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. ©
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
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
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
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
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.
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
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
f(p ,n)
T = l - ( l - l / T ) " , (2.3)
FREQUENCY DETERMINATIONS
(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
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
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
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
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
/ P (x)dx=l.
x
= Px{b)-P {a)x
(2.13)
J
— 00
fb
= / Ρχ( ) dx.
χ
(Χ- μ)
χ /·*,· + Δ * / 2
Ζ = (2.15) = / p (x)dx,
x (2.18)
J
x-kx/2
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)
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
0.999 0.947
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
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
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
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
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
G = (2.27)
w
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
Sample
size
η 5 10 15 20 25 50 75 100 1000
producing characteristics of the watershed should be are available for estimating confidence intervals about
Frequency Determinations 19
Skew
coefficient
c
s 1.0101 2 5 10 25 50 100 200
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
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
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
Τ b h
V Οι/
/ 0 = exp(i/ ).
U T
Return Period ( y r s )
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
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
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
Q = aA ,
b
(2.34)
as 1.0 results in an estimate for a of
T
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
i/E(*i-*) E(r,-y) 2 2
V /=i i-l
Γ ν V S \r( Χ X )
Υ = ' + y, (2.36)
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
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
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
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
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
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
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
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
Example Problem 2.2. Frequency analysis Plot Q versus l/T. Since lognormal probability paper is
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
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
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
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
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
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
Solution: 20 I
First develop a plot of Q /Q T X
versus T. Use Qi
for Q v
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
fore, 3000
β 5 0 = X Q = 5 X 1500 = 7500.
2
1000
The estimated 50-year flow is 7500 cfs.
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
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
^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
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
Σ( Y - Y ) 2
= X(ln Q - 6.594) = 2.145 2
8,340 23,100 55,000 1,200
a p
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)
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
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
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
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
0.5 2 3 6 12 24
The data tabulated below were obtained from TP 40 for F (years) 1
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
- 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
10 1
10° 10 1
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
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
Table 3.3 Synthetic 3-hr, 25-Year Rainfall, Stillwater, Oklahoma, DDF Method
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-
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
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
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
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
E x a m p l e P r o b l e m 3 . 6 . Time distribution
of rainfall—Chicago method
Interception
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
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
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
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
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
Table 3.9 Infiltration Constants for Bluegrass Turf Table 3.10 Infiltration Calculations Based on Horton
(Terstriep and Stall, 1974) and Holtan Models
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
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
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
Θ
A Theory
Β Horton ~~7K >
C Holtan
WETTING FRONT
^^^^^^^^
~H®iF~
-ΔΘ-
- θ β -
/= 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
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
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
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
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
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
F r o m Table 3.3.
a
/ = 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
AR AFa
F
p S AR C
(min) (cm) a
(cm/hr)* (cm) c
(cm)* (cm) e
(cm/ (cm)*
a
From Table 3.3 converted to centimeters.
^Potential infiltration rate from Green-Ampt equation.
c
Potential infiltration volume / Δ / . ρ
^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.
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
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)
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
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
- EFFECTIVE RAINFALL OR
RAINFALL E X C E S S HYETOORAPH
• AREA · V
A Conceptual Model
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
'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
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,
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
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
dA dv dA
~dt * A
T x + V
~dx~ = 9 ( χ , ί )
· <· )
3 36
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 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
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
Time
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
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
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.
,
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
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
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
Figure 3.34 Dimensionless unit hydrograph and mass curve with triangular
unit hydrograph superimposed.
Runoff Estimation 79
—e»-'/'» (3.59)
mated by
1.92
K = 6.5 ρρ (3.61)
A /inches . x
1.0 r -
Curve parameter is Κ
t/t
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
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
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
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
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
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
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
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
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
(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
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
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
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
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
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
/ = 0.25. 5 can be computed from Eq. (3.22) using the 0.35 2.35477 -0.49735 -0.11985
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
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
.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
diversion outlet. The overland flow velocity is estimated from computed from I = 0.25, where 5 is given by
a
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
Premining Postmining
CN = 62.5 CN = 66.6
0.81 1.03
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
5 X 1215
'<•- i90oyo^ = 0 2 5 h r
- V = = 1.00 in.
60.5Λ 60.5 X 100
Use t = 15 min.
L
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
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
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)
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
although finite, are essentially zero for times prior to 11.33 D = 5 min f = 15 min
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
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
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
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
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
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
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
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
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
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.
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
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
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
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
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
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-
{
have
d = 1.23 ft.
c
2 2
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
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
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.
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
bd + Zd 2
b + 2d Vz+ 2
b 1
+
b + Zd
2dVZ + 1 2
2
t - b • 2dZ
Τ - b • 2DZ
2dVz *1
2 2
or
—• approx.
0.67d
Ο \"2
approx.
td
2
2
A = — ( 0 ~ sin 0) (4.24)
(4.25)
CASE m $<Y<D
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
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
0.024
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
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
tan
r-y
^3 = 100 + 5 = 105
Solution:
(320/84) 320 2/3
(400/57) 400
2/3
η
(500/105) 500 2/3
0Ότ3
1.49 //33 \ 2 / 3
U
1/2
((0.04) = 72.4 cfs.
0.024 = 9010 cfs.
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
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
A =bd + zd 2
= 6 X 5 + 3 X 5 = 105 ft
2 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
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
bd + zd 2
6.22(0.69) + 3(0.69) z
-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
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
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
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
''Not recommended.
118 4. Open Channel Hydraulics
Channel vegetation
Soil Bare
texture channel Retardance Poor Fair Good
Clay 2.5 Β 3 5 6
2.5 C 3 4.5 5.5
2.5 D 3 4 5
.4
-.1 .2 .3 .4 .5 .6 .8 1 2 3 4 5 6 8 10 20
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
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
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
/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
(4.28)
= 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
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
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
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
a
Adapted from McWhorter et al. (1968).
Design of Open Channels 125
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
What has been shown thus far is that a channel having a R = = 0.658
1.5r + 4d
2 2
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
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
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
η = 0.0456(D S)
0.159
(4.34) ν = 12MD% . 51
50
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.
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
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
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
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
± = 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
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
Depth to Tractive
Angle of convey force Stability Safety
Manning's repose flow τ factor factor
(ft) η (°) (ft) (lb/ft ) 2
(SF )
(%) b
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)
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
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
oi )
b
(SF ) b
(lb/ft )
2
(η') (SF)
''Use a riprap with a D 5 0 of 2.2 ft for both channel sides and bottom.
= 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
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
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
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
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
D (base)
50
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
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).
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
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
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
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
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 +
AM ¥
Dam
method.
Figure 4.25 Sketch for Example Problem 4.21.
Channel Transitions 137
y v a
V2/2g E b
dx J
(ft) (fps) (ft) (ft) <fty (ft)
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
(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
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
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
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
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
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
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
(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
144
Hydraulics of Flow Control Devices 145
SLUICE GATE
Q- KLy (2gH)
2
0 , 5
Mutt bt V«otikit«d
ΐΤΓ~1 —
1.5 W>2M
Q - 3.087 L H
L - Width
ORIFICE
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
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
IT
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
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
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
a(2gH') l/2
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
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
K = 5087 c n /D 2 4/3
'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).
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
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
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*
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
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
and
Vd 1/2
(5.11)
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
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
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
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
m / s e c . For a width of 2 m,
2
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
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
(9.8)(0.02)(0.4 ) dh 2
dh In this case, it would b e necessary to calculate h for 2
. 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
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
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
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
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
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
(3) O U T L E T U N S U B M E R G E D
Η > H*
Y <d t
(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.
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
.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).
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_
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
Determine Outflow Velocity c. Connect the K point on the length scale to the
e
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 +D c 4.10 + 5
K = — — - 4 5 5 ft.
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
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
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.
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
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
Discharges (cfs)
g = 4.957(// + 2 . 1 0 ) .
fl 1/2
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
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
(2) define t h e required spillway dimension for a given for a rectangular cross section. T h u s Q is uniquely r
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
(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
g 2
Vc
2
(b+zy )y c<z CtZ
2g(A )
Ct2
2
2g 2(b + 2zy , ) c z ' (
' > cc r
% Error = 2 . 2 7 [ G ] = 2.27J0.15]L 14 9
- 0.13%.
Q_ l.5b +zHe c , ζ
t
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
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
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
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
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
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.
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
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
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
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
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 volume = 0 . 2 9 ( 1 . 5 6 ) ( 6 0 / ( 1 . 5 6 ) ) ( 2 0 ) °
30 25 ZU
= 1319 ft . 3
= 5.3 ft.
TW/D
Figure 5.20 Dimensionless rating curves for outlets of circular culverts on horizontal and mild slopes (after Federal
Highway Administration, 1975).
(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
TOR QT NATURAL
8ERM IS REQUIRED
TO SUPPORT WPRAP
" 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.
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
TAILWATER FLOW
JZ
SLOPE=0
SECTION
Figure 5.22 Rock riprap apron (adapted from Environmental Protection Agency, 1976).
DISCHARGE. CFS
Figure 5.24 Design of outlet protection—minimum 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
V dv 10 f t / s e c 1 ft/sec
= 6.2 Χ 1 0 " 4
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
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
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
Q = kS m
(6.18)
±
Q = bA m
(6.19)
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
dQ dA A 4
— + — = 4 (6.21) XQ + γ = a + β, (6.26)
dx dt
4
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
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
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
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
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)
.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
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
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
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
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
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
300 300
υ
200 h 200
ο
100 h 100 h
400 400
300 h 300 h
0)
υ
200 \- 200 h
ο
100 h 100 h
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.
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
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)
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
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
1. A s s u m e 0 = O 2 v
curve.
6. R e p e a t steps 2 - 5 until 0 remains unchanged. 2
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
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
(3)
Ι + I
γ 2 300 + 270 ft 3
(6) Δ5 = - 2 Τ ^ Δ / - - ^ Τ ^ Δ /
(17) 0 = 7(3.218) ' 0 498
= 12.528.
/ 12.08 + 12.08 \ 10 2
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
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
^ 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©
-Η
no
\5
CO
r»;
Q
TF
«-Η
νο
ο ο ο CO 00 -Η
OO OO 00 OO
73
CO CO CO Ο «Ο
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
-«
-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
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
(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
Τ = (6.34)
x
m Σ 9,
where t and q represent hydrograph time a n d flow
t {
^ - ^ - 7 ^ - 84 - 48 - 36 min.
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),
("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.
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.
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
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
TIME / \
CYLINDER
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
ird2
pVs 2
πά 3
(p -p)g,
s (7.1)
C D = 24//? , e (7.2)
.005 .01
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
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
^
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
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
H lF = -^K aR n n [u ] ,s
3
F x F
,2
a (7.7c)
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%.
settling.
Compression settling, like hindered settling, is highly where B , B c m9 and B are coefficients given in Table
s
Values of coefficients
inEq. (7.10) and (7.12)
a
W values from Lara and Pemberton (1963). Β values from Miller (1953).
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)
geometry
7 7
* ο θ
(7.15)
Hence,
tan (g/2)y?2
F = (7.16)
rD 2
ν 3
VT V
3 3 5
ρ — ___ — s
'
= 7 SI
^· (7-17)
W
in t h e size range d is
cone a r e shown in Fig. 7.9. Since t h e displacement of t
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
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
V = D/T, k is defined
0
3600 sec3 3
0.0024-0.0034 0.0030 0.04 2.53 x Ι Ο 5
1.39 x 10* 5
(15.57/12) J
ft 3
= 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
Then
W 0
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
.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
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
Large sizes
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
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.
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)
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
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
UJ
the delivery ratio for subwatershed i. A n example
m e t h o d for calculating D is the Modified Universal u
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
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
or
D2i = -BT /D^
e ti} 9 ( 7 3 0 )
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
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 F c l = 0.26O 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
F are the fractions of sediment in clay, silt, sand, ^sa - " a . ) ' - 0.04(1 - 0.36) = 0 5
l g
(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
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
"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
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
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.
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
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
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
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
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)
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.
gd (SG - 1)
3
9.81 m / m s e c X (0.013 m) (2.65 - 1)
2 3
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
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
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)
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)
100
t
Φ.
CONCENTRATION
Figure 7.24 Einstein's bedload equation (after Einstein, 1950).
and C is sediment concentration. Based on an analogy at a distance a above the channel bed, D is the depth
E
(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
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
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
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
—• —· 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
(Ν
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
.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 ) ©
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
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.
Particles (%) A Β
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
R b
us V CT 1
s vf β' it
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
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
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
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
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
Compare
Total Total
Detached Transport Capacity
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)
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
»
fro
st.c , o n
(8.3)
D r c = CX, (8.4)
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
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
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
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
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
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
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.
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
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
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)
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.
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.
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)*
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
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
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
-—.70
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.
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
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
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
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>
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
% 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
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.
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 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
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
Table 8.5 Experimental Κ Value Estimates for Disturbed Lands (English Units)
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
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
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
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
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.
m = = 0 73 / 210 \ ° 7 3
Table 8.7 Adjustment Factors to LS to Estimate Soil Loss on Irregular Slopes (after McCool
etal, 1993)°
,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
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.
fl
Equation (8.40) χ Eq. (8.43) with λ = 2 1 0 ft.
266 8. Erosion and Sediment Yield
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
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.
Σ= 100 50.08
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.
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
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
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.) (%)
mulch effects. T h e factor is given by total root zone. Values for n and a are given in Table i t
R G > 0.0 (8.55) is the sum of root mass and buried residue, or
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
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
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
On a mass basis, the buried residue in this zone is thus roughness of 0.4 in.,
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
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.
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
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
= (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
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
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
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
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
Table 8.13 Selected USLE Ρ Factors (after Wischmeier and Smith, 1978)
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)
^ 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
10471rt|1.5
(8.66) A n alternative to rotating strips is to include perma
1.667„
(EI ) S
lo
on
Ii \
80
0.8 1
OB
OB
CP
d c
•
60 X W
a Storm Severity ^
ο Adjustment Value ( P ) s
3α 20 i
•
οο ι
ι i
I 1
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
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.
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
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
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.
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
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.
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
(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
=
(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
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
= 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)
(8.81) (8.89)
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
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
(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 )
Table 8.15 Comparison between Predictions from the Foster and Lane Concentrated Flow Model and the
Ephemeral Gully Model.
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
= 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
= 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
/ nQ \ 3 / 8
yS (9803
(9803)(0.02) From earlier calculations, r = 29.2 Pa, which does not a
c f
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)].
/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{
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
t = 1—25. = — = 2.376 X K T * 5
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
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
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
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
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
actual erosion to determine the time to reach the nonerodi Using the value calculated above for (dW/dt) in act
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
cο
3
ο
co 1.00
*4
Ο
3
cd
< 0.10
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).
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
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
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
(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
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
(8.107)
MUSLE Routing Procedures i= l
(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
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
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
( 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
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.
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
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
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
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
(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€
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
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
τ * = [aXl +bX*Y . /3
(8.136) sity ( m / s e c ) , and C is the canopy factor, C is the c gc
mt + b
(8.153)
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
Property
3 Length 60 ft
Value
Steepness 5%
(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
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)
(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
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.
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
can b e integrated t o
C0 C1 Q
In —t (9.5)
V
or
Cl
Q
Co In 1 - (9.6)
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
(9.3) (9.10)
? <c„-c,)-^i
316 9. Sediment Control Structures
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
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
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
Co ι5 CoZ Cq 3
Pool = ΛΛ
// Outlet
= 1 - exp
-[t-(l- P l )t ]d
?
I
1
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
Table 9.3 Equations for Reactor Models with Dead Storage and Short Circuiting*
iti C ' - D ! V fi tj
Co
&=1, t < h u ( 9 2 5
>
Co
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
Time a
E(t) =
(min) (ppm) W = C,/C 0
dFldt l-C./Co
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
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
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
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
CtE(t)dt Ct(dF/dt)dt
h = Λ)
= CtdF. = EIi-WlAi'-Ai'EIi-Wl.
*g -OO in
/ E(t)dt
For this case
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
Σ = 5.51 F(t/t ) = 0.5. Using the data set, t'^ can be found by
d
interpolation, or
£(',') =
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
were made with a spreadsheet with F(t\) = C / C and not from rounded
T 0
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
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
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.
r [(V -
d p V )r + (V-V + V )(r - 7 )]/¥
ds dp p ds mo mi
(9.33)
+ (2.4-0.6 + 0.15)(7 m o -7 m i ).
Hence
Β
Tmo ~T = 1 8 . 7 hr. mi
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
T . This value is actually the detention time of t h e last ples of reservoir shapes that have large dead storage
mi
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
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
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
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 +
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 +
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
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
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
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
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
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
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
1.0
.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
Particle
size Diameter FF Particle
range* v.b
Column 6/Σ Σ size for
SI
(mm) (mm) ΔΧ 2
c
V A X / 1 . 1 5 x 10-Λ
si
J
( l - V /1.15 x l O " ) ^ .
si
4
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
Vertical Flux In = ρ θ ν Δ χ - ρ ε
δ
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
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
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
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
= 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
SI
(mm) (mm) IsX? (ft/sec) nVg d m
Ff
Σ = 0.789 Σ = 0.794
particles.
c
V / 1 . 1 5 x 10" .
jf
4
1.23xl0 V .
rf 3
5/
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
80 L _ /_ 13^f__
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
determined.
0.60
C ^ . - C j F F K , , ) ,
each layer is
0.0
.0001 .001 .01 .10 Layer; C . (mg/liter)
c y
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
Required Diameter d · c
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}
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
Average C Q
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
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
CQRGO
(9.60)
Variable Flow Rate—Modified Overflow Rate Models CA A t
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
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 ) +
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
(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
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
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
D
(9.69) where 5 is the rate of scour due to turbulence. S is
C c
and
dc
U* l Ζ •VJC-.- (9.76)
U(z) = —log 30.27-* (9.70) z= 0
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
%
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
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 9.27 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
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>
RR =
T Q
LA R
(9.80)
RY
: =
[jStorage Volume (Empty)
Mean Runoff Volume
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
4 80 8.00
Parameter Value 5 100 50.00
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
¥ 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
ciency, L , is 1.0,
F
that T is given by the problem statement as 100 hr,
lA
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
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
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
fl
RR = 8.36 V . s
£ =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.
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
Particle Equivalent
settling particle
Time velocity Overflow rate a
diameter''
(min-sec) Depth(cm) (ft/hr) (gal/dayft )
2
(mm)
a
\ gal/dayft = 40.7 liters/daym .
2 2
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
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
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'*
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
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
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
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
Outflow Velocity = V p e r
per'
initially at height H as shown in Fig. 9.38, with n o
Q
p
or
T h e total volume of water, C , p e r unit a r e a that V t P c r
\ 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
= -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
A =
H -L (e " <-l).
Q t
K L
(9.94) tV f ST - (24 hr)(3.54 ft/hr)
time. A maximum constraint for H would also typi - 6050 ft or 0.14 acre. 2
6.0 ft
Q
ϋ Α^
ΨΈ 36,300 ft 3
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
D^A„
'FF^w 36,300 ft 3
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.
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.
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.
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
{
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
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
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
K S - 0.091 + 20.76 ( M E I ) 2
(9.99) 0.074 = — d V ^ — — — j (0.08)" 2
0.106
[/ * = 0.754 ( M E I )
c 2 (9.100) d/ 5 3
f(d ) { =
(2d f + 0.0525) 273
v2/3
of grass that is stiff, and M E I is the stiffness of the (1.5)(0.08) (0.0525) 1/2
(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
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
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
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
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
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
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
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
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)
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
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 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
/ , 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
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)
The incoming sediment load that is coarse material ( > 0.037 Using fl with FF in Fig. 9.43, the average diameter for
vd d
D ^ - l - / · / » - 1 -(0.341)(0.64)
K = 0.021 ft/sec (Fig. 7.3)
= 0.782. StC
0.782
K -2.81dl [Eq.(7.5)]
This distribution is shown in Fig. 9.43. From this distribution,
tf
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
1 + (0.00473)(1 - r ) cs
ft /sec,
2
— = 1306.5
/.o = [/ + / (i -/)](ι -Λ!) +/d (A! -fS) +/d7?-
d
c m
r
From Eq. (9.102), the fraction trapped by settling, T is Thus the filter traps 68.1% of the incoming sediment.
S
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
1. Selecting a vegetation type. The vegetation assumed in L equals L, the fall number becomes
T
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
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
given as FFj in Fig. 9.43. Using Fig. 9.43, the values in the
d
Particle Fraction,
(ft/sec)
0.80. From item 4 above,
class FF 4
(mm)
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
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
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
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
can b e given by
where E and E refer to total energy at the d a m and
d q
Solution:
Since V = cV
d n9
1. Calculate the normal depth from equation 9.133.
2Zy AX 2
n
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
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'
_ 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.
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
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
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
(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
(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)
Ax = (9.149)
t r a t e d in Example Problem 9.19.
380 9. Sediment Control Structures
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.
Size V
v
si
b
vV
s\'· 5.6 gpm/ft 2
0.3 gpm/ft 2
10 gpm/ft 2
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
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
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-
Filter Fabric
Overland Flow
Pipe ( . 6 1 m ) 100
Four Flow
Spoilers
Inlet
(.61x.61m)
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
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
391
392 10. Fluvial Geomorphology: Fluvial Channel Analysis and Design
PRODUCTION
ZONE 1
Status of Variable 0
Geology I I
Water discharge I I X
Channel sinuosity I D X
Channel roughness D D X
a
I; independent variable; D ; dependent variable; X; indeterminate.
Channel Classification 393
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
Δ Β
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.
o
£ 00.01
0)
' " ' T ^ - . ^ J
ο.
J
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
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
η = </ 1 / 6
/21.1, (10.10)
η =</ 1 / 6
/25.7. (10.11)
U/U* =6.74(ΛΛ0 1 / 6
, (10.12)
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 {
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
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
• 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
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
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
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
(r -r)d o
8 5
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
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)],
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
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.
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 \
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
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
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)
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)
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
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
- A — - -500
No -400 to
-300
0. 200
0.025. - 4
No
—___n_
- 3 •a
V
—2
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
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.
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
= 0.000132
F.-0.1. (32.2)(2.57)(0.33)((0.876)(22.2)/10- ) 5
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
(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
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
2000
3000
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
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
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
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
51 and
Slope > 0.004
-0.410
5 = 0.223Q (10.64)
0.001065<4
e
15
and
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
= 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
^ - = 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
d.
0.01
ο
to
T3
0.001
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
"Ό
Χ
C/D
0.001
0.0002
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
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.
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
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.
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
slope; 5 , sinuosity.
U
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
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
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
(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
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)
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
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).
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
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
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
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'=
to Λ to obtain
2
*K{h\ ~ fc?)
Q = (11.12)
ln( r / r , )
2 Ground turfooe
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
(11.18)
prior to determination of specific capacity. T h e well
Basic Concepts of Ground Water Hydraulics 435
(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
5 =
r = 7.8 X 10~ 5
14
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
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
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
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
Dye Dilution
(12.3)
C? — C h
Figure 12.2 Flumes used for flow measurements. ( A ) H-flume, t h e stream, C is the tracer concentration in the injec
x
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
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
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
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
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
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
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
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)
flows, low flows, and runoff volumes, whereas most estimated values of O . N o t e that f may b e trans
p 2
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
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
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
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
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.
400
13 14 15 16
Time (hours)
Figure 13.11 Impact of development C and D lumped.
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
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
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
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
485
486 Appendix 3B
497
498 Appendix 3C
Table 3.2a Runoff Curve Numbers for Urban Areas 0
Average percent
Cover type and hydrologic condition impervious area* A Β C D
*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*
Hydrologic
Cover type Treatment* condition 0
A Β C D
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
Hydrologic
Cover type condition A Β C D
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*
Hydrologic
Cover type condition* A c
Β C D
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).
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).
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
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
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
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
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
.004 .006 .006 01 .02 03 .04 .06 .08 0.1 0.2 0.3 0.4 0.5
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
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
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
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
(continues)
546 Appendix 8A
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
(continues)
R Factor Information 547
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
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
Condition C factor
"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
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)
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
Effective canopy 0
Forest litter*
(% of area) (% of area) C factor 0
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
"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)*
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.
A. Depletion rates
Depletion per
Crop 15-day period
B. Replenishment rates
Replenishment rate _
0.5 + 0.662 ( / » - ! )
per 15-day period
10£P£18
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
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
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+
Table 8D.2 Step effect on soil erosion (after Dissmeyer and Foster, 1984).
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
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)
Small grain
(20% to full maturity) Across slope Upslope and downslope
Table 8E.2 Concentrated Flow Values for Manning's η for Typical Soil
Covers' for CREAMS Equation (after Foster et al., 1980b)
1
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
Good 0.090
565
566 Appendix 8F
Item w m R* s(*. )
c
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
™ - V » - U A 1 - h \ ( 9 A 1 )
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
Figure 9B.1 Regions for E P A model rainfall statistics (after Driscoll et al., 1986).
569
570 Appendix 9B
Rainfall statistics
Q ι d r u
4si ~ 4sd
(9C.2) Given a value of q , o n e simultaneously solves Eqs.
sba
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
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
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
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
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
- 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
and
sb e t
Set
= Of,
5 = 5
e e t - S = 0.0867 - 0.08 = 0.0067.
c *f('f) -
575
576 General Appendix
C o m m o n Equivalencies
Length
Equivalent
Area
Equivalent
Unit Square inch Square foot Square meter Acre Hectare Square kilometer Square Mile.
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
Equivalent
Velocity
Equivalent
Mass
Equivalent
Unit Pound mass Kilogram Metric slug Slug Metric ton Long ton
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
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
'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
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
Physical Constants
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
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