unfractured
CARBONATE ROCKS
Fractured
SILT. LOESS

SILTY SAND

CLEAN SAND

GRAVEL

"I DECKER
Spots J AREA
DIRECTION OF
GROUNDWATER FLOW
CoeJ
EQUIPOTENTIAL LINES Ί OOLSTRIP
8
P° H
J AREA

10" 8
10* 7
1 0 * 10" 6
10 *
-
10" 3
10 2
10' 1
1

Hydraulic Conductivity (meters per day)

Figure 11.0 Typical values of hydraulic conductivity (adapted from


Heath (1982) and Van Voast and Reiten (1988).

SECTION A - A
u r e 11.6 shows the ranges of typical values for hy­
Figure 11.5 Schematic of example equipotential lines for pumping
well and stream. draulic conductivity. T h e equation

Q = VA (11.6)

T h e combination of equipotentials and flow lines is can b e combined with Darcy's equation to obtain a
referred to as a flow net. m e a n flow velocity for the entire cross-sectional area.
H e n r i Darcy applied a theory for water flow in This velocity is a macrovelocity. T h e actual flow veloc­
capillary tubes proposed by H a g e n and Poiseville to ity t h r o u g h t h e pores would b e m u c h larger t h a n this
ground water (Vesilind et al., 1988). (Darcy's law is velocity since t h e t r u e flow area is much smaller than
discussed in C h a p t e r 3 with respect to infiltration using t h e total cross-sectional area.
other mathematical forms of the equation.) T h e rates
and direction of flow between two points are described
by Darcy's law, or Example Problem 11.1. Velocity of water movement
in porous media
Q = -KA(dh/dL), (11.5)
Estimate the actual velocity of ground water movement
where Q is t h e rate of flow through the media, Κ is through an aquifer composed of coarse sand and through an
the hydraulic conductivity of the material, A is t h e aquitard that confines it. Assume that dh/dL equals
total cross-sectional area of the porous medial includ­ 1 m/1000 m for the aquifer and dh/dL equals 1 m / 1 0 m for
ing both pores and particles, h is t h e head, and L is the aquitard. Assume Κ = 50 m/day for the aquifer and
the length of the porous media. According to Darcy's 0.0001 m / d a y for the aquitard.
law, the rate of ground water movement is propor­ Solution:
tional to the hydraulic gradient, dh/dL. T h e propor­ Equations (11.5) and (11.6) can be combined to obtain the
average velocity of the entire cross-sectional area as
tionality factor, K, is known as the hydraulic conductiv­
ity. High values of hydraulic conductivity m e a n that the V = -K(dh/dL),
material readily transmits water. A m o n g o t h e r things,
hydraulic conductivity d e p e n d s on the size, shape, a n d which is also known as the Darcian velocity. Since water only
connectivity of pores and fractures in t h e aquifer. Fig­ flows through the pores in the media, a porosity term must
Basic Concepts of Ground Water Hydraulics 431

be included to obtain heavy clay soils, the procedure and equipment would
need to be modified. The other three methods pro­
K = -K/n(dh/dL),
duced comparable results. Dorsey et al. (1990) sug­
a

where K is pore velocity. Table 11.2 indicates that porosity


a
gested that selection of the best method for a specific
of a sandy material (aquifer) is approximately 25%. Substitut­ application is dependent upon the soil water condi­
ing η into the previous equation along with the hydraulic tions.
conductivity Κ and hydraulic gradient dh/dL yields Probably the simplest and most widely used method
50 / 1 \ for measuring saturated hydraulic conductivity in the
Κ = - r ^ b z : = - 0 . 2 m/day. presence of a water table is the auger hole method. A
0.25 \ 1000 / '
hole is augered to a depth below the water table, and
Similarly, Table 11.2 contains a porosity for clay of 50%. the water is allowed to rise until equilibrium is reached.
Substituting with other values for the aquitard yields The water is then removed, and the water level
0.0001 / 1 \
recorded. Subsequent changes in water level as a func­
Κ = - ~7Γ77Γ~ — = -0.00002 m/day. tion of time are recorded. The water is again removed,
0.50 \ 10 / '
and the process is repeated several times. The method
The negative velocity simply indicates that water flows in the is probably the best alternative for shallow water table
opposite direction of the hydraulic gradient or from high conditions.
potential to low potential. Thus the flow is expected to travel Hydraulic conductivity of aquifers is difficult to mea­
10,000 times faster in the aquifer under these conditions. sure directly but can be determined indirectly if the
Movement in locations that contain limestone caverns or depth of the aquifer is known and the transmissivity is
rock fractures may be much faster. While actual flow veloci­ estimated using a well test. Transmissivity is defined as
ties are difficult to accurately quantify, the potential for
spreading a pollutant over a large area once it reaches an
aquifer is dependent on K . a
Τ = KD, (11.7)

where Κ is hydraulic conductivity of the aquifer and D


Example Problem 11.2. Flowrate using Darcy's law is the aquifer thickness. Transmissivity is a good mea­
in porous media sure of the total ability of an aquifer to transmit water
under a unit hydraulic gradient.
Estimate the flowrate through an aquifer if the aquifer is
30 m thick and 200 m wide and has hydraulic conductivity
equal to 0.8 m/day. Observation wells have been placed Laplace's Equation
1200 m apart in a direction parallel to the flow. The head in
one well is 28 m, and the head in the second is 20 m. Darcy's law is useful for understanding the theory of
fluid movement in porous materials, but it should be
Solution: Since the observation wells are in the direction of recognized that it is limited in direct application be­
water movement, this is a direct application of Darcy's law cause it deals only with movement in one general
such that substituting yields
direction. Water movement in aquifers is quite dy­
dh 0.8 m/day(30 m)(200 m) (28 m - 20 m) namic. It does not occur in only one direction, but may
Q =
' ^ d L =
1200 m change directions depending upon head. For general
situations where the flow is not one-dimensional,
which simplifies as Darcy's equation must be combined with the equation
Q = - 3 2 m /day.
3 of continuity to provide the basis for a general differ­
ential equation. Consider a rectangular volume ele­
ment Ax Ay A ζ of soil having sides parallel to an
x, y, ζ coordinate system as shown in Fig. 11.7. Let the
A variety of methods for measuring saturated hy­ velocity of water flowing in the χ direction through a
draulic conductivity are available. Dorsey et al. (1990) unit area perpendicular to the jc-direction be V . The
x

compared four field methods that are commonly em­ flow of water per unit time into the face Ay Δ ζ is then
ployed for near-surface measurements. These included V Ay Az. On the opposite face, the flow out is V +
x x

the Guelph permeameter, the velocity permeameter, a (dV /dx) Ax, where the term (dV /dx)
x x represents the
pumping test procedure, and the auger hole method. change in flow across the distance Ax. Analysis of flow
They found that the Guelph permeameter produced in the other two directions yields similar terms. The
significantly lower estimates than the other methods. inflow minus outflow in each direction represents the
They indicated that if the instrument is to be used on net volume of water per unit time accumulating in the
432 1 1 . Groundwater

Pumped well

ion

I / j Drawdown ι

I I I i τ*
Figure 11.7 Definition of volume for Laplace's equation deriva­
Μ ) I I I
tion. I I I I I I
l_l I I I L L
Figure 11.8 Schematic defining a simple well flow problem (after
rectangular volume element from flow in the jc-direc- Linsley et al., 1982).

tion as

Inflow, - Outflow,
b e replaced by Κ (i.e., homogeneous and isotropic
= V Ay Δ ζ - (V
x x + Δχβν /3χ)ΔγΔζ,
χ conditions), t h e n the general equation can be reduced
to
which simplifies for the jc-direction to
dh
2
dh
2
dh2

Inflow, - Outflow, = - (dV /dx) Δχ Δγ Δζ. (11.10)


x
+ t-T + t - t = 0,
dx 2

dy
T h e net flow accumulating in the volume from the
y-direction and z-direction is found similarly so that which is referred to as Laplace's equation. Although
the water accumulating in the volume element is Laplace's equation a p p e a r s relatively simple, in actual
ground water flow problems, it is difficult to apply
dB
Δχ Ay Δ ζ because of limitations associated with its simplifying
Tt assumptions.
= -(dVJdx + dV /dy y +3ν /3ζ)ΔχΔγΔζ,
ζ

W e l l Hydraulics u n d e r Equilibrium C o n d i t i o n s
where V , V , and V are the velocities in the χ-, y-,
x y z

and z-directions and 0 is the water content on a A simplified well problem is shown in Fig. 11.8,
volumetric basis. Eliminating the volume element from which shows a well in a homogeneous and isotropic,
each side of the equation leaves the general differential unconfined aquifer (i.e., the hydraulic conductivity does
equation not vary with location or direction) of infinite areal
extent having water that initially is moving horizontally
dV

x

+— + —
dV y dV 2 dd toward the well. For water to e n t e r the well, there
(11.8)
dx dy dz It must be a drawdown at the well that forms a cone of
depression. However, if drawdown is small compared
Since the equation cannot b e solved directly for the
to the total thickness of the aquifer, and if the well is
velocities, modification using Darcy's law allows the
fully penetrating the aquifer, flow streamlines may be
equation to be written in the form
assumed to b e horizontal so that an appioximation of
d I dh\ d I dh\ d I dh \ 3θ well discharge as a function of aquifer characteristics
can be obtained. Dupuit proposed a solution technique
ΊΪ[~ *ΘΊ;}
Κ +
d ^ \ K
^ J +
dl[~ *Jz~)
K
Έ'=

based on Darcy's law and the assumption that flow


(11.9) through a cylindrical surface at radius χ from the well
must equal the discharge of the well. With this assump­
where K , K , and K are the hydraulic conductivities
x y z
tion,
in the JC-, y-, and z-directions, respectively; h is the
hydraulic head at a point; 0 is the water content on a dy_
volumetric basis; and t is time. If 0 is a constant, such Q = -lirxyK (11.11)
as is the case under saturated conditions, and if K , x

K , and K are equal and constant such that they can


y z where lirxy is the cylinder's area and dy/dx is the
Basic Concepts of Ground Water Hydraulics 433

slope of the water table, which is the driving force


behind the water movement. Integration can be per­
formed with respect to χ from r, to r and y from h
2 l

to Λ to obtain
2

*K{h\ ~ fc?)
Q = (11.12)
ln( r / r , )
2 Ground turfooe

where h and h are the heights of the water table


x 2
^ Initial ptaonwtric turfbo*
above the bottom of the aquifer at distances r and r x 2 ConfMrtQ layer
from the p u m p e d well. A similar equation for a con­ Ι Ι Cono of deprMofon
fined aquifer is

2irbK(h -h ) =3 P
fc P=
b

3
2 x

(11.13)
ln( r / r , )
2
3_
Equations (11.12) and (11.13) represent the interre­
lationship between Q, K, h, and r for steady-state (or
equilibrium) conditions. Linsley et al. (1982) e m p h a ­ Figure 11.9 Cone depression and radius of influence for single
sized that low ground water flow velocities cause true well (after Linsley et al., 1982).
equilibrium conditions to occur only after pumping a
very long time at a constant rate.
Generally, ground water movement is not nearly as by drilling test wells into the ground and pumping the
simple as indicated in the preceding analysis. G r o u n d water at a known rate. Test wells can supply much
water moves in the direction of decreasing total head, information about ground water availability, including
which may or may not be the same as that of the the position and thickness of aquifers and hydraulic
decreasing pressure head. Theis sought to account for features, such as hydraulic conductivity, transmissivity,
the effect of time and storage coefficients of the aquifer storage, and specific capacity. F o r flow to occur to a
because he realized that only a fraction of the total well, there must be a gradient to the well. This gradi­
aquifer depth provides flow to a well. H e a t h (1982) ent forms a cone of depression similar to the idealized
describes a technique involving three wells to deter­ shape shown in Fig. 11.9 for a confined aquifer. In a
mine the movement of ground water. T h e m e t h o d confined, artesian aquifer, t h e actual water level does
requires that the wells be arranged in a triangular not drop. In an unconfined aquifer, the water surface
pattern so that relative location and distance between corresponds to the cone of depression. T h e cone of
wells is known along with the total head at each well. depression is d e p e n d e n t upon the pumping rate and
A n o t h e r difficulty that occurs in defining flow is that aquifer characteristics. A n increased pumping rate or
flow in fractured systems is d e p e n d e n t on t h e extent of low transmissivity will increase t h e d e p t h to the cone of
fracturing, the interconnectivity of the fractures, and depression. If m o r e than o n e well is in an area, cones
the mechanisms available for water to enter the frac­ of depression may overlap and increase the d e p t h to
ture systems. All of these factors are highly variable water in regions between wells as shown in Fig. 11.10
and site specific. Figure 11.6 shows the range of values for an unconfined aquifer. Such well interference in-
experienced for hydraulic conductivity in consolidated
materials. This variability is largely d u e to t h e n a t u r e of
the secondary porosity of the material. Highly frac­ Ground Surface
tured rock may have quite high hydraulic conductivities 77s
and thus b e able to rapidly transmit water.
Karst systems are also difficult to define mathemati­
cally. These systems contain n u m e r o u s conduits that
vary in size from a few centimeters to several meters.
Caves are common. It is not unusual to find the major Drawdown
by each well
flow in a conduit moving opposite to the flow direction
dictated by the general piezometric surface.
Figure 11.10 Combining cones of depression using superposition
T o achieve better understanding about ground water and showing well interference for multiple wells (after Linsley et al.,
availability, ground water supplies are often evaluated 1982).
434 1 1 . Groundwater

fluences the available drawdown and also reduces the function is given by
maximum yield of a well. If a large well is installed, t h e
resulting cone of depression may cause previously o p ­
W(u) = - 0 . 5 7 7 2 1 6 - ln(u) + u -
erating wells to go dry because the water level drops 2 X 2!
below the screens of these wells. T h e piezometric sur­
face of confined, artesian aquifers will d r o p in a similar . (11.16)
m a n n e r as a result of well interference. + •3 x 3 ! 4x4! +
T h e form of t h e Theis equation is such that it cannot
Specific C a p a c i t y a n d Transmissivity be solved directly. T o overcome this problem, Theis
developed a p r o c e d u r e that involves plotting the type
Transmissivity, as given by Eq. (11.7), represents the
curve and test data using logarithmic graph p a p e r . A
rate of flow through a section of aquifer of unit thick­
type curve is a l o g - l o g plot of u versus W(u). Values
ness u n d e r a unit head. Specific capacity of a well is
of observed r /t 2
versus s are plotted on log-log
the flow per unit drop of water level in the well.
scales. T h e two curves are superimposed and moved
Specific capacity of a well is d e p e n d e n t upon both an
about until segments coincide with t h e axes parallel.
aquifer's hydraulic characteristics and those of the well
T h e coincident points d e t e r m i n e u, W(u\ r /t, and 5 .
2
C
itself. H e a t h (1982) listed several components that con­
E q u a t i o n s (11.14) and (11.15) can then be used to
trol specific capacity for ground water wells:
d e t e r m i n e the transmissibility and storage coefficient.
(A) Transmissivity of the zone where the water e n t e r s O n e potential problem in applying this p r o c e d u r e is
the well. (This may be much less than transmissiv­ that Theis assumed that the discharging well is fully
ity of the aquifer depending on the size of t h e p e n e t r a t i n g t h e aquifer. Sometimes it is not possible,
screen.) or desirable, to fully p e n e t r a t e the aquifer. T h e impact
(B) Storage coefficient of t h e water-bearing formation. of partial p e n e t r a t i o n on drawdown must then be con­
(C) Duration of pumping. sidered. T h e Theis m e t h o d has b e e n t h e basis for other
(D) Effective radius of the well. m e t h o d s that a r e m o r e easily utilized. However, the
(E) Pumping discharge. assumptions for the Theis m e t h o d are not nearly as
restrictive as for o t h e r m e t h o d s .
All of these factors except pumping discharge can be
C o m p u t e r s permit trial and error estimation of for­
evaluated using the Theis m e t h o d as discussed below.
mation constants using p u m p i n g test data. Estimated
Since pumping discharge influences well loss, it can
values of the formation constants can be substituted
only b e estimated from an aquifer test in which draw­
into the Theis m e t h o d and optimized until the test
downs are m e a s u r e d in both pumping and observation
d a t a are approximated by t h e simulation.
wells.
J a c o b ' s Method
Theis Equation
O t h e r m e t h o d s for analyzing aquifer test data have
Theis developed an equation to relate drawdown to
b e e n developed. O n e that is somewhat easier to use
transmissibility of a confined aquifer (Viessman et al,
was developed by Jacob ( H e a t h , 1982).
1989). T h e equation can also be simplified if the p u m p ­
Jacob's m e t h o d utilizes d a t a collected after a long
ing continues for a relatively long time. H e a t h (1982)
p u m p i n g time. A s time passes, the shape of the cone
presented Theis's equation as
and the drawdown rate vary. Initially after beginning a
test, the cone of depression is rapidly changing. At a
W{u) Q
(11.14) later time, t h e cone changes m o r e slowly. Jacob's
Air s m e t h o d works only for times such that the terms be­
yond ln(w) in Eq. (11.16) are negligible. This condition
where Τ is the transmissivity, Q/s is the specific capac­
is considered met if
ity ( β is t h e pumping discharge and s is drawdown),
and W(u) is the well function such that u < 0.05, (11.17)

r S„
2
Substituting 0.05 into Theis's equation for u and mak­
u = (11.15) ing units conversions, the minimum time at which
477 '
Jacob's equation applies is given by using the form
where r is the effective radius of the well, 5 is the
C

storage coefficient, and t is the length of pumping time 7200r 5


2
c

(11.18)
prior to determination of specific capacity. T h e well
Basic Concepts of Ground Water Hydraulics 435

where t is the time in minutes when steady-state


c per. U n d e r ideal conditions, d a t a plot as a straight line
conditions develop at a distance r in feet from t h e instead of as a curve. A difficulty with Jacob's method
pumping well, 5 is the dimensionless storage coeffi­
C is, unlike Theis's, it is applicable only in those situa­
cient, and Τ is the transmissivity in square feet per tions in which u < 0.05, w h e r e a s Theis's procedure is
day. m o r e generally applicable.
By setting W(u) equal to - 0 . 5 7 7 2 1 6 - ln(w), Jacob
solved for Τ as

23Q E x a m p l e 1 1 . 3 . Jacob's method


T = log 10 (11.19)
4πΔ$
Use the data shown in Fig. 11.11 to find the formation
constants, transmissivity, and storage coefficient for an
where As is t h e drawdown in t h e interval between t 2
aquifer. The pumping rate is 600 gpm, and the observation
and r,. By setting t h e ratio ί / Ί equal to an o r d e r of
2
well is located 500 ft from the pumped well.
magnitude (i.e., one log cycle), Τ becomes
Solution: First plot a straight line through the data points
23Q that are linear on the semilog plot. The points that fall along
T = (11.20) this line were obtained after steady state was approached.
4vAs
Next find As, the difference in drawdown over one log cycle,
and the storage coefficient becomes by reading the difference in drawdown using the straight line
over one log cycle. Thus Δ$ is 6.8 ft. Then use Eq. (11.20)
2.257ΐ η
with appropriate substitution and unit conversions to obtain
(11.21)
f 2.3(600 gpm)
= 2.16 ft /min.
2

where Q is the discharge (cfs), Δ 5 is the drawdown 4ir(6.8 ft)(7.48 g a l / f t )


3

(feet) measured across one log cycle in the straight line Equation (11.21) can then be solved by reading the straight
portion of a plot of drawdown versus lag time, t is t h e 0
line at a point corresponding to zero drawdown to obtain t 0

time where the straight line intersects t h e zero draw­ equal to 4 min. Substitute this value and other knowns into
down line, and r is distance (feet) between the p u m p ­ Eq. (11.21) as
ing and discharge well.
ft
2

T o apply Jacob's method, drawdown is plotted on 2 . 2 5 ( 2 . 1 6 — 1(4 min)


the vertical arithmetic axis versus time on a logarithmic
1
min 1

5 =
r = 7.8 X 10~ 5

horizontal axis. A straight line is a p p r o a c h e d after (500 fty ,


steady conditions are reached as shown in Fig. 11.11.
A similar procedure is described in Heath (1982), which can
T h e straight line has a slope that is proportional to the
be used to estimate distance versus drawdown relationships.
pumping rate and the transmissivity. T h e procedures
In either set of well tests, three observation wells are pre­
are illustrated in Example Problem 11.3. O n e feature ferred. They should be located in a triangular arrangement
of Jacob's method that makes it easier is its use of with varying distances from the pumping wells. Limited infor­
semilog graph paper instead of logarithmic graph pa- mation can also be obtained from a single observation well,
or even a production well. Heath (1982) provides procedural
information on this subject.
16

14

12 / Ground Water Recharge


ζ
I

*i 10 G r o u n d water recharge is the quantity of water


>< ~T
tf
9.6
a d d e d to the ground water reservoir. Natural recharge
occurs as a result of the natural movement of water
5 68
Q 6 through t h e hydrologic cycle as shown in Fig. 11.1.
/ Natural conditions t e n d to have much higher recharge
rates than less pervious areas such as cities. As a result
OA
r C..O of this, t h e r e has recently b e e n considerable interest in
f

1 • 1 the concept of providing artificial recharge in develop­


1 10 100 1000 ing areas in order to increase ground water supplies
Time since Pumping Began, min and provide storm water control by reducing surface
Figure 11.11 Pump test data plotted for drawdown problem. runoff. Natural recharge to ground water aquifers may
436 11. Groundwater
range from negligible amounts u p to 40 or 5 0 % of m o n and widespread m e a n s of natural recharge. Hy­
precipitation. Artificial recharge occurs when water is drologic processes at the surface and in the vadose
added to the ground water reservoir that would not zone largely d e t e r m i n e the quantity of water that be­
have naturally reached the reservoir. Artificial recharge comes d e e p percolation. Rainfall amounts, timing, and
can result from recharge basins, artificial impound­ intensities are influential. Large quantities of rainfall
ments, recharge wells, applying water to the land sur­ occurring at low intensities during high infiltration pe­
face through irrigation, waste disposal, and other riods will maximize the water available for recharge via
means. Recharge enhancement refers to activities that percolation. Evapotranspiration removes a large frac­
increase the rate of natural recharge. Such activities as tion of the infiltrated water before it can become d e e p
land treatment to increase infiltration could constitute percolation. Climatic, plant, and soil factors govern
recharge enhancement. Care must b e taken to ensure evapotranspiration so that it is often nearly equal to
that artificial recharge does not pollute the aquifer. If precipitation. Large and infrequent precipitation events
recharge water contains pollutants, they may be carried contribute most of the water to d e e p percolation. In
into the aquifer and be difficult to eliminate. humid regions, d e e p percolation can be a significant
Some methods of providing artificial recharge sug­ part of the hydrologic budget and may account for
gested by Linsley et al. (1982) include: several percent of the annual precipitation.
T h u s recharge via d e e p percolation is governed to a
(A) Holding storm water in reservoirs located over
large extent by the hydrologic processes that take place
permeable areas.
in the near-surface zone. This zone generally consti­
(B) Containing surface runoff in reservoirs and releas­
tutes the root zone of any actively growing vegetation.
ing it into stream channels at rates that match the
T h e character of the vegetation is a determinant of the
percolation rate of the channel.
a m o u n t of recharge. In evaluating evapotranspiration,
(C) Discharging streamflow into infiltration areas lo­
the impact of vegetation through such things as type of
cated over permeable formations.
vegetation, density of vegetation, leaf area index, root
(D) Constructing recharge basins that reach p e r m e ­
density, and growing season must b e considered. Fac­
able formations.
tors that reduce evapotranspiration tend to increase
(E) Forcing water down wells into an aquifer.
recharge if all other factors remain the same.
(F) Overapplying irrigation water over p e r m e a b l e for­
T h e permeability of the material above an aquifer is
mations.
often an important determinant of the recharge rate.
T o d d (1980) described many applications w h e r e Highly p e r m e a b l e materials that allow rapid movement
recharge methods are being used. Applications dis­ of water vertically are primary contributors to recharge.
cussed include irrigation, spreading basins, overland If the aquifer is part of a layered system, the least
flow, and recharge wells. p e r m e a b l e layer will govern the recharge rate.
T h e combination of hydrologic and geologic settings F r a c t u r e zones and solution channels through rock
that contribute to natural ground water recharge are material may increase recharge if they are located so
many and varied. Some of the major settings include that they come in contact with water having a pressure
general infiltration and percolation over large areas, potential equal to or greater than atmospheric pres­
percolation from bodies of surface water, such as the sures. In such cases they may act as localized sinks and
influent stream recharging a confined aquifer, inflow of rapidly transmit water u n d e r conditions of free surface
ground water from other sources, and rapid movement and pressure flow to the ground water reservoir.
of water from the surface through fractures and solu­ Localized areas overlying an aquifer may contribute
tion channels. Each of these mechanisms is individually much of the recharge to an aquifer. Such areas may
discussed. have m o r e favorable conditions for allowing the rela­
Recharge d e p e n d s on the availability of water for tively rapid movement of water from the ground sur­
recharge, the physical characteristics of soils and rock face to the ground water. T h e s e conditions may be the
material that the water must pass through, and the result of very permeable soils, solution channels, highly
ability of the ground water reservoir to accept the fractured rock, or shallow ground water conditions.
recharge water. Any one of these three major factors Such areas are often termed recharge areas even
may be limiting and thus define the actual recharge. though recharge may be occurring at slower rates over
Obviously the recharge rate cannot exceed the rate at o t h e r parts of the aquifer. Disturbances of these
which water is available to supply the recharge process. recharge areas have a great potential for impacting the
In the case of surface water bodies, water availability actual recharge rate of an aquifer.
will not be limiting as long as surface water is present. Streams, lakes, and ponds may be a source of
D e e p percolation of infiltrated precipitation is a com­ recharge for some aquifers. In humid regions, the
Movement of Pollutants 437

water level of surface water bodies is often a reflection transmission of water and thus p r o m o t e recharge of
of the ground water level with the slope of the ground ground water as shown by t h e hydraulic conductivities
water surface being downward toward the surface wa­ in Fig. 11.6. Hydraulic conductivity of fractured rock
ter. In such instances the surface water is being aug­ may be orders of m a g n i t u d e higher than in unfractured
mented by subsurface or ground water flow. In semi- rock. If fractures are not interconnected, they cannot
arid and arid conditions, the slope of the ground water serve as conduits for water movement. Slightly frac­
surface is often away from the surface body of water tured systems are thus not likely to allow significant
indicating that the surface water is a recharge source movement of water, whereas highly fractured systems
for the ground water. Streams that contribute water to may serve as major conduits.
ground water are often known as influent streams, Fracturing of rock is brought about by stresses ap­
while streams that gain water from ground water are plied to and released from rock formations. Stress
known as effluent streams, as shown in Fig. 11.4. Some relief fractures are common in the Appalachian area
prefer the terminology gaining and losing streams where overlying soil material has gradually e r o d e d away
rather than effluent and influent streams. A particular and removed part of the compression load on the
stream may be a gaining stream over a part of its underlying rock. A s this load is relieved, the rocks tend
length and a losing stream over a n o t h e r part of its to expand and fracture. A fractured zone is very com­
length. A stream may also be a gaining stream part of mon on the u p p e r layer of rock in the Appalachian
the time at a particular location and a losing stream at region. This fractured zone may be u p to 80 ft thick
the same location at another time. T h e determining and can provide pathways for significant movement of
factor as to whether a surface water body is gaining or water.
losing is the relative elevations of the surface water and O n c e water enters a fracture system, it tends to
the ground water. Losing streams are frequently continue its generally downward movement. Fracture
ephemeral; that is, they go dry during droughty periods flow results in ground water recharge, hillside seepage,
because percolation depletes the flow. or seepage into tributary streams. For water to enter
Aquifer characteristics may limit water recharge in any but the smallest fractures, it must be at or above
instances where the potential recharge rate exceeds the atmospheric pressure. A n example situation where such
rate at which the water is transmitted away from the flow may occur is that water from saturated materials
recharge area resulting in the buildup of a ground can move readily into fracture systems. Infiltration
water mound. This mound would continue to build basins located over a fractured zone can provide large
until the hydraulic gradients in the aquifer were suffi­ quantities of recharge.
cient to cause lateral flows in the aquifer equal to the T h o m a s and Phillips (1979) described several in­
recharge rate or until the m o u n d limited the recharge stances where water movement in macropores would
rate itself. p r o d u c e noticeably different impacts than would Dar-
Changes in any of the factors influential in governing cian (diffuse) flow. A n estimate for the Missouri Ozarks
recharge rates may result in an alteration of the actual is that water travelling through macropores contributes
recharge of ground water. Actual recharge will be five times as much to ground water recharge and spring
reduced if the factor currently limiting recharge is flow than does Darcian flow. W h e n much of the water
altered so as to be more restrictive to recharge or if a flows through macropores, pollutants can move from
factor not currently limiting is changed so as to become the surface at a speed that is much greater than is
the limiting factor. expected using Darcian theory. They can easily add
Topography also impacts recharge because it affects contaminants to saturated zones since the contami­
the time available for precipitation to infiltrate. Steeply nants do not have an opportunity to bond or chemically
sloping land will provide less opportunity time than will react with the soil profile. Obviously, potential move­
flat land u n d e r the same cover conditions. In recharge m e n t of contaminants into ground water is of concern
applications that utilize overland flow, a mild slope is to many people for a variety of reasons.
desirable.

MOVEMENT OF POLLUTANTS
FRACTURE ROCK HYDROLOGY
Pollution of ground water is a serious problem be­
Flow systems in fracture zones are very difficult to cause of the difficulties in correcting the problem once
quantify. T h e controlling factors are the extent, size, it occurs. G r o u n d water often contains large amounts
distribution, and degree of interconnectedness of the of dissolved solids. This is a result of t h e ability of
fractures. A highly fractured material may allow rapid water to dissolve some of almost any substance it
438 11. Groundwater
contacts combined with the extremely long residence h u m a n activities is also of concern. Pollution of ground
time of water after it enters the ground water. Table water also results from inadequate disposal of wastes
11.1 lists several natural inorganic substances that are o n the land surface and improper application of fertil­
commonly dissolved in water that may affect its use. izers or other agricultural chemicals. Septic tanks may
O t h e r substances that are not naturally occurring may contribute bacteria and nutrients to ground water. In­
also cause contamination of ground water as shown in dustrial wastes are a particularly difficult problem be­
Table 11.4. This table also shows the relative impor­ cause many chemicals are toxic in extremely small
tance of the different sources. Obviously, the pollutants concentrations. Coastal locations often have problems
of significance vary considerably from region to region. with saltwater intrusion. This results from the in­
Since pollutants cannot easily be removed from ground creased pumping of ground water necessary to supply
water and ground water is the source of much of the water to rapidly increasing populations. Landfills rep­
drinking water in the United States, maintaining or resent another potential source of ground water pollu­
improving water quality is receiving increasing atten­ tion.
tion from regulatory personnel, environmental groups, O n e aspect of the ground water pollution problem is
and individuals. Chemical pollution has recently been that it can b e difficult to detect. Some activities, such
detected in locations that were considered to be free of as industrial plants and wastewater treatment facilities,
pollution problems only a few years ago. Leakage from are point sources for pollution. O t h e r s , such as agricul­
underground petroleum tanks has received widespread tural and silvicultural applications, are diffused over
concern, and legislation has been passed that seeks to large areas and are referred to as nonpoint sources.
alleviate this particular source of ground water pollu­ Chemicals associated with pollution are often diluted
tion. Deterioration of ground water as a result of other so that a small trace of the chemical may impair large

Table 11.4 S o u r c e s o f Ground Water Pollution in the U . S . b y R e g i o n ( A d a p t e d from


0

National Research C o u n c i l , 1 9 8 4 )

Source NE NW SE SC SW

Natural pollution
Mineralization from soluble aquifers
Aquifer interchange
Ground water development
Overpumping/land subsidence
Underground storage/artificial recharge
Water wells
Saltwater encroachment
Agricultural activities
Dryland farming
Animal wastes, feedlots
Pesticide residues
Irrigation return flow
Fertilization

Mining activities 2
Waste disposal
Septic tanks/cesspools 1
Land disposal, municipal and industrial wastes
Landfills 1
Surface impoundments
Injection wells
Miscellaneous
Accidental spills 2
Urban runoff
Highway deicing salts 1
Seepage from polluted surface waters 3

Northeast includes NY, NJ, PA, MD, DE, and New England; Southeast includes AL, FL, GA,
fl

MS, NC, SC; Northwest includes CO, ID, MT, OR, WA, WY; Southwest includes AZ, CA, NV, UT:
South central includes AR, LA, NM, OK, TX. Reports not completed for Great Lakes and North
Central regions, AK, and HI.
^Numbers indicate degree of contamination: 1, high; 2, medium high; 3, medium low; 4, low.
Movement of Pollutants 439

quantities of water. For example, a pollutant from a ( D ) Close proximity with currently operating systems
point source may enter the ground water at a very may lower the water levels such that some systems
specific point. It then moves laterally and longitudinally do not o p e r a t e .
as the aquifer carries it along. Detection of the pollu­ (E) Wellheads should be properly constructed and
tant in observation wells typically shows a plume that protected.
can be traced to locate the source area for the pollu­ (F) R e c h a r g e area should b e protected from contami­
tant. Such is the case with oil-related substances. Since nation.
petroleum products are less dense than the water a n d
do not easily mix with water, they tend to disperse as a G r o u n d water flow has b e e n modeled for many years
thin film over widespread areas. using physical models such as sand tank (porous me­
T h e potential for ground water contamination prob­ dia) models, analog models using heat or electricity,
lems necessitates the need for careful consideration of and m e m b r a n e models. A variety of these models are
operations that may contribute to ground water pollu­ described in T o d d (1980).
tion. H e a t h (1982) described several factors that can b e Recently, mathematical ground water flow and trans­
used to avoid ground water pollution in the selection of port models that assess the movement of potential
waste disposal sites: pollutants into and with ground water have been devel­
oped. Most mathematical models use the finite differ­
(A) Select a site having significant depth of unsatu­ ence method. This computational m e t h o d divides an
rated clay a n d / o r organic material. aquifer into a grid and analyzes the flow using a
(B) Locate near a point of natural ground water dis­ variation of the equation of continuity (Anderson and
charge. Woessner, 1992). Additional information on several
(C) Divert surface runoff and minimize surface infil­ specific models that have recently been developed can
tration. be found in Bedient and H u b e r (1992). This reference
T o avoid potential contamination of ground water, contains detailed descriptions of the equations and
wells must be situated so that they are not s u r r o u n d e d examples of the code used for several ground water
by areas prone to surface or subsurface pollutants. flow problems.
Such areas include locations adjacent to septic tanks, Increasing availability of c o m p u t e r hardware and
agricultural fields and feedlots, and waste disposal sites. software has led to rapid growth in the use of models
Wellhead protection is especially important because to anticipate the potential movement of pollutants or
leakage around the outside of a pipe going into ground track t h e m to a source. A variety of models are avail­
water will provide easy access for pollutants to the able for use ranging from nonpoint-source pollution
water below. Grouting around the pipe reduces the models, such as G L E A M S developed by the U.S. D e ­
opportunity for pollutants to travel down the outside of p a r t m e n t of Agriculture, to commercially available
the pipe to the aquifer or to travel from o n e aquifer to models for pollutant plume analysis. A major difficulty
another. A similar problem occurs if confining layers with many of these models is the lack of an adequate
leak water between aquifers. If o n e aquifer gets pol­ description of the u n d e r g r o u n d media. F u r t h e r growth
luted, a leaky aquifer can provide an easy route to in these applications is anticipated. Additional infor­
another aquifer. A bibliography and n u m e r o u s refer­ mation summarizing the basic concepts of ground wa­
ences dealing with human-caused ground water pollu­ ter hydrology and a list of references grouped by topic
tion is contained in T o d d and McNulty (1974). can be found in H e a t h (1982). T h e National Research
T h e need to protect ground water from pollution is Council (1990b) has p r e s e n t e d a comprehensive discus­
evident. Some measures only allow for control of bac­ sion of ground water modeling.
terial contamination. In most instances, additional pro­
tection and conservation considerations are appropri­
ate (Soil Conservation Service, 1984): Problems
(A) Measures should b e designed to prevent cross-con­ (11.1) D e s c r i b e t h e prevailing g r o u n d w a t e r
tamination where wells p e n e t r a t e two or more province at your location. A r e either unconfined or
aquifers. confined formations commonly used for drinking wa­
(B) Discharge from a flowing artesian aquifer should ter?
be controlled so that large quantities of water are (11.2) At your location, what naturally occurring
not wasted. inorganic chemicals would you expect to be most likely
(C) A n overall m a n a g e m e n t scheme should include to pollute ground water?
analysis of the amount that can be p u m p e d eco­ (11.3) G r o u n d water is the source for baseflow in a
nomically, the importance of its use, and expected small stream located 100 m from a well. A highly
recharge. water-soluble chemical is spilled at the wellhead and
440 11. Groundwater
leaks into the ground water via an improperly grouted than the piezometric surface, and water leaks vertically
casing. How long would you expect it to take for the downward from the unconfined aquifer to the confined
pollutant to reach the stream if the aquifer is com­ aquifer through the aquitard. Using the bottom of the
posed of good sand and the hydraulic gradient is aquitard as a datum, the water table is at an elevation
1 m / 2 5 m? Impermeable rock is located below the of 40 m, the piezometric surface is at 35 m, and the top
aquifer. of the aquitard is at 8 m. Hydraulic conductivity of the
(11.4) Referring back to Problem (11.3), could the unconfined aquifer is 12 m / d a y and for the aquitard is
well be managed to reduce the damage from the pollu­ 0.3 m / d a y . Assume steady-state conditions, and deter­
tant? Explain. mine the leakage velocity in m / d a y .
(11.5) If the aquifer in Problem (11.3) is located (11.14) Plot the pumping test data shown below on
above rock fractures, would you expect a different semilogarithmic paper. Fit a straight line through the
result? points. T h e test data were collected at a pumping rate
(11.6) Estimate the Darcian velocity for water of 2000 m / d a y with drawdowns m e a s u r e d at an obser­
3

flowing through a column having an area of 80 c m , 2


vation well located 50 m away. Estimate the transmis­
hydraulic conductivity of 0.1 m / d a y , and hydraulic sivity and storage coefficient.
gradient of 0.05. W h a t material(s) would you expect
the column to contain?
t Drawdown t Drawdown
(11.7) Use the same data as shown in Fig. 11.11 to (min) (m) (min) (m)
find the formation constants, transmissivity, and stor­
age coefficient for an aquifer. T h e pumping rate is 0 0 10 0.46
480 gpm, and the observation well is located 300 ft 1 0.20 15 0.50
from the p u m p e d well. 2 0.27 20 0.54
(11.8) Calculate the relative pore velocity for a 3 0.32 50 0.64
sand aquifer that has a Darcian velocity of V and a 4 0.36 100 0.72
porosity equal to 3 7 % . 0.90
5 0.38 500
(11.9) Many aquifers are composed of more than
8 0.43
one horizontal layer, each individually isotropic, but
with different thicknesses and hydraulic conductivities.
Consider a case where there are two horizontal layers (11.15) A well that p u m p e d at 350 m / h r for 8 hr
3

that have thickness z and z and hydraulic conductiv­


x 2 was shut down. M e a s u r e m e n t s of the drawdown as it
ities K and K , respectively. Derive an equation for
x 2 recovered were m a d e in an observation well located 30
the hydraulic conductivity in the horizontal direction. m away. Calculate the transmissivity of the system.
Assume a homogeneous system and a unit width. (Hint: Plot t h e residual drawdown s' as a function of
(11.10) Generalize the equation derived in Problem t/t' on semilog p a p e r where t' is minutes since shut­
(11.9) for η horizontal layers. This equation defines the down and t is time since pumping began.)
equivalent horizontal hydraulic conductivity for a strat­
ified material.
f s'
(11.11) Consider again the two-layered, stratified (min) (m)
aquifer. Derive an equation for the equivalent vertical
hydraulic conductivity, K , for the system. Assume
z 1 0.62
homogeneous conditions and a unit width. Generalize 2 0.57
the equation for η horizontal layers. 3 0.54
(11.12) A n alluvial flood plain contains an uncon­ 4 0.53
fined aquifer with ground water flowing parallel to the 5 0.51
stream channel. T h e depth from the water table to the
10 0.46
aquitard below the aquifer is 40 m, and the flood plain
15 0.44
is approximately rectangular with a width perpendicu­
20 0.42
lar to the stream of 1200 m. Estimate the flow in the
30 0.39
aquifer in cubic meters p e r day and cubic meters per
second if the aquifer has a Κ equal to 80 m / d a y and a 60 0.35

hydraulic gradient equal to 0.01. How does the flow in 90 0.32


the aquifer compare to that of a stream? 120 0.31
(11.13) A n aquitard seperates an unconfined aquifer 240 0.27
from an underlying aquifer. T h e water table is higher
Movement of Pollutants 441

References National Research Council (1990a). "Surface Coal Mining Effects on


Ground Water Recharge." National Academy Press, Washington,
Anderson, M. P., and Woessner, W. M. (1992). "Applied Groundwa­ DC.
ter Modeling—Simulation of Flow and Advective Transport." National Research Council (1990b). "Ground Water Models." Na­
Academic Press, San Diego, CA. tional Academy Press, Washington, D C .
Bedient, P. B., and Huber, W. C. (1992). "Hydrology and Floodplain Soil Conservation Service (1984). "Engineering Field Manual," 4th
Analysis," 2nd ed. Addison-Wesley, Reading, MA. printing. Department of Agriculture, U.S. Soil Conservation Ser­
Bouwer, H. (1978). "Ground Water Hydrology." McGraw-Hill, N e w
vice, Washington, D C .
York.
Thomas, G. W., and Phillips, R. E. (1979). Consequences of water
CAST (1985). Agriculture and ground water quality, Agricultural
movement in macropores. J. Environ. Quality 8(2): 149-152.
Science and Technology, Council for Report N o . 103. A m e s ,
Todd, D . K. (1980). "Ground Water Hydrology." Wiley, N e w York.
Iowa.
Todd, D . K., and McNulty, D . E. O. (1974). "Polluted Groundwater."
Dorsey, J. D., Ward, A. D., Fausey, N. R., and Bair, E. S. (1990). A
Water Information Center, Huntington, N e w York.
comparison of four field methods for measuring saturated hy­
draulic conductivity. Trans. Am. Soc. Agric. Eng. 33(6): Van Voast, W. Α., and Reiten, J. C. (1988). Hydrogeologic re­
1925-1931. sponses: Twenty years of surface coal mining in Southeastern
Dunne, T. and Leopold, L. B. (1978). "Water in Environmental Montana, Memoir 62. Montana Bureau of Mines and Geology,
Planning." Freeman, San Francisco, CA. Montana College of Mineral Science and Technology, Butte, MT.
Heath, R. C. (1982). Basic ground-water hydrology, U.S. Geological Vesilind, P. Α., Peirce, J. J., and Weiner, R. (1988). "Environmental
Survey water supply paper 2220. U.S. Geological Survey, Denver, Engineering," 2nd ed. Butterworths, N e w York.
CO. Viessman, W., Jr., Lewis, G. L., and Knapp, J. W. (1989). "Introduc­
Linsley, R. K., Kohler, Μ. Α., and Paulhus, J. L. H. (1982). "Hydrol­ tion to Hydrology," 3rd ed. Intext Educational Publishers, New
ogy for Engineers," 3rd ed. McGraw-Hill, N e w York. York. %
National Research Council (1984). "Ground Water Contamination." Walton, W. C. (1970). "Ground Water Resource Evaluation."
National Academy Press, Washington, D C . McGraw-Hill, N e w York.
Monitoring Hydrologic Systems

A large part of this book presents techniques for with the values of the p a r a m e t e r s used in the model.
estimating various hydrologic quantities in the absence Every model has p a r a m e t e r s that are used to charac­
of any actual measurements of these quantities at t h e terize the particular catchment of concern. T h e true
location(s) of interest. Only a small fraction of small values of these p a r a m e t e r s are not known and must be
catchments throughout the world are actually moni­ estimated. T h e r e is much uncertainty associated with
tored. For the vast majority of these catchments some t h e p a r a m e t e r estimates.
type of model must be used to estimate the quantities Actual d a t a on t h e physical process of concern from
of interest. Generally the estimation technique used a catchment can b e extremely valuable in reducing
has b e e n developed and tested on only a small subset model and p a r a m e t e r uncertainty. D a t a provide a basis
of the small catchments that are actually monitored. for estimating model p a r a m e t e r s reflective of t h e ac­
T h e net result is that estimation techniques used on tual catchment and for testing the model and the
small catchments have only been tested on a very low selected p a r a m e t e r s . Without any actual data on the
n u m b e r of the total existing catchments. Finally, n o catchment of concern, o n e can only infer model perfor­
estimation technique has b e e n found that is 100% m a n c e and model p a r a m e t e r values based on other
accurate for any catchment. Thus, it is frequently desir­ (and hopefully similar) small catchments. Even very
able to monitor hydrologic and water quality variables limited d a t a can prove to b e quite useful in p a r a m e t e r
on a local and site-specific basis. evaluation.
Municipalities and agencies may well find it very cost
effective to o p e r a t e m o d e r a t e catchment monitoring
UNCERTAINTY programs for the purpose of determining locally appli­
cable models and for determining model p a r a m e t e r s
W h e n a model is selected for application to a small applicable to local conditions. For instance, if a munic­
catchment, at least two points of uncertainty regarding ipality consistently overestimates storm water runoff by
the model exist. T h e first is the uncertainty associated 2 0 % , overexpenditure on storm water-related facilities
with the model itself. N o matter how rigorous, every will be on this same o r d e r of magnitude. O n the other
model has associated with it uncertainty regarding how hand, if a consistent underestimation is m a d e , exces­
well the actual catchment processes are being r e p r e ­ sive costs may be incurred d u e to flood d a m a g e and
sented. T h e second source of uncertainty is associated m a i n t e n a n c e costs.

442
Sources of Data (U.S.) 443

G o o d data are the best source of information on properties change. T h u s , for a given d e p t h , flow may
hydrologic response. G o o d d a t a are preferred to t h e b e above or below t h e value indicated by t h e calibra­
same information g e n e r a t e d from a model. Poor d a t a tion curve. This could b e a r a n d o m error. Similarly, the
may be misleading and can b e inferior to model results. d a t a or p r o c e d u r e used t o define t h e depth-flow rela­
T o p r o m o t e similarity in the type of data collected a n d tionship may not b e representative of t h e t r u e relation­
t h e m a n n e r they are collected, many agencies of t h e ship over t h e entire range of interest. Nonuniformities
U.S. G o v e r n m e n t joined efforts to p r o d u c e a " N a t i o n a l may introduce error into the result in certain d e p t h
H a n d b o o k of R e c o m m e n d e d M e t h o d s for W a t e r - D a t a ranges. T h e s e errors a r e systematic and may lead to
Acquisition" (U.S. Geological Survey, 1977; a n d revi­ over- o r underestimation of t h e flow for certain d e p t h
sions). This handbook contains recommendations for ranges.
collecting data on precipitation, surface and ground
water, water quality a n d sediment, soil water, evapora­
tion, hydrometeorology, snow a n d ice, a n d catchment INSTRUMENTS
characteristics. T h e U.S. D e p a r t m e n t of Agriculture
(Brakensiek et al., 1979) has also p r e p a r e d a c o m p r e ­ F o r many years, instruments used to measure hydro-
hensive manual for field m e a s u r e m e n t of hydrologic logically important variables w e r e generally of the me­
data. This manual is divided into sections on precipita­ chanical type using spring w o u n d clocks, floats at­
tion, runoff, climate, sedimentation, geology a n d soil tached to mechanically activated pens, or weighing
conditions, and watershed characteristics. Similarly, t h e devices attached to mechanically driven pens. Al­
U.S. W e a t h e r Bureau has guides for t h e installation of t h o u g h many of these devices are still in use, they are
various hydrometeorologic instruments. In any d a t a being replaced by electrical and electronic devices.
collection effort, it is important to use s t a n d a r d tech­ M o d e r n d a t a recorders use digital clocks and micro­
niques so that t h e data obtained will be consistent with electronic technology to receive a n d / o r transmit data
o t h e r data sources. t o a central receiving station. D a t a are now stored on
All m e a s u r e m e n t s are subject to an unknown a m o u n t magnetic tapes, magnetic disks, and optical disks.
of error. Two major types of errors are r a n d o m errors M a n y recording stations are battery powered, with
and systematic errors. R a n d o m errors result in errors solar energy recharge capability greatly reducing the
that are both positive and negative and have a m e a n time required to actually visit t h e sites a n d service the
value of zero. R a n d o m errors may result from insensi- instruments. Battery-powered sites can transmit d a t a to
tivity of an instrument to the p h e n o m e n a being mea­ central receiving stations for retransmission or storage,
sured, errors in reading scales, or pulsating conditions. t h u s eliminating t h e n e e d to change charts. Instrumen­
R a n d o m errors are generally chance errors. Since these tation technology is changing rapidly. F o r this reason,
errors are r a n d o m with a zero m e a n , r e p e a t e d sam­ this c h a p t e r addresses t h e principles of m e a s u r e m e n t
pling or m e a s u r e m e n t may b e used to r e d u c e this of hydrologic variables but does not detail the actual
source of error. instruments used.
Systematic errors introduce a bias into d a t a in that Reliability of instruments a n d training personnel that
t h e m e a n systematic error is not zero. I m p r o p e r instru­ service t h e instruments must b e considered in selecting
mentation, always reading on the high or low side of a t h e type of instrumentation to b e used for a particular
scale, and faulty calibrations are frequently t h e cause application. Often t h e most valuable data, the data of
of systematic errors. R e p e a t e d sampling c a n n o t over­ real concern, occur during t h e most adverse w e a t h e r
come systematic errors. a n d flooding conditions w h e n instrumentation failure is
D a t a with small r a n d o m errors are said to have high t h e most likely. Simple, reliable instruments that func­
precision in that the data are repeatable. D a t a with tion properly without observer attention are a must
small systematic errors a r e said to have high accuracy u n d e r these conditions.
in that they are representative of the t r u e d a t a value.
Obviously, minimizing both r a n d o m and systematic er­
rors is desirable. SOURCES OF DATA (U.S.)
Faulty calibration may b e the source of e r r o r in
some instruments a n d may p r o d u c e both r a n d o m and Hydrologic d a t a are available from a variety of state
systematic errors. F o r example, streamflow is often a n d federal agencies. T h e primary source of weather-
related to flow depth. Theoretically, t h e r e is a one-to- r e l a t e d d a t a , such as precipitation a n d t e m p e r a t u r e , is
o n e relationship between flow rate and d e p t h for uni­ t h e National W e a t h e r Service, which has a data center
form flow in a prismatic channel with constant p r o p e r ­ in Asheville, N o r t h Carolina. T h e U.S. Geological Sur­
ties. Natural flow may not be uniform and channel vey, with offices in Reston, Virginia, is a primary source
444 12. Monitoring Hydrologic Systems

of data on surface and ground water quantity a n d Recording raingages are of t h r e e main t y p e s —
quality as well as geology, topography, and aquifer weighing, tipping bucket, and float. D a t a from record­
characteristics. Soils and land-use data may generally ing gages are required for storm water computations
be obtained from the Soil Conservation Service of t h e from small catchments since the time distribution of
U.S. D e p a r t m e n t of Agriculture h e a d q u a r t e r e d in rainfall is as important as the volume of rainfall for
Washington, D.C. These large Federal agencies have these catchments. T h e timing of the gages should be
state offices responsible for data pertaining to particu­ such that rainfall time increments reflective of the
lar states. T h e Soil Conservation Service also has many hydrologic response time of t h e c a t c h m e n t can be
local offices scattered throughout the U.S. d e t e r m i n e d . T h e s e time increments are discussed in
O t h e r Federal agencies that collect data of value to C h a p t e r 3 and may be as short as 5 min.
hydrology and sediment studies are the U.S. Forest As the n a m e implies, weighing raingages respond to
Service and Agricultural Research Service of the U.S. the weight of precipitation passing through the collec­
D e p a r t m e n t of Agriculture; the U.S. Army Corps of tor opening. This weight is converted to an equivalent
Engineers; the U.S. Bureau of Reclamation; the Envi­ d e p t h over the area of the collector. T h e cumulative
ronmental Protection Agency; the Office of Surface weight ( d e p t h ) as a function of time is recorded. T h e
Mining; the U.S. Park Service; and the Bureau of L a n d slope of the d e p t h versus time relationships is the
M a n a g e m e n t of the U.S. D e p a r t m e n t of Interior. intensity of t h e precipitation. Weighing gages may be
State and local agencies should not be overlooked as suitable for estimating snowfall if a suitable m e t h o d of
a source of data. Often universities, especially in con­ melting t h e snow without allowing it t o bridge over t h e
junction with agricultural experiment stations, conduct gage openings is used.
special hydrologic studies producing valuable data. Tipping-bucket gages have two small " b u c k e t s " that
State water agencies and environmental agencies also r o t a t e on an axis below a collection orifice. T h e buck­
may be a source of data on water quantity and quality. ets rotate after collecting a designed and very small
Local governments often conduct special studies of volume* of rainfall. W h e n o n e bucket tips, a second,
hydrology and may have data applicable to their partic­ empty bucket is rotated in place u n d e r the collector
ular areas of responsibility. orifice. T h u s o n e bucket is always in place u n d e r the
Private firms are also archiving data collected by collector orifice. During the actual tipping process,
federal and state agencies in the form of microcom­ some error may b e introduced into t h e recording. F o r
puter-compatible disks. These firms also are vendors of light rainfalls, this error will be small and generally not
software useful in analyzing the raw data. exceed 2 % . T h e water spilled from t h e tipping bucket
is collected and m e a s u r e d so that t h e tipping record­
ings can be adjusted as required to reflect the correct
PRECIPITATION total volume of rainfall.
Float-type gages have a cylindrical c h a m b e r into
Precipitation may occur in the forms of rain, drizzle, which t h e precipitation is directed. A float in this
freezing rain, snow, and ice. T h e only active form of c h a m b e r transmits t h e water d e p t h to the recording
precipitation generally considered in the design of device so that a continuous record of rain d e p t h is
storm water or erosion control facilities is rain. Snow obtained.
melt may contribute appreciably to runoff and erosion All types of gages should be protected from freezing
during the melting phase. Snow melt may be an impor­ by using an a p p r o p r i a t e antifreeze solution. Similarly,
tant erosive agent on small catchments but rarely p r o ­ evaporation may be controlled through the use of some
duces flood flows on them. By contrast, snow melt may type of oil as an evaporation suppressant.
be an important contributor to floods on larger catch­ Raingage catch is very sensitive to wind turbulence.
ments. G a g e s must be located to minimize this source of error.
Precipitation may be measured directly with a col­ It is generally r e c o m m e n d e d that the horizontal dis­
lecting gage of some type or estimated via radar. Col­ tance from any vertical obstructions be at least twice
lector-type rain gages are generally cylindrical with t h e height of the obstructions and preferably four
diameters ranging from 2 to 12 in. (5 to 30.5 cm) or times t h e height. G a g e s should also be located on
more. Studies have shown that the diameter of the horizontal or nearly horizontal areas. U p - and down-
gage has little impact on gage accuracy. Collector gages slope drafts can appreciably impact t h e catch of gages
may be recording or nonrecording. Nonrecording gages located o n slopes. G a g e s should not be located on
are simply designed to store the collected precipitation roofs. T h e gages should be installed so that the top of
until it can be manually measured. T h e standard non- t h e collecting cylinder is horizontal. A c a r p e n t e r ' s level
recording raingage in the U.S. W e a t h e r Service has an can be used to periodically check this condition. It is
8- in. (20.3 cm) opening and is read once every 24 hr. also important that the opening of the collector be
Runoff 445

cylindrical and u n d a m a g e d . Any dents in the opening


should be repaired. Standard gage height in the U.S. is
generally 30 in. (76.2 cm) above the ground.
Raingages should be distributed uniformly. T h e den­
sity of the gage network d e p e n d s on the purpose of the
data and the prevalent types of rainfall. W h e n storms
are cyclonic, widespread, and of low intensity, a sparse
network may be a d e q u a t e . A m o r e dense gage network
is required to characterize convective t h u n d e r s t o r m s of
high intensity and variable areal extent. In mountain­
ous areas, orographic effects must be considered in t h e
layout of the gage network.
Often a combination of recording and nonrecording
gages can be used to improve network density a n d
storm definition. T h e network might consist of four
nonrecording gages for every recording gage. A t least
Figure 12.1 Stream-gaging arrangement.
two recording gages and preferably t h r e e should b e
included in each network.
R a d a r had become a standard instrument in meteo­ vegetation, erosion, stream alterations, and vandalism.
rological networks and is especially valuable in provid­ Stream-gaging stations require constant attention,
ing real-time information on the location, intensity, m a i n t e n a n c e , a n d recalibration to ensure the integrity
type, vertical and horizontal extent, and movement of of the m e a s u r e m e n t .
rainstorms. Quantitative estimates of rainfall amounts
and intensities can be e n h a n c e d by incorporating
Stream Stage Determination
recording raingages with data transmission capabilities.
Correlation of data transmitted from t h e gages and the Stream stage is t h e d e p t h of water above some
radar echos can produce reasonable estimates of rain­ arbitrary d a t u m . T h e reference d a t u m should be se­
fall as it occurs. This type of information is valuable in lected so that negative stages are not possible. For
the real-time operation of storm water m a n a g e m e n t weirs a n d flumes, the reference is often selected so
facilities. that a stage of zero corresponds to no flow, and any
positive stage results in some flow. Stage may be mea­
sured with a staff gage, float-activated recorder, pres­
RUNOFF sure-activated recorder, wire weight gage, or a crest
stage gage. Figure 12.1 shows a typical stream gaging
T h e m e a s u r e m e n t of runoff from small catchments is setup using a float recorder.
a complex topic because of the wide range of flow Staff gages may b e used as t h e primary means of
conditions that might exist. T h e references by t h e U . S . measuring stage o r as a check o n o t h e r types of mea­
Geological Survey (1977) and Brakensiek et al. (1979) surements. A staff gage is a scale placed in the flow so
have excellent coverage of the topic and should be that the stage can b e read directly. Staff gages may be
consulted for many details associated with t h e mea­ placed vertically or inclined. Inclined gages must be
surement of runoff. calibrated to provide correct vertical stage reading for
Flow m e a s u r e m e n t s may be d o n e by: t h e particular angle of inclination involved. Staff gages
should be placed w h e r e they can be easily read from
(1) Measuring stream stage and relating stage to dis­
stream banks or bridges and w h e r e they are protected
charge.
from d a m a g e d u e to debris in the flow.
(2) Measuring flow velocity and calculating discharge
Float-activated recorders a r e the most common d e ­
based on velocity arid area.
vice used to m e a s u r e stage. Generally the installation
(3) Using precalibrated weirs and flumes.
for a recorder of this type has an instrument shelter
(4) Using hydraulic characteristics of channels and flow
and stilling well located off to the side of a channel
control structures.
with an intake pipe extending into the flow. T h e shelter
(5) Using model studies of flow control structures.
must be located so that it is above the highest flow
(6) Using dye dilution techniques.
level to be recorded, is accessible during flood flows,
(7) Using a tipping bucket.
does not interfere with the flow, and is not damaged by
Flow measurement is fraught with difficulties. Prob­ floating debris and ice. T h e intake to the stilling well
lems that must be anticipated are debris, ice, sediment, must be located so that it will not become clogged with
446 12. Monitoring Hydrologic Systems

sediment or debris such as leaves, rags, paper, plastic, prevented from floating. If the stick is of t h e p r o p e r
and other trash. T h e intake must also be large enough length, the cap on the pipe may serve this purpose.
so that the water level in the stilling well corresponds O n e of t h e most frustrating flow-monitoring prob­
to the water level in the stream. T h e suitable opening lems is the loss of data during extreme but rare events.
size d e p e n d s on the area of the stilling well and the By definition, a 100-year event is expected to occur just
rate of change of stage in the stream. once every 100 years. A n event of this frequency pro­
O n c e the recorder is installed, care must be taken to duces large flows with a lot of erosive action and
ensure the chart readings are properly related to stream carrying a lot of debris. Installations that are not well
stage, that the recorder will not be subject to vibrations constructed with a d e q u a t e foundations and located
from flow, wind, traffic, etc., and that the float and above t h e flood level will invariably fail during critical
counterweight can o p e r a t e freely through t h e entire events, and the most valuable d a t a of t h e gaging pro­
range of the anticipated stage. gram will b e lost, to say nothing of t h e instrumentation
A bench mark should be established n e a r the gage itself.
so that the reference elevation of t h e gage can b e
checked periodically. It is also desirable to check the
recorded stage against the actual water stage at several
Velocity D e t e r m i n a t i o n
different water surface elevations to ensure correct
data over the range of the expected stages. Flow velocity may be d e t e r m i n e d using rotating ele­
A pressure-activated gage operates on the principle m e n t current meters, floats, and tracers. Occasionally,
that t h e d e p t h of water above a given point is directly electromagnetic and acoustical flow meters, pitot tubes,
related to the water pressure at that point. By measur­ or optical m e t h o d s are used. T h e s e latter m e t h o d s are
ing the pressure, the stage can be determined. T h e m o r e prevalent in laboratory research t h a n they are in
most common method of pressure m e a s u r e m e n t is to field hydrologic studies.
bubble a gas into the flow and record the required gas Rotating element current m e t e r s may be of the pro­
pressure. Bubble gage installations are often less- peller or cup type. T h e rotating element is held in
expensive than installation of a stilling well and can p r o p e r orientation to the flow by a streamlined body
easily measure a 50-ft (15 m) range in stage. Bubble a n d flow vanes. T h e rotations are electronically or
gages can also b e moved from o n e location to a n o t h e r manually timed and related to flow velocity through a
with relative ease. T h e orifice through which t h e gas is calibration curve. T h e velocity obtained is an estimate
bubbled must not be located where it will be covered of t h e velocity at o n e point in the flow. Figures 4.2 and
with m u d or debris or in highly turbulent flow. T h e 4.3 show that t h e velocity varies t h r o u g h o u t t h e cross
most common bubble gage uses a mercury m a n o m e t e r section of t h e channel. T h u s several m e a s u r e m e n t s
for recording pressure. Continuous records can be ob­ must be obtained.
tained using pressure transducers and automatic T h e general p r o c e d u r e is to divide t h e stream width
recorders. into a n u m b e r of sections, d e t e r m i n e the average veloc­
A wire-weight gage consists of a weight suspended ity for each section, c o m p u t e t h e discharge for each
from a wire or cable wrapped on a cylinder. T h e weight section as t h e product of t h e section area and average
is lowered until it contacts the water surface at which velocity, and finally sum t h e section discharges to d e ­
point a counter on the cylinder indicates the length of t e r m i n e t h e total discharge.
wire required. This in turn is related to water stage. If a logarithmic velocity profile is assumed, t h e aver­
Bridges are frequently used to house wire-weight gages age velocity of a vertical profile is approximately equal
as they m a k e excellent platforms from which to m a k e to t h e average of t h e velocities at 0.2 and 0.8 times the
measurements. flow d e p t h . If t h e flow d e p t h is too shallow to reliably
A crest-stage gage is used to d e t e r m i n e t h e maxi­ d e t e r m i n e the 0.2 a n d / o r 0.8 d e p t h velocity, t h e aver­
mum stage reached during a runoff event. A popular age velocity may be taken as t h e velocity at 0.6 times
crest-stage gage consists of a vertical 2 in. (5 cm) pipe t h e d e p t h . Occasionally, the velocity at all t h r e e d e p t h s
that houses a wooden stick or piece of thin-walled is averaged. Field studies of actual velocity profiles
conduit. Burnt cork is placed within the pipe, which is have verified these approximations.
perforated a r o u n d the bottom. T h e cork floats u p with O n small catchments, especially in u r b a n areas, the
the water in the pipe and tends to cling to the measur­ d e p t h of flow may b e changing rapidly, making it
ing stick when t h e water level drops. By removing the difficult to m a k e very many velocity m e a s u r e m e n t s
measuring stick and recording the height of the de­ without a substantial change in stage. In this event,
posited cork, t h e maximum stage can be determined. relatively few vertical profiles can be sampled and the
Measuring sticks m a d e of buoyant materials must b e 0.6 d e p t h reading should b e used.
Runoff 447

Floats may b e used t o approximate average flow rapid rise a n d fall of t h e hydrograph, making it difficult
velocities. Difficulties e n c o u n t e r e d with floats include to complete discharge m e a s u r e m e n t s .
the variation in downstream velocity across t h e channel
and t h e fact that surface a n d not average velocity is
Precalibrated Weirs a n d Flumes
measured. A correction factor of 0.85 times t h e float
velocity is often used to approximate t h e average p r o ­ Precalibrated weirs a n d flumes a r e commonly used
file velocity. Correction factors from 0.80 to 0.95 may to measure runoff from small research catchments.
be applicable depending on t h e particular flow condi­ Such weirs d o not generally require field calibration.
tions. U.S. Geological Survey (1977) a n d Brakensiek et al.
Tracer techniques for determining flow velocities (1979) contain excellent t r e a t m e n t s of this topic. In
consist of injecting a slug of tracer into t h e flow a n d C h a p t e r 5, t h e use of weirs for flow control is dis­
measuring the time it takes for t h e centroid of t h e cussed. Weirs a n d flumes m a k e excellent flow-measur­
c o n c e n t r a t i o n - t i m e curves t o pass two points a known ing devices because they have a one-to-one relationship
distance apart. Very little e r r o r is introduced if t h e b e t w e e n discharge a n d stage.
time between peak concentrations r a t h e r than cen- A weir is a n overflow structure placed across the
troids is used. Occasionally, visual observation of t h e flow. T h e edge over which t h e flow occurs is called the
time of passage of a dye tracer is used; however, actual crest of t h e weir. Stage is generally m e a s u r e d relative
concentration m e a s u r e m e n t s give m o r e reliable results. t o t h e weir crest. T h e crest of a weir may b e sharp
A fluorescent dye such as R h o d a m i n e W T can b e (metal plates) or b r o a d (small dams or humps). If t h e
detected with good accuracy at low concentrations overflow from t h e weir is unobstructed by t h e water
using a fluorometer. level downstream from t h e weir, a free overfall is said
T h e amount of R h o d a m i n e W T 2 0 % solution r e ­ to exist. If t h e weir overflow is partially u n d e r water,
quired for a time of travel study can b e estimated from the weir is t e r m e d submerged a n d a unique stage
(Kilpatrick, 1970), discharge relationship may n o longer exist. Weirs may
have a free overfall u n d e r low flows a n d become sub­
J L\ - 0 9 3

m e r g e d u n d e r high flows. F o r flow m e a s u r e m e n t , sub­


K = 2.0xlO- |G -pJ
d
3
C , m p (12.1) mergence of t h e weir should b e avoided.
T h e s t a g e - d i s c h a r g e relationships for some typical
where V is t h e volume of dye (liters), Q
d is t h e m
sharp-crested weirs a r e given in C h a p t e r 5. Sharp-
maximum discharge (cubic m e t e r s p e r second), L is t h e crested weirs a r e difficult to maintain in natural streams
length of t h e reach (kilometers), V is t h e estimated because t h e edge becomes nicked a n d bent from debris
m e a n velocity (meters p e r second) a n d C is t h e d e ­ p
in t h e flow. T o overcome this problem, t h e U.S. D e ­
sired p e a k concentration (micrograms p e r liter) at t h e p a r t m e n t of Agriculture (Brakensiek et al., 1979) has
downstream e n d of t h e reach. C generally should not
p
developed a series of broad-crested, V-notch weirs that
exceed 5 ^ g / l i t e r . have b e e n widely used on research catchments. Rating
Regardless of t h e m e t h o d employed, velocity mea­ tables a n d dimensions of weirs for measuring dis­
surements should b e m a d e in straight a n d uniform charges u p t o 1300 cfs a r e p r e s e n t e d .
sections of a channel. This m a k e s t h e determination of
Weirs present problems if t h e flow is transporting
discharge based on velocity a n d a r e a m o r e accurate. By
sediment o r debris. Sedimentation tends to occur im­
combining m e a s u r e m e n t s on stage a n d discharge, t h e
mediately u p s t r e a m from t h e weir causing shifts in the
stage-discharge or rating curve for a stream may b e
s t a g e - d i s c h a r g e relationship a n d requiring frequent
determined. Frequently t h e relationship may b e ex­
clean out. This problem may b e largely overcome by
pressed in the form
using a flume. Flumes a r e designed to cause flow to
Q = aH , b
(12.2) pass through critical d e p t h thus producing a unique
s t a g e - d i s c h a r g e relationship. They also a r e designed to
where Q is discharge and Η is stage. F o r some loca­ allow t h e flow t o pass without ponding a n d without a
tions, a separate rating curve is required for the rising b o t t o m obstruction so that sediment and debris tend to
and falling stages of the hydrograph. T h e rating of pass without clogging t h e flume.
streams from larger catchments, especially in t h e pres­ T h e common types of flumes a r e Η-flumes and
ence of alluvial channels, is a specialized topic beyond Parshal flumes (Fig. 12.2). Brakensiek et al. (1979)
the scope of t h e coverage p r e s e n t e d here. F o r small present design information on Η-flumes having capaci­
catchments, problems with shifting controls a n d ero­ ties u p t o 117 cfs a n d Parshal flumes u p to 3000 cfs.
sion and sedimentation within t h e channel a r e not as Η-flumes provide excellent accuracy for very small
severe. A major problem with small catchments is t h e catchments. Parshal flumes have t h e advantage of low
448 12. Monitoring Hydrologic Systems

F o r example, in C h a p t e r 5 t h e hydraulics of culverts


was discussed. Frequently these hydraulic relationships
are used to develop a rating curve for culverts, which
t h e n is used to d e t e r m i n e t h e discharge. A stage-
recording device can be used to monitor flow continu­
ously or high-water marks can b e used to estimate p e a k
discharges.
Occasionally, flow hydraulics may not be amenable
to analytic considerations, and model studies may b e
required to develop the n e e d e d s t a g e - d i s c h a r g e rela­
Sheet metal
tionship. Properly conducted model studies are expen­
Β
sive and a r e generally limited to important installations
that are going to be in operation over a long period of
time. Laboratory facilities a n d the expertise required
for hydraulic similitude model studies are available in a
limited n u m b e r of places—generally government or
Crest university facilities. Often these facilities will contract
Diverging for outside studies.
Section

Dye Dilution

T r a c e r dyes may be used to m a k e flow m e a s u r e ­


ments. If a tracer is injected at a constant rate and
allowed to completely mix with t h e stream, discharge
may be estimated from

(12.3)
C? — C h

w h e r e Q is the stream discharge, q is the rate of tracer


injection, C is the background tracer concentration in
b

Figure 12.2 Flumes used for flow measurements. ( A ) H-flume, t h e stream, C is the tracer concentration in the injec­
x

(B) Parshal flume. tion stream, and C is the m e a s u r e d tracer concentra­


2

tion in t h e stream. This relationship assumes that an


equilibrium condition exists.
Dye dilution techniques require that t h e dye be
head r e q u i r e m e n t s — a b o u t a fourth of that required thoroughly mixed with the flow at the point of sam­
for a weir having the same crest l e n g t h — a n d can thus pling. A manifold system may b e required for injecting
be used in channels of low slopes. Parshal flumes are t h e dye so that complete mixing occurs in a relatively
more expensive and may not provide satisfactory mea­ short distance. C a r e must be taken not to inject the dye
surement of low flows. For example, a Parshal flume in a concentrated fashion so that the dye stream is not
having a maximum capacity of 1000 cfs has a minimum completely mixed with the stream flow. Incomplete
free-flow capacity of 10 cfs. D e p e n d i n g on the purpose mixing will r e n d e r the concentration C nonrepresen-
2

of the gaging program, it may be necessary to install an tative and will invalidate t h e results. If t h e flow con­
additional weir for low-flow m e a s u r e m e n t s on some tains significant quantities of sediment with an ex­
catchments. Weirs and Η-flumes are much easier to change phase with the dye, significant error in flow
construct than Parshal flumes. Η-flumes and Parshal estimation may result.
flumes are available commercially in various sizes and
materials.
GROUND WATER
Flow Control Structures
T h e most fundamental quantity m e a s u r e d when
T h e hydraulic characteristics of some flow control monitoring ground water is the elevation of the water
structures may b e used to develop a s t a g e - d i s c h a r g e table or piezometric surface. O t h e r quantities of inter­
relationship and thus serve as a flow-measuring device. est include physical characteristics of aquifers and
Ground Water 449

measures of ground water quality. G r o u n d water moni­ must be of sufficient diameter, when finished, to ac­
toring sites should be selected to be as representative c o m m o d a t e measuring and sampling equipment. O b ­
as possible of the factors being monitored. They should servation wells must be carefully installed to ensure
be free of influences that are extraneous to the pur­ that the resulting m e a s u r e m e n t s are reflective of the
pose of the monitoring program. For example, if a specific aquifer of interest. Precautions must be taken
ground water level monitoring program is desired to to prevent leakage and possible contamination of one
measure the impact of land t r e a t m e n t on aquifer aquifer by another.
recharge, the monitoring sites should not be located in Piezometers are much like observation wells except
a municipal well field where drawndown d u e to p u m p ­ they are generally smaller in diameter and designed to
ing would obscure the desired data. Monitoring sites collect water at o n e location r a t h e r than throughout
should be located where they are accessible u n d e r all the d e p t h of an aquifer, as in t h e case of an observa­
expected weather conditions. They should also be se­ tion well. Piezometers are most frequently used in fine
cure from external disturbance and d a m a g e . packed material w h e r e changes in the piezometric sur­
Although existing wells may be used for monitoring face occur very slowly.
wells, it is generally preferable to install observation Pits, ponds, streams, etc., may be used as an indica­
wells specifically for data collection. W h e n an observa­ tor of ground w a t e r levels in t h e immediate vicinity.
tion well is installed, the following data should be W a t e r quality determinations from these sources may
collected (U.S. Geological Survey, 1977): not be reflective of g r o u n d water quality because of the
high likelihood of surface water contamination. If these
(1) Aquifer(s) tapped.
surface water sources are used as indicators of ground
(2) Aquifer hydrologic characteristics
water levels, care must be taken to d e t e r m i n e the
(3) Lithologic and geophysical logs.
aquifer that is in hydraulic contact with them. A pit,
(4) Well depth, size and type of casing or finish,
for example, may contain water that has e n t e r e d from
location, and type of perforations.
the surface and reflects a temporary, perched water
(5) Elevation of land surface and measuring point.
table r a t h e r t h a n t h e w a t e r table elevation of an expan­
(6) Diagram and photograph of well showing access
sive aquifer.
to well and measuring point.
If a network of monitoring locations is established,
(7) D a t e the well was drilled.
the elevation of a measuring point or reference point
(8) Well-response data for u n p u m p e d well (specific
should b e established at each location. T h e elevations
capacity tests).
should all be referenced to a common d a t u m . T h e
(9) Local well n a m e and owner.
reference points should be p e r m a n e n t and easily relo­
(10) Location by legal description such as latitude and
cated.
longitude coordinates.
(11) Significant features n e a r well that could affect the
water level. Wcrter-Level M e a s u r e m e n t s
(12) Use of the well.
T h e most c o m m o n m e t h o d s of water-level measure­
G r o u n d water monitoring sites might consist of exist­ m e n t are g r a d u a t e d steel t a p e , electrical measuring
ing wells, observation wells, piezometers, or exposed line, air lines, and float-activated recorders.
features such as pits, ponds, springs, or streams. Exist­ Water-level m e a s u r e m e n t s with a steel tape are
ing wells may prove satisfactory as monitoring loca­ straightforward. T h e t a p e is simply lowered into the
tions; however, pumping from the well may interfere well and a reading taken at the reference mark. T h e
with the monitoring program. It is also essential that t a p e is withdrawn and the wet line on the tape deter­
data of the type listed above be available. For example, mined. T h e difference in these two readings is the
the aquifer from which the water is originating must be d e p t h below the reference mark to the water surface.
known if more than one aquifer exists in the locality. A dark t a p e is easier to use than a bright one. Fre­
A b a n d o n e d wells may also be satisfactory; however, quently t h e t a p e is coated with chalk to aid in deter­
the reason for a b a n d o n m e n t should be known. If clog­ mining the wet line reading. A lead weight should be
ging or well collapse or some other problem is present, attached to t h e down hole end of the tape to ease
faulty data may result. lowering the t a p e and to ensure that the t a p e remains
Specially drilled observation wells are the best straight.
ground water monitoring sites. A s the well is drilled, T h e electrical m e t h o d consists of an electrical probe
detailed lithologic and geologic data can be collected. that is lowered into the well until electrical continuity
D a t a are easy to collect because there is no p u m p to is established. T h e p r o b e may be one wire serving as a
interfere with the m e a s u r e m e n t s . Observation wells positive lead with t h e well casing serving as the nega-
450 12. Monitoring Hydrologic Systems

tive lead. Alternatively, the probe may contain both a ments. Surface m e t h o d s rely on the differential physi­
positive and negative lead. A battery is connected to an cal properties of various geologic formations to trans­
a m m e t e r or light and to t h e electrical probe. W h e n t h e mit electrical signals and elastic disturbances. Many
probe touches the water, the electrical circuit is com­ geophysical m e t h o d s have b e e n developed by petroleum
pleted and the a m m e t e r or light activated. T h e electric geologists and are m o r e a p p r o p r i a t e to that field than
probe should b e weighted to ensure that it hangs to geohydrology.
straight in the well. T h e probe should also b e shielded A c o m m o n electrical geophysical m e t h o d is to apply
to protect against false readings that might occur if the a direct current or very low-frequency alternating cur­
probe contacts t h e walls of the well. T h e p r o b e wire is rent to two electrodes. Between these two electrodes,
marked so that when the circuit is complete, the d e p t h t h e voltage d r o p across two additional electrodes is
from the reference point can be easily determined. recorded. T h e s e m e a s u r e m e n t s permit an estimate to
A n air line consists of a small diameter tube to b e m a d e of the a p p a r e n t resistivity of t h e underlying
which a compressed air source is attached. T h e tube is strata. R e p e a t e d m e a s u r e m e n t s are m a d e as the elec­
lowered into the well. T h e depth to the water surface is t r o d e spacing is increased allowing t h e electrical influ­
obtained by subtracting t h e length of t h e submerged ence to p e n e t r a t e m o r e deeply. A s t h e electrode spac­
tubing from the total length of the tubing. T h e length ing is increased, the a p p a r e n t resistivity of d e e p e r
of the submerged tubing is obtained by measuring the layers is d e t e r m i n e d .
air pressure on the tube and converting this pressure to T h e resistivity of geologic materials varies widely
a hydrostatic h e a d (h = p/y). This calculation assumes ranging from 1 0 " Ω-m for graphite to 1 0 " Ω-m for
6 2

no significant pressure loss in forcing the air through quartzite. Dry materials have a higher resistivity t h a n
the tubing. T h u s air flow rates must be kept small, and the same material w h e n wet. Clays have smaller resis­
clogging of the tubing must be avoided. tivities than gravels. T h u s , based on a p p a r e n t resistivi­
T h e piezometric surface elevation for flowing wells ties, an estimate of underlying formations can b e m a d e .
can be determined by capping the well and measuring If resistivity d a t a a r e correlated to well log data, the
the water pressure in the capped well. T h e pressure interpretation of t h e underlying geology can be greatly
must be allowed to stabilize before this determination improved.
is m a d e . If the pressure in t h e well is small, a short Seismic m e t h o d s rely on the differential travel time
piece of tubing may be used by attaching the tubing to of seismic waves through geologic materials. Loose
the capped well and raising the tubing until t h e flow unconsolidated material transmits t h e waves m o r e
stops. For higher pressures, a mercury m a n o m e t e r or slowly t h a n consolidated material. T h e m e t h o d im­
pressure gage can b e used. poses a s u d d e n disturbance either by an explosive
For continuous records of water level, a float-actu­ device or by a heavy h a m m e r blow to a steel plate lying
ated recorder may be installed on the well. If this on the ground. T h e time it takes for the resulting
approach is used, care must be used to ensure that the seismic wave to travel various distances is d e t e r m i n e d
float and the counterweight do not interfere and that by a series of geophones. By studying these travel
the t a p e does not drag along the side of t h e well. times, the d e p t h to rock layers can b e d e t e r m i n e d . T h e
T h e frequency that water level m e a s u r e m e n t s are p e t r o l e u m industry makes extensive use of seismic re­
m a d e d e p e n d s on the purpose of the m e a s u r e m e n t s flection m e t h o d s , while hydrogeologists often find seis­
and t h e dynamics of the ground water system being mic refraction m e t h o d s m o r e satisfactory. Seismic
monitored. Rapidly responding systems may require m e t h o d s use a travel time versus distance curve for the
daily measurements, while systems that respond slowly seismic waves. T h e slope of various portions of this
may be adequately characterized by weekly or even curve can b e used to d e t e r m i n e velocities. T h e inter­
monthly measurements. In a cyclic system such as a cept can b e used to d e t e r m i n e depths.
p u m p e d well field for municipal or irrigation water, T h e variations in gravimetric and magnetic fields
sampling frequency must be such that the cyclic re­ may also be related to changes in the unconsolidated
sponse of the water table can be completely defined. material such as buried stream channels, which may
prove to b e productive aquifers.
As with electrical m e t h o d s , seismic, gravimetric, and
Geophysical Measurements
magnetic field data are especially useful when used in
A n u m b e r of geophysical methods to assist in deter­ conjunction with well logs. Geophysical well logging
mining t h e m a k e u p of geologic formations have b e e n gives direct access to information on subsurface m a t e ­
developed. T h e s e m e t h o d s are useful in locating a n d rials. Well logs may b e constructed from visual obser­
mapping the extent of aquifers. Geophysical m e t h o d s vation of material p r o d u c e d as a well is drilled, by
may be divided into surface and subsurface m e a s u r e ­ recording t h e speed of p e n e t r a t i o n of the drilling pro-
Water Quality 451

cess, or by more sophisticated logging techniques such Surface Water


as electrical, acoustical, or nuclear logging.
Sample sites must be accessible and must produce
Generally good borehole geophysical d a t a are ob­
samples representative of the objectives of the sam­
tained from boreholes used in conjunction with the
pling programs. Generally they must b e located where
petroleum industry. Only large, municipal water wells
flows are well mixed. Certain types of instrumentation
are generally logged by other than visual means. D o ­
may require external electrical power while others may
mestic water wells are often logged visually. Many
be battery operated. Samples must be taken from points
states require well logs to be submitted to a central
in the flow that are representative of the general flow
state agency where they are filed and available for
situation. For example, if the objective is to sample
others to inspect. T h e following descriptions of bore­
sediment concentrations, several samples at different
hole techniques are summarized from F e t t e r (1980).
d e p t h s at the same location may be required because
Caliper logs: Measures the diameter of uncased of the variation in sediment concentration with depth.
boreholes. Hole diameter may vary because of caving Ideally the sampling location should b e such that re­
or solution-enlarged bedding planes and joints in car­ gardless of where within the stream cross section the
bonate aquifers. sample is collected, the same results would be ob­
T e m p e r a t u r e logs: Best d o n e after fluid in borehole tained.
regains t e m p e r a t u r e equilibrium after boring. May re­ Sampling immediately below a concentrated source
veal geothermal gradients. Aquifers may also have dis­ of lower quality water may result in samples that are
tinct temperatures. not representative of the catchment as a whole. O n the
Single-point resistance: O n e electrode is lowered other hand, if the purpose is to monitor potential water
into borehole on insulated cable. T h e o t h e r electrode quality degradation d u e to a concentrated source of
is grounded at the surface. T h e resistance of all m a t e ­ pollutant, sampling sites might b e located immediately
rial between the lower electrode and the surface is below and above the impacted site.
determined. If the borehole fluid is homogeneous, some If automatic, r e m o t e samples are used, they must be
of the variation in resistance will be d u e to lithology located where they will not be impacted by high flows
near the borehole. or by vandalism. T h e samplers must also be located
Resistivity: Two electrodes are lowered and the re­ w h e r e they can b e easily inspected and where the
sistance between the electrodes determined. Resistivity samples that are collected can b e properly handled.
logs can often be calibrated and used quantitatively. Sampling locations at or n e a r stream-gaging stations
Spontaneous potential: Measures the natural elec­ are desirable if information o n total loads of various
trical potential that develops between the formation constituents is to b e determined. Sometimes correla­
and the borehole fluid. O n e use is to distinguish shale tions between concentrations and flow rates can be
from sandstone lithology. Shale has a positive sponta­ m a d e so that stream-gaging information is useful in
neous potential and sandstone a negative o n e if the estimating concentrations prior to the sampling pro­
salinity of the formation fluid is greater than that of gram and between the actual sampling times.
the borehole fluid. Sampling frequency is d e p e n d e n t on the variability
Nuclear logging: Measures either natural radioactiv­ in the constituents being sampled. Rapid changes in
ity of the rocks and fluid or their attenuation of in­ concentration require frequent sampling, while monthly
duced radiation. Nuclear logging may be d o n e in cased samples may b e sufficient w h e r e concentrations are
or uncased wells and the results are not affected by the relatively stable. If concentrations are d e p e n d e n t on
type of drilling mud. Natural gamma radiation, n e u t r o n flow rates, a sampling schedule tied to the flow rate or
logging, and gamma-gamma radiation are used. stage in the stream may be desirable. Some commercial
automatic samplers can b e set to sample on a flow
basis or on a time interval basis. T h e purpose of the
WATER QUALITY study and the variability in the constituents of interest
will determine the most appropriate sampling scheme.
A good water quality monitoring program requires Since t h e "first flush" of runoff water is often rich in
that representative samples be collected, that samples pollutants, it is generally desireable to include it in the
be handled in a m a n n e r that will maintain their in­ sampling program.
tegrity, and that p r o p e r analytical procedures be used Types of samples range from grab samples to sophis­
during the analysis phase. T h e weakest link in this ticated automatic monitoring. With grab samples, a
triad determines the confidence that can be placed in collecting jar is lowered into the flow, uncovered, and
the results of the monitoring program. the sample collected. Sample containers should be
452 12. Monitoring Hydrologic Systems

rinsed in the flow prior to this p r o c e d u r e . Plastic In sampling for sediment load, it is important to
containers are generally preferable to glass containers obtain representative samples of t h e entire cross sec­
unless the constituents of interest react with or are tion in both a horizontal and vertical context. Using a
absorbed by plastic. Glass containers are preferred for p u m p i n g sampler with a single intake location is not
organic components. Clean sample containers are es­ likely to yield a representative average sediment con­
sential. G o o d laboratory procedures should be used to centration.
ensure that the sample containers are not contami­
nated.
Ground Water
A grab sample is a point sample. A sampling proce­
d u r e that takes water from all d e p t h s at a particular G r o u n d water samples must b e h a n d l e d with t h e
location (depth-integrated sample) is generally p r e ­ same precautions as surface water samples for accurate
ferred to a point sample. Some constituents exhibit and meaningful determination of water quality p a r a m ­
definite concentration profiles. Sediments and con­ eters. T h e siting of sampling points and the frequency
stituents transported by sediments generally have of sampling will be governed by the purpose of the
higher concentrations near the bottom of a channel sampling program. If the purpose of the monitoring is
and lower concentrations n e a r the surface of t h e flow. to evaluate the overall water quality associated with a
A true depth-integrated sample will collect water at a region or an aquifer, sampling points must be selected
rate proportional to the flow velocity. to b e representative of t h e region or t h e aquifer.
Automatic pumping samplers are frequently used to If the sampling program is designed to monitor
collect water quality samples from small catchments. ground water quality impacts from a waste site, m a n u ­
T h e expense of these samplers generally m e a n s that facturing facility, or feedlot, then the monitoring wells
only one, fixed location in that stream is sampled. must be located downgradient from these facilities,
Obviously for valid sampling, this single point must be w h e r e they will truly sample t h e ground water that
representative of the flow in general. Often small flumes might b e impacted. Ideally, both upgradient and down-
are installed to measure the flow rate and to improve gradient locations would be sampled to better quantify
flow mixing so that a point sampled by an automatic impacts of t h e object of the monitoring program.
pumping sampler is more representative. A u t o m a t i c Well-casing material may impact water quality. Shal­
samplers can generally be p r o g r a m m e d to take samples low wells may be cased with plastic. For d e e p e r wells
on a set time interval or on a flow proportional basis. cased with metal, sufficient water should be p u m p e d so
Samples collected over time may be treated as indi­ that n o n e of the water standing in the well and possibly
vidual samples or may b e composited. T h e choice is interacting with the casing is included in the sample.
d e p e n d e n t on the purpose of the study and the re­ Multiple aquifers in the same location require spe­
sources available for sample analysis. cial care. Ideally, a series of wells should be used so
Continuous sensing of certain p a r a m e t e r s such as that a particular well draws water from only one aquifer.
t e m p e r a t u r e , p H , and conductivity is possible. Sensors F e t t e r (1980) outlines a p r o c e d u r e whereby a single
may be located directly in the flow, or water may b e well can b e used to sample water quality from multiple
continuously p u m p e d into a sensing chamber of limited aquifers. T h e p r o c e d u r e described would be useable
volume. only o n e time as it involves sampling an aquifer and
Sediment sampling in streams is an especially diffi­ then increasing the d e p t h of the borehole to the next
cult task (see C h a p t e r 7). T h e difference in density aquifer.
between water and sediment particles results in the G r o u n d water quality generally does not change very
particles settling toward the bottom of the stream. rapidly. U n n a t u r a l disturbances to the system can,
Turbulence acts to suspend particles. T h e net result is however, result in rapid changes. T h e arrival of a
a sediment concentration gradient with increasing con­ pollutant plume, for example can signal a step change
centration n e a r the stream bottom. T h e r e is also a in some water quality p a r a m e t e r s . Spills and other
separation of particle sizes with the larger particles sudden contaminations may also cause abrupt changes
predominating along the stream bottom. in ground water quality. T h e time between the spill
Sediment is transported in streams as suspended and the detection of t h e impact in ground water de­
load and bed load. In reality, t h e r e is not a sharp p e n d s on t h e distance t h e sampling location is from t h e
boundary between the two transport mechanisms. They spill and the rate of movement of ground water through
grade into each other. Devices for sampling the sus­ the intervening formations.
p e n d e d load are better developed than are bedload F o r d e e p aquifers, relatively infrequent sampling
samplers. Brakensiek et al. (1979) should be consulted may suffice as a check on water quality. A frequency of
for a discussion of the various types of sediment sam­ o n e sample per year may b e satisfactory. Shallower
plers. aquifers and recharge areas for d e e p aquifers may
Water Quality 453

require more frequent sampling since the quality in Problems


those areas can be more easily impacted. Wells used
(12.1) T h e following stream-gaging data have been
for ground water sampling should be properly grouted
collected. Estimate the flow rate. Stations are refer­
if p e r m a n e n t and filled and sealed if not p e r m a n e n t to
enced to the left b a n k and measured perpendicular to
prevent surface water contamination of the aquifer.
t h e stream.

Sample Handling Velocity (fps) at


Station Depth
Sufficient volume of sample must b e collected to (ft) (ft) 0.2 day 0.8 day
meet the requirements of all of the analytical tests that
are to be performed. For certain constituents, immedi­ 0 + 00 0.00
ate field testing of the sample must be done to obtain 0 + 25 3.91 2.5 2.3
reliable results. For these constituents, changes occur 0 + 50 8.62 3.1 2.8
rapidly and r e n d e r a delayed analysis of little value. 0 + 80 10.87 3.9 3.4
T e m p e r a t u r e , p H , specific conductance, and dissolved 1+20 12.14 4.1 3.6
gases are examples of constituents requiring immediate
1+80 10.42 3.8 3.3
field analysis.
2 + 20 9.58 3.7 3.2
For some constituents it is best to chill or possibly
2 + 50 5.17 3.2 2.9
freeze the sample for later analysis. O t h e r constituents
2 + 75 2.34 2.5 2.3
may require the addition of a stabilizing agent such as
nitric acid. Biologic activity is greatly reduced by chill­ 2 + 90 0.00

ing the sample to 0.5° C. Frozen samples may yield


unreliable results for certain dissolved organic and (12.2) T h e following are stage-discharge data for
inorganic constituents. Samples for metal analysis Crooked C r e e k n e a r Nye, Kansas. Plot the data on log
treated with nitric acid may be stored for several paper. Estimate the coefficients a and b in Eq. (12.2).
months prior to analysis. Refrigerated samples are Estimate the discharge for a stage of 8.0 ft. Briefly
generally stable in the absence of sediments. Samples discuss the confidence you have in this estimated dis­
should always be stored in a cool place out of direct charge.
sunlight. Regardless of storage life, samples should be
analyzed as soon as possible after collection for best
Stage Discharge Stage Discharge
results. All samples should be clearly and permanently
(ft) (cfs) (ft) (cfs)
marked with sufficient information to identify all perti­
nent variables such as location of sample site, date and 2.03 118 6.28 2880
time of sampling, n a m e of observer, flow rate or gage 3.68 1360 6.72 4370
height, and other n e e d e d information.
4.65 2310 7.40 7400
4.78 2530 7.59 10000
Analytical Procedures 5.66 3970 5.47 3070

Standard analytical techniques should be used for all 6.13 4950 4.49 1550
analyses. Some references to accepted analytical proce­ 4.18 2080 5.98 3730
dures are the American Society for Testing Materials 3.72 1400 5.68 3210
(1981 and revisions), American Public Health Associa­ 3.89 1610 5.10 2370
tion (1980 and revisions), U.S. Environmental Protec­ 5.12 3330 4.47 1320
tion Agency (1979 and revisions), and U.S. Geological 6.93 7100 8.01 13600
Survey (1977 and revisions). These standard techniques
5.28 3490 4.25 2140
must be complemented by good laboratory manage­
6.82 5970 4.07 1840
ment and record keeping.
4.65 2150 4.21 1840
All instruments should be routinely calibrated. A n a ­
5.00 3570 4.56 2290
lytic techniques should be constantly checked through
duplicate sample analysis, spiked samples, alternate 4.89 1820 7.94 13200

techniques, a n d / o r the use of alternate laboratories. 6.20 3930 7.00 5680


Standard reference materials are available from the 6.50 4660 8.98 12000
Office of Standard Reference Materials of the U.S. 7.15 6360 9.00 12100
National Bureau of Standards in Washington, D.C., or 6.70 4910 5.53 2280
the Quality Assurance Branch of the U.S. Environmen­ 6.08 2980
tal Protection Agency in Cincinnati, Ohio.
454 12. Monitoring Hydrologic Systems

(12.3) A float requires 38 sec to travel 150 ft. T h e References


width of the stream at the measuring point is estimated
American Public Health Association, American Water Works Asso­
to 25 ft and the average depth 5.5 ft. Estimate the
ciation, and Water Pollution Control Federation (1980 and revi­
dischange. sions). "Standard Methods for the Analysis of Water and
(12.4) A float requires 29 sec to pass through a Wastewater," 15th ed. American Public Health Association,
100-ft wide rectangular bridge opening. T h e road bed Washington, D C .

is 20 ft wide. T h e depth of flow is estimated at 3 ft. American Society for Testing Materials (1981 and revisions). "An­
nual Book and A S T M Standards," Part 31. Water. Philadelphia.
Estimate the flow rate.
Brakensiek, D . L., Osborn, Η. B., and Rawls, W. J., coordinators
(12.5) A stream reach 1200 m long is available for (1979). "Field Manual for Research in Agricultural Hydrology,"
flow m e a s u r e m e n t using a R h o d a m i n e W T fluores­ Agricultural Handbook 224. U.S. Department of Agriculture,
cence dye. T h e maximum desired dye concentration is Washington, D.C.
4 /xg/liter. T h e estimated mean velocity and maximum Fetter, C. W., Jr. (1980). "Applied Hydrogeology." Merrill, Colum­
bus, O H .
discharge are 1.4 m / s e c and 150 m s e c , respectively.
3

Kilpatrick, F. A. (1970). Dosage requirements for slug injections of


How many liters of dye are required? Rhodamine B A and WT dyes, U.S. Geological Survey Profes­
(12.6) A tracer solution injected at a rate of 1 sional Paper 700-B, pp. 2 5 0 - 2 5 3 .
l i t e r / m i n into a stream that is initially free of the U.S. Environmental Protection Agency (1979 and revisions). "Meth­
ods for Chemical Analysis of Water and Wastes," E P A 6 0 0 / 4 - 7 9 -
tracer. T h e tracer solution is 5 0 % tracer and 5 0 %
020, 3rd ed. U.S. Environmental Protection Agency Environmen­
water. Downstream a distance sufficient for complete tal Monitoring and Support Laboratory, Cincinnati, O H .
mixing, the tracer concentration is found to be 2 X 1 0 " 6

U.S. Geological Survey (1977 and revisions). "National Handbook of


l i t e r / l i t e r . Estimate the stream discharge. R e c o m m e n d e d Methods for Water-data Acquisitions," proposed
(12.7) Derive Eq. 12.3. under the sponsorship of the office of Water Data Coordination,
Geological Survey. U.S. Department of Interior, Reston, V A .
13
Hydrologic Modeling

Hydrologic modeling has become commonplace over practices will mitigate adverse water quality impacts.
the past 25 years. Virtually all hydrologic design is Models a r e generally required to evaluate t h e potential
based on t h e results of applying a hydrologic model. effectiveness of various control efforts such as best
T h e ready availability of models a n d computers and m a n a g e m e n t practices. O n c e a model is operational for
the "user-friendly" n a t u r e of many hydrologic models a particular catchment, various combinations of stor­
ensures continued and virtually absolute reliance on age, channel modifications, and land-use changes can
such models. H a a n et al. (1982) present a detailed be evaluated at very little incremental cost. Often a
treatment of hydrologic modeling of small watersheds. particular combination of designs can b e found to meet
T h e availability of sophisticated hydrologic models a hydrologic a n d economic objective that will result in
has greatly improved our ability to perform complex, considerable savings over t h e life of the project. These
detailed hydrologic analyses. Many different designs cost savings will generally greatly exceed the cost of the
can b e evaluated at minimal cost once baseline d a t a modeling effort.
are collected. Software that not only performs hydro- If a planning body has a development plan for a
logic analyses but also suggests a p p r o p r i a t e model particular area and a timetable for the plan, a hydro-
parameters in t h e form of " p o p u p " screens on micro­ logic model can b e used to d e t e r m i n e the type and
computers is available today. Neat, professional-look­ timing of various storm water control works required to
ing reports can b e p r e p a r e d almost automatically. m e e t an agreed o n flow objective. For example, it may
Models are used for a variety of hydrologic studies. b e possible to d e t e r m i n e that 10 years into the devel­
Possibly the most common use is to evaluate the im­ o p m e n t it will b e necessary to install a detention basin
pact of some physical change within a catchment on at a particular location to provide t h e required level of
the hydrology of that catchment. For a model to b e storm water control. Financial scheduling can be done
useful in this m o d e , it must contain p a r a m e t e r s that to ensure that the available funds are in h a n d for the
are sensitive to the catchment changes that are taking project. T h e land might b e secured prior to it becoming
place. If an internal channel system is modified and the inflated in value d u e to the development.
model has n o way of reflecting this modification, t h e T h e r e is a great tendency for a developer or munici­
model obviously will not be able to define t h e hydro- pality to consider each development unit as a separate
logic impact of the channel modifications. entity a n d design conveyance and storage facilities
Concerns about nonpoint-source pollution have led independently for that particular unit. In C h a p t e r 6,
to regulatory requirements for showing how control t h e advantages of considering storm water manage-

455
456 13. Hydrologic Modeling
m e n t on a regional basis were discussed. This a p p r o a c h the definition of theoretical models to be models that
requires t h e use of a hydrologic model. T h e hydrologic include both a set of general or theoretical principles
model makes it feasible to evaluate the impact of a and a set of statements of empirical circumstances. A
structure locally and its regional impact as well. strictly empirical model is o n e that is based on no basic
A well-conducted model study requires detailed physical laws but contains only a representation of data
knowledge of t h e system being modeled. T h e model (empirical results). Admittedly this is not a very clear
does not replace system knowledge. A model simply distinction, but such a distinction is only of academic
carries out computations. By being relieved of compu­ interest anyway. Suffice it to say that no m a t t e r how
tations, however, the hydrologist or engineer should be sophisticated and detailed, all hydrologic models rely
m o r e willing to d o less lumping of p a r a m e t e r s and to on empirical results to some extent.
give m o r e attention to variability that exists within a F o r o u r purposes a hydrologic model will b e defined
catchment. Small, n o n h o m o g e n e o u s areas can b e in­ as:
cluded as they are rather than lumping t h e m with
larger units. Models also allow and encourage t h e use a collection of physical laws and empirical
of " w h a t if" scenarios and encourage the use of inno­ observations written in mathematical terms
vation. a n d combined in such a way as to p r o d u c e
T o limit t h e scope of this chapter somewhat, it is hydrologic estimates (outputs) based on a set
necessary to define what is m e a n t by a hydrologic of known a n d / o r assumed conditions (inputs).
model. T h e t e r m " m o d e l " brings to mind different
things to different individuals. Websters dictionary pro­ T h e r e a r e many ways of "collecting" physical laws
vides the following as definitions of model: " a general­ a n d empirical observations and of " c o m b i n i n g " t h e m
ized, hypothetical description, often based on an anal­ to p r o d u c e a model. Certain of these ways result in a
ogy, used in analyzing or explaining something." If we formulation in which t h e use of a c o m p u t e r is desir­
substitute for t h e word " s o m e t h i n g " t h e word "hydrol­ able. Such models are called c o m p u t e r models. Com­
ogy," we have a reasonably good definition of a hydro- p u t e r models are the focus of the r e m a i n d e r of the
logic model. discussion in this chapter.
Regardless of how models are classified, they can
Hydrologic models have b e e n classified in many ways.
generally b e r e p r e s e n t e d as
Some of the terms that have b e e n used in model
classification are deterministic, parametric, statistical, O = f(I,P,0 +e, (13.1)
stochastic, physically b a s e d , empirical, blackbox, w h e r e Ο is an η X A: matrix of hydrologic responses to
lumped, linear, nonlinear, distributed, theoretical, p r e ­ b e modeled, f is a collection of / functional relation­
dictive, operational, research, design, similarity, iconic, ships, I is an η X m matrix of inputs, Ρ is a vector of ρ
analog, numerical, regression, event, continuous simu­ p a r a m e t e r s , t is time, e is an η X k matrix of errors, η
lation, and conceptual. is the n u m b e r of d a t a points, k is the n u m b e r of
This chapter is limited to a discussion of m a t h e m a t i ­ responses, a n d m is t h e n u m b e r of inputs.
cal models. Mathematical models range from single Responses in Ο may range from a single number,
prediction equations to complex computer simulation such as a p e a k flow or a runoff volume, to a continuous
algorithms. T h e mathematical basis for a model may b e record of flow, soil water content, evapotranspiration,
theoretical or empirical. A completely theoretical a n d o t h e r quantities.
model would contain only relationships derived en­ Model classification refers to t h e n a t u r e of f. T h e
tirely from basic physical laws. Such laws are t h e con­ distinction b e t w e e n I and Ρ is not always clear and not
servation of mass, conservation of energy, laws of ther­ of extreme i m p o r t a n c e . t o t h e discussion h e r e . G e n e r ­
modynamics, etc. Empirical relationships are based on ally I represents inputs, some of which are time vary­
observations a n d / o r experimentation. Manning's equa­ ing, such as rainfall, t e m p e r a t u r e , a n d land use, while
tion for uniform o p e n channel flow is an empirical Ρ represents coefficients particular to a watershed that
equation. Often empirical equations become so well must b e estimated from tables, charts, correlations,
accepted and entrenched in usage that they are viewed observed data, or some o t h e r m e a n s .
as physical laws. T h e application of Darcy's equation T h e error term, e, represents the difference between
for flow through a porous media is an example of this what actually occurs, O, a n d what t h e model predicts,
in hydrology.
T h e r e exists no completely theoretical operational
O p = f(I,P,i) (13.2)
model in hydrology. All hydrologic models contain
empirical relationships. T h u s we generally liberalize e = Ο - O . p (13.3)
A Brief Look Back 457

A parametric model is a model having p a r a m e t e r s mathematical relationships is observed and im­


that must be estimated in some fashion. T h e p a r a m e ­ proved by using digital computers to carry the
ters may be estimated based on observed d a t a (calibra­ calculations forward in time. . . . As m a t h e m a t ­
tion), tables a n d / o r charts (Manning's η or curve ical relationships are developed, every attempt
numbers), correlation-type relations (regional analysis), is m a d e to realistically reproduce physical pro­
analysis of site-specific information (water-holding ca­ cesses in the model. Experimental results and
pacities of soils, etc.), experience, an edict from an analytical studies a r e used wherever possible
agency, or by some other means. to assist in defining the necessary relation­
A n empirical model is a model containing any empir­ ships.
ical relationship. Empirical here m e a n s data based or
based on observation. Even Darcy's law and Manning's This statement of their research objective has been
equation are empirical equations. p a r a p h r a s e d and r e p e a t e d by others in many locations
A lumped model describes processes on a scale throughout the world.
larger than a point. A completely distributed model T h e S W M defined watershed flow and storage pro­
would be o n e in which all processes are described at a cesses in mathematical terms and combined them so
point and then integrated over three-dimensional space that a continuous record of estimated streamflow was
and time taking into account variations in space and g e n e r a t e d in response to hourly rainfall and daily evap­
time to produce the total watershed response. otranspiration. In developing the governing relation­
Based on these somewhat restrictive definitions, all ships, many model p a r a m e t e r s were defined and a
hydrologic models are to some degree parametric, em­ conceptual basis assigned to them. Estimation of the
pirical, and lumped. many p a r a m e t e r s generally relied on a streamflow
record and manually adjusting the p a r a m e t e r s until a
satisfactory estimate of streamflow was obtained.
A BRIEF LOOK BACK In 1966, Huggins a n d M o n k e (1966) p r e s e n t e d the
first developments of what has become known as
Many of the current developments in hydrologic A N S W E R S (Beasley and Huggins, 1981). This model is
modeling d e p e n d m o r e heavily on computational im­ representative of a class of models that breaks a water­
provements than on improved representations of hy­ shed into a grid system, simulates the hydrology of
drologic processes when compared to models of 25 each grid and the interaction of the grids with their
years ago. neighbors, a n d integrates these results over the water­
T h e 1960s might be thought of as the golden years of shed to p r o d u c e the total watershed response. G e n e r ­
hydrologic modeling. Digital computers were becoming ally with models of this type, the goal is to use physi­
widely available and hydrologic researchers began tak­ cally based p a r a m e t e r s , since a set of p a r a m e t e r s must
ing advantage of their power. T h e Stanford Project in be defined for each grid within the watershed.
Hydrologic Simulation was initiated in 1959 and, u n d e r In addition to general purpose models applicable to
the general leadership of R. K. Linsley, developed a wide range of watersheds, hydrologic models have
approaches to hydrologic modeling that continue to b e e n developed to address specific questions. T h e
play a major role in the field to this day. A n important model developed at Iowa State University ( H a a n
development from this program was the initiative it and Johnson, 1968a, b; D e B o e r and Johnson, 1971;
provided in fostering research in the development of Campbell et al. 1974) to evaluate the impact of subsur­
9

hydrologic models in several locations throughout the face drainage on flood flows in north central Iowa
U.S. represents such a model.
T h e most famous of the models developed at Stan­ Since the 1960s t h e r e have b e e n thousands of pa­
ford was the Stanford Watershed Model, S W M pers, articles, a n d books written dealing with hydro-
(Crawford and Linsley, 1962, 1966). In their 1966 re­ logic modeling. In the 1960s, hydrologic modelers spent
port, Crawford and Linsley stated: a great deal of time justifying why hydrologic models
should b e developed and their applicability to certain
T h e objective of the research is to develop a problems. Today the situation is reversed. Everyone
general system of quantitative analysis for hy­ wants to use models for every conceivable hydrologic
drologic regimes. T h e most effective way for problem, and applications are frequently m a d e well
doing this has been to establish continuous outside the verified domain of the model being used.
mathematical relationships between elements Many m o d e r n hydrologic investigations would not be
of the hydrologic cycle. T h e operation of these possible without c o m p u t e r models.
458 13. Hydrologic Modeling

MODEL SELECTION Computing Facilities Available


This consideration is of less concern t h a n in the past
T h e availability of a large n u m b e r of models, strong but still must be considered. Most hydrologic modeling
advocates of the various models, in some cases vocal on small catchments is currently d o n e using microcom­
detractors from certain modeling approaches, opti­ puters. T h e model selected must b e compatible with
mistic claims by model developers, and conflicting t h e c o m p u t e r system in terms of input and o u t p u t of
modeling objectives all contribute to a great deal of data and results, operating systems, internal memory,
confusion when it comes to model selection for a disk storage, and graphics. F o r some it may b e best to
particular application. T h e r e are at least three possible hire c o m p u t e r services if a limited n e e d for a particular
approaches to modeling: type of system is at hand. R e q u i r e d peripheral devices,
such as m o d e m s , pointers, plotters, graphics, and moni­
(1) U s e of an existing model. tors, must b e considered.
(2) Modification of an existing model.
(3) Development of a new model.
Likelihood of Other Applications for Model
T h e s e options should be considered in the o r d e r If the likelihood of frequent use of a model is
they are listed. A great deal of time and money is a p p a r e n t , m o r e resources can b e used to obtain t h e
involved in model development and testing. Develop­ required h a r d w a r e and software. O n e temptation that
ing a new model can only b e justified for large and must b e resisted is that of redefining problems to
important projects. If an existing model is modified, match an available modeling system so as to recover
reports based on the modified model should clearly m o r e of the investment in the system.
indicate the modification and should not imply that t h e
original, unmodified model was used.
T h e choice of the best model d e p e n d s on: Available Documentation and Other Forms
of Assistance
(1) T h e problem to b e solved. Well-documented models are a must. T h e documen­
(2) C o m p u t e r facilities available. tation should not only explain how to run t h e model,
(3) Likelihood of other applications for t h e model. b u t should explain or reference the algorithms used in
(4) Available documentation and o t h e r forms of t h e model. T h e availability of help screens and training
assistance. sessions also e n h a n c e s a model and helps a user ini­
(5) Characteristics of the model. tially get started with the model. G u i d a n c e in p a r a m e ­
(6) Users modeling experience. ter estimation and error checking should b e provided.

The Problem to Be Solved Characteristics of the Model


T h e model selected should be matched to the prob­ Obviously, a model must b e able to predict the
lem to b e solved. T h e problem should not b e redefined desired quantities with t h e accuracy required for a
to fit a model that is available. T h e problem is fixed; particular study. Model evaluation is discussed later in
models are variable. In evaluating the problem to be this chapter. T h e simplest model that will m e e t the
solved, the required flow characteristics must b e de­ requirements should be selected. O n e should not con­
fined. Some models estimate only peak flow or runoff fuse model complexity with model accuracy. T h e pa­
volumes. O t h e r models generate runoff hydrographs r a m e t e r s of t h e model selected should b e sensitive to
from single events. Continuous simulation of stream- the purpose of the modeling effort.
flow over long periods of time is also possible with
certain models. W a t e r quality p a r a m e t e r s may b e re­
Users Modeling Experience
quired. O n e must be certain that t h e selected model
can provide an estimate of the flow characteristics that A s o n e gains experience in t h e use of hydrologic
are of importance. Finally, the value of the decision to models, p a r a m e t e r selection, model evaluation, error
be m a d e should be considered. T h e r e is little justifica­ detection, and o t h e r modeling attributes are accom­
tion in spending a great deal of money to arrive at a plished m o r e quickly. A n inexperienced user can be
flow to size a culvert if simply using a larger size culvert overwhelmed by t h e n u m b e r of p a r a m e t e r s and com­
is c h e a p e r (than the modeling cost) and is hydrologi- plexity of some of t h e m o r e detailed models. Because
cally acceptable. of p a r a m e t e r correlations a n d t h e interrelationships
Basic Modeling Approaches 459

Precipitation

i
Snow
ice
i
I Interception
Ε

A.
Channel
storage
Channel
* precipitation

Ϋ Throughffall Influent
t effluent
seepage i Stream
Bank
Surface Surface storage
detention storage Influence
.Infiltration ET
Flood plain Μ
Λ Interflow storage
Soil
water Infiltration
Surface iUnsaturated +
RO Percolation γ f percolation
ET
Ground
water ©aseflowl

Figure 13.1 A representation of the hydrologic cycle.

programmed into most models, it generally takes con­ play a major role in governing vegetative growth, which
siderable time and effort to become familiar with some also influences evapotranspiration. This discussion is
of the m o r e complete and complex hydrologic models. limited to the water flow and storage processes.
In building a hydrologic model a r o u n d a diagram,
such as Fig. 13.1, the factors that must b e defined
BASIC MODELING APPROACHES include t h e capacity limits in t h e various storages, the
rate of release of water from the storage, and the rate
O n e way of differentiating hydrologic models is based of movement of water in the various flow phases. It is
on the time scale of importance. This leads to two desirable to define the processes in as physically based
approaches—event simulation and continuous simula­ a m a n n e r as possible so that model p a r a m e t e r s can be
tion. Event simulation refers to modeling t h e hydro- conceptually, if not actually, related to physical, catch­
logic response to a single, isolated storm. Continuous m e n t p a r a m e t e r s . F o r example, soil water storage might
simulation refers to modeling t h e ongoing hydrology of b e related to the available water-holding capacity of
a catchment over long periods of time, such as years. the soil in t h e root zone and soil water flow processes
Event-based models are commonly used in designing might be related to t h e water-transmitting properties
storm water control facilities for small catchments. of the soil.
Continuous simulation models are used w h e r e long- Figure 13.1 could serve as a basis for developing a
term flow volumes and storage considerations are im­ continuous simulation hydrologic model. Some models
portant. omit some of t h e c o m p o n e n t s shown, some include
Hydrologic models are generally written in terms of additional components, and some expand on the com­
flow and storage processes. Figure 13.1 represents the p o n e n t s shown.
hydrologic cycle on a catchment in these terms with t h e F o r models that a t t e m p t to describe t h e same pro­
boxes representing storage processes and t h e arrows cess, event-based models generally require less detail
representing the flow processes. N o t e that in this dia­ t h a n continuous simulation models. Often subsurface
gram, only the inflow and outflow of water to the flow processes and abstractions with the exception of
system is being considered. O t h e r factors such as solar infiltration, evapotranspiration, and baseflow are ne­
radiation and air masses play an important role in that glected in event-based models based on time scale and
they import energy to drive evaporative processes and o r d e r of magnitude arguments. Figure 13.2 might serve
460 13. Hydrologic Modeling
Precipitation define the total catchment response. T h e former ap­
proach is t e r m e d lumped and the latter distributed. In
actuality, distributed models represent averages over
some finite area, often a hectare or so, and thus are
Surface Surface
Storage Detention lumped to some extent. Similarly, lumped models may
Overland b e applied to very small catchments with the results
jFlow
combined and routed. In this way, l u m p e d models may
Soil -'interflow Channel .Stream b e c o m e distributed models.
Water Storage Flow
It is a p p a r e n t that regardless of w h e t h e r a model is
Figure 13.2 A representation of an event-based model. event based or continuous, lumped or distributed, it
will contain p a r a m e t e r s that must be estimated to
define the flow and storage processes.
as a basis for an event model to predict the hydrologic
response of a catchment to a design storm. In C h a p t e r
3, the steps in estimating a runoff hydrograph were PARAMETER ESTIMATION
given as (a) defining the design storm, (b) deducting
losses (abstractions) from the design storm, (c) routing A s the n a m e implies, p a r a m e t e r estimation is the
the rainfall excess to the channel system, and (d) process by which t h e p a r a m e t e r s of a hydrologic model
routing the channel flow to the catchment outlet. T h e s e are estimated for a particular application. Rational
are the processes that are depicted in Fig. 13.2. T h e p a r a m e t e r estimation must b e tied to some criterion if
rainfall excess is the overland flow arrow and is com­ a unique p a r a m e t e r set is to be found. Some criteria
posed of water released from surface storage. This is that might be used include (1) personal j u d g m e n t of
the a m o u n t in excess of infiltration, surface storage goodness of fit of simulated hydrographs to observed
capacity, and interception. hydrographs, (2) direct m e a s u r e m e n t of physical p r o p ­
In C h a p t e r 3, it was shown that there are many ways erties in the field o r in the lab, (3) indirect measure­
of representing these processes. Building an event- m e n t of physical properties through their relationship
based model consists of combining algorithms r e p r e ­ with other hydrologic processes and watershed charac­
senting flow and storage processes in a logical m a n n e r teristics, (4) optimization of some objective function
and providing a means of estimating the various pa­ either computationally or by trial and error, (5) satis­
rameters that are required as a result of the represen­ faction of agency requirements, a n d (6) compliance
tations. with published tables and charts.
F r o m Figs. 13.1 and 13.2, it can be seen that a Some of the things that m a k e p a r a m e t e r estimation
continuous simulation model may be considerably m o r e for hydrologic models difficult are (1) specification of
complex than an event model. T h e complexity is in­ a p p r o p r i a t e criteria for p a r a m e t e r selection, (2) corre­
creased as the time step is decreased. If a time step of lation among p a r a m e t e r s , (3) a m o u n t of computation
1 day is used in a small catchment, then detailed involved in many models, (4) restrictions on appropri­
overland flow routing and infiltration calculations can­ ate values for some of t h e p a r a m e t e r s , (5) n o n u n i q u e -
not be d o n e . In this case, the storage process becomes ness of p a r a m e t e r sets for certain objective functions,
dominant in the model. Generally for an event-based (6) thresholding in some of the model relationships,
model, a short time increment is used. Typically, for and (7) errors in data. P a r a m e t e r estimation is m a d e
small catchments the time increment will be in min­ m o r e difficult by increasing the n u m b e r of p a r a m e t e r s
utes. If a time step of minutes is used in the continuous to b e estimated, the lack of correspondence between
case, the event model becomes basically a submodel to individual p a r a m e t e r s and measurable physical proper­
the overall model with other components of the model ties of the catchment, multiple objectives, limited data,
keeping track of antecedent storage. For an event and p r o n o u n c e d seasonality in hydrologic regimes.
model, antecedent conditions in terms of soil water Problems in p a r a m e t e r estimation have b e e n recog­
storage are either specified or become an implicit as­ nized for some time. Dawdy and O ' D o n n e l l (1965)
sumption of the model. investigated the possibility of obtaining an efficient
A n o t h e r model classification concerns the m a n n e r in automatic p r o c e d u r e for finding numerical values of
which physical processes are represented. In some the various p a r a m e t e r s of an overall watershed model.
models, processes and parameters are defined as aver­ Beard (1967) and DeCoursey and Snyder (1969) ad­
age representations over the entire catchment, while in dressed c o m p u t e r procedures for finding optimal val­
other models the processes are defined at points or ues of p a r a m e t e r s for a hydrologic model. Jackson and
within cells and then integrated over the watershed to A r o n (1971) reviewed p a r a m e t e r estimation techniques
Parameter Estimation 461

in hydrology. Most of the earlier p a p e r s approached choice of the simpler univariate objective function may
p a r a m e t e r estimation from a mathematical rather than b e appropriate; however, potential users of the model
statistical point of view. A n objective function, gener­ or t h e estimated p a r a m e t e r s may rightfully b e skeptical
ally a minimization of a sum of squares, was defined, of the univariate optimization a n d / o r the model itself
and search techniques were employed to find the pa­ when the poorly estimated SCS curve n u m b e r p a r a m e ­
rameter set that optimized the objective function. ter (which governs runoff volume) is c o m p a r e d to more
Sorooshian (1983) reviewed p a r a m e t e r estimation conventional estimates of this p a r a m e t e r as commonly
techniques for hydrologic models. In his review the found in tables. Inclusion of a m e a s u r e of prediction
shift from a deterministic, mathematical interpretation errors on volumes in the objective function overcame
of parameters toward a stochastic, statistical interpre­ this problem and resulted in curve n u m b e r estimates
tation can be noted. Increasingly, causal models are that were in a g r e e m e n t with conventional estimates.
being viewed as "somewhat structured empirical con­ Runoff typically accounts for only 10 to 3 5 % of
structs whose elements are regression coefficients with annual rainfall. Estimation of all model parameters
physical sounding n a m e s " (Klemes, 1982). A d o p t i o n of based solely on runoff ignores 65 to 9 0 % of the pro­
this viewpoint leads to p a r a m e t e r estimation in a statis­ cesses accounting for water loss from a catchment. T h e
tical framework and focuses attention on treating a assumption that if runoff can be predicted well then all
p a r a m e t e r as a r a n d o m variable (rv) with a probability model p a r a m e t e r s must have b e e n adequately deter­
density function (pdf). T r o u t m a n (1985) m a d e an ex­ mined has no clear justification. Including some mea­
tensive investigation of p a r a m e t e r estimation using this sure of performance in the optimization that reflects
approach. some of the flow and storage processes occurring within
T o this point, a vast majority of the research on a watershed in addition to runoff should improve the
p a r a m e t e r estimation in hydrology has b e e n for t h e stability (reduce t h e variance) and accuracy (reduce the
case where a single objective is to b e met. F o r example, absolute error) of the estimated p a r a m e t e r s .
this objective might relate to prediction of peak flows, Traditionally, p a r a m e t e r estimation criteria have
storm runoff volumes, or daily streamflow. Diskin and b e e n tied to some m e a s u r e of how well predicted
Simon (1977) investigated 12 such univariate objective streamflow agreed with observed streamflow. Espe­
functions. Attempts to find p a r a m e t e r sets that m e e t cially for continuous flow simulation, these criteria are
multiple objectives such as peak flows, runoff volumes, difficult to apply. A n alternative m e a s u r e of perfor­
and daily streamflow in some optimal sense have b e e n mance and thus a basis for p a r a m e t e r estimation is
scarce. how well the model performs in a design situation. For
T h e use of multiple objective criteria for p a r a m e t e r example, o n e might select t h e p a r a m e t e r s of a daily
estimation permits m o r e of the information contained flow model so that the estimated capacity of a reservoir
in a data set to be used and distributes the importance to meet some d e m a n d would be as close as possible to
of the p a r a m e t e r estimates among more components of the capacity estimated on the basis of observed stream-
the model. For example, if a continuous flow simula­ flow data using the same capacity estimation algorithm.
tion model is optimized based on p e a k flows, p a r a m e ­
ters related to evapotranspiration ( E T ) may b e poorly
estimated. If the estimation criteria included both peak
Parameter Estimation Criteria
flow and E T , it is likely that the precision of the E T P a r a m e t e r estimation techniques can be divided into
parameters would be greatly improved without an ad­ two major categories—personal and objective. Per­
verse impact on the peak flow parameters. sonal p a r a m e t e r estimation relies solely on the judg­
Edwards (1988) reported an example of improved m e n t of the modeler in arriving at p a r a m e t e r values.
p a r a m e t e r estimates using multiobjective optimization Objective p a r a m e t e r optimization generally deals with
criteria when applied to the Soil Conservation Service some function of t h e error term, e, of Eq. (13.1) and
(SCS) runoff model. W h e n model p a r a m e t e r s were thus requires some observed data on the quantity being
estimated based solely on a minimization of prediction modeled. Generally, a probability density function for
error sum of squares for peak flows, the resulting e, as well as o t h e r properties such as independence,
parameters would do a good job of predicting p e a k constant variance, and zero mean, is assumed. Equa­
flows but gave very poor estimates of runoff volume. tions (13.1)-(13.3) show that e is a function of the
Changing the estimation criteria to include a m e a s u r e p a r a m e t e r s , P.
of error sum of squares on peaks and volumes and
their interactions had n o appreciable impact on predic­ Personal P a r a m e t e r Estimation
tions of peaks but greatly improved runoff volume Possibly the most commonly used p a r a m e t e r estima­
estimates. If interest lies only in runoff peaks, then the tion technique relies on the personal judgment of the
462 13. Hydrologic Modeling
modeler. P a r a m e t e r s are initially assigned on t h e basis maximize L are sought. Again for hydrologic models,
of judgment, published guides, and physical properties numerical search techniques are generally required.
and characteristics of the catchment. T h e s e initial pa­
rameters are then adjusted again based on j u d g m e n t as Arbitrary Objective Functions
to the appropriateness of the model results. If ob­ Any objective function or criterion function, C , can
served data are available, the p a r a m e t e r s may b e ad­ b e used t o find P. In general,
justed several times in an effort to obtain a "satisfac­
tory" fit to the observations. In the absence of observed
data, p a r a m e t e r adjustment d e p e n d s entirely on judg­
C = G[/ (0),/ (O )]
1 2 p -Gl/^./aiP)], (13.6)

ment. This m e t h o d of p a r a m e t e r estimation has t h e


advantage of allowing the user to weight mentally t h e w h e r e G is t h e arbitrary function, f is a function of
x

importance of various flow components such as p e a k t h e observed values of O, and f is a function of t h e


2

flows, low flows, and runoff volumes, whereas most estimated values of O . N o t e that f may b e trans­
p 2

objective estimation techniques can focus only on a formed to a function of t h e p a r a m e t e r s , / , through 3

single objective. T h e method has the disadvantage of E q . ( 1 3 . 2 ) . Using numerical search techniques, Ρ that
sole reliance on the judgment of the modeler and optimizes C is sought. A n example of this a p p r o a c h
generally a poorly defined objective function in t h e might b e setting f (0) x = l n ( O ) , f = l n ( O ) , and G =
2 p

mind of t h e modeler. Different hydrologists would ar­ Π/ι(0) - / ( 0 p ) ] and finding Ρ that minimizes G .
2
2

rive at different p a r a m e t e r estimates on t h e basis of This would b e a minimization of t h e sum of squares of


the same model using the same data. t h e differences in the logarithms of the observed and
predicted outputs.
Least Squares F o r all of the above four estimation procedures, e is
In the least-squares procedure, a sum, 5 , is defined a η X 1 vector, point estimates for t h e p a r a m e t e r s a r e
as t h e sum of squares of t h e e,: obtained, and a single objective is used.

Bayesian Estimation
S = t e f . (13.4)
1= 1 Bayesian estimation is fundamentally different from
Values of the ρ p a r a m e t e r s that minimize S are sought. t h e above p r o c e d u r e s in that it evolves from probabilis­
For hydrologic models, numerical search techniques tic considerations r a t h e r t h a n some arbitrarily specified
are generally employed in p-dimensional space. objective function. Bayesian estimation is concerned
with t h e probability distribution of t h e p a r a m e t e r s
Minimization of Absolute Errors r a t h e r t h a n point estimates. Point estimates, however,
T h e minimization of absolute errors requires a sum, may be derived as t h e m o d e of t h e resulting distribu­
A, be computed as tion.
Some references to Bayesian estimation are Box and
a = Σ kl- T i a o ( 1 9 7 3 ) , Kuczera ( 1 9 8 3 ) , Vicens et al ( 1 9 7 5 ) ,
i=l
E d w a r d s ( 1 9 8 8 ) , a n d E d w a r d s and H a a n ( 1 9 8 8 , 1 9 8 9 ) .
Values of the p a r a m e t e r s that minimize A in p - d i m e n - Bayesian estimation has t h e advantage that multiple
sional p a r a m e t e r space are sought, generally through objectives may be incorporated into the analysis. Box
numerical search techniques. a n d T i a o ( 1 9 7 3 ) show that t h e point estimates for Ρ
may b e found by minimizing t h e d e t e r m i n a n t of
Method of Moments
S(P), |S(P)|, with respect t o P, w h e r e
T h e m e t h o d of moments requires equating the first
ρ sample m o m e n t s of e with the first ρ population
p
S(P) = e'e (13.7)
moments of the pdf of e. In general the population
moments will b e a function of the ρ model p a r a m e t e r s .
Thus ρ equations in ρ unknowns, which must b e for the case w h e r e e is a η X k matrix of errors as in
solved for values of the ρ unknown parameters, result. Eq. ( 1 3 . 1 ) . S(P) is k X k matrix of sums of squares and
cross products of errors. Since |S(P)| is simply a n u m b e r
Maximum Likelihood (the d e t e r m i n a n t of e'e), numerical search p r o c e d u r e s
T h e likelihood function, L , of e is written as can b e used to find t h e Ρ that minimizes |S(P)|. N o t e
that if e is a η X 1 vector, Eqs. ( 1 3 . 4 ) a n d ( 1 3 . 7 ) a r e
L = Πρ(*Ρ)> (13.5) identical. In this case, t h e Bayesian a p p r o a c h provides
i= l a statistical justification for t h e least-squares proce­
where pie^P) is the pdf of e given P. Values of Ρ that dure.
Event Modeling 463

O t h e r Considerations content if observed data were available and the model


James and Burges (1982) present an excellent discus­ provided estimates of soil water.
sion of p a r a m e t e r estimation and estimation criteria. If
a model structure is such that only a subset of p a r a m e ­
ters impact o n e particular aspect of model output, EVENT MODELING
those p a r a m e t e r s may be estimated on t h e basis of an
optimization criteria related to that particular aspect of T o illustrate event models and modeling, the con­
model output. For example a p - p a r a m e t e r model may struction of an event model based on concepts dis­
be structured so that a subset p of the ρ p a r a m e t e r s
x
cussed in C h a p t e r s 3, 4, and 6 is illustrated. As the
governs runoff volumes and a second nonoverlapping model is developed, it will b e c o m e a p p a r e n t that
subset ρ governs p e a k flows. T h e p a r a m e t e r s in p
2 x
the bookkeeping of modeling is often as difficult as the
could then b e estimated on the basis of a criterion modeling itself. In the simple model to b e developed,
related to volumes while t h e p a r a m e t e r s in p could b e
2
the bookkeeping or sequence m a n a g e m e n t will be left
estimated o n t h e basis of a criterion related t o peaks. to the modeler and not incorporated into the model.
If in this case the p a r a m e t e r s in p and p are all
x 2
Recall that event models a r e used to describe the
estimated o n the basis of peaks, t h e p a r a m e t e r s in p x
response of a catchment to a single hydrologic event or
will be poorly determined and likely will have a large rainstorm. T h u s components of the model must deal
variance. While strict division of influence between with the characteristics of the input rainstorm, abstrac­
p a r a m e t e r sets and outputs generally does not occur, it tions from the rainfall, routing of overland and shallow
is common for some p a r a m e t e r s to have relatively little subsurface flow to the c h a n n e l system, a n d finally the
influence on optimization criteria and thus b e poorly routing of the channel flow to the catchment outlet.
determined. J a m e s and Burges (1982) define sensitivity Detailed p r o c e d u r e s for carrying out these steps have
coefficients that can help identify this possibility. If t h e b e e n presented in earlier chapters. T h u s model devel­
model is t h e n used in a situation where the aspect of o p m e n t becomes o n e of choosing algorithms for repre­
t h e model that is highly d e p e n d e n t o n t h e poorly senting t h e various model c o m p o n e n t s a n d combining
defined p a r a m e t e r s is important, a questionable design t h e m in a logical and functional m a n n e r to produce a
may result. F r o m t h e preceding situation, designing a runoff hydrograph.
storage reservoir with a model whose p a r a m e t e r s were T h e model to be used as an illustration will rely on
optimized on peaks would be a situation w h e r e a the SCS type II rainfall distribution to define the storm
potentially poor design could result. U s e of Eq. (13.7) input, the curve n u m b e r a p p r o a c h to determine ab­
partially overcomes this problem since prediction er­ stractions from rainfall, the SCS curvilinear unit hydro-
rors on both peaks and volumes can b e incorporated g r a p h for overland and shallow subsurface flow rout­
into p a r a m e t e r estimation. ing, and the Muskingum routing p r o c e d u r e for channel
As an alternative to Eq. (13.7) to include multiple routing.
objectives, a weighted criterion function could be used. T h e catchment to be modeled is shown in Fig. 13.3.
Such a function might b e Based on land use, soils, and stream configuration, the
overall catchment is divided into four subcatchments.
C = WiEef+ w J>f, (13.8) Physical properties of these catchments are shown in
2

Table 13.1. Table 13.2 shows the identification system


used in differentiating the various hydrographs. Catch­
where the subscripts refer to different objectives (i.e.,
m e n t D is to u n d e r g o urbanization. It is desired to
peaks and volumes), w is a weight, and e is the error.
evaluate t h e feasibility of using a detention basin just
T h e selection of t h e weights is arbitrary a n d may be
below the total catchment to mitigate the impact of
done to reflect the relative importance of t h e two
this urbanization on downstream flooding. Figure 13.4
objectives. Equation (13.8) differs from Eq. (13.7) in
shows the basic model unit ( B M U ) that constitutes the
that interaction between the two objectives is not in­
main routine of the model to be developed. T h e B M U
cluded. For the case where k = 2, Eq. (13.7) becomes
consists of t h e type II distribution, curve n u m b e r pro­
c e d u r e , and unit hydrograph p r o c e d u r e . T h e B M U
must be applied to each of t h e catchments. Figure 13.5
shows how the B M U s must be combined to produce
Equation (13.8) can be generalized to any n u m b e r of the total catchment hydrograph. T h e schematic of Fig.
objectives. It is not necessary that all e used in an
t 13.5 represents the portion of the model that manages
equation like (13.7) or (13.8) be related to flow. For o u t p u t from the B M U . This m a n a g e m e n t aspect is
example, e might refer to some m e a s u r e of soil water
2 highly d e p e n d e n t on t h e particular catchment being
464 13. Hydrologic Modeling

Table 13.2 Catchment Identification S y s t e m

Code Catchment or hydrograph description

A A
Β Β
C C
D D before development
DP D after development
Ε C and D lumped
EP C and DP lumped
Figure 13.3 Watershed schematic. AB A added to Β
ABR AB routed to outlet of C
ABRC ABR added to C

modeled. A good hydrologic model must have an easily ABRE ABR added to Ε
used m a n a g e m e n t scheme. ABREP ABR added to EP
T h e event to b e modeled is a 6-in. rain in 24 hr. T h e ABRCD ABRC added to D
time distribution for t h e rain will be taken as t h e type ABRCDP ABRC added to DP
II distribution of t h e SCS. T h e modeling approach is to
determine t h e runoff hydrograph from catchments A
and B, a d d these hydrographs together, route t h e com­
bined hydrographs through stream segment C, a n d a d d
the hydrographs from catchment C a n d catchment D t o b e r a n d development of t h e catchment does not have a
produce t h e total runoff hydrograph. Figure 13.6 shows large impact o n this curve n u m b e r .
the component hydrographs a n d t h e total resulting In a modeling effort such as this, t h e m a n n e r in
hydrograph. T h e same procedure is followed after t h e which catchments a r e defined can have a p r o n o u n c e d
development of catchment D . T h e only change is in t h e influence o n t h e conclusions reached. In this particular
curve n u m b e r representing catchment D . Figure 13.7 example, it might b e logical t o define catchment Ε as
shows that t h e impact of t h e development on t h e t h e combination of catchments C a n d D (C a n d D
hydrograph from catchment D is an increase in t h e lumped). A predevelopment hydrograph could t h e n b e
runoff volume a n d peak flow rate. Figure 13.8 indi­ developed by applying t h e B M U t o catchment Ε a n d
cates, however, that t h e impact of t h e increase in flow adding t h e result to t h e r o u t e d s u m of t h e flows from
from catchment D is of n o significance as far as total catchments A a n d Β to p r o d u c e t h e total runoff hydro-
catchment flow is concerned. A detention basin would graph. T h e total runoff hydrograph after development
not b e required. T h e reasons for this a r e (1) t h e could b e d e t e r m i n e d as before by considering catch­
relatively small part of t h e total catchment that is m e n t s C a n d D separately. Figure 13.9 shows t h e result
represented by catchment D a n d (2) catchment D in of this approach. If Fig. 13.9 was t h e only information
the undeveloped state has a relatively high curve n u m - available, o n e might conclude that development of

Table 13.1 Catchment Properties

Catchment

Property A Β C D DP

Soil HSG Β Β c C C
CN 72 65 79 79 85
Area (acres) 60 75 80 20 20
Land slope (%) 5 6 7 10 10
Stream slope (%) 1 1 0.5 2 2
Max flow length (ft) 3200 3500 6000 1000 1000
Stream seg. length 3000
Event Modeling 465

Φ 24 hr rain 400

• Figure 13.6 Component and total hydrographs.


Figure 13.4 Basic model unit ( B M U ) .

a p p r o a c h from which o n e would conclude development


catchment D produced a quick hydrologic response so
on catchment D has only minor (negligible) impact on
that runoff from D was nearly over before the flow
the total runoff hydrograph.
from the bulk of the catchment reached the catchment
O n e purpose of this discussion has b e e n to illustrate
outlet. Further, one might conclude that development
t h e importance of catchment definitions a n d t h e im­
of catchment D actually reduced the total runoff peak
pact of lumping. If t h e effect of a small change on a
flow. These conclusions are, of course, contrary to the
large watershed is to be investigated, the lumping of
conclusions reached on the basis of Fig. 13.8. T h e
t h e area undergoing change with t h e rest of the catch­
reason for this disparity is that in Fig. 13.9 in the
m e n t will often mask any impacts. Generally, catch­
undeveloped case, C and D were lumped, while in the
ments with widely varying characteristics should not be
developed case, C and D were treated separately.
lumped. Catchments with different land uses, soils,
Figure 13.10 shows this l u m p / n o lump impact when D
surface topography, stream characteristics, etc., should
is undeveloped. Clearly both modeling approaches can­
not be correct.
A third alternative would be to lump C and D both
400
prior to and after development and generate runoff
hydrographs. Figure 13.11 shows the result of this

BMU BMU BMU


for A for Β forC forD

4±h

Channel
Route

- Total Catchment
Hydrograph
12 13 14 15 16
Time (hours)
Figure 13.5 Management module. Figure 13.7 Impact of development on catchment D .
466 13. Hydrologic Modeling
400 400

ABRE

300 300 h

Μ 01
υ u
* 200 ABRCDPJ * 200 h
ο ABRCC

ABRCD

100 100

11 12 13 14 15 16 11 12 13 14 15 16
Time (hours) Time (hours)

Figure 13.8 Impact of development on total runoff hydrograph. Figure 13.10 Effect of using different catchment definitions on
runoff hydrographs.

be identified and treated separately with the flows 400


combined through a routing process. Certainly differ­
ent degrees of lumping should not be used to delineate
some hydrologic impact. T h e approach of Fig. 13.9 is
clearly invalid. A r g u m e n t s may develop over the pref­
erence for the approach of Fig. 13.8 or Fig. 13.11.
Possibly Fig. 13.8 is more valid as it tends to m o r e
sharply identify changes due to developments on catch­
ment D . This illustration points out that consistently
applied models may be used to evaluate relative im­
pacts, but overreliance on the absolute numerical pre­
dictions should be avoided, especially for uncalibrated

400

13 14 15 16
Time (hours)
Figure 13.11 Impact of development C and D lumped.

model applications (no observed data used to temper


p a r a m e t e r estimates). Hydrologic models produce hy­
drologic estimates. F o r comparisons between estimates
to be valid, a consistent modeling approach must be
used.

A Note on Parameter Estimation


11 12 13 14 15 T h e B M U for this event model requires estimation
Time (hours) of watershed physical characteristics and the runoff
Figure 13.9 Effect of development using inconsistant definitions of curve number. Assuming the physical characteristics
catchments. can be estimated from m a p s and o t h e r data sources,
Continuous Simulation Models 467

the curve number, C N , becomes the only p a r a m e t e r S, making t h e determination of t h e single " b e s t " set
that must b e estimated. difficult. This relates to the nonuniqueness problem.
In the absence of any streamflow data, C N will T h e r e are many examples of currently available event
necessarily be estimated from tables such as those in models. Some of the most widely used in the U.S. are
C h a p t e r 3 and from the judgment of the modeler. the SCS T R - 2 0 model of the Soil Conservation Service
Obviously a good estimate for this p a r a m e t e r is impor­ (U.S. D e p a r t m e n t of Agriculture, 1973); t h e University
tant since it will b e called on to reflect changes in of Kentucky's S E D I M O T II (Wilson et al, 1983); the
hydrology d u e to development of catchment D . Since U S G S model (Carrigan, 1973); the Storm W a t e r Man­
there are n o streamflow records to use in validating a g e m e n t Model, S W M M , of t h e Environmental Protec­
either the model or the estimates for the various CNs, tion Agency ( H u b e r et al., 1981); and the U.S. Army
the modeler must rely on previous validation of the Corps of Engineers H E C - 1 model (U.S. Army Corps of
model for other catchments and assume the model will Engineers, 1985). T h e s e models are all lumped param­
be equally valid for this particular catchment. Changes eter models. Two examples of distributed parameter,
in land use, such as development, must b e reflected in event models are A N S W E R S (Beasley and Huggins,
the estimates for CN. Without validation data, the 1981) and F E S H M (Ross et al, 1979). All of these
modeler can assume that the relative changes in flow event models combine c o m p o n e n t s discussed in C h a p ­
from the pre- to postdevelopment state are accurate if ters 3, 4, and 6. Effective use of t h e models requires a
the model is a valid o n e and is properly used, but basic u n d e r s t a n d i n g of t h e individual model compo­
cannot assume that the quantitative estimates are abso­ nents.
lutely correct.
Proper estimation of model p a r a m e t e r s is absolutely
essential to valid model studies. A model that has
p a r a m e t e r s that can b e estimated from available CONTINUOUS SIMULATION MODELS
sources of data, including streamflow if streamflow
data are available, must b e selected. Continuous simulation models differ from event
If streamflow data are available at the outlet of t h e models in that they a t t e m p t to simulate the hydrologic
total catchment, these data can b e a valuable source of response of a catchment over long periods of time.
information about the correct values for the p a r a m e ­ Continuous input streams of precipitation and fre­
ters. O n e could go through the data and pick out quently t e m p e r a t u r e or solar radiation are required.
isolated storm events. F r o m each of these storm events, Continuous models should represent m o r e of the hy­
a C N could be determined on the basis of Eqs. (3.21) drologic processes that are occurring on a catchment
and (3.22). This C N could t h e n serve as a guide for t h a n event models. Generally, a continuous accounting
estimating the individual CNs for the four individual of soil water content and ground water storage is
catchments. For models where the p a r a m e t e r s ) cannot required. A representation of the interaction between
be explicitly determined, p a r a m e t e r value (s) that mini­ soil water content, evaporative d e m a n d , and stage of
mizes S of Eq. (13.4) can b e sought. T h e e would t plant growth is often required. In general, many more
represent the difference between observed and pre­ p a r a m e t e r s are required by continuous simulation
dicted runoff volumes. Again if the total catchment is models t h a n by event models. Some continuous simula­
considered as o n e unit, a single value for the C N can tion models require that 30 or m o r e p a r a m e t e r s be
be estimated rather easily by a straightforward mini­ estimated before t h e model can be run. Obviously a
mization of S. large n u m b e r of p a r a m e t e r s implies a fairly complex
If it is desired to estimate a C N for each of the four model structure to incorporate their individual im­
catchments that make u p the total catchment, then the pacts. It is not only difficult to visualize the effect of so
estimation process becomes m o r e involved. It still re­ many p a r a m e t e r s , but it is difficult to determine a
mains to find the set of CNs that minimize S; however, unique set of p a r a m e t e r s to fulfill some objective func­
the problems discussed in the p a r a m e t e r estimation tion.
section of this chapter begin to c r e e p in. T h e s e are T o illustrate the working of a continuous simulation
problems of nonuniqueness, errors in data, p a r a m e t e r model, a simple four-parameter model for simulating
correlations, etc. In t h e example given h e r e , a mini­ monthly w a t e r yield from small catchments will be
mization routine that would allow minimizing S over used. Again this is not an attempt to p r o m o t e any
four dimensions (one C N estimate for each c o m p o n e n t particular model. This model has b e e n chosen for
catchment) would b e required. For this particular illustrative purposes so that the basic process of contin­
model and p a r a m e t e r set, o n e may well find that sev­ uous simulation modeling can b e illustrated without
eral p a r a m e t e r sets produce nearly identical values of getting involved in trying to u n d e r s t a n d the multitude
468 13. Hydrologic Modeling

of relationships required in t h e m o r e complex models


having several p a r a m e t e r s .
This four-parameter water yield model was devel­
oped at the University of Kentucky ( H a a n , 1972). T h e
model is representative of many such models. O n e
attractive feature of t h e model is that it has a built in
p a r a m e t e r estimation scheme that d e t e r m i n e s the val­
ues for the four p a r a m e t e r s of t h e model that mini­
mizes the sum of squares between the observed a n d
predicted monthly flows.
T h e water yield model shown schematically in sev­
eral ways in Fig. 13.12 divides the water-holding capac­
ity of the soil into a volume M , which is readily r

available for evapotranspiration, and a volume M,


which is less readily available for evapotranspiration.
T h e maximum capacity of M is 1 in. of water. T h e x

maximum capacity of M, is M L C , a model p a r a m e t e r .


Af and Μ, are used to d e n o t e the current volume of
r

water stored in the readily and less readily available


(for E T ) soil water zones, respectively.
Precipitation input into the model is daily totals.
This daily total is t h e n divided into hourly rainfall on
the basis of the appropriate SCS type curve as shown
in C h a p t e r 3. Hourly precipitation is divided into infil­
tration and runoff. T h e infiltration rate, / , is deter­
mined from

/ = / m a x forP>/ m a x and M < 1r or M, < M L C


/ = P forP</ m a x and Mr < 1 or M, < M L C
/ = 0 for M = 1 r and M, = M L C , (13.10)

where / is the maximum possible infiltration r a t e


m a x

and Ρ is the precipitation rate. All infiltrated w a t e r is


stored in M until the entire 1-in. capacity is filled, at
r

which point any additional infiltrated water is a d d e d


directly to Af,. W h e n both storage are filled to their SRO
capacity, all precipitation is assumed to b e c o m e runoff.
ET
T h e surface runoff volume S R O is d e t e r m i n e d from

SRO = (P-f)t for P>f


SRO = 0 forP</, (13.11) SRO
M,
where t is the time increment involved.
T h e daily evapotranspiration E T is d e t e r m i n e d from S
RO
ET = E n for P = 0;
d 0 <M r < 1.00
Figure 13.12 Schematic of four-parameter water yield model.
M,
ET = £„ for P = 0;
d M = 0
r
P
MLC w h e r e E is t h e potential daily evapotranspiration and
p

ET = 0.5E n for P > 0 . 0 1 ;


d 0 <M r < 1.00 P is the d e p t h of rainfall (inches) that occurred on t h e
d

day in question.
Mi Evapotranspiration is equal to potential evapotran­
ET = 0.5£ for P > 0 . 0 1 ;
d M = 0,
r
P
MLC spiration as long as water is contained in t h e readily
(13.12) available zone a n d t h e n is r e d u c e d by t h e ratio of Μ,
Continuous Simulation Models 469

TIME
Figure 13.13 Results of simulation on Clemson 2 watershed.

to M L C . O n days when precipitation occurs, the evap­ 6-year record were used to obtain optimal p a r a m e t e r
otranspiration rate is reduced by a factor of 2 to values. T h e results were
account for cloudy conditions and low solar radiation.
D e e p seepage 5 or water that does not a p p e a r as /max =
0.30 (inches per h o u r )
streamflow within the watershed is determined from 5 = 0.185 (inches per day)
m a x

M L C = 3.75 (inches)
5 = 5 (13.13)
m a x
MLC' F = 0.40 (-).
where 5 is the maximum possible seepage rate in
m a x
T h e s e p a r a m e t e r s were used to simulate the entire
inches p e r day. A certain amount of return flow R F is
6-year period of record shown in Fig. 13.13.
allowed within the catchment and is calculated from
Figure 13.14 shows a scatter plot of observed versus
RF = F X 5, (13.14) predicted monthly streamflow. If the model were doing
a perfect job of prediction, all the points would fall on
where F is a constant defining the fraction of seepage the line of equal values, the correlation coefficient
that becomes runoff. would be 1.00, and the slope of t h e regression line
T h e total runoff R O is then equal to the sum of between observed and predicted flows would be 1.00.
surface runoff and the return flow Figure 13.14 d e m o n s t r a t e s a common procedure for
evaluating hydrologic models if observed data are
RO = SRO + RF. (13.15)
available. In using this comparison, it is important to
T h u s there are four model p a r a m e t e r s to be estimated: evaluate both t h e correlation coefficient and the slope
of the regression line since either a low correlation
5 , the maximum d e e p seepage rate
m a x
coefficient or a slope significantly different from 1.00
M L C , the soil water storage capacity of the less readily
can indicate a poor fit of the model. Neither measure,
available soil zone
when used alone, provides a d e q u a t e information since
F, the fraction of the seepage that becomes runoff
a slope of n e a r 1.00 and a low correlation coefficient or
/ , the maximum infiltration rate.
a correlation coefficient of near 1.00 and a slope very
m a x

Figure 13.13 shows a time series of t h e results of different from 1.00 could indicate poor modeling re­
applying this model to a 561-acre catchment n e a r sults. T h e slope provides a measure of the bias in the
Clemson, South Carolina. T h e first 2 years of the estimates m a d e by the model. T h e correlation coeffi-
470 13. Hydrologic Modeling

design decisions are m o r e sensitive to certain flow


quantities than others. T h u s what a p p e a r s to be a good
representation of a flow hydrograph may p r o d u c e a
substantially different design if the critical flow quan­
tity is not properly modeled. A n example would b e the
importance of modeling t h e persistence in low flows for
water supply storage requirement determinations.
T h e r e are several currently available continuous flow
simulation models. Continuous flow models are espe­
cially popular as a foundation for w a t e r quality models
because of t h e importance of a n t e c e d e n t conditions at
t h e time of a storm event.
T h e Stanford W a t e r s h e d Model (Crawford and
Linsley, 1966) and its many derivatives a n d t h e precipi­
tation—runoff modeling system ( P R M S ) of t h e U.S.
Geological Survey (Leavesley et al., 1983) represent
two popular continuous simulation models that can
g e n e r a t e daily streamflow hydrographs and storm flow
hydrographs.
S I M U L A T E D MONTHLY RUNOFF (in.) T h e s e and o t h e r continuous simulation models re­
Figure 13.14 Comparison of observed and simulated monthly quire n u m e r o u s computations in t h e form of water
runoff for Clemson 2 watershed. b u d g e t s for t h e various storages in the model. P a r a m e ­
ter estimation, if based on observed data, requires
running t h e model using a n u m b e r of different poten­
tial p a r a m e t e r sets, and selecting t h e p a r a m e t e r set
cient is a measure of how the model results track the that optimizes t h e desired objective function [i.e., mini­
observed results in a relative sense (model predictions mizes S of Eq. (13.4)]. P R M S has an option for a u t o ­
go u p or down when observed results go u p or down, matic p a r a m e t e r estimation based on Eq. (13.4). Expe­
respectively). rience with P R M S and o t h e r models has shown that
T h e r e is no generally accepted, simple criteria to use for some catchments, the model results are not very
in evaluating continuous simulation models. Evaluation sensitive to the values used for some of t h e p a r a m e t e r s .
of these models has always relied to a certain extent on This m a k e s estimation of these p a r a m e t e r s difficult—a
the judgment of the modeler based on comparisons of difficulty that may not be of importance unless some
observed and predicted hydrographs (such as Fig. 13.13) watershed modification or change over time whose
and based on how the model predicts in the face of hydrologic impact should b e reflected in t h e insensitive
extreme or changing conditions in comparison to the parameter(s) is being evaluated. This shortcoming is
modelers expectation. c o m m o n to all continuous simulation models, not just
Some quantitative comparisons that can be m a d e PRMS.
include means; variances; correlation between ob­ U s e of a model such as P R M S or S W M or o n e of its
served and predicted, standard error of estimate, a n d derivatives requires considerable time to b e c o m e famil­
first-order autocorrelations of flows integrated over iar with t h e model, to fully u n d e r s t a n d all of t h e
some finite time period such as 1 day or 1 month; flood internal workings of t h e model, and to c o m p r e h e n d t h e
peaks for various frequencies, such as the 5-year p e a k interactions and interrelationships of t h e various flow
flow; a n d / o r low-flow duration—frequency estimates c o m p o n e n t s . Before a model is selected, the user's
such as the 10-day, 2-year low-flow volume. A n o t h e r m a n u a l should be studied to d e t e r m i n e t h e adequacy
valuative approach is to compare design or operational of the d o c u m e n t a t i o n for the model and w h e t h e r the
decisions that would be m a d e based on model results model is suited for t h e intended p u r p o s e . A complex
and observed data. Such things as required storage model with i n a d e q u a t e d o c u m e n t a t i o n should not b e
capacity in a detention basin, storage to meet a p r o ­ used unless o n e is thoroughly familiar with the model.
jected water supply d e m a n d , or size of bridge opening T h e d o c u m e n t a t i o n should clearly state t h e capability
or culvert to meet a given criteria might be compared. and the limitations of the model and provide guidance
Certainly if the model-predicted flows agree very well on p a r a m e t e r estimation a n d model evalution. I n a d e ­
with observed flows, then design decisions based on q u a t e d o c u m e n t a t i o n is often a stumbling block to
modeled and observed flows should also agree. Certain p r o p e r and efficient use of a model.
Information Systems Technologies 471

DATA LAYERS REPRESENTATION OF OATA OVERLAY ANALYSIS


Μ MAP FORM Μ R A S T E R FORMAT
1 1 1 V3 T
333
1 1 11J 333
1 1 f 3 333
2 1 2 3 1 333
2 2 2 2 2 333
2 2 2 2 2 3 3,3
FOREST TYPES' 2 2 2 2 2 2 3 FOREST
2 2 2 2 2 2 3 3
FOREST

I 1 2 2 2 3
1 1 1 2 2 2 3
1
1 1 t 2 2 3
1 1 1 1 2 2 3
1 1 t 1 2 2 3
0 I t t 2 2 2 2
1 1 1 2 2 2 2
1 1 1 2 2 2 2 SOILS
SOIL T Y P E S
SOILS

aaacoaaa
aa
uuuouuau
(EQBOOGBQdQ
(EE
TOPOGRAPHY TOPOGRAPHY
TOPOGRAPHY

ANALYSIS R E S U L T S
REAL WORLD

Figure 13.15 A representation of GIS.

INFORMATION SYSTEMS TECHNOLOGIES general, have a m e a n s of encoding and converting


geographically (or spatially) oriented data and the ca­
Advances in microcomputer technology has b e e n pacity to retrieve and display the information through a
paralleled by developments in software that promise to variety of media including c o m p u t e r screens, printers,
revolutionize hydrologic and water quality modeling. and plotters.
T h r e e of these technologies are geographical informa­ Two basic techniques of representing spatial data
tion systems (GIS), expert systems (ES), and visualiza­ are employed. T h e first is a vector format where the
tion techniques (VT). basic unit of data is a single vector or m a p line. Vectors
form polygons that enclose like information within a
data layer. A second data model is a raster format
Geographical Information Systems
m a d e u p of cells or pixels. T h e raster format is com­
A characteristic of sophisticated, distributed p a r a m ­ monly called a grid-cell system. Figure 13.15 depicts
eter, hydrologic models is the large a m o u n t of spatially t h e basis of a raster G I S system. T h e figure shows how
oriented data that they require. G I S is a technology for the data layers are converted to cells having attribute
manipulating spatial data ( Z h a n g et aL, 1990a, b). D a t a values. T h e s e layers can then form the required data
of various types are collected in the form of layers. base for input into a hydrologic model or they can be
Typical layers might b e soils, topography, and land use. combined by overlaying techniques to generate the
T h e ability to extract, overlay, and delineate land char­ input data.
acteristics makes G I S eminently suited to the delin­ A hydrologic model can perform an analysis on the
eation of hydrologically homogeneous subareas. basis of the imported data and export results by cell
A complete G I S includes the h a r d w a r e and software back to the G I S . T h e G I S can then p r e p a r e visuals or
used to perform geographic analyses as well as the d a t a m a p s of the resulting hydrologic analysis. Z h a n g et aL
bases and the people who use the system to m e e t a (1990b) have interfaced a root-zone chemical transport
specific set of objectives (Brown, 1986). A G I S will, in model with a G I S to p r o d u c e m a p s showing the proba-
472 13. Hydrologic Modeling
bility of the applied chemical exceeding the health resulting anticipated flood flow elevations and areal
advisory limit of the U.S. Environmental Protection extent of flooding along streams draining the catch­
Agency at a depth of 1 m. m e n t are u n d e r development.
Most G I S are data managers and do not themselves Visualization technology will encompass G I S and
contain hydrologic modeling capability. Hydrologic hydrologic models along with high-speed parallel pro­
models must be interfaced with the G I S . T h e G I S cessing to c o m p u t e r g e n e r a t e the required graphic
manages data for the model. T h e model in turn pro­ information. This technology will b e useful in flood
vides the G I S with the results of an analysis that can be forecast centers and in explaining emergency opera­
m a d e into a G I S layer and mapped in various ways. tional plans to citizens. It may also help show citizens
G I S application will grow rapidly over the next sev­ t h e expected extent of flooding as a storm develops and
eral years as data bases become more readily available thus the likelihood of their particular property being
and the concepts and software become better under­ impacted by the flood.
stood.

Problems
Expert Systems
(13.1) Investigate m a t h e m a t i c a l techniques for
E x p e r t systems a r e a s u b s e t of "artificial finding the maximum or minimum of an objective func­
intelligence." E S are computer software that offer ad­ tion with respect to a single unknown p a r a m e t e r . Con­
vice to the software user based on its own store of sider both analytical and numerical approaches. Dis­
knowledge and the users response to a n u m b e r of cuss the applicability and relative merits of the ap­
if—then rules or questions. E S obtain their knowledge proaches to hydrologic modeling.
from the developers of the software and the users of (13.2) Same as Problem (13.1) but with multiple
the software. Some E S have a learning ability in that unknown p a r a m e t e r s .
they accumulate knowledge in response to the experi­ (13.3) (a) Write c o m p u t e r coding for the hydrologic
ence of their users. model shown schematically as Fig. 13.12. (b) Select a
E S are potentially valuable in hydrologic and water hydrologic record of at least 1 year in length from a
quality modeling as a means of selecting appropriate humid region catchment and estimate the p a r a m e t e r s
modeling approaches based on available data and of for this model, (c) Discuss quantitatively and qualita­
estimating the appropriate values for model p a r a m e ­ tively how well the model describes t h e hydrology of
ters. ES, when properly formulated and used, m a k e it the selected catchment.
possible for a novice to make modeling decisions simi­ (13.4) Discuss the merits of using t h e model de­
lar to those of a modeling expert. For this to h a p p e n , picted in Fig. 13.12 for evaluating the hydrologic im­
the software making u p the E S must have captured pact of forest clear cutting on stream hydrology for a
within it expert knowledge. Often this knowledge is 250-acre (100-ha) catchment. Include in your discus­
gained by interviewing experts and coding their re­ sion how the model might be used, your opinion as to
sponses to key questions in a form that captures the w h e t h e r the model would p r o d u c e reasonable results,
required knowledge. T h e best E S are based o n t h e and t h e hydrologic quantities (water yield, p e a k flow,
knowledge of more than one expert. etc.) that likely could and could not b e evaluated with
ES or any modeling tool should not be totally relied this model. W h a t aspects of the model would be the
upon. T h e user of any model, including those contain­ most important in this application? H o w are these
ing ES, has t h e ultimate responsibility for ensuring that important aspects reflected in t h e model in terms of
the modeling is d o n e correctly using appropriate model p a r a m e t e r s and model structure?
parameters. (13.5) Select a hydrologic model and a catchment.
Estimate the p a r a m e t e r s for the model and t h e catch­
m e n t . Select four of t h e p a r a m e t e r s of the model. Vary
Visualization Technology the values of the p a r a m e t e r s by 10, 20, and 5 0 % from
Visualization technology is the use of computer their estimated values, and run the model using these
graphics to e n h a n c e the understanding of computer- p a r a m e t e r values. Vary t h e p a r a m e t e r s individually.
generated results. Television weather forecasters m a k e Discuss the sensitivity of the p a r a m e t e r s with respect
good use of visualization technology when they show to hydrologic estimates that might b e m a d e with the
storm systems sweeping across the country and use this model.
visual impression to explain the weather forecast. Simi­ (13.6) D o Problem (13.5) except vary the p a r a m e ­
larly, techniques that will show the detailed movement ters simultaneously in pairs, triplicates, and all simulta­
of a severe storm across a catchment along with the neously.
Information Systems Technologies 473

(13.7) Select a hydrologic model. Discuss the basic (13.20) Describe desirable characteristics of a hydro-
structure of the model, the n u m b e r of p a r a m e t e r s , how logic model that is going to be used as a framework for
t h e p a r a m e t e r s can be estimated in the absence of a water quality model.
stream flow data, situations where the model could a n d (13.21) Describe at least two potential modeling ap­
could not be expected to produce reliable hydrologic proaches for generating runoff hydrographs from im­
estimates. pervious parking lots. W h a t are the advantages and
(13.8) P r e p a r e the computer coding for the basic disadvantages of each approach. Which approach do
model unit of Fig. 13.4. your prefer? Why?
(13.9) Apply the coding developed for Problem (13.22) Develop c o m p u t e r coding for one of the
(13.8) to a selected catchment of a r o u n d 50 acres models described for Problem (13.21). Test the coding
(20 ha). by simulating the runoff from a hypothetical parking
(13.10) How would the impact of a land-use change, lot.
such as surface mining on runoff hydrographs, b e re­
flected in the model depicted in Fig. 13.4?
(13.11) For a selected hydrologic model, discuss t h e
References
approach used to properly sequence the hydrology of
subwatersheds (i.e., discuss the model m a n a g e m e n t Beard, L. R. (1967). Optimization techniques for hydrologic engi­
neering. Water Resources Res. 3(3):809-815.
approach for combining and routing hydrographs).
Beasley, D . B., and Huggins, L. F. (1981). " A N S W E R S Users
(13.12) Select a particular catchment for which a
Manual," E P A - 9 0 5 / 9 - 8 2 - 0 1 . U.S. Environmental Protection
streamflow record is available. Without any reference Agency Region V, Chicago, IL.
to the streamflow, use a hydrologic model to estimate Box, G. E. P., and Tiao, G. C. (1973). "Bayesian Inference in
the hydrologic record for the same period as the avail­ Statistical Analysis." A d d i s o n - W e s l e y , Reading, MA.
able record. Discuss the difficulties encountered. Dis­ Brown, C. (1986). Implementation of Geographic Information Sys­
tem. What makes a new site a success? In "Proceedings, Geo­
cuss how well the estimated records resemble the ac­
graphic Information System Workshop, Bethesda, M D . " Am.
tual record of streamflow. Soc. Photogrammetry and R e m o t e Sensing.
(13.13) Use the available streamflow record of Prob­ Campbell, Κ. K., Johnson, H. P., and Melvin, S. W. (1974). Mathe­
lem (13.12) to improve the estimates of the model matical modeling of drainage watersheds, American Society of
p a r a m e t e r s and r e p e a t the estimation and discussion. Civil Engineers preprint 2373. Annual and National Environmen­
A r e the estimated flows more in a g r e e m e n t with tal Engineering Convention, Kansas City, M O .
Carrigan, P. H. (1973). Calibration of the U.S. Geological Survey
the observed flows after modifying the p a r a m e t e r s ?
Rainfall-Runoff Model for peak flow synthesis—Natural basins,
Why? U.S. Geological Survey Computer Report. U.S. Department of
(13.14) Discuss the similarities and differences Commerce NTIS.
among deterministic, parametric, and stochastic hydro- Crawford, Ν. H., and Linsley, R. K. (1962). Synthesis of continuous
logic models. U n d e r what conditions would each of streamflow hydrographs on a digital computer, Technical report
12. Stanford University Department of Civil Engineering, Stan­
these modeling approaches be the most a p p r o p r i a t e ?
ford, CA.
(13.15) Discuss the procedure that one might use to Crawford, Ν. H., and Linsley, R. K. (1966). Digital simulation in
verify the results of an application of an event-based hydrology: Stanford Watershed Model IV, Technical report 39.
hydrologic model (a) with a good record of streamflow Stanford University Department of Civil Engineering, Stanford,
and (b) in the absence of any streamflow data. CA.
(13.16) Discuss the procedure that one might use to Dawdy, D . R., and O'Donnell, T. (1965). Mathematical models of
catchment behavior. Am. Soc. Civil Eng. Proc. 91(HY4):123-127.
verify the results of an application of a continuous
D e B o e r , D . W., and Johnson, H. P. (1971). Simulation of runoff from
simulation hydrologic model (a) with a good record of depression characterized watersheds. Trans. Am. Soc. Agric.
streamflow and (b) in the absence of any streamflow Eng. 14(4):615-620.
data. DeCoursey, D . G., and Snyder, W. M. (1969). Computer oriented
(13.17) Describe the basic mathematical structures method of optimizing hydrologic model parameters. / . Hydrol.
9:34-56.
of a selected hydrologic model such as the S W M or
Diskin, Μ. H., and Simon, E. (1977). A procedure for the selection of
PRMS. objective functions for hydrologic simulation models. J. Hydrol.
(13.18) Discuss the advantage of objective p a r a m e ­ 34:129-149.
ter estimation based on a mathematical fitting criteria Edwards, D . R. (1988). Incorporating parametric uncertainty into
as compared to reliance on the judgment of the model flood estimation methodologies for ungaged watersheds and wa­
user. tersheds with short records, Ph.D. dissertation. Oklahoma State
University Library, Stillwater, OK.
(13.19) U n d e r what condition would p a r a m e t e r esti­
Edwards, D . R., and Haan, C. T. (1988). Confidence limits on peak
mation based on personal judgment be preferred over flow estimates for ungaged watersheds. In "Proceedings, Interna­
an objective mathematical fitting criteria? tional Symposium on Modeling Agricultural, Forest, and Range-
474 13. Hydrologic Modeling
land Hydrology." American Society of Agricultural Engineers, St. Leavesley, G. H., Lichty, R. W., Troutman, Β. M., and Saindon,
Joseph, MI. L. G. (1983). "Precipitation-Runoff Modeling System: User's
Edwards, D . R., and Haan, C. T. (1989). Incorporating uncertainty Manual," Water Resources Investigations Report 83-4238. U.S.
into peak flow estimates. Trans. Am. Soc. Agric. Eng. Geological Survey, Denver, CO.
32(0:113-119. Ross, Β. B., Contractor, D . N., and Shanholtz, V. O. (1979). A finite
Haan, C. T. (1972). A water yield model for small watersheds. Water element model of overland and channel flow for assessing the
Resources Res. 8 ( l ) : 5 8 - 6 8 . hydrologic impact of land use changes. / . Hydrol. 41:11-30.
Haan, C. T., and Johnson, H. P. (1968a). Hydraulic model of runoff Sorooshian, S. (1983). Surface water hydrology: On line estimation.
from depressional areas. I. General considerations. Trans. Am. Reu. Geophys. Space Phys. 21(3):706-721. [U.S. National Report
Soc. Agri. Eng. ll(3):364-367. to International U n i o n of Geodesy and Geophysics 1979-1982.]
Haan, C. T., and Johnson, H. P. (1968b). Hydraulic model of runoff Troutman, Β. M. (1985). Errors and parameter estimation in precipi­
from depressional areas. II. Development of model. Trans. Am. tation-runoff m o d e l i n g 1: Theory. Water Resources Res.
Soc. Agric. Eng. l l ( 3 ) : 3 6 8 - 3 7 3 . 2 1 ( 8 ) : 1 1 9 5 - 1 2 1 2 ; 2: C a s e Study. Water Resources Res.
Haan, C. T., Johnson, H. P., and Brakensiek, D . L. (1982). Hydro- 21(8):1214-1222.
logic modeling of small watersheds, A S A E Monograph 5. Ameri­ U.S. Army Corps of Engineers (1985). "HEC-1 Flood Hydrology
can Society of Agricultural Engineers, St. Joseph, MI. Package Users Manual." Hydrologic Engineering Center, Davis,
Huber, W. C , et al. (1981). "Storm Water Management Model users CA.
manual," Version III, EPA-600/2-84-109a. Environmental Pro­ U.S. Department of Agriculture, Soil Conservation Service (1973).
tection Agency, Cincinnati, O H . "Computer Program for Project Formulation Hydrology." Tech­
Huggins, L. F., and Monke, E. J. (1966). The mathematical simula­ nical Release 20, Washington, D C .
tion of the hydrology of small watersheds, Technical Report 1. Vicens, G. J., Rodriguez-Iturbe, I., and Shaake, J. C. (1975). A
Purdue University Water Resources Research Center. Bayesian framework for the use of regional information in hy­
Jackson, D . R., and Aron, G. (1971). Parameter estimation in hydrol­ drology. Water Resources Res. 11(3):405-414.
ogy: The state of the art. Water Resources Bull. 7(3):457-472. Wilson, Β. N., Barfield, B. J., and Moore, I. D . (1983). " A Hydrology
James, L. D . , and Burges, S. J. (1982). Selection, calibration and and Sedimentology Watershed Model. I. Modeling Techniques,"
testing of hydrologic models. In "Hydrologic Modeling of Small Special publication. Agricultural Engineering Department, Uni­
Watersheds" (Haan et a/., eds.), pp. 4 3 7 - 4 7 2 . A S A E Monograph versity of Kentucky, Lexington, KY.
5, American Society of Agricultural Engineers, St. Joseph, MI. Zhang, H., Haan, C. T., and Nofeiger, D . L. (1990a). Hydrologic
Klemes, V. (1982). Empirical and causal models in hydrology. In M o d e l i n g with G I S : A n Overview. Appl. Eng. Agric.
"Scientific Basis of Water Resources Management," pp. 9 5 - 1 0 4 . 6(4):453-458.
National Academy Press, Washington D C . Zhang, H., Nofziger, D . L., and Haan, C. T. (1990b). Interfacing a
Kuczera, G. (1983). Improved parameter inference in catchment root-zone transport model with GIS, A S A E Paper 903034 pre­
models 1: Evaluating parameter uncertainty. Water Resources Res. sented at the 1990 International Summer Meeting of A m . Soc.
19(5):1151-1162. Agr. Engrs., Columbus, O H , June 2 4 - 2 7 , 1990.
Appendix 2
Cumulative Standard Normal Distribution

475
476 Appendix 2

Appendix 2
Cumulative Standard
Normal Distribution

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09

-3.4 00003 0.0003 0.0003 0.0003 0.0003 0.0003 0.0003 0.0003 0.0003 0.0002
-3.3 0.0005 0.0005 0.0005 0.0004 0.0004 0.0004 0.0004 0.0004 0.0004 0.0003
-3.2 0.0007 0.0007 0.0006 0.0006 0.0006 0.0006 0.0006 0.0005 0.0005 0.0005
-3.1 0.0010 0.0009 0.0009 0.0009 0.0008 0.0008 0.0008 0.0008 0.0007 0.0007
-3.0 0.0013 0.0013 0.0013 0.0012 0.0012 0.0011 0.0011 0.0011 0.0010 0.0010
-2.9 0.0019 0.0018 0.0017 0.0017 0.0016 0.0016 0.0015 0.0015 0.0014 0.0014
-2.8 0.0026 0.0025 0.0024 0.0023 0.0023 0.0022 0.0021 0.0021 0.0020 0.0019
-2.7 0.0035 0.0034 0.0033 0.0032 0.0031 0.0030 0.0029 0.0028 0.0027 0.0026
-2.6 0.0047 0.0045 0.0044 0.0043 0.0041 0.0040 0.0039 0.0038 0.0037 0.0036
-2.3 0.0062 0.0060 0.0059 0.0057 0.0055 0.0054 0.0052 0.0051 0.0049 0.0048
-2.4 0.0082 0.0080 0.0078 0.0075 0.0073 0.0071 00069 0.0068 0.0066 0.0064
-2.3 0.0107 0.0104 0.0102 00099 0.0096 0.0094 0.0091 0.0089 0.0087 0.0084
-2.2 0 0139 0.0136 0.0132 0.0129 0.0125 0.0122 0.0119 0.0116 0.0113 0.0110
-2.1 0.0179 0.0174 0.0170 0.0166 0.0162 0.0158 0.0154 0.0150 0.0146 0.0143
-2.0 0.0228 0.0222 0.0217 0.0212 0.0207 0.0202 0.0197 0.0192 0.0188 0.0183
-1.9 0.0287 0.0281 0.0274 0.0268 0.0262 0.0256 0.0250 0.0244 0.0239 0.0233
-1.8 0.0359 0.0352 0.0344 0.0336 0.0329 0.0322 0.0314 0.0307 0.0301 0.0294
-1.7 0.0446 0.0436 0.0427 0.0418 0.0409 0.0401 0.0392 0.0384 0.0375 0.0367
-1.6 00548 0 0517 0.0526 0.0SI6 0 0505 0.0495 0.0485 0 0475 0.0465 0.0455
-1.5 0.0668 0.0655 0.0643 0.0630 0.0618 0.0606 0.0S94 0.03S2 0.0571 0.0559 I
-1.4 0.0808 0.0793 0.0778 0.0764 0.0749 0.0735 0.0722 0.0708 0.0694 0.0681
-1.3 0.0968 0.0951 0.0934 0.0918 0.0901 0.0885 0.0869 0.0853 0.0838 0.0823 I
-1.2 0.1151 0.1131 0.1112 0.1093 0.1075 0.1056 0.1038 0.1020 0.1003 0.0985
-1.1 0.1357 0.1335 0.1314 0.1292 0.1271 0.1251 0.1230 0.1210 0.1190 0.1170
-1.0 0.1587 0.1562 0.1539 0.1515 0.1492 0.1469 0.1446 0.1423 0.1401 0.1379
-0.9 0.1841 0.1814 0.1788 0.1762 0.1736 0.1711 0.1685 0.1660 0.1635 0.1611
-0.8 0.2119 0.2090 0.2061 0.2033 0.2005 0.1977 0.1949 0.1922 0.1894 0.1867
-0.7 0.2420 0.2389 0.2358 0.2327 0.2296 0.2266 0.2236 0.2206 0.2177 0.2148
-0.6 0.2743 0.2709 0.2676 0.2643 0.2611 0.2578 0.2546 0.2514 0.2483 0.2451
-0.5 0.3085 0.3050 0.3015 0.2981 0.2946 0.2912 0.2877 0.2843 0.2810 0.2776 I
-0.4 0.3446 0.3409 0.3372 0.3336 0.3300 0.3264 0.3228 0.3192 0.3156 0.3121
-0.3 0.3821 0.3783 0.3745 0.3707 0.3669 0.3632 0.3594 0.3557 0.3520 0.3483
-0.2 0.4207 0.4168 0.4129 0.4090 0.4052 0.4013 0.3974 0.3936 0.3897 0.3859
-0.1 0.4602 0.4562 0.4522 0.4483 0.4443 0.4404 0.4364 0.4325 0.4286 0.4247
-0.0 0.5000 0.4960 0.4920 0.4880 0.4840 0.4801 0.4761 0.4721 0.4681 0.4641
0.0 0.5000 0.5040 0.5080 0.5120 0.5160 0.5199 0.5239 0.5279 0.5319 0.5359
0.1 0.5398 0.5438 0.5478 0.5517 0.5557 0.5596 0.5636 0.5675 0.5714 0.5753
0.2 0.5793 0.5832 0.5871 0.S9I0 0.5948 0.5987 0.6026 0.6064 0.6103 0.6141
0.3 0.6179 0.6217 0.6255 0.6293 0.6331 0.6368 0.6406 0.6443 0.6480 0.6517
0.4 0.6554 0.6591 0.6628 0.6664 0.6700 0.6736 0.6772 0.6808 0.6844 0.6879
0.5 0.6915 0.6950 0.6985 0.7019 0.7054 0.7088 0.7123 0.7157 0.7190 0.7224
0.6 0.7257 0.7291 0.7324 0.7357 0.7389 0.7422 0.7454 0.7486 0.7517 0.7549
0.7 0.7580 0.7611 0.7642 0.7673 0.7704 0.7734 0.7764 0.7794 0.7823 0.7852
0.8 0.7881 0.7910 0.7939 0.7967 0.7995 0.8023 0.8051 0.8078 0.8106 0.8133
0.9 0.8159 0.8186 0.8212 0.8238 0.8264 0.8289 0.8315 0.8340 0.8365 0.8389
1.0 0.8413 0.8438 0.8461 0.8485 0.8508 0.8531 0.8554 0.8577 0.8599 0.8621
1.1 0.8643 0.8665 0.8686 0.8708 0.8729 0.8749 0.8770 0.8790 0.8810 0.8830
1.2 0.8849 0.8869 0.8888 0.8907 0.8925 0.8944 0.8962 0.8980 0.8997 0.9015
1.3 0.9032 0.9049 0.9066 0.9082 0.9099 0.9115 0.9131 0.9147 0.9162 0.9177
1.4 0.9192 0.9207 0.9222 0.9236 0.9251 0.9265 0.9278 0.9292 0.9306 0.9319

1.5 0.9332 0.9345 0.9357 0.9370 0.9382 0.9394 0.9406 0.9418 0.9429 0.9441
1.6 0.9452 0.9463 0.9474 0.9484 0.9495 0.9505 0.9515 0.9525 0.9535 0.9545
1.7 0.9554 0.9564 0.9573 0.9582 0.9591 0.9599 0.9608 0.9616 0.9625 0.9633
1.8 0.9641 0.9649 0.9656 0.9664 0.9671 0.9678 0.9686 0.9693 0.9699 0.9706
1.9 0.9713 0.9719 0.9726 0.9732 0.9738 0.9744 0.9750 0.9756 0.9761 0.9767

2.0 0.9772 0.9778 0.9783 0.9788 0.9793 0.9798 0.9803 0.9808 0.9812 0.9817
2.1 0.9821 0.9826 0.9830 0.9834 0.9838 0.9842 0.9846 0.9850 0.9854 0.9857
2.2 0.9861 0.9864 0.9868 0.9871 0.9875 0.9878 0.9881 0.9884 0.9887 0.9890
2.3 0.9893 0.9896 0.9898 0.9901 0.9904 0.9906 0.9909 0.9911 0.9913 0.9916
2.4 0.9918 0.9920 0.9922 0.9925 0.9927 0.9929 0.9931 0.9932 0.9934 0.9936

2.5 0.9938 0.9940 0.9941 0.9943 0.9945 0.9946 0.9948 0.9949 0.9951 0.9952
2.6 0.9953 0.9955 0.9956 0.9957 0.9959 0.9960 0.9961 0.9962 0.9963 0.9964
2.7 0.9965 0.9966 0.9967 0.9968 0.9969 0.9970 0.9971 0.9972 0.9973 0.9974
2.8 0.9974 0.9975 0.9976 0.9977 0.9977 0.9978 0.9979 0.9979 0.9980 0.9981
2.9 0.9981 0.9982 0.9982 0.9983 0.9984 0.9984 0.9985 0.9985 0.9986 0.9986

3.0 0.9987 0.9987 0.9987 0.9988 0.9988 0.9989 0.9989 0.9989 0.9990 0.9990
3.1 0.9990 0.9991 0.9991 0.9991 0.9992 0.9992 0.9992 0.9992 0.9993 0.9993
3.2 0.9993 0.9993 0.9994 0.9994 0.9994 0.9994 0.9994 0.9995 0.9995 0.9995
3.3 0.9995 0.9995 0.9995 0.9996 0.9996 0.9996 0.9996 0.9996 0.9996 0.9997
3.4 0.9997 0.9997 0.9997 0.9997 0.9997 0.9997 0.9997 0.9997 0.9997 0.9998
Appendix 3A
Rainfall Maps

477
478
479
480
481
482
483
Appendix 3B
Hydrologic Soil Groups

Aasted Β Alderwood Β Angie C Assumption Β Barney A


Aberdeen D Aldino D Angola C Astoria Β Barnhardt A
Abilene C Aleen C Ankeny A Athelwold Β Barnstead Β
Abington Β Alexandria Β Annandale Β Athena Β Barrancas D
Acadia D Alexis Β Anoka A Atherton D Barron C
Acme Β Alford Β Ansel mo Β *Athol Β Barronett C
Acton Β Algarrobo D Antanus C Atkins D Barth Β
Adair D Algiers D Anthony Β Atterberry Β Bartle D
Adams A Alicel Β Antigo Β Atwood Β Basher Β
Adamville C Allard Β Apache Β A u Cree C Bass Β
Ade A Alleghany Β Apakuie A Auburn C Baster Β
Adel Β Allen Β Apishapa C Auburndale D Bastrop Β
Adelanto Β Allendale C Apison C Augusta C Batavia Β
Adelphia Β Allenwod Β Applegate C Aurora C Bates Β
Adler C Allie C Appling D Austin Β Bath C
Adolph D Alligator D Arch Β Ava C Baudette Β
Afton D Allison C Archer C Avalanche Β Bayard A
Agar Β Allovers Β Arenzville Β Avery Β Bayboro D
Agate C Alma C Argyle Β Avon Β Bayside C
Agawam Β Almena C Ark C Avonburg D Baysmon Β
Agency C Almirante D Arkport Β Axtell D Beadle C
Agnew Β Almo D Arland Β Ayr Β Bear Lake D
Aguadilla A Almy Β Armagh D Babylon A Bear Prairie Β
Aguilita C Alonso C Armer C Baca Β Bearden C
Ahmeek Β Alpe C Armour Β Bagnell D Beardstown C
Ahnberg C Altamaha D Armuchee C Baigh C Beatty Β
Aiken Β Altamont C Arnold A Bainville Β Beaucomp C
Airmont Β Altavista Β Arnot C Baker C Beauford D
Akaka Β AltoC Arredondo A Balch A Beaumont D
Akan C Alton Β Artesia C Baldock Β Beauregard C
Akan C Altoona C Arvada D Baldwin C Beaver Β
Akaska Β Altura Β Arveson D Balfour Β Beaverhead Β
Akaska Β Altvan Β Arzell D Balm A Beaverton Β
Alachus Β Alvin Β Asa Β Balmorhea D Beckett C
Alaeola Β Alvira C Asboy Β Bancombe A Beckton D
Alama C Amalu D Asbury Β Bangor Β Beckwith C
Alamance C Amarillo Β Ascalon Β Banks A Bedford C
Alamosa C Amenia Β Ash Springs C Bannerville C Bedington Β
Albaton D Americua A Ashby C Barabook Β Beecher C
Albermarle* Β Amerlia C Ashe Β Barbour Β Beechy D
Albertville C Ames C Ashkum C Barbourville Β Belfast D
AlbiaC Amite Β Ashley C Barclay C Belfore C
Albion Β Amity C Ashton Β Barnard C Belgrade Β
Alcester Β Amsterdam Β Ashuelot C Barnes Β Belknap C
Alcoa Β Andover D Ash wood D Barneston A Bell D
Alden D Andrea Β Asotin Β Barnet D Belle Β

485
486 Appendix 3B

Bellinghouse D Blanket D Brazoe Β Buse Β Carrington Β


Belmont Β Blanton A Brecknock C Butler D Carrizo A
Beltrami Β Blencoe C Breece Β Butlertown C Carroll D
Beltsville C Blichton C Breese D Butte Β Carson D
Belvoir C Blockton C Bremo C Buxin D Carstairs C
Benefield C Blodgett Β Brennan Β Buxton C Carver A
Benevola Β Blomford Β Brennar C Byars D Carytown D
Benld D Bloomfield A Brenton Β Byrds D Casa Grande C
Bennington C Bloomington Β Bresser Β Cabinet C Cascade C
Benoit D Blount C Brewater C Cabo Rojo C Casco A
** Benson C Blue Earth D Brewer Β Cabot C Casey C
Bentonville C Bluffton D Brewster D Cacapon Β Cashmere A
Beotia Β Bluford D Briacoe Β Cacapon Β Cashton D
BergC Bobtail C Brianfield C Caddo C Cass A
Bergland D Bodine Β Brickton C Cagey D Cass Β
Berka C Bogota C Bridgehampton Β Caguas D Cassville Β
Berkley C Bohemian Β Bridgeport Β Cahaba Β Castana Β
Berkshire Β Bold Β Bridger Β Cajon A Castle D
Bermudian Β Bolivia Β Bridgeville Β Calais Β Castner C
Bernard D Bolton Β Briggs A Caldwell Β Catalina C
Bernardston C Bombay C Briggsdale C Calhoun D Catalpa C
Berrien Β Bo η accord D Brill Β Califon C Catano A
Berthoud Β Bonaparte A **Brimfield Calloway C Cataula C
Bertie C Bonham Β Brimley Β Calvert D Cathcart C
Bertolotti A Bonilla Β Brinkerton D Calverton C Catlett C
Bertrand Β Bonita D Brittain C Calvin C Catlin Β
Berwyn C Bonner Β Broadbook C Camaguey D Catoctin C
Bethany C Bonneville A Broadview C Camas A Catron D
Bethel D Bonnie D Brockport C Cambridge C Catskill C
Beulah Β Bono D Brooke C Camden Β Cattaraugus Β
Beverly A Bon pas Β Brookfield Β Cameron C Cave C
Bewleyville Β Boomer C Brookings Β Camillus C Cavode C
Bibb D Boone A Brooklyn D Camp Β Cavot Β
Bickleton Β Bordeaux Β Brookston Β Campo C Cavour D
Biddeford D Bosket Β Brooksville D Camroden C Cayagus c
Bienville Β Boswell D Broughton D ** Canaan C Caylor Β
Big Horn C Bow D Broward C Canadaigua C Cayucos D
Biggs A Bowamnville D Brownfield A Canadian Β Cayuga C
Biggsville Β Bowdoin D Browning Β Canadice C Cazenovia C
Billett A Bowdre C Brownlee C Canaserage Β Cecil Β
Billings D Bowie Β Bruno A Cane Β Celina Β
Binnaville Boyd D Bryce D Caneadea C Center C
Bippus Β Boyer Β Bub C Caneyville C Centerton Β
Birds C Boynton C Buchanan C Canfield Β Central A
Birdsall D Bozarth C Buckingham C Cannett D Chagrin Β
Birdsboro Β Bozeman C Buckland C Canoncito D Chalfont D
Birkbeck Β Braceville Β Buckley D Canyon C Chama Β
Birkhardt Β Bracken D Buckner A Cape Fear D Chamber D
Biscay D Brackett C Bucks Β Capeshaw C Chamberlino C
Bitterroot C Braddock Β Bucoda D Capron Β Chamokane Β
Blacklock D Braden C Bude C Captina C Chandler C
Blackwater D Bradenton C Bunkerville D Capulin Β Channahon Β
Bladen D Bradley Β Burchard Β Carbo C Chariton C
Blago D Brady Β Burdett C Cardiff Β Charleston C
Blain C Braham Β Burgess Β Cardinton C Charloe Β
Blair C Brallier A Burgin D Carey Β Charlotte D
Blairton C Brandon Β Burke Β Caribou Β Charlton Β
Blakeland A Brandywine C Burleson D Carisle D Chaseburg Β
Blakely Β Branford Β Burnham D Carlsborg A Chastain D
Blanc C Brashear C Burnside D Carlton Β **ChatfieldC
Blanchard A Bratton Β Burnsville Β Carnegie Β Chatsworth D
Blanco Β Braxton Β Burnt Fork Β Carnero C Chattahoochee
Blandford C Brayton C Burrell C Caroi Β Chauncey C
Blanding Β Brazito A Burton Β Caroline C Chehalis Β
Hydrologic Soil Groups 487

Chelsea A Cleman Β Conent Β Crofton Β Denrock D


Chemawa Β Clement C Conestoga Β Croghan Β Denson C
Chenango Β Cleora Β Conesus C Croom C Denton C
Cheney Β Clermont D Congaree Β Crosby C Depew C
Chenoweth Β Clifton Β Conley C Crossville Β Derby A
Cherette A Climax D Conottom Β Croton D Descalabrado Β
Cherokee D Clinton Β Conover Β Crow C Detour C
Cherry Β ClioC Conowingo C Crowder D Detroit Β
Cherryhill Β Clipper C Constable A Crowley D Dewart Β
Cheshire Β Clodine Β Continental C Crown Β Dewey Β
Chester Β Cloquallum C Conway C Crystal Β Dexter Β
Chesterfield Β Cloquet Β Cook D Culleoka Β Dick A
Chetek A Cloud D Cookeville Β Cullo C Dickey A
Chewacla C Clover Creek C Cookport C Culpeper Β Dickinson Fine
Chewelsh Β Clovia Β Coolidge Β Culvers C Sandy Loam A
Cheyenne Β Clyde C Coolville C Cumberland Β Dickinson Loam Β
Chichasha Β Clymer Β Cooney C Curran C Dickson C
Chiefland A Coamo C Cooper Β Curtis Β Dill Β
Chigley C Cobb Β Copaka Β Cushman Β Dilldown Β
Chilgren C Cochise Β Copas C Custer D Dillinger Β
Chilhowie D Cocoa Β Copeland C Cut Bank Β Dillon D
Chili Β Cody A Copelia C Cuthbert C Dilman
Chillisquaque C Coeburn C Coplay D Dade A Dimmick D
Chillum C Cogswell C Coral Β Daggett Β Disco Β
Chilmark C Cokedale C Corcege D Dakota Β Dishlo D
Chilo D Coker D Corduroy Β Dalbo Β Dixmont Β
Chipeta D Cokesbury D Corkindale C Dale C Dodgeville
Chippewa D Colbert D Corley C Dalhart Β Shallow Phase C
Chiricahua D Colby Β Cornutt D Dalton C Dodgeville D e e p
Choctaw Β Colden D Corvalis Β Dana Β Phase Β
Choptank A Coldwater D Corwin Β Dandridge C Doland Β
Choteau C Colebrook Fine Corydon C Daniela Β Dominguito D
Christian Β Colebrook Loamy Cossayuna C Danley C Dominic Β
Christiana C Fine Sand A Cotaco C Dannermors C Donerail C
Christianburg D Coleman Β Goto C Danvers C Donlonton C
Churchill D Colemantown D Cottonwood C Dariem C Dorchester Β
Cialee Β Colfax C Cougar C Darling C Dorsey D
Cialitoe C Colinas C Couparle D Darnell C D o s Cabezas Β
Ciane D Collamer Β Coupeville Β Darret Β Doty Β
Ciapus A Collington Β Course C Darwin D Dougherty Β
Cicero C Collins C Courtland Β Davidson Β Douglas Β
Cie Elum Β Collinsville C Courtney D Davie D Dover C
Cincinnati Β Colo C Cove D Dawes C Dowelton D
Cinebar A Coloao D Coveland C Dayton D Dowling D
Cintrona D Coloase A Coverytown C D e Soto C Downs Β
Clackamana C Coloma A Covington D Deachutes A Doylestown D
Claiborne Β Colonie A Cowden D Deary C Dragston C
Clallam C ColpD Cowiche Β Decatur Β Drake Β
Clarence D Colrain Β Cowling Β Deckerville D Dreeden Β
Clareville C Colta Neck Β Crago Β Decorrs Β Driping Springs D
Clarinda D Colton A Craig C Defiance D Drummer Β
Clarion Β Columbia Β Crandon Β * Dekalb Β Drummond D
Clark Fork Β Colville Β Crane Β D e l Rey C Drury Β
Clarksburg C Colwood D Craven C Delanco C Dryad C
Clarksdale C Colyar C Crawford D Delfina C D u Page Β
Clarkson C Comfrey C Creal D DellC Duane Β
Clarksville Β Comly C Creedmoor C Dellrose Β Dubbe Β
Clary Β Commerce C Cresbard C Delphi Β Dubois C
Clatsop D Comoro A Crescent Β Delpine D Dubuque Β
Claverack Β Compton Β Crestmore C Delray D Dubuque D e e p
Clawson C Conassuga C Crete D Demers C Phase Β
Clayton C Concord D Crevasse A Denham D Dubuque
Cleaver D Condit D Crider Β Dennis C Shallow Phase C
Cleburne Β Condon Β Crockett D Denny D Duffield Β
488 Appendix 3B
Duffy C Eldorado C Exline D Fort Pierce Β Gilt Edge D
Dukes A Elfort C Eylar D Fortuna D Ginat D
Dulac C Elfrida C Faceville Β Fox Β Gird Β
Dunbar C Elioak Β Fahey Β Foxhome Β Givin C
Duncan C Elk Β Fairfax Β Frankford D Glasagow C
Duncannon Β Elkina D Fairhaven Β Franklinton C Glenbar C
Duncom D Elkinsville Β Fairhope C Frankstown Β Glencoe D
Dundas C Elkton D Fairmount D Fraternidad D Glendale C
Dundee C Ellery D Fajardo C Frederick Β Glendive A
D u n e Sand A Elliber A Falaya C Fredon C Gleneig Β
Dunellen Β Elliott C Falcon Β Freehold Β Glenfield C
Dungeness Β Ellis D Falkner D Freeland C Glenford C
Dunham Β Ellison Β Fall Β Freeon Β Glenoma Β
Dunkirk Β Ellsberry C Fallbrook Β Freer C Glenville C
Dunlap C Ellsworth D Fallsburg C Fremont C Gloucester Β
Dunmore C Elmmert A Fallsington D Frenchtown D Godwin Β
Dunning D Elmo C Falun Β Freneau D Goessel D
Duplin Β Elmore C Fannin Β Frio Β Gogebic Β
Dupo C Elsinboro Β Fargo D Frost D Goldridge A
Dupont D Elsmere A Farland Β Fruita Β Goldsboro Β
Durant D Elswood C **Farmington C Frye C Goldston C
Durham Β Elwha C Farragut C Fullerton Β Goldvein C
Durkee C Emmet Β Farum C Fulton C Goliad C
* Dutchess Β Emory Β Fauguier Β Gage D Gooch C
Dutson C Empey C Faunce A Gainesville A Gore D
Duval Β Empeyville C Fawcett C Gale Β Gorua Β
Dwight D Enders C Faxon D Galen Β Goshen Β
Dwyer A Enfield Β Fayette Β Galestown A Gosort C
Dyke Β Englund D Fe D Gallatin D Gothard D
Easton C Ennis Β Felda D Gallion Β Gowen C
Eastonville A Enon C Felida Β Galveston A Grady D
Ebbe Β Ensenada C Fellowship C Galvin Β Graham C
Ebbert D Ensley D Fergua C Gann Β Grail C
Ebeys Β Enstrom Β Fidalgo C Gara C Granby D
Eckman Β Enterprise A Fillmore D Grande Roda C
Gardnerville C
Ector C Enuclaw Β Fincastle C Grant Β
Garfield D
Edalgo D Ephrata A Fitch A Grantsdale Β
Garner D
Eddy C Epping D Fitchville C Granville Β
Garrison A
Eden C Era Β Fitzhugh Β Grayling A
Garwin D
Edenton C Eram D Flamingo D Great Bend Β
Gasconade D
Edge D Erie C Flander C Greeham Β
Gaviota A
Edgeley C Ernest C Flandreau Β Greeley Β
Gayville D
Edgemont Β Escondido C Flanagan Β Green Bluff Β
Gearhart A
Edgewick A Eskin Β Flasher A Green River Β
Geary Β
Edgington C Espinosa Β Flathead Β Greenbush Β
Geer C
Edina D Esquatzel Β Fleetwood Β Greendale Β
Geiger D
Edinburg C Essex C Fletcher C Greenfield Β
Gem C
Edisto C Estacion Β Flint C Greenport C
Genesee Β
Edith A Estellin Β Flora D Greensboro Β
Genoa D
Edmonds D Estevan C Florence C Greenville Β
Genola Β
Edna D Estherville Β Florsheim D Greenwater A
Georgetown Β
Edneyville Β Esto C Floyd Β Greer C
Georgetown C
Edom C Etowah Β Fluvanna C Grenada C
Gerald D
Edwards D Ettrick D Foard D Grenville Β
Germania Β
EelC Eubanks Β Foley C Greybull C
Geronimo Β
Efland C Eufaula A Folsom C Greys Β
Gila Β
Egam C Eulonia C Fordney A Griffin C
Gilcrest Β
Egeland Β Eustis A Fordville Β Grimstad C
Gilead C
Eibbard C Eutaw D Fore D Groeclose C
Giles A
Elbert D Evans Β Forestdale D Groton A
Gilford D
Elburn Β Evendale C Forrest D Grove A
Gilligan Β
Elco Β Everett A Fort Collins Β Groveland Β
Gilman Β
Eld Β Everson C Fort Lyon Β Grover Β
Gilpin C
Eldon C Ewing A Fort M e a d e A Groveton Β
Gilson Β
Hydrologic Soil Groups 489

Grundy C Hartselle Β Hoble Β Humbarger Β Jonesville A


Guadalupe Β Harwood Β Hockley C Humeston C Joplin Β
Guanica D Hasel C Hoffman C Humphreys Β Josefa D
Guyayabo C Hasen Β Hogansburg Β Hunt D Josephine C
Guaysma C Haskill A Hoko C Huntaville Β Joy Β
Guckeen C Hassel D Holbrook D Hunters Β Juana Diaz Β
Gudrid Β Hastings C Holcomb D Hunters Β Judith Β
Guelph Β Hatchie C Holdrege Β Huntington Β Judson Β
Guernsey C Haven Β Hollad Β Hurst D Jules Β
Guin A Havre Β Hollinger C Hutchinson C Juliaette A
Guthrie D Haxtun A **Hollis Hyde D Juncoe C
Gypremort C Hayden Β Hollister C Hymon C Juniata Β
Habersham Β Hayesville Β Holloway Β Hyrley D Junius C
Haccke D Haymond Β Holly C Hysattsville Β Juno A
Hackers Β Haynie Β Hollywood D Iberia D Kaena D
Hackettstown Β Haytar Β Holmdel Β Ida Β Kahana Β
Hadley Β Heach Β Holston Β Idena C Kaines Β
Hagener A Heath Β Holt Β Ilion C Kalamzoo Β
Hagerstown Β Hebo D **Holyoke C Ilion C Kalihi D
HaigC Hector Β Homer C Illiopolis Β Kalispell Β
Haiku C Hedville C Hondo C Ima Β Kalkaska A
Hainea Β Heisler Β Honeoye Β Immokalee Kalmia Β
Halawa C Heitt C Honokas Β Imperial D Kaloko D
Haleakala A Helena C Honolua Β Ina C Kamnanui C
Halewood Β Hemmni C Honomanu C Independence A Kanab C
Half Moon C Hempstead Β Honoulfuli C Inglefield C Kanapaha Β
Halfway C Henderson D Hood Β Ingomar D Kaneohe Β
Halii C Hennepin Β Hoodsport A Inman C Kapapela Β
Haliimaile Β Henry D Hooker Β Inola D Karnak D
Hall Β Henshaw C Hoosic Β Ιο Β Karro Β
Halsey D Herando Β Hopewell C Iola A Kars Β
Hamburg Β Herbert Β Hopper Β Iona Β Kasota C
Hamilton Β Herkinser Β Hoquiam Β Ipeva Β Kasson C
Hamlin Β Hermiston Β Hord Β Irdell D Kato Β
Hammerly C Hermitage Β Hornell C Irion D Katy D
Hammond D Hermon Β Hortman C Irish D Kaufman D
Hampshire C Hermosa C Horton Β Iron River Β Kawaihae C
Hampton Β Herdon C Hosmer C Irurena D Kawsihapai A
Hanaka C Hero Β Houdexk Β Irvia D Keansburg D
Hanalei C Herrick C Houghton D Irving D Keating C
Hanceville Β Hershal Β Houlka D Irvington C Kedron C
Hand Β Hesseltin A Housetonic C Isagore C Keelakekua Β
Hanford Β Hesson C Houseville C Isanti D Keene C
Hanipoe Β Hialeah D Houston Black Β Iso Β Keith Β
Nannahatchee Β Hiawatha A Houston D Isolte Β Kelly D
Hanover Β Hibbing Β Houton C Isom Β Lelso C
Hanson Β Hickory C Hovde Β Issaquah Β Kelton A
Harbin Β Hicks Β Hoven D Istokpoga D Kempsville Β
Harbourton C Hidalgo Β Howard Β Iuka C Kempton Β
Harlem Β Hidewood C Howell Β Ivanhoe D Kenansville Β
Harley C Highfield Higley Β Hoye C Ives Β Kendaia C
Harlingen D Hiko Springs D Hoyleton C Jacana C Kendall Β
Harmon D Hildsboro Β Hoypus A Jackson Β Kennebec Β
Harmony C Hilgar Β Hubbard A Jacob D Kennedy Β
Harpeter Β Hilliard Β Huckabee A Jaffreyu A Kenney A
Harriet D Hillsdale Β Huckleberry C Jaucas A Kenspur Β
Harris D Hilo Β Hudson C Jeanerette C Kent D
Harrisburg C Hilton C Huey D Jefferson Β Keomah C
Harrison C Hinckley A Huff Β Jerome D Kerby Β
Harstine A Hinman D Huffine Β Jersuld D Keri C
Hartford Β Hiwassee Β Huggins C Jessup D Kerrtown Β
Hartland Β Hiwood A Hugo Β Joe Creek Β Kershaw A
Hartleton Β Hixton Β Huikau A Johnston D Kessau D
Hartsburg Β Hobble Β Humacao C Joliet C Kettle Β
490 Appendix 3B
Kettleman C Lagonda D Lawhorn D Lintonia Β Lynchburg C
Keyport C LaGrande C Lawrence C Lisbon Β Lynden A
Keysport C LaHogue Β Lawrenceville Lismas D Lynndyl C
Keystone A Lahontan D Clubbock C Lismore Β Lyons D
Kibbie C Laidig C Lawson Β Little Horn Β Lystair A
Kickerville Β Laidlaw A Lawton Β Littlefield D Mabi D
Kilauea A Laie D Lax C Littleton Β Machete C
Kilbourne A Lairdsville Β Leadvale C Litz C Mack Β
Kimbrough D Lajas C Leaf D Liverty C * Macomber c
Kinghurst Β Lake Charles D Leavenworth A Livingston D Macon D
Kings C Lake Creek Β Leavitt Β Llave C Madalin C
Kinrose D Lakehurst A Leavittville Β Lloyd Β Maddock A
Kipling D Lakeland A Lebanon D Lobdell Β Maddox Β
Kipp Β Lakemont C LeBar Β Lobelville C Madison Β
Kipson D Lakeville Loam Β Leek Kill Β Lockhard C Madras C
Kirkland D Lakeville Sandy Lee Β Lockhart Β Madrid Β
Kirvin C Loam A Lee D Lockport C Maginnis C
Kistler Β Lakewood A Leeds Β Locust C Magnolia Β
Kitsap C Lakin A Leel Β Lodi Β Mahaska Β
Kittitas Β LaLande Β Leetonia Β Lofton C Mahnomen Β
Kittson Β Lamington D Legore C Logan C Mahoning D
Kiwanis Β Lamont Fine Lehigh C Logendale C Maiden C
Klaberg C Sandy Loam A Leicester C LoloB Maile Β
Klamath C Lamont Loam Β Lela D Lomax Β Makalapa D
Klaus A Lamonta C Lempster D Lone Rock Β Makawso Β
Klej Β Lamoure C Lena D Lonepine Β Makena Β
Kline A Lamson C Lenoir Fine Longford C Malago Β
Klinesville C Lanark Β Sandy Loam Β Longrie C Malays C
Knappa Β Lancaster Β Lenoir Fine Lonoka Β Maleza Β
Knight C Land C Sandy Loam C Lookout C Mamala C
Know Β Landes Β Lenox C Loon A Manalapan D
Koch C Landisburg C Leon C Lorain C Manana C
Koehler Β Lane C Leona D Lordstown C Manassa Β
Kohala Β Lane C Leonardtown D Lorella C Manassas Β
Kokokahi D Langford C Leota C Lorenso A Manatash C
Kokomo D Langley b Leshara Β Loring Β Manatee D
Konokti A Lanham D Lester Β Lorsdale Β Manchester A
Koolsu D Lanrgrell Β LeSueur Β Los Guineoa D Mangua C
Kopish D Lansdale Β Letcher D Los O s o e C Manhattan Β
Kosmos D Lansdowne C Letort Β Loudon C Manhelm C
Koster C Lansing Β Levan A Loudonville C Manor Β
Kranzburg Β Lantz D Lewisberry Β Louisa Β Mansfield D
Krause Β Lapine A Lewiston C Louisburg Β Mansic C
Kreamer C Lapon D Lewisville Β Loup D Mansker Β
Kresson C Laporte D Lexington Β Lowell C Mantachie C
Krum C Laredo Β Lick Β Loy D Manteo D
Kukaisu Lares C Lick Creek Β Loysville C Manvel Β
Kunia Β Largent C Lickdale D Lualualei D Maple D
Kutztown C Largo C Lightning D Lucas C * * Mapleton c
La Brier C Larimer Β Lignum C Lucien C Marble A
La Palma Β Larkia Β Lihen A Ludlow C Marcus D
La Prairie Β Larry D Likes A Lufkin D Marcy D
La Rose Β Las Animas A Lima Lukin C Mardin C
La Verkin Β Las Casa C Limerick C Lumni C Marengo C
LaBelle C Las Lucas D Lincoln A Lun C Marhsall Β
Labette C Las Piedras D Lindley C Lunt C Mariana C
Labounty C Las Vegas D Lindsborg D Lupton D Marias D
Lacamas D Lashley Β Lindside C Lura D Marina A
Lackawanna Β Lassen C Linecroft A Luray C Marion D
Ladd C Latah C Linganore C Luton D Marissa C
LaDelle Β Lauderdale C Link Β Luverne Β Markland C
Ladoga C Laurel Β Linker Β Luzena D Marksboro Β
Ladysmith D Lauren A Linneus Β Lycoming C Marlboro Β
Lafe D Laveen Β Lino C * Lyden Marietta Β
Hydrologic Soil Groups 491

Marlow C Mench Β Montara D Navajo D Nosbig D


Marlton C Menfro Β Montell D Navesink Β Nuby C
Mama D Menlo D Monteola D Navesota D Nuckolls C
Marquette Β Mentor Β Montesano Β Naylor C Nueces A
Martha C Mercedita D Montevallo C Nebish Β Nunda C
Martin Pena D Mercer C Montgomery D Neble A Nunn C
Martinsdale Β Mereta D Monticello Β Needmore C Nutley D
Martinton C Meridian Β Montoya D Negley Β Nymore A
Masada Β Meros A Moody Β Nehalem Β Oahe Β
Mason Β Merrimac Β Moreau D Nellie Β Oakford Β
Massena C Mertz Β Moree D Neosho D Oakland Β
Massillon Β Mesa Β Morley C Neptune A Oasie C
Matansas C Meskill D Morman Mesa Nereson Β O'Fallon D
Matapeake Β Metes Β Moro Cojo A Neshaminy Β O'Neill Β
Matawan C Methow Β Moro May D Nesika Β Ochlockonee Β
Matlock D Metolum A Morocco C Nester D Ochopee D
Matmon C Mexico D Morrill Β Neubert Β Ockley Β
Matney Β Mhoon C Morris C Nevada D Oconee C
Mattapex C Miami Β Morrison Β Neville Β Odessa C
Maumee D Micres A Morrow C N e w Cambria C Odin C
Maunabo D Middlebury Β Morton Β Newark C Ogemaw Β
Maury Β Midland D Moscow Β Newart Β Okav D
Maverick D Mifflinburg Β Moshannon Β Newberg Β Okeechobee D
May Β Miquel D Mossyrock A Newbery C Okeelanta D
Mayhew D Milaca Β Mottsville A Newfane Β Okemah C
Maynard Lake Β Milam Β Mount Carroll Β Newkirk Β Okenee D
Mayo Miles Β Mount Lucas C Newport Β Okoboji Β
Mayodan Β Milford c Mountview Β Newton D Oktibbeha D
Maytown C Mill Creek Β Mucars Β Newtonia Β Olequa C
Mazeppe Β Millbrook Β Muir Β Nicholson Β Olinda Β
McAfee Β Miller D Muirkirk Β Nicholville Β Olivier C
McAllister C Millington Β Mukilteo A Nickel D Olmitz Β
McBride Β Millsdale Β Mullins D Nicollet Β Olmsted C
McDonald C MiloD Munising Β Niles C Olsa Β
McDowell Milroy D Munuacong D Ninigret Β Olympic Β
McEwen C Mimosa C Murrill Β Ninrod C Omaha Β
McGary C Minatare D Muscatine Β Nipe Β Omega A
McKamie C Minco Β Muse Β Nisqually A Ona C
McKay D Mineola Β Muskingnum C Nixa Onalaska Β
McKenna Β Miner D Muskogee C Nixon Β Onamia Β
Mckenzie D Minnequa Β Musselshell Β Nixonton Β Onarga Β
McLain C Minora C Myatt D Nobescot A Onaway Β
McMurray A Minvale Β Myersville Β Noble Β Ondawa Β
McNeal C Miota D Naches Β Nodaway Β Oneida C
McPherson D Mission C Nacimiento C Nogalee C Onslow Β
Meadin A Mitchell Β Macogdoches C Nohili D Ontario Β
Meadowville Moca D Naiwa Β Nolan C Ontonagon C
Meadville Β Modale C Nakelele Β Nolichucky Β Onyx Β
Mecklenburg C Moenkopie C Nantucket C Nolo C Ookala Β
Meda Β Moffat Β Nanum Β Nonopahu C **Oquaga C
Medary C Mohave C Napa D Nookachamps D Oquawka A
Medford C Mohawk Β Napier Β Nooksack C Ora C
Medina Β Moiese Β Nappanee D Nora Β Oracle Β
Medio C Moira C Naranjito Norden Β Orange D
Meeteetee C Mokena C Narcisse Β Norfolk Β Orangeburg Β
Mehlhorn C Molena A Narragansett C Norge C Orcas A
Meigs C Molina D Nasel C Norma C Orchard Β
Melbourne C Monarda C Naslehu Β North Powder C Ordway D
Mellenthin Β Monee D Nason C Northport C Orelia D
Melrose C Monmouth C ** Nassau C NorthumberlndC Orells D
Mervern C Monona Β Natalie C Northville D Orient Β
Melvin D Monongahala C Natches Β Norton Β Orienta Β
Memphis Β Monroeville C Natchitoches D Norwich D Orio C
Menahga A Montalto Β National A Norwood Β Orion C
492 Appendix 3B
Orlando High Patit Creek Β Pittstown C Puna Β Reparada D
Phase A Patoutville C Pittwood Β Purdy D Retsof C
Orlando Low Patrick Β Placentia D Purgatory D Rex C
Phase Β Patton C Plainfield A Puu O o Β Reynolds Β
Orman D Paulding D Piano Β Puu Pa Β Rhinebeck C
Orrville C Pauwela C Plata C Puysllup Β Rhoades D
Ortello A Pawlet Β Platea C Quamba D Richfield Β
Orting C Pawnee D Platner C Quandahl Β Richland C
Osage D Paxton C Plattsmouth Β Quay C Richview C
Osceola D Paymaster Β Plattville Β Quicksell C Richwood Β
Oshawa D Payne D Pledger D Quincy A Ridgebury C
Oshtemo Β Peace River D Plummer D Quinlan Β Ridgely Β
Osmund Β Peacham D Plymouth Β Quonset A Ridgeville Β
Oso A Pearman Pocomoke D Raber C Riesel D
Ostrander Β Pearson C Podunk Β Rabun Β RifTe A
Otero Β Pecetonics Β Poinsett Β Racine Β Riga C
Othello D Pedernales C Poland Β Racoon D Riggs D
Otisville A Pekin C Poison C Radford Β Riley Β
Otsego C Pelan Β Pomello A Radnor D Rillito Β
Ottawa A Pella Β Pompano D Ragnar D Rimer C
Otter D Pence Β Pomroy Β Rago C Rinard D
Otterhold Β Penn C Pope Β Raina D Ringling C
Ottokee A Pennington Β Port Β Rainbow C Ringold Β
Otway D Penoyer C Port Bryon Β Ralston Β Rio Arribe D
Ovid C Penrose C Portales Β Ramona C Rio Canas C
Owaneco D Pen wood A Porters Β Ramsey Β Rio Lajas A
Owen Creek C Peoh C Portland D Randall D Rio Piedras D
Owens D Peona C Portsmouth D Ranger C Ritchey Β
Ozona C Peotone C Portugues D Rankin C Rittman C
Pace Β Pequea C Poskin C Rantoul D Ritzville Β
Paden C Perkinsville Β Poso Blanco C Rapidan C Riverside A
Page D Perks A Post D Rarden C Riverton C
Pahranagat C Perrine D Potamo D Raritan Β Roane C
Pahroc D Perry D Potter C Raub Β Roanoke D
Paia Β Persayo D Pottsville D Rauville D Robbe D
Painesville Β Pershing C Poultney Β Ravalli D Robertsville D
Paiute Β Peru C Poverty C Ravenna C Robinsonville Β
Palestine Β Peshastin A Powder Β Ravola Β Roby C
Paletine C Patoskey A Powell C Ray Β Rockaway Β
Palm Beach A Petrolis D Poygan D Rayne Β Rockbridge Β
Palmas Atlas D Pettis Β Prather C Reagan C Rockdale D
Palmdale C Pewbroke Β Pratt A Reaton Β Rockmart C
Palmyra Β Pheba C Prentiss C Reaville C Rockport Β
Palouse Β Phelpe Β Prescott D Rebuck C Rockton Β
Palso C Phillips D Presque Isle Β R e d Bay Β Rockwood Β
Pandura Β Philo C Preston A Red Hook C Rocky Ford Β
Panton D Piagah Β Prewitt C Redfield Β Rodman A
Papago Β Picacho D Prieta Β Redington C Roe Β
Papakating D Pickford D Princeton Β Redlands Β Roebuck D
Papineau C Picksway C Pring Β Redmond Β Rogers D
Parishville C Pickwick Β Proctor Β Reed D Rohrersville D
Parkdale A Pierce Β Progresso Β Reeser C Rokeby D
Parker Β Pierre D Promise D Reeves C Rolfe C
Parks Β Pilchuck A Prospect Β Regent C Romeo C
Parkwood C Pilot Β Prosser Β Regnier D Rosario C
Parnell D Pilot Rock C Providence C Reinach Β Roscoe C
Parr Β Pilhonua Β Provo Β Reliance C Roscommon D
Parsons D Pima C Prowers Β Renfrow D Rosebud Β
Pasco Β Pinal D Ptarmingan Β Reno D Rosedell D
Pasloe Β Pinckney C Puget Β Renohill C Roselms D
Paso Seco C Pinones D Puhi Β Renova Β Rosemount Β
Pasquotank D Pinson Β Pulaski Β Renshaw Β Roseville Β
Pasuhau Β Pintura A Pulehu A Renslow Β Rositas A
Patent C **PittsfieldC Pullman D Rentide C Rosyln C
Hydrologic Soil Groups 493

Ross Β Sassafras Β Shoshone Β St. Clair D Sunsweet C


Rosschi D Sauble A Shouna Β St. Helens A Superstition A
Rossmoyne C Saugstuck C Shrewsbury D St. Joe Β Surry Β
Rosulus D Sauk Β Shubuta C St. Johns D Susquehanna D
Round Butte D Savage C Shuvah C St. Lucie A Sutherlin C
Routon D Savannah C Sicily Β St. Marys C Sutphen D
Rowe D Sawaill C Sidell Β St. Paul C Sutton Β
Rowland C Sawyer C Sierrs C Staatsburg C Swaim C
Rowley C Saybrook Β Sifton Β Stacum D Swanton C
Rox Β Scandia Β Signal C Stambaugh Β Swantown A
Roy Β Scantic C Siler A Stamford D Swartswood C
Royalton C Scarboro D Silerton Β Stanfield C Sweden Β
Roza D Schapville C Silver Creek D Stanton D Sweeney C
Rozetta Β Schoharie C Sima D Starks C Sweetwater D
Ruark C Schooley D Simcoa Β Starr Β Swims Β
Rubicon A Schumacher C Simla Β Staser Β Switzer D
Rubio C Schuylkill Β Sinai C State Β Swygert C
Rucker Β Scio Β Sinclair Β Steekee C Sylvan Β
Rudyard C Sciotoville C Singsaas Β Steinauer Β Symerton Β
Rumford Β Scipio D Sioux Β Steinsburg Β Tabernash Β
Rumney C Scituate C Sipple Β Stendal C Tabler D
Rupert A Scobey Β Siskiyou Β Stephensburg C Tabor D
Rushtown A Scott D Sites C Stephenville Β Taft C
Rushville D Scott Lake Β Skaggs C Stetson Β Tohoe C
Ruskin C Scowlale Β Skagit Β Stevenson Β Talanta D
Russell Β Scranton C Skalkho Β Steward D Talbott D
Russellville C Searing Β Skamanis Β Stidham Β Talcot D
Ruston Β Seaton Β Skames C Stimmon Talihina D
Rutlege D Sebeka D Skerry C Stissing C Talladega C
Ryder C Sebewa D Skiyou Β Stockbridge C Tallula Β
Sabana C Sebring D Skokomish A Stockland Β Tally Β
Sabana Seca D Sedan C Skyberg C Stockton Β Taloka D
Saco D .Sediu C Skykomish A Stoneham Β Tama Β
Saffell Β SegalD Sleeth C Stonington Β Tammas C
Sage D Segno Β Sloan D Stono C Tanama C
Sagemoor C Selah C Slocum D Stookey Β Tanberg D
Salal Β Selkirk D Smoky Butte C Storden Β Taneum C
Salem Β Selle A Smolan C Story C Tanwax A
Salemsburg Β Selma Β Snow Β Stough C Taos Β
Salisbury D Semiahmoo C Soda Lake Β Stoy D Tarrant D
Salix Β Sequatchie Β Sodus C Strasburg C Tate Β
Salkum D Sequois C Soga C Strauss C Tatum Β
Salmon Β Serrano D Soller D Strawn Β Taylor C
Salol D Sewtooth C Solomon D Stronghurst Β Teague D
Saltillo C Sexton D Somers Β Stryker D Teas C
Saluvia C Seymour C Somerset Β Stukal C Tedrow A
Salvisa D Shannon Β Sonoita Β Stumpp D Teja C
Samish D **ShapleighC Sontag D Sudbury Β Tell Β
Sammamish Β Sharkey D Souva D Suffield C Teller Β
Sams Β Sharon Β Sparta A Sula Β Tellico Β
Sharpeburg Β Spearfish Β Sulphurs C Tenino A
San Antonio D
Shavano Β Spencer Β Sultan Β Tepee D
San German C
Shelburn C Sperry C Sumas C Teresa D
San Jose Β
Shelby C Spile D Sumit C Terril Β
San Josquin D
Shelbyville c Spooner C Summerville C Terry Β
San Juan Β
Shelmadine c Spottswood Β Sumner A Tescott Β
San Saba D
Shelocata Β Spring Creek D Sumter D Teton Β
Sango C
Shelton Β Spring D Sun D Tetonka C
Santa C
Sheppard A Springer A Sunbury Β Thackery Β
Santa Clara D
Sheridan Β Springfield D Suncook A Thatuna C
Santa Isabel D
Sherman Β Springtown Β ** Sunderland C Thayer Β
Santa Lucia C
Shiloh C Spur Β Sunniland C Thomasville C
Santiago Β
Shoals C St. Albans Β Sunnyside Β Thompson A
Sargeant D
Shook Β St. Charles Β Sunrise Β **Thorndike C
Sarpy A
494 Appendix 3B
Thornton D Travis C Usine D Wagner D Waverly D
Thornwood A Treadway D Utica Β Waha C Wayland C
Thoroughfare A Trego C Utuado Β Wahee C Wayne Β
Thorp C Trempealeau Β Uvalde Β Wahiawa Β Waynesboro Β
Thurman A Trenary Β Vader Β Wahtum C WeaB
Thurmont Β Trent Β Vaiden D Waialua Β Weaver C
Thurston Β Trexler C Vale Β Waikaloe Β WebbC
Tiburones D Trinity D Valentine A Waikapu Β Webster C
Tice C Tripp Β Valers C Wailes Β Weeksville Β
Tickfaw D Tromp Β Vallecitos C Waimanalo D Wehadkee D
Tijeras Β Trout River A Valois C Waimea Β Weikert C
Tilden C Trowbridge Β Vance C Waipahu C Weinbach C
Tillman Β Troxel Β Vandalia C Waiska Β WeirD
Tilsit C Troy C Vanderville C Waits Β WeldC
Timmer Β Truman Β Vanoss Β Wakeland Β Weller D
Timmerman A Trumbull D Varna C Wakonda C Wellington D
Timpahute D Tuacumbia D Vaucluse C Walla Walla Β Wellman C
Timula Β Tubac D Vayma D Wallace Β Wellsboro C
Tinton Β Tucker D Vebad Β Waller C Wemple Β
Tiocano D Tucumcari Β V e g a Alta C Wellington C Wenaa C
Tioga Β Tuffit Β Vega Baja D Wallkill C Wenatchee Β
Tippah D Tughill D Vekok C Wallpack Β Wesley C
Tipparary A Tujunga A Velma Β Walpole C Wessington Β
Tippecanoe Β Tuller D Venango D Walsh Β West Point D
Tipton Β Tully C Venedy D Walters D Westbury C
Tirah Β Tumacacori Β Verdel D Walton C Westfall C
Tisbury Β Tumbez D Verdigris Β Wampeville Β Westland D
Tisch C Tumwater A Verdun D Wann A ** Westminister C
Tishominto C Tunica D Vergennes D Wapato C Westmoreland C
Tiston Tofton Β tunkhannock Β Verhalen D Wapping Β Weston C
Titusville C Tupelo D Vernon D Ward D Westphalia Β
Tiwoli A Turbotville C Verona Β Warden Β Westport A
Toa C Turis C Veyo D Warman D WestviHe C
Tobin Β Turkey Creek Β ViaC Warne D Wethersfield C
Tobosa D Turnbow D Vicksburg Β Warners C Weymouth C
Todd Β Turner Β Victor Β Warrenton A Whalan Β
Toddville Β Turnerville C Victoria D Warrior C Wharton C
Tokul A Tuscan Β Vienna Β Warsaw Β Whatcom D
Toledo D Tuscarora C Vilas A Wartrace C Whately D
Tolley Β Tuscola Β Viola D Warwick Β Wheeling Β
Tolo Β Tusquites Β Vira Β Washjburn D Whidbey A
Toltec Β Tuxedo D Virden C Washington Β Whippany C
Tombigbee A Twin Creek Β Virgil Β Washoe C White H o u s e C
Tonawanda C Twin Lakes Β Virgin River D Washougal C White Store D
Tongue River C Two Dot Β Virtue C Washtenaw C White Swan C
Tonopah Β Tyler C Vista Β Wassaic C Whitefish Β
Toppenish C Udolpho C Viston A Wassuk D Whiteford C
Topton Β Uinta Β Vivee C Wagauga Β Whitelaw
Torres C Ulen Β Vivi Β Watchaug Β Whitesburg C
Tortugas D Ulm Β Vlassaty C Waterboro D Whiteson D
Tours C Ulupalakum Β Volga D Wateloo C Whitetail Β
Toutle A Ulysses Β Volin Β Waterville C Whitlock A
Tower D Umapine Β Volke C Watsaka A Whitman D
Townsbury Β Una D Volney Β Watson C Whitson D
Townsend Β Unadilla Β Volperite C Watsonvile D Whitwell C
Toxaway D Uncompahgre Β Voluaia C Watt D Wibaux C
Toyah Β Ungers Β Vona A Watton C Wibaux C
Toyey Β Union C Vrooman Β Waubay Β Wichita C
Traer C Unison Β Wacoasta C Waugh C Wickersham C
Trailer Β Unity A Wade D Waukeesha Β Wickham Β
Transylvania Β Upehur C Wadell D Waukegan Β Wickiup A
Trapper Β Urbana Β Wadena Β Waukon Β *Wilbraham C
Travessilla Β UrboD Wadesboro Β Waumbek Β Wilcox D
Travessilla D Ursula D Wadsworth D Wausson D Wildwood D
Hydrologic Soil Groups 495

Wilkee C Winema Β Woodbridge C Wurtsboro C Zaca D


Wilkeson Β Winfield C Woodglen D Wykoff Β Zahl Β
Will D Wingville Β Woodinville Β Wynoose D Zaleeki Β
Williamette Β Winifred D Woodlyn C Xenia Β Zaneis C
Willard C Winlock D Woodson D Yabucoa D Zapeta D
Williams Β Winnett D Woodstown C Yadkin Β Zell Β
Williamsburg C Winooski Β Woodward Β Yahola Β Zimmerman A
Williamson Β Winslow Β Wooster Β Yakim A Zion C
Willoughby C Winston A Woostern Β Yale C Zipp C
Willow Creek Β Winterset C Worland Β Yauco D Zita Β
Wilson D Witt Β Worsham D Yeoman Β Zook D
Winchester A Wolcottsburg C Worth C Yoder Β Zuber Β
Wind River A Woldale D Worthen Β Yonaba C Zwingle D
Windom Β Wolf Β Worthington Β Yordy Β
Windsor A Wolfever C Wortman C YorkC
Windthorst C Wood River D Wrightsville D Yunes D
Appendix 3C
Runoff Curve Numbers

497
498 Appendix 3C
Table 3.2a Runoff Curve Numbers for Urban Areas 0

Curve numbers for


Cover description hydrologic soil group

Average percent
Cover type and hydrologic condition impervious area* A Β C D

Fully developed urban areas (vegetation established)


Open space (lawns, parks, golf courses, cemeteries, etc.) c

Poor condition (grass cover < 50%) 68 79 86 89


Fair condition (grass cover 50% to 75%) 49 69 79 84
Good condition (grass cover > 75%) 39 61 74 80
Impervious areas
Paved parking lots, roofs, driveways, etc.
(excluding right-of-way) 98 98 98 98
Streets and roads:
Paved; curbs and storm sewers (excluding
right-of-way) 98 98 98 98
Paved; open ditches (including right-of-way) 83 89 92 93
Gravel (including right-of-way) 76 85 89 91
Dirt (including right-of-way) 72 82 87 89
Western desert urban areas
Natural desert landscaping (pervious areas only)** 63 77 85 88
Artificial desert landscaping (impervious weed
barrier, desert shrub with 1- to 2-inch sand or
gravel mulch and basin borders) 96 96 96 96
Urban districts
Commercial and business 85 89 92 94 95
Industrial 72 81 88 91 93
Residential districts by average lot size
1/8 acre or less (town houses) 65 77 85 90 92
1/4 acre 38 61 75 83 87
1/3 acre 30 57 72 81 86
1/2 acre 25 54 70 80 85
1 acre 20 51 68 79 84
2 acres 12 46 65 77 82
Developing urban areas
Newly graded areas (pervious areas only, no vegetation) 77 86 91 94
Idle lands (CN's are determined using cover types
similar to those in Table 3.2c).

Note. Source: Soil Conservation Service (1986).


a
Average runoff condition, and I = 0.2S.
a

*The average percent impervious area shown was used to develop the composite CN's. Other assumptions are as
follows: impervious areas are directly connected to the drainage system, impervious areas have a CN of 98, and
pervious areas are considered equivalent to open space in good hydrologic condition. CN's for other combinations of
conditons may be computed using Eq. 3.24.
c
C N ' s shown are equivalent to these of pasture. Composite CN's may be computed for other combinations of open
space cover type.
^Composite C N ' s for natural desert landscaping should be computed using Eq. 3.24 based on the impervious area
percentage (CN = 98) and the pervious area CN. The pervious area CN's are assumed equivalent to desert shrub in poor
hydrologic condition.
'Composite CN's to use for the design of temporary measures during grading and construction should be computed
using Eq. 3.24, based on the degree of development (impervious area percentage) and the CN's for the newly graded
pervious areas.
Runoff Curve Numbers 499
Table 3.2b Runoff Curve Numbers for Cultivated Agricultural Lands*

Curve numbers for


Cover description hydrologic soil group

Hydrologic
Cover type Treatment* condition 0
A Β C D

Fallow Bare soil 77 86 91 94


Crop residue cover (CR) Poor 76 85 90 93
Good 74 83 88 90
Row crops Straight row (SR) Poor 72 81 88 91
Good 67 78 85 89
SR + CR Poor 71 80 87 90
Good 64 75 82 85
Contoured (C) Poor 70 79 84 88
Good 65 75 82 86
C + CR Poor 69 78 83 87
Good 64 74 81 85
Contoured & terraced (C&T) Poor 66 74 80 82
Good 62 71 78 81
C&T + CR Poor 65 73 79 81
Good 61 70 77 80
Small grain SR Poor 65 76 84 88
Good 63 75 83 87
SR + CR Poor 64 75 83 86
Good 60 72 80 84
C Poor 63 74 82 85
Good 61 73 81 84
C + CR Poor 62 73 81 84
Good 60 72 80 83
C&T Poor 61 72 79 82
Good 59 70 78 81
C&T+CR Poor 60 71 78 81
Good 58 69 77 80
Close-seeded SR Poor 66 77 85 89
or broadcast Good 58 72 81 85
legumes or C Poor 64 75 83 85
rotation meadow Good 55 69 78 83
C&T Poor 63 73 80 83
Good 51 67 76 80

Note. Source: Soil Conservation Service (1986).


a
Average runoff condition, and I = 0.2S.
a

b
Crop residue cover applies only if residue is on at least 5% of the surface throughout the year.
h y d r o l o g i c condition is based on combination of factors that affect infiltration and runoff, including (a) density and
canopy of vegetative areas, (b) amount of year-round cover, (c) amount of grass or close-seeded legumes in rotations,
(d) percent of residue cover on the land surface (good > 20%), and (e) degree of surface roughness. Poor: Factors
impair infiltration and tend to increase runoff. Good: Factors encourage average and better than average infiltration and
tend to decrease runoff.
500 Appendix 3C
Table 3.2c Runoff Curve Numbers for Other Agricultural Lands' 1

Curve numbers for


Cover description hydrologic soil group

Hydrologic
Cover type condition A Β C D

Pasture, grassland, or range—continuous Poor 68 79 86 89


forage for grazing* Fair 49 69 79 84
Good 39 61 74 80
Meadow—continuous grass, protected from — 30 58 71 78
grazing and generally mowed for hay.
Brush—brush-weed-grass mixture with brush Poor 48 67 77 83
the major element* Fair 35 56 70 77
Good 30* 48 65 73
Woods—grass combination (orchard or tree farm)' Poor 57 73 82 86
Fair 43 65 76 82
Good 32 58 72 79
Wood/ Poor 45 66 77 83
Fair 36 60 73 79
Good 30 55 70 77
Farmsteads—buildings, lanes, driveways, and — 59 74 82 86
surrounding lots.

Note. Source: Soil Conservation Service (1986).


"Average runoff condition, and I = 0.2S.
a

b
Poor: <50% ground cover or heavily grazed with no mulch. Fair: 50 to 75% ground cover and not heavily grazed.
Good: >75% ground cover and lightly or only occassionally grazed.
c
Poor: <50% ground cover. Fair: 50 to 75% ground cover. Good: >75% ground cover.
^Actual curve number is less than 30; use CN = 30 for runoff computations.
'CN's shown were computed for areas with 50% woods and 50% grass (pasture) cover. Other combinations of
conditions may be computed from the CN's for woods and pasture.
fPoor: Forest litter, small trees, and brush are destroyed by heavy grazing or regular burning. Fair: Woods are grazed
but not burned, and some forest litter covers the soil. Good: Woods are protected from grazing, and litter and brush
adequately cover the soil.

Table 3.2d Runoff Curve Numbers for Arid and Semiarid Rangelands*

Curve numbers for


Cover description hydrologic soil group

Hydrologic
Cover type condition* A c
Β C D

Herbaceous—mixture of grass, weeds, and Poor 80 87 93


low-growing brush, with the minor element Fair 71 81 89
Good 62 74 85
Oak-aspen—mountain brush mixture of oak brush, Poor 66 74 79
aspen, mountain mahogany, bitter brush, maple, Fair 48 57 63
and other brush Good 30 41 48
Pinyon-juniper—pinyon, juniper, or both; Poor 75 85 89
grass understory Fair 58 73 80
Good 41 61 71
Sagebrush with grass understory Poor 67 80 85
Fair 51 63 70
Good 35 47 55
Desert shrub—major plants include saltbrush, Poor 63 77 85 88
greasewood, creosotebrush, blackbrush, bursage Fair 55 72 81 86
palo verde, mesquite, and cactus Good 49 68 79 84

Note. Source: Soil Conservation Service (1986).


"Average runoff condition, and I = 0.2S. For range in humid regions, use Table 3.2c.
a

Poor: <30% ground cover (litter, grass, and brush overstory). Fair: 30 to 70% ground cover. Good: >70% ground
b

cover.
Appendix 5A

DISCHARGE (CFS)
Figure 5A.1 Culvert capacity of circular concrete pipe, groove-edged entrance, 1 8 - 3 6 in. diameter (after Bureau of
Public Roads, 1965a).

0 50 100 150 200 250 300 350


DISCHARGE (CFS)
Figure 5A.2 Culvert capacity of circular concrete pipe, groove-edged entrance, 3 6 - 6 6 in. diameter (after Bureau of
Public Roads, 1965a).
502 Appendix 5A
14 1 r s L "' 1
Jc J6 M20)r
ET 12
LU
L.
10
0d / 7 <tyyX >^ ^ ^ ^ ^
Ld
8
<
AD

6
UJ
X
4 . 1 1 1 1 1 1
100 200 300 400 500 600 700 800 900 1000 1100 1200 1300
DISCHARGE (CFS)
Figure 5A.3 Culvert capacity of circular concrete pipe, groove-edged entrance, 6 0 - 1 2 0 in. diameter (after Bureau of
Public Roads, 1965a).

500 1000 1500 2000 2500 2800


DISCHARGE (CFS)
Figure 5A.4 Culvert capacity of circular concrete pipe, groove-edged entrance, 7 2 - 1 8 0 in. diameter (after Bureau of
Public Roads, 1965a).

DISCHARGE (CFS)
Figure 5A.5a Culvert capacity of circular corrugated metal pipe, projecting entrance, 1 8 - 3 6 in. diameter for low
values of L / 1 0 0 0 5 (after Bureau of Public Roads, 1965a).
0
Appendix 5A 503

Ρ 7

UJ ~

r
Ld Ο
ϋ. 5

ι;
LJ

<
UJ ο
I L·

1 10 20 30 40 50 60
DISCHARGE (CFS)
Figure 5A.5b Culvert capacity of circular corrugated metal pipe, projecting entrance, 1 8 - 3 6 in. diameter for high
values of L / 1 0 0 0 S (after Bureau of Public Roads, 1965a).
0

DISCHARGE (CFS)
Figure 5A.6a Culvert capacity of circular corrugated metal pipe, projecting entrance, 3 6 - 7 2 in. diameter for low
values of L / 1 0 0 0 S (after Bureau of Public Roads, 1965a).
0

\ I I L
50 100 150 200 250 300 350
DISCHARGE (CFS)
Figure 5A.6b Culvert capacity of circular corrugated metal pipe, projecting entrance, 3 6 - 7 2 in. diameter for high
values of L / 1 0 0 0 S (after Bureau of Public Roads, 1965a).
0
504 Appendix 5A

100 200 300 400 500 600 700 800 900 1000 1100 1200 1300
DISCHARGE ( C F S )
Figure 5A.7 Culvert capacity of circular corrugated metal pipe, projecting entrance, 6 0 - 1 8 0 in. diameter (after
Bureau of Public Roads, 1965a).

DISCHARGE ( C F S )
Figure 5A.8 Culvert capacity of circular corrugated metal pipe, projecting entrance, 8 4 - 1 8 0 in. diameter (after
Bureau of Public Roads, 1965a).
Appendix 5B

505
506 Appendix 5B

180
HW ENTRANCE
ΤΥΡΕ
(D (2) (3)
, 10.000 -fr~"SCALE
168 rr- 6
156 SQUARE EDGE I
144 (D WITH HEADWALL 5
132
120 GROOVE END 4
(2) WITH HEADWALL
• 108
96 • 1000 X
GROOVE END
; 8οο (3) PROJECTING
• 84 • 600 CO­
OT
TO USE SCALE (2) OR (J) PROJECT LU
• 72 • 500 • • o ^ ^ a l l y T0SCALE (l). r —
F- 1.5 t L 1.5
• 60 Ο • 200
• 54 <
• 48 ξ 18°
1.0 1.0
- 42 - 40 X 1.0
.9 .9
• 36
- 30
I—
L_- _a_
• 33 Ο Lu
-20 Q
- 30 <
Χ — 10 or
• 27 ο = 8 Ld-
CO I—
- 24
• 21
if I
- 18 - 3 ο
- 2 δ"
- 15 X L _ .5
1.0
- 12
Figure 5B.1 Headwater depth for circular concrete pipe culverts with inlet control (after Federal Highway
Administration, 1985).

— 156 Τ
=— 144 i
* 0

(0
ENTRANCE
TYPE
0 )

6
(2)
(3)
P-5
44 j Ξ-3000 (2)
MITERED TO CON- Q
FORM TO SLOPE \ " b-4 1

5
- 6

132 1-2000 — 4
CO d PROJECTING ^ E- 4
UJ (3) -3 =
120 ju X^
X CO £==1000 — 3
Ο 108 J 800 on
b - 896
4 *
15
Ο
PS!
ι-
TO USE SCALE (2) OR (5) PROJECT
HORIZONTALLY TO SCALE (1), THEN
USE STRAIGHT MCUNE UNE THROUGH ^ 2
E., E- 2
Fs— 300 0 AND Q SCALES. OR REVERSE.
< -1.5 Ε
I— 6 0 C-1.5
• 54
200
Q F- 1.5
0_ • 48 LU E-100 Ζ
Ο
• 42 E_80 X
ο
_l < Ι­
T3"
α: ID X —30 Ο.
< (J • 36 Ο Lu
Q • 33 CO =-20 Q ( - 8
u_
Ζ
<
ο
•30 Q
=--10
ι—8
• 27 = 8 \—.7 ^.7
• 24
Ld F=ir
— 21
<
< h - .6
— 18 _r-2 UJ
χ ,
— 15
1
—.5
1.0 '— .5

12
Figure 5B.2 Headwater depth for circular corrugated metal pipe culverts with inlet control (after Federal Highway
Administration, 1985).
Appendix 5B 507
Appendix 5C

509
510 Appendix 5C
Appendix 5D

511
512 Appendix 5D

Values of H , ft
p

Figure 5D.1 H cc versus H p for emergency spillways of various lengths L. Case 1, b = 100 ft, ζ = 2, η = 0.04. (Soil
Conservation Service, 1968).
Appendix 5D 513

Volues of H p , ft
Figure 5D.2 H cc versus H p for emergency spillways of various lengths L. Case 2, b = 100 ft, ζ = 2, η = 0.04. (Soil
Conservation Service, 1968).
514 Appendix 5D

Reservoir Control
15.0-r Woter Section
'5.0 Surface

10.0

-9.0
IfflWtt
~β.Ο

uiiwv — 7 . 0
5.0-
VAW\ - 6 . 0 NGMEHCIATURE
Hec = Critical specific energy
-5.0 head - ft
4.0- Hp s Biergy head of the water
ο in the reservoir above the
Φ spillway crest - ft
I -4.0 Bp = Difference in the elevation
of the water surface in the
3.0- reservoir and the spillway
to crest - ft
φ
L s Length of the spillway
-3.0 °* upstream from the control
section - ft

20-

"2.0
Appendix 5D 515
Control
Section

*2.0

Figure 5D.4 H ec versus H p for emergency spillways of various lengths L. Case 4, b = 100 ft, ζ = 2, η = 0.04. (Soil
Conservation Service, 1968).
516 Appendix 5D

Figure 5D.5 H cc versus H p for emergency spillways of various lengths L. Case 5, b = 100 ft, ζ = 2, η = 0.04. (Soil
Conservation Service, 1968).
Appendix 5D 517
518 Appendix 5D

Values of H p i ft
Figure 5D.7 H ec versus H p for emergency spillways of various lengths L. Case 7, b = 100 ft, ζ = 2, η = 0.04. (Soil
Conservation Service, 1968).
Appendix 5D 519

Values of H D t ft
Figure 5D.8 H cc versus H p for emergency spillways of various lengths L. Case 8, b = 100 ft, ζ - 2, η = 0.04. (Soil
Conservation Service, 1968).
520 Appendix 5D

Values of H p t ft
Figure 5D.9 H cc versus H p for emergency spillways of various lengths L. Case 9, b = 100 ft, ζ = 2, η = 0.04. (Soil
Conservation Service, 1968).
Appendix 5E

521
522 Appendix 5E

4 5 6 7 θ 9 10

Critical Specific Energy Head, H c . ft e

Figure 5E.1 Effect of η on friction head loss. Case 1, b = 100 ft, ζ = 2 (Soil Conservation Service, 1968).
Appendix 5E 523
Reservoir

0.49 0.5 0.6 0.7 0.8 0.9 I 2 3 4 5 6 7 8 9 10 15


Critical Specific Energy Head, H e C t ft
Figure 5E.2 Effect of η on friction head loss for η = 0.04. Case 2, b = 100 ft, ζ = 2 (Soil Conservation Service, 1968).
524 Appendix 5E

Critical Specific Energy Head, H ,ft


e c

Figure 5E.3 Effect of η on friction head loss for η = 0.02 and 0.08. Case 2, b = 100 ft, ζ = 2 (Soil Conservation Service,
1968).
Appendix 5E 525

Q45Q5 0.6 0.7 0.8 0.9 I 2 3 4 5 6 7 8 9 10 15

Critical Specific Energy Head, Hec, ft


Figure 5E.4 Effect of ζ on friction head loss. Case 1, b = 100 ft, η = 0.04 (Soil Conservation, 1968).
526 Appendix 5E

Critical Specific Energy Head, H c, ft e

Figure 5E.5 Effect of b on friction head loss. Case 1, ζ = 2, η = 0.04 (Soil Conservation Service, 1968).
Appendix 5E 527

NOMENCLATURE

the c r i t i c a l slope for a discharge


o f 0 / 4 , where Q i s i n c o r r e s p o n d e n c e
w i t h Hec - f t / f t

the permissible v e l o c i t y in the


e x i t channel - f t / s e c

i
.01 .02 .03 .04 .05 .06 .08 0.1 0.2 0.3· 0.4 0.5
Bottom Slope of Exit Channel, s , ft/ft 0

Figure 5E.6 Permissible Hcc for various s and u with b = 100 ft, ζ = 2, and η = 0.04 (Soil Conservation Service, 1968).
0 p
528 Appendix 5E

NOMENCLATURE

S
C,Q/4
the critical slope for a discharge
of Qjh, where Q is in correspondence
with H^c - ft/ft

the permissible velocity in the


exit channel - ft/sec

Note The value of η in the block in the


upper right hand corner does not
apply to the s ^ Q / -lines for
c 4

η = 0.03 and η = 0.01*.

.004 .006 .006 01 .02 03 .04 .06 .08 0.1 0.2 0.3 0.4 0.5

Bottom Slope of Exit Channel, S o , f t / f t


Figure 5E.7 Permissible Hcc for various sQ and vp with b = 100 ft, ζ = 2, and η = 0.02 (Soil Conservation Service, 1968).
Appendix 8A
R Factor Information
530 Appendix 8A

Figure 8A.1 Isolines of R factor for Western U.S. (after Renard et al., 1993b). Units on R are ft · tonsf · i n . / a c r e · hr ·
year. T o convert to metric, MJ · m m / h a • h · year, multiply by 17.02.
R Factor Information 531

Figure 8A.2 Isolines of R factor for California (after Renard, et β/., 1993b). Units on R are
ft · tonsf · i n . / a c r e · hr · year. T o convert to metric, MJ · m m / h a · h · year, multiply by 17.02.
532 Appendix 8A

Figure 8A.3 Isolines of R factor for Washington and Oregon (after Renard et al., 1993b). Units on R are
ft · tonsf · i n . / a c r e · hr · year. T o convert to metric, MJ · m m / h a · h · year, multiply by 17.02.
534 Appendix 8A

Figure 8A.5 Isolines of 10-year return period single storm R factor for the Eastern U.S. (after Renard et al, 1993b). Units
on R are ft · tonsf · i n . / a c r e · hr · storm. T o convert to metric, MJ · m m / h a · h · storm, multiply by 17.02.
R Factor Information 535

Figure 8A.6 Isolines of 10-year return period single storm R factor for the Western U.S. (after Renard et aL, 1993b.)
Units on R are ft · tonsf · i n . / a c r e · hr · storm. T o convert to metric, MJ · m m / h a · h · storm, multiply by 17.02.
536 Appendix 8A

Figure 8A.7 Isolines of 10-year return period single storm R factor for California (after Renard et aL y

1993b). Units on R are ft · tonsf · i n . / a c r e · hr · storm. T o convert to metric, MJ · m m / h a · h · storm


multiply by 17.02.
R Factor Information 537

Figure 8A.8 Isolines of 10-year return period single storm R factor for Washington and Oregon (after Renard et al.,
1993b). Units on R are ft · tonsf · i n . / a c r e · hr - storm. T o convert to metric, MJ · m m / h a · h · storm, multiply by 17.02.
538 Appendix 8A

Table 8A. 1 R Factor Distributions for Zones in Fig. 8.9*

Geographic Area

Parti

Date 1 2 3 4 5 6 7 8 9 10

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1/15 4.3 4.3 7.4 3.9 2.3 0.0 0.0 0.0 0.8 0.3
2/1 8.3 8.3 13.8 7.9 3.6 0.0 0.0 0.0 3.1 0.5
2/15 12.8 12.8 20.9 12.6 4.7 0.5 0.0 0.0 4.7 0.9
3/1 17.3 17.3 26.5 17.4 6.0 2.0 0.0 0.0 7.4 2.0
3/15 21.6 21.6 31.8 21.6 7.7 4.1 1.2 0.9 11.7 4.3
4/1 25.1 25.1 35.3 25.2 10.7 8.1 4.9 3.6 17.8 9.2
4/15 28.0 28.0 38.5 28.7 13.9 12.6 8.5 7.8 22.5 13.1
5/1 30.9 30.9 40.2 31.9 17.8 17.6 13.9 15.0 27.0 18.0
5/15 34.9 34.9 41.6 35.1 21.2 21.6 19.0 20.2 31.4 22.7
6/1 39.1 39.1 42.5 38.2 24.5 25.5 26.1 27.4 36.0 29.2
6/15 42.6 42.6 43.6 42.0 28.1 29.6 35.4 38.1 41.6 39.5
7/1 45.4 45.4 44.5 44.9 31.1 34.5 43.9 49.8 46.4 46.3
7/15 48.2 48.2 45.1 46.7 33.1 40.0 48.8 57.9 50.1 48.8
8/1 50.8 50.8 45.7 48.2 35.3 45.7 53.9 65.0 53.4 51.1
8/15 53.0 53.0 46.4 50.1 38.2 50.7 64.5 75.6 57.4 57.2
9/1 56.0 56.0 47.7 53.1 43.2 55.6 73.4 82.7 61.7 64.4
9/15 60.8 60.8 49.4 56.6 48.7 60.2 77.5 86.8 64.9 67.7
10/1 66.8 66.8 52.8 62.2 57.3 66.5 80.4 89.4 69.7 71.1
10/15 71.0 71.0 57.0 67.9 67.8 75.5 84.8 93.4 79.0 77.2
11/1 75.7 75.7 64.5 75.2 77.9 85.6 89.9 96.3 89.6 85.1
11/15 82.0 82.0 73.1 83.5 86.0 95.9 96.6 99.1 97.4 92.5
12/1 89.1 89.1 83.3 90.5 91.3 99.5 99.2 100.0 100.0 96.5
12/15 95.2 95.2 92.3 96.0 96.9 99.9 99.7 100.0 100.0 99.0

Part 2

Date 11 12 13 14 15 16 17 18 19 20

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1/15 5.4 3.5 0.0 0.7 0.0 0.0 0.0 0.0 1.0 9.8
2/1 11.3 7.8 0.0 1.8 0.0 0.0 0.0 0.0 2.6 18.5
2/15 18.8 14.0 1.8 3.3 0.5 0.5 0.7 0.6 7.4 25.4
3/1 26.3 21.1 7.2 6.9 2.0 2.0 2.8 2.5 16.4 30.2
3/15 33.2 27.4 11.9 16.5 4.4 5.5 6.1 6.2 23.5 35.6
4/1 37.4 31.5 16.7 26.6 8.7 12.3 10.7 12.4 28.0 38.9
4/15 40.7 35.0 19.7 29.9 12.0 16.2 12.9 16.4 31.0 41.5
5/1 42.5 37.3 24.0 32.0 16.6 20.9 16.1 20.2 33.5 42.9
5/15 44.3 39.8 31.2 35.4 21.4 26.4 21.9 23.9 37.0 44.0
6/1 45.4 41.9 42.4 40.2 29.7 35.2 32.8 29.3 41.7 45.2
6/15 46.5 44.3 55.0 45.1 44.5 48.1 45.9 37.7 48.1 48.2
7/1 47.1 45.6 60.0 51.9 56.0 58.1 55.5 45.6 51.1 50.8
7/15 47.4 46.3 60.8 61.1 60.8 63.1 60.3 49.8 52.0 51.7
8/1 47.8 46.8 61.2 67.5 63.9 66.5 64.0 53.3 52.5 52.5
8/15 48.3 47.9 62.6 70.7 69.1 71.9 71.2 58.4 53.6 54.6
9/1 49.4 50.0 65.3 72.8 74.5 77.0 77.2 64.3 55.7 57.4
9/15 50.7 52.9 67.6 75.4 79.1 81.6 80.3 69.0 57.6 58.5
10/1 53.6 57.9 71.6 78.6 83.1 85.1 83.1 75.0 61.1 60.1
10/15 57.5 62.3 76.1 81.9 87.0 88.4 87.7 86.6 65.5 63.2

11/1 65.5 69.3 83.1 86.4 90.9 91.5 92.6 93.9 74.7 69.6
11/15 76.2 81.3 93.3 93.6 96.6 96.3 97.2 96.6 88.0 76.7
12/1 87.4 91.5 98.2 97.7 99.1 98.7 99.1 98.0 95.8 85.4
12/15 94.8 96.7 99.6 99.3 99.8 99.6 99.8 100.0 98.7 92.4

(continues)
R Factor Information 539
Part 3

Date 21 22 23 24 25 26 27 28 29 30

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

1/15 7.5 1.2 7.9 12.2 9.8 2.0 0.0 0.0 0.6 0.0

2/1 13.6 1.6 15.0 23.6 20.8 5.4 0.0 0.0 0.7 0.0

2/15 18.1 1.6 20.9 33.0 30.2 9.8 1.0 0.0 0.7 0.0

3/1 21.1 1.6 25.7 39.7 37.6 15.6 4.0 0.2 0.7 0.0

3/15 24.4 1.6 31.1 47.1 45.8 21.5 5.9 0.5 1.5 0.2

4/1 27.0 1.6 35.7 51.7 50.6 24.7 8.0 1.5 3.9 0.8

4/15 29.4 2.2 40.2 55.9 54.4 26.6 11.1 3.3 6.0 2.8

5/1 31.7 3.9 43.2 57.7 56.0 27.4 13.0 7.2 10.5 7.9

5/15 34.6 4.6 46.2 58.6 56.8 28.0 14.0 11.9 17.9 14.2

6/1 37.3 6.4 47.7 58.9 57.1 28.7 14.6 17.7 28.8 24.7

6/15 39.6 14.2 48.8 59.1 57.1 29.8 15.3 21.4 36.6 35.6

7/1 41.6 32.8 49.4 59.1 57.2 32.5 17.0 27.0 43.8 45.4

7/15 43.4 47.2 49.9 59.2 57.6 36.6 23.2 37.1 51.5 52.2

8/1 45.4 58.8 50.7 59.2 58.5 44.9 39.1 51.4 59.3 58.7

8/15 48.1 69.1 51.8 59.3 59.8 55.4 60.0 62.3 68.0 68.5
9/1 51.3 76.0 54.1 59.5 62.2 65.7 76.3 70.6 74.8 77.6

9/15 53.3 82.0 57.7 60.0 65.3 72.6 86.1 78.8 80.3 84.5

10/1 56.6 87.1 62.8 61.4 67.5 77.8 89.7 84.6 84.3 88.9
10/15 62.4 96.7 65.9 63.0 68.2 84.4 90.4 90.6 88.8 93.7

11/1 72.4 99.9 70.1 66.5 69.4 89.5 90.9 94.4 92.7 96.2
11/15 81.3 99.9 77.3 71.8 74.8 93.9 93.1 97.9 98.0 97.6
12/1 88.9 99.9 86.8 81.3 86.6 96.5 96.6 99.3 99.8 98.3
12/15 94.7 99.9 93.5 89.6 93.0 98.4 99.1 100.0 99.9 99.6

Part 4

Date 31 32 33 34 35 36 37 38 39 40

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1/15 0.0 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
2/1 0.0 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
2/15 0.0 0.1 0.0 0.0 0.0 0.0 0.0 1.1 0.0 0.0
3/1 0.0 0.1 0.0 0.0 0.0 0.0 0.0 4.3 0.0 0.0
3/15 0.2 0.6 0.6 1.8 2.5 0.9 0.0 7.2 1.6 1.5
4/1 1.0 2.2 2.3 7.3 10.2 3.4 0.0 11.0 6.5 6.2
4/15 3.5 4.3 4.2 10.7 15.9 6.7 1.0 13.9 11.0 10.1
5/1 9.9 9.0 8.8 15.5 22.2 12.7 3.9 17.9 17.8 16.3
5/15 15.7 14.2 16.1 22.0 27.9 18.5 9.1 22.3 24.7 23.3
6/1 26.4 23.3 30.0 29.9 34.7 26.6 19.1 30.3 33.1 32.5
6/15 47.2 34.6 46.9 35.9 43.9 36.3 26.7 43.1 42.8 42.2
7/1 61.4 46.3 57.9 42.0 51.9 46.0 36.3 55.1 50.3 50.1
7/15 65.9 54.2 62.8 48.5 56.9 53.5 47.9 61.3 54.9 55.6
8/1 69.0 61.7 66.2 56.9 61.3 60.2 61.4 65.7 59.7 60.5
8/15 77.2 72.9 72.1 67.0 67.3 68.3 75.1 72.1 68.9 67.5
9/1 86.0 82.5 79.1 76.9 73.9 75.8 84.5 77.9 78.1 74.3
9/15 91.6 89.6 85.9 85.8 80.1 82.6 92.3 82.6 83.6 79.4
10/1 94.8 93.7 91.1 91.2 85.1 88.3 96.0 86.3 87.5 84.1
10/15 98.7 98.2 97.0 95.7 89.6 96.3 99.1 90.3 93.0 91.1
11/1 100.0 99.7 98.9 97.8 93.2 99.3 100.0 93.8 96.5 95.8
11/15 100.0 99.9 98.9 99.6 98.2 99.9 100.0 98.4 99.2 99.1
12/1 100.0 99.9 98.9 100.0 99.8 100.0 100.0 100.0 100.0 100.0
12/15 100.0 99.9 98.9 100.0 99.8 100.0 100.0 100.0 100.0 100.0

(continues)
540 Appendix 8A
Part 5
Date 41 42 43 44 45 46 47 48 49 50

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

1/15 0.1 0.0 0.0 1.7 0.2 0.0 0.0 0.0 0.0 0.0

2/1 0.2 0.0 0.0 2.3 0.2 0.0 0.0 0.0 0.0 0.0

2/15 0.2 0.0 0.0 2.4 0.3 0.0 0.0 0.0 0.0 0.0

3/1 0.2 0.0 0.0 2.4 0.3 0.0 0.0 0.0 0.0 0.0

3/15 0.2 0.0 0.0 2.4 0.4 0.0 0.0 0.0 0.0 0.1

4/1 0.2 0.0 0.0 2.4 0.6 0.0 0.0 0.0 0.0 0.4

4/15 0.4 0.2 0.1 2.7 0.8 0.6 0.4 0.0 0.7 2.4

5/1 1.1 0.9 0.4 3.5 1.4 2.6 1.6 0.0 2.7 8.2

5/15 6.8 5.2 2.7 7.6 3.7 7.5 5.8 2.0 8.3 13.7

6/1 22.9 17.3 9.5 18.5 10.2 19.6 17.0 8.1 20.0 23.8

6/15 40.1 33.8 21.9 34.3 22.6 32.9 33.0 15.4 27.5 38.8

7/1 54.9 53.2 42.7 52.5 41.8 48.9 52.5 27.8 35.6 55.1

7/15 63.8 66.5 58.6 64.0 54.0 63.0 66.4 40.7 44.6 66.1

8/1 70.7 75.9 71.1 72.3 64.5 73.5 75.7 52.6 46.0 73.6

8/15 81.5 87.6 84.6 83.3 78.7 83.3 85.5 61.1 70.2 81.8

9/1 89.8 93.7 91.9 90.0 88.4 89.5 91.3 69.3 81.3 87.7

9/15 96.3 97.5 97.1 95.1 96.0 95.6 96.5 82.6 89.2 93.8

10/1 98.7 99.0 99.0 97.3 98.7 98.3 98.8 92.0 93.6 97.0

10/15 99.2 99.7 99.8 98.5 99.4 99.6 100.0 98.0 98.5 99.4

11/1 99.3 100.0 100.0 98.9 99.7 100.0 100.0 100.0 100.0 100.0

11/15 99.4 100.0 100.0 98.9 99.7 100.0 100.0 100.0 100.0 100.0

12/1 99.4 100.0 100.0 98.9 99.8 100.0 100.0 100.0 100.0 100.0

12/15 99.7 100.0 100.0 99.2 99.9 100.0 100.0 100.0 100.0 100.0

Part 6
Date 51 52 53 54 55 56 57 58 59 60

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

1/15 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

2/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

2/15 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

3/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

3/15 0.3 0.0 0.0 0.2 0.0 0.0 0.0 0.2 0.0 0.0

4/1 1.0 0.0 0.0 0.7 0.0 0.0 0.1 0.9 0.0 0.0

4/15 3.1 0.6 0.8 2.4 1.3 1.3 1.0 2.9 2.2 0.4

5/1 8.7 2.5 3.0 7.2 5.4 5.1 3.5 8.0 8.9 1.5

5/15 18.8 6.8 9.5 14.7 13.3 11.4 9.2 13.2 15.6 4.0

6/1 35.8 17.5 24.2 27.2 25.5 22.3 21.5 21.0 24.2 9.5

6/15 49.6 29.8 35.3 37.2 31.6 29.5 31.0 29.1 31.1 13.3

7/1 60.4 46.1 48.0 47.3 38.8 38.5 43.5 38.0 38.3 20.5

7/15 70.2 60.5 63.1 58.8 52.5 51.1 60.4 45.9 46.0 33.6

8/1 77.0 72.7 76.1 67.6 66.8 65.2 75.1 54.5 54.9 52.8

8/15 84.0 86.0 87.7 74.0 75.5 77.8 86.1 65.4 64.2 66.5

9/1 88.8 92.8 93.5 79.2 81.2 85.6 91.6 74.8 73.2 76.7

9/15 93.8 96.8 97.2 86.7 87.9 91.7 96.2 82.1 81.9 88.1

10/1 96.6 98.4 98.6 92.6 92.8 95.0 98.1 87.5 88.5 94.2

10/15 99.1 99.7 99.5 97.9 98.3 98.7 99.4 95.4 95.7 98.6

11/1 100.0 100.0 99.8 99.8 100.0 100.0 99.9 98.8 98.6 100.0

11/15 100.0 100.0 99.9 99.9 100.0 100.0 99.9 99.7 99.4 100.0

12/1 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 99.7 100.0

12/15 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 99.7 100.0

(continues)
R Factor Information 541
Part 7

Date 61 62 63 64 65 66 67 68 69 70

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1/15 0.0 0.0 0.0 0.0 3.6 0.0 0.0 2.3 2.0 0.5

2/1 0.0 0.0 0.0 0.0 7.0 0.0 0.0 4.5 3.7 0.7
2/15 0.0 0.1 0.0 0.7 9.6 0.0 0.0 7.8 5.7 1.0

3/1 0.0 0.3 0.0 2.8 11.4 0.0 0.0 10.4 7.8 1.3
3/15 0.0 0.8 0.0 7.4 13.0 0.1 0.1 12.0 10.5 1.7
4/1 0.0 2.1 0.0 12.4 14.4 0.5 0.4 13.3 12.4 2.2
4/15 1.3 3.6 0.9 14.4 16.3 1.1 0.9 16.3 13.7 2.8
5/1 5.0 6.5 3.7 15.6 17.7 2.2 1.6 17.7 14.3 3.4

5/15 8.5 9.7 7.8 17.3 18.4 3.6 1.9 18.1 14.7 3.9

6/1 15.5 13.7 13.3 19.4 19.3 6.0 2.4 18.2 15.1 4.7
6/15 29.8 16.5 15.8 21.0 20.5 7.6 5.0 18.3 15.7 5.4
7/1 41.8 20.8 19.9 24.4 23.6 11.1 12.1 18.4 17.1 7.4
7/15 46.0 27.3 29.0 32.3 32.0 19.8 24.8 19.9 22.7 15.7
8/1 49.2 40.1 46.8 48.0 50.0 38.9 48.3 24.5 36.7 36.5
8/15 56.0 56.9 64.7 61.4 66.2 59.7 73.6 35.0 50.4 55.8
9/1 65.1 72.6 78.3 72.1 77.2 74.4 86.5 54.4 63.6 70.3
9/15 71.6 83.4 88.8 81.9 85.4 83.2 92.0 69.4 75.0 80.9
10/1 78.6 89.4 93.9 87.0 88.8 88.1 94.3 78.6 81.8 86.4
10/15 91.1 95.5 98.5 90.1 90.4 94.6 96.6 85.7 87.8 90.9

11/1 97.3 98.1 100.0 92.4 91.3 97.7 97.9 89.2 90.8 93.4
11/15 99.3 99.6 100.0 98.1 92.7 99.4 99.5 91.9 93.2 96.4
12/1 100.0 100.0 100.0 100.0 94.8 100.0 100.0 93.9 94.9 98.1
12/15 100.0 100.0 100.0 100.0 97.0 100.0 100.0 97.0 97.5 99.4

Part8

Date 71 72 73 74 75 76 77 78 79 80

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1/15 0.7 0.0 0.0 0.0 0.1 0.0 0.2 0.0 0.0 0.6
2/1 1.2 0.0 0.1 0.0 0.1 0.0 0.3 0.0 0.0 1.2
2/15 1.6 0.0 0.1 0.0 0.1 0.0 0.3 0.0 0.0 1.6
3/1 2.1 0.0 0.2 0.0 0.2 0.0 0.4 0.0 0.0 2.1
3/15 2.8 0.0 0.2 0.1 0.5 0.1 0.8 0.0 0.2 2.5
4/1 3.3 0.1 0.3 0.2 1.3 0.2 1.5 0.2 0.7 3.3
4/15 3.6 0.2 0.6 0.5 1.9 0.6 2.0 0.5 1.3 4.5
5/1 4.0 0.7 1.3 1.2 3.0 1.3 2.8 1.6 2.7 6.9
5/15 4.5 0.8 4.1 2.7 4.1 2.0 3.9 3.8 5.8 10.1
6/1 5.6 1.3 11.5 6.4 6.6 3.5 5.9 8.9 12.7 15.5
6/15 6.5 3.5 18.1 10.2 10.0 4.9 7.2 13.2 18.8 19.7
7/1 9.1 9.9 28.3 18.4 17.6 8.4 10.3 21.8 28.8 26.6
7/15 18.5 24.7 40.2 31.0 28.3 17.4 21.5 35.8 41.6 36.4
8/1 40.6 51.4 54.1 50.7 44.7 37.3 46.5 56.6 58.4 51.7
8/15 59.7 71.5 67.0 68.7 59.4 57.5 66.3 75.4 75.7 67.5
9/1 74.0 83.6 77.2 81.2 71.6 72.9 78.3 86.0 86.5 79.4
9/15 86.3 93.8 87.7 91.6 83.9 83.7 86.5 92.9 94.2 88.8
10/1 91.7 97.7 93.3 96.1 90.3 89.5 90.8 95.9 97.3 93.2
10/15 94.7 99.2 97.5 98.4 94.7 95.8 96.0 98.2 98.9 96.1
11/1 96.0 99.8 99.1 99.2 96.7 98.4 98.2 99.2 99.5 97.3
11/15 96.7 99.9 99.6 99.8 98.8 99.6 99.1 99.8 99.9 98.2
12/1 97.3 99.9 99.8 100.0 99.6 100.0 99.5 100.0 100.0 98.7
12/15 98.8 100.0 100.0 100.0 99.9 100.0 99.8 100.0 100.0 99.3

(continues)
542 Appendix 8A
Part 9

Date 81 82 83 84 85 86 87 88 89 90

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1/15 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0
2/1 0.1 0.1 0.1 0.1 0.0 0.0 0.0 0.0 1.0 2.0
2/15 0.2 0.1 0.1 0.1 0.0 0.0 0.0 0.0 1.0 3.0
3/1 0.4 0.2 0.1 0.2 0.0 0.0 1.0 1.0 2.0 4.0
3/15 0.5 0.2 0.3 0.3 0.0 0.0 1.0 1.0 3.0 6.0
4/1 0.8 0.5 0.9 0.6 1.0 1.0 2.0 2.0 4.0 8.0
4/15 0.9 1.2 1.6 1.7 2.0 2.0 3.0 3.0 7.0 13.0
5/1 1.5 3.1 3.5 4.9 3.0 3.0 6.0 6.0 12.0 21.0
5/15 3.9 6.7 8.3 9.9 6.0 6.0 10.0 13.0 18.0 29.0
6/1 9.9 14.4 19.4 19.5 11.0 11.0 17.0 23.0 27.0 37.0
6/15 12.8 20.1 30.0 27.2 23.0 23.0 29.0 37.0 38.0 46.0
7/1 18.2 29.8 44.0 38.3 36.0 36.0 43.0 51.0 48.0 54.0
7/15 30.7 44.5 59.2 52.8 49.0 49.0 55.0 61.0 55.0 60.0
8/1 54.1 64.2 72.4 68.8 63.0 63.0 67.0 69.0 62.0 65.0
8/15 77.1 83.1 84.6 83.9 77.0 77.0 77.0 78.0 69.0 69.0
9/1 89.0 92.2 91.2 91.6 90.0 90.0 85.0 85.0 76.0 74.0
9/15 94.9 96.4 96.5 96.4 95.0 95.0 91.0 91.0 83.0 81.0
10/1 97.2 98.1 98.6 98.2 98.0 98.0 96.0 94.0 90.0 87.0
10/15 98.7 99.3 99.5 99.2 99.0 99.0 98.0 96.0 94.0 92.0
11/1 99.3 99.7 99.8 99.6 100.0 100.0 99.0 98.0 97.0 95.0
11/15 99.6 99.8 99.9 99.8 100.0 100.0 100.0 99.0 98.0 97.0
12/1 99.7 99.8 100.0 99.8 100.0 100.0 100.0 99.0 99.0 98.0
12/15 99.9 99.9 100.0 99.9 100.0 100.0 100.0 100.0 100.0 99.0

Part 10

Date 91 92 93 94 95 96 97 98 99 100

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1/15 0.0 0.0 1.0 1.0 1.0 2.0 1.0 1.0 0.0 0.0
2/1 0.0 0.0 1.0 2.0 3.0 4.0 3.0 2.0 0.0 0.0
2/15 0.0 0.0 2.0 4.0 5.0 6.0 5.0 4.0 1.0 0.0
3/1 1.0 1.0 3.0 6.0 7.0 9.0 7.0 6.0 1.0 1.0
3/15 1.0 1.0 4.0 8.0 9.0 12.0 10.0 8.0 2.0 1.0
4/1 1.0 1.0 6.0 10.0 11.0 17.0 14.0 10.0 3.0 2.0
4/15 2.0 2.0 8.0 15.0 14.0 23.0 20.0 13.0 5.0 3.0
5/1 6.0 6.0 13.0 21.0 18.0 30.0 28.0 19.0 7.0 5.0
5/15 16.0 16.0 25.0 29.0 27.0 37.0 37.0 26.0 12.0 9.0
6/1 29.0 29.0 40.0 38.0 35.0 43.0 48.0 34.0 19.0 15.0
6/15 39.0 39.0 49.0 47.0 41.0 49.0 56.0 42.0 33.0 27.0
7/1 46.0 46.0 56.0 53.0 46.0 54.0 61.0 50.0 48.0 38.0
7/15 53.0 53.0 62.0 57.0 51.0 58.0 64.0 58.0 57.0 50.0
8/1 60.0 60.0 67.0 61.0 57.0 62.0 68.0 63.0 65.0 62.0
8/15 67.0 67.0 72.0 65.0 62.0 66.0 72.0 68.0 72.0 74.0
9/1 74.0 74.0 76.0 70.0 68.0 70.0 77.0 74.0 82.0 84.0
9/15 81.0 81.0 80.0 76.0 73.0 74.0 81.0 79.0 88.0 91.0
10/1 88.0 88.0 85.0 83.0 79.0 78.0 86.0 84.0 93.0 95.0
10/15 95.0 95.0 91.0 88.0 84.0 82.0 89.0 89.0 96.0 97.0

11/1 99.0 99.0 97.0 91.0 89.0 86.0 92.0 93.0 98.0 98.0
11/15 99.0 99.0 98.0 94.0 93.0 90.0 95.0 95.0 99.0 99.0
12/1 100.0 100.0 99.0 96.0 96.0 94.0 98.0 97.0 100.0 99.0
12/15 100.0 100.0 99.0 98.0 98.0 97.0 99.0 99.0 100.0 100.0

(continues)
R Factor Information 543
Part 11

Date 101 102 103 104 105 106 107 108 109 no

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1/15 0.0 0.0 1.0 2.0 1.0 3.0 3.0 3.0 3.0 1.0
2/1 0.0 1.0 2.0 3.0 3.0 6.0 5.0 6.0 6.0 3.0
2/15 1.0 2.0 3.0 5.0 6.0 9.0 7.0 9.0 10.0 5.0
3/1 2.0 3.0 4.0 7.0 9.0 13.0 10.0 12.0 13.0 7.0
3/15 3.0 4.0 6.0 10.0 12.0 17.0 14.0 16.0 16.0 9.0
4/1 4.0 6.0 8.0 13.0 16.0 21.0 18.0 20.0 19.0 12.0
4/15 6.0 8.0 10.0 16.0 21.0 27.0 23.0 24.0 23.0 15.0
5/1 9.0 11.0 14.0 19.0 26.0 33.0 27.0 28.0 26.0 18.0
5/15 14.0 15.0 18.0 23.0 31.0 38.0 31.0 33.0 29.0 21.0
6/1 20.0 22.0 25.0 27.0 37.0 44.0 35.0 38.0 33.0 25.0
6/15 28.0 31.0 34.0 34.0 43.0 49.0 39.0 43.0 39.0 29.0
7/1 39.0 40.0 45.0 44.0 50.0 55.0 45.0 50.0 47.0 36.0
7/15 52.0 49.0 56.0 54.0 57.0 61.0 53.0 59.0 58.0 45.0
8/1 63.0 59.0 64.0 63.0 64.0 67.0 60.0 69.0 68.0 56.0
8/15 72.0 69.0 72.0 72.0 71.0 71.0 67.0 75.0 75.0 68.0
9/1 80.0 78.0 79.0 80.0 77.0 75.0 74.0 80.0 80.0 77.0
9/15 87.0 85.0 84.0 85.0 81.0 78.0 80.0 84.0 83.0 83.0
10/1 91.0 91.0 89.0 89.0 85.0 81.0 84.0 87.0 86.0 88.0
10/15 94.0 94.0 92.0 91.0 88.0 84.0 86.0 90.0 88.0 91.0

11/1 97.0 96.0 95.0 93.0 91.0 86.0 88.0 92.0 90.0 93.0
11/15 98.0 98.0 97.0 95.0 93.0 90.0 90.0 94.0 92.0 95.0
12/1 99.0 99.0 98.0 96.0 95.0 94.0 93.0 96.0 95.0 97.0
12/15 100.0 100.0 99.0 98.0 97.0 97.0 95.0 98.0 97.0 99.0

Part 12

Date 111 112 113 114 115 116 117 118 119 120 121

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1/15 1.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0 2.0 1.0 8.0
2/1 2.0 0.0 2.0 2.0 2.0 3.0 2.0 2.0 4.0 2.0 16.0
2/15 3.0 1.0 3.0 4.0 3.0 5.0 3.0 3.0 6.0 4.0 25.0
3/1 4.0 2.0 4.0 6.0 4.0 7.0 4.0 5.0 8.0 6.0 33.0
3/15 5.0 3.0 5.0 8.0 5.0 9.0 5.0 7.0 12.0 7.0 41.0
4/1 6.0 4.0 6.0 11.0 6.0 12.0 7.0 10.0 16.0 9.0 46.0
4/15 8.0 5.0 8.0 13.0 8.0 15.0 9.0 14.0 20.0 12.0 50.0
5/1 11.0 7.0 10.0 15.0 10.0 18.0 11.0 18.0 25.0 15.0 53.0
5/15 15.0 12.0 13.0 18.0 14.0 21.0 14.0 22.0 30.0 18.0 54.0
6/1 20.0 17.0 17.0 21.0 19.0 25.0 17.0 27.0 35.0 23.0 55.0
6/15 28.0 24.0 22.0 26.0 26.0 29.0 22.0 32.0 41.0 31.0 56.0
7/1 41.0 33.0 31.0 32.0 34.0 36.0 31.0 37.0 47.0 40.0 56.5
7/15 54.0 42.0 42.0 38.0 45.0 45.0 42.0 46.0 56.0 48.0 57.0
8/1 65.0 55.0 52.0 46.0 56.0 56.0 54.0 58.0 67.0 57.0 57.8
8/15 74.0 67.0 60.0 55.0 66.0 68.0 65.0 69.0 75.0 63.0 58.0
9/1 82.0 76.0 68.0 64.0 76.0 77.0 74.0 80.0 81.0 72.0 58.8
9/15 87.0 83.0 75.0 71.0 82.0 83.0 83.0 89.0 85.0 78.0 60.0
10/1 92.0 89.0 80.0 77.0 86.0 88.0 89.0 93.0 87.0 88.0 61.0
10/15 94.0 92.0 85.0 81.0 90.0 91.0 92.0 94.0 89.0 92.0 63.0
11/1 96.0 94.0 89.0 85.0 93.0 93.0 95.0 95.0 91.0 96.0 66.5
11/15 97.0 96.0 92.0 89.0 95.0 95.0 97.0 96.0 93.0 97.0 72.0
12/1 98.0 98.0 96.0 93.0 97.0 97.0 98.0 97.0 95.0 98.0 80.0
12/15 99.0 99.0 98.0 97.0 99.0 99.0 99.0 97.0 97.0 99.0 90.0

(continues)
544 Appendix 8A

Part 13

Date 122 123 124 125 126 127 128 129 130 131

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1/15 7.0 4.0 4.0 7.0 9.0 8.0 8.0 9.0 10.0 8.0
2/1 14.0 8.0 9.0 12.0 16.0 15.0 15.0 16.0 20.0 15.0
2/15 20.0 12.0 15.0 17.0 23.0 22.0 22.0 22.0 28.0 22.0
3/1 25.5 17.0 23.0 24.0 30.0 28.0 29.0 27.0 35.0 28.0
3/15 33.5 23.0 29.0 30.0 37.0 33.0 34.0 32.0 41.0 33.0
4/1 38.0 29.0 34.0 39.0 43.0 38.0 40.0 37.0 46.0 38.0
4/15 43.0 34.0 40.0 45.0 47.0 42.0 45.0 41.0 49.0 41.0
511 46.0 38.0 44.0 50.0 50.0 46.0 48.0 45.0 51.0 44.0
5/15 50.0 44.0 48.0 53.0 52.0 50.0 51.0 48.0 53.0 47.0
6/1 52.5 49.0 50.0 55.0 54.0 52.0 54.0 51.0 55.0 49.0
6/15 54.5 53.0 51.0 56.0 55.0 53.0 57.0 53.0 56.0 51.0
7/1 56.0 56.0 52.0 57.0 56.0 53.0 59.0 55.0 56.0 53.0
7/15 58.0 59.0 53.0 58.0 57.0 53.0 62.0 56.0 57.0 55.0
8/1 59.0 62.0 55.0 59.0 58.0 53.0 63.0 57.0 58.0 56.0
8/15 60.0 65.0 57.0 61.0 59.0 54.0 64.0 57.0 59.0 58.0
9/1 61.5 69.0 60.0 62.0 60.0 55.0 65.0 58.0 60.0 59.0
9/15 63.0 72.0 62.0 63.0 62.0 57.0 66.0 59.0 61.0 60.0
10/1 65.0 75.0 64.0 64.0 64.0 59.0 67.0 61.0 62.0 63.0
10/15 68.0 79.0 67.0 66.0 67.0 63.0 69.0 64.0 65.0 65.0
11/1 72.0 83.0 72.0 70.0 71.0 68.0 72.0 68.0 69.0 69.0
11/15 79.0 88.0 80.0 77.0 77.0 75.0 76.0 73.0 74.0 75.0
12/1 86.0 93.0 88.0 84.0 86.0 83.0 83.0 79.0 81.0 84.0
12/15 93.0 96.0 95.0 92.0 93.0 92.0 91.0 89.0 90.0 92.0

Part 14

Date 132 133 134 135 136 137 138 139 140*

1/1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
1/15 10.0 8.0 12.0 7.0 11.0 10.0 11.0 8.0 13.0
2/1 18.0 16.0 22.0 15.0 21.0 18.0 22.0 14.0 28.0
2/15 25.0 24.0 31.0 22.0 29.0 25.0 31.0 20.0 43.0
3/1 29.0 32.0 39.0 30.0 37.0 30.0 39.0 25.0 56.0
3/15 33.0 40.0 45.0 37.0 44.0 39.0 46.0 32.0 65.0
4/1 36.0 46.0 49.0 43.0 50.0 46.0 52.0 37.0 69.0
4/15 39.0 51.0 52.0 49.0 55.0 51.0 56.0 42.0 69.4
5/1 41.0 54.0 54.0 53.0 57.0 54.0 58.0 47.0 69.7
5/15 42.0 56.0 55.0 55.0 59.0 57.0 59.0 50.0 70.1
6/1 44.0 57.0 56.0 57.0 60.0 58.0 60.0 53.0 70.4
6/15 45.0 58.0 56.0 58.0 60.0 59.0 61.0 55.0 70.8
7/1 46.0 58.0 56.0 59.0 60.0 59.0 61.0 56.0 71.1
7/15 47.0 59.0 56.0 60.0 60.0 60.0 61.0 58.0 71.5
8/1 48.0 59.0 57.0 61.0 61.0 60.0 61.0 59.0 71.9
8/15 49.0 60.0 57.0 62.0 61.0 60.0 62.0 61.0 72.2
9/1 51.0 60.0 57.0 63.0 61.0 61.0 62.0 63.0 72.6
9/15 53.0 61.0 57.0 65.0 62.0 62.0 62.0 64.0 73.0
10/1 56.0 62.0 58.0 67.0 63.0 63.0 63.0 66.0 73.3
10/15 59.0 64.0 59.0 70.0 64.0 64.0 64.0 68.0 73.6
11/1 64.0 68.0 62.0 74.0 67.0 67.0 66.0 71.0 74.0
11/15 70.0 74.0 68.0 79.0 71.0 72.0 71.0 76.0 76.0
12/1 80.0 83.0 77.0 85.0 78.0 80.0 78.0 85.0 81.0
12/15 90.0 91.0 88.0 92.0 89.0 90.0 89.0 93.0 89.0

(Data provided by Renard, USDA-ARS, Tucson, AZ. Units on R are ft»tonsf»in./acre«hr-year. To convert to met­
e

ric, MJ«mm/ha*hr*year, multiply by 17.02.


*Zone 140 is for Pullman, WA and should be used for winter wheat and other dryland grain crops in the Northwest.
R Factor Information 545

Table 6A.2 Return Period Single-Storm R Factors for Selected Cities in the U . S . e

Index values normally exceeded once in (year) Index values normally exceeded once in (year)

Location 1 2 5 10 20 Location 1 2 5 10 20

Alabama Des Moines 31 45 67 86 105


Birmingham 54 77 110 140 170 Dubuque 43 63 91 114 140
Mobil 97 122 151 172 194 Rockwell City 31 49 76 101 129
Montgomery 62 86 118 145 172 Sioux City 40 58 84 105 131
Arkansas Kansas
Fort Smith 43 65 101 132 167 Burlingame 37 51 69 83 100
Little Rock 41 69 115 158 211 Coffeyville 47 69 101 128 159
Mountain Home 33 46 68 87 105 Concordia 33 53 86 116 154
Texarkana 51 73 105 132 163 Dodge City 31 47 76 97 124
California Goodland 26 37 53 67 80
Red Bluff 13 21 36 49 65 Hays 35 51 76 97 121
San Luis Obispo 11 15 22 28 34 Wichita 41 61 93 121 150
Colorado Kentucky
Akron 22 36 63 87 118 Lexington 28 46 80 114 151
Pueblo 17 31 60 88 127 Louisville 31 43 59 72 85
Springfield 31 51 84 112 152 Middlesboro 28 38 52 63 73
Connecticut Louisiana
Hartford 23 33 50 64 79 New Orleans 104 149 214 270 330
New Haven 31 47 73 96 122 Shreveport 55 73 99 121 141
District of Columbia 39 57 86 108 136 Maine
Florida Caribou 14 20 28 36 44
Appalachicola 87 124 180 224 272 Portland 16 27 48 66 88
Jacksonville 92 123 166 201 236 Skowhegan 18 27 40 51 63
Miami 93 134 200 253 308 Maryland
Georgia Baltimore 41 59 86 109 133
Atlanta 49 67 92 112 134 Massachusetts
Augusta 34 50 74 94 118 Boston 17 27 43 57 73
Columbus 61 81 108 131 152 Washington 29 35 41 45 50
Macon 53 72 99 122 146 Michigan
Savannah 82 128 203 272 358 Alpena 14 21 32 41 50
Watkinsville 52 71 98 120 142 Detroit 21 31 45 56 68
Illinois East Lansing 19 26 36 43 51
Cairo 39 63 101 135 173 Grand Rapids 24 28 34 38 42
Chicago 33 49 77 101 129 Minnesota
Dixon Springs 39 56 82 105 130 Duluth 21 34 53 72 93
Moline 39 50 89 116 145 Fosston 17 26 39 51 63
Rantoul 27 39 56 69 82 Minneapolis 25 35 51 65 78
Springfield 36 52 75 94 117 Rochester 41 58 85 105 129
Indiana Springfield 24 37 60 80 102
Evansville 26 38 56 71 86 Mississippi
Fort Wayne 24 33 45 56 65 Meridian 69 92 125 151 176
Indianapolis 29 41 60 75 90 Oxford 48 64 86 103 120
South Bend 26 41 65 86 111 Vicksburg 57 78 111 136 161
Terre Haute 42 57 78 96 113 Missouri
Iowa Columbia 43 58 77 93 107
Burlington 37 48 62 72 81 Kansas City 30 43 63 78 93
Charles City 33 47 68 85 103 McCredie 35 55 89 117 151
Clarinda 35 48 66 79 94 Rolla 43 63 91 115 140

(continues)
546 Appendix 8A

Table 8A.2 (Continued)

Index values normally exceeded once in (year) Index values normally exceeded once in (year)

Location 1 2 5 10 20 Location 1 2 5 10 20

Springfield 37 51 70 87 102 Oklahoma


St. Joseph 45 62 86 106 126 Ardmore 46 71 107 141 179
Montana Cherokee 44 59 80 97 113
Great Falls 4 8 14 20 26 Guthrie 47 70 105 134 163
Miles City 7 12 21 29 38 McAlester 54 82 127 165 209
Nebraska Tulsa 47 69 100 127 154
Antioch 19 26 36 45 52 Oregon
Lincoln 36 51 74 92 112 Portland 6 9 13 15 18
Lynch 26 37 54 67 82 Pennsylvania
North Platte 25 38 59 78 99 Franklin 17 24 35 45 54
Scribner 38 53 76 96 116 Harrisburg 19 25 35 43 51
Valentine 18 28 45 61 77 Philadelphia 28 39 55 69 81
New Hampshire Pittsburgh 23 32 45 57 67
Concord 18 27 45 62 79 Reading 28 39 55 68 81
New Jersey Scranton 23 32 44 53 63
Atlantic City 39 55 77 97 117 Puerto Rico
Marlboro 39 57 85 111 136 San Juan 57 87 131 169 216
Trenton 29 48 76 102 131 Rhode Island
New Mexico Providence 23 34 52 68 83
Albuquerque 4 6 11 15 21 South Carolina
Roswell 10 21 34 45 53 Charleston 74 106 154 196 240
New York Clemson 51 73 106 133 ' 163
Albany 18 26 38 47 56 Columbia 41 59 85 106 132
Binghamton 16 24 36 47 58 Greenville 44 65 96 124 153
Buffalo 15 23 36 49 61 South Dakota
Marcellus 16 24 38 49 62 Aberdeen 23 35 55 73 92
Rochester 13 22 38 54 75 Huron 19 27 40 50 61
Salamanca 15 21 32 40 49 Isobel 15 24 38 52 67
Syracuse 15 24 38 51 65 Rapid City 12 20 34 48 64
North Carolina Tennessee
Asheville 28 40 58 72 87 Chattanooga 34 49 72 93 114
Charlotte 41 63 100 131 164 Knoxville 25 41 68 93 122
Greensboro 37 51 74 92 113 Memphis 43 55 70 82 91
Raleigh 53 77 110 137 168 Nashville 35 49 68 83 99
Wilmington 59 87 129 167 206 Texas
North Dakota Abilene 31 49 79 103 138
Devils Lake 19 27 39 49 59 Amarillo 27 47 80 112 150
Fargo 20 31 54 77 103 Austin 51 80 125 169 218
Williston 11 16 25 33 41 Brownsville 73 113 181 245 312
Ohio Corpus Christi 57 79 114 146 171
Cincinnati 27 36 48 59 69 Dallas 53 82 126 166 213
Cleveland 22 35 53 71 86 Del Rio 44 67 108 144 182
Columbia 20 26 35 41 48 El Paso 6 9 15 19 24
Columbus 27 40 60 77 94 Houston 82 127 208 275 359
Coshocton 27 45 77 108 143 Lubbock 17 29 53 77 103
Dayton 21 30 44 57 70 Midland 23 35 52 69 85
Toledo 16 26 42 57 74 Nacogdoches 77 103 138 164 194
San Antonio 57 82 122 155 193

(continues)
R Factor Information 547

Table 8A.2 (Continued)

Index values normally exceeded once in (year) Index values normally exceeded once in (year)

Location 1 2 5 10 20 Location 1 2 5 10 20

Temple 53 78 123 162 206 West Virginia


Victoria 59 83 116 146 178 Elkins 23 31 42 51 60
Wichita Falls 47 63 86 106 123 Huntington 18 29 49 69 89
Vermont Parkersburg 20 31 46 61 76
Burlington 15 22 35 47 58 Wisconsin
Virginia Green Bay 18 26 38 49 59
Blacksburg 23 31 41 48 56 LaCrosse 46 67 99 125 154
Lynchburg 31 45 66 83 103 Madison 29 42 61 77 95
Richmond 46 63 86 102 125 Milwaukee 25 35 50 62 74
Roanoke 23 33 48 61 73 Rice Lake 29 45 70 92 119
Washington Wyoming
Spokane 3 4 7 8 11 Casper 4 7 9 11 14
Cheyenne 9 14 21 27 34

a
(After Wischmeier and Smith, 1978). Units on R are ft tonsf*in./acre»hr«storm. To convert to metric, MJ»mm/ha*h«storm, multiply by 17.02.
e
Appendix 8B
Universal Soil Loss Equation C Factors

Table 8B. 1 Typical C Factor Values Reported in the Literature for


Construction Sites and Disturbed Lands (after Israelson et al., 1980)

Condition C factor

1. Bare soil conditions


Freshly disked to 6 - 8 in. 1.00
After one rain 0.89
Loose to 12 in. smooth 0.90
Loose to 12 in. rough 0.80
Compacted root raked 1.20
Compacted bulldozer scraped across slope 1.20
Same except root raked across 0.90
Rough irregular tracked all directions 0.90
Seed and fertilize, fresh, unprepared seedbed 0.64
Same except after 6 months 0.54
Seed, fertilize after 12 months 0.38
Undisturbed except scraped 0.66-1.30
Scarified only 0.76-1.31
Sawdust 2 in. deep, disked in 0.61
2. Asphalt emulsion
1210 gal/acre 0.01-0.019
605 gal/acre 0.14-0.57
302 gal/acre 0.28-0.60
3 . Dust binder
605 gal/acre 1.05
1210 gal/acre 0.29-0.78
4. Other chemicals
Aquatain 0.68
Aerospray 7 0 , 1 0 % cover 0.94
PVA 0.71-0.90
Terra-Tack 0.66
5. Seedings a

Temporary, 0 to 6 0 days* 0.40


Temporary, after 60 days 0.05
Permanent, 2 to 12 months 0.05
6. Brush 0.35

"If plantings are used with mulches, use the minimum C values.
lf dry weather occurs at planting and emergence is a problem, extend the
b

0 - 6 0 days to a period when rainfall normally occurs.


550 Appendix 8B
Table 6B.2 C Factors for Permanent Pasture, Rangeland, Idle Land, and Grazed Woodlands
(after Wischmeier and Smith, 1978)°

Vegetal canopy Cover that contacts the surface


Percentage ground cover
Type and height of Canopy
raised canopy* cover (%) Type? 0 20 40 60 80 95-100

N o appreciable canopy G 0.45 0.20 0.10 0.042 0.013 0.003


W 0.45 0.24 0.15 0.090 0.043 0.011

Canopy of tall weeds or 25 G 0.36 0.17 0.09 0.038 0.012 0.003


short brush (0.5-m fall height) W 0.36 0.20 0.13 0.082 0.041 0.011
50 G 0.26 0.13 0.07 0.035 0.012 0.003
W 0.26 0.16 0.11 0.075 0.039 0.011
75 G 0.17 0.10 0.06 0.031 0.011 0.003
W 0.17 0.12 0.09 0.067 0.038 0.011

Appreciable brush or bushes 25 G 0.40 0.18 0.09 0.040 0.013 0.003


(2-m fall height) W 0.40 0.22 0.14 0.085 0.042 0.011
50 G 0.34 0.16 0.085 0.038 0.012 0.003
W 0.34 0.19 0.13 0.081 0.041 0.011
75 G 0.28 0.14 0.08 0.036 0.012 0.003
W 0.28 0.17 0.12 0.077 0.040 0.011

Trees, but no appreciable 25 G 0.42 0.19 0.10 0.041 0.013 0.003


low brush (4-m fall height) W 0.42 0.23 0.14 0.087 0.042 0.011
50 G 0.39 0.18 0.09 0.040 0.013 0.003
W 0.39 0.21 0.14 0.085 0.042 0.011
75 G 0.36 0.17 0.09 0.039 0.012 0.003
W 0.36 0.20 0.13 0.083 0.041 0.011

α
Α11 values shown assume: (1) random distribution of mulch or vegetation and (2) mulch of appreciable depth where it
exists. Idle land refers to land with undisturbed profiles for at least a period of 3 consecutive years. Also to be used for
burned forest land and forest land that has been harvested less than 3 years ago.
^Average fall height of waterdrops from canopy to soil surface in meters.
Tortion of total surface area that would be hidden from view by canopy in a vertical projection (a bird's-eye view).
^G, cover at surface is grass, grasslike plants, decaying compacted duff, or litter at least 2 in. deep. W, cover at surface
is mostly broadleaf herbaceous plants (as weeds with little lateral root network near the surface) and/or undecayed residue.
Universal Soil Loss Equation C Factors 551
Table 8B.3 C Factors for Mechanically Prepared Woodland Sites (after Wischmeier and Smith, 1978)

Soil condition and weed cover* 7

Excellent Good Fair Poor


Percentage of soil covered with
residue in contact with soil surface NC WC NC WC NC WC NC WC

None
A. Disked, raked, or bedded 0
0.52 0.20 0.72 0.27 0.85 0.32 0.94 0.36
B. Burned* 0.25 0.10 0.26 0.10 0.31 0.12 0.45 0.17
C. Drum chopped* 0.16 0.07 0.17 0.07 0.20 0.08 0.29 0.11

10% cover
A. Disked, raked or bedded 0
0.33 0.15 0.46 0.20 0.54 0.24 0.60 0.26
B. Burned* 0.23 0.10 0.24 0.10 0.26 0.11 0.36 0.16
C. Drum chopped* 0.15 0.07 0.16 0.07 0.17 0.08 0.23 0.10

20% cover
A. Disked, baked or bedded 0
0.24 0.12 0.34 0.17 0.40 0.20 0.44 0.29
B. Burned* 0.19 0.10 0.19 0.10 0.21 0.11 0.27 0.14
C. Drum chopped* 0.12 0.06 0.12 0.06 0.14 0.07 0.18 0.09

40% cover
A. Disked, raked or bedded 0
0.17 0.11 0.23 0.14 0.27 0.17 0.30 0.19
B. Burned* 0.14 0.09 0.14 0.09 0.15 0.09 0.17 0.11
C. Drum chopped* 0.09 0.06 0.09 0.06 0.10 0.06 0.11 0.07

60% cover
A. Disked, raked, or bedded 0
0.11 0.08 0.15 0.11 0.18 0.14 0.20 0.15
B. Burned* 0.08 0.06 0.09 0.07 0.10 0.08 0.11 0.08
C. Drum chopped* 0.06 0.05 0.06 0.05 0.07 0.05 0.07 0.05

80% cover
A. Disked, raked, or bedded 0
0.05 0.04 0.07 0.06 0.09 0.08 0.10 0.09
B. Burned* 0.04 0.04 0.05 0.04 0.05 0.04 0.06 0.05
C. Drum chopped* 0.03 0.03 0.03 0.03 0.03 0.03 0.04 0.04

''Multiply A values by following values to account for surface roughness: 0.40, Very rough, major effect on runoff and sedi­
ment, storage depressions greater than 6 in.; 0.65, moderate; 0.90, smooth, minor surface sediment storage, depressions less
than 2 in. The C values for A are for the first year following treatment. For Λ-type sites 1 to 4 years old, multiply C value by 0.7
to account for aging. For sites 4 to 8 years old, use Table 8B.2. For sites more than 8 years old, use Table 8B.4.
*The C values for Β and C areas are for the first 3 years following treatment. For sites treated 3 to 8 years ago, use Table
8B.2. For sites treated more than 8 years ago, use Table 8B.4.
T^C, no weed cover, WC, weed cover.
552 Appendix 8B

Table 8B.4 C Factors for Undisturbed Woodlands


(after Wischmeier and Smith, 1978)

Effective canopy 0
Forest litter*
(% of area) (% of area) C factor 0

100-75 100-90 0.0001-0.001


75-40 85-70 0.002-0.004
35-20 70-40 0.003-0.009

e
W h e n effective canopy is less than 20%, the area will be consid­
ered as grassland or idle land for estimating soil loss. Where wood­
lands are being harvested or grazed, use Table 8B.2.
^Forest litter is assumed to be at least 2 in. deep over the percentage
ground surface area covered.
'The range in C values is due in part to the range in the percentage
area covered. In addition, the percentage of effective canopy and its
height has an effect. Low canopy is effective in reducing raindrop
impact and in lowering the C factor. High canopy, over 13 m, is not
effective in reducing raindrop impact and will have no effect on the C
value.

Table 8B.5 USLE Mulch Factors and Length Limits for Construction Sites
(after Wischmeier and Smith, 1978) a

Type of Mulch rate Land slope Length limit*


mulch (ton/acre) (%) Factor C (ft)

Straw or hay, tied 1.0 1-5 0.20 200


down by anchoring 1.0 6-10 0.20 100
and tacking equipment 1.5 1-5 0.12 300
1.5 6-10 0.12 150
2.0 1-5 0.06 400
2.0 6-10 0.06 200
2.0 11-15 0.07 150
2.0 16-20 0.11 100
2.0 21-25 0.14 75
2.0 26-33 0.17 50
2.0 34-50 0.20 35

Wood chips 7.0 <16 0.08 75


7.0 16-20 0.08 50
12.0 <16 0.05 150
12.0 16-20 0.05 100
12.0 21-33 0.05 75
25.0 <16 0.02 200
25.0 16-20 0.02 150
25.0 21-33 0.02 100
25.0 34-50 0.02 75

"Developed by an interagency workshop group on the basis of field experience and limit­
ed research data.
^Maximum slope length for which the specified mulch rate is considered effective. Wnen
this limit is exceeded, either a higher application rate or mechanical shortening of the effec­
tive slope length is required.
c
When the straw or hay mulch is not anchored to the soil, C values on moderate or steep
slopes having Κ values greater than 0.30 should be taken at double the values given in this
table.
Appendix 8C
Supplemental C and Ρ Parameters
for Revised Universal Soil Loss Equation

Table 8C.1 Typical Crop Parameter Values for Residue Decomposition in Eqs. (8.54) and (8.58)
(after Yoder era/., 1993)*

Residue Row Plant


to grain a d
Yield* spacing population
U/R* w
Crop ratio a
(acre/lb) per acre (in.) (plants/acre)

Alfalfa 0.15 0.0040 30 0.00056 6 tons drilled 180,000


Brome grass 0.15 0.0040 80 0.00056 4 tons 7 drilled 330,000
Corn 1 0.0017 62 0.00038 130 bu 30 25,000
Cotton 1 0.0022 40 0.00022 9001b 38 35,000
Oats 2 0.0015 80 0.00106 65 bu 7 drilled 890,000
Peanuts 1.3 0.0022 30 0.00033 2600 lb 36 58,000
Rye 1.5 0.0040 80 0.00056 30 bu drilled 890,000
Sorghum 1 0.0022 60 0.00034 65 bu 30 41,000
Soybeans 1.5 0.0022 31 0.00058 35 bu 30 110,000
Sunflowers 1.5 0.0012 39 0.00024 11001b 30 20,000
Tobacco 1.8 0.0030 80 0.00034 22001b 48 6,000
Wheat (spring) 1.7 0.0018 107 0.00060 30 bu 7 drilled 890,000
Wheat (winter) 1.3 0.0018 107 0.00060 45 bu 7 drilled 890,000
Wheat (PNW) 1.3 0.0018 107 0.00060 70 bu 7 drilled 890,000

*Larson era/. (1978).


^Constant for the effect of residue size on decomposition.
c
Carbon to nitrogen ratio of the residue at harvest.
''Ratio of area covered by residue to its weight.
Typical yield value for the crop indicated.

553
Table 8C.2 Typical Field Operations and Associated RUSLE Parameter Values
(after Yoder etal., 1993)

R K Percentage buried
random residue Tillage
Field operations roughness (in.) (%) (in.

Moldboard (8 in. deep) 1.9 90 8


Moldboard (5-7 in. deep) 1.3 70 6
Chisel (2-in.shovels) 0.9 25 8
One-way disk (24 to 26 in. disks) 1.2 50 4
One-way disk (18 to 24 in. disks) 1.1 40 4
Large offset disk 2.0 50 6
Blades (36-in. width) 0.5 10 4
Sweeps (24-36 in.) 0.65 15 4
Field cult, (shovels) 0.7 25 4
Field cult. (16-18 in. sweeps) 0.6 20 4
Tandem disk 0.75 50 4
Harrow (tine) 0.4 15 4
Harrow (spike) 0.3 20 4
Fertilizer applicator 0.5 10 4
Rodweeder (plain) 0.4 10 4
Rodweeder (shovel) 0.45 15 4
Broadcast planting 0.4 0 0
Conventional drill 0.4 10 2
Semi-deep furrow drill 0.45 15 2
Deep furrow drill 0.5 20 3
No-till drill 0.4 10 2
Row planter 0.4 15 4
No-till planter 0.4 5 2
Row cultivator 0.6 30 4
Lister 1.0 80 4
Mulch treader 0.6 90 4
Supplemental C and Ρ Parameters for Revised Universal Soil Loss Equation 555
Table 8C.3 Growing Season Soil Moisture Depletion
and Replenishment Rates for Arid Western Drylands
(after Yoder etal., 1993)

A. Depletion rates

Depletion per
Crop 15-day period

Winter Wheat and 1.0


Other Deep Rooted Crops
Spring Barley 0.7
Spring Peas and Lentils 0.67
Shallow Rooted Crops 0.50
Fallow 0.0

B. Replenishment rates

Replenishment rate _
0.5 + 0.662 ( / » - ! )
per 15-day period

10£P£18

P, annual precipitation in inches. Replenishment rate is


inches per 15-day period.

Table 8C.4 RUSLE Terrace Factor Values Ρ Λ for Conserva­


tion Planning (after Foster et al, 1993)*

Open outlet, with percentage grade of 6

Horizontal terrace Closed


interval (ft) outlets* 7
0.1-0.3 0.40-0.7 >0.8

Less than 110 0.5 0.6 0.7 1.0


110-140 0.6 0.7 0.8 1.0
14O-180 0.7 0.8 0.9 1.0
180-225 0.8 0.8 0.9 1.0
225-300 0.9 0.9 1.0 1.0
More than 300 1.0 1.0 1.0 1.0

Multiply these values by other Ρ factor values for contouring, strip


cropping, or other supporting practices on the interterrace interval to obtain
a composite P factor value.
*The channel grade is measured on the first 300 ft of terrace or the one-
third of total length closest to the outlet, whichever distance is less.
c
Values for closed outlet terraces also apply to terraces with under­
ground outlets and to level terraces with open outlets.
Appendix 8D
Supplemental C Subfactors
for Disturbed Forests

557
558 Appendix 8D
Table 8D. 1 Effect of bare soil, fine root mat of trees, and soil
reconsolidation on C factor (after Dissmeyer and Foster, 1984).

A. Untitled Soils
Percentage of bare soil with dense mat of fine
roots in top 3 cm of soil
Percentage
bare soil 100 80 60 40 20 0

0 0.0000 0.000 0.000 0.000 0.000 0.000


1 0.0004 0.0005 0.0007 0.0010 0.0014 0.0018
5 0.003 0.003 0.005 0.007 0.009 0.012
10 0.005 0.006 0.009 0.013 0.017 0.0123
20 0.011 0.014 0.020 0.028 0.038 0.050
40 0.023 0.027 0.042 0.058 0.079 0.104
60 0.037 0.043 0.067 0.092 0.127 0.167
80 0.055 0.066 0.098 0.141 0.192 0.252
100 0.099 0.117 0.180 0.248 0.342 0.450

B. Tilled soils with good initial fine root mat in topsoil and subsoil has
good structure and permeability
Time (months) since tillage
Percentage
bare soil 0 6 12 and 72+ 24+ thru 6 0

0 0.0000 0.0000 0.0000 0.0000


1 0.0014 0.0017 0.0018 0.0020
5 0.009 0.011 0.012 0.013
10 0.019 0.022 0.023 0.026
20 0.037 0.045 0.049 0.056
40 0.083 0.095 0.104 0.117
60 0.137 0.157 0.172 0.194
80 0.212 0.244 0.267 0.301
100 0.360 0.414 0.450 0.510

C. Tilled soil with poor initial fine root mat in topsoil (subsoil has good
structure and permeability)
Time (months) since tillage
Percentage
bare soil 0 6 12 to 36 48 72+

0 0.0000 0.0000 0.0000 0.0000 0.0000


1 0.0021 0.0023 0.0025 0.0022 0.0018
5 0.014 0.015 0.016 0.015 0.0117
10 0.027 0.031 0.033 0.029 0.023
20 0.058 0.065 0.069 0.060 0.049
40 0.122 0.135 0.144 0.129 0.104
60 0.201 0.224 0.239 0.213 0.171
80 0.313 0.348 0.352 0.330 0.266
100 0.530 0.590 0.630 0.560 0.450
Supplemental C Factors for Disturbed Forests 559

Table 8D.2 Step effect on soil erosion (after Dissmeyer and Foster, 1984).

Percentage of total slope in steps


Percentage
slope 0 10 20 30 40 50 60 70 80 90 100

5 1.00 0.99 0.99 0.98 0.97 0.96 0.95 0.94 0.94 0.93 0.92
6 0.97 0.94 0.92 0.89 0.86 0.84 0.81 0.78 0.76 0.73
i.oo
7 1.00 0.96 0.92 0.88 0.84 0.80 0.75 0.71 0.67 0.63 0.59
8 1.00 0.95 0.90 0.85 0.80 0.75 0.69 0.64 0.59 0.54 0.49
9 1.00 0.94 0.89 0.83 0.77 0.71 0.65 0.60 0.54 0.48 0.42
10 1.00 0.94 0.87 0.81 0.75 0.68 0.62 0.56 0.49 0.43 0.36
12 1.00 0.93 0.85 0.78 0.71 0.63 0.56 0.49 0.42 0.34 0.27
15 1.00 0.92 0.84 0.75 0.67 0.59 0.51 0.43 0.34 0.26 0.18
20 1.00 0.91 0.82 0.74 0.65 0.56 0.47 0.38 0.29 0.20 0.11
30+ 1.00 0.91 0.81 0.72 0.63 0.53 0.44 0.35 0.25 0.15 0.06

Table 8D.3 Contour tillage subfactors for forestlands


(after Dissmeyer and Foster, 1984).

Degrees off contour


Percentage On contour
slope 0 15 30 45 60 90

0-2 0.80 0.88 0.91 0.94 0.96 1.00


3-7 0.70 0.82 0.87 0.91 0.94 1.00
8-12 0.80 0.88 0.91 0.94 0.96 1.00
13-18 0.90 0.94 0.96 0.97 0.98 1.00
19+ 1.00 1.00 1.00 1.00 1.00 1.00
Appendix 8E
Roughness Values and Critical Tractive Force
Values for CREAMS Equations

0.045 Void Ratio


CM
Ο · - 0.8
\* 40.0 L
0.040 • .
Δ · - 1.0
m 1 2
Δ
0)
09
• < - 1.4
CO
20.0 L 0.035 - 1.6 ^ / Δ

m Ο
u
u
10.0 L
ο 0.030
ο fa
+>
ο
ed
0.025 y

fa
ο 0.020
ed
ο
33
u
Η
cd
ο 0.015
a* n
ο 0.010
u
Class for Tillage 0.005
u
a n d Time Since Tillage 10 15 20 25 30 35
Figure 8E.1 Effect of tillage on critical shear stress. Class 1, long-
term without tillage; Class 2, 1 year since seedbed tillage; Class 3,
Plasticity Index
primary tillage in land 1 year since seedbed; Class 4, typical seedbed, Figure 8E.2 Effect of plasticity index and void ratio on critical
Class 5, finely pulverized seedbed (after Foster et al., 1980b). shear force (after Lyle and Smerdon, 1965).

561
562 Appendix 8E

Table 8E. 1 Overland Flow η Values for CREAMS Equations (after Foster et al, 1980b)

Treatment Manning's η Treatment Manning's η

Cornstalk residue applied to fallow surface Grass


1 ton/acre 0.020 Sparse 0.015
2 tons/acre 0.040 Poor 0.023
4 tons/acre 0.070 Fair 0.032
Good 0.046
Cornstalk residue disk-harrow incorporated Excellent 0.074
1 ton/acre 0.012
2 tons/acre 0.020 Dense 0.150
4 tons/acre 0.023 Very dense 0.400

Wheat straw mulch Rough surface depressions


0.25 ton/acre 0.015 4 to 5 in. deep 0.046
0.5 ton/acre 0.018 2 to 4 in. deep 0.023
1 ton/acre 0.032 1 to 2 in. deep 0.014
2 tons/acre 0.070 N o surface depressions 0.010
4 tons/acre 0.074

Crushed stone mulch


15 tons/acre 0.012
6 0 tons/acre 0.023
135 tons/acre 0.046
240 tons/acre 0.074
375 tons/acre 0.074

Small grain
(20% to full maturity) Across slope Upslope and downslope

Poor stand 0.018 0.012


Moderate stand 0.023 0.015
Good stand 0.032 0.023
Dense 0.046 0.032
Roughness Values and Critical Tractive Force Values for CREAMS Equations 563

Table 8E.2 Concentrated Flow Values for Manning's η for Typical Soil
Covers' for CREAMS Equation (after Foster et al., 1980b)
1

Cover Cover density Manning's η

Smooth, bare soil; roughness elements Less than 1 in. deep 0.030
1-2 in. deep 0.033
2 - 4 in. deep 0.038
4 - 6 in. deep 0.045

Corn stalks (assumes residue stays 1 ton/acre 0.050


in place and is not washed away) 2 tons/acre 0.075
3 tons/acre 0.100
4 tons/acre 0.130

Wheat straw (assumes residue stays 1 ton/acre 0.060


in place and is not washed away 1.5 tons/acre 0.100
2 tons/acre 0.150
4 tons/acre 0.250

Grass (assumes grass is erect and Sparse 0.040


as deep as the flow) Poor 0.050
Fair 0.060
Good 0.080
Excellent 0.130
Dense 0.200
Very dense 0.300

Small grain (20% to full Poor, 7-in. rows 0.130


maturity rows with flow) Poor, 14-in. rows 0.130
Good, 7-in. rows 0.300
Good, 14-in. rows 0.200

Rows across floors Good 0.300

Sorghum and cotton Poor 0.070

Good 0.090

Sudangrass Good 0.200

Lespedeza Good 0.100

Lovegrass Good 0.150


"Does not include effects of submergence or product of velocity-hydraulic radius.
564 Appendix 8E

Table 8E.3 Critical Shear Stress Values as a Function of


Tillage and Consolidation for Moderately Erodible Agricultural
Soils (after Foster et al, 1980b) fl

Critical shear stress


Tillage-consolidation condition (lb/ft )2

Moldboard plowed 0.20


Chisel or disk for primary tillage 0.5
Disking: common corn seedbed or crop cultivation 0.10
Finely pulverized seedbed 0.05
1 month after last tillage of common seedbed 0.20
2 months after last tillage of common seedbed 0.30
3 months after last tillage of common seedbed 0.40
Long term, undisturbed 0.60

°Ίο be used as a first estimate when soil information is not available.


For soils with dispersion ratio data available, Foster et al. recommend
that the Smerdon and B e a s l e y ( 1 9 6 1 ) equation be u s e d , or x =
c

0.213/dJ*63. These values must be corrected for tillage by factors from


Fig. 8E.1.
Appendix 8F
Conveyance Function and Equilibrium Channel
Properties for Concentrated Flow Equations

565
566 Appendix 8F

Table 8F. 1 Concentrated flow channel parameters (after Lewis, 1990).*

Item w m R* s(*. )
c

1 0.00 0.7436 0.1510 0.2523 35.0000 —


2 0.01 0.7287 0.1524 0.2572 26.3603 73.75
3 0.02 0.7138 0.1535 0.2622 13.3985 29.21
4 0.03 0.6990 0.1544 0.2671 9.0856 17.19
5 0.04 0.6841 0.1552 0.2721 6.9351 11.89
6 0.05 0.6692 0.1558 0.2770 5.6497 9.00
7 0.06 0.6543 0.1562 0.2820 4.7971 7.21
8 0.07 0.6395 0.1565 0.2870 4.1918 6.00
9 0.08 0.6246 0.1565 0.2919 3.7413 5.14
10 0.09 0.6097 0.1564 0.2969 3.3941 4.50
11 0.10 0.5949 0.1562 0.3018 3.1194 4.01
12 0.11 0.5800 0.1557 0.3068 2.8974 3.62
13 0.12 0.5651 0.1551 0.3117 2.7152 3.31
14 0.13 0.5502 0.1542 0.3167 2.5637 3.06
15 0.14 0.5354 0.1533 0.3216 2.4365 2.85
16 0.15 0.5205 0.1521 0.3266 2.3288 2.66
17 0.16 0.5056 0.1507 0.3315 2.2371 2.53
18 0.17 0.4908 0.1492 0.3365 2.1587 2.40
19 0.18 0.4759 0.1475 0.3415 2.0915 2.29
20 0.19 0.4610 0.1457 0.3464 2.0340 2.20
21 0.20 0.4461 0.1436 0.3514 1.9849 2.12
22 0.21 0.4313 0.1414 0.3563 1.9430 2.05
23 0.22 0.4164 0.1390 0.3613 1.9076 1.99
24 0.23 0.4015 0.1364 0.3662 1.8781 1.95
25 0.24 0.3867 0.1336 0.3712 1.8539 1.90
26 0.25 0.3718 0.1307 0.3761 1.8345 1.87
27 0.26 0.3569 0.1276 0.3811 1.8197 1.83
28 0.27 0.3420 0.1243 0.3861 1.8092 1.81
29 0.28 0.3272 0.1208 0.3910 1.8029 1.79
30 0.29 0.3123 0.1172 0.3960 1.8005 1.77
31 0.30 0.2974 0.1134 0.4009 1.8021 1.76
32 0.31 0.2826 0.1094 0.4059 1.8076 1.76
33 0.32 0.2677 0.1052 0.4108 1.8172 1.76
34 0.33 0.2528 0.1009 0.4158 1.8309 1.76
35 0.34 0.2379 0.0964 0.4207 1.8490 1.77
36 0.35 0.2231 0.0917 0.4257 1.8718 1.78
37 0.36 0.2082 0.0868 0.4307 1.8998 1.80
38 0.37 0.1933 0.0817 0.4356 1.9334 1.82
39 0.38 0.1785 0.0765 0.4406 1.9736 1.85
40 0.39 0.1636 0.0711 0.4455 2.0212 1.89
41 0.40 0.1487 0.0655 0.4505 2.0778 1.93
42 0.41 0.1338 0.0598 0.4554 2.1454 1.98
43 0.42 0.1190 0.0538 0.4604 2.2267 2.05
44 0.43 0.1041 0.0477 0.4654 2.3260 2.13
45 0.44 0.0892 0.0414 0.4703 2.4497 2.24
46 0.45 0.0743 0.0350 0.4753 2.6083 2.37
47 0.46 0.0595 0.0283 0.4802 2.8209 2.56
48 0.47 0.0446 0.0215 0.4852 3.1272 2.82
49 0.48 0.0296 0.0144 0.4903 3.6289 3.26
50 0.49 0.0145 0.0072 0.4954 4.7181 4.19
51 0.50 0.0000 0.0000 0.5000 5.7322 —

g(X.) values less than 1.8 indicate insufficient tractive forces to cause scour and the forma­
tion of an incised channel. For selected values of g(X* ), X. is a double-valued function. Use
c c

of the smaller value will yield a conservatively large erosion estimate.


Appendix 9A
Predicting Turbulence in Sediment Ponds

Characteristics of Interest Conversely, a parabolic velocity would result in (Dob­


bins, 1944)
Turbulent characteristics a r e important t o prediction
and flocculation of sediment. P a r a m e t e r s of interest dU l-y/H
are mean square velocities a n d turbulent diffuswities.
dy Η
Diffusivities a r e important for prediction of diffusion of
sediment. M e a n square velocities a r e important t o
w h e r e U is t h e velocity at t h e water surface. T h e
s
prediction of flocculation a n d dispersion.
resulting t u r b u l e n t diffusion coefficient for t h e
parabolic profile is given by
Models of Turbulence
HUl
Using t h e definition sketch in Fig. 9A.1 t h e m o m e n ­ ε = (9A.4)
t u m equation, for gradually varied flow in a sediment ~2ΪΓ
p o n d can b e written as (Wilson a n d Barfield, 1986b)
T h e turbulent diffusion coefficient is predicted by Eq.

™ - V » - U A 1 - h \ ( 9 A 1 )

where is t h e shear velocity a n d is equal to (gHS ) , t 05


X =Z;
3 U3 - W

u'u' is t h e Reynold's shear stress, ν is kinematic viscos­


ity, U is average velocity at any distance y above t h e
X, = X ; Uj = U
bed, a n d Η is t h e total depth. A n analogy between
Reynolds' stresses a n d turbulent diffusion has b e e n
typically assumed for most flows. Using this analogy,
the turbulent diffusion coefficient can b e estimated as

u υ
ε = (9A.2)
dU/dy 9

><2=Y; U2=V
where ε is t h e turbulent diffusion coefficient.
Using a logarithmic velocity profile, dU/dy =
U*/(ky\ where k is von K a r m a n ' s constant (ap­ Χϊ=Χ; U 1 = U
proximately 0.4), t h e turbulent diffusion coefficient can
thus b e predicted by

Ml-
DATUM-

ε = U*ky (9A.3)
Figure 9A.1 Definition sketch for the momentum equation.

567
568 Appendix 9A

(9A.3) for t h e log velocity model and by Eq. (9A.4) for T h e assumption of a horizontal water surface is a
the parabolic velocity model. standard approximation in reservoir routing. T h e as­
sumption that the average velocity is Q/A is also
Model Parameters reasonable unless major recirculation occurs.
A w a t e r surface velocity is required for the parabolic
T h e velocity profile models require that the energy
velocity model. Wilson and Barfield (1986a, b) assumed
slope be calculated. This p a r a m e t e r can be estimated
a value for U of 1.5 times the average velocity at the
s
at each point by assuming that (a) the water surface
cross section.
elevation is i n d e p e n d e n t of χ and (b) the velocity at a
cross section is equal to the inlet volumetric flow rate
divided by the corresponding cross-sectional area. U n ­ Model Evaluation
der these conditions, the slope of the energy gradient
becomes Wilson a n d Barfield (1986b) evaluated t h e accuracy
of t h e models o n d a t a collected in a model sediment
d(U /2)
2

p o n d . In general, t h e log model was m o r e accurate


5. = (9A.5)
dx t h a n t h e parabolic model.
Appendix 9B
Rainfall Statistics for EPA Model

Figure 9B.1 Regions for E P A model rainfall statistics (after Driscoll et al., 1986).

569
570 Appendix 9B

Table 9B. 1 Rainfall Statistics for EPA (1986) Model

Rainfall statistics

Q ι d r u

Volume (in.) Intensity (iph) Duration (hr) Interval (hr)

Zone Period Mean CV. Mean CV. Mean CV Mean CV.

1 Annual 0.28 1.46 0.051 1.31 5.8 1.05 73 1.07


Summer 0.32 1.38 0.082 1.29 4.4 1.14 78 1.07
2 Annual 0.36 1.45 0.066 1.32 5.9 1.05 77 1.05
Summer 0.40 1.47 0.101 1.37 4.2 1.09 77 1.08
3 Annual 0.49 1.47 0.102 1.28 6.2 1.22 89 1.05
Summer 0.48 1.52 0.133 1.34 4.9 1.33 68 1.01
4 Annual 0.58 1.46 0.097 1.35 7.3 1.17 99 1.00
Summer 0.52 1.54 0.122 1.35 5.2 1.29 87 1.06
5 Annual 0.33 1.74 0.080 1.37 4.0 1.07 108 1.41
Summer 0.36 1.71 0.110 1.39 3.2 1.08 112 1.49
6 Annual 0.17 1.51 0.045 1.04 3.6 1.02 277 1.48
Summer 0.17 1.61 0.080 1.16 2.6 1.01 425 1.26
7 Annual 0.48 1.61 0.024 0.84 20.0 1.23 101 1.21
Summer 0.26 1.35 0.027 1.11 11.4 1.20 188 1.15
8 Annual 0.14 1.42 0.031 0.91 4.5 0.82 94 1.39
Summer 0.14 1.51 0.041 1.1: 2.8 0.80 125 1.41
9 Annual 0.15 1.77 0.038 1.35 4.4 1.20 84 1.24
Summer 0.10 1.74 0.058 1.44 3.1 1.14 78 1.13
Appendix 9C
Equations for Predicting the Advance
of the Deposition Wedge in Vegetative Filter Strips

T h e deposition wedge in t h e grass filter strip is that tion wedge,


area where bedload material is deposited, as shown in
3.57
Figs. 9.40 and 9.41. T h e following material is a proce­ K
( ^sba^et)
d u r e to correct calculated trapping efficiencies for this tfsba ~ λ 2.07
(9C.3)
"pba
effect.

w h e r e Κ is a constant given by Eq. (9.119). Also


Deposition in Zone Bit)
Z o n e Bit) is the area w h e r e deposition occurs uni­ ^ mba = (HW3S1/2 ( f t / s e c ) ) ( 9 C 4 )

formly along a deposition wedge as shown in Fig. 9 C . 1 .


P a r a m e t e r s are estimated at t h e midpoint of zone
where
B(t). T h e incoming sediment load to zone Bit) is that
coming into the filter, q , since it is assumed that n o
si
s fbau

sediment is deposited in zone Ait). T h e sediment load (9C.5)


exiting Bit) is equal to the transport capacity for Bit),
q^, as given by Eq. (9.118) or (9.115). T h u s , the aver­
F r o m continuity, the flow p e r unit width, q , wba is given
age sediment load on the deposition wedge, q , should sba
as
b e the average of q and q , or
si sd

#wba Knba^fba* (9C.6)


4si + <?sd
#sba (9C.1)
It is assumed that n o infiltration occurs in Ait) and
Bit); hence
and the fraction of sediment t r a p p e d in zone Bit) is
given by #wba — (9C.7)

4si ~ 4sd
(9C.2) Given a value of q , o n e simultaneously solves Eqs.
sba

( 9 C . 3 M 9 C . 7 ) for 5 and d^. T h e resulting equation


e t

for d^ is
O t h e r than q
sba and / , the major variable of interest is
t h e slope of t h e deposition wedge, which will b e used 1/6
^fba^s
in subsequent equations to predict the location of the *fba = C1,
leading edge of the deposition wedge, Xit). T h e slope 2<*fba + Ss

of the deposition wedge in zone Bit) is different from


channel slope and is defined as the equilibrium slope, w h e r e C I is a constant given by
5 . Conceptually, it is t h e slope required for t h e flow
e t

tfwba
t o
transport the sediment load q . Given t h e xnq^K 014

sba
CI = 1 C ^ 0 . 1 4 ^0.2898 (9C.8)
value for q^, a value for 5 can b e calculated from a
e t A-^tfsba " p b
combination of sediment transport rate from Eq.
(9.118), velocity from (9.96), and continuity from (9.98). Κ is defined by Eq. (9.119), q is cfs/ft, q is l b / s e c ·
w s

Using these equations and substituting an additional ft, a n d d


pb is in mm. Equation (9C.8) must be solved by
subscript b a for variables that a r e unique to t h e deposi­ trial a n d error.

571
572 Appendix 9C

L o c a t i o n o f t h e L e a d i n g E d g e o f t h e Deposition 1/2
W e d g e X(t) *f('f) = — ^(if-O+^ii'i) 2

T h e location of t h e leading edge of the deposition


wedge determines t h e effective length, L ( i ) , of t h e for Y (t ){ { <H (9C.11)
suspended load deposition area, D(t\ as given in Eq.
Y (t ) = Η if Y (t ) >Hin E q . (9C.10) (9C.12)
(9.105). Its location can b e determined by sediment
f f f {

mass continuity relationships, as given by Tollner et al.


(1977). Using the definition given in Fig. 9.41 a n d t h e *f('f)=*f('f)M'f-0^
concept of mass continuity, Tollner et al. derived t h e ^r b S

following equations for the d e p t h of deposition a n d t h e


if y > Η in E q . ( 9 C . 1 0 ) (9C.13)
advance distance X(t\ or
f

Sc = S c t -S .
c (9C14)
s b
(9C.9)
2flsiS c
In t h e above terms, t* is t h e time required for deposi­
1/2 tion to reach t h e height of t h e media, H, y is t h e bulk sb

-{fQsSJ^itf-^ + YXtiY density of t h e deposited sediment, Y (t ) a n d X (t ) are


f { f {

the d e p t h of deposition a n d advance distance, respec­


for Y (t ) <H (9C.10) tively, at time t , Yj(fj) a n d -Yj(fj) a r e t h e d e p t h of
f
f {

deposition a n d advance distance, respectively, at an


Equations for Predicting the Advance of the Deposition Wedge in Vegetative Fitter Strips 573

earlier time t , and {/^ ba^etiavg indicates that these


{ S
Solution:
values are averaged on the time interval t - t^ {
1. Parameters. From Problem 9.15, 5 = .0525 ft, xn = S

T o calculate values for Y (tf) and X (t ) at any t ,


f f { {
.056, Κ = 3.68 Χ 10 , d = .17 mm, q = q = .074 cfs/ft,
7
pd wi wb

o n e divides the inflow hydrograph and sedimentgraph <? = .211 lb/sec · ft, q% = .32 lb/sec · ft. From Table 9.8,
c
sd

Η = 4 inch = .33 ft
into discrete time increments; calculates q , / , S ,
2. Determining values for equilibrium slope S in zone B(t)
sba e t

and S for each time increment; and then calculates e(


e
(coarse particle only). From Eq. (9C.8), using a superscript c
Y (t ) and X (t ) from Eqs. ( 9 C 9 ) - ( 9 C . 1 4 ) .
f { { {
to denote coarse particles,

Deposition in Zone 4(f) and Upstream of the VFS CI = 0 1 4


JO.2898

Z o n e A(t) is the region w h e r e sediment has de­


posited u p to the level of the V F S ; hence all sediment rfpb = rfpd = 0.17 mm.
that enters this zone flows into zone B(t). Backwater
effects from the deposition wedge in zone A(t) cause From Eq. (9C.1),
deposition upstream of the V F S ; as shown in Fig. 9C.1.
q$ + <7 c
sd 0.32 + 0.211
Studies by Hayes et al. (1984) showed that the slope of 4sba — = 0.266 lb/sec · ft
the upstream delta was zero. Using this fact and the
assumption that the d e p t h of the delta at the interface (0.056)(0.074)(3.68 Χ 10 ) 7

with the V F S equals the d e p t h of deposition given by r\ = — — - = η 064


1.5(0.266)° (0.17)° 14 2 8 9 8

Eq. (9C.10), Hayes et al. showed that the rate of


sediment deposition in the upstream delta was given by From Eq. (9C.8),
1/6
t < t; (9C.15)
2(i -ii)Sc
f <*fba = CI = 0.064
2<*fba + Ss

where y is bulk density of deposited material. T h e


s b

sediment load entering the V F S then is given by or


-,1/6
<7sa = ? s i - 4 s u > ' < ' * > (9C.16) (^ )(0.0525)
a

- 0.064 = 0.0.
where q is the sediment load coming from upslope. It
si
2d<u» + 0.0525
should b e noted that after the d e p t h of deposition
By trial and error, d is found to be 0.122 ft or 1.464 in.
reaches the height of the filter media, q goes to zero f d a

Solving for R and 5 ,


su

and
sb e t

4sa = <7si, t*t*. (9C.17) ^fda^s (0.122)(0.0525)


= 0.0216.
T h e fraction of material trapped in the u p s t r e a m 2
^ f d a + $s (2)(0.122) + 0.0525
delta can then be calculated as
From Eq. (9C.3), 5 e t can be calculated as
/ u = 4su/4si. (9C.18)

Again, it is important to point out that the only m a t e ­ . q&^K* 7


(0.266) (0.17) - 028 0 5798

rial assumed to be deposited in the upstream delta and i e t


* K° 2 8
(0.0216X3.68 X 1 0 ) 7 0 2 8
s b

zones A(t)-C(t) is coarse particles larger than 0 . 0 3 7


mm. Therefore, the equations for these zones are for 3.'Ratio between sediment deposition in upstream delta and
coarse particles only, requiring that the inflow sedi­ Bit). Prior to Υ = Η (t < t*) in Eq. (9C.9), it is necessary to
ment load b e divided into coarse and fine particles. correct / in Eq. (9C.10) for deposition in the upstream delta.
Procedures for accomplishing this were discussed in It can be shown from geometry that the ratio between
sediment deposited in the upstream delta ¥ j , to that in zone
the body of C h a p t e r 9 and illustrated in Example
B(t\ ¥ is
Problem 9 . 1 5 . 2 >

Set
= Of,

Example Problem 9C.1. Rate of advance


of the deposition wedge which is the ratio of sediment deposited in B(t) to the total
sediment deposited. Also
Estimate the advance distance for the sediment wedge in
Problem 9.15 at the end of 5 minutes. Assume y is 1.25 X sb

62.4 = 78 lb/ft . Assume that the VFS is mowed.


3
V, + ¥ 2
1 + a
574 Appendix 9C
The ratio / in Eq. (9C.2) must be modified by this ratio for was given as 0.32 lb/sec · ft from Example Problem 9.15.
time prior to r*. From earlier calculation, 5 = 0.0867 and e t
Using y = 78 l b / f t as given, 3
sb

5 is given as 0.08. Hence


C

0.0867 - 0.08 (0.33) (78) 2

a = — = 0.084. t* = = 79,237 sec or 1321 min.


(2)(0.025)(0.32)(0.0067)
0.08
If / is the total fraction of coarse particles trapped in the Obviously, at 5 min, the deposition had not reached the
deposition wedge, then the portion trapped in zone Bit) media height.
would be 5. Advance distance at time 5 min. t = 0, X^t) = 0. At x

/ a \ I 0.084 \ t = 300 sec, X (t) = ? At t - 5 min (300 sec), Y(t) < H\


{ f

hence Eq. (9C.11) applies. Using / ' instead of / in Eq.


J
\ 1+ α / \ 1 + 0.084 / (9C.11),
4. Time to deposit to top of vegetation. From Eqs. (9C.14),
1/2

5 = 5
e e t - S = 0.0867 - 0.08 = 0.0067.
c *f('f) -

From Eq. (9C.9), using / ' in place of / , 1/2


"/ 2 \ (0.025)(0.32)
" 78 0.0067 J
( 3
" ° - ° ) +
"
r= = 3.03 ft.
General Appendix
Common Equivalences

575
576 General Appendix

C o m m o n Equivalencies

Length

Equivalent

Unit Millimeter Inch Foot Meter Kilometer Mile

Millimeter 1 0.039 37 0.003 281 0.001 000 1 E-6 0.621 4 E - 6


Inch 25.40 1 0.083 3 0.025 40 25.40 E - 6 15.78 E - 6
Foot 304.8 12 1 0.304 8 304.8 E - 6 189.4 E - 6
Meter 1,000 39.37 3.281 1 0.001 621.4 E - 6
Kilometer 1,000,000 39,370 3,281 1,000 1 0.621 4
Mile 1,609,000 63,360 5,280 1,609 1.609 1

Area

Equivalent

Unit Square inch Square foot Square meter Acre Hectare Square kilometer Square Mile.

Square inch 1 0.006 944 645.2 E - 6 0.159 4 E - 6 64.52 E - 9 645.2 E - 1 2 249.1 E - 1 2


Square foot 144 1 0.092 90 22.96 E - 6 9.290 E - 9 92.90 E - 9 35.87 E - 9
Square meter 1,550 10.76 1 247.1 E - 6 1 E-4 1 E-6 386.1 E - 9
Acre 6,273,000 43,560 4,047 1 0.404 7 0.004 047 0.001 563
Hectare 15,500,000 107,600 10,000 2.471 1 0.01 0.003 861
Square kilometer 1.550 Ε + 9 10,764,000 1,000,000 247.1 100 1 0.386 1
Square mile 4.014 Ε + 9 27,880,000 2,590,000 640 259 2.590 1

Volume

Equivalent

Unit Cubic inch Liter U.S. gallon Cubic foot Cubic yard Cubic meter Acre-foot Second-foot-day

Cubic inch 1 0.016 39 0.004 329 578.7 E - 6 21.43 E - 6 16.39 E - 6 13.29 E - 9 6.698 E - 9
Liter 61.02 1 0.264 2 0.035 31 0.001 308 0.001 810.6 E - 9 408.7 E - 9
U.S. gallon 231.0 3.785 1 0.133 7 0.004 951 0.003 785 3.068 E - 6 1.547 E^6

Cubic foot 1,728 28.32 7.481 1 0.037 04 0.028 32 22.96 E - 6 11.57 E^6
Cubic yard 46,660 764.6 202.0 27 1 0.764 6 619.8 E - 6 312.5 E - 6
Cubic meter 61,020 1,000 264.2 35.31 1.308 1 810.6 E - 6 408.7 E - 6

Acre-foot 75.27 E + 6 1,233,000 325,900 43,560 1,613 1,233 1 0.504 2

Second-foot-day 149.3 E+6 2,447,000 646,400 86,400 3,200 2,447 1.983 1


Common Equivalencies 577

Discharge (Flow Rate, Volume^ime)

Equivalent

Unit Gal/min Liter/sec Acre-ft/day Ft /sec


3
Million gal/day M /sec
3

Gal/min 1 0.06309 0.004419 0.002228 0.001440 63.09 E^6


Liter/sec 15.85 1 0.07005 0.03531 0.02282 0.001
Acre-ft/day 226.3 14.28 1 0.5042 0.3259 0.01428
Ft /sec
3
448.8 28.32 1.983 1 0.6463 0.02832
Million gal/day 694.4 43.81 3.069 1.547 1 0.04381
M /sec
3
15,850 1,000 70.04 35.31 22.82 1

Velocity

Equivalent

Unit Ft/day Km/hr Ft/sec Mile/hr M/sec

Ft/day 1 12.70 E - 6 11.57 E - 6 7.891 E - 6 3.528 E - 6


Km/hr 78,740 1 0.9113 0.6214 0.2778
Ft/sec 86,400 1.097 1 0.6818 0.03048
Mile/hr 126,700 1.609 1.467 1 0.4470
M/sec 283,500 3.600 3.281 2.237 1

Mass

Equivalent

Unit Pound mass Kilogram Metric slug Slug Metric ton Long ton

1 0.4536 0.04625 0.03108 453.6 E - 6 446.4 E - 6


Kilogram 2.205 1 0.1020 0.06852 0.001 984.2 E - 6
Metric slug 21.62 9.807 1 0.6721 0.009807 0.009651
Slug 32.17 14.59 1.490 1 0.01459 0.01436
Metric ton 2,205 1,000 102.0 68.52 1 0.9842
Long ton 2,240 1,016 103.7 69.63 1.016 1
578 General Appendix
Conversion Factors

Temperature Pressure
Celcius + 273.15 = Kelvin Atmosphere x 1.013 Ε 05 = pascal
Fahrenheit + 459.67 + 1 . 8 = Kelvin B a r x 1.000 Ε 0 5 = pascal
Rankline + 1.8 = Kelvin Feet of water (39.4°F) χ 2989 = pascal
Fahrenheit - 32 + 1.8 = Celcius Inches mercury χ 3386 = pascal
Celcius x 1.8 + 32 = Fahrenheit Inches water x 249.1 = pascal
Kilogram (force) per square meter χ 9.807 = pascal
Viscosity Millibar χ 100 = pascal
Centipoise χ 0.001 = pascal-seconds Millimeter mercury χ 133.3 = pascal
Centistokes χ 1.000 E - 0 6 = square meter per second Pounds per square foot χ 47.88 = pascal
Square foot per second χ 0.09290 = square meter per second Pounds per square inch χ 6895 = pascal
Poise χ 0.100 = pascal-seconds
Pound (force) -second per square foot χ 47.88 = pascal-seconds
Slug/foot-second χ 47.88 = pascal-seconds
Stokes χ 1.000 E - 0 4 = square meter per second

Conversion Factors for Universal Soil Loss Equation (USLE) Factors

To convert from: U.S. customary units Multiply by: To obtain: SI units*

Rainfall intensity, / or / in./hr 25.4


hr

fftonf megajoule MJ
Rainfall energy per unit of rainfall, e 2.638 x 10" 4

icre-in. hectare«millimeter

ft'tonf
Storm energy, Ε 0.006701

tnegajoute*miUiineter
Storm erosivity, El 0.1702
acre · hr hectare* hour

hundreds offt*tonf*in* megajoulernillimeter


Storm erosivity, El 17.02

hundreds of ft»tonf»in. megajoulerni 11 imeter


Annual erosivity, R 17.02
scre*hr*yev hectare*hour»year
ton*acre*hr
Soil erodibility, K c
0.1317
hundreds of acre*ft*tonf*in. hectare«megajoule*milUmeter

Soil loss, A 2.242

Soil loss, A 0.2242 kg/m 2

'The prefix mega (M) has a multiplication factor of 1 x 10 . To convert ft-tonf to MJ, multiply by 2.712 χ 10" . To convert to hectare, multiply by
6 3

0.4071.
*This notation, "hundreds of," means numerical values should be multiplied by 100 to obtain true numerical values in given units. For example, R = 125
(hundreds of ft-ton-in./acre-hr) = 12,500 ft-tonf-in./acre-hr. The converse is true for "hundreds of" in the denominator of a fraction. Erosivity, El or fl, can
be converted from a value in U.S. customary units to a value in units of newtons per hour (N/hr) by multiplying by 1.702.
c
Soil erodibility, K, can be converted from a value in U.S. customary units to a value in units of metric ton»hectare per newton*hour (ton»hr/ha«n) by
multiplying by 1.317.
Common Equivalencies 579

Properties of Water

Specific Kinematic
Temperature weight Density Viscosity viscosity

English units
Ί μ χ 10 v x 10
5 5
Ρ
(°F) (lb/ft )
3
(slugs/ft ) 3
(lb · sec/ft ) 2
(ft /sec)
2

32 62.42 1.940 3.746 1.931


40 62.43 1.940 3.229 1.664
50 62.41 1.940 2.735 1.410
60 62.37 1.938 2.359 1.217
70 62.30 1.936 2.050 1.059
80 62.22 1.934 1.799 0.930
90 62.11 1.931 1.595 0.826
100 62.00 1.927 1.424 0.739
110 61.86 1.923 1.284 0.667
120 61.71 1.918 1.168 0.609
130 61.55 1.913 1.069 0.558
140 61.38 1.908 0.981 0.514
150 61.20 1.902 0.905 0.476
160 61.00 1.896 0.838 0.442
170 60.80 1.890 0.780 0.413
180 60.58 1.883 0.726 0.385
190 60.36 1.876 0.678 0.362
200 60.12 1.868 0.637 0.341 ·
212 59.83 1.860 0.593 0.319

SI units

y Ρ μ χ 10 3
v x 10 6

(°C) (kN/m ) 3
(kg/m )3
(N«sec/m ) 2
(m /sec)
2

0 9.805 999.8 1.781 1.785


5 9.807 1000.0 1.518 1.519
10 9.804 999.7 1.307 1.306
15 9.798 999.1 1.139 1.139
20 9.789 998.2 1.002 1.003
25 9.777 997.0 0.890 0.893
30 9.764 995.7 0.798 0.800
40 9.730 992.2 0.653 0.658
50 9.689 988.0 0.547 0.553
60 9.642 983.2 0.466 0.474
70 9.589 977.8 0.404 0.413
80 9.530 971.8 0.354 0.364
90 9.466 965.3 0.315 0.326
100 9.399 958.4 0.282 0.294
580 General Appendix

Physical Constants

English units SI units

Acceleration of gravity 32.2 ft/sec 2


9.81 m/sec 2

Standard pressure at sea level 14.7 psia 101.32 k N / m 2

29.92 in. Hg 760 mm Hg


33.9 ft water 10.3 m water
Index

Absolute errors, 462 Bernoulli process, 6 Channel liners


Abstractions, 52 Bernoulli's equation, 105, 139 flexible, 122
Accuracy, 443 Binomial distribution, 6 temporary, 123-125
Aerobic microbes, 359 Blench regime method, 406, 409 Channel models, See also Channel
Aggregate, 204, 206 Braided channel, 393, 394 dynamic models, 415
stability, 216 Brune's model, 349 Chang's quantitative model, 417
Alluvial channel bedform, 397 FLUVIAL, 407, 418
Brush barrier, 382
Alternate depth, 106 Lane's geomorphic model, 415, 416
Bulk density, 212, 213
Anaerobic microbes, 359 qualitative predictors, 415, 416
Angle of repose, 128 Schumm's qualitative model, 416, 417
Annual series, 10 Channel morphology, See also
ANSWERS, 457 C factor, See Cover management factor Hydraulic geometry, 394
Antecedent moisture Calibration, 443 Channel sinuosity, 397, 398
erodibility factor effects, 257 Caliper logs, 451 Channel transitions, 137, 138
Antiseep collars, 354 Canaliform, 394 Channelized flow, 373
Apparent roughness, 400 Canals vegetative filter strips, 362
Aquifer, 422, 429 stable alluvial, 408, 409 Channels, See also regime channels
artesian, 433 Chang's rational method, 407-409 gravel, 401, 412
confined, 429 Channel regime, 405
perched, 429 Check dams, 375
armoring, 233
unconfined, 429 Chemical bridging, 208
depositing, 393
water table, 429 Chemical flocculation, 350
Armoring, 238 eroding, 393
Chen model, 335
channel, 233 graded, 391, 393 Chezy equation, 108
Average soil loss, 249 regime, 391 Chicago hyetograph, 49-51
stable, 391, 393 Churchill's model, 350
Channel analysis Chute, 167
Backwater curves, See Flow profiles fluvial, 391 Circular conduit, 110-112
Baffles, 327 Channel bends, 133 Circulation patterns, See Reservoirs
Bank storage, 53 Channel classification, 393 Classification, 223
Base time, 68, 75, 81, 82 description, 394 particle size, 213
Baseflow, 75 geomorphic, 393 Closed form equation, 244
BASIN model, 343 planform, 393 Coefficient of variation, 9
Bayesian estimation, 462 Combination sediment-stormwater, 354
Channel design
Bed factors, 407 Complex response, 393
fluvial, 391
Bedform Concentrated flow equations, 565
Channel dynamic models, 418, 419
alluvial channel, 397, 398 Concentrated flow erosion, 285
antidunes, 398, 399 HEC-6, 418
conveyance function, 286
chutes and polls, 398, 399 sediment routing, 418 DYRT model, 291
dunes, 398, 399 water routing, 418 equilibrium geometry, 285
fall diameter, 400 Channel erosion controls, 311 potential channel erosion, 292
flat, 398, 399 Channel flow erosion, 285 potential-actual relationship, 292
ripples, 398, 399 Channel forming discharge, 396 shear distribution, 285
transition/plane, 398, 399 Channel gradient, 396, 397 stage I, 287
Bedload, 223-225, 229, 409 Channel hydraulics, 392 stage II, 288

581
582 Index
Concentration mining, RUSLE, 275 Delivery ratio, See Sediment delivery
sediment, 227, 232 USLE, 266 ratio
Cones of depression, 431 mulch factors, 552 Deposition, 243
Confidence interval, 18-21 permanent pasture, 550 vegetative filter strips, 366
Conservation of mass rangeland, 550 DEPOSITS model, 336
groundwater, 427 residue decomposition, 273 Depth-duration-frequency, 40-42, 45
Conservation practice factor, 249, 280 root mass, 272 Detention basin, 201
RUSLE, 555 RUSLE crop parameters, 553 Detention storage, 38, 53, 198, 200
RUSLE subfactor tabulation, 282 RUSLE field operation parameters, Detention storage time, See abo
RUSLE subfactors, 280 554 Reservoirs, 187, 322, 352
subfactor, contour support, 280 step effect subfactor, forest, 559 plug flow model, 324
rangeland, 284 subfactor approach, agriculture, 270 DEWOPER, 187
strip cropping, 283 undisturbed woodlands, 552 Diameter
terracing, 284 USLE, 549 geometric mean, 216
USLE tabulation, 280, 281 USLE tabulation, 266 mean, 216
Constructed wetlands, See Wetlands CREAMS, 219-221, 241, 299 Diffusion, 227
Continuity equation, 38, 70, 93, 104, storm erosivity, 300 Diffusion wave, 70
183, 186, 187 transport capacity, 300 Discharge
reservoir sedimentation, 341 CREAMS parameter channel forming, 396, 397
sediment, 242, 315, 333-334, 419 critical tractive force, 561 Dispersion, 329
Continuous stirred reactor, See CSTRS Critical depth, 106, 107, 168 Diversions, 166
Continuous stirred tank reactor, See Critical flow, 106 Dobbin-Camp model, 334, 335
CSTRS Critical shear stress, 247 Double layer, 208, 329
Control Critical slope, 107 Downdrain, 166
inlet, 159, 161-163 Critical tractive force Drag
outlet, 159-161, 164 plasticity index factor, 561 partitioning, 401
Conversion factors, 575-580 tillage effect, 561, 564 Drawdown time, 355
Conveyance factor, 77 Critical velocity, 106 duBoys, 223
Conveyance function, 286 CSTRS, See also Reactor models, 315, Duration, 68, 72, 73, 75, 77, 83
open channel flow, 565 316 Dye dillution, 448
Cover management factor, 249, 267 variable flow rate, 341 Dye tracer tests
agriculture, USLE, 267 CSTRS model, 319 continuous injection, 315
agsubfactor, canopy cover, 270, 271 Culvert, 156, 158, 159, 160-164 slug injection, 315
prior land use, 273 capacity chart, 162, 163 Dynamic flows, 345
soil moisture, 273 classes, 156, 157, 158 Dynamic headwall, 239
surface cover, 270, 271 critical depth, 160 Dynamic wave, 70
surface roughness, 273 ditch relief, 166 DYRT model, 291
tabulation, 271, 272 Culvert capacity charts, 501-510
construction and disturbed land, 549 Cumulative deposition,
construction, RUSLE, 275 vegetative filter strips, 368 Ecotone, 358
USLE, 266 Curve number, 63-66, 89, 90, 92, Eddy diffusivity, 332
contour tillage subfactor—forest, 559 497-500 Effective particle diameters
disturbed forest, 558 Cutoff trenches, 354 vegetative filter strips, 368
disturbed lands, RUSLE, 275 Effluent concentration, 352
disturbed subfactor, density, 276 EI index, 250
30

PLU, 276 D , 127


5 0
Einstein, 223-225, 228, 229, 232
time, 276 Dam Einstein bedload function
disturbed woodland sites, 551 rockfill, 153 vegetative filter strips, 368, 370
distured forest and woodlands, 277 Dam hazard classification, 353 Einstein-Barbarossa method, 401, 402
forest subfactor, bare soil, 277 DAMBRK, 187 Electrical conductivity, 329
canopy, 278 Darcy's law, 356, 429, 430 Emergency spillway design, 353
consolidation, 277 Darcy-Weisbach Energy, 105
deposition storage, 278 friction factor, 152, 412 loss, 105
fine root, 277 Data, 442, 443 Energy dissipator, 167, 175, 178
step, 278 sources, 443, 444 Energy grade line, 106
tillage, 279 Dead storage, 317, 328, 342 Engelund Method, 401
USLE, 266 Deep seepage, 38 Environmental Protection Agency, 355
grazed woodlands, 550 Delivery ratio, 217, 293 EPA reservior model, 340
Index 583

EPA urban model Flocculation, 206, 209, 329, 350 Green-Ampt equation, 59, 60, 62
rainfall statistics, 569, 570 Flood frequency analysis, See Groundwater
EPA urban reservoir model, 345 Frequency analysis monitoring, 448
Ephemeral gullies, 239 Flood peak reduction, 198 mound, 437
Ephmeral gully erosion, See also Flood storage volume, 313 movement, 428
Concentrated flow erosion, 285 Flow occurrence, 428
Ephmeral gully model, 289 channel, 144 pollution, 423
Episodic events, 392 open channel, 164, 165, 166 Gullies
Equilibrium channel geometry, 285 Flow control emphemeral, 239
Equilibrium transport orifice, 146, 148, 155 headwall, 239
vegetative filter strips, 368 pipe, 144, 147
Equipotential lines, 429, 430 Flow control
Erodibility factor, 249 rock, 151 Hazard classification, 353
antecedent moisture effects, 257 spillway, 150 Head, 165
average annual, 255 weir, 144 elevation, 106
definition, 254 Flow depth pressure, 106
disturbed lands, 262 vegetative filter strips, 365 velocity, 106
rock fragment effects, 259 Flow profiles, 135-136 Head loss, 147, 148, 149, 152
seasonal variation, 256 direct step method, 136 Head loss
textural classification, 260, 261 Flow resistance, 400 bend, 147,148
Wischmeier nomograph, 256 Flow routing, See Routing entrance, 147-149, 161, 164
Erodible channel, 113, 114 Flow velocity, 75, 90 friction, 147, 169
Erosion Flume, 167, 447, 448 transition, 147
interrill, 239, 245 Fluvial channels, 391-393 velocity, 147
rill, 239, 246 Fluvial geomorphology, 391 Head discharge, 145, 167
Erosion control, 311 Fluvial system, 391, 392 Headwall
Erosion control dependent variables, 392 cutting, 248
philosophy, 238 independent variables, 392 gullies, 239
Erosion model time frame, 391 Headwater, 164
RUSLE, 249 FLUVIAL, 407, 413, 418 Hiding factor, 225
CREAMS, See also CREAMS, 299 FMO, 241 Historic data, 27, 34
USLE, 249 Form roughness, 400 Holtan equation, 58
WEPP, See also WEPP, 300 Formations Horton equation, 57, 61
Errors, 443 consolidated, 428 Hortonian flow, 56
ESP, See Exchangeable sodium unconsolidated, 428 Hybrid reactor, See reactor models
percentage Foster-Lane model, 285 Hydraulic conductivity, 56, 430
Evaporation, 93-95 Fractured systems, 433, 436, 437 sand filter, 356
Evapotranspiration, 38, 52, 93, 95-97, Frequency analysis, 5, 8-11, 16, 20, 23, Hydraulic geometry, 394
468 29, 31, 86 mean depth, 394
potential, 96, 97 Frequency factor, 16, 17, 19, 20 mean velocity, 394
Event-based, 38 Frequency histogram, 11 point, 394
Exceedance, 7 Friction factor, 108, 152 reach, 395
Exchangeable sodium percentage, 329 Darcy-Weisbach, 412 suspended sediment load, 394
Expert systems, 471, 472 Friction slope, 70 width-to-depth ratio, 395, 396
Froude number, 70, 106, 113, 134, 138, Hydraulic grade line, 106
139 Hydraulic jump, 138-140
Fall number Hydraulic radius, 108, 110, 224
vegetative filter strips, 367, 372 partitioning, 401
Federal highway administration, 401 Geographical information systems, 471 Hydraulic response, 314
Filter, 131, 132 Geomorphology, 391 Hydraulics
sand bed, 355 Geotextiie, 166 flow control, 144
Filter fence, 375 Graded channel, 391, 393 ground water, 427
Filtration Gradually varied flow, 133, 134 structures, 144
first flush, See First flush filtration Grain roughness, 400 Hydrograph, 67-73, 78, 83, 89, 91, 92,
Filtration basin, 357 Gravel bed channel, 401 182
First flush filtration, 354, 357 Gravel bed streams, 413 Hydrologic cycle, 38
First flush runoff, 354 Gravel channels, 412 Hydrologic soil group, 63-66, 89,
FLDWAV, 187 Gravitational potential, 56 485-495
584 Index
Hydrometer, 213 Meandering, 397 Outlet, 151, 155, 156, 166
Hyetograph, 44-52, 68, 69 Meandering channel, 393 culvert, 174
Meanders, 393 pipe, 176
Measurement, 443 Outlier, 28, 29
Imhoff cone, 211, 212 geophysical, 450 Overflow rate, 331
Index flood, 25, 26 seismic, 450 Overflow rate model, 329, 340
Inertial separation, 384 water level, 449 quiescent flow, 329
Infiltration, 38, 54-67 MEI, 364, 365 turbulent flow, 332
rock fragment effects, 259 Method of moments, 462 Overland flow, 70, 71, 75
Infiltration rate Microrelief, 239 Overland flow controls, 311
vegetative filter strips, 366 Minimum stream power, 405, 411 Overland flow roughness
Information sytems, 471 Model, 25, 27 CREAMS data, 562, 563
Inlet, 147, 150, 156, 166 classification, 456 Manning's n, 562, 563
Instruments, 443 continuous simulation, 467 Overtopping
Intensity-duration-frequency, 41, 42 distributed, 457 roadway, 156, 157
Interception, 38, 52 empirical, 457
Interrill erosion event, 463
hydrologic, 455, 456 Ρ factor, See conservation practice
splash, 245
lumped, 457 factor
Isochrone, 68
mathematical, 456 Parabolic cross section, 110
Isohyetal method, 39
parametric, 457 Parameter estimation, 460-463, 466
selection, 458 Parameters, 9
water yield, 468 Partial duration series, 10
Jacob, 434
Model evaluation, 470 Particle
Modeling aggregates, 204
approaches, 459 density, 222
Κ factor, See Erodibility factor Modified USLE, See MUSLE primary, 204
Karst systems, 433 Momentum, 107 shape, 206
Kinematic viscosity, 401 Momentum equation, 183, 187 size classification, 213
Knickpoint, 285 Monitoring, 442 Particle size distribution
KYERMO, 242 ground water, 448, 452 pond performance, 313
sediment, 452 Partitioning drag, 401
water quality, 451, 453 Einstein-Barbarossa method, 401,
Lag time, 68, 75-77, 83 Musgrave equation, 240 402, 403
Laminar sublayer, 401 MUSLE, 217, 297 Engelund method, 401, 402, 403, 404
Lane's geomorphic model, 415, 416 lumped parameter, 297 Partitioning hydraulic radius, 401
Lateral inflow, 70, 71 routing parameters, 298 Peak flow, 68,77-79, 81-88, 91, 93
Least squares, 462 size distribution, 298 Peaks over threshold, 10
Logarithmic velocity profile, 402, 403, Peizometric surface, 429, 434
567, 568 Percolation velocity, 356
Logarithmic velocity profile Nonerodible channel, 112 Permanent pool, See also Reservoirs,
reservoirs, 332 Nonuniform flow, 108, 133 342
Normal distribution, 475, 476 Permanent pool volume, 313, 352
Nuclear logging, 451 Permeability, 436
Macropores, 437 Phi index, 60
Manning's equation, 109, 363, 375, 376, Pipe, 147, 150, 155, 159, 164, 165
378, 400 Objective function, 462 Pipette, 214, 221
Manning's n, 75, 76, 109, 112, 115, 116, Old channel, 393 Planform
118-121, 172, 364, 563, 400 Open channel flow braided, 393
Maryland reservoir sedimentation Conveyance function, 565 meandering, 393
model, 350 models, 415 nonsinuous braided, 394
Matric potential, 56 Open channel flow, See also channel sinuous, braided, 394
Mature channel, 393 models canaliform, 394
Maximum likelihood, 462 Open channel models point bar, 394
Mean, 9, 14 FLUVIAL, 407, 418 straight, 393
Mean areal precipitation, 39 Orifice, 146, 148, 150, 156, 164 Plotting position, 12
Mean square velocities, 567 Outflow concentration, 352 Plug flow, 198
Index 585

Plug flow model, 320 annual, California, 531 Reservoir design procedure
Plug flow model, See Reactor models Hawaii, 533 peak outflow rate, 352
Plug flow reactor, See Reactor models Washington/Oregon, 532 sediment storage volume, 352
Plug flow-diffusion model, See also western US, 530 storage volume, 352
Reactor models, 319 average annual, 251 Reservoir sedimentation
PMP, 353 historical single storm, 545 quiescent flow, 329
Point bar, 394 monthly distribution, 252, 538 Reservoir sedimentation model
Pollution Pacific Northwest, 254 Churchill's model, 350
ground water, 423, 437-439 ponding adjustment, 252 Maryland model, 350
Pond sediment, 210 single storm, 252 plug flow, 336
Pond models, See Reservoir BASIN, 343
synthetic storm, 253
sedimentation models Brune's method, 349
Radius of influence, 433
Pond performance Chen, 335
Rainfall
reservoir shape effects, 326 CSTRS variable flow, 341
24-hour, 477-483
water chemistry effects, 326 DEPOSITS, 336, 337, 339
effective, 61, 62, 66-69, 72, 73, 90, 92
Ponds, See Reservoirs Dobbin-Camp, 335
SCS pattern, 46, 47
Pore velocity, 152 dynamic flows, 345
Porosity, 152, 428, 429 time distribution, 44-52
Rainfall energy factor, See also R factor, E P A 340
check dam, 379 EPA urban methodology, 345
Porous structures 250
Rainfall excess, See effective rainfall evaluation, 344
brush barrier, 382 long term trapping, 347
check dams, 375, 377 Rainfall/runoff factor, 249
Raingage, 444 Modified overflow rate, 340
filter fence, 375
Rating function, 186 SEDIMOT II, 336, 343
gabions, 381
size distribution, 339
infiltration impacts, 381 Rational method, 83, 84
Tapp model, 340
mechanical filtration, 381 Rational regime relationships
turbulent flow, 332
rock fill dams, 381 natural channels, 410, 412
Vetter, 335
straw bale, 383, 375, 382 Reactor models, 314, 318, 342
Reservoir shape, 324
Potential energy barrier, 326 CSTRS, 315, 318, 319, 334, 341
Reservoir survey method, 297
Potential evapotranspiration, 468 Reactor models
Reservoir type
Precipitation, 39 dead storage, 317, 318
dewatered, 325
mean annual, 94, 95 hybrid reactors, 317, 318
permanent pool, 325
monitoring, 444 parameter estimation, 318
Precision, 443 Reservoir volume
plug flow, 317, 318, 320, 334, 336-338
Primary particles, 204 flood storage, 313
plug flow-diffusion, 318, 319, 321
Probability, 6, 7, 12, 14, 28, 29 permanent pool, 313
short circuiting, 317, 318
paper, 12, 21 sediment storage, 313
Recharge, 436, 437 Reservoirs
plot, 10, 11, 12 enhancement, 436
Probability distribution, 13 baffles, 327
Recurrence interval, 6 circulation patterns, 322
extreme value I, 16, 17, 21, 31, 41
Regime, 114 dead storage, 328
log Pearson III, 16, 17, 21, 31
Regime channels, See also canals, 391, detention storage time, 322, 323
lognormal, 16, 17, 21, 31
405, 409 hydraulic response, 314
normal, 13-17
bed factors, 407 momentum factor, 328
Probable maximum precipitation, 43,
Blench method, 406, 409 overflow rate model, 329
353
Provinces Chang's rational method, 407, 408, permanent pool, 323
ground water, 423 409 reactor models, 314
Puis method, 191, 192 gravel bed, 413, 414 sediment scour, 329
natural channels, 410, 412 Residence time, 438
side factors, 407 Residence time distribution, 315
Qualitative channel models, 415 Simons and Albertson's method, 406, Resistance to flow
409 gravel channels, 412
Regional analysis, 26, 32 Resistivity, 450, 451
R factor, See also Rainfall energy factor, Relative frequency, 14 Retardance class, 115-117, 364
250 Relative submergence, 413 Return period, 5-11, 24
10 year, California, 536 Relief length ratio, 294 Revised USLE, 241
eastern US, 534 Removal fraction, 346 Reynold's number, 152, 153, 205
Washington/Oregon, 537 Removal ratio, 346 vegetative filter strips, 365, 366, 370
western US, 535 Reservoir, 147, 172, 211 Reynold's shear stress, 567
586 Index
Richards equation, 56 RUSLE R factor, See Rainfall energy Sediment yield, 293, 294
Rigid channels, 392, 393 factor MUSLE, 297
Rill density, 246 RUSLE S factor, See Slope steepness reservoir survey method, 297
Rill erosion, See also Erosion, rill factor time distribution, 299
models, 248 Sedimentation, 204, 238
Rill geometry, 247 Sedimentation basin, 354, 357
Rill growth and development, 247 S factor, See Slope steepness factor SEDIMOT II, 241, 336, 343
Rill incision, 247 S-curve, 73, 74 Seepage, 93, 354
Rill-interrill model, 300 Safety factor, 126 Sequent depth, 108, 139
Rill networks, 246 Sand bed filter, 355 Settleable solids, 352
Riparian vegetation, 359 Settling, 204
Santa Barbara hydrograph, 69
Riprap, 126-132, 166, 167, 175-178 compression, 204, 210
SAR, See Sodium adsorption ratio
Riprap lined wiers, 383 discrete, 204, 205
Schoklitsch, 223
Riprap porous structures, 383 flocculent, 204, 207
Schumm's Qualitative model, 416, 417
Risers hindered, 204, 210
Scour hole, 174, 175
multistage, 155, 156 tube, 209
Sedigraph, 299
single-stage, 155 zone, 204, 210
Sediment Settling velocity, 355
Risk, 6, 7
closed form equation, 244 Shear, 108
Rock, 151, 175-178
Rockfill, 151-155 deposition, 243 Shear distribution, 285
Roughness discharge, 223 Shear intensity, 225, 368, 370
fixed bed, 400 properties, 204 Shear stress
fluvial bed, 401 transport, 204, 222 critical, 247
form, 400 Sediment basin baffles, 327 Sheet erosion, See Erosion, interrill
grain, 400 Sediment basins, 311 Shield's parameter, 329
partitioning, 401 Sediment control, 311 Shield's diagram, 222, 223
Routing, 71 off-site, 311 Short circuiting, 317
channel, 184 on-site, 311 Side factors, 407
convex, 185, 189, 192 philosophy, 238 Sieve, 218, 219
flow, 182 systems approach, 385 Simon and Albertson's method, 406,409
graphical, 191 Sediment control structures, 311 Simulator
hydraulic, 187 Sediment delivery ratio, 293 rainfall, 218, 221
interval, 191 area relationship, 293 Single step method, 379
kinematic, 70, 186, 190 channelization relationship, 294 Size
Muskingum, 184, 185 Forest Service model, 294, 295, 296 distribution, 211, 214, 215, 217, 218
Muskingum-Cunge, 185, 187 Sediment detention basins, 312 eroded, 218, 220
numerical, 193, 194, 197 Sediment discharge Size distribution
reservoir, 190, 196 vegetative filter strips, 368 reservoir sedimentation models, 331
storage, 183, 184, 187, 189 Skewness, 9, 18, 33
Sediment graph, 299
Williams, 218 Slope length factor, 249, 261
Sediment pond, 172
RTD, See Residence time distribution rill-interrill ratio, 262, 263
turbulence, 568
Runoff, 67 slope length exponent, 261, 263
Sediment scour, 329
monitoring, 445 watersheds, 266
Sediment size distribution, 339, 368, Slope steepness factor, 249, 260
volume, 67, 79, 91
369, 372 Slurry flow rates, 379
Runoff coefficient, 84, 85
reservoir sedimentation models, 331 Sodium adsorption ratio, 329
Runoff diversions, 311
Sediment storage volume, 313, 351 Soil
Runoff hydrograph, See hydrograph
Sediment structures erodibility, 219
RUSLE, 249
RUSLE, See also Revised USLE, 241 check dams, 375 erosion, 238
RUSLE C factor, See Cover filter fence, 375 matrix, 220
management factor straw bales, 375 Soil erosion equations
RUSLE erodibility factor, See Sediment trapping Musgrave, 240
Erodibility factor vegetative filter strips, 366 Soil water content, 56
RUSLE erodibility factor, See Sediment trapping efficiency Soil water potential, 56
Erodibility factor check dam, 379 Solids
RUSLE Κ factor, See Erodibility factor vegetative filter strips, 362, 363 settleable, 210, 211, 212
RUSLE L factor, See Slope length Sediment trapping, long term, 347 suspended, 209
factor Sediment traps, 383 Specific capacity, 434
Index 587

Specific energy, 106, 138 Γ-year event, 5, 6, 7, 11, 24 USLE C factor, See Cover management
Specific force, 107 Γ-year flood, See Γ-year event factor
Specific momentum, 107 Temperature logs, 451 USLE erodibility factor, See Erodibility
Specific retention, 428, 429 Test well, 433, 434 factor
Specific yield, 428, 429 Texture, 214, 215 USLE Κ factor, See Erodibility factor
Spillway, 150, 168 Theis, 434 USLE L factor, See Slope length factor
Theoretical detention time, See USLE R factor, See Rainfall energy
broad-crested, 167, 169
Detention storage time factor
drop inlet, 147
Thiessen method, 39 USLE S factor, See Slope steepness
emergency, 167, 172, 173
factor
principal, 173 Thresholds
trickle tube, 164 channel, 392
Spillway charts, 511-528 Time of concentration, 68, 75, 76, 90
Vegetated channels, 115, 117, 122, 359,
Spontaneous potential, 451 Time to peak, 68, 75, 77-79, 81, 82
361
Stable alluvial canals, 408, 409 Top width, 110 Vegetative filter strip design, 373
Stable channel, 391 TP 40, 40 water quality impacts, 375
Staff gage, 445 TR 20, 85 constructed, 359
Stage, 445 TR 55, 85-88, 93 deposition wedge advance, 571, 572,
Tractive force 573
Stage discharge, 150,151,154,155,159,
allowable, 113 fall number, 367, 372
165, 166, 173, 191, 195
limiting, 113, 114 natural, 359
Stage-discharge curves, 403
Transmissivity, 434 retardance class, 364
Stage storage, 191, 195
Transport riparian, 360
Standard deviation, 9, 14
bedload, 225 sediment size distribution, 368, 370,
Stanford Watershed Model, 457 371, 372
sediment, 204, 223
Stationary time series, 6, 10 sediment trapping, 366, 370
Transport rate function, 368
Statistics, 9 trapping efficiency, 360, 362, 363
Trapezoidal cross section, 110
Steady flow, 108 vegetative stiffness, 364
Trapping efficiency, 210, 217, 312, 331
Stiff diagram, 295 Velocity
check dam, 375, 377, 378
Stilling basin, 140 allowable, 113, 115, 117, 118
overflow rate model, 329
Stilling well, 167 average, 104
size distribution effects, 331
Stokes' equation, 205, 206, 222 Darcian, 430, 437
turbulence effects, 334
Storage characteristic curve, 192, 194, fall, 205
water chemistry effects, 326 flow, 232
195 Travel time, 75, 76
Storm water, 38 limiting, 113, 114, 172
Trend, 10 measurement, 446
Straight channel, 393 Triangular hydrograph, 199 profile, 105, 227
Straw bales, 375 Trickle tuble, 164, 165 settling, 205, 206, 227, 232
Stream Turbulence, 206 Velocity distribution, See Velocity
effluent, 429, 437 Turbulence models, 567 profile
influent, 429, 437 Turbulent diffusion coefficient, 567 Vetter equation, 335
Stream gaging, 445, 446 Turbulent diffusivity, 567 VFS, See Vegetative filter strips
Stream power, 229, 243, 405 reservoirs, 332, 334 Visualization technology, 472
alluvial channels, 405 Volume
Streamflow, 38 runoff, 68, 74
Stress Uncertainty, 442 von Karmon's constant, 332, 567
shear, 222 Uniform flow, 70, 108
Strickler formula, 400 Unit hydrograph, 71-83, 89
Subcritical flow, 106, 107 Washload, 223
dimensionless, 78-80
Submergence, 164 Water
double triangle, 80, 81
Supercritical flow, 106, 107 capillary, 429
synthetic, 75
Support practice factor, See ground, 422
triangular, 78 vadose, 429
conservation practice factor Unit stream power, 229 Water balance, 93
Surface storage, 53 Universal soil loss equation, 241 Water chemistry, 326
Suspended load, 223, 228, 229 Unsteady flow, 108, 182 dispersion, 329
Swales, 311 USLE, See aho Universal soil loss double layer theory, 329
Swirl concentrator, 384 equation, 241, 249 electrical conductivity, 329
Systems approach, 385 rainfall energy factor, 250 flocculation, 329
588 Index

Water quality deposition parameter, 302 Yalin, 229, 232


sediment basin impact, 358 downslope variability, 303 Yang, 229, 232
Water surface profiles, See Flow profiles equilibrium width, 301 Youthful channel, 393
Water yield model, 468 interrill detachment, 302
Weir, 144-146, 150, 155, 164, 447 rill detachment, 302
broad-crested, 145, 156 sediment yield, 303
rectangular, 145, 146 transport capacity, 302
riprap lined, 383 Wetlands, 358 Zeros, 28
sharp-crested, 144, 145, 156 constructed, 358, 359 Zone
trapezoidal, 145 Wetlands saturated, 422
triangular, 145, 146, 447 native, 358 unsaturated, 422
Well interference, 433 Wetted perimeter, 108, 110 vadose, 422
WEPP, 219, 242, 300
basic equations, 301

Anda mungkin juga menyukai