Guidelines

For

Determining

Bulletin7#VB
of the
Hydrology

Subcommittee

Revised September
Editorial Corrections

1981
March

1982

INTERAGENCY ADVISORY COMMITTEE

ON WATER DATA
US. Department
of the interior
Geological Survey
Office of Water Data Coordination
Reston, Virginia 22092

FOREWORD
An accurate
an effective,
an acute

estimate

nationwide

need for

of the nation's
levels

flood

water

widespread

application

determining

flood-flow

In a pioneer

of flood

enterprise.

attempt

published

Bulletin

to promote

for

Federal

planning

and private

organizations.

Adoption

to develop

methodological

presented

tion.

flow

to flood-flow

and
for

Council

stations

was adopted

water

Flood

and related

land

by::State,

local

was based upon the clear

improvements

Flow

by the Council
resources.
government,

understanding

in the technique

No. 15 was published

for

Flood

accepted
with

Determining

methods

sufficient

was a synthesis

improvements

Determining

of Bulletin

No. 17, "Guidelines

the currently

in December 1967

would

when appropriate.

and update

The guide

ological

approach

recommended use of the technique

and adopted

data at gaging

acceptance,
technique

for

therein

involving

also

as Bulletin

Technique

presented

The Council

An extension

both a consistent

and accurate

a consistent

No. 15, "A Uniform

The technique

be continued

is

among various

development,

the U.S. Water Resources

Frequencies."

efforts

there

frequencies.

determination,

that

To obtain

consistent

to

because management

is shared

requires

of a uniform,

Further,

program.

resources

losses

is a key element

to such estimates

land

and private

frequency

use in all

approach

and related

estimate

damage potential

damage abatement

a consistent

of government

and accurate

of the flood

and a survey

for
detail

of studies
of existing

analyzing

Flow Frequency."
peak flow

to promote
undertaken
literature

in March 1976

uniform

It

frequency
applica-

to findmethodon peak flood

determinations.
The present

guide

is the second revision

i

of the original

publication

and improves

the methodologies.

techniques

in

the

previous

explanation

of other

to develop

a coherent

potentials.

techniques.
set

to submit

peak

Data Coordination
Federal
activities

agencies
involving
organizations

to assure

more uniformity,

that

all

Bulletin

and offers

result

of a continuing

this

is the

study

required

will

is

be fully

frequency

before

determinations

by the

are

to use these

and related
are

concerned

Office

effort
flood

who are

interested

encouraged
of Water

guidelines

resources.

in all
State,

also

in

must use for

planning

local

guidelines

and comparability

and citizens

a further

Subcommittee.

to use these

compatibility,

two goals

are

I-lydroloqv

land

encouraged

agencies

the
All

to the

the

defining

attained.

and proposals

requested

some of

accurately

consideration

water

and private

of

for

criticism
for

and expands

It

flood-flow

comments,

revises

of procedures

and consistency

improving

values

editions

Much additional

of accuracy
in

It

the

frequency

many vital

decisions.

This
T hat review

present

of these

and by examination
be

revision

is adopted

procedures
and testing

will

with

the knowledge

continue.

of new techniques,

publlshed.

ii

and understanding

When warranted

other

revisions

by experience
will

Ll

HYDROLOGY SUBCOMMITTEE
*
Robert
Robert
Walter

Department

Agency

Member

Soi 1 Conservation

E. Rallison
G. Delk
J. Rawls

Agriculture
II
II

Service

Forest Service
Science Education
Administration

Vernon K. Hagen
Roy G. Huffman

Corps of Engineers
II

Arty

Allen

NOAA, National
Service
II

Commerce

F. Flanders

John F. Wilier

Weather

II
Housing and Urba
Development

Truman Goins
Porter Ward
David F. Gudgel
Don Willen
Ewe11 H. Mohler,

Jr.

Sidney J, Spiegel
Pat Uiffy
Leo Fake
Victor Berte'
Irene L. Murphy

Bureau of Indian Affairs

Bureau of Mines
Fish and Wildlife
Service
National
Park Service
Heritage,
Conservation
and
Recreation
Service

D. c. woo
Philip
L. Thompson

Federal

Timothy

Horn

Steve

Parker

Patrick
Brian
William

Highway
II

Administrat

it

on

Transportation
II
Environmental
Protection
Agent
II

Stuart

Robert

Interior
II
II

Geological
Survey
Bureau of Reclamation
Office
of Surface Mining
Office of Water Research
and Technology

Federal Energy
Regulatory
Commission
II

Jefferson

Federal EmergentManagement Agent,

I,

Mrazik
S. Bivins

Nuclear Regulatory Commission

Edward F. Hawkins
...

111

.Y

HYDROLOGY
w,.
. SUBCOMMITTEE

Department

Agency

Member
*

- Con't

Tennessee
Authocity

Donald W. Newton
Larry

Valley

14. Richardson
Water Resources
Council

Ron Scullin
WORKGROUPON REVISION OF BULLETLN 17

Department

Agency

Member
Soil

Roy G. Huffman

Corps of Engineers

Army

John F. Miller*

NOAA, National
Service

Commerce

William
H. Kirby
Wilbert
0. Thomas,
Frederick
A. Bertle

Geological

Jr.

Conservation

II

Bureau of Reclamation

Interior

6,
11

Tennessee
Authority

* Chairman

as of September

Weather

Survey

Donald W. Newton

YMembership

Service

Agriculture

Roger Cronshey

1981

iV

Valley
*

The following

pages contain

"Guidelines

for

1, 4, 8-2,

and 13-1

The revised

material

The following

Determining

is included

from material
Flood

presented

in

Flow Frequency."

on the lines

pages of Bulletin

13-2 through

enclosed

by the +

17 have been deleted:

13-35

The following
Bulletin

revisions

pages contain

revisions

from the material

in either

17 or 17A,
i,

ii,

iii,

iv,

20, 26, 7-1,

6-2,

6-3,

7-9,

9-l

through

The following

Editorial

vii,

1, 3, 10, 11, 12, 13, 14, 15, 17, 18, 19,

l-3,

l-4,

2-3,

l-2,

6-1,

The revised

v, vi,
6-5,

6-6,

9-10,

material

6-7,
10-1,

2-8,

7-1,

7-2,

10-2,

10-3,

is included

page of Bulletin

corrections

2-7,

4-1,

7-3,

5-1,

7-4,

7-5,

12-2 through

on the lines

enclosed

5-2,

5-3,

7-6,

17B were incorporated

7-7,

7-8,

12-37 and 14-1

by the *

77 and 77A has been deleted

to Bulletin

5-4,

from 17B:

into

this

CONTENTS
Foreword. . . . . ..I.......................................
.,...............................
Hydrology
Subcommittee
Page Revisions
to Bulletin
17 and 17A ..................

46
I.
II.

...........................................

Summary ................................................
A.
Information
to be Evaluated ........................
B. Data Assumptions ...................................
of the Frequency
Curve ...............
C. Determination
...........................
D. Reliability
Applications
..........................................
E. Potpourri
F. Appendix ...........................................

Information
to be Evaluated ............................
A.
Systematic
Records .................................
B. Historic
Data ......................................
C. Comparisons
with Similar
Watersheds ................
.................
From Precipitation
D. Flood Estimates

III.

IV.

V.

Introduction

Data
A.
B.
C.
D.
E.

.......................................
Assumptions
Climatic
Trends ....................................
Randomness of Events ...............................
Watershed
Changes ..................................
Mixed Populations
..................................
Reliability
of Flow Estimates ......................

Determination
of Frequency
Curve .......................
...................................
A. Series
Selection
B. Statistical
Treatment ..............................
The Distribution
...............................
::
.......................
Fitting
the Distribution
3.
Estimating
Generalized
Skew ....................
.................
4. Weighting
the Skew Coefficient
Broken Record ..................................
5.
Incomplete
Record
7: Zero Flood Years .............................................................
a. Mixed Populations ..............................
.......................................
9. Outliers
10.
Historic
Flood Data ............................
to Frequency
Curve .....................
C. Refinements
1.
Comparisons
with Similar
Watersheds ............
.............
2.
Flood Estimates
From Precipitation

vi

Page
. . i.
111

VI.
*

VII.

pa%
22

.................................
Reliability
Application
A. Confidence
Limits ...................................
6.
Risk ................................................
................................
C. Expected
Probability
...............................................
Potpourri
A.
Non-conforming
Special
Situations
...................................
B. Plotting
Position
C. Future Studies ......................................

23
24
24
...................

Appendices
1.
References .................................................
2.
......................................
Glossary
and Notation
3. Table of K Values
.........................................
+) 4. Outlier
Test K Values
...................................
.........................
5. Conditional
Probability
Adjustment
6.
Historic
Data ..............................................
7. Two-Station
Comparison .....................................
8. Weighting
of Independent
Estimates .........................
Confidence
Limits ..........................................
1::
Risk .......................................................
.......................................
Expected
Probability
1::
Flow Diagram and Example Problems ..........................
13.
Computer Program ...........................................
14.
"Flood
Flow Frequency
Techniques"
Report Summary ...........

vii

25
25
26
27

7-1
8-l
1:~;
11-l
12-l
13-l
14-1

I.
In December
Flood

Flow

Water

Resources
distribution

III

distribution)

As pointed

with
in

flow

Frequency"

that

frequency

by the

Water

of Bulletin

guide

flood

frequency

It

was limited

and exceedance
flood

flows

from
studies

Bulletin

earlier

this

fidence

limits

revision

were

frequency
accurately
analysis

with

where

Flood

various

address

Center

curve.

Flow
in

based

Research

(summarized

of

peak

was selected

1976,

on

in Water

in Appendix
+

Frequency.

June 7977 as Bulletin

and expand

upon the

procedures

comparisons,

of the orlginal

for

weighting

skew study,

detect-

and computing

The Work Group that

the suitability

discharge

record

a generalized

two station

application.

peak

to

revised

from

accepted

procedures

to improve

17B provides

making

of

for

and reissued

results

uniform

prior

on Flood

effort

Flow
was an

currently

of

literature
at the

revised

The guide

a systematic

set

Work Group

the

prepared

con-

this

distribution

skew map.
are

There

which,

define
will

encountered

determinations.

alone

studies.

Determining

in terms

at Austin

a frequency

problems

can be adopted

Type
Type

covering

to promote

Texas

latest

outliers,

did not

the

Bulletin

or ,the generalized
Major

of

by the

about

Pearson

were needed

incorporating

potentials

purpose

17B is the

and treating

the

It provided a more complete

detail

in

University

skew value

of

(log-Pearson

Council.

The recommended

publications.

a station
ing

sufficient

used or described

The "Guidelines"

17A.

analysis

at locations

and on studies

the

frequency

for

No. 15.

probability

of the

flow

Resources

flood

for

data

studies

to defining

conducted

Resources
14)
=i%

with

is available.

those

flood

"Guidelines

17,

and update
methods

for

Committee

use of

Determining

for

determinations.

extension
technical

Hydrology

of the

further

Bulletin

flow

Technique

recommended

transformation

report,

was issued

for

by the

The report
log

'A Uniform

15,

as a base method

out
of

No.

was issued

Council,

In March 1976,

Bulletin

Frequencies,"

III

aspects

1967,

Introduction

when developing
is no procedure

when rigidly

the
not

flood

applied

potential

resolve

all

frequency

for

flood

flow

or set of procedures

to the available

of any given
flood

guides

watershed.
problems.

data,

that

will

Statistical
As discussed

in subsequent
are

sections

inherent

in any

based on properly
considering
sional

flood

in

frequency

is

procedures

of

and proper

analysis

to standarize

and uncertainty

decisions

must

be

of

results

interpretation
the

will

and promote

risk

User

Therefore,

hydrologic
analysis

possible

elements
analysis.

and uncertainty.

experienced
It

guide,

frequency

applied

risk

of a flood

of this

judgment

enhance

the

appropriate

many elements

of

a profes-

usefulness

application.

of

flood

frequency

analysis,

describes
each major element of the process of defining
..This
..-^ . guide
_
flood
potential
at a specific
location
in terms of peak discharge
exceedance
records

probability.
are

situations
dures

Use is confined

adequate

to warrant

may require

other

of this

guide

are

appropriate

study

recommended

procedures.

As a further
the

one analyst
rare

followed,

and accompanied

is working

limited,

at each location,

the

determination

Thus,

an estimate

may be outdated

flood

data

records

flood

porate

his
they

When making

study
should

a review

of

flow

guide

reason

frequency

length

(at

as the

basis

curves

least
for

the
where

10 years)
determination.

data

proceby

using

the

results,

when more than

same location.

of

record

This

probabilities
become available

potential

for

after

may change.
it

is made.

a fresh

a new assessment,

be acknowledged

describes

results

exceedance

a few years

of earlier

the

more latitude.
flood

Additional

assessment
the

estimates.

analyst

of
should

Where differences

and explained.
I I.

This

the

As more years

may be sufficient

potential.

in

appear,

alone

allows

where

Special

be supported

be coordinated
for

data.

and improving

when defining

guide

of

and

available

the

cases
must

consistency

particularly

are

the

deviations

on data

this

Flood

In those

studies

where

of

by a comparison

currently

holds

events,

that

where

analysis

approaches.

not

recommends

recommendation
for

statistical

means of achieving

Work Group

to stations

the

Summary

and procedures

systematic

stream

to warrant

statistical

The procedures
2

for
gaging

computing
records

analysis
do not

cover

flood
of

sufficient

are available
watersheds

incor-

where

flood

where

the

flows

are

appreciably

possibility

of

altered

unusual

events,

considered.

The guide

annual

peak discharge.

It

be used to treat

other

flood

could

also

volumes.

by reservoir

or

such as dam failures,

was specifically

developed

is recognized

for
that

hydrologic

however,

Such applications,

regulation
the
the

elements,

were

not

must

be

treatment

of

same techniques
such as flood

evaluated

and are

not

intended.
The guide

is

divided

into

six

broad

sections

which

are

summarized

below:
A.

Information

to

be Evaluated

The following

categories

records,

historic

data,

estimates

from

precipitation.

potential

is

briefly

of

flood

comparison

data

with

are

recognized:

similar

systematic

watersheds,

and flood

Mow each can be used to define

the

flood

described.

13. Data Assumptions

A brief
to those
data

developing

errors.

mixed

c.

discussion

Procedures
flood

the

The main

are

data

also

assumptions

frequency
of

the

is

recommended

of

flow

presented

to be aware

events,

as a reminder
of potential

watershed

estimates

are

changes,

briefly

discussed.

Curve

basic

thrust

is

curves

randomness

Frequency

provides

guide

for

determination

determination

to convert

of the
an annual

of

annual
to

the

flood

freseries.

partial-duration

series.
The Pearson

flood

data

the

the

%ered.

are

statistical

to incorporate

is

parameters
are
for
are

of the

information

gained

from

the

used to modify

described

transformation

the

refining

station

systematic
from

the
record

comparisons

precipitation.

of the

basic

distribution

of moments

distribution

of most flood
for

as the

The method
of

treatment

log

recommended

series.

analysis

estimates

with

III)

flood

statistical
proposed

distribution
Type

annual

Proce dures

and flood

III

relations

4Generalized
Methods

Type

(log-Pearson

defining

termine

from

flow

trends,

of

section

curve.

basic

, and reliability

Determination

quency

for

flood

Natural

populations

This

of

from

is

used to

station

de-

data.

skew coefficient.

record
basic

problems
curve

-95

encoun-

determined

and historic

flood

with

watersheds

similar

data
~

uo

Kellabl

Ilt;y

AppllCatlOnS

Procedures
provided

along

ability

for
with

computing

confidence

limits

to the

those

calculating

risk

and for

for

frequency

curve

are

making expected

prob-

adjustments.

E. Potpourri
This
guide,

section

including

positions,
F.

provides

information

a discussion

and suggested

of interest

of non-conforming

future

but not essential

special

to the

situations,

plotting

studies.

Appendix
The appendix

provides

a list

ymbols, tables
of K values,
B
of the recommended procedures
program

for

handling

made at the University

all

treating

how to obtain

Techniques")

guided

of
most

a computer
and a c

of data,
describing

selection

studies

of some of the pro-

Systematic
Annual

Specific

records

are obtained

periodic

observations

frequency

curve,

The four

general

frequency

applications

only

agencies

analysis

the

ana7yst

types

should

are discussed

con-

of data which

are described

either

information

is observed

and private

enterprises.

from a continuous

can

in the follow-

in subsequent

sections.

of a crest-stage

are available

in U.S.

gage.

files,

form is available

Geological

but additional
from other

systematically
of river

Crest-stage
base.

Survey
information
sources.

by many

Most annual

trace

on peaks above some preselected

and computer

unpublished

flow

peak discharge

and state

data

flow

Records

Federal

information

Information to be Evaluated

information.

in the flood

ing paragraphs.

Papers

for

and treatment

Flow Frequency

a flood

available

be included

these

about

and list

proposed,

When developing

A.

details

analysis

of Texas which

III,
sider

, information

("Flood

a glossary

computational

the statistical

summary of the report

cedures

the

of references,

peak

stages

records
A major

or from
may provide

portion

(USGS) Water Supply

in published

or

of

A statistical
determination
B.

analysis
of the flow

Historic

plain

for

which

occurred

either
This

the largest

there

before

or after

often

either

computer

files.

recorded

flood

of other

intensive

inquiry

frequency

information

Use of historic

With

data assures

Similar

between

computed

being

are useful

Studies

have been made and published

*Numbers

gross

and

files
which

from

or by
the flood

and documented
record

estimates

fit

curves

is relatively

community

experi-

region

and downstream,

frequency

[e.g.,

estimates

(l),

events

(2),
with

(3),
with

regional

(4)1*

which

generalized

information

be made, particularly

to promote

and for

determinations.

at a site

Comparisons
should

in a hydrologically

of unusual

flow

region.

and maximum flood

and those

identification

of flood

a homogeneous

upstream

prevent

for

frequency

immediate

that

investigated

the reasonableness
flood

be obtained

frequency

testing

stations

reports

Watersheds

region

in the

The USGS

determinations.

similar

stations

for

where the systematic

the frequency

data of the watershed

for

which

be obtained

from newspaper

near the site

of

during

published

can sometimes

should

particularly

comparing

period
are known.

in its

or extracted

information

possible,

Comparisons

data

is needed,

flood

Comparison

or historic,

and investigation

ence and improves

of systematic

an extended

information

agencies

floods

be used to make estimates

information

Additional

the files

help

the

the flood

about major

the period

defines

some historic

estimates

for

each station.

is information

also

floods,

Historic

period,

can often

includes

permit

for

where man has occupied

information
It

peak discharge.

C.

curve

particularly

an extended

collection,

short.

frequency

basis

Data

At many locations,

whenever

data is the primary

of these

at
at gaging

consistency

and

errors.

in parentheses

refer

to numbered references

in Appendix

1.

D.

Flood Estimates

From Precipitation

Flood discharges
snowmelt)

estimated

can be a useful
however,

valid

model for

Unless

such models

effort

may be required
Whether

Necessary

to prepare

This

of events,

watershed

to the watershed,

considerable

depend upon the availabilit,

of the existing

records,

and the exceedar

a statistical

section

discusses

frequency

are that

and representative

time

of the adequacy

a necessary

changes,

analysis

first

the effect

step

sample of

and applicability
in flood

of climatic

mixed populations,

the array

frequency

trends,

randomnes

and reliability

of flow

analysis.

Trends
is much speculation

indicates

that

of years.

assume flood

flows

time

major

about

climatic

changes

occur

In hydrologic

analysis

are not affected

invariance

changes.
in time
it

Available

scales

involving

is conventional
to
._-. .
trends or cycles.

by climatic

was assumed when -developing

this

of annual

rates

guide.

Randomness of Events
In general,

considered
tical

data and a

to discharge.

are made will

Assessment

is therefore

on flood

There

B.

for

events.

analysis,

Climatic

climatic

such estimates.

is a reliable

records

thousands

measurements.

Data Assumptions

assumptions

of flood

evidence

adequate

and/or

is most important,

information

Climatic

(rainfall

precipitation

calibrated

the adequacy

random homogeneous

A.

at least

or not such studies

which

estimates

streamflow

converting

IV.

of flood

to direct

require

are already

of the information,
probability

data

adjunct

Such estimates,
watershed

from climatic

tests

deviation
estimation

an array

maximum peak flow

a sample of random and independent

of the serial
from this
of future

The nonrandomness

correlation

assumption,
flood

coefficients

the annual

activity

if

of the peak series

events,

other

will,

may be

Even when statisindicate

a significant

peak data may define

assumptions
however,

an unbiase

are attained.

increase

the degree

of uncertainty
data will

(5),

Watershed
It

increasingly

flow

difficult

conditions

the construction

from a lesser

sample

to find

by man's

include

watersheds

activity.

Man's activities

urbanization,

of reservoirs,

in which

channelization,

diversions,

and alteration

of

conditions.
Watershed

assure

that

record.
All

attainable

has not been altered

can change

cover

based upon nonrandom

Changes

regime

levees,

a relation

(6).

is becoming

the flow
which

is,

have a degree of reliability

of random data
C.

that

in the relation;

history

no major

documents

watershed

which

unlisted,

years.

regime

records
changes

accompany

from year

records

record

examined

during

often

however,

reservoirs

might

of urbaniza-

over

a period

changes may not significantly

but the cumulative

effect

of

such changes.

be the effects

small

to

the period

list

homogeneity,

might

of numerous
to year

be carefully

have occurred

instance,

Such incremental

should

flood

affect

for

and the construction

several
flow

watershed

changes which

not be listed;
tion

and flood

of

alter

can after

the

several

years.
Special

effort

not homogeneous.
watershed

should
Only

conditions

be made to identify

records
should

which

those

represent

be used for

records

relatively

frequency

which

are

constant

analysis.

13. Mixed Populations

At some locations
For example,

flooding

flooding

is created

in some watersheds

or by combinations

of both

not be homogeneous

and may require

E.

Reliability
Errors

values.
flood

exist

flows.

in flood

is created

and rainstorms.
special

types

of events.

by snowmelt,

rainstorms,

Such a record

may

treatment.

of Flow Estimates

Errors

introduced

snowmelt

by different

in streamflow
in flow

estimates

Measurement
is usually

flows.

records,

small

The effects

errors

as in all

are generally
are usually

in comparison
of measurement

other

greatest

measured
during

maximum

random3 and the variance

to the year-to-year
errors,

therefore,

variance
may

normally

be neglected

estimates

of historic

uncertainty

with

will

discussion

found

(7).

A. Series

flood

A partial-duration

peaks equal

to or greater

series

events

events.

flood

agency

measurement

is

Curve

events
A study

per year

use.

the annual

convert

from the annual

as summarized

encountered
to ensure

partial-duration
but less

desirable

duration

series.
specifically

series

that

an empirical
will

developed

to assure
one event

all

events

for

for

separating

are recommended for

if

a consistent

series

which

defining

relationship
could

series.

from observed

data.

from the annual

choice

be used&to
Based on this
the

An alternative
to the partial-

is to use a conversion

the hydrologic

per

series.

is to convert
the first

that

are independent,

14, the Work Group recommends that

be developed

For this,

flood

depend upon the investigator

to the partial-duration

solution

all

a partial-duration

basis

guidelines

and partial

peak for

a partial-

at least

when using

in a partial

in Appendix

by taking

must be considered,

(8) was made to determine

between

or partial-duration

base flood.

including

No specific

existed

is obtained

separation

to be included

annual

The base is selected

to establish
for

either

a predefined

events

The basis

and the intended

factor

than

problem

is common practice

study

the

A more

is based on the maximum flood

are evaluated

A major

is to define

flood

using

may be appropriate.

of interest

time period.

flood

substantial,
value,

in streamflow

series

more than one flood

duration

It

If

a corrected

of Frequency

series

each year.

series

relationships.

for

of error

can be analyzed

The annual

series.

all

because of the

Selection

Flood events

If

in error

of the data collecting

and a request

Determination

Peak flow

or suspected.

to the attention

of sources

V.

analysis.

can be substantially

be apparent

evidence

complete

frequency

and stage-discharge

be brought

supporting
in

stage

errors

should

flow

floods

in both

At times
errors

in flood

region

in which

the

gage is located.

[e.g.,

WI

is to use published

relationships

Except
series,

The second choice

for

the preceding

the procedures

discussion

described

of the the partial-duration

in this

guide

apply

to the annual

flood

series.

6.

Statistical
1.

events

Treatment

The Distribution--Flood
which,

as far

known statistical

tractable
a study

This

from this

study

log transformation
should

it

studies

2.

that
for

skew coefficient
Fitting

a log-Pearson

goods

of the data

which

mr&l

Type III

distribution

of defining

flood

a distribution.

of many possible

the Pearson

distribu-

The Work Group concluded

Type III

of annual

as described

any one specific

to assign

14.

analysis

of natural

best m@t the purposes of this

(log-Pearson

compute the base 10 logarithms

probability,

fit

in Appendix

the Distribution--The
Type III

however,

to find

is summarized

be the base method

generalized

is necessary,

fitting

and other

do not

To make the problem

was sponsored

tions and alternative

guide.

are a succession

as can be determined,

distribution.

probabilities
Therefore,

events

distribution

with

distribution)

series

data using

in the following

section.

recommended technique

for

to observed
of the discharge,

annual

a
fitting

peaks is to

Q, at selected

exceedance

P, by the equation:
Log Q=X+KS

where x and S are as defined

of the skew coefficient
K can be obtained

and selected

from Appendix

The mean, standard

may be computed

below and K is a factor

using

exceedance

(1)
that is a function

probability.

Values

of

3.

deviation
the following

and skew coefficient

equations:

of station

data

(34

=[

(XX;;

NX(X-X

fX)'/N

0.5

(W

13

(N - l)(N

- 2)S3

N*( ZX3)

- 3N(C

(44
X)(X

X2) -I- 2Lr:

X = logarithm

of annual

(4b)

Xl3

N(N-l)(N-2)S3

in which:
N = number of items

peak flow

in data set

x = mean logarithm
S = standard

deviation

G = skew coefficient
Formulas

for

computing

and G are given

equations
to the

3.

computation
digits,

errors

Estimating
(station

skew) is sensitive

accurate

estimated

skew coefficient

generalized

The following

skew estimates

guidelines

of values

only

provide

x,

S,

computed with

equations

used in their

3a and 4a

calculation,

When

for

number

a limited

3a and 4a are preferable.

Skew--The

skew coefficient

to extreme
from small

can be improved

skew estimated

the statistics

than with

facilities

Generalired

for

The precision

digits

equations

to obtain
with

2.

number of significant

of significant
record

in Appendix

of logarithms

3b and 4b is more sensitive

the available
*

the standard

of logarithms

by pooling

are recommended for

10

events;
samples.

by

weighting

information
estimating

of the station

thus

it

is difficult

The accuracy
the station
from nearby
generalized

of the
skew
sites.
skew.&

Guidelines
next

on weighting

section

of this

station

and generalized

mile
It

that

may be necessary.

skew coefficients

stations

within

have 25 or more years

in some locations

a relaxation

procedure

includes

drawn on a map; 2)

the mean of the station

generalized

or all

used should

The actual

skew isolines

developing

40 stations,

The stations

is recognized

1)

for

the use of at least

radius.

in the

bulletin,

The recommended procedure

requires

skew are provided

skew prediction

skew values.

of record.

of these
analysis

a JOOcriteria

by three

equation;

Each of the methods

methods:

and 3)

are discussed

separately.
To develop
troid

thecysoline
map, plot each station
-._drainage
basin and examine the plotted

of its

or topographic

trends.

and the average

values,
then

an isoline

record.

equation

used to calibrate
will

should

homogeneous

of about

skew coefficient.
geologic

Select

observed

with

that

20 stations

would

accuracy

situations,

should

either
map

variables.

estimating

The

the skew

the range of data

error

of estimate

of the preciction

equation.

of the skew coefficients

of the

40 stations

runoff

with

the arithmetic

may be used to estimate

The drainage

relate

considered.

of the station

are within

mean and variance

In these

is evident,

from the isoline

The MSE (standard

obtaining

and isoline

not further

that

be used for

variables

are drawn
be used in

no pattern

and climatologic

preferably

as to preclude

characteristics

the station

If

In some cases the variability

hydrology.

isolines

the skew coefficient

watershed

the arithmetic

may be so large

any geographic

The MSE will

be developed

affect

the equation.

stations.

between
map.

be used to evaluatethe

Determine

data for
then

or the differences

include

at stations

variance

should

These would
equation

at the cen-

be drawn and is therefore,

that

coefficient

all

of the isoline

variables

prediction

squared)

differences

skew coefficients

to predictor

is evident,

(MSE), is computed.

map cannot

A, prediction

for

error

the accuracy

the station

a pattern

of the squared

mean-square

appraising

If

skew value

areas and meteorologic,

be representative

regime

reasonably
mean and

the generalized
topographic,

of the region

around

of interest.
the method that

provides

the most accurate

11

skew coefficient

and

estimates.
diction

Compare the MSE from the isoline

equation.

of the data.

If

method with
for

If

neither

more accurdte

MSE should

then

the MSE is significantly

the smaller

MSEc

ance,

The smaller

MSE should

map to the MSE for

the isoline

map nor the prediction

of the

than

be used and that

MSE is not significantly

estimate

be compared to the variance

smaller

the smaller

skew coefficient

than

be used aS 'it

estimate

should

be used in equation

In the absence
read from Plate
of generalized

of detailed

I found

skew was developed

for

the

the procedures
still

not to develop

an alternative
their

The accuracy
not comparable
given.

of a regional

to Plate

A comparison

approximately

for

use with

of subregions

within

the United

be made between
area.

generalized

a given

can be combined

watershed.

unbiased

of the weighted
generalized
errors.

This

from Tasker

concept
(39) which

is expressed
should

efficient:

station
that

to their

Gw =
MSEg + MSEC
12

that
would
that

cover
be

are

and generalized
of skew for
skew is

error

the station
individual

in the following

+ MSEC(Q

of Plate
are not

the generalized

be used in computing

MS+(G)

States

estimate

by weighting

proportion

is generally

accuracy

skew, the mean-square

is minimized

skew in inverse

who prefer

I accuracy

to form a better

of station

estimate

those

skew relationships

Under the assumption

and independent

I is

relationships

Plate

applicable
to the entire
country.
4. Weighting
the Skew Coefficient--The
skew coefficient

for

to

Plate

skew relationship
the average

to other

aspects

However,

While

the same geographical

comparable

introduced

skew procedures.

generalized

only

map

the statistical

in all

the guide

can be

This

was first

I accuracy.

should

accurate

aEgo

guide.

bulletin

guide.

Vale.

skew (c)

of this

do not conform

own generalized

the accuracy

directly

stations

most

5 for

used to develop

recommended in the current

considered

is given,

when this

The procedures

individual

tne

the vari-

does the mean

provides

pocket

than

provides

the generalized

in the flyleaf

and has not been changed.

analysis

studies

the

MSE used in equation

equation

should

variance

the variance,

smaller

The mean skew coefficient

and the

the pre-

(MSE)
and

mean-square

equation
a weighted

adopted
skew co-

i(- where Gw

= weighted

skew coefficient

= station

= generalized

skew

MSEc

= mean-square

error

of generalized

= mean-square

error

of station

MSEG
Equation

skew

of generalized

of MSEc = 0.302
skew for

from the results

(40).

Their

is a function
ing Gwl this
accuracy

skew,

provided

When generalized

skew can be estimated.

station

skew

5 can be used to compute a weighted

of the source
the value

skew

should

log-Pearson

of record
function

skews are read from Plate

Type III

show that
length

regardless

the MSE of the generalized

be used in equation

of Monte Carlo

results

skew estimate

5.

random variables

experiments

The MSE of the

can be obtained

by Wallis,

Matalas,

the MSE of the logarithmic

and population

skew.

(MSEG) can be approximated

I,

and Slack

station

skew

For use in calculat-

with

sufficient

by the equation:

[A - B ~Lw10(W10)9]
MSEG"-10
Where A

(6)

-0.33

f O.OSlGl

if

IGI LO.90

-0.52

f 0.3OlGI

if

IGi

0.94

- 0.26IGI

if

IGI 51.50

if

IGJ >I.50

0.55
in which

IGJ is the absolute

estimate

of population
adjustment

historically

adjusted

described
skew,%,

G and N, respectively,

computations,
values

for

equation
selected

of the station

skew (used as an

skew) and N is the record

the historic
for

value

>0.90

in Appendix

and historic

in equation

6.

lengths

If

6 has been applied,

the

H, are to be used

For convenience

and station

13

in years.

period,

6 was used to produce

record

length

table

1 which

skews.

in manual
shows MSEG

TABLE 1, - SUbiMARY OF MEAN SQUARE ERROR OF STATION

0.6
0.1
3.2
0 3
(3:;
0. 5
0 l et
0.7
3-c
it.9
1

1.1
z

1-2
I.3
1.4
1:;
1

1.7

1.E
I.9
2.0
2. f
2.2
2.3
2.4
2.5
2.6
2.7
2 .P
2.9
3.0
*

SKEW AS A FUNCTIONOF RECORD LENGTH AND STATION SKEW. Jt

Application

of equation

values

(logs)

greater

little

weight

to the station

give

improper

6 and table

than

2 and long

Broken

of conditions

removal.

In this
is

not
case,

record

total

6.

in which

or too

high

period

because

different

length

equal

change

in the

Record--An

some peak

to record,
of

5 may also

of the watershed

certain

flood

relatively

generalized
and station
an examination
of

to the station

for

to

gives

should

skewe

years

may be missing

magnitude,

such as gage

record

segments

to the

sum of

watershed

skew

are

analyzed

both

between

as

records,

unless

segments

which

may

nonhomogeneous.

Incomplete

record

related

with

record

given

peaks

the

some physical

make the

weight

Record--Annual

because

a continuous

characteristics

greater

absolute

of equation

to the generalized
skew if the
In these situations,
by more than 0.5.

be made and possibly

with

of record

Application

value.

the data and the flood-producing

there

periods

weight

skews differ

5.

1 to stations

incomplete

flows

or the

are

record

missing

because

gage was out

flooding.

Missing

refers

high

to a streamflow

they

were too

of operation

for

and low data

require

low

a short
different

treatment.
When one or more high
record

have not

from

which

data

collecting

been

the

peak

should

be documented

some years.

adjustment
for

Zero

Flood

the

entire

year.

++distribution

because

gage sites

the

the

Years--Some

streams

Thus,

annual

the

the
flood

of

flood

values.
bf

zero

with

zero

is recommended
flood

years
15

If

not,

analysis,

and
it

not

reached

probability
+e

5.
regions

have

for

these

precludes

the

recommended
is minus
for

the

advised.

gage is

series

This

logarithm

for

instances

conditional

in arid

the

curves

the

in Appendix

using

adjustment

frequency
agency

data

probability

available

In most

use of the

as described

of systematic

information

flood

bottom

period

such estimates.

collection

situation

tional

records

of

data

of the

the

is usually

provides

and the

have one or more zero

analysis

during

can be estimated.

routinely

is recommended

statistical

there

discharge

For this

7.
will

peaks

is made as part

At some crest
*in

recorded,

agency

such an estimate

annual

streams
normal

log-Pearson

infinity.

determining

as described

no flow

Type

III

The condifrequency

in Appendix

5.

Mixed

8.

different

Population--Floodin-g

types

abnormally

large

when plotted
the

of

separate

with

for

each type

of

(1)
cyclonic

ating

during

general

those

operating

Nevada

during

region

rain

through

March,

and for

of April

through

July.

caused
example

in some instances

floods

been separated,

thereby

When it

the

causes

data
data

sets

by calendar

using

reasonable

clearly

by different

the

unless

of this
exception

developed
If

guide

the

tions

cannot

ally

meaningful

flood

that
for

be identified
criterion,

the

that

events

are

believed

record

one population.
16

by cause-caused

events

where
have

in

in

the

(11).

Separation

is not

considered

separate

periods

series
cannot

separately
be used

examined.

to comprise
by an objective

shall

are

The fitting

each flood
being

the

to combine

conditions.

fit

and genera

segregate

and then

skew coefficients

type

during

Coasts,

to

by events

events

and separated
the

segregated

two or more distinct

hydrometeorologic
generallzed

occur

and Gulf

separately,

can be used to

months

estimates.

of separation
unless

been made
the

which

from

Sierra

predominately

such as described

specific

events

are

oper

summer

the

during

and nonhurricane

each set

in lieu

local
have

may be more reliable

procedures

periods

caused

there

there

were

Atlantic

frequency

it

analyze

the

storms

different

is in

and those

hurricane

improving

hydrologically

with

from

of floods

by cause,

procedures

along

can be shown that

independent
flood

is

quite

floods,

Peak flows

by snowmelt

tropical

and relationships

principally

snowmelt

combined.

flood-producing

or during

studies

occur

then

of

population

Frequency

changes

by computing

intense

usually

floods

of mixed

which

Another

are

snowmelt

California.

predominately

by rain.

spring

(2)

factors

flood

floods

of November
those

rain

are

types

and

Hydrologic

One example

of
for

the months

storms.

slope

with

In some situations

The curves

snowmelt

by

curves

be described

of different

with

winter

floods.

separately

rain

created

frequency

paper.

can best

event.

of combinations

is

by abnormal

probability

events

general

cloudburst

normal

of annual

include:

in flood

reflected

curve

Two examples

some w,atersheds

results

skew coefficients

curves

events

This

events.

on logarithmic

frequency

in

be treated

two or more popul

and hydrologicas coming

from

-Ifi-

9.

Outliers--Outliers

from the

trend

of

deletion

of these

the

computed

procedures

for

treatment
on the

flow

If

the

are considered

If

low outliers

are

considered

2 0.4,

for

both

eliminating

any outliers

The following

the

statistical

small

samples.

require

judgment

considerations.
described

12 (figure

12-3),

than

station

tests

skew is less
Where the

the

below,

+0.4,

and low outliers

from

data

is used

All

involving

The detection

are

first.

equation

and are

for

high

than

station
should

and

-0.4

outlined

outliers
tests

for

skew is between
be applied

before

set,

to detect

high

outliers:

= x + KWS

xH
where

for

significantly

modification,

affect

ultimately

the

high

depart

The retention,

skew is greater

first,

which

especially

outliers

in Appendix

station

tests

data,

and low outliers

chart

the

data,

and hydrologic

of high

points

can significantly

from

treating

mathematical

data

remaining

outiiers

parameters
both

are

XH

high

outlier

mean logarithm
events9

(7)

threshold

in

log

units

of systematic

peaks

below

peaks

gage base,

(X's)

excluding

and outliers

zero

flood

previously

detected.
S
KN
If

the

standard

K value

logarithms

7 then

they

outliers

are

be compared

sites.

If

is

is treated

as historic

then

the

hl'storic
they

ment of all
in the

of X's

Appendix

considered

outlier(s)
useful

from

of peaks

should

at nearby

deviation

4 for

in a sample
high
with

information

sample
are

historic

flood

available

information

should
historic

be retained
flood

data

as part
data

of

of

and high

analysis.

in

time,

should

a high

the

Section

outlier(s)

V.B,lO.
for

high
information

indicated

systematic

outliers

equation

considered

to adjust

the

XH in

and flood

which

period

available

peaks
data

as described

is not

N
than

Flood

maximum in an extended
flood

greater

outliers.
is

size

high

record.
be well

If
outliers,
The treatdocumented
*

17

The following

equation

is

used to detect

low outliers:

XL= x - KNS
XL = low outlier

where
defined

for
If

then

threshold

equation

for

following

flood

is

used to detect

equation

XL = low outlier
= K value

KH

If

units

historic

xL
where

log

and the

=x

threshold
from

dai;a

in

log

Appendix

4 for

(8b)

units
period

used to compute%

mean logarithm

-Y

= historically

adjusted

standard

of any annual

8a or b, then

considered

low outliers

probability

adjustment

If multiple
recommended
desirable

they
are

to test

that
are

the

deleted

in a sample
from

very

not

close

sensitivity

the

been
to the

of

the

are

less

low outliers.

in Appendix

have

and?

deviation

considered

described

values

procedure

peaks

are

been made,

- KH:

adjusted

equation

are a:

low outliers:

= historically

logarithms

terms

has previously

the

other

7.

an adjustment

the

in

h-4

record

than

XL in

Flood

peaks

and the

conditional

5 is applied.
identified

as outliers

threshold

results

value,

by th

it

to treating

may be

these

valu

as outliers.
Use of
that

detects

K values

are

equation

outliers.
testing
flood
the
from

K values

outliers
based

In this

liers.
the

the

from

at the

Bulletin,

the

simulated
these

and comparing
data
populations

were

test

of this

procedures

between
rangina

level

of
for

is applied

once

outlier

detection

between

with
-3.67

values

t
Th

of single

and all

out
above

8a or b are consid
procedure

was based

III

and observ

log-Pearson

The population

(38).

detection

from equation

4 1.5,

to a one-sided

significance

that

on simulated
results.

4 is equivalent

distribution

or below

The selection
data

10 percent

on a normal

7 threshold

several

Appendix

Type

skew coefficients

skews for
and +3.25.

samples

for

selected

The skew value

'for

the

observed

evaluated

included

The selected
cedures

Flood

which

occurred

are maximums

in

computations.
the
curve

in Appendix

to this

historically

two analyses,
suggest

tended
C.

that

record.

All

Refinements

analysis
those

of

flood

directly

partly
data

and partly

data

is

ating

flood

decisions

data

be used in

frequency

of the

on the

data

are

data,

over

the

computed

frequency

described

used.

The underlying

from

the

systematic

record

between

the

systematic

and

results

from

systematic

be used unless
of

the

data

observed

are

not

made should

the

and

comparison

peaks,

indicative

or other
of

be thoroughly

the

ex-

documented.

Curve
probability

estimates

deteriorates

based

for

probabilities

by the

period

of

systematic

the

sampling

error

of

the

because

the

of

record

basic

underlying

upon statistical
more rare

record.

statistics
distribution

from

than

This

is

the

station

of flood

known exactly,

Although
and flood

historic

defined

because
not

the

systematic

be made.

should

magnitude

of flood

of

should

the

data

the

any flood

conditions

thasc

period

Comparison

that

The adjustment

the

intervening

to Frequency

The accuracy

of

Based on

4 3.

reliability

analyst.

is that

information

the

pro-

use with

of skews,

in watershed

when historic

analyses

for

indicates

skew.

other

appropriate

the

effects

the

to apply.

should

used,

by the

lengths.

The hjstoric
factors

and the

the

adjusted

of time

changes

than

range

procedures

and zero

and easy

or after

are

adjustment
of

record

of the

time,

be evaluated

is representative
historic

data

SUCII

6 is recommended

assumption

the

test

weighted,

which

period

magnitude,
of

must all

considered

during,

an extended

period

simple

over

before,

Other

or better

- Information

Data

peak discharge

extended

is

distribution

Before

and t2.80.

as well
being

procedure

III

Type

Historic

peaks

-2.19

generalized,

performed
same time

this

log-Pearson

between

use of station,

at the

results,

10.

were

procedure

while

these

data

other

data

from

levels

procedures0
adjoining
at

high

for

estimating

watersheds
flows

and rare

19

floods

can sometimes
exceedance

on a watershed
be used for
probabilities;

evalu-

procedures

for

procedures

discussed

information

doing

is

so cannot
thus

needed

can justifiably

far.

will

be spent

of the

recommendations
data

0.01

exceedance

Analyses

following

Analysis

Comparisons

with

types

magnitudes
sections
in the

for

from

guides

frequency

with

flood

frequency

and effort

that

with

water-

other

The remainder

for

use of

these

analysis.
the

length

flood

magnitudes

of systematic

with

record

Length of Record
10 to 24 25 to 50
Watersheds

Precipitation

of analyses
exceedance

explain

of time

as the

as shown

tabulation:

Similar

Estimates
All

are

Include

to

Statistical
Flood

flood

vary

the

precipitation.

when determining

probability

by an X in the

to include

which

amount
from

section

the

for

same extent

and make comparisons

estimates

in this

to refine

The analyses

the

to obtain

and make and use flood

to the

The purpose

determine

sheds,
additional

be standardized

how to

X
--

should

be incorporated

probabilities
include

of

the

various

X
---

when defining

less

than

types

Available
50 or more

0.01.

flood

The following

of flood

information

analysis.
1.

Comparisons

and storm
nearby

records

with
(see3

hydrologically

and interpreting
analysis
unusual

both

of a given
a given

Similar
e.g.,

(12))

watersheds

unusual

flood

event,

will

flow

greater

the
will

between

frequency

often

experience

The shorter
the

comparison

and flood

similar
watershed,

flood

Watersheds--A

aid

flood

ana!yses

at

in evaluating

and the

flood

flood

record

and the more

need for

such com-

be the

frequency

parisons,
Use of the
one form
and are

of

weighted

regional

described

skew coefficient

comparison.
in

the

following

recommended

Additional
paragraphs.

20

comparisons

by this

guide

is

may be helpful

Several
a short

mathematical

record

to

Such procedures
are on the

reflect

usually

apart.

given

in Appendix

at least

10 percent
in

are

prediction
method

the

should

appraisal,

the

accuracy

exceedance

probability

Flood

several
estimates,
model,

from

than

Observed

of

upon flood

can improve

(13),

(15),

(14),

area

plots.

to

its

of specific

appraising

the

accuracy

guide.

In the

absence

estimate

estimates

of the

two

given

variance

as

procedures

for

of

such estimates

of

an accuracy

of a flood

equivalent

the

As these

The weight

proportional

can

(16)],

average

alone.

mc:st

parameters

probabilities

either

a regional

based

exceedance

Precipitation--Floods
(rainfall

to that

with
from

(e.g*)

be made of the

watershed

maximum event

flood

such procedures

from

0.01
an analysis

in

potential.

Such

watershed
for

are

flood

estimates.

upon the

procedures

observed

can be used

calibrated

relationships)

to make these
depends

snowmelt)

of watershed

a procedure
Unless

estimated

and/or

definition

require

is warranted
use to

in accuracy

curve:

a weighted

hydrograph,rainfall-runoff

such effort

is

record.

cipitation
to discharge.
effort
may be required

a missing

analysis,

this
of

improve

however%

and on the

drainage

Recommendations

precipitation

unit

or simple

can be assumed

Estimates

ways to

50

to those

of statistical

selected

be inversely

of

station

or estimated

with

[e.g.,

or for

scope

2.

is confined

watersheds

equations

8.

the

of a lo-year

stations

such adjustments

and an improvement

similar

be more accurate

comparisons

beyond

gaging

not more than

making

of a frequency

station

in Appendix

regional
are

(17),

of

to each estimate

for

station,

when the

centers

Comparisons

of flows

will

described

short

determinations.

independent

estimates

are

hydrologically

distribution
flood

long-term

only

with

adjusting

for

can be demonstrated.

nearby

be made using

results

procedure

and adjustment

frequency

index

useful

been proposed

The use of such adjustments

records

Comparisons

of the

at a nearby

or in watersheds

73

where

flood

experience

The recommended

situations

experience

have

yield

same stream

miles

procedures

converting

available,
Whether
and data

considerable
or not
available

estimate.

precipitation

can sometimes

in an incomplete
21

flood

record,

pre-

be used to estimate

Observed
stations

watershed

in a meteorologically

a synthetic

record

records

are

of this

guide.

studies,

Consequently,
for

developed

Because

of

at this

time.

from

Analysts

methods

for

probability

making

computing

flood

based

at the

exceedance

a guide

for

adjusting
flood

of
these

are

certain

are

discussed
It

of risk

sections
the

curves
in

ity

the
that

period

first

of

evaluate
experience
Plotting

region.

determined
the

stand-

provides
Any adjust-

curve.

flood

estimate

and

would

of years.

presented

recommended

curve

solution

at a gaged

of water

considerations

which

procedures
location.

resource

must

for
When

problems,

be kept

there

in mind.

These

section.

to make a distinction

If

a specified

Application

frequency

in hydrology

between

the

concepts

(18).

is a permanent
risk

should

frequency

accuracy

have

to the

this

useful

such as floods.
then

flood

and uncertainty

Risk

recommended

and storm

thus
the

Reliability

additional

is

is

data.

The preceding
applying

should

values.

be used and then

probability

relative

such

records

experienced

homogeneous

and extrapolating

VI.

determination

to

upon flood

flood

other

beyond

procedures

the

making

scope

can be used to adjust

use of such

of the

the

estimates
procedure

and meteorologically

the

flood

no specific

in a hydrologically

recognize

for

the

been evaluated.

extrapolation

of occurrence

ments must

not

precipitation

is outside

procedures

precipitation

the many variables,

as adequate

when available
have

at nearby

can be used to generate

technique

alternative

including

curves,

as many years

deciding

observed

region

of the

by such procedures

frequency

ardize

for

Appraisal

or criteria

Floods

or precipitation

homogeneous

of floods

available.

be extended

its

precipitation

the

population
population

be exactly
flood
Risk

is

property

of any random

distribution
known.

magnitude
inherent

The risk

were

be exceeded

in the

phenomenon

be avoided.
22

known for

is stated

will

phenomenon
floods,

as the

probabil-

in a specified
itself

and cannot

Because
because

use is made of data

population

properties

some technique,
into

the

various

flood

properties

and estimates

constitute

uncertainties.

water

resources

be decreased
statistical

which

frequency

losses

are

by

introduced

from

the

data

by various

uncertainties

data

population

sample

or minimized

while

more or better

data

between

derived

them.

adjusting

purpose

should
of

A,

Confidence

the

can

and by using

better

of

an estimate

the total

provide

are

inherent

specific
the

guides

instruction

time

expected
as to how

on when to

methods

on the

it

How well

(population)

depends

or not

the

either

a measure

that

is not

a sample.

depend

on

curve

an exact

this

sample

upon the

underlying
of

the

the

at a selected

exceedance

can be computed

representation.
predict

sample

size,

is known.

uncertainty

discharge

is

will

distribution

of a selected

discharge

of

the

or a measure

of

probability.

by the

procedure

9.

of confidence
upon

!iitilits

in

the needs

of the

frequency

curve

that

the

only

best

estimate

become available,
confidence

be aware

curve;

discharge

to emphasize
more data

should

probability

Appendix

depends

today's

curves

is only

of the

limits
in

than

for

to obtain

given

computing

uncertainties

risk

curve

rather

population

and whether

Application
decision

frequency

experience

uncertainty

described

frequency

be applied

exceedance

ConFidence

calculating

(a)

estimate.

limits

estimated

the

of the

flood

Confidence

(b)

the

on when to use each of the

record

-its accuracy,

to use for

Limits

The user
A streamflow

procedures

The recommendations

Decisions

the

outline

can be used to evaluate

estimate.

the procedures

and the

these

can be decreased

determination,

and (c)

probability

the

properties

and measures,

sections

limits

periods,

only

Risk

from

and

techniques.

confidence

apply

of these

or biased,

Differences

determination.

by obtaining

The following
in the

and information

developments
only

are deficient,

must be estimated

errors

frequency

which

limits

of
the

the

reaching

user.

This

developed

flood

estimate

narrowed.
23

resource

discussion

using

frequency
will

water

this

planning
is presented

guide

distribution.

normally

be improved

is
As

B.

Risk
As used in this

one or more events

period

will

of years.

representing

guide,

risk

exceed a given

Accepting

the flood

the flow

exceedance

may be computed

for

the probability

of exceedance,

interval
however,
during

flood

ignores

will

the lifetime

time period.
annual

or more times
Procedures

for

C.

Expected

probabilities

between

the confidence

represents

The study

calculations

can also
during

be

a specifiec

that

the flood

be exceeded

are described

hydrology

is defined

texts

one

in Appendix

or in (19)

adjustments

mented as to whether
curves

for

samples

and (20)

of the true

any specified

flood

of a specified

of the central

tendency

size

frequent
[(B),

of the spread

the Work Group (8) and summarized

adjustments

procedures

final

as the average

estimates

correct

published

be the

for

that

recommends that

If

curve

will

occur

limits.

committee

made.

will

chance

that

statement

event

exceedance

of 1 percent

these

a measure

are approximately

would

a rare

This

chance

70 years.

magnitude

conducted

14 indicates

the statistical

It

year.

is a 50 percent

be made from successive

It

bution

that

of a flood

probability

(21)].

Appendix

of risk

is a 1 percent

The frequency

in most standard

of all

might

as accurately

Probability

The expected
that

a specifiec

For a l-year
period
reciprocal
of the recurrence

in a given

probability

that

within

an estimate

there

risk

there

making

10 and can be found

curve

Thus,

the probability

in the next

frequency

is the

of a structure.

exceedance

magnitude

period.

which

be exceeded

For instance,

flood

time

the risk.

as the probability

probability,

the considerable

used to estimate
with

any selected

T, expresses

the loo-year

is defined

for

frequency

described
if

or not the expected

are plotted,

guide.

probability

to the normal

in the frequency
they

for
curves

in this

an expected

applicable
step

[(21),(Z)]

probability

must be appropriately
24

the normal
computed

distri-

using

Therefore,

the

adjustment

is made,

distribution

analysis.

in

be used.
It

must be docu-

adjustment
labeled,

is

It

should

ment that

be recognized
adjustments

such

of uncertainty
curve

expected

of confidence

limits

independent

and risk

The decision

about

decision

in estimates

and in obtaining
frequency

use of the expected

flood

flood

frequency

used in computation
weighted

averages

of

discharge.
probability
It
guide.

beyond the scope of this

of annual

adjust-

the effects

The basic

is the curve

of flood

probability

to incorporate

of the curve.

probability

estimates

the expected

are an attempt

in application

without

a policy

when using

adjustment
is
is most often used

damages and in establishing

design

flood

criteria.
Appendix
bility

11 provides

and further

precedures

description

sections

but not essential

A,

This
flood

guide

potential

the

as expressed
the

point

approaches

and that

are not followed,

information

that

is of interest

guide,

set of procedures
by a flood

is made that

in those
deviations

a comparison

proba-

Situations

describes

Introduction

including

provide

Special

the expected

Potpourri

to use of this

Non-conforming

computing

of the concept.
VII.

The following

for

flow

special

cases where

recommended for
frequency

situations

of this

by appropriate

obtained

with

In the

may require

the procedures

must be supported

of the results

curve.

defining

those

other

guide
study,

obtained

using

the recommended procedures.

It

is not anticipated

approaches
are limited
historic

will

future

modifications

of the station

reasons

for

should
departing

are to be documented

comparison

situations

special

of results.

be called
of

data

including

be followed

guide.
25

on analysis

For consideration

of

there

are

procedures.

by appropriate

Subcommittee

other

records

unless

from the guide

and supported

attention

warranting

recommendations

The Hydrology

to its
this

situations

and specific

These procedures

technical

These deviations
including

Detailed

to the treatment
events.

compelling

these

occur.

that'many

asks
in

study3
that

The map of
encouraged
the

Major

by the

flood

their
analysis

estimates

record.

for

interest

in

the

at gaged

using

locations

probabilities

For these

incorporate

of

V.B.3.

frequency

flood

region

more rare

situations

analysis

the

but

are

allows

guide

than

described

considerable

Position

Calculations
of a plotting
to permit

specified

in this

position.

data, states

studied

for

Users are

estimate.

in analysis.

Plotting

data.

in

to

a generalized

studies

available

information

is

in Section

when making

latitude
t3.

outlined

problems

encountered
the

I)

to make detailed

procedures

defined

skew (Plate

that

the

an evaluation

The merits

Section
results
of the

for

require

describing

analysis

effect

of alternative
formula

do not

V.B.TO.,
of the

and no recommendation
A general

guide

plotting

treatment

should

on the

of historic

be shown graphically

analysis

position

designation

of including
formulae

historic

were

not

is made.
computing

plotting

positions

(23)

is
(9)

pJ!!xL-

(N-a-b+l)
where
m = the
the
*

orderedsequence
largest

IJ = number

of

upon the
distributions

of

equal
items

to
in

flood

values

with

data

distribution.
a=b and the

set

and a and b depend

*
For symmetrical
formula

reduces

(10)

p=(m-_a)

(IWa+l)

26

to

The Weibull

plotting

position

0 was used to illustrate

use of the

and has been incorporated

13,

to

This

facilitate
easily

Plotting

program

comparisons

Position

(18,

10 equals

adjustment

computer

was used because

understood

a in equation

historic

and analysis

position

intuitively

Weibull

in the

data

plotting

in which

it

of

figure

referenced
by the

6-3

in Appendix

program

is analytically

user.

simple

and

24).

formula:
p=m

(11)
N+f

C.

Future

that

Studies

This

guide

is designed

the

Federal

Government

accurately

defining

flood

damage abatement.
the

twin

hoped

goals

that

guide

of a coherent

Although

selection

of the
members,
the
of

as needed

set

the

and special

studies
studies

Selection
fitting
(b)

Initially
the

the

flood

the

proper

determination.

made based

upon a review

experience

of Work Group

in

the

of alternative

selection

process,

is a list

distribution

conditions.

Time did

not

and

warranted.

had expected
for

upon watershed

procedures
distributions

is believed

Work Group

27

essential

to be used in each step

and fitting

idea.

is

by the Work Group.

depending
this

frequency

Following

identified

procedures

before

by defining

recognized.

study

required

effort

to aid

of distribution

Continued

flood
It

practical

studies

are

of

for

be obtained.

procedures

is

procedures

will
for

conducted

demand

in programs

has been carefully

needed

(a)

of proedures
analytical

of

and data

to this

considerable

additional

some additional

study

contributes

analysis

literature,

1.

set

a coherent

and consistency

of the

of the

need for

develop
potentials

accuracy

elements
or element

ever-pressing

Much additional
of

this

to meet a current,

a watershed

to find

that

would

vary

and hydrometeorological
permit

exploration

of

(c)

More adequate

criteria

are needed for

selection

of a distribution.
(d)

Development

of techniques

homogeneity

of series

2.

The identification

3.

The treatment

of outliers

computational

procedures.

4.

Alternative

5.

More adequate

6.

Procedures
into

7.

for

Guides

for

III

with

defining

by urbanization

of mixed distributions.
as to identification

treating

historic

procedures

for

and

data.

confidence

limits

distribution.
flood

estimates

from precipitation

analysis.

defining

and watersheds
8.

computation

to incorporate

frequency

Guides

both
for

evaluating

is needed.

and treatment

procedures

to the Pearson

for

flood
limited
flood

potentials
gaging
potentials

and by reservoirs*

28

for

ungaged watersheds

records.
for

watersheds

altered

Appendix

REFERENCES
1.

Brater,
E. F. and J. D. Sherrill,
Fre uencies of Floods in Michi an
--%-r-3--+
Univers ty o Mlc igan, Ann r or,

2.

Gann, E. E.,
in Missouri,"

3.

Thomas, C. A., W. A. Harenberg,

and 3. M, Anderson,
"Magnitude and
Frequency of Floods in Small Drainage Basins in Idaho,"
USGS Water
Resources Inv. 7-73, NTIS, Springfield,
VA., 1973.

4.

Todorovic,
Analysis,"
1648.

5,

Carrigan,
P. H., Jr., and C. S. Hutzen,
Annual Floods,"
International
Hydrology
September 1967.

6.

Matalas,
N. C.,
U.S. Geological

7.

Pacific
Southwest Interagency
Commission Report of the Hydrology
Subcommittee,
"Limitations
in Hydrologic
Data as Applied to Studies
in Water Control Management," February 1966.

8.

Beard, L. R,, --Flood Flow Frequency

in Water Resources,
The Unlverslty

9.

Langbein,
W. B., "Annual Floods and the Partial
Transactions,
American Geophysical
Union, Vol.

"Generalized
Flood-Frequency
Estimate for
USGS Open-File
Report, 18 pp., 1971.

Urban Areas

R. and E. Zelenhasic,
"A Stochastic
Model for Flood
Water Resources Research, Vol. 6, No. 6, 1970, pp. 1641"Serial
Correlation
of
Symposium, Fort Collins,

"Autocorrelation
of Rainfall
and Streamflow
Survey Prof. Paper 434B, 1963.

Techni ues Center

+o Texas at Austin,

Minimums,"

for Research
1974.

Duration
Series"
30, 1949, p. 879.

10.

Hardison,
C. H., "Generalized
Skew Coefficients
of Annual Floods in
the United States and Their Application"
Water Resources Research,
Vol. 10, No. 5, pp. 745-752.

11.

U.S. Army Engineer District,

Sacramento,
California,
Civil Works
Investigations
Project
CW-151, Flood Volume Studies-West
Coast.
Research Note No. 1, "Frequency of New England Floods,"
July 1958,

l-l

Engineers,
n Washington,

in the United

Q&of;

13.

Benson, M. A., "Evaluation

of Methods for Evaluating
of Floods,"
USGS Water Supply Paper 1580-A, 1962,

14.

Benson, M. A., "Factors

llh$d Region of Diverse
.

15,

Benson, M. A., "Factors Affecting

the Occurrence of Floods
Southwest,"
USGS Water Supply Paper 1580-D, 1964.

16,

Bock, P,, I. Enger, G, P, Malhotra,

and D. A, Chisholm,
"Estimating
Peak Runoff Rates from Ungaged Small Rural Watersheds,"
National
Cooperative
Highway Research Program Report 136, Highway Research
Board, 1972.

17.

Dalrymple,
pt "Flood

T., Flood Frequent

Anal ses Manual of H drolo
Flow mques,"
Uf+G ater SumPaper ----b8$i96~

18.

Yevjevich,
Resources

Vujica,
Probabilit
Publications,'-l?X+o~ns,

the Occurrence

Influencing
the Occurrence of Floods in a
Terrain, " USGS Water Supply Paper 1580-8,

and Stat;s;;;so+H
C 1
d ,

Gumbel,
Science

20.

Riggs, H. C.,
Division,
&

in the

Part

drolo
e

Risk in Flood Control,"

E. J., "The Calculated
Research, Section A, Vol. 5, 1955, The Hague.

19.

U.S. Army,
1945.

"Storm Rainfall

12,

"Frequency of Natural
Eventseli Journal
&,
Vol. 87, No. HYl, 1961,m-x.-

Water
Applied
of the Hydraulics
34

21.

Beard, L. R., "Probability

Estimates
tion Samples," Journal -of Geophysical

22.

Hardison,
C., and M. Jennings,
"Bias in Computed Flood Risk,"
Journal
of the Hydraulics
Division,
PFoc. ASCE, Vol. 98, No. HY3, Marc!
l~i%
Ecussion
ana*-,
=415-427.

23,

Harter,
H. L., "Some Optimization
Problems in Parameter Estimation,"
edited by Jagdish S. Rustagi,
Optimizing
Methods $I- Statistics,
Academic Press, New York, 1971, pp. 32-62.

24,

Chow, V. T,, Handbook of A lied

New York, 196mpp,8-zB
--!hz9*
an

25.

Characteristics,"
Hardison,
C., "Accuracy of Streamflow
Professional
Paper 650-D, 1969, pp* D210-D214.

1-2

Based on Small Normal-DistribuResearch, July 1960.

Hydrology,

McGraw-Hill

Book Co.,
USGS

26.

Wilson,
E. B. and M. M. Hilferty,
"The Distribution
National
Academy of Science,
Proc.,
Vol. 17, No.
pp. 684-688.

12,

of Chi-Square,"
December 1931,

27.

McGinnis,
David G., and William
H. Sammons, Discussion
of Paper
by Payne, Neuman, and Kerri.
"Daily
Stream Flow Simulation,"
Journal
Division,
Proc. ASCE, Vol. 96, No. HY5,
--of the Hydraulics
May 1970.

28.

Jennings,
M. E. , and M. A. Benson, "Frequency
Curves for Annual
Flood Series with Some Zero Events or Incomplete
Data,"
Water
Resources
Research,
Vol. 5, No. 1, 1969, pp. 276-280.

29.

Matalas,
Hydrologic

30.

Gilroy,

31.

Beard, L. R.,
of Engineers,

32.

Natrella,
Handbook

33.

Hardison,

34.

U.S.
ized

N.,

and B. Jacobs,
"A Correlation
Data: USGS Professional
Paper

E. 3.

personal

communication

to

Procedure
for
434-E, 1964.
C. Hardison,

Statistical
Methods in Hydrology
Civil
Works Investigation
Projeci

M. G., Experimental
91, 1963.
C. H.,

personal

Corps of Engineers,
Computer Program,

Statistics,
communication,

Augmenting

1974.

U. S. Army Corps
CW-151, 1962.

National

Bureau

of Standards

1974

"Regional
Frequency
Computation,"
The Hydrologic
Engineering
Center,

General-

July

1972.
35.

Water Resources
for Determining
D.C.,
1967.

36.

Harter,
H. L., and A. H. Moore, "A Note on Estimation
I Extreme-Value
Distribution,"
Technometrics,
Vol.
1967, pp. 325-331.

37.
* 38.

Council,
Hydrology
Committee,
A Uniform
Technique
Flood Flow Frequencies,
BulletTn
15, Washington,
--

Thorn, H. C. S., "A Note on the Gamma Distribution,"

Review,
Vol. 86, 1958, pp. 117-122.

9,

from a Type
NO.
2, May

Monthly

Weather

Grubbs,
Frank E, and Glenn Beck, "Extension
of Sample Sizes and
Percentage
Points
for Significance
Tests of Outlying
Observations,"
Technometrics,
Vol. 14, No. 4, November 1972, pp. 847-854.

39.

Tasker,
Gary
Coefficient,"
pp. 373-376.

D., "Flood
Frequency
Analysis
with a Generalized
Water Resources
Research,
Vol. 14, No. 2, April-

Skew
1978,

40.

Wallis,
J. R., N. C. Matalas,
and J. R. Slack,
Just a Moment,"
Water Resources
Research,
Vol. 10, No. 2, April
1974, pp. 211-219.
36
l-3

41.

Resnikoff,
G. J., and G. J. Lieberman,
Tables of the Non-Central
t-Distribution,
Stanford
University
Press, Stanford,
California,
1957.

42.

Zelen, k, and N. C. Severo, "Probability

Functions,"
of Mathematical
Functions,
Applied Mathematics
Series
U. S. National
Bureau of Standards,
1904.

43,

for
Owen, D. B., "Factors
Variables
Sampli,ng Plans,"
March, 1963.

One-Sided Tolerence
Sandia

Corporation

Handbook
No. 55,

Limits and for

Monograph SCR-607,
=x

l-4

Appenaix

GLOSSARY AND NOTATION

Glossary
The terms
ences listed
Manual

used in this

guide

in the Bibliography

43, American

especially

prepared

statistical

terms*

and W. R. Buckland,

Society
for

include

Engineers,
of Statistical

Publishing

Flood

The maximum momentary

Hydraulics,"

and from definitions

definitions

Terms"

A list

peak discharge

(Sometimes

discharge
FZood

from refer-

of

by M. G. Kendall
1957.

Definition

of record.

Annual

1962,

Company, New York,

TERM
Annual

for

For more technical

guide.

see "Dictionary
Hafner

taken

or from "Nomenclature

of Civi'l

this

definitions

in each year

the maximum mean daily

is used,)

of annual

floods.

b!MVkS
Annual

Series

Array

A general

term for

which

each item

A list

of data

frequency
largest

Broken Record

in order
it

first,

the smallest

fdrst.

A systematic

record

ate continuous
discontinuation
of time.

2-l

of data in

is the maximum or minimum in a year.

analysis
value

a set of any kind

of magnitude;
is customary

to list

in a low-flow

which

segments

of recording

for

the

frequency

Is divided
because

in flood-

into

analysis

separ-

of deliberate
significant

periods

coef;faczent of
Skezl?ne88

n numerical
metry

measure

or inaex

In a frequency

the thtrd

0T tne

distribution.
Also

sym-

OT

Function

moment of magnitudes

measure of asymmetry.

lack

about
called

of

their

mean, a

"coefficient

of

skew" or "skew coefficient."

Confidence
Limits

Computed values
a parameter
bility

DistributionFree

a specified

the range

in which

the true

describjng

Requiring

the relative

of various

distribution

Frequency

magnitude,

Exceedance
ProbabiZity

Probability

that

specified

magnitude
unless

The average
magnitude

might

of proba-

exceed a specified

in a given

otherwise

for

probability.
will

exceed a

time period,

usually

indicated.
probabilities

of all

any specified

flood

be made from successive

fresamples

size.

Generalized Sketi
Coefficient

A skew coefficient

Homogeneity

Records

integrates

that

a random event

estimates

a specified

with

occur.

the kind

exceedance

of the true

that

frequency

about

of values

100 times

one year

quency

of the

a set of data may have,

The percentage

Probability

probavalue

magnitudes

no assumptions

Exceedance

Expected

of

lies.

events

bility

of an estimate

show for

Function
which

sides

that

parameter
Distribution

on both

values
from

2-2

derived
obtained

by a procedure
at many locations,

the same populations.

which

of

Incomp Zete

A streamflow

record

Record

are missing

because

high

to record

for

LeveZof
Significance

a short

level

they

some peak flows

were too low or

or the gage was out of operation

period

The probability
when it

in which

because

of flooding.

of rejecting

is in fact

true.

of significance

a hypothesis
At a "10-percent"

the probability

is

l/10.

Mean-Square
Ex~or

Sum of the squared

differences

true

values

and estimated

by the

between

of a quantity

number of observations.

defined

as the bias

the

It

squared

plus

divided

can also

be

the variance

of the quantity.

Method of
Mome&s

A standard
ating

statistical

computation

for

the moment of a distribution

estim-

from the

data of a sample.

Nonparametric

The same as distribution-free.

NopmaZ
lxs -haLtion

A probability
about
It

distribution

the mean, median,

is the most studied

tistics,

even though

normally

distributed,

in theoretical
distributions
It

is also

2-3

is symmetrical

and mode (bell-shaped).

distribution

in sta-

most data are not exactly

because

of its

value

work and because many other

can be transformed
known as Gaussian,

The Gauss-Laplace,
tribution,

that

into

normal.

The Laplacean,

or the Laplace-Gauss

or the Second Law of Laplace.

dis-

Outliers

OU-t;ZiC?X

(extreme

which

depart

from

A characteristic

Parameter

Chance

PopuZation

A probability

(Return

Period,
ante

Excsed-

Interva

2I

multiplied

from which

a sample

is taken

number

present,

on a river

lation

of

that

floods

floods

for

are

not measured

time

occurrences

of a hydrological

given

or greater

flood

series,

interval

average

even

if

between

actual

event

of a

In an annual
interval

in which

size

is exceeded

as an

In a partial

interval

regardless

popu-

average

of a given
maximum.

is the

or recorded.

magnitude.
the

and future

location

The average

duration

between
of

their

or any other

period

distinction

holds

recurrence

intervals

are

part,

or fragment

even

floods

series,
of a given

relationship

the year

both

Skew Coefficient

past,

of data

or collected.

at a location

size,

Sarrp i!e

of

number

floods

the

of data.

by 100.
infinite)

annual

rest

such as a mean

(usually

a flood

points

of the

The entire

the

Interval

trend

data

deviation.

The total

Recuxxence

the

are

descriptor,

or standard

Percent

events)

to

of time.

though
nearly

for

The

large
the

floods

same for

series.

An element,
Every

hydrologic

longer

record,

See "coefficient

2-J

record

of

of a "popUlat.iOn."

is a sample

skewness."

of a much

Standard

A measure

Deviation

of a series

of

the

dispersion

of statistical

as precipitation
the

square

of the

of the

deviations

from

mean divided
or events

in the

number

order
the

Standard

Error

by the

to get

from

Distribut<on
(t-distribution)

like

the

standard

from

it

values
is now
to divide
one in

estimate

sample

standard

Often

set

of

minus

the

of the

a single

arithmetic

statistics

an unbiased

is

squares

It

of values

variance

An estimate

the

number
in

of a statistic,

Student's

such

It

sum of

series.

practice

by the

values

or streamflow.
root

standard

or precision

of

data.
deviation

calculated

from

of observations.

Calculated

deviation

but

differing

in meaning.

A distribution

used in evaluation

variables

which

involve

deviation

rather

than

of

sample

standard

population

standard

deviation.

Test

of

Significance

A test

made to learn

result

is accidential

differs
the

from

many types

formulas
it

significance,"
generally

of

10 percent

2-5

another

of

not

For all

there

are

In making

to choose
choice

standard

a test

a "level
being

of
arbitrary

than

the

low level

nor more than

the

high

1 percent.

less

that

a result

result.

of tests

the

but

probability

or that

and tables.

is necessary

level

the

I flt:

WldlKJf2

U.1.

make later
a plot

it

UdLd

I.0

to linearize

a skewed

more nearly

changing

their

roots;

distribution

a normal

many others

of a set

of the
of

distri-

equal
A value
series
dividing

around

square

obtained
of values
the

2-6

of spread
their

the mean value

to the

or cube

or dispersi
mean,

of the

from

the

mean,

of

the

standard

by multiplying
by its

sum of those

weights.

roots

values

possible.

amount

deviations

sum of the

numerical

square
are

values

by calculating
of the

ordinary

logarithms,

A measure

Means

WI

The most common transformations

are those

WeCghted

VdlUfZb

easier,

or to normalize

bution.

Variance

computations

by making

into

flUlllt:IICd

assigned
products

obtaine
squares

and hence
deviation

each of a
weight

and

by the

notation
Appendix
is consistent,

notation
slight

is

described

variations

Fitting

parameter

Variate

in equations

B
b

Fitting

parameter

Variate

in equation

Skew coefficient

Generalized

Historically

H
KH
K
*

KN

equation

*:

6.

9 and 10 which

depends

upon the

used in equation
9 which

6.

depends

upon the

(23)

Gw

used in

(23).

distribution

most notation

Explanation

distribution
%-

While

do occur.

Notation

46

in each Appendix.

of

logarithms

of annual

peak discharges

skew coefficient
adjusted

skew coefficient

Weighted

skew coefficient

Historic

record

length

K value

from

Appendix

Pearson

Type

III

K value

from Appendix

4 for

historic

4 for

sample

deviate

74

Historically

adjusted

MSE

Mean-square

error

MSEZ

Mean-square

error

of

generalized

MSEG

Mean-square

error

of

station

Ordered

sequence

period

size

mean logarithm

of flood

skew

skew
values,

with

the

largest

equal

to 1

Number of

items

Exceedance

Peak discharge,

Standard

Historically

in data

set

probability
cfs

deviation

of

adjusted

logarithms
standard

of annual

peak

discharges

deviation
-It

2-7

Iqota-cion
sEG

Standard
samples

error
from

of

sample

a normal

skew coefficient,

distribution

Standard

error

estimated

as:

of

sample

for

can be estimated

6N(N 1)
(N - 2)(N + l)(N
SES

which

standard

-I- 3)
deviation,

can be

SES =
Am-

SEX

Standard

Recurrence

Logarithm

of

peak

Mean logarithm

of

XH

High

xL

Low outlier

outlier

error

of

sample

interval

mean,

can be estimated

in years
flow
peak flows

threshold
threshold

in

log

i n log

2-a

units
units

as:

as:

Appendix

'ABLES OF K VALUES
The following

table'

skew coefficients,
bilities,

K values

for

use in equation

G, from 0 to 9.0 and 0 to -9.0

P, from 0.9999
Approximate

formation

contains

values

and exceedance

of K can be obtained

(26) when skew coefficients

from the following

are between

where K, is the

standard

normal

Because of the

limitations

(27)

table

proba-

deviate
invo'lved

trans-

1.0 and -1.0:

use of the

for

to 0.0001.

- ;) ;+ 113- 1

forms,

(1),

and G is the

(3-l 1
skew coefficient.

in use of this

and other

trans-

is preferred.

1
This table was computed by Or, H. Leon Harter and published
in
Technometrics,
Vol. 11, No. 1, Feb. 1969, pp. 177-187, and Vol. 13, No.
'I Feb. 1971, pp. 203-204, "A New Table of Percentage
Points of the
Points of the
PLarson Type I II Distribution"
and "More Percentage
Pearson Distribution,"
respectively.
These publications
describe
values only for positive
coefficient
of skew.
Values for negative
coefficient
of skew were obtained
by inverting
the positive
table and
changing signs.
The latter
work was performed by the Central Technical
Unit,
SCS, Hyattsville,
Md.

3-l

0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9800
0.9750
0.9600
0.9500
0.9000
0.8000
oe7000
0.6000
0.5704
0.5000
0.4296
oc4ooo
0.3000
0.2000
0.1000
0.0500
0.0400
0.0250
0.0200
0.0100
0.0~050
0.0020

0.0010
0.0005
0.0001

G =o.o
-3.71903
-3.29053
-3.09023
-2.S7416
-2.57393
-2.32635
-2.05375
-1.95996
-1.75069
-1 .h4485
-1.28155
-0.84262
-0.52440
-0.25335
-0.17733
0.0
0.17733
0.25335
0.52440
O.A4162
1.28155
I.64485
1 e75039
1.05996
2.05375
?.32635
2.57583
2.87815
3.09023
3.29053
3.71902

G =O.l
-3.50703
-3.12767
-2.94834
-2.75706
-2.48187
-2.25258
-1.99973
-1.91219
-1.71580
-1.61594
-1.37037
-0.84611
-0.53624
-0a261382
-0.19339
-0.01662
0.16111
0.33763
0.51207
0.83639
lo29178
1.67279
1.78462
2.00688
2.10637
2.33361
2.66965
2a9YY78
3.33322
3.45513
3.93453

G =0.2
-3.29921
-2.96695
-2.80786
-2.63672
-2.38795
-2.17840
-1.94499
-1.86360
-1.67999
-1.58607
-1.25824
-0.84986
-0.54757
-0928403
-0.20925
-0eO3325
0,14472
0.22168
0.49927
0.83044
1.30105
1.69971
1.81756
2.05290
2,15935
2.47226
2,76321
3.12169
3.37703
3.62113
4.15301

G =0.3
-3.09631
-2.80889
-7.66915
-2.51741
-2.29423
-2.10394
-1.88959
-1.81427
-1.64329
-1.55527
-1.24516
-0.85285
-0.55839
-0e29897
-0e22492
-0eo4993
0.12820
0.20552
0.48600
0.82377
1.30936
1.72562
1.84949
2.09795
2.21081
I 2.54421
2.85636
3,24371
3a52139
3.78820
4.37394

G =0.4
-2.89907
-2.65390
-2.53261
-2.39942
-2.20092
-2.02933
-1.83361
-1.76427
-1.60574
-1.52357

G =0.5
-2.70836
-2.50257
-2.39867
-2.28311
-2.10825
-1.95472
-1.77716
-1.71366
-1.56740

-1.23114

-1.49101
-1.21618

-0.85508
-0.56867
-0.31362
-0.24037
-0.06651
0.11154
0.18916
0.47228
0.81638
1.31671
1.75048
1.88039
2.14202
2.26133
2.61539
2.94900
3.36566
3.66608
3.95605
4.59687

-0.85653
-0.57840
-0.32796
-0.25558
-0,08302
0.09478
0.17261
Oe45812
0.80829
1.32309
1.77428
1.91022
2.18505
2.31084
2.68572
3.04102
3.48737
3.81090
4.12443
4.82141

G =0.6
-2.52507
-2.35549
-2.26780
-2.16884
-2.01644
-1.88029
-1.72033
-1.66253
-1.52830
-1.45762
-1.20028
-0.85718
-0.58757
-0.34198
-0.27041
-0.09945
0.07791
0.15589
O-44352
0.79950
1.32850

1,797Ol
1.93896
2.22702
2.35931
2.75514
3.13232
3.60872
3.95567
4.29311
5.04718

Y
W

0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9800
0.9750
0.9600
O"9500
0.9000
0.8000
Oe7000
0.6000
0.5704
0.5000

0.4296
0.4000
0.3000
0.2000
OelOOO
0.0500
0.0400
0.0250
0.0200
0.0100
0.0050
0.0020
0.0010
0.0005
0.0001

G =3.7

-2.35015
-2.21323
-3.14053
-2.05701
-l.g2S8C!
-1.80621
-1.66325
-1.61099
-1.48R52
-1.42345
-1.18.347
-0.85703
-0.59615
-0.35565
-0,285ll;
-0,11579
O.06097
oe13901
0.42851
0.7~002
1.33;P9*

l.PlS64
1.96660
2.26790
2a.40670
2.8235Q
3.22281
3,72957
4.10022
4.46189
S.27389

G =0.8

-2.1'344i!
-2.07661
-2.01739
-1.948d6
-1.R366C)
-1.73271
-1.60604
-1.55914
-1.44813
-1.39855
-1.16574
-0.85607
-0.60412
-0.36889
-0e29961
-0e13199
0.04397
0.12199
oe41309
0.779d6
1.33640
I.93916
1.99311
2.30764

2.45298
2oR3101
3.31243
3.84981
4.24439
4.63057
5.50124

G =0.9

G =l.O

-2.02891
-1.94611
-1.Q9894
-1.84244
-1.74919
-1.66001
-1.54886
-1.50712
-1.40720
-1.35299
-1.14712
-0.85426
-0.61146
-0.35186
-0.31368
-0.14807
0.02693
O.lOadS
0.39723
0.76902
1.33883
1.Y5856
2.01848
2.34623
2.49811
2.95735
3.40109
3.96932
4.38RO7
4.79899
5.72899

-1.88410
-1.82241
-1.78572
-1.74062
-1.66390
-1.58538
-1.49188
-1.45507
-1.36584
-1.316t34
-1.12762
-0.85161
-0.61815
-0.39434
-0.32740
-0.16397
0.00987
0.08763
0.38111
0.75752
1834039
1.87663
2.04269
2.38364
2.54206
3.02256
3.48874
4.0(3(302
4.53112
4.96701
5.95691

G =l.l

-1.75053
-1.70603
-1.67825
-1.64305
-1.58110
-1.51808
-1.43529
-1.40314
-1.32414
-1c28019
-1.10726
-0.84809
-0e62415
-0.40638
-0.34075
-0.17968
-0eoo719
0.07032
0.36458
oe74537
1.34092
l.l39395
2.06573
2.41984
2.58480
3.08660
3.57530
4.20582
4.67344
5.13449
6.18480

G =1.2

-1.62838
-1.59738
-1.57695
-1.55016
-1.50114
-1.44942
-1.37929
-1m35153
-1.28225
-1.24313
-1.08608
-0.84369
-0.62944
-0.41794
-0.35370
-0.19517
-0.02421
0.05297
0.34772
!'%z
1:90992
2.08758
2045482
2.62631
3.14944
3.66073
4.32263
4.81492
5.30130
6.41249

G =1.3
-1.51752
-1.49673
-1.48216
-1.46232
-1.42439
-1.38267
-1.32412
-1.30042
-1.24028
-1.20578
-1.06413
-0.83841
-0.63400
-0.42899
-0.36620
-0.21040
-0.04116
0.03560
0.33054
0*71915
1033904
1.92472
2.10823
2.48855
2.66657
3.21103
3.74497
4.43839
4.95549
5.46735
6.63980

P
0.9999

Ti"
8

0.9995
0.9990
0.9980
0.9950
0.9900
0.9800
0.9750
0.9600
0.9500
0~9000
0.8000
O.?OOO
0.6000
0.5704
0*5000
Oe4296
0*4000
0.3000
0.2000
0I1000
0.0500
0*0400
0.0250
0.0200
0.0100
0.0050
0.0020
0.0010
0.0005
0.0001

G =1.4
-1.41753
-1.40413
-1.39408
-1.37981
-1.35114
-1.31915
-1.26999
-1.25004
-1.19842
-1.16827
-1.04144
-0.83223
-0.63779
-0a.43949
-0.37824
-0.22535
-0.05803
0.01824
0.31307
0.70512
1.3366G
1.93836
2.1276R
2.52102
2.70556
3.27134
3.82798
4.55304
5.09505
5.63252
6.86661

G =l.S
-1.32774
-1.31944
-1.31275
-1.30279
-1.28167
-1.25611
-1.21716
-1.20059
-1.15682
-1.13075
-1.01810
-0.82516
-0.64080
-0.44942
-0.38977
-0.23996
-0.07476
0.00092
0.29535
0.69050
1.33330
1.95083
-2.14591
2.55222
2.74325
3.33035
3.90973
4.66651
5.23353
5.79673
7.03277

G =I.6

G =l.?

-1.24728
-1.24235
-1.23805
-1.23132
-1.21618
-1.19680
-1.16584
-1.15229
-1.11566
-1.09338
-0.99418
-0.81720
-0.64300
-0.45R73
-0.40075
-0.25422
-0.09132
-0.01631
0.27740
0.67532
1.32900
1.96213
2.16293
2.58214
2.77964
3.39804
3.99016
4.77875
5.37087
5.95990
7.31818

-1.17520
-1.17240
-1.16974
-1.16534
-1.15477
-1.14042
-1.11628
-1.10537
-1.07513
-1.05631
-0.96977
-0.80837
-0.64436
-0.46739
-0.41116
-0.26808
-0.10769
-0.03344
0.25925
0.65959
1.32376
1.97227
2.17873
2.61076
2.81472
3.44438:.
4.06926
4.88971
5.50701
6.12196
7.54272

G =1.8
-1.11054
-1.10901
-1.10743
-1.10465
-1.09749
-1.08711
-1.06864
-1.06001
-1.03543
-1.01973
-0.94496
-0.79868
-0.64488
-0.47538
-0.42095
-0.28150
-0.12381
-0.05040
0.24094
0.64335
1.31760
1.98124
2.19332
2.63810
2.84848
3.49935
4.14700
4.99937
5.64190
6.28285
7.76632

G =1.9
-1.05239
-1.05159
-1.05068
-1.04898
-1.04427
-1.03695
-1.02311
-1.01640
-0.99672
-0.98381
-0.91988
-0.78816
-0.64453
-0.48265
-0.43008
-0.29443
-0.13964
-0.06718
0.22250
0.62662
1.31054
1.98906
2.20670
2.66413
2.88091
3.55295
4.22336
5.10768
5.77549
6.44251
7.98888

G =2.0
-0.99990
-0.99950
-0.99900
-0.99800
-0.99499
-0.98995
-0.97980
-0.97468
-0.95918
-0.94871
-0.8.9464
-0.77686
-0.64333
-0.48917
-0.43854
-0.30685
-0.15516
-0.08371
0.20397
0.60944
1.30259
1.99573
2.21808
2.68808
2.91202
3.60517
4.29832
5.21461
5.90776
6.60090
8.21034

i'

P
0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9800
0.9750
0.9600
0.9500
0.9000
0.8000
0.7000
0.6000
0.5704
0.5000
0.4296
0.4000
0.3000
0.2000
0.1000
0.0500
0.0400
0.0250
0.0200
0.0100
0.0050
0.0020
0.0010
0.0005

0.0001

G =2.5
-0.71429
-0.71429
-0.7142a

-0.71426
-0.71425
-0.71415
-0.71377
-0.7134A
-0.71227
-0.71116
-0.70209
-0.66603
-0.60434
-0.51.274
-0.47588
-0.35353
-0.262@2
-0.20255,
6.05746
0.45990
1.21013
7.00992
2.27470
2.54134
3.1139Q
3.97301
4.84669
6.01558
6,91505
7.81339
9.93643

G =2.9
-0.68966
-0.68966
-0.h9965
-0.68955
-0.6ir964
-0.6r3959
-0.h893S
-0.68917
-0.68834
-1).4e759
-0.68075
-0.6508s
-0.53634
-0.51212
-0.4t3037
-0.35991
-0.27372
-0.31523
0.03997
0.44015
1.19539
2.00710
2.?7676
2.85492
3.13356
4.01286
4.90884
6.11254
7.03.443
7.96431
lo.14602

G =3.0
-0.66667
-0.64447
-0.64667
-!I.64667
-0.646b6
-0.66663
-0.66644
-0.66638
-0.66585
-0.66532
-0.66023
-0.h35b9
-0.58783
-0.51073
-0.4i3109
-0.39554
-0.23395
-0.22726
0.02279
0.42040
1.1~006
2.00335
2.27780
2.86735
3.15193
4.0513t)
4.96959
6.20506
7.15235
8.10836
10.35418

6 =3.1
-0.64516
-0.64516
-0.44516
-0.6451b
-0.64514
-0.64514
-0.64507
-0.44500
-0.64465
-0.64429
-0.64056
-0.62060
-0.57887
-0.501363
-0.48107
-0.40041
-0.29351
-0.23868
0.00596
0.40061
1.16416
1.99863
2.27785
2.87865
3.16911
4.08859
5.02897
6.29613
7.26881
8.25115
10.54090

G =3.2
-0.62500
-0.62500
-0.62500
-0.62500
-0.62500
-0.62499
-0.42495
-0.62491
-0.62469
-0.62445
-0.62175
-0.60567
-0.56953
-0.50585
-0.48033
-0.40454
-0.30238
-0.24946
-0.01050
0.38081
1.14772
1.99314
2.27493
2.R8884
3.18512
4.12452
5.08697
6.38578
7.38382
8.39248
10.76618

G =3.3
-0.60606
-0.60606
-0.60606
-0.60606
-0.60606
-0.60606
-0.60603
-0.60601
-0.60587
-0.60572
-0.60379
-0.59094
-0.55989
-0.50244
-0.47890
-0.40792
-0.31055
-0.25958
-0.02654
0.3b104
1.13078
1.98474
2.27506
2.89795
3.20000
4.15917
5.14342
6.47401
7.49739
8.53236
10.97001

G =3.4
-0.58824
-0.58824
-0.58024
-0.58824
-0.58824
-0.58823
-0.58822
-0.58821
-0.58812
-0.58802
-0.58466
-0.57652
-0.55000
-0.49044
-0.47682
-0.41058
-0.31002
-0.26904
-0.04215
0.34133
1.11337
1.97951
2.27229
2.90599
3.21375
4.19257
5.19892
6.56084
7.60953
8.67079
11.17239

0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9800
0.9750
3.9600
3.9500

Y
cn

0.9000
0.8000
0.7000
0.6000
0.5704
0bS000
Oe4296
0.4000
0.3000
0.2000
0.1000
0*0500

OeO400
3.0250
0.0200
0.0100
0.0050
B.0020
ooool4
0.0005
0*0001

G =2.1
-0.95234
-0.95215
-0.9518Q
-0.95131
-0.94945
-0.94607
-0.93978
-0.93495
-0.92295
-0.91458
-0.86938
-0.76482
-0.64125
-0.49494

-0.44628
-0.31872
-0.17030
-0.09997
0.18540
0.59183
1.29377
2.00125
2.22986
2.71234
2.94181
3.65600
4.37186
5.32014
6.03865
6.75798
8.43064

G =2.2
-0.90908
-0.90899
-0.90885
-0.90854
-0.90742
-0.90521
-0.90009
-0.89728
-0.88814
-0.08156
-0.94422
-0.75211
-0.63833
-0.49991
-0.45329
-0.32999
-0.18504
-0.11590
0.16682
0.57383
1.26412
2.00570
2.23967
2.73451
2.97028
3.70543
4.44398
5.42426
6.16816
6.91370
8.64971

G =2.3
-0.86956
-0.86952
-0.86945
-0.86929
-0.R6863
-0.86723
-0.86371
-0.86169
-0.H5486
-0.04976
-0.t31929
-0.73880
-0.63456
-0.50409
-0.45953
-0.34063
-0.19933
-0.13148
0.14827
0.55549
1.27365
2.00903
2.24831
2.75541
2.99744
3.75347
4.51467
5.52694
6.29626
7.0bR04
8.86753

G =2.4
-0.83333
-0.83331
-0.83328
-0.83320
-0.83283
-0.83196
-0.82959
-0.82817
-0.52315
-0.81927
-0.79472
-0.72495
-0.62999
-0.50744
-0.46499
-0.35062
-0.21313
-0.14665
0.12979
0.53683
1.26240
2.01128
2.25581
2.77506
3.02330
3.80013
4.58393
5.62818
6.42292
7.22098
9.08403

G =2.5
-0.80000
-0.79999
-0.79998
-0.79994
-0.79973
-0.79921
-0.79765
-0.79667
-0.79306

-0.79015
-0.77062
-0.71067
-0.62463
-0.50999
-0.46966
-0.35992
-0.22642
-0.16138
0.11143
0.51789
1.25039
2.01247
2.26217
2.79345
3.04787
3.84540
4.65176
5.72796
ii.54814
7.37250
9.29920

G =2.6
-0.76923

-0.76923
-,0.76922
-0.76920
-0.76909
-0.76878
-0.76779
-0.76712
-0.76456
-0.76242
-0.74709
-0.69602
-0.61854
-0.51171
-0.47353
-0.36852
-0.23915
-0.17564
0.09323
0.49872
1.23766
2.01263
2.26743
2.81062
3.07116
3.88930
4.71815
5.82629
6.67191
7.52258
9.51301

G =2.7
-0.74074
-0.74074
-0.74074
-0.74073
-0.74067
-0.74049
-0.73987
-0.73943
-0.73765
-0.73610
-0.72422
-0.68111
-0.61176
-0.51263
-0.47660
-0.37640
-0.25129
-0.18939
0.07523
0.47934
1.22422
2.011177
2.27160
2.92658
3.09320
3.93183
4.78313
5.92316
6.79421
7.67121
9.72543

P
0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9800
0.9750
0.9600
0.9500
0.9000

0.8000
0.7000
0.6000
005704
Od5000
004296
0.4000
0.3000
0,200o
001000
0.0500
0.0400
0.0250
oeo200
0.0100
I.0050
5.0020
O.OOl@
0.0005
0.0001

G =3.s
-0.57143
-0.57143
-0.57143
on.57143
-0.57143
-0.57143
-0.5714?
-0.57141
-0.57136
-0.57130
-il.57035
-0.56242
-0.53993
-0.49391
-0.47413
-0.41253
-0.32479

-.O*27782
-0.05730
r),.32171
1.09s52
T.97147
2.26862
2.911290
3.22641
4.22473
5.25291
6.64627
7.72020
RI80779
11.37334

G =3.h
-0.55s54
-0.55556
-0.55556
-0.55556
-0.55556
-0.55556
-0.55555
-0.55555
-0.55552
-0.55548
-0.53453
-0.548h7
-0.52975
-0.48888
-0.47088
-0.41381
-0.33035
-0.23592
-0.fI71YS
0.30223
1.07726
1.95266
2.26409

2.91898
3.23800
4.25569
5.30559
6.73032
7.82954
8.94335
11.57284

G =3.7
-0.54054
-0.54054
-0.54054
-0.54054
-0.54054
-0.54054
-0.54054
-0.54054
-0.54052

-0.54050
-0.54006
-0.53533
-0.51952
-0.48342
-0.46711
-0.41442
-0.33623
-0.29335
-0.08610
0.2!?t290
1.058h3
1.95311
2.25872
2.92397
3.24853
4.23545
5.35698
6.51301
7.93744
9.07750
11.77092

G =3.9

-0.52632
-0.52632
-0.52632
-0.52632
-0.52632
-0.52632
-0.52631
-0.52631
-0.52630
-0.52629
-0.52600
-0.52240
-0.50929
-0047758
-0.46286
-0.41441
-0.34092
-0.30010
-0.09972
0.2637h
I.03965
l-94283
2.25254
2092799
3.25803
4.31403
5.40711
6.89435
8.0439"s
9.21023
11.96757

G =3.9

-0.51282
-0.51282
-0.51282
-0.51282
-0.51282
-0.51282
-0.51282
-0.51282
-0.51281
-0.51281
-0.51261
-0.50990
-0.49911
-0.47141
-0.45819
-0.41381
-0.34494
-0.30617
-0.11279
0.24484
1.02036
1.93186
2.24558
2.93107
3.26653
4.34147
5.45598
6.97435
8.14910
9.3415%
12016280

G =4.0

-0.50000
-0.50000
-0.50000
-0.50000
-0.50000
-0.50000
-0.50000
-0.50000
-0.50000
-0.49999
-0.49986
-0.49784
-0.48902
-0.46496
-0.45314
-0.41265
-0.34831
-0*31159
-0.12530
0.22617
1.00079
1.92023
2.23786
2.93324
3.27404
4.36777
5.50362
7.05304
8.25289
9.47154
12.35663

G =4.1
-0.48780
-0.48780
-0.48780
-0.48780
-0.48780
-0.48780
-0.48780
-0.48780
-0.48780
-0.48780
-0.48772
-0.48622
-0.47906
-0m45828
-0.44777
-0e41097
-0.35105
-0.31635
-0.13725
0.20777
Oe98096
1.90796
2.22940
2.93450
3.28060
4.39296
5.55005
7*13043
8.35534
9.60013
12.54906

$-

P
0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9800
0.9750
0.9600

Y
03

0.9500
0.9000
0.8000
0.7000
Oe6000
0.5704
0.5000
0.4296
0.4000
0.3000
0.2000
0.1000
0.0500

0.0400
0.0250
0.0200
0.0100
0.0050
0.0020
0.0010
0.0005

0.0001

G =4.2
-0047619
-0.47619
-0.47619
-0.47619
-0.47619
-0.47619
-0.47619
-0.47619
-0.47619
-0.17619
-0.47614
-0*47504
-0.46927
-0.45142
-0.44212
-0.40881
-0.35318
-0.32049
-0m14861
0.18967
0.96090
1.89508
2.22024
2.93489
3.28622
4.41706
5;59525
7.20654
8.45646
9.72737
12.74010

G =4.3
-0.46512
-0.46512
-0.46512
-0.46512
-0.46512
-0.46512
-0.46512
-0.46512
-0.46512
-0.46511
-0.46508
-0.46428
-0.45967
-0.44442
-0.43623
-0.40621
-0.35473
-0.32400
-0.15939
0.17189
0.94064
1.98160
2021039
2093443
3.29092
4.44009
5.63934
7.28138
8.55627
9.85326
12.92977

G =4.4
-0.45455
-0.45455
-0.45455
-0.45455
-0.45455
-0145455
-0.45455
-0m45455
-0.45455
-0.45454
-0.45452
-0.45395
-0,45029
-0.43734
-0043016
-0.40321
-0e35572
-0.32693
-0al6958
0.15445
0.92022
1.86757
2.19988
2.93314
3.29473
4.46207
5068224
7.35497
8.65479
9.97784
13.11808

G =4.5
-0.44444
-0.44444
-0.44444
-0.44444
-0.44444
-0.44444
-0.44444
-0c44444
-0.44444
-0.44444
-0.44443
-0.44402
-0.44114
-0.43020
-0.42394
-0.39985
-0.35619
-0e3292a
-0.17918
0.13737
Ocai39964
1.85300
2.18874
2.93105
3.29767
4.48303
5.72400
7.42733
8.75202
10.10110
13.30504

G =4.6
-0.43478
-0.43478
-0.43478
-0.43478
-0.43478
-0.43478
-0.43478
-0o43478
-0.43478
-0.43478
-0.43477
-0.43448
-0.43223
-0.42304
-0.41761
-0.39617
-0.35616
-0.33108
-0.18819
0.12067
0.87895
1.83792
2.17699
2.92818
3.29976
4.50297
1 5.76464
7.49047
a.84800
10.22307
13.49066

G =4.7

G =4.8

-0.42553
-0.42553
-0.42553
-0.42553
-0.42553
-0.42553
-0.42553
-0~42553
-0.42553
-0.42553
-Or42553
-0.42532
-0.42357
-0.41590

-0.41667
-0-41667
-0.41667
-0841667
-0.41667
-0041667
-0e4i667
-0.41667
-0.41667
-0,41667
-0.41666
-0.41652
-or841517

-0.41121
-0.39221

-0.40477
-0.38800
-0.35475
-0.33315
-0e20446
0.08847
0.83731
1.80631
2.15174
2.92017
3.30149
4.53990
5.84265
7.63718
9.03623
10.46318
13.85794

-0.35567
-0033236
-0.19661
0.10436
Oe85817
1.82234
2.16465
2.92455
3.30103
4.52192
5.80418
7.56842
8.94273
10.34375
13.67495

-0.40880

P
0.9999

0.9995
0.9990
0.9900
0.9950
0*9900
0.9000
0.9750
Oa9600
0.9500
0.9000
0.0000
0.7000

0.6000
0.5704
0.5000
0.4296
0.4000
0.3000
0.2000
0.1000
0..0500
0.0400
0.0250
0.0200
0.0100
0.0050
0.0020
0.0010
0.0005
0.0003

G =4.9
-0.40816

-0.40816
-0.40816
-0.40816
-0.40816
-0.40816
-0.40A16
-0.40816
-0.40816
-0.40816
-0.40816
-0.40806
-0.40703
-0.40177
-0.39833
-0038359
-0.35343
-0.33347
-0.21172
O,O7300
0.81641
1.78982
2e13829
2891508
3930116
4.55694
5.88004
7e70479
9.12852
10.58135
14.03963

G =S.O

-0.40000
-0.40000
-0.40000
-0.40000
-0.40000
-0.40000
-0.40000
-0.40000
-0.40000
-0.40000
-0.40000
-0.39993
-0.39914
-0.39482
-0e39190
-0.37901
-0-35174
-0.33336
-0.21843
Oc.95798
0.79548
1.77292
2.12432
2.90930
3.30007
4.57304
5.91639
7.77124
9.21961
10.69829
14.22004

G =S.l
-0.39216

-0.39216
-0.39216
-0.39216
-0.39216
-0.39216
-0.39216
-0.39216
-0.39216
-0.39216
-0.39216
-0.39211
-0.39152
-0.3879Y
-0.38552
-0.37428
-0.34972
-0.33284
-0.22458
0.04340
0.77455
1.75563
2.10985
2.90283
3,29R23
4.58823
5.95171
7.83657
9.30952
10.81401
14e39918

G =5.2
-0.38462
-0.38462
-0.38462
-0.38462
-0.38462
-0.38462
-0.38462
-0.38462
-0.38462
-0.38462
-0.38462
-0.38458
-0.38414
-0.38127
-0.37919
-0.36945
-0.34740
-0.33194
-0.2301Y0.02927
0.7S364
1.73795
2.09490
2.89572

3.29567
4.60252
5.98602
7.90078
9.39827
10.92853
14.57706

G =5.3
-0.37736
-0.37736
-0.37736
-0.37736
-0.37736
-0.37736
-0.37736
-0-37736
-0.37736
-0.37736
-0.37736
-0.37734
-0.37701
-0.37469
-0.37295
-0.36453
-0.34481
-0.33070
-0.23527
OeO1561
0.73277
1.71992
2.07950
2.08796
3e29240
4.61594
6.01934
7.96390
9.48586
11.04186
14.75370

G =5.4
-0.37037
-0.37037
-0.37037
-0.37037
-0.37037
-0.37037
-0.37037
-0.37037
-0.37037
-0.37037
-0,37037
-0.37036
-0.37011
-0.36825
-0,36680
-0.35956
-0.34198
-0.32914
-0.23984
0.00243
0.71195
1.70155
2.06365
2.87959
3.28844
4,62850
h.OS169
5.02594
9.57232
11015402
14.92912

G =5.5
-0.36364
-0.36364
-0.36364
-0.36364
-0.36364
-0,36364
-0.36364
-0.36364
-0.36364
-0.36364
-0.36364
-0.36363
-0.36345
-0.36196
-0.36076
-0.35456
-0,33895
-0.32729
-0.24391
-0.01028
0.69122
1.68287
2.04739
2.87062
3.28381
4.64022
6,08307
8.08691
9.65766
11026502
15.10332

P
0.9999
009995
oa9990

0.9980
0.9950
0.9900
0.9800
0.9750
0.9600
0.9500
0.9000
0.8000
0.7000
0.6000
0.5704
0.5000
0.4296
0.4000
0.3000
0.2000
0.1000
0.0500

0.0400
3.0250
0.0200
0.0100
0.0050
0.0020

0.0010
0.0005
0.0001

G =3.6

-0.35714
-0.35714
-0.35714
-0035714
-0.35714
-0.35714
-0.35714
-0.35711
-0.35714
-0.35714
-0.35714
-0.35714
-0.35700
-0.35583
-0.35484
-0.34955
-0.33573
-0.32519
-0.24751
-0.02252
0.67058
1.66390
2.03073
2.86107
3.27a54
4.65111
6*11351
9.14hH3
9674190
11.34487
15.27632

5 =s.7
-0.35Obd
-0.35088
-0e35088
-0.35088
-0.3SOY8
-0.35088
-0.35088
-0.3soat3
-0.3SOHH
-0.35088
-0.35088
-0.35087
-0.35078
-0.3491if;
-0.34903
-0e34455
-O.33236
-0.32285
-0.?5064
-0.03427
0.65006
1.64464
2.01369
2.85096
3.27263
4.66120
6.14302
8.20572
9,R2505
P1.48360

15.44813

G =!T.cs
-0.34483
-0.34483
-0*344c33
-0.34483
-0.34483
-0634463
-0.34483
-0.344a3
-0.34453
-0.34483
-0*344n3
-0e344t33
-0.34476
-0.34402
-0.34336
-0.33957
-0.328336
-0.32031
-0e25334
-0.04553
0.62966
I.62513
I*99629
2.84030
3.26610
4.67050
6.17162
8.26359
9.YO713
.I1.59122
15.4187d

G =5.3
-0,3389d

-0.33898
-0.3389t3
-0.33898
-0.33838
-0.33898
-0.33898
-0.33898
-0,3389H
-0.33aYu
-0.33898
-oo33a98
-0,33RY3
-0.33836
-0,337t32
-0.33463
-0e32525
-0e31753
-0.25562
-0.05632
0.60941
1.60538
I.37855
2.82912
3.25898
4.67903
6.19933
R.32046
9.98915
11069773
15.78826

G =6.0
-0.33333
-0.33333
-0.33333
-0.33333
-0.33333
-0.33333
-0.33333
-0.33333
-0.33333
-0.33333
-0.33333
-0.33333
-0.33330
-0.33285
-0.33242
-0.32974
-0.32155
-0.31472
-0.25750
-0.06662
0.58933
1.58541
1.96048
2.81743
3.25128
4.66680
6822616
8.37634
10.06812
11.80316
15.9S660

G =b.l
-0.32787
-0.32787
-0.32787
-0e32787
-0,32787
-0.32787
-0.32787
-0.32787
-0.32787
-0.32787
-0.32787
-0.32787
-0.32784
-0.32750
-0.32715
-0.32492
-0.31780
-0.3Pl71
-0.2!i901
-0.07645
0.56942
1.56524
1,94210
2.80525
3.24301
4.69382
6.25212
8.43125
10.14706
11.90752
16.12380

G =6.2

-0,32258
-0.32258
-0.32258
-0.32258
-0.32258
-0.32258
-0.3225%
-0.32258
-0.32258
-0.32258
-0.32258
-0.32258
-0832256
-0.32230
-0.32202
-0.32016
-0.31399
-0.30859
-0.26015
-0.08580
0.54970
1.54487
1.92343
2.79259
3.?3419
4.70013
6.27723
6.48519
10.22499
12.010b2
16.28939

S-4
d

0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9880
0.9750
0.9600
0.9500
0.9000
0.8000
0.7000
0.6000
0.5704

0.5000
0.4296

0.4000
003000
O*ZQOO
0,1000
0.0500
0.0400
0.0250
0.0200
0.0100
0.0050
0.0020
0.0010
0.0005
0.0001

G =6.3
-0.31746
-0.31746
-0.31746
-0.31746
-0.31746
-0.31745

-0.31746
-0.31746
-0.31746
-0.31746
-0.31746
-0.31746
woe31745
-0.31724
-0c.31702
-0.31549
-0.31016
-0.30538
-0.26097
-0.09469
0.53019
1.52434
1.90449
2.77947
3.22481
4.70571
6.30151
9.53820
10.30192
12.11307
16.45487

G =6.4

G =6.5

-0.31250
-0.31250
-0.31250
-0.31250
-0.31250
-0.31250
-0.31250
-0.31250
-0.31250
-0.31250
-0-31250
-0.31250
-0.31249
-0.31234
-0,31216
-0*31030
-0.30631
-0.30209
-0.?6146
-cr,l0311
0,51089
le50365
1.88528
2,76591
3,21497
4m71061
6.32497
8,59027
lOe37785'
12.21429
16.61875

-0.30769
-0.30769
-0.30769
-0.30769
-0.30769
-0,30769
-0.30769
-0.30769
-0.30769
-0.30769
-0.30769
-0.30769
-0.30769
-0e30757
-0.30743
-0.30639
-0.30246
-0.29575
-0,2hfe7
-0.11107
Oe49182
1.482Bl
1,86584
2.75191
3.20460
4.71482
6.347.62
Be64142
lo,45251
12.31450
16.78156

G =6.6

-0.30303
-0.30303
-0.30303
-0.30303
-0.30303
-0.30303
-0.30303
-0.30303
-0.30303
-0.30303
-0.30303
-0.30303
-0.30303
-0.30294
-0.30253
-0.30198
-0.29862
-0.29537
-0.26160
-0,118SY
0.472YY
1.46186
1.84616
2.73751
3.19374
4.71836
6.36948
8.69167
10.52681
12.41370
16.94329

G =6.7
-0.29851
-0.29851
-0.29851
-0.29051
-0.29851
-0.29851
-0.29851
-0.29851
-0.29851
-0.29851
-0.29851
-0.29851
-0.29850
-0.29844
-0.29835
-0.2.9766
-0.29400
-0.29196
-0.2612d
-0.12566
0.45440

1.44079
1.82627
2.72270
3.18241
4.72125
6.39055
R.74102
10.59986
12.51190
17.10397

G =6.8

-0.29412
-0.29412
-0.29412
-0.29412
-0.29412
-0.29412
-0.29412
-0.29412
-0.29412
-0.29412
-0.29412
-0.29412
-0.29412
-0.29407
-0.29400
-0.29344
-0o29101
-0.28854
-0,26072
-0.13331
0,43608
1.41963
1.80618
2.70751
3.17062
4,72350
6e41036
8.78950
10.67197
32.60913
17.26361

G =6.9
-0120986
-0.28986
-0.28986
-0.28986
-0.28986
-0.28986
-0.29986
-0.28986
-0.28986
-0.28906
-0.28986
-0.28986
-0.28985
-0.28982
-0.28977
-0.28931
-0.28726
-0.28511
-0-25995
-0.13853
O-41803
1.39839
1,78591
2.69195
3.15838
4-72512
6.43042
8.H3711
10.74316
12.70539
17,42221

P
0.9999
0.9995
0.9990
0.9980

Y
z

. ..

0.9950
0.9900
0.9800
0.9750
0.9600
0.9500
009000
0.8000
0.7000
0.6000
0.5704
0*5000
0.4296
0.4000
0.3000
0.2000
0.1000
0.0500
0.0400
0.0250
0.0200
0.0100
0.0050
0.0020
0.0010
0.0005
0.0001

G =?.O

-0.28571
-0.28571
-0.28571
-0.28571
-0.28571
-0.28571
-0.28571
-0.28S71
-0.28571
-0.28571
-0.29571
-0.28571
-0.28571
-0.28569
-0.28565
-0.28528
-0.28355
-0.28169
-0.25899
-0.14434
0.40026
lo37708
1.76547
2.67603
3.14572
4.72613
6.44924
8.88387
10.81343
12.80069
17.57979

G =7.1

-0.28169
-0.28169
-0.28169
-0.28169
-0.28169
-0.28169
-0.28169
-0.28169
-0.38169
-0.28169
-0.28169
-0.28169
-0.28169
-0.28167
-0.28164
-0.28135
-0.27390
-0.27829
-0.25785
-0.14975
0.38277
1.35571
1.74487
2.65977
3.13263
4.72653
6.46733
8.92979
PO.88281
PZ.R3SO5
17.73636

G =7.2
-0.27778
-0.27778
-0.27778
-0.27778
-0.27778
-0.27778
-0.27778
-0.27778
-0.27778
-0.27778
-0.27778
-0.27778
-0.27778
-0.27776
-0.27774
-0.27751
-0.27629
-0.27491
-0.25654
-0.15478
0.36557
1.33430
1.72412
2.64317
3. 11914
4.72635
6.48470
8.9740%
10.95129
12.98848
17.59193

G =7.3
-0.27397
-0.27397
-0.27393
-0.27397
-0.27397
-0.27397
-0.27397
-0.27397
-0.27397
-0.27397
-0.27397
-0.27397
-0.27397
-0.27396
-0.27394
-0.27376
-0.27274
-0.27156
-0.25510
-0.15942
0.34868
1.31257
1.70325
2.62626
3.10525
4.72559
6.50137
9.01915
11.01890
13.08095
lR.04652

G =7.4
-0.27027
-0.27027
-0.27027
-0.27027
-0.27027
-0.27027
-0.27027
-0.27027
-0.27027
-0.27027
-0.270'27
-0.27027
-0.27027
-0.27026
-0.27025
-0.27010
-0.26926
-0.26825
-0.25352
-0.16371
0.33209
1.29141
1.68225
2.60905

3.09099
4.72427
6.51735
9.06261
11.08565
13.17258
18.20013

G =7.5
-0.26667
-0.26667
-0.26667
-0.26667
-0.26667
-0.26667
-0.26667
-0.26667
-0.26667
-0.26667
-0.26667
-0.26667
-0.26667
-0.26666
-0.26665
-0.26654
-0.26584
-0.26497
-0.25183
-0.16764
0.31582
1.26995
1.6b115
2.59154
3.07636
4.72240
6.53264
9.10528
11.15154
13.26328
18.35278

G =7.6
-0.26316
-0.26316
-0.26316
-0.26316
-0.26314
-0.26316
-0.26314
-0.26316
-0.26316
-0.26316
-0.26316
-0.26316
-0.26316
-0.26315
-0.26315
-0.26306
-0.26248

-0.26175
-0.25005
-0.17123
0.29986
1.24850
1.63995
2.57375
3.06137
4.71998
6.54727
9.14717
11.21658
13.35309
18.50447

P
0.9999
0.9995
0.9990
0.9900

0.9950
0.9900
0.9800
o.9750
0.9600
0.9500
0.9000
0.8000
0.7000
Oe6000
0.5704
0.5000
0.4296
0.4000
OS3000
0.2000
0.1000
0.0500
0.0400
0.0250
0.0200
0.0100
0.0050
0.0020
0.0010
0.0005
0.0001

G =7.7
-0.25974
-0.25974
-0.25976
-0.25974
-0.25974
-0.25974
-0.25974
-0.25974
-0.25974
-0.25974
-0.25974
-0.25974
-0.25974
-0.25974
-0.25973
-0.25966
-0.25919
-0.25657
-0.24817
-O.P745r!
0.213422
1.22706
1.61867
2.55569
3.04604
4.71704
6.56124
9.1882R
11.28080
13.44203
15.65522

G =7.R
-0.25641
-0.25641
-0.25641
-0.25641
-0.25641
-0.25641
-0.25641
-0.2564'1
-0.25641
-0.25641
-0.25641
-0.25641
-0.25641
-0.25641
-0.25640
-0.25635
-0.25596
-0.35544
-0.24622
-0.17746
0.26992

1.2OSb5
1.59732
2.53737
3.03038
4.71359
_ 6.57456
9.22863
11.34419
13.53009
lY.80504

G =7.9
-0.25316
-0.25316
-0.25316
-0.25316
-0.25316
-0.25316
-0.25316
-0.25316
-0.25316
-0.25316
-0.25316
-0.25316
-0.25316
-0.25316
-0.25316
-0.25312
-0.25200
-0.25236
-0.2442lL
-0.lRO12
0.25394
1.19427
1.57591
2.51581
3.01433
4.70961
6.58725
9.26823
11.40677
13.61730
18.95393

G =R.O
-0.25000
-0.25000
-0.25000
-0.25000
-0.25000
-0.25000
-0.25000
-0.25000
-0.25000
-0.25000
-0.25000
-0.25000
-0.25000
-0.25000
-0.25000
-0.24996
-0.24970
-0.24933
-0.24214
-r).lP249
0.23929
2.16295
1.55444
2.50001
2.99RlO
4.70514
6.59931
9.30709
11.46855
13.70366
19.10191

G =8.1
-0.24691
-0.24691
-0.24691
-0.24691
-0.24691
-0.24691
-0.24691
-0.24691
-0.24691
-0.24691
-0.24691
-0.24691
-0.24691
-0.24691
-0.24691
-0.24689
-0.24667
-0.24637
-0.24003
-0.18459
0.22498
1.14168
1.53294
2.48099
2.98150
4.70019
6.61075
9.34521
11.52953
13.78919
19.24898

G =8.2
-0.24390
-0.24390
-0.24390
-0.24390
-0.24390
-0.24390
-0.24390
-0.24390
-0.24390
-0.24390
-0.24390
-0.24390
-0.24390
-0.24390
-0.24390
-0.24308
-0.24371
-0.24345
-0.23783
-0.1a3643
0.21101
1.12048
1.51141
2.46175
2.96462
4.69476
6.62159
9.38262
11.58974
13.87389
19.39517

G =8.3
-0.24096
-0.24096
-0.24096
-0.24096
-0.24096
-0.24096
-0.24096
-0.24096
-0.24096
-0.24096
-0.24096
-0.24096
-0.24096
-0.24096
-0.24096
-0.24095
-0.2408P
-0.24060
-0.23571
-0.18803
0*19737
1.09936
1.48985
2.44231
2.94746
4.68887
6.63153
9*41931
11.64917
13.95778
19.S4046

P
0.9999
0.9995
0.9990

0.9980
0*9950
0.9900

0.9800
0.9750
0.9600
0.9500
0.9000

0.8000
0.7000
0.6000
085704
0.5000
0.4296
0.4000
0*3000
0.2000
0.1000
0.0500
0.0400
0.025rJ
0.0200

O.OlOC
0~0050
0.0020

0.0010
0.0005
0.0001

=8.4

-0.23910
-0.23810
-0.23810
-0.23810
-0.23!310
-0.23RlO
-0.23810
-0.23910
-0.23310
-0.23810
-0.23@10
-0.23810
-0.23810
-0.23810
-0.239OY
-0.238OR
-0.23797
-0.23779
-0.23352
-0*18939
0.18408
1.07@32
1.46329
2.42268
2.93002
4.68253

6.64148
9.45530
11.70785
14.04086
19.68489

G =R.S
-0.23529
-0.23529
-0.23529
-0.33529
-0.23529
-0.23529
-0.23529
-0.23529
-0.23529
-0.23529
-0.23529
-0.23529
-0.23529
-0.23tJ29
-0.2d523
-0.23S28
-0.23520
-0*?3505
-0.23132
-0.19054
0.17113
1.05738
1.44673
2.40287
2.91234
4.67573
6.65056
9.a30so
11.76576
14.12314
13.RZd45

G =9.6
-0.23256
-0.23256
-0.23256
-0.23256
-0.23256
-0.23256
-0.23256
-0.23256
-0.23256
-0.23256
-0.23256
-0.23256
-0.23256
-0.23256
-0.23256
-0.232SS
-0.23248
-0.23236
-0.2291%
-0.19147
0.15851
1.03654
1.42518
2.38288
2.89440
4.66850
6.65907
9.52521
11.82294
14.20463
19.Y7115

G CR.7
-0.22989
-0.22989
-0.22989
-0.22989
-0.229iiY
-0.22989
-0.229BY
-0.22959
-0.22989
-0.22983
-0.22989
-0.22989
-0.22989
-0.22988
-63).22988
-0.22988

-0.229a2
-0.22972
-0.22690
-0.19221
0.14624

1.01581
1.40364
2.36273
2.87622
4.66085
6.66703
9.55915
11.87938
14.2s534
20.113OO

G =8.8
-0.22727
-0.22727
-0.22727
-0.22727
-0.22727
-0.22727
-0.22727
-0.22727
-0.22727
-0.22727
-0.22727
-0.22727
-0.22727
-0.22727
-0.22727
-0.22727
-0.22722
-0.22714
-0.22469
-0.19277
0.13431
0.99519
1.38213
2.34242

2.85782
4.65277
6.67443
9.59243
11.93509
14.36528
20.25402

-6 =8.9
-0.22472
-0.22472
-0.22472
-0.22472
-0.22472
-0.22472
-0.22472
-0.22472
-0.22472
-0.22472
-0.2,?472
-0.22472
-0.2L472
-0.22+72
-0.22472

-0.22472
-0.2246d
-0.22461
-0.22249
-0.19316
0.12272
0.97471
1.36065
2.32'197
2.83919
4.64429
6.6td130
9.62504
11.99009
14.44446
30.3Y420

G =9.0
-0.22222
-0.22222
-0.22222
-0.22222
-d.22222
-0.22222
-0.22222
-0.22222
-0.22222
-0.22222
-0.22222
-0.22222
-0.22222
-0.22222
-0.22222
-0.22222
-0.22219
-0.22214
-0.22030
-0.19339

0.11146
0.95435
1.33922
2.30138
2.@2035
4.63541
6.68763
9.65701
12.04437
14.522pR
20.53356

0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9800
oe9750
0.9600
0.9soo
3.9000
0.8000
0.7000
0.6000
0.5704
0.5000
Oe4296
!I!.4000
0.3000
0.2000
0.1000
0.0500
0.0400
0.0250
0.0200
0.0100
0.0050
0*0020
0.0010
0.0005
0.0001

G =-i).n

-3.71302
-3.29r353
-3.03023
-2.P7414
-2.575H3
-3.3263q
-2.05375
-1.95996
-1.75069
-1.61485
-1.28155
-0.ir4163
-n,s244r!

-0.2s335
-n.17733
000
0.17733
oezs335
(r.52440
Ooi\4163

1*2P%SS
I,64495

I.75069
1.959Qh

?,05375
?.3263%
2.57583
i?.H7Ylk
3.09023
3.?9051
3.7~0!3~

G =-0.1

G =-0.2

G =-0.3

G =-0.4

-3.93453
-3.45513
-3.?3322
-2.9997ri
-2.h69bs
-2.39331
-2.13697
-.?.@ObH.Q
-1.75462
-1.67279
-1.29178
-0.R.3639
-0.51207
-O.?37b3
-@.lhlll
0.01662
0.19339
O.Zbisti2
0.53624
0.44611
1.?7037
1.61594
1.71580
1.91219
1.99573
2.2525R
?.lYlt)7
?.757Ob
2.94934
3.12767
3.st1703.

-4.l53dl
-3.62113
-3.37783
-3.12163
-2.74321
-2.47226
-2.15935
-2. r)s230
-1.$1756
-1.69971
-1.30105
-0.83044
-0.49027
-0.22165
-0.14472
0.03325
0,20925
Oe28403
0,54757
O,R49d6

-4.37394
-3.78P20
-3.52139
-3.24371
-2.85636
-2.54422
-2.21081
-?.0979S
-1.34944,
-1.72562
-1.30936
-0.82377
-0.48600
-0.2oss2
-0.12820
0.04393
0.22492
0.29897
c1.55933
0.85285
1.24516
%,SSS27
1.64329
1081427
1.889S9
2.10394
2.29423
2.51741
2.66915
2.80R8Y
3.09631

-4.59687
-3.95605
-3.66608
-3.36566
-2.94900
-2.61539
-2.26133
-2.14202
-1.88039
-1.75048
-1.31671
-0.81638
-0.47228
-0.18916
-0.lllS4
0.06651
0.24037
0.31362
0.56867
oea5508
1.23214
1852357
1.60574
1.76427
1.83361
2.02933
2.20092
2.39942
2.53261
2.65390
2.89907

Io%Si324

1*5~607
1.67939
1.H6360
1.94499
2.17840
2.3;r735
2.63672
2e50786
2e96638
3.3Y921

G =-0.5

G =-0.6

-4.82141
-4.12443
-3.81090
-3.48737
-3.04102
-2.68572
-2.31084
-2.18505
-1.91022
-1.77428
-1.32309
-0.80829
-0.45812
-0.17261
-0.09478
0.08302
0.2SS58
0.32796
0.57840
0.85653
1.21618
le49101
1.56740
1.71366
1.77716
1.95472
2.10825
2.28311
2.39867
2.50257
2.70836

-5.04718
-4.29311
-3.95567
-3.60872
-3.13232
-2.75514
-2.35931
-2.22702
-1.93896
-1.79701
-1.32950
-0.79950
-0.44352
-0.15589
-0.07791
0.0994s
0.27047
0.34198
O,S8757
0.85718
lm2002R
1.45762
1.52830
1.66253
1.72033
1.88029
2,01644
2.16884
2.26760
2.35549
2.52507

oe9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9aoo
0.9750
0.9600
0.9500
0.9000
008000
0.7000
0.6000
0.5704
0.5000
0.4-296
0.4000

0.3000
0.2000
0.1000
0.0500
0.0400
0.0250
0.0200
0.0100
0.0050
0.0020
0.0010
0.0005
0.0001

61 z-0.7
-5.27389
-4.46189
-4.10022
-3.72957
-3.22251
-2.82359
-2.40570
-2.26790
-1.966bO
-1.RlR64
-1.33294
-0.79ooil
-0.42852
-0.13901
-fI.0609?
O.ll57R
n.28516
0.3556cl
n.59615
O.R5703
1.19347
1.42345

1.4!!$52
1.61099
1.66325
1.80621
1.92590
3.05701
2.14053
2.21328
2.35015

1; 1 = -0.8

-5.50124
-4.63057
-4.34439
-3,!?49t)l
-3.31243
-2.a9101
-2.45298
-2.30764
-1.93311
-1.93916
-1.33640
-0.77986
-0.41309
-0.12199
-0.04397
0.13199
0.29961
0.3sHd9
O.hO412
0.85607
l.lh574

1.3Qti55
1.44913
1.55914
1.60604
1.73271
1.R3660
1.94804
2.01739
2.07661
'%2.1.544%

lsl =-0.9

-5.72899
-4.79899

-4.38807
-3.96932
-3.40109
-2.9S735
-2.49811

-2.34623
-2.0184s
-1.BCiRb6
-1.33889
-0.76902
-0*39729
-0.10486
-0.02693
0.14807
0.32365
0.3HlHh
0.61144

O.t+5426
I.14712
1.35299
1.40720
1.50712
1.54R56
l.66001
1.74919
1.?!4244
1.89R94

1.94611
2.O~SSl

Gl =-1.0
-5.95691
-4.96701
-4.53112
-4.08802
-3.48874

-3.02256
-2.54206
-2.3R364
-2.04269
-1.r37683
-1.34039
-0.75752
-0.3ail1
-0.08763
-0.009a7

0.16397
0;32740
0.394.34
11.61915
fl.85161
1.12752
1.31hH4
1836584
1.45507r
1.49188
h588\3a
1.66390
1.74062
-1.78572
1.82241
1.88410

Gl =-1.1
-6.18480
-5.13449
-4.67344
-4.20582
-3.57530
-3.08660
-2.58480
-2.41984
-2.06573
-1.a9395
-1.34092
-0.74537
-0.36458
-0.07032
0.00719
0.17968
0.34075
0,4063S
0.e2425

0.84809
1.10726
1.2#019
1.32414
' 1.+0314*
1.43529
I... 5 1 a 0 a

1.58110
1.64305
1.67825
1.70603
1.75053:

Gil =-1.2
-6.41249
-5.30130
-4.81492

-4.32263
-3.66073
-3.14944
-2.62631
-2.45482
-2.08758
-1.90992
-1.34047
-0.73257
-0.34772
-0.05297
0.02421
0.19517
0.35370
0.41794
0.62944
O.H43b9

1.08608
1.24313
1.28225
1.35153
1.37929
1.44942
1.50114
1.55OP6
1.57695
1.59738
1.62838

Gl =-1.3
-6.63980
-5.46735
-4.95549
-4.43839
-3.74497
-3.21103
-2.66657
-2.48855
-2.10823
-1.92472
-1.33904
-0.71315
-0.33054
-0.03560
0.04116
0.21040
0.36620
0.42899
0.63400
o.r3s41
1.05413

i.2u57a
1.2402H
1.30042
1.32412
I.38267
1.42439
1.46232
1.48216
1.49673
1.51'752

Y
d
-4

0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9800
0.9150
0.9600
0.9500
0.9000
0.8000
0.7000
0.6000
0.5704
0.5000
0.4296
0.4000
0.3000
002000
0*1000
0~0500
0.0400
0.0250
0.0200
0.0100
0*0050
0.0020
0.0010
0.0005
0.0001

G z-1.6

r, =-1.5

-6.86661
-5.63252
-5.09505
-4.55304
-3.82799
-3.27134
-2.70556
-2.52102
-2.12768
-1.93836
-1.33665
-0.70512
-0.31307
-0.01824
0.05803
0.22535
0.37524
0.43949
0.63779
0.83223

-7.09277
-5.79673
-5.23353
-4.66651
-3.90973
-3.33035
-2.74325
-2.55222
-2.14591
-1.95083
-1.33330
-0.69050
-0.29535
-0.00092
0807476
0.23996
0.38977
O-44942
Qeh4080
0oR25lh
l.01810
1.13075
l.156t32
1.20059
1.21716
1.25611
1.28167
1830279
1.31275
le31944
1.32774

1.04144
1.16827

1.19842
1.25004
1.26999
I.31815
le35l14
1.37981
1.39408
1.40413
1*4l753

G =-1.6

-7.31918
-5.95990
-5.37057
-4.77875
-3.99016
-3.38804
-2.77964
-2.58214
-2.16293
-1.96213
-1.32900
-0.67532
-0.27740
0.01631
0609132
0.25422
0*4007s
0.45873
Oe64300
0*8P7t?O
0.99418
1.09338
l.ll566
1.15229
1.16584
1.19680
1.21618
1.23132
1.23AOS
1.24235
1.24728

G =-1.7

G =-La

G s-l.9

G =-2.0

-7.54272
-6.12196
-5.50701
-4.8R971
-4.06926
-3.44438
-2.81472
-2.61076
-2.17873
-1.97227
-1.32376
-0.65959
-0.25925
0.03344
0.10769
0.26808
0.41116
0046739
0.64436
O.dO837
0.96977
1.05631
1.07513
1.10537

-7.76632
-6.28285
-5.64190
-4.99937
-4.14700
-3.49935
-2.84848
-2.63810
-2.19332
-1.98124
-1.31760
-0.64335
-0.24094
0.05040
0.12381
0.28150
0.42095
0.47538
0.64488
0.79868
0.94496
I.01973
1.03543
1.06001
1.06864
1.08711
1.09749
1.10.465
1.10743
1.10901

-7.98888
-6.44251
d5.77549
-5.10768
-4.2233'6
-3.55295
-2.88091

-8.21034
-6.60090
-5.90776
-5.21461
-4.29832
-3.60517
-2.91202
-2.68888
-2.21888
-1.99573
-1.30259
-0.60944
-0.20397
0.08371
0.15516
0.30685
0.43854
0.48917
0.64333
0.77686
Oe89464
0.94871
0.95918
0.9746a
0.97980
0.98995
0.99499
0.99800
0099900
0099950
0.99990

1.11628
1.14042
1.15477
: 1.16534

1.16974
lel7240
1.17520

l.llOS4

I$:;;$-;;
-1.98906
-1.31054
-0.62662
-0.22250
0.06718
0.13964
0.29443
Oe43008
0.48265
0.64453
0.7a916
0.91988
0.98381
0.99672
1.01640
1.02311
1.03695
1.04427
1.04898
1.05068
l.OSl59
1.05239

P
0.9999
0.9995
0.9990
0.9900
0.9950
0.9900
0.9000
009750
0.9600
0.9500
0.9000
Oe8000
0.7000
Oa6000
0.5704
oe5000
0.4296
0.4000
0.3000
0.2000
0.1000
0.0500
0.0400
0.0250
.Y.0.0200

0.0100
0.0050
0.0020

G =-2.1
-8.43064

-8.64971

-6.7579R
-6.03865
-5.32014
-4.37186
-3.65600
-2.94181
-2.71234

-6.91370
-6.16616

-2.22986
-2.00128
-1.29377
-0.59383
-0.18540
0809997
0~17030
0.31872

0.44628
0.49494
0.64125
0176482
0.86938
Oe91458
0892295
0.93495

0.93878
0.94607
Oa94945
0.95131

0.0010

0.95baR

0.0005

0.95215
0.95234

0,0001

li =-2.2

-5.4242b

-4044338
-3.70543

-2.97028
-2.73451
-2.23967
-2.00570
-1.2d412
-0.57383

-0mlb682
r)*11590
0.18504
Oa32939
0.45329
0.49991
0.63833
0.7SZPl
0.94422
0.8el56
0,88814
0.89728
0,9c)oo9
0.90521
0.90742
0.908S4
0.90835
0.90399
0.90308

G =-2.3
-d.a67S3
-7.06804

-4.29626
-5.52694
-4.51467
-3.75347
-2.99744
-2.75541
-2.24831
-2.00903
-i.i?-?3b5

-0.55549
-0.14827
0*13148
0*19933

0.34063
0 es5953
0.50409
0.63456
0.73880
Oed1929
0.84976
0.?~5486

0.86169
0,863710.36723

U.d6863
0.66929
0.86945
O.H69S2
Il.&6956

G =-2.4

G =-2.5

6 =-2.6

-9.08403

-9.29920

-9.51301

-7.22098
-6.42292

-7.37250
-6.54814
-5.72796
-4.65176
-3.84540
-3.04787
-2.79345
-2.26217
-2.01247
-1.25039
-0.51789
-0.1%143
0.16136
0.22642
0.35992

-7.52258
-6.67191

-5.b2818
-4.58393
-3.80013
-3.02330
-2.77506
-2.25581
-2.01128
-1.26240
-0053683
-0.12979
0.14665
0.21313

0.35062
0.46499
0.50744
0.62999
0.72495
0.79472

0.81927
0.82315
0.82817
0.82959
0.83196
0.83283
0.83320

0.83328
0.83331
0.83333

0.46966
0.50999
0.62463
0.71067
0.77062

-5.82629
-4.71815
-3.88930
-3.07116
-2.81062
-2.26743
-2.01263
-1.23766
-0.49872
-0.09323
0.17564
0.23915
0.36852
0.47353
oe51171
0.61854
0.69602
0.74709

0.79015

0.7b242

0079306
0.796b7
0.79765
0.79921
0*79973
0.79994
0.79998
0.79999
0.80000

0.76456
0.76712
0.76779
0.76878
0.76909
0.76920
0.76922
0.76923
0.76923

Gl=-2.7
-9.72543
-7.67121
-6.79421
-5.92316
-4.78313

-3.93183
-3.09320

-2.82658
-2.27160
-2.01177
-1.22422
-0.47934
-0.07523
0.18939
0.25129

Oe37640
Oe47660
0.51263
0.61176

0.68111
Oe72422
0.73610
0.73765
0.73943
0.73987
0.74049
0.74067
0.74073
0.74074
0.74074
0.74074

G =-2.4

=-2.9

0.9999
0.9995

-9.93643
-7.81039

0.9990
0.9980
0.9950
0.9900
0.9800

-6.91505
-6.0185f?

-6.11254

-4.84669
-3.97301

-4.90884
-4.01286

-3.11399
-2.54134

-3.13356
-2,-ss492
-2.27676
-2.00710
-1.19539
-0.44015

0.9750
0.9600
0.9500
0.9000
0.8000
O.?QOO
0.6000
0.5704
0 .SQOO
0.4296
0.4OQO
0*3000
0.2800
0.1000
0.0500
0.0400
0.0250
0.0200

-2.27470
-3.00992

-1.21013
-a.45980
-0.05746
0.2025Q
0.26282
!I*35353
0.47858
0.51276
Q160434
0.66603
0.70209
O.tlll6
O.71Z27
0.71345
0.71377

0.0100

0.91415

0.0050
0.0020
0.0010
0.0005
0.0001

0.71425
0.71428
0.7142Q
0.71423
0.71a29

-10.14602
-7.96411
-7.03443

-0.#3997

0.21523
0.27372

9.38991
0.48037

0.51212
0.59634
0.45086
0.68075
0.68759
0.68836
0.68917

0.6a935
0.68959
0.68964
0.68965
0.48965
0. fit3966

B.6a966

G =-3.0

-10.3541R
-0.10836
-7.15235
-6.20506
-4.96959

-4.05138
-3.15133
-2.86735
-2.27780
-2.00335

-1.18006
-0.42040
-0.02279
0.22726

0.28395
0.39554
0.48109
0.51073
0.55783
0.63569
0.66023
0.66532
0.66585
0.66630
0.66649
0.66663
0.66666
0.666at
0.66667
0.66467
0.66667

G =-3.1

-10.56090
-8.25115
-7.26881
-6.29613

G =-3.2
-10.76618

-8.39248
-7.38382
-6.38578
-5..08697

G =-3.3

-10.97001

-11.17239

-8.53236
-7.49739

-8.67079

-3.16911

-4.12452
-3.18512

-6.47401
-5.14362
-4.15917
-3.20000

-2.87865
-2.27785

-2.88884
-2.27693

-2.89795
-2.27506

-1.99869
-1.16416

-1.99314

-0.40061
-0.00596

-0.38081
0.01050
0.24946
0.30238

-1.98674
-1.13078
-0.36104
0.02654
0.25958
0.31055

-5.02897
-4.08859

0.23868
0.29351
0.40041
0.48107
0.50963

-1.14772

0.62060
0.64056
0.64429
0.64465

0.40454
0.48033
0.50585
0.56953
0.60567
0.62175
0.62445
0.62469

0.64500
0.64507

0.62491
0.62495

0.64514
0.64516
0.64516
0.64516
0.64516
0.64516

0.62499
0.62500
0.62500
0.62500
0.62500
0.62500

0.57887

Gl=-3.4

0.40792
0.47890
0.50244

0.55989
0.59096
0.60379
0.60572
0.60587

0.60601
0.60603
0.60606
0.60606
0.60606
0.60606
0.60606
0.60606

-7.609.53
-6.56084
-5.19892
-4.19257
-3.21375
-2.90599
-2.27229

-1.97951
-1.11337
-0.34133
0.04215
0.26904
0.31802
0.41058
0.47682

0.49844
0.55000
0.57652

0.58666
0.58802
0.58812
0.58821
0.58822
0.58823
0.58824
0.58824
0.58824
0.58824
0.58824

0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9800
0.9750
0.9600
0.9500
0.9000
0.8000
0.7000
0.6000
0.5704
065000
Oe4296
084600
om3000
0.2000
0.1000
O.OSOO
0.0400
0.0250
0.0200
0.010-0
0.005O
0.0020
0.0010
0.0005
O.OOOl

G =-3.5

G =-3.6

G =-3.7

G =-3.8

G =-3.9

-11.37334
-8.80779
-7.72024
-6.64427
-5.25291
-4.22473
-3.22641
-2.91299
-2.26.062
-1.97147
-1.09552
-0.32171
0.05730
0.27782
c1.32479
0.41253
0147413
0.49491
0.53993
0.56242
0.57035
0.57130
Oe57136
0.57141
0.57142
0*57143
0.57143
0.57143
0.57143
3.57143
0.57143

-11.57284
-8.94335
-7.82954
-6.73032
-5.30559
-4.25569
-3.23000
-2.91098
-2.264O9
-1.96266
-1.07726
-0.30223
0.07195
0.28592
0.33005
0.41381
0.47000
0.~888R
(9.52975
0.54867
0.55403
0.5S548
0.55552
0.55555
0.55555
0.55556
0.55556
Oe55556
0.55556
0.55556
Oe55556

-11.77092
-9.07750
-7.93744
-6.81301
-5.3569d
-4.28545
-3.24853
-2.92397
-2.25872
-1.95311
-1.05863
-0.28290
0.05610
0.29335
0.33623
0.41442
0.4b71%
0.48342
0.51952
0.53533

-11.96757
-9.21023
-0.04395
-6.89435
-5.40711
-4.31403
-3.25803
-2.92799
-2.25254
-1.94203
-1.03965
-0026376
0.09972
0.30010
0.34092
0.41441
0.46206
0.47750
0.50929
0.52240
0.52600
0.52629
0052630
0.52631
0.52631
0.52632
0.52632
0.52632
0.52632
0.52632
0.52632

-12.16280
-9.34158
-8.14910
-6.97435
-5.45598
-4.34147
-3.26653
-2.93107
-2.24550
-1.93186
-1.02036
-0.24484
0.11279
0.30617
0.34494
0.41381
0.450r9
0.471.41
0.49911
0.50990
O.Sl261
0.51281
0.51281
Q.51282
0.51282
0.51282
0.51282
0.51282
0.51282
0.51282
0.51202

0.54006
0.54050

0.54052
0.54054
0.54054
0.54054
0.54054
0.54054
0.54054
0.54054
0.54034

G t-4.0
-12.35663
-9.47154
-0.25209
-7.05304
-5.50362
-4.36777
-3.27404
-2.93324
-2.23706
-1.92023
-1.00079
-0o22617
Oe12530
0.31159
0.334031
O-41265
0.45314
0.46496
0.48902
0.49784
0.49906
0.49999
0850000
0.50000
0.50000
0.50000
0.50000
0,50000
0.50000
0.50000
0.50000

G =-4.1
-12.54906
-9.60013
-8.35534
-7.13043
-5.55005
-4.39296
-3.28060
-2.93450
-2.22940
-1.90796
-0.98096
-0.20777
0.13725
0.31635
0.35105
0.4%097
0.44777
0.45820
0.47906
0.40622
0.48772
0.40780
0.40780
0.48780
0,4070o
0.48700
0.4878O
o.40700
0.48700
0.40700
0.48700

..,

P
0.9499
0.9995
0.9990
0.9980
0.99SO
0.9900
0.9800
0.9750
0.9600
0.9500
0.9000

0.8000
0.7000
0.6000
0.5704
0.5000
004296
oe4ooo
0*3000
0.2000
O.lOOO
0.0500
0.0400
oao250
0.8200
oco1oo
000050
h

OeQO20
0.0010
0.0005
000001

G =-4.2

-12.74010
-9.72737
-8.45646
-7.20654
-5.5952R
-4.41706
-3.28622
-2.93489
-2.22024
-1.89508
-0.96090
-0.18967
0.14861
0.32049
0.35318
Oe40581
0.44212
0845142
0 a46927
0047504
0.47614
0.47619
0.47619
0.47619
0.47619
0.47619
0.47619
0.47419
Oe47619
0.47619
Oe47619

G =-4.3

-12.92977
-9.95326
-8.55627
-7.28138
-5.63934
-4.44009
-3.29092
-2.93443
-2.21039
-1.88160
-0.94064
-0.17189
0.15939
0.32400
0.35473
O.kO621
0.43623
0.44442
0.45967
0.46428
0.46508
0.46511
0.46512
0.46512
0.46512
0.465112
0.46512
0.46512
0.46512
0.46512
0.46512

G =-4.4

-13.111308
-9.97784
-8.65479
-7.35497
-5.63224
-4.46207
-3.29473
-2.93314
-2.19988
-1.86757
-0.92022
-0.15445
0.16958
0.32493
0.35572
0.40321
0.43016
0*43734
0.45029
0.45395
0.45452
0.45454
0.45455
0.45455
0.45455
0.45455
0.45455
0.45455
0.45455
0.45455
0.45455

G =-4.5

-13.30504
-10.10110
-8.75202
-7.42733
-5.72400
-4.48303
-3.29767
-2.93105
-2.18874
-1.a5300
-0.89964
-0.13737
0.17918
0.32928
0.35619
0.39985
0.42394
Oe43020
0.44114
Oe44402
0.44443
0.44444
0.44444
oe44444
0.44444
0.44444
0.44444
0.44444
0.44444
0.44444
0.44444

G =-4.6

-13.49066
-10.22307
-8.84800
-7.49847
-5.76464
-4.50297
-3.29976
-2.92018
-2.17699
-1.83792
-0.87895
-0.12067
0.18819
0.33108
0.35616
0139617
0141761
0.42304
0.43223
0.43448
0.43477
0.43478
0.43478
0.43478
0.43478
0.43478
0.43478
0.43478
0.434is

0.43478
0.43478

G =-4.7

-13.67.495

-10.34375
-0.94273
-7.56042
-5.80418
-4.52192
-3.30103
-2.92455
-2.16465
-1.82234
-0.85817
-0.10436
0.19661
0.33236
0035567
Q.39221
8.41121
0.41590
0042357
0.42532
0.42553
0.42553
0.42553
O-42553
0.42553
0.42553
0.42553
0.42553
Cl,42553
0.42553
0.42553

G =-4.8

-13.85794
-10.46318
-9.03623
-7.63718
-5.84265
-4.53990
-3.30149
-2092017
-2.15.174
-1.80631
-0.83731
-0.00847
0.20446
0.33315
0.35475
0.38800
0.40477
0.40880
0.41517
0.41652
0.41666
0.41667
0.41667
0.41667
0041667
0.41667
0.41667
-0.41667
0.41667
0.41667
0.41667

0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9800
0.9750
0.9600
0.9500
0.9000
0.8000
0.7000
0.6000
0.5704
0*5000
0.4296
0.4000
0.3000
0.200~0

0.1000
0.0500
0.0400
0.0250
0.0200
oeo1oo
0.0050
0.0020
0.0010
0.0005

0.0001

-14.03963
-10.53135
-9.12852
-7.70479
-5.B8004
-4.55634
-3.30115
-2.Ql50~
-2.13229
-1.70982
-0.51441
-0.07300
0.21172
0.33347
0.35343
0.38359
0.39833
O.40177
0.40703
ll.40ROcl
0.4OSlh
0.40316
0.4OSl5
0.40516
0.40'316
0.40816
0.40816
0.40AlC(
0.4Oi3lc,
0.4OSlt;
0.*0914

-14.22004
-10.59t329
-9.21~~1
-7.77134
-5.91639
-4.57304
-3.30007
-2.Q0930
-2.12432
-1.77232
-0.7Y548
-0.05798
0.21w3
0.33336
0.35174
0.37901
0.39190
Oe39482
0.3YY14
0.39993
0.40000
0.40000
0,4(i000
0.40000
0.40000
0.40000
0.4000(!
0.40000
0.60000
0.40000
0.40000

G =-5.1

G =-5.2

G =-5.3

G =-5.4

G =-5.5

-14.39919
-10.81401
-9.30952
-7.53657
-5.95171
-4.59533
-3.29823
-2.902d3
-2.10985
-1.75563
-0.77455
-0.04340
0.22451)
0.33284
0.34972
0.37428
0.38552
0.38799
0.39152
fi.39211
0.39216
0.39216
0.33216
0.39216
0.39216
0.39216..
0.39216
0.39216
0.39216
0.39216
0.39214

-14.57706
-10.92853
-0.39827
-7.90078
-5.98602
-4.60252
-3.29567
-2.89572
-2.09490
-1.73795
-0.75364
-0.02927
0.23019
0.33194
0.34740
0.36945
0.37919
0.38127
0.3841.4
0.30459
0.39462
0.39452
0.38462
0.38462
0.38462
0.38462
0.38462

-14.75370
-11.04186
-9.48586
-7.96390
-6.01934
-4.61594
-3.29240
-2.88796
-2.07950
-1.71932
-0.73277
-0.01561
0.23527
0.33070
0.34481
0.36453
0.37295
0.37469
0.37701
0.37734
0.37736
0.37736
0.37736
0.37736
0.37736
0.37736
0.37736
0.37736
0.37736
0.37736
0.37736

-14.92912
-11.15402
-9.57232
-8.02594
-6.05169
-4.62850
-3.28844
-2.87959
-2.06365
-1.70155
-0.71195
-0.00243
0.23904
0.32914
0.34I98
0.35956
0.36680
0.36825
0.37011
0.37036
0.37037
0.37037
0.37037
0.37037
0.37037
0.37037
0.37037
0.37037
0.37037
0.37037
0.37031

-15.10332
-11.26502
-9.65766
-8.08691
-6.08307
-4.64022
-3.28381
-2.87062
-2.04739
-1.68287
-0.69222
0.01028
0.24391
0.32729
0.33895
0.35456
0.36076
0.36196
0.36345
0.36363
0.36364
0.36364
0.36364
0.36364
0.36364
0.362.64
0.36364
0.36364
0.36364
0.36364
0.36364

Ct.38462

0.38462
0.38462
0.38442

P
0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9800
0.9750
0.9600
0.9500
0.9000
0.8000
007000
0.6000
0.5704
oe5000
0.4296
0.4000
0.3000
0.2000

0.1000
0.0500
0.0400
0.0250
0.0200
0.0100
0.0050
o.lI.020
0.0010
0.0005
0.0001

G z-S.6
-15.27632
-11.37487
-9.74190
-8.14683
-6.11351
-4.65111
-3.27854
-2.86107
-2.03073
-1.66390
-0.6705F?
0.02252
0.24751
0.32519
0033573
0834955
0.35484
0.35583
oe35700
oat35714
0.35714
0.35714
0.35714
0.35714
0.35714
0.35714
oe35714
0.35714
0035714
om35714
0.35714

=-3.7

-15.44813
-11.4S360
-9.82505
-9.20572
-6.14302
-4.6blt?o
-3.27263

-2.85096
-2.01369
-1.6,4+64
-0.65006
0.03427
0.25064
0.32285
0.33236
oe34455
0.34903
0.34985
oe3507l3
0*35087
0.35088
0.35088
0.35088
0.35058
?.35088
0035088
0*3sot3a
0.35t)88
0.75oaa
0.3SU~B
u.3tj08H

G =-!?a.8
-15.61878
-11.59122
-9.90713
-8.26353
-6.17162
-4.,67050
-3.26610
-2.84030
-1.99629
-1.62513
-0.62966
0.04553
0.25334
0.32031
0.32886
0.33957
0.34336
0.34402
0.34476
0.34433
oe344t13
O-34483
0.34483
0,34493
0*34483
0.34483
0.34433
0.34483
0.34433
0,34483
0*344a3

=-5.9

-15.78826
-11.69773
-9.98815
OR.32046
-6.19933
-4.67903
-3.25998
-i?.&291Z
-1.97855
-1.60538
-0.60941
0.05632
0.25562
0.31759
0.32525
0.33463
0.33782
0.33836
0.33893
0.33898
0.33898
0.33898
0.33898
0.33598
0.33898
0.33898
0.33898
0.33898
0.33898
0.33898
0.33896

G =-6.0
-15.95660
-11.80316
-10.06812
-8.37634
-6.22616
-4.68680
-3.25128
-2.81743
-1.96048
-1.58541
-0.58933
0.06662
0.25750
0.31472
0.32155
0.32974
0.33242
0.33285
0.33330
0633333
0.33333
0.33333
0.33333
0.33333
0.33333
0.33333
0.33333
0.33333
0.33333
0.33333
0.33333

G 2-6.1

G =-6.2

-16.12380
-11.90752
-10.14706
-8.43125
-6.25212
-4.69382
-3.24301
-2.80525
-1.94210
-1.56524
-0.56942
0.07645
0.25901
0.31171
0.31780
0032492
Oa32715
0.32750
0.32784
0032787
0.32787
0.32787
0.32787
0.32787
0.32787
0.32787
0.32787
0.32787
0.32787
0.32787
0.32787

-16.28989
-12.01082
-10.22499
-8.48519
-6.27723
-4.70013
-3.23419
-2.79259
-1.92343
-1.54487
-0.54970
0.08580
0.26015
0.30859
0.31399
0.32016
0.32202
0.32230
0.32256
0.32258
0.32258
0.32258
Oe32258
0.32258
0.32258
0.32258
0.322513
0.32258
0.32258
0.32258
0.32258

P
0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9800

0.9750
0.9600
0.9500
0.9000
0.8000
0.7000
0.6000
0.5704
0*5000
004296
0.4000
0.3000
0.2000
0e1000
0.0500
0.0400
0.0250
0.0200

0.0100
0.0050
0.0020
0.0010
0.0005
0.0001

G =+.3

G =-6.4

G =-6.5

-16.45487

-16.61875

-16.713156

-12.11307
-10.30192

-12.21429
-10.37785

-12.31450
-10.45281

-8.53320
-6.30151
-4.70571
-3.22484
-2.77347

-9.59027
-6.32497
-4.71061
-3.21497
-2.76591
-1.R8528
-1.50365

-&.64142
-6.34762
-4.71482

-1.90449
-1.52434
-0.53019
0.09469
0.26097
0.30530
0.31016
oe31549
0.31702
0.31724
oe31745
0.31746
0.31746
0.31746,
0.31746
0.31746
0.31741;
0.31746
0.31746
0.31746

0.31746
0.31746
Oe31746

-0.5108Q
0.10311
0.26146
0.30209
0.30631

0.31040
0,312lh
0.31234
0.31249
0.312s,o
0.31250
0.31250
0.31250
0.31250
0.31250
0.31250
0.31250
0.31250
0.31250
0.31250
0.31250

-3.204bo

-2.75191
-1.86584
-1.48281
-0.49182

0.11107
0.26167
0.29875
0.30246
0.30639
OS30743
0.30757
0.30769
0.30769

Oe30769
Oe.30769
o.jo7b9

0.30769
0.30769
0.30769
0.30769
0.30769
0.30769
0.30769
0.30769

G =-6.6

-16.94329
-12.41370
-10.52681
-8.69167
-6.36948
-4.71836
-3.19374
-2.73751
-1.84616
-1.46186
-0.47299
0.11859
0.26160
0.29537
0.29862
0.30198
0.30283
0030294
0.30303
0.30303
0030303
0.30303
0.30303
0.30303
0.30303
0.30303
0.30303
0.30303
0.30303.
0.30303
0.39303

G =-6.7

-17.10397
-12.51190
-10.59986
-8.74102
-6.39055
-4.72125
-3.18241
-2.72270
-1.82627
-1.44079
-0.45440
0.12566
0.26128
0.29196
0.29480
0.29766
0.29835
0.29844
0.29850
0.29851
0.29851
0.29851
0.29851
0029851
0.29851
0.29851
0.29851
0.29851
Oo29851
0.29851
0.29851

G =-6.8

G =-6.9

-17.26361
-12.60913
-10.67197

-17.42221
-12.70539
-10.74316

-8.78950
-6.41086
-4.72350
-3.17062
-2.70751
-1.80618
-1.41963
-0.43608
0.13231
0.26072
0.28854
0.29101
0.29344
0.29400
0.29407
0.29412
0.29412
0.29412
0.29412
0.29412
0.29412
0.29412
0.29412
0.29412
0.29412
0.29412
0.29412
0.29412

-8.83711
-6.43042
-4.72512
-3i15838
-2.69195
-1.78591

-1.39839
-0.41803
0.13853
0.25995
0.28511
0.28726
0.28931
0.28977
0.28982
0.28985
0.28986
0028986
0.28986
0.28986
0.28986
0.28986
0.28986
0.23986
0.28986
0.28986
0.28986
0.28996

0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
Oa9800
0.9750
019600
0.9500
0.9000
0.8000
0.7000
0.6000
0.5704
0.5000
Oe4296
0.4000
0.3000
0~2000
oe100o
0.0500
000400
0.0250
0.0200
0.0100
0.0050
0.0020
0.0010
0.0005
0.0001

6 =-7.0

G =-7.1

-17.57979
-12.80069
-10.81343
-rj.88387
-6.44924
-4.72613
-3.14572
-2.67603
-1.76547
-1,377OR
-0.40026
0.14434
0.25899
0.28169
0.28355
0.28528
0.28565
0.28569
0.28571
0.28571
0.28571
0028571
0.28S71
0.28571
O-28571
0.28571
0.28571
Op28571
0.28571
0.28571
0.28571

-17.73636
-12.89505
-10.Rt328a.
-8.92979
-6.46733
-4072653
-3.13263
-2.65977
-1.74487
-1.35571
-0.38277
0.14975
0.25785
Oe27829
0.27990
0.28135
0.28164
0.28167
0.28169
Oe28169
0.28169
0,28169
OS28169
Oe28169
0.28169
0.28169
0.28169
0.28169
0.28169
0.28169
0.28r69

G =-7.2
-17.89193
-12.95848
-10.95129
-8.97488
-6.48470
-4.72635
-3.11914
-2.64317
-1.72412
-1c33430
-0.36557
0.15478
0.25454
0,27431
0.27629
0.277Sl
0.27774
0.27774
Oe27778
0.27778
Oe27778
0827778
0.27778
0.27778
0.27778
Oe27778
0.27778
0.27778
Om27778
0,2777t3
0.27778

G =-7.3

G c-7.4

-18.04652
-13.08098
-11.01@90
-9.01915
-6.50137
-4.72559
-3.10525
-2.42426
-1.70325
-1.31287
-0.34R68
0.15942
0.25510
0.27156
0.27274
0.27376
0.27394
0027396
0.27397
0.27397
0027397
O-27397
0.27397
0.27397
0.27397
0.27397
0.27397
0.27347
0.27397
0.27397
0.27397

-18.20013
-13.17258
-11.08565
-9.06261
-6.51735
-4.72427
-3.09099
-2.60905
-1.48225
-1.29141
-0.33209
0.16371
0.25352
0.24825
0.24924
0.27010
0.27025
0.27026
0.27027
0.27027
0.27027
0.27027
Oe27027
0.27027
0.27027
0.27027
0.27027
0.27027
0.27027
0.27027
0.27027

G z-7.5
-18.35278
-13.26328
-11.15154
-9.10528
-6.53264
-4.72240
-3.07636
-2.59154
-1.66.115
-1.26995
-0.31582
0.16764
0.25183
0.24497
0.26584
0.26654
0.26665
0.26466
0.26647
0.26467
0.24667
0 e26667
0.26667
0.26467
0.26667
0.26667
0.26667
Oe26667
0.26667
0.26667
0.26667

G =-7.b
-18.50447
-13.35309
-11.21658
-9.14717
-4.54727
-4.71998
-3.04137
-2.57375
-1.63995
-1-24850
-C.29984
0.17123
0.25005
0.26175
0.26248
0.26306
0.26315
0.26315
Oe26316
0.26316
Oe26316
Oe26316
Om26316
0.26316
0026316
0.26314
0.26316
0.24314
0.26316
OS26316
0.26316

Y
N
Q,

0.9999,
0.9995
0.9990
Oe9980
0.9950
0.9900
0.9800
0.9750
0.9600
0.9500
0.9000
0.8000
0.7000
0.6000
0.5704
0.5000
0.4296
0.4000
0.3000
0.2000
0.1000
0.0500
0.0400
0.0250
0.0200
0.0100
0~0050
0.0020
0.0010
0.0005
0.0001

=-7.7

G =-7.8

G =-7.9

G =-8.0

G =-8.1

-1Q.65522
-13.44202
-ll.ZRi)80
-9.18828
-6.56124
-4.71704
-3.04604
-2.55569
-1.61867
-1.22706
-0.28422
0.17450
0,24817
Om25857
0.25919
Oe25966
0825973
0.25974
Oe25974
Oe25974
r),2Sc)74
0.25974
0.25974
0.25974
0.25974
0*25974
0.25974
Oa25974
Oe25974
0.25974
0.25974

-1B.ROS04
-13.53009
-11.34419
-9.22863
-6.57456
-4.713%
-3.03038
-2.53737
-1.59732
-1.20565
-0.26892
0.17746
0.24622
0.25544
0.25596
0.25635
0.25640
Oe25641
OePS641
0.25641
O.?S641
0.25e41
0,25641
0,25641
0.25641
Oe25641
0.2S641
Oe25641
0.25641
0.2S641
0.25641

-18.95393
-13.61730
-11.40677
-9;26823
-6.58725
-4.70961
-3.01439
-2.51881
-1.57591
-1.18427
-0.25394
0.18012
Oe24421
Oe25236
0.25280
0.25312
0.25316
Oe.25316
0.25316
0.25316
0.25316
0.25316
0.25316
0.25316
0.25316
0.25316
0.25316
0.25316
Oe25316
0.25316
ri 0.25316

-19.10191
-13.70366
-11.46ass
-9.30709
-6.59931
-4.70514
-2.99810
-2.50001
-1.55444
-1.16295
-0.23929
0.18249
0.24214
0.24933
0.24970
0.24996
0.25000
0.25000
0.25000
0.25000
0.25000
0~25000
Oe25000
0.25000
0.25000
0.25000
0.25000
0.25000
0.25000
0.25000
0.25000

-19.24898
-13.78919
-11.52953
-9.34521
-6.61075
-4.70019
-2.98150
-2.48099
-1.53294
-1.14168
-0.22498
0.18459
0.24003
0.24637
0.24667
0 :24689
0.24691
0.24691
0024691
0.24691

0024691
0.24691
0.24691
0.24691
0.24691
0.24691
0.24691
0.24691
0.24691
0.24691
0.24691

fj

c-8.2

-19.39517
-13.87389
-11.58974
-9.38262
-6.62159
-4.69476
-2.96462
-2.46175
-1.51141
-1.12048
-0.21101
0.18643
0.23788
0.24345
0.24371
0.24388
0.24390
0.24390
0.24390
0.24390
0.24390
Oe24390
Oe24390
0.24390
0.24390
0.24390
0.24390
0.24390
0.24390
0.24390
0.24390

G =-8.3
-19.54046
-13.95778
-11.64917
-9.41931
-6.63183
-4.68887
-2.94746
-2e44231
-1.48985
-1.09936
-0.19737
0.18803
0.23571
0.24060
0.24081
0.24095
0.24096
Om24096
0.24096
0.24096
Oa24036
Oe24096
Be24096
0.24096
0.24096
Oe24096
0.24096
0024096
0.24096
0.24096
O-a24096

0.9999
0.9995
0.9990
0.9980
0.9950
0.9900
0.9000
0.9750
0.9600
Q.95QQ
0.9000
Q.0QQQ
Q.i'QQQ
0.6000
0.5704
Qe5QQQ
0.4296
Q.4QQQ
0.3080
0.2000
QelQQQ
Q.Q!SQQ
0.0400
Q.Q25Q
0.0200
Q.QlQQ
Q.QQSQ
0.0020
Q.QQlQ
Q,QQQ5
Q.QQQl

G =-8.4
-19.68489
-14.04056
-11.70785
-9.45530
-6.64148

G =-8.5
-19.82845
-14.12314
-11.76576
-9.49060

-2.93002
-2.42268

-6.65056
-4.67573
-2.91234
-2.40287

-1.46829

-1.44673

-1.07832
-0.184QR

-1.05738
-0.17113
0.19054
0.23132
0.23505
0.23520
0.23526
0.23529
Qe23529
0.23529
0.23529
0.23529
0.23529
0.23529
0.23529
0.23529
0.23529
0.23529
0.23529
Q-23529
0.23529
0.23529

-4.68253

0.18939

0.23352
0.23779
0.23797
0.23809
6.23809
0.23816
0.23818
0.23810
0.23510
Q.23810
Q.23SlO
0.23810
Q.23810
0.23810
0.23810
0.23610
0.23810
O-23310
0.23EiPt.I

G =-8.6
-19.97115
-14.20463
-11.82294
-9.52521
-6.65907
-4.66850
-2.09440
-2.38280

-1.42518
-1.03654
-0.15851
0.19147
0.22911
0.23236
0.23248
0.2325s
0.23256
0.23256
8.23256
0.23256
0.23256
Oa23256
0.23256
0:23256
0.23256
Qe23256
0.23256
0.23256
0.23256
Q,23?56
0.23256

G =-8.7

G =-a,8

G z-c),9

G =-9.0

-20.11300
-14.28534
-11.87938
-9.55915
-6.66703
-4.66085
-2.87622
-2.36273
-1.40364
-1.01581
-0.14624
0.19221
0.22690
0.22972
0.22982
0.22988
0.22988
0022988
0022989

-20.25402
-14.36528
-11.93509
-9.59243
-6.67443
-4.65277
-2.05782
-2.34242
-1.38213
-0.99519
-0.13431
0.19277
0.22469
0.22714
0.22722
0.22727
0022727
8.22727
0.22727
0.22727
0.22927
0.22727
0.22727
0.22727
0.22727
0.22727
0.22727
0.22727
0.22727
0.22727
0.22727

-20.39420
-14.44446
-11.99009
;9.62504
-6.68330
-4.64429
-2.83919
-2.32197
-1.36065
-0.97471
-0.12272
0.19316
0.22249
0.22461
0.22468
0.22472
Qe22472
0.22472
0.22412
0422472
0.22472
0.22472
0.22472
0.22472
0.22472
0.22472
0022472
0.22472
0.22472
0.22472
0.22472

-20.53356
-14.52288
-12.04437
-9.65701
-6.68763
-4.63541
-2.82035
-2.30138
-1.33922
-0.95435
-0.11146
0.19330
0.22030
0.22214
0.22219
0.22222
0.22222
Qe22222
0.22222
0.22222
8.22.222
0022222
0.22222
0.22222
0.22222
0.22222
0.22222
0.22222
0.22222
Q,22222
0.22222

oe229a9

0022989
Qe22989
a)022989
0.22989
Qa22989
0.22989
0.22989
0.22909
0.22989
Qa22909
0022989

Appendix

OUTLIER TEST K VALUES

10 PERCENT SIGNIFICANCE LEVEL K VALUES

The table below contains

one sided 10 percent significance
1evelKN
Tests conducted fo select
the outvalues for a normal distribution
(38).
lier detection
procedures
used in this report alndicate these KIJ values are
applicable
to log-Pearson
Type III
distributions
over the testEd range of
skew values.
Sample
size

KN
value

Sample
size

KN
value

2.036
2,088
2.134
2.175
2.213
2.247
2,279
2.309
2.335
2.361
2.385
2.408
2.429
2.448
2.467
2.486
2.502
2.519
2.534
2.549
2.563
2.577
2.591
2.604
2.616
2.628
2.639
2.650
2.661
2.671
2.682
2.692
2.700
2.7ilO
2.719

45

2.727
2.736
2.744
2.753
2.760
2.768
2.775
2.783
2.790
2.798
2.804
2.811
2.818
2.824
2.831
2.837
2.842
2.849
2.854
2.860
2.866
2.871
2.877
2.883
2.888
2.893
2.897
2.903
2,908
2.912
2.917
2.922
2.927
2.931
2.935

ii
48
49
50
51
z:
54
55
56
57
:;
60
::
63
64
65
66
67
ii
5;
72
33
75
:;
78
79

Sample
size
i:

82
ii2
iii"6
z3
89
ii:
i%
94
95
i;
98
1::
101
102
103
104
105
106
107
108
109
110
111
112
113
114
4-1

"N
value

Sample
size

$j
value

2.940
2,945
2.949
2.953
2.957
2.961
2.966
2.970
2.973
2.977
2.981
2,984
2.989
2.993
2.996
3.000
3.003
3.006
3,011
3.014
3.017
3.021
3.024
3.027
3.030
3.033
3.037
3.040
3.043
3.046
3,049
3.052
3.055
3.058
3.061

115
116
117
118
119
120
121
122
123
124
125
126
127
128
129

3.064
3.067
3.070
3.073
3.075
3,078
3.081
3.083
3.086
3.089
3.092
3.095
3.097
3.100
3.102
3.104
3.107
3.109
3.112
3.114
3.116
3.119
3.122
3.124
3.126
3.129
3,131
3,133
3,135
3.138
3.140
3.142
3.144
3.146
3.148

;::
132
133
134
135
136
137
138
139
140
141
142
143
144
i45
146
147
148
149

Appendix

COi4DITIONAL ~KOLj~UI L.ITY ADJ,USTblEblT

For stations
omission

of peaks

outlier

where

the

below

a gage base,

criterion,

reference

the

(28)

cedures

is

should

only

will

may not

quency

curve
Prior

should
peaks

fit

for

historic
have

step

data

and the

the

for

silould

shocrn on the

been completed,

from

III

The steps

flows,

stations

all

the

any computed

fre-

values.
the

statisapplicable

except
12 (page

conditional

data

gage-base

recomputing

in Appendix

discharge
pro-

and incorporating

chart

record

by this

above

All

pro-

low outliers,

adjustment,

outliers,

the

total

treated

probability

been completed.
in

These

the

of observed

level,

flow

of

in

as the minimum

zero

fur

low

described

distribution,

a plot

by the

and/or

curve.

2% percent

detecting

have

flow,

frequency

statistics

truncation

zero

truncated

adjustment

gage base,

with

is

is defined

conditional

Procedures
above

over

Type

be compared

information

computation

the

a log-Pearson
the

peaks
with

the

level

Because

been reviewed

peaks

to obtain

below

to applying

computed.

tics

are

peaks

should

have

years

be used when not

discharges.

cedure

annual

probability

A truncation

exclude

and no other

of

conditional

recommended

has been truncated.

that

record

probability

the

last

12-3)

should

adjustment

as follows:
1.

exceed

Calculate
the

the

truncation

estimated
level

probability7

by the

that

any annual

peak will

formula:

(5-la)
in which
the

total

included,

N is the

number

number

of years

then

equation

of peaks
of
b-lb

above

record.
should

the
If

truncation

historic

be used rather

5-1

level

information
than

S-la.

and n is
has been
*

H-WL

$Zr
where

H is the

and W the

historic

systematic

record
record

length,

weight

&-lb)

L the

number of peaks

as computed

in Appendix

probabilities,

P, for

truncated
6,

equation

6-l.
2.
Pd, on the

Recompute

the

exceedance

frequency

curve

using

equation

selected

points,

5-2:
(5-2)

P ='i;xPd
This

accounts
3.

are

for
not

Therefore,

it

exceedance

synthetic

of peaks

probabilitfes,

needed

is necessary
log

below

truncation

P, computed

to compute

the

to interpolate
discharge

the

values

by equation

synthetic
either

for

the

level.

sample

5-2

statistics.

graphically

or mathe-

0.01,

and 0.50

0.10,

probabilities.
Since

4.

three

those

to obtain

does not

omission

The exceedance

usually

matically

the

have

the

conditional

probability

known statistics,

statistics
exceedance

will

synthetic

ones will

be determined

probabilities

adjusted
based

determIned

in

frequency
be computed.

on the
step

curve

values

3, using

for
the

These
the
following

equations.
Gs

-2.50

+ 3.12

Log(Q.ol/ QJO)

(5-3)

Lg(Qelo/ Q.50)
s, = Log (Q.ol/Q.50)
K*ol

where
ard

G,,

S,,

deviation,

and Ts

are the

and mean,

- K*50

synthetic

respectively;

(5-4)

logarithmic
Qeol,

5-2

QelQ,

skew coefficient,
and Qe5Q are

standdischarges

with

0.01,

and 0.10,

and 0.50 exceedance

and Keol and I(,50 are Pearson

probabilities
Equation
of +2.5

Type III

probabilities
deviates

of 0.01 and 0.50 respectively,

5-3 is an approximation

respectively;

for

exceedance

and skew coefficient

appropriate

for

use between

Gs.

skew values

and -2.0.

5.
should

The frequency

curve

be compared with

position

should

developed

the observed

from the synthetic

annual

be based upon the total

statistics

peak discharges.

number of years

The plotting

record,

n or H,

as appropriate.
The minimum additional
curve

is the determination

Appendix
using

the conditional

a complete
data,

12 illustrate
analysis

or other

requirement

of the weighted
the basic

steps

probability
might

techniques

to arrive

include
described

skew.

Examples

in computing

adjustment.
two-station
in this

5-3

at a final

Other

3 and 4 of

a frequency
considerations

comparison,
report.

frequency
curve
in

use of rainfall
*

NOTATION

= synthetic

= historic

K.O1'

logarithmic
record

skew coefficient

length

= Pearson type III

!50

exceedance
respectively,

deviate from Appendix 3 for

probabilities
of 0.01 and 0.50
and skew coefficient
6,.

= number

of peaks

truncated

= number

of peaks

above

= total

= exceedance

7-f

= estimated
probability
exceed the truncation
= selected

'd

Q.OV ?10*

number

Q.50

of years

the
of

truncation

level

record

probabilities

points

that an annual
level.

on the

frequency

peak will

curve

= discharges
with exceedance
probabilities
0.01,
0.10,
and 0.50,
respectively
= synthetic
'u
= systematic

logarithmic

= synthetic

logarithmic

record

5-4

standard
weight

from

mean

of

deviation
Appendix

Appendix

HISTORIC DATA
+

Flood

information

used to extend
longer

than

the

that

ing analytical
Historic

record,

from the

assumption

that

systematic

largest

a weight

log-

longer

of 1.0.

a weight

is representative

period

largest

in

The remaining

of (H-Z)/(N+L)

of the

on the
*
remaining

(H-Z)

period.

can be done directly

and the

floods

much

adjusted

are known to be the

are given

ach individual
year's
data using equations
4
Figure 6-l is an example of this procedure
three

be

the follow-

the historically

that

record

longer

The computations

record

period

a historically

"HI' are given

distribution

of the historically

systematic

to a historic

In such a situation,

is used to define

period

their

can often

The number "Z" of events

longer

record

curve.

knowledge

"N" events

2.

events

are used to compute

frequency

the historically

years

in the systematic

of the largest

techniques

of "HI' years.
*

record

that

of the systematic

Pearson Type III

10

outside

by applying
6-1,

6-2a,

in which

in the 77 year

period

6-3a,

there

1897, 1919 and 1927 floods

the weights

to

and 6-4a.

are 44 years

of

are known to be the

1897 to 1973.

If

statistics

have'

been previously
computed for the current
continuous
record,
they can be
+
+
adjusted
to give the equivalent
historically
adjusted
values using equations
6-1,
+

6-2b,
3,

probability
individual

6-3b,

and 6-4b,

The historically
paper
flood

as illustrated
adjusted

through
events

points
should

in Figure

frequency

curve

established
also

6-2.
is sketched

on logarithmic-

by use of equation

be plotted

for

6-5,

comparison,

The

The histor-

ically
adjusted
plotting
positions
for the individual
flood events are
+ computed by use of equation
6-8, in which the historically
adjusted
order
number of each event
putations

Yii" is computed

are illustrated

in Figures

7s shown in Figure 6-3. +

+
4. The following
example
historic

peak adjustment

weighting

with

by this

procedure

only.

the generalized
is appropriate

from equations
6-l

illustrates
It

and 6-2,

6-6 and 6-7.

does not include

skew.

The historically

to use in developing

6-4

in application
the

final
adjusted

The com-

and the completed

the steps

piotting
of the

step

of

skew developed

a generalized

skew.

+X

DEFINITION OF SYMBOLS
= event number when events are ranked in order from greatest
magnitude
to smallest
magnitude.
The event numbers "E" will
range from 1 to
(Z + N).
= logarithmic
magnitude of systematic
peaks excluding
events,
peaks below base, high or low outliers

xz

= logarithmic
magnitude of a historic
that has historic
information

= number of X's

+M

flood

a high

outlier

= mean of X's

ti

= historically

tfi

= historically
to compute

= standard

= historically

= skew coefficient

+z

peak including

zero

adjusted

mean

adjusted
order number of each event for use in formulas
the plotting
position
on probability
paper
+

deviation

of the X"s

adjusted

= historically

standard

deviation

of the X's
adjusted

skew coefficient

= Pearson Type III coordinate

expressed in number of standard
deviations from the mean for a specified
recurrence
interval
or percent
chance

= computed
chance

flood

I%

= plotting

position

flow

for

a selected

= number of historic
information

+z

interval

or percent

in percent

= probability
that any peak will
in step 1, Appendix 5)

*:

recurrence

exceed the truncation

level

(used
46

peaks including

high outliers

that

have historic
+

*l-l

= number of years

+L

= number of low values to be excluded,

such as: number of zeros,
number of incomplete
record years (below measurable
base), and low +
+t
outliers
which have been identified

*
a

*w

in historic

period

= constant
that is characteristic
For Weibull
formula,
a = 0; for
Hazen formula,
a = 0.5
= systematic
record weight
6-2

of a given plotting
position
formula.
Beard formula,
a = 0.3; and for
Jc

EQUATIONS
(6-J)

7=wzx
+xxz

(6-2a)

H-WL

-2

wC(X - ii) + x (Xz - r;l,

(6s3a)

(H-WL-1)
H-WL
(H-WL-1)

WZ(X - i)3

(H-WL-2)

L
F;i =

- iQ3

WNM+2X,

(6-4a)

(6-2b)

H-WL
W (N - 1)s'

-2

tqx,

t WN (M - i)2

+x(X,

- M)2
(6-3b)

(H-WL-1)
H-WL
(H-WL-1)

W(N - 1) (N - 2)s G t 3w (N

(H-WL-2)i

- 1) (M - r;l,s'

N
1
+ WN (M - ii)3 +x(X,

*
Log
+

Q-itKi
G= E; when: 1 2 E 2 2
(j= WE - (W - 1) (Z t 0.5);

fp=

-E;i)3

(6%

-I (6-5)
(6-6)

when:

(Z +I)

m-a

2 E 5 (Z t N+L)

(6-7)
(6-8)

H t 1 - 2a loo

6-3

*
+

Figure
StatIon:

6-1.

HISTOKLCALLY

7-finfir,

Rlc
Record:

WEIGHTED
Sanrlv
iR97,

Uistorlcal
N

River
1919,
neriorl:
= 44;

LOG

PEARSON

at
Rrwrton,
19?7,
1970-1973
11197-1977
z = 7;
11 = 77

TYPE
TN.
177

n.
(47
wars)

III

A.
?fli
vears)

ANNUAL
sqnar~

PEAKS
miles

Figure

6-l.

(Eq.

6-2a)

Solving

HISTORICALLY

WEIGHTED LOG PEARSON-TYPE III

Solving

(Eq.

SX = 162.40155
WCX

= 273.13018

xx2

= 3.09755
= 5.20952

= 1.13705

286.11751

E!

6.34657

= 286.11751177

x2

= 3.71581

s
Solving

(Eq.

xx3

= -0.37648

WCx3

= -0.63317

Xx,3 =

6-3a)

W&i2

12.98733

x x,=

- ANNUAL PEAKS (Continued)

6.34657/(77

- 1)

= 0.28898

T3

=
=

0.08351
0.02413

6-4a)
(77)

g=

(76)

0.70802

(0.07485)
(75)

= 0.0418
-

(0.02413)

0.07485
Solving

(Eq.

N=

77

A=
B =
,sEG

6, Page 131

Solving

-0.33

f 0.08

(0.0418)

= -0.32666

0.94

- 0.26

(0.0418)

10[-0.32666
(Eq.

9.5,

0.92913

- 0.92913[0.88649]]

= ,o[-1.150325]

Page 12)

0.302(0.0418)

t 0.07074(-0.2)
= -0.00409

Gw=
-

.302 + 0.07074

Solving

Solving

= o,07074

(Eq.

(Eq.

6-5)-

Solving

6-6)

(Eq.

6-7)

2=3
For E = 1; iii = E = 1
ForE=2;iiT=E=2
For E = 3; I?= E = 3

iii=

(1.682)

(E)

- (0.682)

Solving

i-5 = (1.682)

(E)

(Es.

For Weibull:

[: ': ;j 2 ;7
For 4< E<47:

6-a
a = 0. FP = (100)

(3.5)

2.387

(%)/(78)

6-5

Figure

6-2.

Results
Big Sandy

of Standard

River

N = number
= standard

at Bruceton,
#3-6065

TN.

the

Current

Continuous

Record

DA - 205 square miles

(44 years)

= 3.69094

deviation

of logarithms

= 0.26721

S3 = 0.01908

G = coefficient

of skewness
Adjustment

(logs)

=-0.18746

to Historically
Historic

;I yz W3/s) i

Year

for

- ANNUAL PEAKS

used = 44

logarithms

S2 = 0.07140

WEIGHTED LOG-PEARSON TYPE III

Computation

of observations

M = mean of
S

HISTORICALLY

Weighted

Peaks

(Z = 3 Years)

Log Yz = x,

77 Years

lxz
I

I, (xz

- K

I
- i;i)2

'
I

1897

25,000

4.39794

10.68213

1 0.46531

1919
1927

I
'

21,000
18,500

1
1

4.32222
4.26717

0.60641
I 0.55136

1 0.36774
1
0.30400

Summation

12.98733
N = 44

1.83990

(Eq.

6-l):

W = (77-3)/44

= 1.68182

Solving

(Eq.

6-2b):

i = (1.68182)

(44)

Solving

(Eq.

6-3b):

(M - $

= -0.02487;

= (1.68182)(43)(0.07140)

T2 = 0.08351
Solving

FpG=

(Eq.

(3.69094)
77

= 0.000619;

+ (12.98733)

(M - f)3

0.16762
0.70802

z = 0.28898

= 3.71581

= -0.0000154

+ (1.68182)(44)(0.000619)
76

+ (1.13705)=

o 08351
1

T3 = 0.02413

6-4b):
(1.68182)

(76)

0.22300

H = 77

Solving

-2

- q3

0.31740

1.13705

z-3

(M - i)2

(xz

(7573 (0.02413)

(3)(1.68182)(43)(-0.02487)(0~07140)

(43)

>

(0.01908)

(-.0.18746)

1:
+ (1.68182)(44)(-0.0000154)

"G = 0.0418

+- (0.70802)]
*

6-6

0
0

EXPLANATION
Histortcal peaks (1897,1919 and 1927) plotted as
largest m 77 year perrod (1897-1973).
Recorded peaks (44 years, 1930-1973)plotted
such that each point represents 1.68 years
in the longer 77 year period.
Points plotted by Weibull Plotting Position formula.
Weighted Log PearsonType III compged curve.

N=44, Z=3, H=77

G = 0.04

5=3+71581

Gw=-0.004

;=0.289

I
1,000
k I
99.99

#
99.9 99.8

BIG SANDY RIVERAT BRUCETON,TN

ANNUAL PEAKS
HISTORICALLYWEIGHTEDLOG
PEARSPNTYPE Ill
FIGURE6-3

._ FIB
5
2

99.5

EXCEEDANCEPROBABILITY,IN PERCENT

0.5

0.2

Appendix

TWO STATION COMPARISON

INTRODUCTION

The procedure
the

logarithmic

the

basis

outlined

of a regression
basis

The first
the

for
long

for

step

the
record.

equations

the

procedure

record

with

by the

b =

use in adjusting

of a short

record

a nearby

long-term

provided

herein

is to correlate

concurrent

The regression

can be computed

with

for

on

record.
were

developed

(29).

of

short

the

recommended

deviation

analysis

and Jacobs

by Matalas

is

mean and standard

The theoretical

flows

herein

observed

and correlation

following

observed
peak

peak

flows

coefficients,

for
respectively,

two equations:

xxlyl -I=xlDl/Nl
XX;

(7-l)

- @Xl 12/Nl

(7-2)

where

the

If
certain

terms

are

defined

the correlation
criteria,

standard

deviation

when the

variance

evaluated

coefficient

then

of that

in the
case

in the
long

1 plus

estimates

record.

Both

statistic

only

procedure

In each discussion,
contained

improved

end of this
defined

can be made.

separately,

and adjustment

at the

of the
reduced.

one adjustment
for

by equation

of these
is

each statistic
considered:

record,

(2)

only

for

case 2 include

an additional

one.

short

record

statistics

part

As each

statistic

is

The criterion
separately.

(1)

entire

short

of the

short

record

all

mean and

can be improved

are discussed

are

The steps

7-2 meets

may be worthwhile.

two cases
long

Appendix.

of those

record
contained
for

CRITERION

The variance

of the
Var(V)

AND ADJUSTMENT PROCEDURE FOR MEAN

adjusted

mean (v)

(1-r')
-l-&T

=
I

Since

is

(Syl)2/Nl

a better

estimate

the

by equation
q

can be determined

variance

of the

true

the

of yl,
mean than

yl

if

the

(7-3)

)J
mean, V will

short-record
term

7-3:

r2 -N

1-r

be

in

1equation

7-3

is

equation

7-4.

positive.
If

Solving

the

correlation

r > l/(N1
then

an adjustment

inequality

critical

logarithmic

coefficient

for

satisfies

r yields
equation

7-4,

(7-4)

mean is worthwhile.

the minimum

values

relationship

- 2)l"

to the

represents

minimum

this

critical

of r for

value

various

mean can be computed

The right

values

using

of r.
of Nl.

equation

7-5a

side

of this

Table

7-l

contains

The adjusted
or 7-5b.

N2

J=Yp-

(7-5a)

N1+N2
y = Yl

Equation
for

7-5b

saves

each short

record

+ b(=K3 - xl)

recomputing

record

station.

station

While

the

(7-5b)

a newF2
that

is

adjusted

estimate

period,

be an improvement

may not

in case 2.
mean (equation
record

period

It

(N3).

over

to compare

to the

variance

Compute

the

the

of
varfance

long

record

correlated
equation

of the mean obtained

is necessary
7-3)

being

mean from

may be an improved
it

at the

the

from
entire

variance

the mean cv,)

station

with

the

long

7-5a

or 7-5b

the

concurrent

short

record

of the

adjusted

for

of the mean Y3

the
using

mean

entire

short

equation

7-6:
Var(Y3)

s 2
(---Y3 )

(7-M

N3
7-2

where

S
is the standard
y3
record site for the

short
smaller

than

final

the

estimate

equation

7-5a

deviation
period

variance

of

the

of
N3.

the
If

of v given

the

of

variance

flows

for

of equation

in equation

7-3,

use the

value

Otherwise,

mean.

logarithms

then

the
7-6

is

use y3 as the

of y computed

in

or 7-5b.

EQUIVALENT YEARS OF RECORDFOR THE MEAN

As illustrated
is inversely
equation
for

in equations

proportional

7-3 it

7-3 and 7-6,

to the record

can be shown that

the adjusted

the variance

length

of the mean

at the site.

the equivalent

years

Using

of record,

N,,

mean is:

N, =
(7-7)

It may be seen from equation

(r=O),

then

N, is less

technique

can actually

r satisfies

equation

Ne

= N, + N2, the
Although

mean, it
years

than

N,.

decrease
7-4.

total

record

of record

for

length

is no correlation
that
years

correlation
at the long

the equivalent

that

years

exceedance

limits

then
site.

of record

probability

and in applying

the correlation
of record unless

(r=l),
record

N, be used as an estimate

the various

of confidence

when there

This indicates
the equivalent

For perfect

N, is actually

is recommended

computation

7-7 that

for

the

of the equivalent
floods

the expected

dn the
probability

adjustment.
CRITERION AND ADJUSTMENT PROCEDUREFOR THE STANDARD DEVIATION
The variance
viation)

of the adjusted

can be determined

variance
by equation

7-3

Sy2 (square
7-8:

of the standard

de-

E(S )$
Var(Sy2)= N1l
j.

1-

(N,+N2-,)2

where A, B, and C are defined

below and the other

at the end of the appendix.

In equation

variance

of S ' (the short-record

Yl
in equation
7-8 is negative,
then

than

following

equation:

Irl

where

>[-B

,,

are defined

2(S

)4/(N,-,)
is the
Yl
If the second term

variance).

Solving

this

dB2:4AC
2A

]112

(7-9)

the variance

(N2+2)(N,-6)(N,-8)
(Nl-3)(N,-5)

7-8,

terms

2N2(N, -4) 2 +12N N (N 1 -4)'

- (N,-3)2
(N,-3)2(N,-2]

8(N,-4)
- -(N1-3)

4(N,-4)
+ ---(Tip

B 6(N2+2)(N,-6)
= (N,-3)(N,-6)

2(N;-N,-14)
*

2(N,+l)

2N2N,-41

(N,-3)2

(Nl-3)2

2N,N,(N,-4)2

(N1-3)

NJ-3

+ 2N2(N, -4) (N, -5)

(N,-3)

2(N,-4)(N,+3)

C =

(7-8)

of S 2 will be less
Y
relationship
for r yields
the

2
of S
Yl'

the variance

(N1-3)2(N,-2)

++I)

WJ2+21

(NJ-3)(Nl-5)

2(N,-4)(N,+l)
(Nl-33

7-4

(2N,+N2-2)
N1-1

+ N,~~2(N,-4)

(N, -3J2(N, -2)

The right

side

value

of

r for

various

they

r.

are

the

Table
values
affect

the

adjusted

inequality

7-l

solutions

in N2 only
If

of

gives

The table

of equation

7-9

table

correlation

values

The adjusted
variance

standard

in equation

7-10

is

This

adjustment

value

of flow

an adjustment

While
estimate

the
of the

period,

record

variance

variance

is
at

current

compare

7-10.

the

(standard

variance

from

(standard

7-9,

' + !!&

the

then

b2 (x

N1+N2

square

in

the

_ -,,)2](740)
2

root

brackets

of the

adjusted

in equation

estimate

random
equation

deviation)

of Sy2.

noise

to each estimated

7-10

may be an improved

obtained

over

be an improvement
deviation)

of the

The variations

site.

variance

variance

as

7-10:

an unbiased

to adding

Jt may not

in

adjusted

case 2.

variance

from

the entire
It

is

(equation

(S 2, for the entire

Y3
of the short-record
variance
(S

of the

the variance

adjusted

of

+ (N2-l)b2~x22

term

to give

short-term

N2.

Yl

Sy equals

equivalent

the

)Sy12

The third

factor

values

equation

(l-r2)S

(N,-2)

deviation

critical

an approximation

a constant

by equation

+ N2(Nl-4)(Nl-1)
(Nl-3)

critical

are

satisfies

(Nl-1

values

minimum

slightly.

coefficient

= -+j

the

minimum

for

can be computed

S2
y

represents

approximate

of Nla

the

variance

(7-9)

variance

period
2, using

the

con-

short

necessary
7-8)

to

to the

(N3)" ': Compute

equation

7-11.

y3

(7-11)
N3 - 1

7-5

where
7-11

all
is

terms

are

smaller

previously

than

the

defined.

variance

use S
as the final
estimate
y3
use the value of Sy determined

If

the

of Sy2 given

of

the

from

variance

of equation

in equation

standard

7-8,

deviation.

equation

then

Otherwise,

7-10.

FURTHER CONSIDERATIONS

The above
concurrent
a joint
this

observations
normal

of

obtain

the

standard

that

deviation.

site

the

length
r,

equation

but

as described

the

above

the

V.B.4.

7-s

sites

of zero.
are

In addition,

the

the
have
When

not
the

period,

be at least

adjusted

that

equations

concurrent

skew coefficient

skew with
in Section

a skewness

to estimate

only

assumption

caution,

N, should

necessary

generalized

and long-term

of the

is not

The adjusted

short

the

be used with

of

the

with

violated,

on the

under

at the

distribution

estimate
it

developed

flows

seriously

r depends

regression

by weighting

of

should

a reliable
Notice

record

is

technique

ability

were

probability

assumption

and this

from

equations

NT.

exact
reliTo

10 years.
actual

annual

logarithmic
should

skew computed

peaks

mean and

be computed
from

the

short

NOTATION
Nl = Number OF years
N2
N3

when flows

= Number of years
but not observed

Equivalent

observed

when flows were observed

at the
at the short record
site

= Number of years'of

Ne=

were concurrently

flow

years

of

at the

record

S = Standard
deviation
of the
Y
at the short record
site
S
= Standard
deviation
x1
during
concurrent

of

short
the

record

adjust

logarithm

of logarithm
period

longer

two sites

record

site

site

ed mean

of

of

at the

f lows

flows

for

at the

the

extended

long

record

period
site

S" = Standard
"2
for the

deviation
of logarithm
of flows at the long record
site
period
when flows were not observed
at the short record site

S
= Standard
y1
for the

deviation
concurrent

= not

of the
period

logarithm

of

flows

at

the

short

record

site

used

Y2
s

= Standard

'3

xl
5

short

deviation

record

= Logarithms

of flows

= Mean logarithm
period

x2 = Mean logarithm
flow records

x3
yl

= Mean logarithm
= Logarithms

of

of

of

from
flows

of

long

record

at the

long

of flows at the
are not available
of
flows

flows
from

= Mean logarithm
site

= Mean logarithm
of flows
record site
(concurrent

y2 = not

logarithm

flows

for

entire

the

period

at

the

site

of

flows

for

the

short
for

the

for the
period)

during

concurrent

record

site

for

period
the

concurrent

long record site for the

at the short record
site
entire
record
extended
period

used

7-7

period
during

at the

long

concurrent

period

at

of observed

period

the
flow

when

record

site

period
short

record

at the

short

*
5

= Mean logarithm

= RegreSsion

= Correlation
periods

of

flows

coefficient
coefficient

for
for

the

entire

period

at the

short

record

for

concurrent

site

Y, on X,

of the

flows

at the

two sites

TABLE 7-l

MINIMUM r VALUES FOR IMPROVING

MEAN OR STANDARDDEVIATIONESTIMATES

CONCURRENT
RECORD

MEAN

10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35

0.35
0.33
0.32
0.30
0.29
0.28
0.27
0.26
0.25
0,24
0.24
0.23
0.22
0.22
0.21
0.21
0.20
c-J.20
0.20
0.19
0.19
0.19
0.18
0.18
OJ8
0.17
7-9

STANDARD

DEVIATION
0.65
0.62
0*59
0.57
0.55
0.54
0,52
0.50
0.49
0,48
0.47
0.46
0.45
0.44
0.43
0.42
0.41
0.41
0.40
0.39
0.39
0.38
0.37
0.37
0.36
0.36

Appendix

WEIGHTINGOF INDEPENDENT ESTIMATES

The following
estimates
nearby

is suggested

based upon short

hydrologically

*generalization
purposes

similar

that

is true

If

records

the variance

z =

adjusting

to reflect

flood

using

in V.C.1,

the estimates

flow

frequency

experience

in

any one of the various

The procedure

are independent,

is based upon

which

for

practkal

in most situations.

two independent
variance,

for

watersheds,

methods mentIoned

the assumption

their

procedure

estimates

the variance

are weighted

of the weighted

estimate.

of either

According

inversely
average,

proportional
z, is less

to Gilroy

(30),

to

than

if

x(VJ + Y(Vx)
(8-J

vy + vx
then
vv
XY

v, =
ox
in which

Vx + Vy + 2rdm

Vx, Vy, and Vz are the var%ances

and r is the cross

values

:L

+ Vy12

of y.

correlation

Thus,

if

(8-2)

of x, y,

coefficient

two estimates

.
I

and z respectively,

between

values

are independent,

of x and

r is zero

vz = vxvy
vx + v
y <:
As the variance
computed

of flood

(8-3)

(8-3)
events

at selected

by the Pearson Type III

the number of annual

and

events

procedure

exceedance
is inversely

probabilities
proportional

used to compute the statistics

(25),

to
equation

can be written
(WJx)

(C/NY)

C/N, =

(8-4)
C/N, + C/N

in which

Y
C is a constant,

N, and NY are the number of annual

to compute x and y respectively,

would

be required

probabilities
equation

to give

with
8-l.

and N, 'fs the number of events

a flood

a variance

events

event

equivalent

Therefore,
8-l

at the selected
to that

that

exceedance

of z computed

by

used

Nz= N,
From equation

f NY

(8-5)

8-1,
(8-6)

Equation
of flood

8-6 can be used to weight

discharges

to appraise
discharge

the

accuracy

estimated

from the station

stations
miles

at selected

probabilities

of the weighted

by generalization

data,

such weighting

station

for

stations

their

period

which

the estimate

covering

the weighted

a smaller

average

is being

by equation

the estimates

station

data are not independent,

reduced

depending

Given

on the cross

a peak discharge

from a generalization
lo-year
+lS

record,

annual

from which

be obtained

to samples

N should

be that

in the preceding
for

probability

a sample of size

the

that

100

at the
based on

the accuracy
given

of

by equation

8-5.

and from the

estimate

is

of the estimates.
exceedance

equivalent

probability

to an estimate
estimate

the weighted

of 0.02
based on a

of 2,000 cfs

average

would

from

be given
+

2000) = 3 181
.

By equation

estimate

is as good

peaks.

to probability

is to be applied
should

distributions

gives

equivalent

probability

from a normal

8-2

to a weighted

be the same as that

as described

that

the expected

25 taken

8-5 this

adjustment

a sample size

example,

if

of over

of the weighted

and an independent

from 25 annual

from normal
for

the accuracy

1000) f 15(log
25

the adjustment

than

obtained

8-6 as follows:

is 1,520 cfs.

an expected

estimate,
cable

Q.,,

lO(log

with

from the generalization

at the site,

in equation

Log Q,,, =

If

example,

peaks observed

by substitution

as would

for

a radius

records,

8-6 is less

an accuracy

of that

For generalizations

shorter

correlation

frequency

particularly

in comparison

of 1,000 cfs with

with

justified

made,

logarithms

8-5 can be used

As a flood

an area with

is long

of the

to be independent

is often

area or with

given

For cases where

average.

cover

of record

estimates

and equation

tends

used in the generalization

or if

independent

in Appendix
accuracy.

adjustment

distribution,

would

appli11, but
Thus,
be that

Appendix

CQNFIUENCE LIMITS.
The record
underlying

of annual

population

frequency

curve

be selected

of

from

underlying

that

a different

flood

frequency

frequency

curve

of

gauge

accuracy
of

called

the

contains

confidence

period
curve

the

the

would

of

of annual

curves

that,

with

frequency

and their

curve.

end points

are

the

estimate
result.
flood

, one may construct

a high

the

sample

an approximation

population

population

sample

random

a different

probably

approximation

intervals

same size

can be only

frequency

is a random

and can be used to estimate

of time,

underlying
this

a site

If

curve

hypothetical

confidence,

peaks

frequency

of

at

population.

estimated

or range

flows

of annual

population

the

peak

could
of the

Thus,

an

to the

true

peaks.

To

an interval
degree

of

Such intervals
called

are

confidence

limits.
This
flood
end,

appendix

explains

discharges
let

that

Xt denote

the

exceedance

probability

confidence

level

on the
up,c(x)

below
property

observed
1 ies

X6 with

have

c, are

above

specified

true
P.

flood

how to construct

defined

records,

X6 with

probability

exceedance

or population
Upper

confidence

to be numbers
X, such that
That

Up ,(X)
the

c and the
is,

the

intervals

probabilities.

logarithmic

and lower

probability
c.

confidence

;pper
lower

confidence

To this

discharge
limits

for
that

for

X;

and Lp ,(X),
confidgnce
limit
limits

has
) with
based

limit

Lp ,(X)
ha;e

lies

the

that
Probability

(up,c(xl

q.!i~

= c

(g-la)

Probability

bp , ,(x1

5.

xfq

= c

(g-lb)

9-I

*
Explicit

formulas

for

the above formulas

computing

simply

the confidence

explain

limits

the statistical

are given

meaning

below;

of the confidence

limits.
The confidence
limits

because

limits

defined

above are called

each of them describes

of the population
can be formed

p-probability

a bound or limit

or union

confidence

on just

A two-sided

discharge,

from the overlap

one-sided

one side

confidence

of the two one-sided

interval

Intervals,

as follows:
Probability
Thus,

the union

{ Lp ,(X)
,

f x*p 2 up,c(x)

of two one-sided

go-percent

sided

so formed may not be the narrowest

with

that

confidence

use with
*
xp

these

It
if

and only

if

be called

an upper

proportion

(1-P)
limit

terminology

tions

Up c (XI lies

general

In quality

tolerance

limit,

intervals

that

possible

the twointerval

is considered

to proportions
refers

directly

formulas

for

satisfactory

Up ,1X)
(1IP)

(Pb,

for

Lp ,(X1

would

tolerance
whereas

to the discharge
in these
limits

between

of interest,

possible
the
be a lower
limit
the confidencethe

guidelines.
are derived

additional
limits

by specifying

simplifying

sample statistics

form of the confidence

above

UP ,(X1 would
c,,;fo;

of the population,

and making

of all

level

Because tie

the confidence

can lie

terminology,

Similarly,

the relationships

The general

control

is adopted

form of the limits

that

at confidence

the proportion

terminology

to analyze

statistics.

be noted

above a fraction

of the population.

refers

terminology
Explicit

it

in the above equations

for

confidence-limit
the

nevertheless,

in the population.

tolerance

should

(9-2)

guidelines.

may be noted

floods

limit

level;

It

= zc-1

confidence

is a two-sided
interval

interval.

g&percent

assumpand population

is specified

as:

up,c(x) = x f s KP
c
9
(
>

(9-3a)

Lp,clx)

(9-3b)

= x+

s
(

9-2

K"P c
9

in which x

the

final

are upper

and S are the logarithmic

estimated

log Pearson

and lower

confidence

The confidence
The non-central
the process

Type III

approximate

can be obtained

is cumbersome when G,

of the following
approximation

frequency

deviation

curve

of
and KF c and KL
P9C
9

coefficients.

coefficients

t-variate

mean and standard

approximate

is

in tables

non-zero.

formulas

to the non-central

the non-central

(32,

(41,

t-distribution.
321, although

More convenient

pp- 2-156,

t-distribution

Is the use

based on a large

(42):

Kit -KG,,P+l(K:wgP=ab

(9-4a)

sample

KGwgP - 4"'Gw.P
KL
P,c =

(9-4b)

in which
2
zC

a=l-

(9-5)

TJm-

(g-6)

and zc is the standard

with

cumulative

record

length

the estimated
limits

normal

probability

N is deemed to control
frequency

curve

95-percent

years

confidence

limits

when the estimated

deviation,

Pearson

probabflity

the statistical

Type III

l-c).

The systematic

reliability

and is to be used for

deviate)
of

calculating

confidence

information

has been used to estimate

9-3 through

9-6 is illustrated

the

curve.

The use of equations

flood,

(zero-skew

c (exceedance

even when historic

frequency

deviate

and skewness

of systematic

for

X; Olg the 0.01 exceedance

frequency'curve
of 3*00,

by calculating

0.25,

has logarithmic
and 0.20,

respectively

probability
mean, standard
based on 50

record.
*
9-3

*
z
C

= 1.645
=

= 2.4723
KG\4$p
(1.645)*
98

'-

= 0.9724

2 _

b = (2.4723)

(l$

= 6.058
e

K; o1
.

o g5 =
,

2.4723
3.026

o g5 =
,

- (0.9724)(6.058)

Kb o1
.

-I- d(;.V;;"

2.4723

4(2.4723+

- (0 9724)(6

0.9724

2.059

Uo.ol,

o.95

(X) = 3.00

+ (0.25)(3.026)

= 3.756

Loo,,

0.95

(Xl = 3.00 + (0.25)(2.059)

= 3.515

The corresponding
are 3270 and 5700;
4150 cubic
Table

058)

feet

the

in natural

estimated

0.01

units

(cubic

exceedance

feet

per

probability

second)

flood

is

per second.

t tables
(43) for
L
a skew of zero and can be used to compute KUp c and K ,, c for selected
values of P and c when the distribution
of loiarithms
oi the annual
peaks

9-l

limits

is

is normal

a portion

(i.e.,

An example

of the

Gw=O).

of using

table

follows:

Assume the

95-percent

the

exceedance

probability

0.01

mean,

standard

based

on 50 years

non-central

deviation

9-1

to compute

confidence
flood

for

and ske\rrness

of systematic

record,

9-4

of

limits

confidence

aru desired

a frequency
3.00,

limits

0.25

curve

is
for

with

as

X*o olg
0
logarithmic

and O.OU, respectively,

*Ku 0.01,

0.95 = 2.862

Found by entering
level

0.05,

exceedance
KL0.01,

level

"o,ol,

o.g5

(X) = 3.00 +0.25(2.862)

Loeol,

o.g5

(x> = 3.00 + 0.25tl.936)

3820 cubic

feet

limits

the estimated

with

confidence

length

50 and

0.01.

table

probability

9-l
record

probability

0.95,systematic

exceedance

are 3050 and 5190;

systematic

Found by entering

0.95 = 1.936

The corresponding

table

9-l
record

with

confidence

length

50 and

0.01.

= 3.715
= 3.484

in natural

units

0.01 exceedance

per second.

9-5

(cubic

feet

probability

per second)
flood

is

Appendix

9 Notation

up ,c (X>

= upper

confidence

limit

in log

units

LP,ctx)

= lower

confidence

limit

in

units

= exceedance

probability

= confidence

level

*
xP

= population

logarithmic

= mean logarithm

= standard

KGw,P

= Pearson

from

probability

GW

III

Type

deviations

discharge

of peak

deviation

of

logarithms

coordinate
the

for

exceedance

probability

flows

mean for

of annual

expressed
weighted

peak

in number

discharges
of standard

skew (Gw) and exceedance

(P).

= weighted

skew coefficient

KU
P,c

= upper

confidence

coefficient

KL
P,c

= lower

confidence

coefficient

= systematic

i!

= is the

log

record

standard

length
normal

deviate

9-6

CONFIDENCE LIMIT

TABLE 9-l
DEVIATE VALUES FOR NORMAL DISTRIBUTION
EXCEEDANCEPROBABILITY

Confidence
Level

Systematic
Record
Length
N

.002

,005

.OlO

.020

.040

,100

.200

.500

.Ol

10
15
20
25
30
40
50
60
70
80
90
100

6.178
5.147
4.675
4.398
4.212
3.975
3.826
3.723
3,647
3.587
3.538
3.498

5.572
4.639
4.212
3.960
3.792
3.577
3.442
3.347
3.278
3.223
3.179
3.143

5.074
4.222
3.832
3.601
3.447
3.249
3.125
3.038
2.974
2.924
2,883
2.850

4.535
3.770
3.419
3.211
3.071
2.893
2.781
2.702
2.644
2.599
2.561
2.531

3.942
3.274
2.965
2.782
2.658
2.500
2.401
2.331
2.280
2.239
2.206
2.179

3.048
2.521
2.276
2.129
2.030
1.902
1.821
1.764
1.722
1.688
1.661
1.639

2.243
1.841
1.651
1.536
1.457
1.355
1.290
1.244
1.210
1.183
1.160
1.142

.892
.678
.568
.498
.450
.384
.340
.309
-285
0265
.250
.236

-.107
-.236
-.313
-.364
-.403
-.457
-.496
-.524
-.545
-.563
-.578
-.591

.05

10
15
20
25
30
40
50
60
70
80
90
100

4.862
4,304
4.033
3.868
3,755
3.608
3.515
3.448
3.399
3.360
3.328
3.301

4.379
3.874
3.628
3.478
3.376
3.242
3.157
3.096
3.051
3;016
2.987
2.963

3,981
3,520
3.295
3.158
3.064
2.941
2.862
2.807
2,765
2.733
2.706
2.684

3.549
3,136
2.934
2.809
2.724
2.613
2.542
2.492
2.454
2.425
2.400
2.380

3.075
2.713
2.534
2.425
2.350
2.251
2.188
2.143
2.110
2.083
2.062
2.044

2.355
2.068
1.926
1.838
1.777
1.697
1.646
1.609
1.581
1.559
1.542
1.527

1.702
1.482
1.370
1.301
1.252
1.188
1,146
1.116
1.093
1.076
1.061
1.049

.580
,455
.387
.342
.310
.266
.237
.216
.199
.I86
.175
.166

-.317
-.406
-a460
-. 497
-.525
-.565
-.592
-.612
-.629
-,642
-.652
-.662

.800

.900
-.508
-.629
-.705
-.757
-.797
-.854
-.894
-. 924
-.948
-. 968
-.984
-.998
-.712
-,802
-,858
-.898
-.928
-.97O
-1.000
-1.022
-1.040
-1.054
-1,066
-1.077

.950

.990

-.804
-1.314
-. 929 -1.458
-1.008
-1.550
-1.064
-1.616
-1.107
-1.667
-1.169
-1.741
-1.212
-1.793
-1.245
-1.833
-1.272
-1.865
-10 293 -1.891
-1.311
-1.913
-1.326
-1.932
-1.017
-1.114
-1.175
-1.217
-1,250
-1.297
-1.329
-1.354
-1.374
-1.390
-1.403
-1.414

-1,563
-1.677
-1.749
-1.801
-1.840
-1.896
-1.936
-1.966
-1.990
-2.010
-2.026
-2.040

CONFIDENCE LIMIT

TABLE 9-l
(CONTINUED)
DEVIATE VALUES FOR NORMAL DISTRIRIJTION

EXCEEDANCEPQORARILITY
Confidence
Level

Systematic
Record
Length
N

.002

.005

.OlO

.02O

,040

.lOO

.200

.500

.800

.900

.950

,990

.lO

10
15
20
25
30
40
50
60
70
80
90
100

4.324
3.936
3.743
3.623
3.541.
3.433
3.363
3.313
3.276
3.247
3.223
3.203

3.889
3.539
3.364
3.255
3.181
3.082
3.019
2,974
2.940
2.913
2.891
2.873

3.532
3.212
3.052
2.952
2.884
2.793
2.735
2.694
2.662
2,638
2.618
2.601

3.144
2.857
2.712
2.623
2.561
2,479
2.426
2.389
2.360
2,338
2.319
2.305

2.716
2.465
2.338
2.258
2.204
2,131
2.084
2.051
2.025
2.006
1.989
1.976

2.066
1.867
1.765
1,702
1.657
1.598
1.559
1.532
1.511
1.495
I.481
I.470

1.474
1.320
1.240
1.190
1.154
1.106
1.075
1.052
1.035
1,021
1.010
1.001

.437
.347
.297
.264
.239
.206
al84
.l67
.I55
-144
,136
.l29

-.429
-.499
-.541
-.570
-.593
-.624
-.645
-.662
-,674
-,684
-,693
-.701

-.828
-.901
-.946
-.978
-1,002
-1.036
-1.059
-1.077
-1,091
-1,103
-1.112
-1.120

-1.144
-1.222
-1.271
-1.306
-1.332
-1.369
-1.396
-1.415
-1.431
-1.444
-1.454
-1.463

-1.715
-1.808
-1.867
-1.908
-1.940
-1.986
-2,018
-2.042
-2,061
-2.077
-2.090
-2.101

.25

10
15
20
25
30
40
50
60
70
80
90
100

3.599
3.415
3.320
3.261
3.220
3.165
3.129
3.105
3.085
3.070
3.058
3.048

3.231
3.064
2.978
2,925
2.888
2.838
2.805
2.783
2.765
2.752
2.740
2.731

2.927

2.596
2.460
2.390
2.346
2.315
2.274
2.247
2,227
2.213
2.202
2.192
2.184

2.231
2.112
2.050
2.011
1.984
1.948
1.924
1.907
1.893
1.883
1.875
1.868

1.671
1.577
1.528
1.497
1.475
1.445
1.425
1.411
1.401
1.392
1.386
1,380

1.155
1.083
1.045
1.020
1.002
.978
.962
-950
.942
.935
.929
.925

.222
.179
.154
-137
.125
,108
.096
.088
.081
.076
.071
.068

-.625
-.661
-.683
-.699
-.710
-.726
-.738
-.747
-.753
-.759
-.763
-.767

-1.043
-1.081
-1.104
-1.121
-1.133
-1.151
-1.164
-1.173
-1.181
-1.187
-1.192
-1.196

-1,382
-1.422
-1.448
-1.466
-1.479
-1.499
-1.513
-1.523
-1.532
-1.538
-1.544
-1.549

-2.008
-2.055
-2.085
-2.106
-2.123
-2.147
-2.163
-2.176
-2.186
-2,194
-2.201
-2.207

2.775
2.697
2.648
2.614
2.568
2.538
2.517
2.501
2.489
2.478
2.470

CONFIDENCE LIMIT
*

mm
9-1 (CONTINUED)
DEVIATE VALUES FOR NORMAL DISTRIBUTION
EXCEEDANCEF'RORABILITY

Confidence
Level

Systematic

Record
Length
N

.002

.005

.OlO

.020

.040

.lOO

.200

.500

.800

.900

,950

.990

.75

10
15
20
25
30
40
50
60
70
80
90
100

2.508
2.562
2.597
2.621
2.641
2.668
2.688
2.702
2.714
2.724
2.731
2.739

2.235
2.284
2.317
2.339
2.357
2.383
2.400
2.414
2.425
2.434
2.441
2.447

2.008
2.055
2.085
2.106
2.123
2.147
2.163
2.176
2.186
2.194
2.201
2.207

1,759
1.803
1.831
1.851
1.867
1.888
1.903
1.916
1.925
1,932
1.938
1.944

1.480
1.521
1.547
1.566
1.580
1.600
1.614
1.625
1.634
1.640
1.646
1.652

1.043
1.081
1.104
1.121
1.133
1.151
1.164
1.173
1.181
1.187
1.192
1.196

.625
,661
.683
.699
.710
.726
.738
.747
.753
.759
.763
.767

-.222
-. 179
-.154
-.137
-.125
-.108
-.096
-.088
-.081
-.076
-.07f
-.068

-1.155
-1.083
-1.045
-1.020
-1.002
-.978
-.962
-a 950
-. 942
-.935
-.929
-.925

-1.671
-1.577
-1.528
-1.497
-1.475
-1.445
-1.425
-1.411
-1.401
-1.392
-1.386
-1.380

-2.104
-1.991
-1.932
-1.895
-1.869
-1.834
-1.811
-1.795
-1.782
-1.772
-1.764
-1.758

-2.927
-2.775
-2.697
-2.648
-2.614
-2.568
-2.538
-2.517
-2.501
-2.489
-2.478
-2.470

-90

10
15
20
25
30
40
50
60
70
80
90
100

2.165
2.273
2.342
2.390
2.426
2,479
2.517
2.544
2.567
2.585
2,600
2.613

1.919
2.019
2.082
2.126
2.160
2,209
2.244
2.269
2.290
2.307
2.321
2.333

1.715
1.808
1.867
1.908
1.940
1.986
2.018
2,042
2.061
2.077
2.090
2.101

1.489

1.234
1.314
1.364
1.400
1.427
1.465
1.493
1.513
1.529
1.543
1,553
1.563

.828
.9Ol
.946
.978
1.002
1.036
1.059
1.077
1.091
1.103
1.112
1.120

.429
.499

-.437
-.347
-.297
-.264
-.239
-.206
-.184
-.167
-.155
-.144
-.136
-. 129

-1,474
-1.320
-1.240
-1.190
-1.154
-1.106
-1.075
-1.052
-1.035
-1.021
-1.010
-1.001

-2.066
-1.867
-1.765
-1.702
-1.657
-1.598
-1.559
-1.532
-1.511
-1.495
-1.481
-1,470

-2.568
-2.329
-2.208
-2.132
-2.080
-2.010
-1.965
-1.933
-1.909
-1.890
-1.874
-1.861

-3.532
-3.212
-3.052
-2.952
-2.884
-2.793
-2.735
-2.694
-2.662
-2.638
-2.618
-2.601

1.576
1.630
1.669
1.698
1.740
1.770
1.792
1.810
1.824
1.836
1.847

,541
,570
.593
.624
.645
.662
.674
.684
.693
.7Ol

CONFIDENCE LIMIT

TABLE 9-1 (CONTINUED)

DEVIATE VALUES FOR NORMAL DISTRIBUTION
EXCEEDANCEPROBABILITY

Confidence
Level

Systematic
Record
Length
N

.002

,005

,010

.020

.040

.I00

.200

.500

.800

.900

.950

.990

.95

10
15
20
25
30
40
50
60
70
80
90
100

1.989
2.121
2.204
2.264
2,310
2.375
2.421
2.456
2.484
2.507
2.526
2.542

1.757
1.878
1.955
2.011
2.053
2.113
2.156
2.188
2.214
2,235
2.252
2.267

1.563
1.677
1.749
1.801
1.840
1.896
1,936
1.966
1.990
2.010
2.026
2.040

1.348
1.454
1.522
1.569
1.605
1.657
1.694
1.722
1.745
1.762
1.778
1.791

1.104
1.203
1.266
1.309
1.342
1.391
1.424
1.450
1.470
1.487
1.500
1.512

.712
.802
.858
.898
.928
.970
1.000
1.022
1.040
1.054
1.066
1.077

.317
.406
.460
.497
.525
.565
.592
.612
.629
a642
,652
.662

-,580
-.455
-.387
-.342
-.310
-.266
-.237
-.216
-.199
-.186
-.175
-.166

-1.702
-1.482
-1.370
-1.301
-1.252
-1.188
-1,146
-1.116
-1.093
-1,076
-1,061
-1,049

-2.355
-2.068
-1.926
-1.838
-1.777
-1.697
-1.646
-1.609
-1.581
-1.559
-1,542
-1.527

-2.911
-2.566
-2.396
-2.292
-2.220
-2.125
-2.065
-2.022
-1,990
-1.964
-1.944
-1.927

-3.981
-3.520
-3.295
-3.158
-3.064
-2,941
-2,862
-2.807
-2.765
-2.733
-2.706
-2.684

.99

10
15
20
25
30
40
50
60
70
80
90
100

1.704
1.868
1.974
2.050
2.109
2.194
2.255
2.301
2.338
2.368
2.394
2.416

1,492
1.645
1.743
1.813
1.867
1,946
2.002
2.045
2.079
2.107
2.131
2.151

1.314
1.115
1.251
1.458
1.550
1,336
1.616
1.399
1.667
1.446
1.741
1,515
1.793 ' 1.563
1.833
1.600
1.865
1.630
1.891
1.653
1.913
1.674
1.932
1.691

.886
1.014
1.094
1.152
1.196
1.259
1.304
1.337
1,365
1.387
1.405
1.421

.508
.629
.705
.757
,797
.854
.894
.924
.948
.968
.984
.998

-107
.236
.313
.364
-403
.457
.496
.524
.545
.563
.578
.591

-.a92
-.678
-.568
-. 498
-.450
-.384
-.340
-. 309
-.285
-.265
-.250
-.236

-2.243
-1.841
-1.651
-1.536
-1,457
-1.355
-1.290
-1.244
-1,210
-1.183
-1.160
-1.142

-3.048
-2.521
-2.276
-2.129
-2.030
-1.902
-1.821
-1.764
-1.722
-1.688
-1.661
-1.639

-3.738
-3.102
-2.808
-2.633
-2.515
-2.364
-2.269
-2.202
-2.153
-2,114
-2.082
-2.056

-5.074
-4.222
-3.832
-3.601
-3.447
-3.249
-3.125
-3,038
-2.974
-2.924
-2.883
-2.850
*

Appendix

10

RISK
This
the risk

appendix
incurred

used in this
events

describes

will

the recommended procedures

when a location

guide,

risk

is occupied

is defined

exceed a given

flood

for

for

a period

as the probability

magnitude

estimating

within

of years,

that

As

one or more

a specified

period

of

years.
Two basic
methods

approaches

[(e.g.,

methods

which

annual

(19)]

and parametric

use the binomial

exceedance

methods

may be used to compute

frequency

is not great,

consequently,

expression

events

RO

flood

require
known.

(20)) e

Parametric

assuming

that

The difference

distribution

the

between

interest;

is recommended because of

for

estimating

risk

is:

N-I

risk

(10-l)

of obtaining

exceeding

a flood

in N years

magnitude

with

exactly
annual

I
exceedance

P.

When I equals
in which

[(e.g.,

1n the range of usual

(1-P)

RI is the estimated

number of flood

nonparametric

and application.

RI = *PI

probability

is exactly

use of the binomial

The binomial

in which

distribution

particularly

ease of comprehension

methods

risk,

0 equation

10-l

reduces

to:

= (1-P)N

(lo-a)

R, is the estimated

magnitude

in N years.

probability
From this

of nonexceedance
the risk

of the selected

R of one or more exceedance

becomes
R (1 or more)

= 1 - (~MP)~

Risk of 2 or more exceedances,

R(2 or more)
*

Some solutions

(10-3)
R (2 or more),

is

= R-R, = R-NP (l-P)N*l

are illustrated

by the following

10-l.

10-l

(10-4)
table"and

figure

BINOMIAL RISK TABLE

TIME

NONE

ONE OR
MORE

35
12
4

65
88

tif

;69

82
92

50
60

;
0

iti

i
0
0

1;:
110
120
150
200
TIME

10
20
iii
6";
70
80
12
110
120
150
200

NOTE:

** RISK (PERCENT) **
P=O.O50

** RISK (PERCENT) **
P=O.lOO

ii
ii

1;1:
100
100
100
100
100
100
100
100

TWO OR
MORE

NONE
60
1;

1::
100
100
100
100
100
100

** RISK (PERCENT) **
P=O.O40
NONE

ONE OR
MORE

66
44
29
20
13
ii
4

34
56
71
80
87
91
94
96

;
1

ii;
99

!I
0

1::
100

TWO OR
MORE

40
64

TWO OR
MORE
9

i;
92
ii;

~~
60
72
81

%
;

98
;;

87
91
94

100

;ii

0
0

100
100

190:
100

13
8
5

ii'9

ONE OR
MORE

** RISK (PERCENT) **
P=O.O20
NONE

ONE OR
MORE

TWO OR
MORE

82
67

2
6

2
60
70
78
83
88
91

ii
ii

45
70
5:

:%
26

;;

80
76

ii48

:"3

87
84

i"o

;:

11
z
2

89
91
iii

ii;
80
91

1;

lFl"o

TABLE VALUES ARE ROUNDEDTO NEAREST PERCENT

IO-2

BINOMIAL RISK TABLE

TIME

** RISK

F"R"o:No" **
b
NONE ON; OR TWO OR
MORE
MORE

::
3:
:;

TIME

NONE

O;E *OR
MORE

TWO OR
MORE

ii"0

1:

86
82
78

ii

E
30
33
fig

4"

0
ii

26
33

i
6

4359

70
80

1::
110
120
150
200

** RISKp (JEOROC$NT
**

lf
1;
23

33

2
60
63
67

;t
67
64

5:

58
61

42

1;

;2"
13

E
87

4":
60

i;
37

45
;3

:7"
26

** RISK (PERCENT) **
P=O.O02
NONE

ONE OR
MORE

:
8

** RISK (PERCENT) **
P=O.OOl

TWO OR
MORE

ONE OR
MORE

TWO OR
MORE

6
1:
:i
15
16
18

NOTE:

TABLE VALUES ARE ROUNDEDTO NEAREST PERCENT

*.

10-3

1:
20
30

40
50
60
70

80

,Y
92
93
94
95
96
97

98

99

Figure 10-l. RISK OF ONEOR MOREFLOODEVENTSEXCEEDING

A FIOOD OF GIVEN ANNUALEXCEEDANCE
FREQUENCY
WITHIN A F'EEUODOF YEARS

1 o-4

Appendix

11

EXPECTED PROBABILITY
The principle
applied

of gambling

to water

probabilities

resources

based upon estimated

development

must be inferred

and mathematical
uncertainty

decisions.

cannot

do not necessarily

be computed

compensate.

based on sample data is that

a certain

flood

it

could

or four

per hundred

than

zero

direction

times

due to uncertainty

errors

in the other
half

consider

the relative

It

when dealing
normal

with

or probability

samples
curves
it

11-l

The vertical

scale

the value

is the curve
normal

that

population.

on the average

30 times

percent

that

chance

or a 5 percent
thousand
If

that

be

in one
impact

to simply

of
be

is necessary

considerable

to

accuracy

Therefore,

closely

when

to the logarithmic
uncertainty

of frequency

of the range of uncertainty

drawn from a normal

curve
frequency

the true

The curve

frequency

a magnitude

magnitude

labeled

estimate

of 2 would
The figure

also

is two times

.50

of a log-

be exceeded

is 150 or more times

frequency

The

of any flood

scale,

events.

population.

of streamflow.

frequency

the best

per thousand

the true

chance

shows a 5
per thousand

or less

per

events.
a magnitude

the best estimate

for

the true

be used for

From this

with

the logarithm

shown on the frequency

would

can never

or too low,

representation

can represent
that

It

distribution.

based on samples

show the likelihood

it

exceedance

flows.

is a generalized
estimates

half.

to delineate

of flood

but

from the

too high

the estimate

the true

years,

due to

be exceeded

different

fairly

is possible

that

is not adequate

a normal

conform

will

of errors

uncertainty

from

estimates

in probability

exceeds

it

of being

to delineate

distribution,

curves

Thus,

impacts

frequency

Figure

can be quite

if

The impact

the time and too low the other

is possible
flow

years.

direction.

too high

flood

per hundred

is possible

are uncertain

as errors

magnitude

once in 100 years,

more times

exactly

can be

because

they

For example,

on the average
less

However,

from random sample data,

expectation

be three

probabilities

each location.

of 2,O were selected

for

the frequency
The estimated

at 20 independent

is 3 exceedances
total

exceedance

locations,

per hundred
for

all

years

20 locations

would

frequencies
total

However,

be 60 per 100 years.

for

a magnitude

exceedances

20

.9

12

.8

10

.5

.3

.1

from

from Figure

using

11-l

using

exceedances

the best

probability

bility

were used instead

function

however,

could

read the expected

value

of about

would

the low direction,

probability
If

the parent
observed

this.

frequency

population
flood

would

events

number of observations

give

Thus,

the heavier
for

curve

It

of all
estimate

could

be defined--the

would

approach

probabilities

are not accurately

probabilities

as illustrated

the parent

known.
in Figure

11-2

at this

the

when applied
per

probability
than

in

in the low direction

point,

that

known--that
distribution

population

This
Howeverp
11-l

we

expected

probabilities,

frequency

infinity.

estimate

more frequently

be accurately

could

proba-

of 90 exceedances

the expected

true

at

obtain

value

of being wrong

estimated

approaches

This

direction

can be noted,

expected

ll-ltto

an estimate

impacts

probability

"best"

of Figure

while

or 30 more than

derived

of the traditional

in the high

is the average
a flood

increments

as the true

per 100 events.

20 locations.
be wrong

compensate

value

the mathematically

4.5 exceedances

at all

90

90 per 100 years

estimate

probability

to each of the 20 locations

would

be represented

= Approximately

0.05 parameter

is about

If,

estimate

and

at Each of 20 Locations*

Exceedances

each location.

100 years

might

true

.975.

of these

obtained

at each location

at the 20 locations

Per 100 Years

Total

.025 through

The total

differ

uncertainties,

tabulation.

Exceedances

*Determined

of 2.0 would

per 100 years

by the following

due to sampling

is,
of

as the

is not the case where

if

the expected

can be computed,

observed

flood

frequency

for

the estimated
infinity

a large

flood

number of independent

frequency

appears

probability

that

should

the answer

If

answer to the above question

concepts
would

return.

should

indicate
It

that

for
it

a single

= Prob

$j-l
c

tistic

with

The actual

degrees

across
value

11-1,

enter

with

with

of reasoning

equations

exceedance

distribution

the expected

(11-l)
I

of the desired
and tN-1

values

is the

probability
Student's

the sample size

correction
(34) which

probabilities

the appropriate

or the equations

out using

shown in Table

plotting
may also

t-sta-

PN value

11-3

tables

of

11-l

(31).

1 and read

probability.
be calculated

than 0.500
equals

The

position.

are based on Table

greater

11-2,

minus

exceedance

is the corrected

from the following

11-l

to

probability

l/2

(-&)

the desired

probability

Table

be proportional

the same line

can be carried

The expected

sample size,

the

of freedom.

read from the table

given

must9 then

the expected

the normal

variate

or the modified

to the column

For selected

Kn

normal

calculations

the t-statistic,
To use Table

for

N is the sample size,

N-l

gamble?"

from the formula

where K, is the standard

of exceedance,

whether
location,

that

PN can be obtained

pN

a single

to

should.

has been shown (21)

probability

wager for
it

expected

be identical

the wager should

In determining

apply

would

and ordinarily

is that

approaches

as to whether

location

"What is a fair

approach

infinity.

to the question

the gamble must be undertaken,

the expected

approaches

be used at a single

the answer to the question,

will

as the number of observations

and the number of locations

It

locations

11-l.
and a

1 minus the value

in

Exceedance

Probability

Expected

Probability,

PN

.OOOl

.OOOl (1.0

,001

.OOl (1.0

.Ol

a01 (1.0

+ 26/N1*16)

(ll-2c)

.05

.05 (1.0

+ 6/N1004)

(11-2d)

.lO

,l

.30

.3 (1.0

For floods

with

samples of 20 annual
of exceedance
Use of Table

an exceedance
peaks,

from equation
11-l

gives

for

(1.0

+ 280/N1*55)

(11-2b)

(11-2e)

+ 0.46/Noog25)

(lL2f)

probability

example,

of 0.01 based on

the expected

Comparable

(1.0

probability

+ 26/32.3)

equations

or 0.018.
adjusting

upward to give

a discharge

for

probability

the exceedance

probability

are available

11-4

which

for

computed discharge
equals

(11-*a)

+ 3/Nloo4)

11-2~ ds (.Ol)

0.0174.

+ 1600/N1'72)

the expected
(22).

the

4
I

Note: Parameter IS relative frequency with which

true value exceedsthe indicated value as the
number of random samplesof this sizeapproaches Infinity.
w

I/Y

Y.

Figure 11-l
PRO6ABILlTYESTIMATES
FROM
NORMALDISTRIBUTIONSAMPLE
N=lO

99.9

99.8

99.5

99

98

95

90

80

60

30

20

EXCEEDANCE
FREQUENCY,IN PERCENT

10

II
2

Table 11-l
TABI
For use with

OF PN VERSUSPO0

samples drawn from a normal populatSon

?XOTE: pN values above are usable approximately

with
distributions
having small skew coefficients.

Pewson Ty-pe III

Appendix

12

FLOW DIAGRAM AND EXAMPLE PROBLEMS

The sequence
potentials

of procedures
(except

the following
A.

B.

and flow

available

1.

Previous

2.

Gage records

3.

Historic

4.

Studies

5.

Watershed

Evaluate

for

defining

is described

flood
in

diagrams.

data and data to be used.

studies
data
for

similar

watersheds

model

data.

Record homogeneity
Reliability

and accuracy

Compute curve
flow

guide

the case of mixed populations)

outline

Determine

1.
2.
c.

for

recommended by this

following

diagrams.

techniques

follow

guide

procedures

Example problems
the flow

showing

diagram.

12-l

as outlined

in following

most of the computational

ZEROCFLOOO

INCOMPLETE

RECORD

SEE APPENDIX 5,
cotaITloNAL
PROBAl3lLll-Y
ADJUSTMENT,
FOR
OUTLIERS SEE
PAGES 17 TO 19
AFY, APPENDIX
BAND 6

COMPUTE

COMPLETE

RECORD

COMPUTE

STATION

STAnSTlCS

EXTENDED

RECORD APPENDIX

4t- IF SYSTEMATIC
RECORD LENGTH IS
LESS THAN 50 YEARS
THE ANALYST
SHOULD
CONSIDER
WHETHER
THE
USE OF THE PROCEDURES
OF
APPENDIX
7 IS APPROPRIATE.

NOR3

IS FURlHER

ANALYSIS

WARRANTED@

STEPS TO THIS POINT ARE BASIC

STEPS REQUIRED IN ANALYSIS
OF
READILY AVAllpBLE
STAnON
AND
HISTORIC OATA. AT THIS POiNT A
DECISION
WHETHER

SHOULD BE MADE
FUTURE FURTHER

AS TO
REFINE-

hwr
w niiz FREQUENCY
mwm
IS JUSTIFIED.
MIS DECISION WlLL
DEPEMJ BOTH UPON TIME AND
EFFORT REQUlREO FOR REFINEMENT
ANO UPON THE PURPOSE OF THE
FREQUENCY

FLOW

Lrzl
FINAL CURVE

bl

IF DESIRED

ESTIMATE.

DIAGRAM

FOR FLOOD FLOW FREQUENCY ANALYSIS

12-2

*FLOW

DIAGRAM

FOR

HISTORIC

AND

OiJTLIER

ADJUSTMENT

RECOMPUTE
sTm+m;;~s
LOW
OUTLIERS

1YES

YES

YES

NO
\r

iRECOMPUTE
STATISTICS
ADJUSTEb
FOR
HISTORIC
PEAKWHIQH
OUTLIERS
APPENDIX
8

RECOMPUTE
ST;;+m;:~s

RECOMPUTE
ST;;+/W;;~S

LOW
OUTLIERS

OUTLI

LOW
ERS

L
I
CONDITIONAL
PROBABILITY
ADJUSTMENT
APPENDZX

The following
techniques
four

examples

recommended in this

statifns

(Table

1.

Fitting

2.

Adjusting

3.

Testing

4.

Adjusting

in Appendix

12-l)

for

as this

for
*skew

years.
expected

In actual
sources

off

may differ

These examples
recommended in Bulletin
the following

errors

flood

data is given

the analysis

An example
of an incomplete

in Example 4, adjusting
limits

for

zero

and the adjustment

in Example 1.

examples

was taken

The generalized

from Plate

skew may be obtained

made for

I.

from other

the region.

in the computational

beyond the

second decimal

have been completely

17B.

the following:

distribution

is provided.

for

the generalized

Because of round

to illustrate

historic

are described

study

values

for

of confidence

used in these

practice,

peak data for

years

computation

The computation
probability

flood

low outliers

is applied

or a special

computed

flood

specifically

technique

coefficient

for

adjusting

has not been included

flood

zero

6 and an example

Type III

of most of the

outliers

and adjusting

for

Annual

have been selected

high

for

application

guide.

the Log-Pearson

The procedure

record

illustrate

Specific

pages:

12-4

revised
changes

procedures,
point.

using

the procedures

have not been indicated

on

TABLE 12-l
ANNUAL FLOOD PEAKS FOR FOUR STATIONS IN EXAMPLES

fear
1929
I930
1931
1932
1933
1934
1935
I936
I937
I938
1939
I940
I941
1942
I943
1944
I945
1946
1947
1948
1949
I95C
1951
I952
1953
I954
1955
1956
1957
195E
1959
l96C
1961
1962
1963
1964
1965
1966
1967
I968
1969
197c
1971
1972
197:
*Not

Fishkill
Creek
01-3735
Example 1

Floyd River
06-6005
Example 2

Back Creek
01-6140

--Yw15500
4060
l-

2290
1470
2220
2970
3020
1210
2490
3170
3220
1760
8800
8280
1310
2500
1960
2140
4340
3060
1780
1380
980
1040
1580
3630
-"I-

included

in example

1460
4050
3570
2060
1300
1390
1720
6280
1360
7440
5320
1400
3240
2710
4520
4840
8320
13900
71500
6250
2260
318
1330
970
1920
15100
2870
20600
3810
726
7500
7170
2000
829
17300
4740
13400
2940
5660
computations.

22000*
6;OO
3130
4160
6700
22400
3880
8050
4020
1600
4460
4230
3010
9150
5100
9820
6200
10700
3880
3420
3240
6800
3740
4700
4380
5190
3960
5600
4670
7080
4640
536
6680
8360
18700
5210

Orestimba
Creel
11-2745
Example 4
-T-

-L

4260
345
516
1320
1200
2150
3230
115
3440
3070
1880
6450
1290
5970
782
00
335
175
2920
3660
147
0
56::
1440
10200
5380
448
174:
8300
156
560
128
4200
0
5080
1010
584
151:

EXAMPLE 1
FITTING THE LOG-PEARSONTYPE III
a.

Station

DISTRIBUTION

Description

Fishkill

Creek at Beacon,

New York

USGS Gaging Station:

01-3735
Lat: 41"30'42",
long:
73'56'58"
Drainage Area:
190 sq. mi.
Annual Peaks Available:
1945-1968
b.

Computational

Step 1 - List data,

the cubes.

Year
194
194:
1947
1948
1949

Procedures
transform

to logarithms,

and compute the squares

TABLE 12-2
COMPUTATION OF SUMMATIONS
Annual Peak
Logarithm
(cfs)
(x)
2290
3.35984
1470
3.16732
2220
3.34635
2970
3.47276
3020
3.48001
'

X2
10:031:2

and

11.19806
12.06006
12.11047

X3
37.92764
31.77429
37.47262
41 .88170
42.14456

3.08279
3.39620
3.50106
3.50786
3.24551

9.50359
11.53417
12.25742
12,30508
lo,53334

29.29759
39.17236
42.91397
43.16450
34018604

1950
1951
1952
1953
1954

1210
2490
3170
3220
1760

1955
1956
1957
1958
1959

8800
8280
1310
2500
1960

3.94448
3.91803
3.11727
3.39794
3.29226

15.55892
15.35096
9,71737
11.54600
10.83898

61.37186
60.14552
30,29167
39.23260
35.68473

1960
1967
1962
1963
1964

2140
4340
3060
1780
1380

3.33041
3.63749
3.48572
3.25042
3e13988

11.09163
13.23133
12.15024
10.56523
9.85885

36.93968
48.12884
42.35235
34.34144
30.95559

1965
1966
1967
1968
N=24

980
1040
1580
3630
MS

2.99123
3.01703
3.19866
3.55991
C 80.84043

8.94746
9.10247
10.23143
12.67296
273.68646

26.76390
27.46243
32.72685
45.11459
931.44732

12-6

Example
Step

1 - Fitting

the

2 - Computation

Log-Pearson

I II

Type

of mean by Equation

Distribution

(continued)

2:

= !%@$!t?= 3.3684
Computation

of standard

(12-l)
deviation

by Equation

3b:

0.5

-N(EX)2/N
S = px2
1
[

0.5

- (80.84043)'/24
23

S =

Computation

G=

d-

(12-Z)

= 0.2456

of

skew coefficient

N2EX3)

- 3N(CX)EX2)

by Equation

4b:

f 2 &Xl3

N(N-lj(N-2)S3
= (2412(931.44732)

- 3(24)(80.84043)(273.68646)

24124-l)

(24-Z)

6.24561)3

= 536513.6563 - 1592995.0400 f 1056612.7341

(24) (23) (22)
131.3504
= im

t.014816)

= 0.7300

12-7

+ 2(80.84043)3

(12-3)

Example
Step

I - Fitting

the Log-Pearson

3 - Check for
xH = i-+

test

+ 2.467

value

value

detecting

is used in

1.2456)

(continued)

= 3.9743
=

(3.9743)

recorded
for

Distribution

KNS

QH = antilog

the

III

Outliers:

= 3.3684

The largest

Type

does not

possible

equation

(12-4)

9425 cfs

exceed

the

low butliers

8a to compute

threshold

value.

is applied.

the

Next,

The same K,,,

low outlier

threshold

(0,~:

XL = x - KNs
= 3.3684

- 2.467i.2456)

QL = antilog
There
were

are

generalized
study
Equation

the

skew of 0.6

is

within
error

Table

below
high

the

or low tests.

1 or computed

A - 0
c

A=

+ 0.08

-0.33

0 = 0.94

other

error

No outliers

I.

example

In actual

practice

or from

a special

sources

skew is computed

of the

by Equation

6.

station

by use of

skew can be found

Computation

of mean-square

skew by Eq. 6:

MSEC ci 1o
Where:

from

value.
For this

from Plate

A weighted

The mean square

of station

threshold

determined

region.

(12-5)

579 cfs

this

skew may be obtained

made for
5.

values

by either

a generalized

(2.7625)

no recorded

detected

= 2.7625

- 0.26

MSEC p' IO Cse2716

Cl oglOW~)

IGI

= -0.33

IGI
- -7502

0.94

+ 0.08(.730)
- 0.26(.730)

~~qok4)+

12-8

= -.2716

(12-6)

(12-7)

.7502

1o s.55683

5 0.277

(12-R')

Example

1 - Fitting

the

The mean-square

Computation

4 - Compute

error

.302( .73;;4

0.7

the

frequency

Distribution

generalized

(continued)

skew from

Plate

I is 0.302.

5:

3.

III

using

KS

11500

K values

tenth)

for

a skew coefficient

computation

Equation

3.3684

(12-9)

0.6678

coordinates.

An example

.Ol

Q=

to nearest

of
x

The discharge

+ .277(.6)

curve

Type

Q =

rounded

the

III

skew by equation

(rounded

in Appendix

probability
log

Type

MSPg (G) -I- MSEG(G)

MSEE + MSEG

The log-Pearson
found

of

of weighted

G, =

Step

Log-Pearson

for

of 0.7

are

an exceedance

1 follows:

2.82359(.2456)

4.0619

(12-10)

cfs
values

to three

in

significant

this

computation
figures.

12-9

and those

in Table

12-3

are

Example

1 - Fitting

the

Log-Pearson

III

Type

Distribution

(continued)

TABLE 12-3
COMPUTATION OF FREQUENCY CURVE COORDINATES
'Gw,P
P
-

for

.99
.90
.50
.lO
.05
.02
-01
.005
.002

-1.80621
-1.18347
-0.11578
1.33294
1.81864
2.40670
2.82359
3.22281
3.72957

log

Q
cfs

The frequency

Step

Gw = 0.7

5 - Compute

curve

the

The upper
.05 and
Appendix

2.9247

is plotted

confidence

and lower

define

the

on the

curve

Nine

in Figure

are

limits

computed

exceedance

confidence

12-l.

limits.
confidence

.95 percent
9.

841
1200
2190
4960
6530
9110
11500
14500
19200

3.0777
3.3399
3.6957
3.8150
3.9595
4.0619
4.1599
4.2844

limit

at an exceedance

by the

probabilities
curves.
probability

12-10

for

levels

of significance

procedures

outlined

of
in

(P) have been selected

The computations
of 0.99

for
are given

two points
below.

to

-Computed
---With

Frequency
Expected

Curve

Probability
Confidence

Limit

:CEEDANCE

PR4BABiITY-

Figure
Frequency
Fishkill
Beacon,

12-1
Curves for
Creek at
New York

Example

12-l 1

Example 1 - Fitting
Equations

the Log-Pearson

in Appendix

The normal

KP,c-

The Pearson

Distribution

9 are used in computing

deviate,

skew coefficient

Type III

of zero.

Type III

is found

zc,

an approximate

by entering

For a confidence

Appendix

level

p are found

deviates,KG

(continued)

of 0.05,

in Appendix

value

for

3 with

zc = 1.64485.
3 based on

WP

the appropriate

skew coefficient,

and skew coefficient

For an exceedance
KG p =

of 0.7,

probability

of 0.99

-1.80621.

22
l-,b

K2

= 1 - .q-$$yp* =
22
+

G,,

(-1.80621)2

KGw,P+dF

#
KP,c

0.9412

(12-U)

b644W2
24

3,14g7

(12-12)

-l.80621)2-(.9412)(3.1497)
.9412

(12-13)
pl.80621

-I-

.9412
The discharge
Log Q =

3.3684

3.0395

1100

Q
For the

I?P,C

value

lower

confidence

-,

. 33g2

is:
+

(-1,3392)(,2456)

(12-14)

coefficient:
"&,P

KGwaF

.5458

-ab

-1.80621 - .5458 = -* e4989

.94-i*

12-12

(12-75)

Example 1 - Fitting

the

The discharge

Log-Pearson
value

Type III

Distribution

(continued)

).'.,

is:

Log Q = 3.3684 + (-2.4989)(.2456)

(12-16)

= 2.7546
Q = 568
The computations
dence limits
in Figure

showing

are given

the derivation
in Table

of the upper and lower

12-4.

The resulting

curves

confi-

are shown

12-l.
TABLE 12-4
COMPUTATION OF CONFIDENCE LIMITS

KGw,B

for

.99
.90
-50
.lO
.05
.02
.Ol
.005
.002

G, = 0.7

-1.80621
-1.18347
-0.11578
1.33294
1.81864
2.40670
2.82359
3.22281
3.72957

0.05 UPPER LIMIT CURVE

U
Q
log Q
'b SC
cfs
-1.3392
-0.7962
0.2244
1.9038
2.5149
3.2673
3.8058
4.3239
4.9841

Step 6 - Compute the expected

The expected
Table

11-l

probability

3.0395
3.1728
3.4235
3.8359
3.9860
4.1708
4.3031
4.4303
4.5925

1100
1490
2650
6850
9680
14800
20100
26900
39100

probability
plotting

0.05 LOWER LIPiIT CURVE

L
Q
log Q
cfs
9C

Ib

-2.4989
-1.7187
-0.4704
0.9286
1.3497
1.8469
2.1943
2.5245
2.9412

2.7546
2.9462
3.2528
3.5964
3.6998
3.8220
3.9073
3.9884
4.0907

adjustment.
positions

based on N - 1 of 23.

12-13

are determined

from

568
884
1790
3950
5010
6640
8080
9740
12300

Example

1 - Fitting

the

Log-Pearson

Type

III

Distribution

(continued)

TABLE 12-5
EXPECTED PROBABILITY ADJUSTMENT

.99
.90

841
1200
2190
4960
6530
9110
11500
14500

.50
.lO
.05
.02
.Ol
.005
.002
values

The frequency

curve

Figure

.9839
.889
.50
.111

0060
.028*
.0161
e0095
* 0049

19200

*Interpolated

in

Expected
Probability

adjusted

12-l.

12-14

for

expected

probability

is

shown

EXAMPLE 2
ADJUSTING FOR A

a.

Station
Floyd

HIGH OUTLIER

Description
River

at James,

Iowa

USGS Gaging Station:

06-6005
Lat:
42o34'30",
long:
960 18'45"
Drainage Area:
882 sq. mi.
Annual Peaks Available:
1935-1973
b.

Computational

Step 1

Procedures

Compute the statistics.

The detailed
computations
for
have been omitted;
the results

the systematic
record 1935-1973
of the computations
are:

Mean Logarithm
Standard Deviation
of logs
Skew Coefficient
of logs
Years

3.5553
0.4642
0.3566
39

At this point,
the analyst
may wish to see the preliminary
frequency
curve based on the statistics
of the systematic
Figure 12-2 is the preliminary
frequency
curve based
record.
on the computed mean and standard deviation
and a weighted
skew of 0.1 (based on a generalized
skew of -0.3 from Plate I).
Step 2

Check for

Outliers.

The station
skew is between + 0.4; therefore,
the tests for
both high outliers
and low oaliers
are based on the systematic
record statistics
before any adjustments
are made. From
Appendix 4, the KN for a sample size of 39 is 2.671.
The high
xH

outlier

threshold

(QH) is computed

by Equation

7:

= si;+ KNS
= 3.5553

QH = antilog

f 2.671(.4642)
(4.7952)

12-15

= 4.7952
= 62400 cfs

(12-17)

rti 0

Observed

-Preliminary
(Systematic
weighted

Annual

Peaks

Frequency
Curve
record
with
skew)

#XEDANCE

PR~BADilTY

Figure

12-2

Preliminary
Frequency
Curve for
Floyd River
at James,
Iowa
Example
L

12-16

Example

2 - Adjusting

for

a High

Outlier

(continued)

The 1953 value of 71500 exceeds this value.

Information
from local
residents
indicates
that the 1953 event is known to be the largest
event since 1892; therefore,
this event will
be treated
as a high
outlier.
If such information
was not available,
comparisons
with
nearby stations
may have been desirable.
The low-outlier

threshold

xL

QL
There
Step

are

Recompute

x - KNS

3.5553

antilog

no values
the

(QL)

is computed

- 2.671(.4642)

below

this

threshold

Use historic
Application
statistics

(12-18)

207 cfs

value.

statistics.

Mean Logarithm
Standard
Deviation
Skew Coefficient
Years
-

8a:

2.3154

(2.3154)

The 1953 value is deleted

and the
remaining
systematic
record:

Step 4

by Equation

data

statistics

statistics

and plotting

of the procedures
in Appendix
to be adjusted
by incorporation

The historic
period
(H) is 1892-1973
and the number of low values
excluded

(2)

The systematic
or 38 years.

(3)

There

(4)

Compute

(N)

is one event

(Z)

weighting

factor

14 =
=

the

is

positions.

6 allows
of the

(1)

period

from

3.5212
0.4177
-0.0949
38

of logs
of logs

to modify

recomputed

the computed
historic
data.

or 82 years
(L) is zero.

1935-1973

known to be the
(W) by Equation

(with
largest

1953 deleted)
in 82 years.

6-l:

E
82-l
38 -+ 0

= 2.13158
12-17

(12-19)

Example 2 - Adjusting

for

a High Outlier

Compute adjusted

mean by Equation

6-2b:

WNM-I- cXz

M =
x

(continued)

H-WL
M =

3.5212

WNM =

285.2173

cxz =

4.8543
290.0716

'L
M =

290.0716/(82-O)

Compute adjusted
:2 =

= 3.5375

standard

W(N-l)S2

deviation

(12-20)

by Equation

6-3b:

%L2
%2
WN(M-M) SC (Xz- M)

H-WL-1

s =
W(N-l)S2

.4177

13.7604

%L2
WN(M-M)
=

.0215

:2
c(Xz-PI)

1.7340
15.5159

15*515g
%
s

Jg,fj

(12-21)

.4377

Compute adjusted

First

82-O-l

compute

skew:

adjusted

skew on basis
12-18

of record

by Equation

6-4b:

Example

2 - Adjusting

for

a High

Outlier

(continued)

H - WL

'L
+ 3W(N-l)(M-M)S

G =

(H-&l)(H-WL-2)'
%3
f WN(M-M)

G =

+X(X,

IL3
- M)

-0.0949

W(N-1
)(N-2)S3G=

-.5168

3W(N-l)(M-;)S2

-.6729

-a0004

2.2833
1 .a932

%3
WN(M-M)

H
(12-22)

= *I509
(H-WL-l)(H-WL-;)33
G =
Next

compute

.1509

(1.0932)

weighted

.1650

skew:

For this example,

a generalized
skew of -0.3 is determined
from
Plate
I.
Plate
I has a stated
mean-square
error
of 6.302.
Interpolating
in Table I, the mean-square
error
of the station
skew,
based on H of 82 years,
is 0.073.
The weighted
skew is computed by
use of Equation
5:

G, = .302(.1650)
.302 f

GW

= 0.1

(rounded

+ +073(-.3)
.073
to nearest

12-19

o 0745
0

tenth)

(12-23)

Example
Step

2 - Adjusting

5 - Compute
For

the

iii,=
For

the

for

High

adjusted

largest

Outlier

plotting

event

(continued)

positions

(Equation

historic

data.

6-6):

1
succeeding

events

(Equation

i;; = W E - (W-l)(Z
-I- 0.5)
d
- (2.1316-1111
m2 = 2.1316(2)
= 2.5658
For the

for

Weibull

pp= -L

Distribution

6-7):
* .5)

(12-24)

a = 0; therefore,

by Equation

6-8

(100)

H+l
P"p = -

(100)

= 1.20

(12-25)

(100)

= 3.09

(12-26)

82+1

PT2 = y

Exceedance
probabilities
Equation
12-26 by 100.

are

computed

by dividing

values

obtained

from

TABLE 12-6
COMPUTATION OF PLOTTING POSITIONS

Year

Event
Number

Weighted
Order

71500
1 .oooo
1
1 .oooo
1953
__I____----_-----_--_____________I______------------------------------------20600
2.1316
2
2.5658
1962
17300
2e1316
3
4.6974
1969
15100
2,1316
4
6.8290
1960
8.9606
13900
2.1316
5
1952
11.0922
13400
2.1316
6
1971
2.1316
8'
13.2238
1951
8320
15.3554
7500
2.1316
1965
2.1316
lo"
17.4870
1944
7440
1966
7170
2.1316
19.6186
Only

the

first

10 values

are

shown for

this

12-20

example

Weibull
Plottino
Position
Percent
Exceedance
Chance
Probability
p"p

66

1.20

.0120

3.09
5.66
8.23
10.80

-0309
.0566
.0823
.1080

13.36

.1336
.1593

15.93
18.50
21.07
23.64

.1850
.2107
.2364

Example
Step

2 - Adjusting

6 - Compute

the

for

a High

frequency

Outlier

(continued)

curve.
TABLE 12-7

COMPUTATION OF FREQUENCY CURVE COORDINATES

KG,A
P

for

G, = 0.1

log

cfs

.99

-2.25258

2,5515

356

.90

-1.27037

2.9815

958

.50

-0.01662

3.5302

3390

.lO

1.29178

4.1029

12700

.05

1.67279

4.2697

18600

.02

2.10697

4.4597

28800

.Ol

2.39961

4.5878

38700

.005

2.66965

4.7060

50800

.002

2.99978

4.8504

70900

The final

frequency

curve

is plotted

12-a

on Figure

12-3.

Observed
Peaks with
Weighted
Plotting
-Final

Frequency

II
II
II

I
I

II

11

I
I1

Positions

II

Illltt

'

Curve

I
I

I
I

II
II
I,

III
III
I,,,

II

II

CEEDANCE

PRO.BAllilTI-

Figure
Final
Floyd

Frequency
River
at
Example

12-22

12-3
Curve

for

James, Iowa
2

EXAMPLE 3
TESTING AND ADJUSTING

a.

Station

FOR A LOW OUTLIER

Description

Back Creek

near

Jones

Springs,

West Virginia

USGS Gaging Station:

01-6140
Lat: 39030'43",
long:
78oO2'15"
Drainage
Area:
243 sq. mi.
Annual Peaks Available:
1929-31,
b.
Step

Computational
1 -

Procedures

Compute

the

statistics

of

The detailed

computations

computations

are :

Mean Logarithm
Standard
Deviation
Skew Coefficient
Years
At this

point

frequency
Figure

the

curve
12-4

is

the

generalized

2 -

Check for
As the

the

KN for

systematic

record.
the

may be interested

on the

statistics

preliminary

and a weighted
from

Plate

in seeing
of the

frequency

deviation

results

of

the

3,722O
0.2804
-0.7311
38

of logs
of logs

skew of 0.5

curve

the

systematic
based

skew of

preliminary
record.

on the
-0.2

computed

(based

on a

I).

outliers.

computed

detecting

the

have been omitted;

analyst

based

mean and standard

Step

1939-1973

skew coefficient

possible
a sample

is

low outliers
size

of

less

than

is made first.
38 is 2.661.

12-23

-0.4,

the

test

From Appendix

for
4,

Observed

-Preliminary
(Systematic

Annual

Peaks

Frequency
Curve
record
with

ICEEDANCE

?i?~DABl;lTY~

Figure

Back

Example

12-24

12-4

Preliminary
Frequency
Curve for
Creek nr. Jones Springs,
3

W. VA.

Example

3 - Testing

and Adjusting

The low outlier

XL

-ii

3.7220

antilog

The 1969 event

and will

Step

- DelPte

threshold

QL

- Check for

of

536 cfs

computations

Step

- Compute

in Step

(.2804)

is below

2.9759

946 cfs

the

8a:

(12-27)

threshold

and recompute

value

of 946 cfs

the

statistics.

3.7488
0.2296
0.6311
37

of logs
of logs

threshold

the

is computed

3 and the
threshold

the

size

be 22,760
of

12-25

based

on the

No recorded

1 and 2 for

the

threshold.)

curve.

is computed
by the

cfs

37 events.

(See Examples

frequency

curve

modified

to

high-outlier

conditional

frequency

3 and then

sample
value.

to determine

and adjust

A conditional

by Equation

as a low outlier.

in Step
exceed

(continued)

outliers.

The high-outlier

events

computed

(2.9759)

low outlier(s)

high

statistics

is

2.661

Mean Logarithm
Standard
Deviation
Skew Coefficient
Years

Step

a Low Outlier

- KNS

be treated

the

for

conditional

based

on the
probability

statistics
adjustment

Example

3 - Testing

and Adjusting

(Appendix

a Low Outlier

The skew coefficient

5).

in computation.
'L
P

for

has been rounded

The adjustment

N/n

37/38

ratio

(continued)

computed

from

to 0.6

for

Equation

0.9737

ease

5-la

(12-28)

TABLE 12-8
COMPUTATION OF CONDITIONAL

FREQUENCY CURVE COORDINATES

Kc"d
pd
.99
.90
.50
.lO
.05
.02
.Ol
.005
.002

for

G = 0.6

-1.88029
-1.20028
-0.09945
1.32850
1.79701
2.35931
2.75514
3.13232
3.60872

log

3.3171
3.4732
3.7260
4.0538
4.1614
4.2905
4.3814
4.4680
4.5774

The conditional
frequency
curve% along
curve,
is plotted
on Figure
12-5.

12-26

is:

Adjusted
Exceed? se
Probabr r sty
WPd)

cfs

.9639
.876
.487
,097
.049
.0195
* 0097
s0049
.0019

2080
2970
5320
11300
14500
19500
24100
29400
37800
with

the

adjusted

frequency

11

-Conditional
-

-Frequency
Probability

I
I

I
I

II
II

III
III

II

III,,,

Frequency

Curve

Curve with Conditional

Adjustment

CEEDANCE

PRqlABlLlTY

Figure
Adjusted
Back Creek

Frequency
nr. Jones
Example

12-27

12-5
Curves
Springs,
3

for
W. VA,

Example

3 - Testing

Step

and Adjusting

Compute

the

synthetic

The statistics
The use of
with

of

4.01
4.10
4.50
Next,

adjusted

(continued)

frequency

statistics
Type

III

from

= 23880

cfs

= 11210

cfs

compute

GS

the

discharges

a Low Outlier

statistics.

synthetic

a log-Pearson

and Q.5,

for

curve

provides
shape.

the

adjusted

are

unknown.

a frequency

First
curve

determine
on Figure

curve
the

Q~,,,Q~,,

12-5.

5230 cfs
the

= -2.50

synthetic

+ 3.12

skew coefficient

by Equation

5-3.

1w~4~,,/4~,,)
'"g(Q.10'Q.50)

= -2.50

+ 3.12

= -2.50

+ 3.12

:;;;;;

0.5948

12-28

(12-29)

Example

3 - Testing
Compute

sS

s,

Compute

Step

- Compute

and Adjusting
the

synthetic

a Low Outlier

standard

log

.6595/2.8546

synthetic

mean by Equation

log

(Qs5,)

log

(5230)

- (-.09945)(.2310)

3.7185

Gw =

(W-30)

0.2310

the

5-4.

(23880/5230)/~.75514-(-.09945)]

+ .0230

weighted
error

5-5.

I(s50(Ss)
(12-31)

= 3.7415

skew coefficient.
of

the

on n = 38 and using

GW

by Equation

o~(Q~o,/Q~50~/(K,ol-K~50>

the

(continued)

deviation

The mean-square
based

for

station

Gs for

skew,

0.6

tenth)

to

nearest

12-29

Table

1,

is 0.183

.302(0.5948)
+ .183(.5)
.302 + .183
(rounded

from

o 55go
.

(12-32)

Example
Step

3 - Testing

8 - Compute

and Adjusting
the

final

for

frequency

a Low Outlier

(continued)

curve.

TABLE 12-9
COMPUTATION OF FREQUENCY CURVE COORDINATES

for

P
.99
.90
.50
.lO
.05
.02
"01
.005
,002
The final
Note:

frequency

G, = 0.6

-1.88029
-1.20028
-0.09945
1.32850
1.79701
2.35931
2.75514
3.13232
3.60872
curve

log

cfs

3.3072
3.4642
3.7185
4.0484
4.1566
4.2865
4.3780
4.4651
4.5751
is plotted

2030
2910
5230
11200
14300
19300
23900
29200
37600
on Figure

12-6

A value of 22,000 cfs was estimated

for 1936 on the basis of data
from another
site.
This flow value could be treated
as historic
data and analyzed
by the producers
described
in Appendix
6.
As
these computations
are for illustrative
purposes
only,
the remaining
analysis
was not made.

12-30

0
-Final

Observed

Annual

Peaks

Frequency

Curve

:CEEDANCL

PROBABi,TI

Figure

12-6

Final
Frequency
Curve for
Back Creek nr. Jones Springs,
Example

12-31

W. VA,

EXAMPLE 4
ADJUSTING

a.

Station

FOR ZERO FLOOD YEARS

Description

Orestimba

Creek

near

Newman,

California

USGS Gaging,Station:
11-2745
Lat:
37O19 Ol", long:
12107'39"
Drainage
Area:
134 sq. mi.
Annual Peaks Available:
1932-1973
b.
Step

Computational
1 - Eliminate
There

Step

Procedures
zero

are

2 - Compute

flood

6 years

the

statistics

with

years.
zero

of

flood

the

events,

non-zero

Mean Logarithm
Standard
Deviation
of logs
Skew Coefficient
of logs
Years (Non-Zero
Events)

Step

3 - Check the

conditional

frequency

leaving

36 non-zero

events.

events.
3.0786
0.6443
-0.8360
36

curve

for

outliers.

Because the computed skew coefficient

is less than -0.4,
the test for
detecting
possible
low outliers
is made first.
Based on 36 years,
the
(See Example 3 for low-outlier
low-outlier
threshold
is 23.9 cfs.
threshold
computational
procedure.)
The 1955 event of 16 cfs is
below the threshold
value;
therefore,
the event will
be treated
as a
low-outlier
and the statistics
recomputed.
Mean Logarithm
Standard
Deviation
of logs
Skew Coefficient
of logs
Years (Zero and low
outliers
deleted)

12-32

3.1321
0.5665
-0.4396
35

Example

4 - Adjusting

Step 4 - Check

for

for
high

Zero

Flood

Years

(continued)

outliers

The high outlier

threshold
is computed to be 41,770 cfs based on the
statistics
in Step 3 and the sample size of 35 events.
No recorded
events
exceed the threshold
value.
(See examples 1 and 2 for the
computations
to determine
the high-outlier
threshold.)
Step

5 - Compute

and adjust

the

conditional

frequency

curve.

A conditional
frequency
curve is computed based on the statistics
in step 3 and then adjusted
by the conditional
probability
adjustment
(Appendix
5).
The skew coefficient
has been rounded to -0.4 for ease
in computation.
The adjustment
ratio
is 35/42 = 0.83333.
TABLE 12-10
COMPUTATION OF CONDITIONAL

FREQUENCY CURVE COORDINATES

KG,P
for G = -0.4

'd

log

cfs

.99

-2.61539

1.6505

.90

-1.31671

2.3862

243

.750

.50

0.06651

3.1698

1480

.417

.lO

1.23114

3.8295

6750

.083

.05

1.52357

3.9952

98900

.042

.02

1.83361

4.1708

14800

.017

.Ol

2.02933

4.2817

19100

.0083

.005

2.20092

4.3789

23900

.0042

.002

2.39942

4.4914

31000

.0017

Both

frequency

curves

are

plotted

on Figure

12-33

12-7.

44.7

Adjusted
Exceedance
Probability
(P.P,)

.825

Observed
Based
-

Peaks
on 36 Years

Conditional

Frequency
Curve
zero and low-outlier

(Without

events)
Frequency
Curve with
Conditional
Probabilitv
Adjustment

II
II

I
I
I

I
I
I

I
I
I
I

I
I
I
I,

I
Irl

ICEEDANCE

PRO.BABlLlTY

Figure
Adjusted
Orestimba

Example

12-34

12-7

Frequency
Curves for
Creek nr. Newman, CA
4

Example
Step

4 - Adjusting

6 - Compute
First
curve

the

Q.50

the

Flood

Years

the Q,O,,Q.,O,
12-7.
17940

synthetic

skew coefficient

Gs

-0.5

(rounded

Compute

the

G
Y

the

adjusted

1060 cfs

+ 3.12

sS

from

6000 cfs

-2.50

ss

and Q-50 discharges

cfs

the

(continued)

statistics.

G,

Compute

7 - Compute

Q.10

Compute

Zero

synthetic

determine
on Figure
Q.01

Step

for

synthetic

!og(17g40/6000)
log(6000/1060)

log(l7940/1060)/(1.95472

0.6564

2.9708

the weighted

-0.5287

by Equation

5-5.

(12-35)

by Equation

A generalized
skew of -0.3 is determined
from Plate
the mean-square
error
of the station
skew is 0.163.
Gw

.302(-.529)
+ .163(-.3)
.302 + .163

Gw

-0.4

tenth)

(rounded

(12-34)

- (.08302)(.6564)

skew coefficient

to nearest

12-35

(12-33)

5-4.

- .08302)

mean by Equation

log(lO60)

5-3.

tenth)

deviation

to nearest

standard

synthetic

by Equation

-o 4487
.

5.
I

From Table
(12-36)

I,

Example
Step

4 - Adjusting

8 - Compute

for

the

final

Zero

Flood

frequency

Years

(continued)

curve.

TABLE 12-11
COMPUTATION OF FREQUENCY CURVE ORDINATES

KG
a
wp
P

.99
.90
.50
.lO
.05
.02
.Ol
.005
.002

for

Gw = -0.4

log

-2.61539
-1.31671
0.06651
1.23114
1.52357
1.83361
2.02933
2.20092
2.39942

1.2541
2.1065
3.0145
3.7789
3.9709
4.1744
4.3029
4.4155
4.5458

Q
cfs
17.9
128
1030
6010
9350
14900
20100
26000
35100

This frequency
curve is plotted
on Figure
12-8.
The adjusted
frequency
derived
in Step 4 is also shown on Figure
12-8.
As the generalized
skew
may have been determined
from stations
with much different
characteristics
from the zero flood
record station,
judgment
is required
to determine
the
most reasonable
frequency
curve.

12-36

0
-

Observed
Based
Final

Peaks
on 42 Years

Frequency
Frequency

Curve
Curve

at

CEEDANCE

PROIAIILITI

Figure

12-8

Frequency
Curves for
Orestimba
Creek nr. Newman,
Example

12-37

CA

Appendix

13

COMPUTERPROGRAM
+
Programs

have been developed

distribution

from systematically

a single

station

included

for

distort
adjust

--

and other

managing

the curve

be obtained

curve

that

large

annual

of higher

to represent

small

Special
flows

flows.

routines

(outliers)

An option

expected

procedures

Type III

maximum streamflows

known events.

and very

incorporate

from either

compute a log-Pearson

recorded

flows

in the range

the computed

agency programs

zero

that

are
that

probability.

recommended by this

Guide may

of the following:

Phone:

Phone:

programs.

860-6879

is no specific
Other

may have developed

federal

to

Copies of

Hydrologic
Engineering
Center
U.S, Army Corps of Engineers
609 2nd Street,
Suite I
Davis, CA 95616

There

would

is included

Chief Hydrologist
U,S. Geological
Survey, WRD
National
Center, Mail Stop 437
Reston, VA 22092
(703)

at

recommendation
and state

individual

to utilize

agencies

programs

(916)

as well

to suit

13-l

their

756-1104
these

particular

as private
specific

computer

organizations
needs.

Appendix

14

"FLOOD FLOW FREQUENCYTECHNIQUES"

REPORT SUMMARY
Following
report

is a summary of "Flood

by Leo R. Beard,

Resources,

Technical

The University

Resources

Research

and a majority

Director,

are taken

of Texas at Austin

of the Water Resources

Water Resources

Council

Research.

where gage records

methods currently
than

The study

to select

selection
United

of long-record
States

programs

for

statistical
tions

and development
testing

and fitting
outliers,

zero flow

years,

partial-duration
testing

trends,

frequency

flood

recommendations

of background

determinations.
testing,

with

Work Group.

14-l

regard
for

alternate

distribufor
with

maximum and
adjustment,
long-term

to each procedure
being

basic

stations

probability

the gufdes

computer

criteria

annual

consistent

(b)

in the

eight

treating

of expected
data exh%bft

material

testing

between

technical

and current

of alternate

relationships

at locations
and consistency

streamflows

for

develop-

the best

for

including

of procedures

(g) testing
if

(c)

analysis

(d) testing

testing

to determine

and procedures

of

for

analysis

of the literature

procedures3

(e) testing
(f)

a basis

of data management and analysis

techniques,

series,

and (i)

development

for

of the Office

frequency

data of natural

alternate

methods

managing

(h)

methods

Committee,

reliability

flow

of

the general

Hydrology

incorporate

greater

in flood

station

of and under

frequency

would

(a) a review

candidate

in Water Resources

was to provide

which

been available

included:

Research

flood

of Water

from the report,

the auspices

for

in Water

Much of the text

Flow Frequency,

known and would yield

has heretofore

practice

for

The purpose

are available

Council.

Research

the Office

directly

through

ment by the Work Group of a guide

for

for

at the request

of the Work Group on Flood

Techniques,"

Center

and the Water Resources

was made at the Center

The University
guidance

Flow Frequency

of Texas at Austin,

of the exhibits

The study

tested

developed

and

by the

Data
In all,

300 stations

unregulated.
records

Record length

longer

than

best feasible
to include
year.

were used in the testing.

40 years.

coverage

30 years

The stations

of drainage

a substantial

Table

exceeded

with

most stations

were selected

area size

with

the number of stations

no flow
by size

having

to give

and geographic

number of stations

14-1 lists

Flows were essentially

the

location
for

and

an entire

and geographic

zone.
Split

Record Testing
A primary

function

concern

and fitting

from annual

technique

peak flow

used in the fitting

selecting
split

it

fit

is very

to elim-lnate

of annual

events

Frequency

estimates

happens

irregularities.
procedure,

past events

criterion

occur

for

Consequently,

of actual

from the fitting

would

frequencies

to the data

frequencies,

conditions

particularly

for

application

computation

in practice,
to eliminate

was divided

one half

of any general
should

and

Goodness of
methods whose

trend

adequately

and found

happened

that

criteria

in the reserved

record
reserved

and verification
however*

or operator

criteria

simulates

of water

are representative

in order

possibly

practical

exist,
situations

of each other,

of a record

and tested

half,
is complicated,

half

also

because what

is subject

data cannot

to sampling

be used as a silmple,

must be probabilistic,
faced

projects,

nor what future

14-2

might

independent

condltIons

resource

two halves,

the other

simulate

in the other

of verification

into

and even for

were made from each half

Consequently,

target

designer,

flood

maximum flows

were tested

what actually

Development

test

a valid

that

numbers

procedure

as annual

accurate

flow

is not necessarily

of a record

the effect

splitting

actually

flood

of a function

be used,

odd sequence

against

estimates

Goodness of fit
estimates

events

of a mathematical

poor.

Each record

This

best

was used to simulate

as "future"

can nevertheless

was selection

that

best

a portion

fit

using

that

testing

by reserving

data,
process

a method

record

using

of the study

The

by the planner,,

who knows neither

events

will

be.

that

The ultimate
to estimate

objective

the most likely

observed

data,

decision

is formulated.

their

attendant

to forecasting
of testing

but rather

Frequency

Computation

Basic methods

agencies
corresponds
programs
1.
that
for

study

should

distribution

generated
for

which

a
and

an intermediate

step

record

be more rigorous

technique

and direct

than

tests.

tested

in this

study

were

and the WRC Work Group on Flood Flow Frequency

of the literature
those

and in the attached

that

(35).

and experience

were tested

in the various

are listed

number of the methods

below.

Numbering

in the computer

tables.

Type III

(LP3).

The technique

The mean, standard

each data set are computed

in accordance

used for

deviation,
with

this

is

and skew coefficients

the following

equations:

(14-l)
S2 =
!J

the

functions

the split

techniques

to the identification

in

events

is not

Methods

represented;

described

process

that

is ordinarily

goodness-of-fit

by the author

Log-Pearson

future

Accordingly,

and fitting

review

distribution

forecast

criteria

events.

used in this

careful

to best

estimation

Use of theoretical

future
theoretical

after

theoretical

reliability

alternative

selected

of any statistical

C X2 -

(14-Z)

(CX)'/N
N-l

N2CX3 - 3NCXCX' + 2(CX)3

(N-2)9
N(N-1)

(14-3)

where
X = logarithm

of peak flow

N = number of items

in the data

set

x = mean logarithm
s=

standard

deviation

9 = skew coefficient

114-3

of logarithms
of logarithms

Flow logarithms
following

to these

statistics

by use of the

equation:

specified

normal

exceedance

distribution

approximate
2.
identical

(14-4)

X =X+kS
probabilities

Exceedance
for

are related

for

probabilities

routines

transform
Log Normal

to Pearson
(LN).

This

to exceedance
available
3.

(a value

in computer

relates

rocal

of exceedance

estimates

function

are obtained

(36).

libraries

except

applies),

and the

in reference

that

(31).
function

the skew coefficient

and values

by use of the normal

is the Fisher-Tippett

linearly

with

probability

of k are related

distribution

transform

using

The initial

logarithms).

(location

procedures

extreme-value

function,

the log of the log of the recip-

(natural

of the mode and slope

reference

by use of the

method uses a 2-parameter

of zero

This

magnitude

made by iteration

given

and scale

described

estimates

Maximum likelihood
parameters)

by Harter

of the location

are

and Moore in
and scale

as follows:
M= x - 0.45005
B=

Magnitudes

of k

libraries.

Gumbel (G).

which

of k and values

in computer

deviates

III

probabilities

values

are calculated

available

to the log-Pearson

is not computed

specified

are related

(14-5)

.7797 s
to these

(14-6)
statistics

as follows:

X = M + B(-ln(-1nP))

(14-7)

where
M = mode (location
B = slope

(scale

statistic)
statistic)

X = magnitude
P = exceedance
S = standard

14-4

probability
deviatlon

of flows

statistics

Some of the computer

the Central
4.

routines

Technical

Unit

of the Soil

Log Gumbel (LG).

technique

except

5.

that

used in this

This

below,

to zero.

The shape parameter

NHrlund's

(37)

expansion

the following

except

1 + /l

that

estimate

(lnq

are used.
to the 3-parameter

the location
directly

parameter

is set

by solution

of the maximum likelihood

+ $

to the Gumbel

is identical

is determined

as an approximate

CL

is identical

This

by

Service.

(base 10) of the flows

Gamma (62).

Gamma method described

Conservation

technique

logarithms

Two-parameter

method were furnished

equation

of

whfch

gives

of ~1:

- i ElnQ)

(14-8)
-

Aa

V - $.ZlnQ)

4 (In
where

8=

average

annual

peak flow

N = number of items

in the data set

Q = peak flow
Aa = correction
f3 is estimated

factor

as follows:
(14-9)

6.

Three-parameter

statistics

for

procedures

described

if

the

the calculated

to those
bound,

for

Gamma (632.

3-parameter

the

yp is initialized

of record,

Probabilities

using
for

directly
corresponding

by use of a library

the minimum flow

zero,

smaller

value

the

to specified

gamma routine.
14-5

of the lower

(38).

equations

than

the lower
value

bound is derived

Then the parateters

magnitudes

or

are identical

the .lowest

above replacing

Magnitudes

using

is zero,

the statistics
Otherwise,

slfghtly

in reference
using

is accomplished

Gamma distribution.

at a value

criteria

If
than

and the maximum likelihood

by iteration
are solved

(38).

bound is less

2-parameter

of maximum likelihood

Gamma distribution

in reference
lower

Computation

Q with

are computed

corresponding

a and S
Q-y,
directly

to specified

probabilities
7,

are computed
Regional

Log-Pearson

to the log-Pearson
taken

Type

III method,

Type

from Figure

Regionalized

by iteration

14-l

instead

using

III (LPR).
except

of using

skew coefficients

the inverse
This

that

solution.

method is identical

the skew coefficient

the computed

is

skew coefficient.

were furnished

by the U.S.

Geological

Gumbel (BLI).

This

is the same as

Survey.
8.
for

the Gumbel method,

(BLIE)

are used for

likelihood
for

Invariant

Best Linear

testad

that

the function

estimates

this

except
(MLE).

best

linear

invariant

statistics

instead

An automatl"c

censoring

method only,
for

so there are no altenative

Statistics
are computed
method.

this

method

estimates

of the maximum
routine

outlier
techniques
as follows:

M =. C(X(I)U(N,J,I))
B

is used

(14-10)

(14-U)

= c(X(I).V(N,J,I))

where
U = coefficient

UMANN described

in reference

(39)

V=

BMANN described

in qeference

(39)

coefficient

J = number of outliers
deleted plus I
I = order number of flows arranged %"nascending-magnitude
order
N=
Since weighting
only

for

treated
years
cients

coefficients

sample sizes
by this

outliers

sets

by chance

sequential
less

of this

than

(or half

chronological

of weighting

that

tests

might

show that
technjque

are not

average

that

coeffiif

Up to two

array.

Each one is

extreme

would

of the BLIE statistics.

are contained

25

be obtained

were avallable.

a value

1 time IO on the basis

censoring

otherwise

end of the flow

study

of more than

and weighted

coefficients

at the upper

in this

samples

records)

groups

of coefficients

are censored

removed If

from 10 to 25, &year

and records

into

used in lieu

as censored,
U and V were made available

ranging

method,

are divided

more complete

Details

sample size

in refer-

occur

Weighting

ence (40).

coefficients

method were furnished

tion

and most of the routines

by the Central

Technical

Untt

used in this

of the Soil

Conserva-

Service.

Outliers
Outliers

were defined

whose ratio
negative)

to the next
direction

extreme

value

keeping

b. reducing

d.

for

the value

to the product

the value

root

of that

study

the ratio

values
or

of the next

most

value.
consisted

of

of the second largest

to eighth

to the product
ratio,

as extreme

in the same (positive

outliers

as is,

discarding

should

handling

the value

largest

event

event,

of the second largest

event

and

the value.

In the cases of outliers

(c)

value
than

of the second largest

and the square

of this

most extreme

tested

reducing

C.

most extreme

to the eighth

and the ratio

purpose

is more extreme

The techniques
a.

for

at the low end,

be changed

the words

largest

in (b) and

to smallest.

Zero Flow
Two techniques
years

of no flow
(a)

estimates,

1 percent

purposes

Ities

all

zeros

that

with

some complete

to all

values

for

amount from subsequent

and multdplying

by the ratlo

number of values,

described
A secondary

estimated

exceedance

of the number of non-zero

This

in reference

Partial-Duration
annual

stations

of the mean magnitude

and subtracting

of the remaining

the total

handling

and

(b) removing
cies

for

as follows:

Adding

computation

were tested

is the procedure

frequen-

values

of combining

to

probabil-

(27).

Series
concern

maximum flow

of the study

frequencies

Because a partial-duration
specified

magnitude#

definition

normally

it

was the relationship

and partfal-duration
series

is necessary

consists
to define

depends on the application

14-7

between

flow
of all
separate

frequencies,

events

above a

events.

The

of the frequency

study

as

well

as the hydrologic

separate

events

characteristics

were arbitrarily

as many days as five

drainage

area,

with

below 75 percent
This

plus

of the stream.

defined

the natural

logarithm

the requirement

of the lower

is considered

as events

that

separated

study

by at least

of the square miles

intermediate

of the two separate

representative

For this

of separation

flows

must drop

maximum daily
criteria

of

flows.

appropriate

for

many applications.
Maximum daily

flows

there

were insufficient

flows

for

believe
flows

events
that

than

readily

smaller

for

daily

frequency

for

annual

complete

daily

flows

at each station

total

number of years

peak

corresponding

records

at each station

0.3,

0.4,

positions

frequencies

0.5,

0.6,

event

and 0.7 are

by linear

interpolation

(M/(n+l))

for

the annual
flows

number of independent

above each magnitude

station.

Ratios

were averaged

for

derived

as to

of partial-duration

the total

ratios

to develop

to annual-maximum

0.2,

plotting

In order

maximum
by the

of partial-duration

all

for

and dividing
stations

certain

are

to

in each USGS

theoretical

conditions

(9).

Probability

The expected
probabilities

Estimation
probability

of all

is defined

magnitude

be made from successive

flow

magnitude,

it

is considered

or frequency

planning

and management

use.

a probability

for

samples

of probability
Is also

as the average

estimates

estimate
It

to

event

of any assumptions

of 0.1,

at that

frequencies

frequencies.

independent

by counting

zone and compared with

specified

event

magnitudes

simply

might

for

to annual

established

that

be different

of partial-duration

Corresponding

Expected

is no reason

as a ratio

maximum events.

by Langbein

would

There

peak

and partial-duration

probability

event

because

the maximum annual

between expected

annual

maximum.

relationship

probabilities

established

of the study,

data on instantaneous

the annual

relationships

functions,

exceedance

available

part

between

at selected

partial-duration

this

flows.

was expressed

frequencies

than

the frequency

The relationship
series

were used for

estimate
14-8

any specified
of a specified

flood
site.

frequency
For any

to be the most appropriate

of future
that

of the true

flows

for

is theoretically

water

resources
easy to

verify,

because

the observed

number of stations
estimates

approach

that

record

expected

Similar

techniques

distribution

functions.

expected

should

normal

To do this

it

was

be used in the

average

used in this

making the accuracy

of the various

bias

that
methods

as described

for

the other

an empirical
a calibration
normal

This

in (21).

transform

is derived

was determined

transform

adjustment,

probabilities

for

empirical

transform

are the main basis

for

for

threoretical

constant

in estimating

study.

tests

has been developed

distribution

by the theoretical

removed the observed

300 stations

probabilities

Consequently,

when multiplied

Trends

Accordingly,

estimates

are not available

each distribution.

adequacy

or frequency

tests.

samples drawn from a Gaussian

which,

data at a large

probability

increases.

probability

A method of computing

for

in reserved

the computed

as the number of stations

considered
split

should

frequencies

the

was used in

judging

the relative

tests.

and Cycles
There

(longer

is some question

than

nature

1 year)

as to whether

exist

in nature

can be used to improve

specific

long-term

such that

forecasts

trends

knowledge

of flood

flow

times

in the future.

As a part

of this

1 autocorrelation

coefficients

of annual

peak flows

were computed.
data,

there

stations.

If

should

trends

or cycles

be a net positive

A statistically

exist

of their

frequencies

research
for

all

average

lag

stations
part

autocorrelation

positive

for

project,

in any substantial

average

significant

and cycles

for

of the
all

autocorrelation

was not found.

Accuracy

and Consistency

Criteria
theoretical
computed
that

used in judging
distribution

Accuracy

tests

frequencies

occurred

Tests
the adequacy

of each method for

fitting

were as follows:
consisted

of the following

dn one-half

In the reserved

the record

comparisons
with

between

frequencies

of events

data.

a.

Standard

deviation

of observed

reserved

data for

magnitude

estimates

frequencies
corresponding

(by count)

fn

to exceedance

probabilities

of 0.001,

record

This

used.

individual

observed

0.1,

and 0.5 computed

is the standard

stations

average

0.01,

that
bias

would

error

occur

of a frequency

if

in each group

from the part

a correction

of stations

estimate

at

is made for

for

of the
the

each selected

frequency

and method.
b.

Root-mean-square

plotting

position

event

(M/(n+l))

in a half

frequency

record

of that

standard

error

difference

of the largest,

and the computed

respective

event

of a frequency

any bias adjustment

between expected

for

upper decile

expected

half.

exceedance

This

at individual

each method and for

and median

probability

in the other

estimate

probability

is the

stations

the frequency

without

of each selected

event.
Root-mean-square

c.
computed

probability

plotting

position

turn)

record.

except

frequencies

are not

tests

that

involved

of a record

to the

and median event

(in

is similar

are biased

to that

toward

the following

in each half

the two record

halves

median events,

in turn.
that

half

of the

of the preceding

predicting

small

favored.

Root-mean-square

of estimates

1.0 and the ratio

upper decile

criterion

methods

frequencies

a.

between

in the opposite

This

that

Consistency
computed

of flow

of the largest,

in a half

paragraph

difference

of the record

difference

for

full

would

between

record

This

comparisons
with

is an indicator

be made with

the total

computed

extreme,

between
record.

probabilities

largest,

upper decile

of the relative

various

from

random samples

and

unliformity
for

the same

location.
b.

Root-mean-square

the larger
record

computed

extreme,

is essentially
that
value

largest,

The extreme

event

It

to the

decile

predicting

be recognized
or larger

small

multiplied

for

that

for

halves

and median events,

criterion,

by the square
the full
sampling

errors

to

full
This
methods

favored.

is an arbitrary
root

of the ratio

record,

the same sample size

14-10

that

are not

tests

for

in turn.

except

frequencies

used in the consistency

to the medfan event

should

are as large

upper

largest

of the smaller

from the two record

the same as the preceding

toward

the largest

of 1.0 minus the ratio

probabilities

are biased
equal

value

in the reserved
as are sampling

data
errors

of

of computed
estimates

values.

Similarly,

based on opposite

Consequently,

a great

due to sampling

method that

is biased

standard

method that
The latter
ation

error

percentage

data point

tests.

in order
data,

to reduce
Further,

in comparison
the bias

with

is not removed.

errors.

Accordingly,

consl"der-

for

each of the eight

:

methods

estimate

alternate

without

procedures

any expected

were printed

the

too low may have a

if

of the analyses.
curve

consistency

frequencies

frequencies,

for

the complete

probability

adjustment,

out,

of Distributions

Table
stations

14-2 shows for

each method and each USGS zone the number of

where an observed

discharge.

With

that

about

This

comparison

14,200

14 observed

(method

exeeeded the computed

station-years
events

indicates
2),

discharge

that

would
the

and log-Pearson

of record,

it

exceed true
log-Pearson
Type III

might

be expected

1,000"year

magnitudes,

Type III
with

l,OOO-year

(method

generalized

l),

log

skew (method

are the most accurate,

Table

reserved
0.01,

14-3 shows average

portions

0.1,

observed
It

used for

frequencfes

high

frequency

computed

and the plotted

7),

toward

are comparable

is necessary

means of evaluating

results,

Evaluation

halves

estimating

may have smaller

As a further

errors

in the reserved

of estimating

is biased

must be a component

normal

errors

toward

of the average

record

record

number of tests

uncertainty
small

sampling

of half

records

is difficult

from their

of the results
This

for

methods with

closest

computed

to those

probabilities

deviations

(accuracy

for

data.

generalized

and the smallest

that

test

a) of the

frequency.

Figure

14-2 shows

estimates

which

the log normal

skew have observed

standard

in the
of 0.001,

each computed

from these

indicates

(by count)

aids

and logfrequencies

deviations

except

method 4,
Table

tests

computed

the 0.01 probabiljty

comparison

Pearson Type III

frequencies

averages

to draw conclusions

in comparison.

for

for

and 0,5 and the standard

frequencies

a plotting

observed

14-4 shows the average

b and c.

Results

results

are not definitive,

for

all

stations

but again

of accuracy

the log normal

(method

2) and log-Pearson

show results
in figure

as favorable

displays
racy

the results

results.
the

skew (method

method as illustrated

From Figure
generalized

for

test

test

a.

tests.

appears

skew yields

14-4

test

results

The consistency

from or more definitive

14-4 it

Figure

than

the accu-

that

the log-Pearson

Type III

considerably

more consistent

results

Testing

14-6 shows results

consistency

tests

these

show that

tests

and log-Pearson
techniques

for

for

the four
for

Type III

all

stations

different

with

can greatly

computed

computed

frequencies

or lower

end are generally

and low for

for

between

half

[log

normal

(method

7)1,

of outliers

Outliers

having

for

the first

three

this

is caused primarily

end.

Average
at the

outlier

Values

that

upper

techniques

observed

by outliers

plotting

position

Average

computed

probability,

method a

0.059

Average

computed

probability,

method b

0,050

Average

computed

probability,

method c

0.045

Average

computed

probability,

method d

0.038

more discriminatory

appears

to be the most logical

Results

of Zero Flow Testings

of maximum flow

outlier

tests

studies

and justifiable

14-7 shows the average

and consistency

at the

are as follows:

Average

Table

end

the fourth.

is considered

Until

at the

at the lower

outliers

high

2)

outlier

at the upper-end.

records

of

no discrimination

treatment

arrays.

frequencies

all

Results

Unfortunately,

ends of the frequency

increase

techniques.

skew (method

a and b are most favorable.

upper and lower

and

methods

generalized
tests

of the accuracy

outlier

the.favorable

was made in the verification

accuracy

7)

log normal.

Table

lower

for

different

of Outlier

It

generalized

of the consistency

graphically

substantially

Results

as any other

14-5 shows the results

method with
than

with

14-3.

Table
are not

Type III

for

for

14-12

all

0.042

are made, method a

to use.

stations

the two different

of the results

of

zero flow

techniques.

These test
normal
7)1,

(method

indicate

b is slightly

better

of Partial-Duration

Results
zones and that
theoretical

there

occur

average

values.

events

tend

to cluster

other

years,

larger

than

a large

14-8 would

the average

values

theoretical
that

values

that

after

smoothing

relation

to the Langbein

Expected

Probability

adjustment
equal

It will

zone and for

however,

different
flood

study)

than

the
large

to cluster

be expected

in

to be

any given

should

region

be used for
they

(not

stations

in

have a constant

function,

expected

probability

theoretical

to compute average

each zone are shown in Tables

that

these

vary

considerably

intervals.

due to vagaries
shown on the last

methods.

for

such that

in order

for

for

tend

the

Results

exceedance

each distribution

various

computed

the normal

be noted

is believed

the loo-year
for

Adjustment

observed

is small,

such that

events

14-8 would

were

(random)

to be smaller

and small

the

values

per year

are not independent

different

than

The theoretical

be expected

the values

must be multiplied

and 14-11.

for

obtained

number of independent

theoretical

by which

to those

in values

'It

values.

zones as used in this

The ratios

14-8.

in Table

necessarily
region

are shown in Table

by Langbein.

in some years

It was concluded

[log

skew (method

zones is somewhat greater

If the events

theoretical

generalized

If the number of events

in Table

the

of all

that

each year.

values

methods

a.

studies

developed

based on the assumption

events

than

the favorable

with

is some variation

the average

values

for

Studies

of partial-duration

can be seen that

that
Type III

2) and log-Pearson

technique

Results

comparisons

the purpose

Much of this
of sampling.
Average
line

in Table

of comparing

accuracy

1.

Log-Pearson

2.

Log Normal

0.9

3.

Gumbel,

3.4

4,

Log Gumbel

5.

2-parameter

2.1

MLE

-1.2
gamma

3.4
14-13

14-9,

14-10,

from zone to
variation,
ratios
for

14-10 were adopted

These are as follows:

Type III

probabilities

and the

6.

3-parameter

7.

Regional

8.

Gumbel,

Results
normal
mately

of the study

bias

tabulation

indicates

that
distribution

shown range around

do not greatly

It is particularly

multiples.
loo-year

values)

Expected
Sample
Avg.

adjustment

this

that

an adjustment

Probability

The

method,

skew

for

the

Coefficients
only

one

in terms of standard-error

significant

indicate

Adjustment

the most reliable

factor
Ratios

data

near 1.0.
for

All

Zones

lOO-Yr

lOOO-Yr

Err.

Avg.

0.81

0*17

0.94

0.12

1.01

0.13

10

0.60

OS22

1,12

0.20

1.45

0.27

23

0.17

0.27

1.14

0.23

1.68

0.28

Results

of Test

Results
shown in Table
positive
cluster

for

Trends

autocorrelation,
would

observed

are not seriously

chance.

It is considered
in streams

studies
that

indicating
occur

shown are multiples

and it

Err.

Avg.

Std.

Err.

and Cycles

I autocorrelation
14-12.
It is apparent

coefficient,

process

Std.

of lag

more than

t values
tion

Std.

7)

approxi-

factor.

of 1.0 and, with

from it

lo-Yr

Size

for

value

depart

correspond

method wSth regional

probability

approximately

the theoretical

the log

skew (method

the adjustment

expected

applies

only

regional
should

Type III

log-Pearson

that

be the case l"f the distribution

influence

the theoretical

with

zero bias

of 1.0 as would
for

indicate

Type III

because

do not greatly

following

(the

portion

2) and log-Pearson

characteristics

1.1
5.7

of substantial

exception,

Type III

BLIE

to a coefficient

normal

2.3

log-Pearson

of this

(method

are free

gamma

there

a tendency

in a completely
of the standard

is obvious

that

different
that

used in this

to test

annual
study.

14-14

for

trends

is a tendency
for

flood

toward

years

random process.
error

extreme

from variations
peak flows

that
approximate

to
The

of the lag
correlation

are

I correlacoefficients

would

occur

a random

by

Conclusions
Although
there

split

record

are sufficient

Conclusions

clearcut

regional
C.

For methods
as recorded)

d.
flows

e.
this

in terms

methods

computed

for

less

frequencies.
method 2.

(retaining

the

of computed

weight

"b" (discarding

is slightly

to

to outliers.
superior

by the 300 stations

autocorrelated;

use in frequency

zero
to zero

frequencies

should

be used rather
significance

computed

from theoretical

adjusted

to reflect

are related
regions;
than

more frequent

thus,

a single

extreme

adjustment

made from samples

for
need not

events

event

empirical

regional

theoretical
that

in the classical

number of cases are to average

estimates

records

to annual

is the conclusion

functions

For the recommended method,

thus,

selected

analysis.

in different

Of particular

relationship.

frequencies
manner must be

if

frequencies

the same as observed

equal

to the theoretical

drawn from a normal

computed
frequencies.
adjustment

population

is approxi-

correct.
Of interest

record

techniques

to be definitive.
reserved
case if

from a research
require
This

data obscure
curve-fitting

degree than
tial

than

of ratio

technique

frequencies)

differently

relationships

mately

flow

as represented

Partial-duration

in a great

give

future

results
"a"

Type III

"a."

be continuous

for

that

are not substantially

frequencies

in estimating

is more accurate
2 and 7, zero

Streamflows

f.

biased

technique

For methods

study

recommendations,

and method 7 (log-Pearson

2 and 7, outlier
than

technique

definite

somewhat more consistent

frequencies

and adjusting

flow

normal)

skew) are not greatly

Method 7 gives

observed

to support

as anticipated,

can be drawn are as follows:

b.
outlier

were not as definitive

results

Only method 2 (log

a.
with

that

results

standpoint

more than
study

300 records

showed that

the results
functions

is the finding
random variations

to greater
could

of about

reduce

degree

that

50 events

each

in the

than would

uncertainty

split

be the

to a greater

has been possible,

In essence,

then,

regardless

uncertainty

in frequency

of the methodology

estimates

from station

employed,

substan-

data will

exist,

but the log-Pearson

produce

unbiased

is employed,
methods

estimates

and will

when the

reduce

for

Future

It is considered
comprehensive
lower

that

study

that

Treatment

hurricane

this

to expected

will

probability
other

of flood

flow

recorded

data as well
basin

f.
for

progressively

locations
reliable

at the upper and

composed of different

for

in a given

great

types

and those

of events

from snowmelt,

or

differences

in frequency

character-

region;

of systematic
estimates

procedures

for

regional

and applications

as to locations
of procedures

conditions,

Development

of outliers

floods;

frequency

Development

phase of a more

include

from rainfall

explanation

Development

modified

is an initial

curve;

resulting

among streams

e.

study

of sequences

flows

Physical

d.

without
for

to locations

recorded

deriving

coordination
with

data;

frequency

curves

for

such as by urbanization;

of a step-by-step
with

various

results

the amount of effort

expended

cl- Preparation
use in training

of a text

for

skew coefficients

as much as or more than

in the treatment

and nonhurricane

c.

curves

adjustment

uncertainty

should

Differentiation

such as flood

istics

regional

Study

ends of a frequency
b.

method with

tested.

Recommendations

a,

III

type

and practical

procedure

for

amounts and types

can be obtained
is increased;
on flood

flow

application,

deriving

frequency

of data such that

on a consistent

basis

and
frequency

determinations

as

FIGURE

GENERALIZED

SKEW

14-l

COEFFICIENTS

STREAMFLOW

OF
LOGARITHMS

14-l

ANNUAL

MAXIMUM

FIGURE
ACCURACY

COMPARISON

FOR

0.01

14-2
PROBABILITY

ESTIMATE

(TABLE

14-3)

8
L
i

.06
I

/
POSSIBLE COMPARISObj

LINE
I
I

.05
3
, %-

METHOD

NUMBER

.04

.03

.02

.Ol

0.00

-L

0.00

.Ol

.02

.03

AVERAGE

OBSERVED
FOR

0.01

.04

PROBABILITY
COMPUTED

14-18

.05

.06

IN
PROBABILITY

TABLE

.07

14-3

.08

FIGURE
ACCURACY

COMPARISON

14-3
FOR

(TABLE

MAXIMUM

14-4,

TEST

OBSERVED

FLOW

8)

.09

.08

zw

ot
Y
LL
Y
0

.05

0.00

-I
0.00

.Ol

.02

AVERAGE

.03

OBSERVED
FOR

0.0

.oa

.05

PROBABILITY
COMPUTED
14-l

.06

IN
PROBABILITY

TABLE

.07

14-3

.08

FIGURE
CONSISTENCY

COMPARISON
(TABLE

--

14-4
FOR

14-5,

MAXIMUM
TEST

OBSERVED

FLOW

A)

.2

-?hI

i
0.00

.Ol

.02

AVERAGE

.03

OBSERVED
FOR

0.01

.04

PROBABILITY
COMPUTED

14-20

.05

IN
PROBABILITY

.06

.07

TABLE

14-3

.08

Table
Numbers of Verification
USGS

14-1

Stations
Drainage

by Zones and Area Size

area category

(sq.

ml.)

Total

200-1000

1ooo+

10

27

12

24

16

25

3
6

15

14

20

13

24

12

21

11

23

18

10

12

11

13

12

17

13

10

17

14

15

15

16

12

13

13

53

73

142

32

300

ZONE

o-25

- 4

Total

*Zero-flow

25-200

stations

(zones

a, 10 & 11 only)

J-4-21

Table

14-2

NUMBEROF STATIONS WHEREONE OR MORE OBSERVEDFLOOD EVENTS

EXCEEDSTHE lOOO-YR FLOW COMPUTEDFROM COMPLETE RECORD
STATIONYEARS OF

METHOD
6

10

26

10

19

22

1%

19

1124

18

852

'87

969

10

19

920

16

10

636

3
1

10

11

594

11

12

777

13

911

14

14

761

15

15

120

16
*

637

12

495

12

14

18

77

68

56

20

253

RECORD

1414

2
1

1074

1223

703

990

ZONE
---i-

TOTAL

14,200

Based on the 14,200

about

14 observed

*Zero-flow

station-years
events

would

of record,
exceed

stations

14-22

the true

5.

it

might

be expected

lOOO-year

magnitudes.

that

Table 14-3
STANDARD DEVIATION COMPARISONS
AVERAGE FOR ZONES 1 TO 16
COMPUTED
PROBABILITY
.OOl
.Ol
.l
.5
-001
.Ol
.1
.5
Note:
record
lation

.0105
.0041
-0232
-0153
.1088
.1007
.5090
.5149
STANDARD DEVIATION OF
.0290
-0134
-0430
.029
.086
.084
-132
-131

METHOD
3
4
5
6
7
8
AVERAGE OBSERVED PROBABILITIES
.0109
.OOOl
.OllO
.0092
.0045
.0009
.0315
.0023
.0309
.0244
.0170
.0015
.1219
.0707
-1152
.1047
.1020
.0029
.4576
.6152
.4713
.4950
.5108
.0037
OBSERVED PROBABILITIES
FOR SPECIFIED COMPUTED PROBABILITIES
.0244
.0025
.0239
.0218
.0150
.0222
,045
.OlO
,043
.039
-032
.035
.089
.074
,089
.084
.084
.067
.142
.133
,133
.141
.130
.123

Averages
and standard
deviations
are of observed
frequencies
corresponding
to computed mangitudes
based on half
records,
to averages
indicate
more reliable
estimates.

in the reserved
portion
Low standard
deviations

of
in

each
re-

Table
Evaluation
Accuracy
Test

b--Root

Tests

of Alternative

b and cb Average

mean square

computed

14-4

difference

probability

Methods
Values,

All

Stations

between

plotting

half

of record.

in other

position

and

Method
Maximum

1.
.062

2
,060

3
,067

4
,056

5
,070

!i
,069

L
.061

8
,061

Decile

.084

.080

.097

,063

e098

* 094

.081

,082

Median

.254

.105

,657

-193

,518

.295

,120

,727

Test

c--Root

mean square

computed

probability

to plotting
perfect

difference

bewteen

of flow

position,

A zero

1.0 and ratio

in opposite
value

half

would

of

of record

indicate

forecast.
Method

Maxtmum

1
.53

2,
.51

2
.56

Q
.45

ii
.56

5.
.56

L
.51

!i
.59

Decile

*37

.34

.38

a27

.37

937

,34

.40

Median

.40

.12

065

.19

.59

.44

.14

.52

Table
Evaluation
Consistency
Test

a--Root

Tests

decile

of Alternative

Methods

a and b, Average

mean square

the two record

14-5

difference
halves

Values,

between

for

full

and median events.

All

Stations

computed

record

probabilities

extreme,

A zero value

would

largest,
indicate

from
upper
perfect

consistency.
Method

Event
Extreme

1
.003

2
.006

3
.OOl

a
,010

5
e.001

s
,002

L
.003

s
-002

Maximum

,023

,019

,008

,016

,008

.OlO

.OlO

.012

Upper Decile

,072

,047

,043

.025

0037

.033

.025

,048

Median

,119

,076

,072

.047

e049

-045

0041

.131

Test

b--Root

mean square

to the larger
for

full

events.

record

value

of (1.0

computed

probabilities

extremeb

A zero value

minus

largest,

would

the ratio

of the smaller

from the two record

upper decile

indicate

perfect

halves)

and median
consistency.

Method
Event
Extreme

1
a87

2
.54

a
,46

I
026

5
039

!i
.35

a
e29

6
975

Maximum

.74

*45

.$I

421

.34

,30

,24

.72

Upper Decile

.50

.32

.31

.16

.24

.21

.17

.58

Median

.21

.14

.12

010

,08

.08

.oa

,24

Table
Evaluation
Average

14-6

of Outlier
Values, All

Techniques
Stations
Method

Accuracy Test b
Outlier
Technique

1.

,061

.062

,071

,057

.074

.073

,062

.056

,055

.060

.063

.062

.055

,052

,050

.054

,053
,048

,057

,055

.051

,047

.045

,048

,044

,051

.050

.045

1
.53

2
.55

2
.57

!?.
.47

5.
.58

6
.58

2"
.54

.57

.5g

.59

.49

.62

-60

.58

.58

.60

452

.64

.63

.60

.65

.61
.65

.64

.38

.68

.65

.64

2
.005
,004

2
,001
,001

4
.009
-008

2.
. 000

a
.002
.002

2
.002
,002

Accuracy

Test

Outlier
Technique

Consistency
Test a
Outlier
Technique
a

1
.002

,002

.003

.ooo

,007

,003

.003
.003

. 000
* 000

.ooo

.007

D000

-002
-002

.002
.OOl

1
.87

2
056

P
.46

!I
027

5
.39

f!
.36

I
.30

.86

$56

.45

.28

.38

.35

.85
.88

056
.59

.45
.45

.29

.38

.31

.38

.35
.35

.30
.30

Consistency

Test b

Outlier
Techniques
a

d
A zero value would
Method 8 includes
not included

indicate
its

in these

unique

perfect

consistency.

technique

tests,

14-26

.32

for

outliers

and was, therefore,

Table 14-7
Evaluation
of Zero Flow Techniques
Average Values, All Stations
Accuracy

Test b

Technique
a
b
Accuracy

1
,057
.064

2
.057
.060

s.
*OS9
.070

Method
9
5
.057
.062
.057
,068

6
,055
.061

z
*OS9

.061

Test c

Technique
a
b
Consistency

1
646
.51

.32
.30

.59
.59

Method
4
5
.32
-40
.30
.40

6
.340
.4f

2.
.32
.31

Test a
Method

Technique
a
b
Consistency
Technique
a
b

L
.007
.007

.012
.008

.ooo
e000

5
-001
.ooo

6
.ooo
.OOl

7
.006
.004

Method
4
5
.21
.39
.40
.19

6
.34
.38

1
.24
.23

9
.014
.012

Test b
1
.89
.86

2
.e3
.43

3
-44
.44

Method 8 was not tested because logarithms

are not used in its
fitting
computations
and therefore
zero flows are not a problem.

14-27

Table

14-8

Summary of Partial-Duration

Ratios

Partial-duration
for

annual-event

frequencies
frequencies

of

2m.w..
3

A 4

A 5

.6

.7

.475

,641

.844

1.10

.209

-328
.353

.517

c759

1.001

1.30

.094

.206

,368

.507

.664

.862

1.18

4 (8 sta)

,095

.218

.341

.535

.702

D903

1.21

5 (17 sta)

.093

.213

.355

,510

.702

,928

1.34

6 (16 sta)

.134

.267

.393

,575

.774

1.008

1.33

7 (9 sta)

.099

.248

,412

,598

"826

1.077

1.42

8 (12 sta)

,082

,211

.343

.525

.803

1.083

1.52

9 (15 sta)

,106

.234

.385

.553

.765

10 (12 sta)

.108

.248

.410

,588

.776

1.022

1.34

11 (12 sta)

.094

.230

.389

,577

.836

1.138

1.50

12 (12 sta)

.103

.228

-352

.500

,710

.943

1.21

13 (16 sta)

,095

.224

0372

.562

,768

0986

1.30

14 (14 sta)

.lOO

.226

,371

.532

.709

* 929

1.22

15 (3 sta)

,099

.194

,845

1.05

.106

,232

.410
.522

'16 (13 sta)

.301
.355

.696

,912

1.27

Average

,099

.243

,366

.532

.733

e964

1.28

Langbein

,105

.223

-356

.510

,693

.917

11.20

Zone
1 (21 sta)

L 1
.094

L 2
.203

2 (17 sta)

,093

3 (19 sta)

Note:

Data limited

to 226 stations

14-28

originally

o982

609

selected

for

the study.

1.26

TABLE 14-9
ADJUSTMEMT RATIOS FOR lo-YEAR FLOOD
SAMPLE
SIZE
METHOD
5-YR
lo-YR
l/2-REC

ZONE 1
1

.54

.38

.75

l 45

1.21

076

1.11

ZONE 2
METHOD
5-YR

1
.4a

27 STATIONS
3
4
1.02
2.21

2
.42

3
1.06

1.01

.94

1.91

l/P-REC

1.33

1.33

2.76

lo-YR
l/2-REC

ZONE 3
1

.28

-1 .a5

.95
2.01

.37
1.01

.34
1.03

4.56
4.09

AVG l/2

4
.64

5
1.03

.68

.a0

5.70

.49

.54

7.14

25 STATIONS
3
4

1.41
-98

.a1
.14

1.80
1.65

.oo
-1.88

1.87
1.17

1.05
-.52
.45

.94
.02

ZONE 5
1
.55

l/2-REC

.40
.a1
ZONE 6

METHOD
5-YR

1
.80

2
.35

3
1.20

096
.21
AVG l/2

4
.a5

AVG l/2
6
1.79

15 STATIONS

-.50

-1.85

1.3-i

1.95

IO-YR

1.90

1.02

7
.41

1.60

1.92

6
093

RECORD = 22 YRS

-1.58

1.32

METHOD
5-YR

METHOD
5-YR
10"YR

.57

1.41

ZONE 4

l/E-REC

-.27
-1.04

RECORD= 26 YRS
7
8

.a2

24 STATIONS

lo-YR

METHOD
5-YR

.29

AVG l/2
5
6

5
1.29

1.15

-.a5

-.Ol

-.54

1.63

-3.07

1.63

-1.85

1.01
.39

5.39
4,ao

RECORD = 23 YRS

AVG l/2
6

.94
-.45
.25

a
-1.85
3.68
5.57

RECORD = 25 YRS
7
a

.98

1.40
-.96
2.91
-3.61
24 STATIONS

.61
1.42

2
.36

3
1.19

5
1.11

6
.95

7
.45

a
-9 .a5

.la

2.26

1.78

.96

.33

5.64

1.94

-07
AVG l/2
6

4
.15

.a8

.15

-.03
-,40

1.03

.46

1.4D

.12

20 STATIONS
3
4

RECORD = 24 YRS
7
a

.42
699
AVG l/2

.47

-1.85

.19
7*37
.67
6023
RECORD = 23 YRS

lo-YR

1.43

l/P-REC

-.45

METHOD

1.08
ZONE 7
1

5-YR

1.15

1.19

1.69

1.29

1.62

1.59

1.29

-1.85

lo-YR

1.58

1.36

2.34

.12

1.99

1.62

1.57

5.78

l/2-REC

1.97

1.00

2.45

2.87

.92

1.17

7.11

METHOD

ZONE 8
1

5-YR
IO-YR
l/L-REC

.a9

2
.a9

-.66

-1.02

-.13

-.a7

-.98

.2.94
-3.93
21 STATIONS
3
4

-.74

23 STATIONS
3
4
9.71

.79

.29

-2.04

2.28

-3.08

14-29

AVG l/2
6

1.41

1.36

-.35

-.43

.74

.66

-.04
6.14
RECORD= 20 YRS
7
a

RECORD= 21 YRS
7
a
.79

-1.85

-1.02

4.52

-.a7

7.88

TABLE 14-9 CONTINUED

ZONE 9
METHOD

18 STATIONS

5-YR

1
1.38

lo-YR

1.95

1.54

2.54

.45

-.36

.97

l/L-REC
METHOD
5-YR
IO-YR
l/L-REC

2
1.02

3
2.05

ZONE 10
1

6
1.78

.75

2.49

2.22

1.69

.45

-.07

-027

-3.36

-.79

-.80

-.41

-.83

-.03
.08

-.42
-1.27

.90
1.24

-1.16
-5.10

5
-.43
.71
.5B

13 STATIONS

RECORD = 25 YRS

5
1.96

12 STATIONS
3
4

ZONE 11
METHOD

AVG l/2

4
.96

AVG l/2
6

8
-1.85
6.76
4.07

RECORD = 26 YRS
7
8

-.43

-.77

-1.85

.35
-.27

-.22
-1.27

4.24
2.97

AVG l/2
4

7
1.10

RECORD = 23 YRS

5-YR

1.29

1.21

1.89

1.20

1.93

1.75

1.11

-1.85

lo-YR

1.11

1.03

2.21

.04

1.87

1.25

1,03

6.78

l/2-REC

.04
-,23
ZONE 12
1
2

1.99
-2.93
17 STATIONS
3
4

1.20

1.34

1.34

1.51

METHOD
5-YR
lo-YR
l/P-REC
METHOD
5-YR

.73

.79
.41
.19
-.31
ZONE 13

.57

.86
-.45
.54
-2.94
17 STATIONS

1.20
AVG l/2
5
6
1.03

.92
.92

.80

-.44
-.35
AVG l/2

IO-YR
l/2-REC

1.27
.26
-.31

1.16
.22
-1.52

1.65
.88
.21

.96
-.83
-4.89

METHOD
5-YR

ZONE 14
1
2
1.72
1.65

15 STATIONS
3
4
2.12
1.61

5
2.19

lo-YR

2.60

2.50

3.17

1.88

2.82

1.87

-51

.61

1.83

-1.47

1.30

.29

l/2-REC

5
1.77
.67
.I7

7
1.19
038
-1.12

METHOD
5-YR

AVG l/2
5
6
2.66
2.28

IO-YR

1.27

1*27

1.58

1.27

1.58

l/IREC

3.29

3.29

3.29

2.79

3.29

METHOD
5-YR

lo-YR
l/P-REC

069
.58
1.41

.76
-42
-07

METHOD

ALL ZONES
1
2

1.03
.66
.83
-.21
1.68
-3.43
287 STATIONS
3

1058
1.90
AVG l/2
5
6

1.09
.76
1.25

8
-1.85
4.60
2,88

AVG 1;/2 RECORD = 25 YRS

6
7
8
2,oo
1.65
-1.85

3 STATIONS
3
4
2.74
2.55

13 STATIONS
3
4

-1.85

.57
4.06
-.19
2.81
RECORD = 26 YRS

1.52
.42
-.97

ZONE 15
1
2
2.47
2.47

ZONE 16
1

-.23
5.32
RECORD = 23 YRS
7
8

1.05
.07
.64
AVG l/2
6

2.56
075

6.80
5.22

RECORD = 20 YRS
7
8
2.28
-1.85
1,27

2,65

3.29

6.33

RECORD = 24 YRS
a
8
.75
.42
.07

-1.85
4.24
5.29

RECORD = 23 YRS
7
8

5-YR
IO-YR

.94

.79

1.38

1.37

1.21

.81

-1.85

.87

.52

1.52

-.29

1.26

.72

.60

5.27

l/2-REC

.77

.04

1.93

-2.66

1.34

.40

.17

5.36

Values

shown are ratios

distribution
samples
puted 0.1 probability
data.

See note

table

by which

.71

the theoretical

adJustment

for

Gausslan-

must be multiplied
In order to convert from the comto average observed probabilities
in the reserved
14-11.
u-30

TABLE 14-10
ADJUSTMENT RATIOS FOR lDO-YEAR FLOOD
SAMPLE
AVG l/2 RECORD = 26 YRS
5
B
6
7

SIZE
METHOD
5-YR

ZONE 1
1
1.35

1.11

1.27

.39

1.61

10.YR
l/2-REC

1.50

1.10

2.05

a.25

2.83
ZONE 2
1

2.84

3.90
-1.06
24 STATIONS

2.42
4.89

METHOD
5-YR
lo-YR
l/P-REC
METHOD

27 STATIONS
3
4

.91

,79

1.05

.31

1.44

1.40

1.00
ZONE 3

1.08

2.48
3.69

.63
-,B2

25 STATIONS
3
4

1.12
1.73

3.67
AVG l/2
5
6

1.27
2.41
2e97

2.43

2.90

1.41

METHOD

3.36
-1.12
15 STATIONS
3
4

3.71

ZONE 4
1

5-YR
lo-YR

1.67
,57

1.45

2.27
1.07

l/P-REC

1.86
ZONE 5
1

.4B

1.03

.64
.57
.21

METHOD

1.22
2.97
ZONE 6
1

5-YR

1.15

,67

1.02

lo-YR
l/L-REC

2.30
1.20

,55
-,23

1.67
3.22

lo-YR
l/L-REC

ZONE 7
METHOD
5-YR
10.YR
l/2-REC
METHOD
5-YR
101YR
l/2-REC

2
1.07

1.18
3.10

1.09
.47

*04
-,27
-1.24

3.92
1.80
23 STATIONS
3

,27

2.08

1.30

a14

1.59
4.36

-a32

4
eo1
-.35
-1,13

14-31

.46

.BB
AVG l/2
5
6

1.19
lo27
2.97
5

1.12
1.09
2.39
AVG l/2
6

1.17

.8B

1.78

1.10
.79

2.45

4
.28
0.19

2.76
AVG l/2
6
2.02

2.83

3
2.23
2.66

.57
,82

AVG l/2

21 STATIONS

1
1.04

ZONE 8
1

1.37
e24
1.42
-.29
4.38
-1.24
24 STATIONS

7
1.32

1.15

METHOD
5-YR

5.40
1.37
,14
7.16
RECORD = 24 YRS

1.62

2.42

l/E-REC

.59

RECORD= 22 YRS
7
8
-.25
.63

2.84

lo-YR

.56
-.4B
1.54
-1.15
20 STATIONS
3
4

3.42
5.28

1.29

1.76

1.48
.35

.73
1.66

6
1.86

1.18

-.25

5
2.05

S-YR

-.04

2.07
2.46
AVG l/2

1
1.80

.41

1.13

.BB

5
2020

6
2016

B
0.25
4.79
5.53

2.30
RECORD* 23 YRS
7
B
1.64
.42

-.25
, 1.60

1.03
3.81
RECORD- 25 YRS
8
7
-025
.B2
.80
5,65
1.68
7.25
RECORD- 23 YRS
B
7
.76
.66
046

-.25
4.43
5,09

RECORD = 20 YRS
7
8
1620
-025

1.53
5.40
2.20
8.33
2.29
1074
AVG 112 RECORD- 21 YRS
B
5
6
7
-.25
1.66
1,52
a27

2.54
2,99

l"15
2.16

.93

.I4

4.17

2.16

-.32

8.49

TABLE 14-10
ZONE 9
METHOO

18 STATIONS

5-YR

1.07

1.33

1.90

lo-YR

2.45

2.23

l/2-REC

1.07

.39

ZONE 10
METHOO
5-YR
lo-YR

-.15

3.29
-.27
ZONE 11

METHOO
5-YR
IO-YR
l/P-REC

3.21

.90

2.90

-1.72
3

4.39

.66

4.49

.96

-.59

1.06

.75

1.63
-1.79
13 STATIONS

2.42

1.32
AVG l/2

1.79

2,41

T.51

4.14

.30

.79

5040

1.16
1055

l/2-REC

3.77

1.65

2.12

2
.87
.37
-.07

-1.08

-1.33

3.43

1.28

6.64

5
1.56
2.27

2.57

AVG l/2
6

1.63
1.41
2.57

AVG l/2
6
2.21

2.92

2.58

3.53

1.98

4.40

METHOD

3.10
AVG l/2
6

5-YR

2*09

2.24

2.24

1.24

2.76

.26

.26

.26

-069

1.80

1.80

.93

-1.31

METHOD
5-YR
IO-YR
l/L-REC

.61
1.87

*2
.55
1.23

4.21
1.17
ALL ZONES

$90
1.63

3.96
-1.27
287 STATIONS

1.16
1.64

.90
1.03

1.45
2.01

l/P-REC

2.12

.87

3.40

.18
-.59

-1.23

0.01 probablltty
See note table

to average
14-11.

observed
14-32

1.82

1.04
.60
.81

-.25
3.28
2.69

RECORD = 25 YRS
7
8
1.44
-.25
2.32

5.16

2.80
5.37
RECORD = 20 YRS
7
8
-.25

4.84

1.84

.26

1.72

4.37

3,16

.93

.93

AVG l/2
5
6
1.30
1.83

1.22
.99

4.41

2.90
AVG l/2
5

RECORD - 24 YRS
7
8
.62
1.33

-.25
3.64

2.13
4.46
RECORD = 23 YRS

1.66
2.20

1.45
1.62

.94
1.12

0.25
4.25

3.35

2.30

1.14

5.66

Values shown are ratlos

by which the theoretical
dlstrlbutlon
samples must be multlplled
In order
puted
data.

1.86

1.50

4
032
-007

-625
2060

1.98

13 STATIONS
3
4

METHOD
5-YR
IO-YR

1.19
2.11

RECORD = 26 YRS
7
8

1.26
1.25
1.61

5
2.43

9,77

RECORD = 23 YRS
7
8

1.19
l,D4

15 STATIONS
3
4
1.79
065

ZONE 16
1

050

AVG l/2
6

.23

l/P-REC

2.43

3.76
-1.52
3 STATIONS
3
4

IO-YR

3.76

ZONE 14
1
2
1.54
1.44
2.22

2.55

3.05

2.11
2.80
ZONE 15
1
2

lo-YR
l/P-REC

.15

-.27
4.40
RECORD = 23 YRS
-025

4.39

.50
-.28
-1.81

8
-.25

.66

17 STATIONS
3
4
1.12
1.27
2.20

7
-.06

1.54

17 STATIONS
3
4
.44
613

RECORD = 26 YRS

1.58

.17

1010
1.93

ZONE 13
1

4
.ll

1.81
1.99

5-YR

2.57

2.38
6

5-YR
lo-YR

METHOD

3.55

3.78

.29

ZONE 12
1

lo-YR
l/P-REC

3.75

AVG l/2
4

8
-.25

7
1.50

.29

.7p

-.25

RECORD = 25 YRS

2.11

.27

.68

1.63
.58
1.01

5
2.11

METHOD

METHOD
5-YR

4
.72

2
-.lO

.21

l/2-REC

AVG l/2

12 STATIONS

1
-.lO

CONTINUED

adjustment
to convert

probabll!tles

for Gaussianfrom the com-

In the reserved

TABLE 14-11
ADJUSTMENT RATIOS FOR lOOO-YEAR FLOOD
SAMPLE
SIZE

ZONE 1

METHOD

27 STATIONS

AK
4

RECORD s 26 YRS

l/2

5-YR

2.03

1.10

1.19

.2l

%a12

1.44

.85

-.04

lo-YR

2.30

-88

2.21

-.14

2.98

1.87

.52

4.06

l/P-REC

5.01
ZONE 2

1.66

8.54

METHOD

4.13

6.94
-.56
24 STATIONS
2

lo,11

8.16
AVG l/2

RECORD = 22 YRS

5-YR

1.31

.83

1.18

.15

1.57

1.35

.68

10"YR

1.98

2.85

3.85

.64

4.45

3.66

2.07

7..41

3.56

l-58

8.81

l/P-REC

la93

2.11

4.47

ZONE 3
METHOD

-.45

3.56

25 STATIONS
2

AVG l/2

-.04

RECORD s 24 YRS
7

5-YR

2.42

1.22

2.18

-.Ol

2.54

2.08

1.24

IO-YR

6.06
7.41

2.20
2.44

3.06
6.77

-.14
-.51

3.89
7.06

1.82
4.82

2.20

7.11

2,77

11.16

l/GREG

ZONE 4
METHOD
5-YR

1
1.88

15 STATIONS
2
1050

lo-YR

1.24

.54

l/L-REC

2.86

.80

ZONE 5
METHOD
5-YR

1
1.84

IO-YR

2,75

l/2-REC

5.51

5-YR

3.99

l/2-REC

2.88

2.05

7
1.63

8
-,04

-.14

1.13

.36

.7P

1.33

-.48

3.60

3.60

2.40

2.81

AVG l/2

.49

RECORD = 25 YRS

l-92

1,45

2.90

-014

2,43

2.00

.91

6.02

5.76

-.52

5089

5.30

3.22

11.70

24 STATIONS
2
057
1.38

AVG l/2

1.08

.07

1.54

1.73

-006

2.47

-.48
3

RECORD = 23 YRS

.79

-.04

2.33

1.57

2006

1.63

I"12
1.24

4.53
8.92

AVG l/2
4

-,04

1.13

21 SBATIONS
2

1,32

656

.61

947
2.11

2.48

1.36

ZONE 7
METHOD

RECORD = 23 YRS

.94

IO-YR

4
.30

20 STATIONS

1.39

1.91

3
1.46

ZONE 6
METHOD

AVG l/2

-004

RECORD a 20 YRS
7

S-YR

1.19

082

9.91

.-I9

2.18

1.89

11040

-004

IO-YR
l/2-REC

2.33
5.99

096
1.48

3.58
5.36

0'13
.I6

3.25
3.90

2,15
3.90

'1.63

6.52

ZONE 8
METHOD
5-YR
IO-YR
l/P-REC

23 STATIONS

.83

.09

1.28

-.Ol

.83

.83

.14

-.04

2.79

,42

2068

-.14

1.78

1.78

842

5.90

2.70

.84

7.62

-.49

3.54

ZONE 9
1
5-YR
lD-YR
l/P-REC

AVG l/2

2.34
12.61
RECORD * 21 YRS

090
3.61
3.69

18 STATIONS
2

3.54
AVG l/2

1.32

13.61

RECORD = 25'YRS
7

1030

I,37

049

2.33

2.33

1.55

-.04

3059
.59

3.22
3.97

.42
-.53

6.86
2.68

5.85
1.04

3.90
1.07

6.24
6.92

14-33

TABLE 14-11
ZONE 10
METHOD
lo-YR
l/2-REC

12 STATIONS

5-YR

.02

-.04

.44

-.14

7.21

.27

-.04

.70

-.14

3.04

-.56

2
2.15

lo-YR

4.31

2.44

l/2-REC

1.74

.91

ZONE 12

.67

.43

-.14

1.95

1.95

0
9.04
3.79

.27
RECORD

6
1.78

4.50
23 YRS

2.13

5.95

.72

5.06

3.58

1.90

10.41

6.38

-.46

5.01

4.24

,91

15.65

2.84

1*22

1.31

lo-YR

4.30

2.17

2.52

l/L-REC

8.58

.75

.75

4
045
010
-.46

5
2.03

RECORD - 23 YRS

1.27

4027

1.40

2.17

3.37

2.20

1.34

,75

4.59

AVG l/2

-.04

1.51

17 STATIONS
2

a94

AVG l/2

5-YR

7
-.04

.20

ZONE 13

6
.22

17 STATIONS

RECORD = 26 YRS

AVG l/2

1.01

l/2

5
.22

13 STATIONS

1.13

METHOD

.25

5-YR

METHOD

AK

ZONE 11
METHOD

CONTINUED

-.04

RECORD - 26 YRS

5-YR

1.89

1.21

1.11

1.92

1.79

1.21

-.04

lo-YR

1.27

.36

1.39

-.14

1.77

1.77

,53

3.56

2.83

-,57

3.65

.55

4.96

l/2-REC

4.01

-.57

ZONE 14
METHOD

.32

2.43

15 STATIONS

AVG l/2

RECORD = 25 YRS

5-YR

1.91

1.45

1.56

2.66

2.03

1.45

-.04

lo-YR

5.41

2.35

2.81

-.14

4.63

2.17

2.35

5.56

3.45

1.04

5.12

-.53

9.90

1.04

6.69

l/P-REC

ZONE 15
METHOD

.47

6.99

3 STATIONS

AVG l/2

RECORD = 20 YRS

5-YR

2.67

3.00

2.54

-,04

3.51

1.25

1.77

-,04

10"YR

-.14

-.14

-.14

-.14

1.87

1.87

-.14

-.14

l/2-REC

2.17

2.17

-.3B

-.3B

6,15

6.15

0.38

-;38

i3

ZONE 16
METHOD

5-YR

.69

lo-YR

4.02

l/2-REC

8.74

AVG l/2

062

1,15

-.04

1.56

3.05

-.I$

2.37

7.24

-051

ALL ZONES
METHOD

STATIONS
5
1.4D

069

-,04

3.90

1.97

2901

4.46

8.30

602%

3.76

7.24

AVG.1/2

1.18

287 STATIONS

RECORD m 24 YRS

RECORD . 23 YJ&

5-YR

1.60

.95

1.40

.21

1.89

1.54

1.01

-,04

101YR

a,13

1,40

2.66

.04

3.22

2.19

1,45

6.36

l/L-REC

4.66

1.49

4.81

4.99

4.02

1.68

8.80

Values

shown are

dlstribution
computed
served

ratios

samples
0.001

by which
must

probability

-.45
the

be multiplied
to average

data.

14-34

theoretical
In order
observed

adjustment

for

to convert

from

probabilities

In

Gaursianthe
the

re-

Table
Values

in table

14-11 CONTINUED

14-11 are obtained

as follows:

a.
Compute the magnitude corresponding
to a given
exceedance probability
for the best-fit
function.
that

Count proportion
b.
exceed this magnitude,
C.

Subtract

of values

the specified

d.
Compute the Gaussian
to the specified
probability.
size
d.

in remainder

probability
deviate

that

e.
Compute the expected probability
(record length used) and the Gaussian
f.

Subtract

g*

Divide

the specified

probability

of record

from b.
would

correspond

for the given sample

deviate determined
in
from e.

f by c.

*U.S.

14-35

GOVERNMENT

PRINTING

OFFICE:1983-

X91-614/209

GENERALIZED

SKEW COEFFICIENTS

OF ANNUAL

MAXIMUM STREAMFLOW LOGARITHMS*

The generalized
prefer

not

skew map was developed

to develop

was developed

from

detailed

studies

outlined

in Section

as more data

their

own generalized

readily

for

available

their

region

V,B-2.

It

stations.

are

These

drainage

areas

more years

equal

of

to or less

essentially

Periods

when the

natural

flow

by more than

using

lowest

annual

6 from

cient.
test

only

the

identify

where

determine

prediction

characteristics
affect

the
for

one-degree

groups

of

latitude

miles

relating

stream

through

flow

water

used.

from

by equation
the

for

one or more years

the

skew coefficients
not

used in defining

and longitude.

low outlier

No attempt

Preliminary

in areas

skew coeffi-

flood

covering

was

informa-

curve.

was developed

would

year

At 144 stations

to use historic

such relations

with

had 25 or

differed

not

gaging

files

that

used in computing

guide.

Averages

position.

2,972

of each frequent!;

in the

completed,
of annual

skew coefficients.

of 15 or more stations

quadrangles

square

peak

were

evaluation

that

are

used in computing

annual

be revised

on USGS tape

map of skew coefficients

indicated
isopleth

at

were

outiiers,

equations

I will

to be a low outlier

logarithmic

described

procedures

studies

likely

the

or to make a detailed
technique

the

skew coefficients

15 percent

high

ercouraged

Plate

peak discharge

years

and treat

averaging

that

3,000

and was not

the

using

peaks

about

remaining

The generalized

were

annual

14-1

and in computing

made to
tion,

figure

to make

are

annual

peak was judged

At 28 stations

was zero,

than

who

Users

available

unregulated

1973.
the

stations

users

The map

on skew coefficients

the

guide

skew relationships.

interest

logarithmic

is based
all

of

those

and more extensive

The map is of generalized

It

data.

is expected

become available

peak discharge.

for

using
attempts

the
to

to basin
appreciably
the
four

isopleths
or more

The average
degree

quadrangle

are also
degree

of

latitude

shown on the

of one-degree
for

skew coefficients

quadrangles

various

minimum

values

Average

gaging

to the

of quadrangles

shown by the

stations

and the

stations

selected

averages

of stations.

for

groups
one-

The averages

were used to establish

isopleths

of

for

by weighting

number

in each one-

number

skew coefficients

were computed

according

groups

all

and longitude

map.

quadrangles

for

the maximum and

and to position

the

intermediate

15 or more stations

with

lines.
Because
years

the

of record

stations

are

few tenths
values

average
is

subject

closely

from

skew for
to time

grouped,

some group

sampling

the

smoothed

averages.

of skew coefficient

about

error,
lines

especially
are

The standard

the

isopleth

25 or more
when the

allowed

to depart

deviation

line

of station

is about

0.55

nation-

wide.
Only

enough

interpolation

isopleths

between

The generalized
is the

average

for

are shown to define

isopleths

of 0.33

shown for

10 stations

in

part

for

the

in the
nine

remainder

*This

of Alaska

stations

was arbitrarily
shown for

that

the

average

conterminous
the

all

is the

of Hawaii

The coefficient

The average

skew

average

skew of 0.85

for

period

g@neraliZed
skew map was originally
utilizing
in 1976.
It has not been revised
Bulletin
17B.

States

sampled

in

view

may be too

of the

the

shown

at nine

to the maximum generalized

United

for

of 0.70

on skew coefficients

reduced
for

Linear

The generalized
Alaska

State.

is based
area.

shown for

stations.

southeastern

of the

Anchorage-,Fairbanks

coefficient
bility

that

of -0.05

gaging

coefficient

variations.

is recommended.

skew coefficient
30 stream

the

stations

these
skew
possi-

large.

prepared
for Bulletin
17 published
the techniques
recommended in