nreca water balance

7/25/2019 NRECA Water Balance

1/54

Hydrologic

estimates

forsmall hydroelectric

projects

NRECA Small

Decentralized

Hydropower

(SDH) Program


2/54

Hydrologic

estimates

for small

hydroelectric

projects

byDr.

Norman

H.Crawford

and

StevenM.

Thurin

Hydrocomp,

Inc.

September

1981

Prepared

for

NRECA

undera

Cooperative

Agreement

with

the

U.S.

Agency

for

International

Development

Small

Decentralized

Hydropower

Program

International

Programs

Division

NationalRural

Electric

Cooperative

Association

1800Massachusetts

Avenue

NW,

Washington,

DC 20036


3/54

Small

decentralized

hydropower

program

Thispublicationis

oneof aseries that

fosterstheeffective

use

ofsmall

decentralized

hydroelectricpowersystems.

Theseries is

published

by

the Small

DecentralizedHydropower

(SDH)

Program,

International

Programs Division,

NationalRural

Electric

CooperativeAssociation(NRECA).

NRECAoperates

the

SDt,Program

underthe termsof

Cooperative

AgreementAID/DSAN-CA-0226

with

the

Officeof

Energy,Science and

Technology

Bureau,

U.S.

Agencyfor

International

Development.

Under theagreement,

begun inMay 1980,

NRECAprovides a

broad

range

of technical assistance

to developing countries. NRECA

provides such

technical assistanceby--

Designing

and implementingregional

workshops in

Africa,Asia,

andLatinAmerica

Scopingand

managing

in-countryresourcesurveys

andsite

assessments

Providing engineering,design,

supervision,

andspecialized

assistance

Developing specialized

publications,such

as state-of-the-art

reports, inventoriesof

manufacturers, and

assessment

methodologies'

Conducting specialstudies

intosubjects

of

finance,

management,

and

evaluation

Providing training services

in such

topicsas

operation

and

maintenance,

resourceassessment,

managemert, and fabrication

Carrying

out

specializedservices, such

as

tours

ofU.S.

manufacturing

plants

Creatingspecialized

products, suchas

productive-useplans

for

energy

from

small decentralized

hydropower.

Formoreinformationon

the SDHProgram,

pleasecontact:

Information Specialist

SmallDecentralizedHydropower

Program

International

ProgramsDivision

NRECA

1800

Massachusetts

Avenue

NW

Washington,DC 20036

Telephone: 202-857-9622

Telex:

64260

Cable:

NATRECA


4/54

Contents

Section

1.

Introduction,

1

A.

The

purpose

of

this

manual,

1

B.Thebasis

of

hydrologic

estimates,

2

C.

Limitations

of the

hydrologic

methods

presented

in

this

report,

3

2.

Hydrologic

processes,

4

D.

Peakdischarges,

4

E.

Minimum

discharge,

7

F.

Frequency

of

highand

low

flows,8

3.

Estimating

peak

flow

at

asite,

13

G.

Field study,

13

1.

Physical

signs

of

high

water,

13

2.

Interviews

with

residents

andhistorical

accounts,

14

3.Calculating

discharge

fromwater

level, 16

H.

Calculation

of peak

flows

fromwatershed

data,

17

1.

Calculating

the

flowtime for

water

tomove

through

the

watershed,

19

2.Estimating

rainintensity,

20

3.

Estimating

final

peak

flow

value,

21

I.Determining

final

peakflow

value,

22

4. Estimating

the

flow-duration

curveat

a

site,

24

J.

Field

investigation

for

estimation

of

flow-duration

curves

at

a

project

site,

24

K.

Calculations

of

a

flow-duration

curve

frommeteorological

data,

27

1.

Rainfall

and potential

evapotranspiration

(PET)

data, 29

2.

Estimating

watershed

characteristics,

30

3.

Calculations

of

monthly

runoff,32

L.Determination

of

the

flow-duration

curve,

36

5.

Determining

the

final

answer, 37

Appendix

A.

Anexample

of

peak-flow

calculation,

41

B.A

Fortran

program

for

calculatingmonthly

runoff

volumes,

43

C.

NRECA

flow-duration

model,

47

Glossary,

49

iii

Contents

'6


5/54

Tables

1

Monthly

runoffvolumes,

10

2 Monthly

runoff

volumee

inorder of

magnitude,

10

3 Watershed

loss

ratesand correction

factor for

predominant

vegetation,

22

4 NRECA

flow-duration

model:

Tabular calculation

of

monthly

runoffvolumes,

31

5 Example

of

Rio

Targaestimates,

37

6

Peak-flow

sensitivity,

39

7

Duration-curve

sensitivity,

40

Figures

1

The

hydrologic

cycle,

2

2

Flow

routing through

eorage

inchannels

or reservoirs,

6

3 Monthly

flow

volume

in

astream,

7

4 Flood

frequency:

Peak

flowvs.return

period,

9

5

Flow-duration

curve

for

monthly

flow

volume,

11

6

An

illustration

ofpeak-flow

field observations,

15

7

Stream

cross

scction

at flood

stage,

16

8 Finding

flood velocity,

18

9

An

example

of

rainfall

intensity

vs. duration

of rainfall

for

a

50-year

return

period,

20

10

Flowtime

forwater

tomove through

thewatershed

from

watershed

characteristics,

21

11 Flow-duration

curve for

monthly

flow

volumes,

25

12 Stage-discharge

rating

curve,

26

13

Asketch

ofmonthly

runoff

calculations

from

rainfall

and

potential

evapotranspiration

data,

28

14 AET/PET

ratio

as afunction

ofPRECIP/PET

and

soil-moisture

ratio,

34

15 Soil-moisture

storage

ratio,

35

16 Selecting

a

final answer

for

peak flow

from

sensitivity

analysis,

40

iv Contents


6/54

Section1

Introduction

A.

The purpose

of this

manual

Technologically sophisticatedprocedures

are

available

andwidely

used forthe selection,

planning,anddevelopment ofhydropower.

Most

of themarenot

useful

for

smallwatersheds

in

developing

countries,wherefewstreams

havebeengaged and there

is

little

long-term

data onstreamflowcharacteristics. Potentialinvestors

justifiablywantdetailed,

comparativehydrologic

analyses of

candidate sites,

but

the costof gaging

all possiblesites could

make the front-end costs

of mini-hydropowerdevelopment

excessive,

not to

mentiontime delays.

The

methods

presented

include

techniques

forvisualdetectionof

clues tohistorical flood levels,

guidelines

for

questioning

inhabitants in

the

neighborhoodof

the

stream,

and

equations

and

tabulardata for

calculating

peak

dischargefrom

the

scanty

data

likelytobeavailable.

It is hoped that

this

approachwillbe tested

on sites

inmany

countries

and that criticismand

constructivesuggestions

willbe

passedon toNRECAso that

these

procedures can

be further

adapted,

refined,and

improved. Commentsonthis document* are

welcome.

This

manualdescribes methodsand includes samplecalculations

for

(1)

Estimating the

peakstreamflow

at

asite,

and

(2) Estimating

the

flow-durationcurveat

a

site.

It describes

ways

to

collectfielddata

on

historicflows

forpeak

flow

and flow-durationestimates

and calculation techniques for

peakflowandflowduration

basedonrelatedmeteorologic dataand

watershedcharacteristics. The

methods described

canbe

applied

to

watersheds of

up

to 1000

square

kilometers in areaswheresnow

accumulation

and

melt is minimal

and

where streamflowsarenot

regulated

by large lakesor reservoirs.

These

methods

will

providereasonable

hydrologic

estimates when

usedby technicalor

professionalpersons.

The

hydrologic

analysis ofasite should

require less than5days of

technical or professional time.

Thegoalof thismanual

is

to

detail

selected,

simplemethods

for

obtaining hydrologic

estimates for

small

projectsthatuse field

*This

reportwas preparedforNRECA

by Dr.

Norman

H.Crawfordand

StevenM.ThurinofHydrocomp,Inc.

1

Introduction


7/54

Figure

1

The

hydrologic

cycle

Precipitation

Interception

E

Evapotr

apiration

1701,i

Evaporation

C ndwa-.'

datausually

available in

developingcountries. Comprehensive

methods thatusemoreextensive

field

data canbe

usedfor larger

hydroelectric

and

waLer resourceprojects.

There

is

a

short glossary

at

the end

of

this

document.

B.Thebasis

of

hydrologic estimates

The hydrologiccycle (figure

1)

operates

continuously

in

all

watersheds.

The hydrologicinformation

needed

for

a

hydroelectric

project

are

apeak

flowand the flow

duration at the site.

The

basichydrologicprocesses

that

producepeakflowand flow

duration

at

a

siteare describedbelow.

Peakflowsare causedbyhigh intensity stormrainfall. Rainfall

that

moves as surface runoff into streamchannelsproduces

a

"floodhydrograph."

The

restof therainfall

infiltrates

into the

soil and

is

"lost."

It will later evaporate or transpire

orwill

providegroundwater flow

into the stream.

The

factors that must

be estimated

or

calculated to

estimatepeak

flowsare the-

(1)Rainfall intensity

and duration

on

the

watershed

(2)Amount

of

"losses"

or

infiltrationduring

the storm,

and

(3)

Flow

time

and storage

in

streamchannels.

2

Introduction


8/54

Flow

duration for

awatershed

is

the percent of time

that flows

exceedspecific

levels.

The factors

that

control

flow

duration

are

the-

(1)

Annual cycleof

precipitation,

potential

evapotranspiration,

and actual

evapotranspiration

in

the watershed.

(2)

Amount

of

rainfall

that

infiltrates

and

moves

on

subsurface

flowpaths into stream channels. Tnfiltrationrates dependon

the

permeability

and

depth of

thewatershed

soils.

(3)

Subsurface

flow

velocities

and

the storage

capacity

of

subsurface

aquifers.

C.

Limitations

of

the hyrrologic

methods

presented

in

this

report

These

methods

operate

with

limited

fielddata

and

produce

results

of

moderate accuracy.

The

most important

factor

i n the

accuracy

of

a

hydrologic

estimate

is

the

accuracy

of the

meteorologic

data

on

which

it

is based.

If

the

24-hour--100-year frequencyrainfall

on

awatershed

is

60 mm,

and

a

45

mm

amount

is

usedin

calculations,

the

calculation

for

peak

streamflow

will

give

anircorrect

result.

The

calculation

techniques

in

this

manual are

based

on

sound

hydrologic

principles

and

represent

the

key

processes

inflood

flows

and

continuous

monthly streamflows.

They

donot

include

processes like

snowmelt

or

detailed

hydraulic

routing

and should

not

beused

where

snowmelt

is

akey

process

or where

there

are

large

lakes or

reservoirs.

They

apply

tc

small

watersheds,

generallyof

less than

1000

square

kilometers.

3

Introduction


9/54

Section

2

Hydrologic

processes

The

hydrologic

cycle

was

sketched

in

figure

1.

The

water

in

the

system

comes from

precipitation.

It

willeither infiltrate

through

the land

surface

into

the

soil or

move

toward stream

channels as surfacerunoff. Waterthat infiltratesmaymove as

subsurfaceflow ormay

be

evaporated or transpiredby

vegetation.

In

mostwatersheds,

the

amount ofwater

lostby

evapotranspiration

islarger than theamount of

water

that

becomes

runoff. Runoffwatermoves over the surfaceas directflow

to

streams ormoves subsurface

as

delayed

or

groundwater

flow.

Directrunoff

maycause

floods,

whilethe delayedsubsurface flows

provide

continuous

or low

flows to the

rivers.

Two types

of

informationare neededfor

a

hydroelectric

site.

First, the flood flowor

expected

maximumwater level is

needed

to

size

a

spillway

(if any),

to

locate;

turbines-and

generators

above

thehighest

expectedwater

level,

and to design

diversion

structures

or canals.

Second,

the'statisticaldistribution

of

monthly

streamflow

volumes

isneeded

to

estimate

the

reliability

of the site for theproduction of

a

givenamountofelectrical

power and to sizethe turbine.

D.Peakdischarge

Peak flows resultfromacombination

of

heavy

rainfall andhigh

soil-moisturelevels,

which

prevent

thewater frommoving

into

the

soil. Peak

flows

on

small

watersheds are

frequentlycausedby

thermal

or

thunderstormrainfall. Peak flows onlarger

watersheds

arecausedbyaseriesof rtorms

or

bysnowmelt.

Flood

flows

are

plotted

on specialsemilog scales. If flood flows

aremeasuredfor

aperiodof years, this

historicdata

canbeused

to

estimate themagnitudeof floods statistically. Hydrologists

refer to50-year

or

1-year

floods,

whichare theflood levels that

wouldprobablybeequaled

or

exceeded

onlyonce

in

50 or

100

years. Toestimatepeakdischarges,

a

hydrologistmust estimate

themaximumrainfall rates thatwill occurinawatershedfor the

duration

or

length

of

storms

expected to

cause

amaximum

flow.

Therunoff

from thisrainfallmust thenbeestimated.

The

lifferencebetween the rainfall andtherunoff

will

be the

infiltrationthrough

thesoil

surfaceduring

the

storm

event.

Whenrainfall

exceeds thecapacity

of the soil to absorb

water,

surfacerunoff

occurs andenters theriverchannels, duringand

immediatelyafter therainstorm. Itflows through the river

channels

andis

measuredat

a

streamgage as

aflood

hydrograph and

apeak flaw. When

rainfall

onawatershedisplotted together

with

streamflow fromthe

watershed,the

close

relationshipbetween

rainfall

andpeak flowis

clear.

4

Hydrologic

processes


10/54

The

factors

that determine

the

peak

flow or

maximum

flow

in a

flood

hydrograph

are the intensity

of the rainfall

that causes

the

runoff;

the

amount

ofwater

that

infiltrates

and follows

the

slower,

subsurface

paths

to the

stream

channel;

and the amount

of

attenuation

or subsidence

of

the peak

flow as

it moves

through

the

river

channels

to the

streamgage.

The

first

factor,

the intensity

of

rainfall

over the

watershed,

is acharacteristic

of

the climate

of the region.

In

the

United

States,

for

example,

a

rainfall

of

30

umi per hour

is

a

"heavyintense

storm"

in

Seattle,

Washington,

but it is

a

"light

rainstorm"

inHouston,

Texas.

Data

on

the

intensity

of rain

that

is

to be

expected

in

thewatershed

is

needed.

The

second

factor

that determines

flood

magnitude

is the

amount

of

water

that

is

absorbed

through

the soil

to

follow

the

slower

subsurface

flow

paths.

Infiltration

rates or

losses

depend

on

soil properties

and

on soil

moistures.

Soils may

range

from

tight,

lowpermeability

clays,

tohigh

permeability

silt

and

sand.

Soils

with

high

permeability,

like

sandy loam

soils and

forest

soils that

have

thick organic

layers

of decaying

vegetation,

absorbwater

quickly.

In

watersheds that havehigh

infiltration

capacities,

surface

runoff

may

be

unknown.

Soils

of

low

permeability,

such

as

clay,will

absorb

very

littlewater

and

there

will

be surface

runoff.

When

arainstorm

breaks a

long dry

spell,

soil

moistures

will

be

low

and the amount

of water

that infiltrates

or

is absorbed

by the

soil willbe

high.

If a

rainstorm

is

the

latest

ofa

series

of

storms,

soilmoistures

will

behigh

and the

amount

ofwater

absorbed

by the

soil will

be

low.

The

soilmoistures

that can

be

expected

at different

times

of

the year in

a

watershed

arealso

a

function

of the

climate.

Inhumid

climates

where

the rainfall

is

alwayswell

in

excess

of

the potential

evapotranspiration,

soil

moistures

remain

high.

In

arid

climates

where

the

potential

evapotranspiration

exceeds

the

rainfall,

soil

moistures

are

usually

low.

The

runoff

that

enters

stream

channels

from the

land

surface

may

bemodified

substantially

as

watermoves

through

stream

channels

towarda

gaging

site.

Flood

waters that

enter a

natural

lakeor

reservoir

are "routed"

or

attenuated

as

theymove through

the

reservoir.

The peak

flow

leaving

the

storage

may

bemuch

less

than

the peak

flow

entering

the storage.

This

is illustrated

in

figure

2.

Flow

along

natural channels

in

a

river

basin

also tends

to

attenuate.

Themaximum

rate

of discharge,

expressed

in

units of

flow

per unit

ofwatershed

area,

tends to decrease

as

the

flood

moves

downstream.

Hydrologists

account

for this

attenuation

process

by

using

"flood

routing"

or

"flow routing"

procedures.

The

amount

of

attenuation

of aflood

inawatershed

depends

on the

length,

shape,

androughness

of the

channels.

If the channels

have

broad

vegetated

floodplains

and if

the flood

moves

out

of

the

5

Hydrologic

processes


11/54

Figure

2

Flow

routing throughstorage

Discharge(cubicmters persecond)

50

Inflow

40

30

IntflC

W

20 .* - _ _ _

30 t c

10

___________

0 1 2

3 4

5 6

7

Time

hours)

incised

channels

onto the floodplains, the

attenuation

of peak

flowswillbe dramatic. If

the channels

in

the

watershed

are

narrow and

steep, and flowvelocities

are

high,

the

attenuation

of

flood

peakswill

be

much

less.

If

flowsmove

into

natural

lakes,

swamps, marshes, or

manmade

reservoirs, reduction

in

peakflow

due

to

"reservoirrouting"

like that

shown

in

figure

2

can

be

expected.

In

sumary,

the

problem

of

estimatingpeak

flowon

a

natural

watershedrequires

(1)

Data

on

rain

storms thatwilloccur in

the

region,

andonthe

maximumrainfall intensities

that will occur.

(2)

Estimates ofthe

infiltration

losses that

willoccur

during

the

stormevent. This infiltrationwaill depend

on

typical soil

moisture

levels and

on

the

characteristics of

soils in the

watershed.

(3)The

attenuation

of peakflows

as

theymovefrom theheadwaters

of the

watershed

through

the

basin.

6

Hydrologic

processes


12/54

Figure 3

Monthly flowvolumeinastream

Volumeof

discharge

(cubic

meters)

300,000.

250,000

200,000

.

150,000

100,000

50,000-

A

J J IF 1

IA IOIFI AIMJ J ASIOIN D

3 IFIIIA

IM

tJIA

IS1OIN ID

JIFNA

HIJJJJAISIOINID

1976

1977

1978

1979

E.

Minimum

discharge

Low flowperiods in

streamflow

canbe seen

infigure 3.

The

lowest flows

arereached

duringdroughts

when

very

little

rainfall

has

occurredfor an extended

periodof

time. Justas

the

peak

flows

ormaximum

flowsare reduced

by

water that isabsorbed

through the

soil surface,

the

low

flows

are increased

by

this

infiltrating

water

as

waterissupplied

to

the stream

channelsby

thesubsurface

or groundwater

flowpaths

that

wereshown in

figure 1.

Thus, the factors

that control the

lowflow inastream

are the length

of

theminimum

rainfall

periods, the

amount of

water that is

absorbed through

the

soilsurface

during rainstorms,

andthe

time needed

forwater to

flowalong

the subsurface

or

groundwater

flowpaths.

The climatein

the watershed

controls the

length

of drought

periods. In areas that

have seasonalrainfall,

there

may

be

a

drought

for

several

months

of

each year.

Inthese

climates, the

lowest

flow

willoccur toward the

end ofthe annual

dry

seasonand

7

Hydrologicprocesses


13/54

will

be

much

lower than the

minimum

flow in anareawhererainfall

occurs throughout

the

year.

In

arid

climates

where the potential

evapotranspiration

exceeds the rainfall,

minimumflows

willbe

less

than

in

humid

climateswhere

the

rainfall

is greater

than

the

potentialevapotranspiration.

The

subsurface

flowpaths in awatershed

are stable

and

predictable.

If

awatershed

is

known to have

aminimum

flow over

a

10-year

period

of 10

cubic

meters/second,

it

is

unlikely

that

the

minimum flowwill suddenly

drop

to 2cubic meters/second.

Low

flows

respond to cumulative climate

changes.

One or two weeks of

dryweather

will

notcause

unusually

low

flows inawatershed,

but

lowrainfall

foraperiod

of6

to

36

monthsmay

result

in

the

minimum observed

lowflow. Alarge subsurface

groundwater

storage

will takemonths

rather thanweeks

to react to a

change

in

climate. Even

limited observations

of low

flows areveryhelpful

in

documenting

the low-flow

regime inawatershed.

The hydrologicprocesses

important

to lowflow

in streams occurat

and

below

the land

surface.

Water that is infiltrated at

the

land

surfacemaymove

along the subsurface flowpaths

into the

stream

channelwith

time delays

ranging

from

weeks

to

years.

These

time

delays

aremuch longer

than

the time

delays

in

channel

flow. When

water

enters a

river channel, the

typical time

delay

to

move to a

downstream

measurement

point

is

hours to days.

But

an

important

exception

to this rule

is

a

channel system

that

contains

major lakes or

marshes.

They

may

store waterduring

floodperiods

andrelease it slowly

over aperiodof

months,

causing

lowflows

to increase. Alteration

of

natural channel

storage to increase

low

flows

in

astream is

the

goal

of

astorage

reservoir.

Insumary, to

estimate the

minimum discharge

in

a

stream, the

important

factors

are

the

(1)Climate

of the

watershedand the

durationof

drought periods

(2)

Infiltration

of

rainfall

through thesoil surface

and

water

movement

into

the

subsurface

groundwater flow

paths

(3)

Typical

ti'e

delays

or

the subsurface

flowpaths,

and

(4)

Marshes,

lakes, or reservoirs (if

any) that

are present

in

the

watershed.

F.Frequencyof

high

and

low

flows

The continuous

measurements

of streamflowplottedin

figure 3

showedboth

the highand the

lowflows

from

a

watershed. It

is

sometimes convenient

to reorganize

these data

to show the

frequency

of high

or lowflows

separaLely on

semiloggraphs as in

figure

4.

For

example,

for

high flows, the

peak flow

measured

8 Hydrologicprocesses


14/54

- -

Figure

4

Floodfrequency: Peak

flowvs.

return

period

Peak

flow (cubic

meters

per

second)

700 - - - - -: -

* *1" * , .'

. .....

. I... .

.....

...

50

r-tt f ::fl

- -

....

H--4ILL... ... ..

400 ,-7- - - K

.,.. . .. .~ . . . . ....

. ,.. . .,......

I

i 11

I:2

300

S... ....... . ...

.. ....

........

J........................

...

...... ...

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

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

-

..

".

,+-

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

....

. ... .

.... - -. i ii- .

I.... ...

..

.

n

. .

...........- *

.

|...

..

-

-

...

. .,

......

...

..

....

1

2

3

5

10

20

30

50

100

Returnperiod (years)

eachyear for

aseries ofyears

canbe plotted.

If 20

yearsof

data

are

available, the

flood flows are listed

in

order

of

magnitude

and the

highest

flow is

assumed

to recur oncein 20

years, the second

highest

flow, once

in

10 years,

the third

highest

flow once

in

6.67

years,

and

so on. This

allows the

points to

be

plotted

on

the

graph, as

shown

in figure4.

Peak

flows formanywatersheds produce

linear

or

near linear

curveswhenplotted

on semilogpaper.

These flood frequency

curves areused

to

estimate

floods for

project

design.

A

designer

9 Hydrologic processes


15/54

might

choose

to

design

a

structure

for

aonce-in-100-year

flood,

a

flood

that

would

be

equalled

or

exceeded

an average

of once

every

100

years.

Flood

frequency

curves

like

figure

4are

strongly

influenced

by

themaximum

historic

flood

that

has

been

observed.

When

historic

flows

in

a

watershed

are

not

directly

measured,

newspaper

accounts

of

the

floodmarks

or

memories

of

local

residents

may

be

sufficient

to

estimate

the maximum

discharge

for

historic

floods.

Minimum

flows

can

be treated

like

maximum

flows.

A

frequency

curve

of

theminimum

discharge

each

year

could

be

prepared

and

plotted

using

the

same

procedure

that

was

described

for

peak

flows.

Avery

useful

technique

for

graphically

representing

the

continuous

flow

measured

in

a

watershed

is

the

flow-duration

curve.

To

construct

aflow-duration

curve,

assume

that

table

1

is

the

monthly

volume

of

runoff

for

ayear

in

awatershed.

These

runoff

volumes

are

rearranged

in

order

of

magnitude

in

table

2.

The data in table

2

are

plotted

in

figure

5.

It

canbe seen

in

figure

5

that

the

runoff

equals

or

exceeds

100

million

cubic

meters

40

percent

of the

time.

The

runoff

exceeds

175

million

cubic

meters

20

percent

of

the time.

Table

1

Table

2

Monthly

runoff

volumes

Monthly

runoff

volumes

in

order

of

magnitude

Volume

of

runoff

Volume

of

runoff

Month

(cubic

meters)

Rank

(cubic

meters)

January

102,300

1

291,900

February

189,000

2

211,600

March

291,900

3

189,000

April

211,600"

4

113,400

May

98,700

5

102,300

June

32,000

6

98,700

July

12,600

7

96,100

August

8,700

8

68,700

September

14,500

9

32,000

10

14,500

November

96,100

11

12,600

December

113,400

12

8,700

October

68,000

10

Hydrologic

processes


16/54

Figure 5

Flow-duration

curve

for

monthly

flowvolume

Monthly

flow

volume

(cubic

meters)

350, 000

300,000

250,000

200,000

0

150,000

100,000

50,000

0

20

40

60

90

100

Percent

of time

monthly

flow

volumes

are

equalled

or

exceeded

The

flow-duration

curve is

veryuseful

for

study

of

projects

that

divert

water

from

ariver

for

water

supply,

irrigation,

or

hydroelectric

power

production.

Flow-duration

curves

can

be

constructed

from any

period

of

streamflow

data

but are

more

reliable

if

they

are

constructed

for

several years

of

streamflow

data. Flow-duration

curves

use

ill of

the

streamflow

information

that

is

available

at

a

site,while

peak-flow

and

low-flow

frequency

studies

selectout

maximumor

minimum

flows

for

analysis. In

locations

where

very

little

streamflow

data

is

11 Hydrologic

processes


17/54

available,

it maybe feasible

to construct the flow-durationplot

butnot

peak-flow

and

low-flow

frequency curves. Ofcourse,both

peak-flow

and low-flow

frequency curves

and flow-duration

curves

increase

in

accuracy

as the

lengthof the

basic

data observations

increases.

12 Hydrologicprocesses


18/54

Section

3

Estimating

peak

flow

at asite

To

estimate

the

peak

flowthat

could occur

at

a

potential

hydroelectric

site,one should

(1)study thewatershed

to findout

what

high flows have

occurred in

the past,

and

(2)

study

the data

on thewatershed

to

calculatewhat

high flowsmight

occur in

the

future. Section

Gdescribes

how

to perform

field studies

at

a

siteto findout about

pasthigh flows.

Section

Hshowshowto

calculate future

peak flowsusing

the available information

about

thewatershed.

Section

I

tells

how tocombinethe

answers from

sections

GandHtoproduce

afinal

result.

G. Fieldstudy

Thefieldstudy

of

the

watershedwillanswer

the

question,

"How

largehave past

floods

been?"

Twomethodswill

beused to find

outhowhigh

thewaterhasbeen.

Themethods

are to-

(1) Study

the

physical

features near the

streamto findsigns

of

high water, and

(2) Talkto peopleliving

near

the stream

and

to

officials in

the

areaabout

past floods.

Once the

high

watermark

has beendetermined,

youcanusesome

calculations

to

find

the flowrate for

that

water

level.

These are explainedbelow:

1. Physical signs of

highwater

Thepurpose

of this section is

togainan

understanding

ofhow

the

water

level inthe streamvaries.

Writedown

the important

observations that

you

make

as shown

in

figure

6.

Flowingwater

moves

many

objects thatit

touches. Grass,

branches, and other

light objects

float alongnear thewater

surface.

Earthand sand areeroded

by

fast-moving

water

and

carried

downstream, orare

left onbeaches

or inpools.

To

begin

yoursurveyof

the stream

channel, stand on the

bankof the stream

andwatch

the edge of the

streamas

itflowsagainst

plants

and

rocks

or

sand.

Imagine

that

the

water

level

is

higher

by

1

or

2

meters,and

seewhatwould

beunder water. Look

for signs,

like

eroded

tree roots

or

sand

or

gravelbeaches,which

indicate

that

the

water

has been

at that level.

If

water

is

not movingvery

fast or is

flowing

through

bushes and

trees,

light

objects

which

are floating

willoftenbecome

caught

inthebranches

or lefton

floodplain

landswhenthe water

level

goesdown.

Lookforclumps

ofgrass and

sticks

caught

in

the

trees

oronbeaches

asyoumove away from

the stream.

13

Estimating

peakflowat asite


19/54

Lookat the

general slope

of

the

land towards

the

stream.

Water

will

always

form

a

flat surface.

If

you find

ahigh

deposit

of

oldfloatedmaterial,

imagine

that

the water

isat

that level

everywhere,

and

determine

what

areawould

be

under

water.

Look

at

the

ground

at the same

elevation

asyour

high-water

level.

Are

there signs

that

the

ground

has

beenunder

water?

Look

for

large

branches

and logs

whichmight

have floated

to

the

edge of

the

stream

during

a

flood.

Once

you

have

foundwhat

appears

to

be the

highest

deposit

of flotsam,

look

forother

signs that

the

water

was

once

at

or

near

that

level.

Look

for

markson

trees

or

buildings

or

more

branches

or

logs at

the

sameheight.

Look

at

the

deposited

material

and

try

to determine

how

long

it

has

been

there.

Is

the

organic

material

totally

decayed?

If

so,

itis

probably

more

than2years

old. Look

for

growth

of

vegetation

which

has

occurred

since the floating

material

was

laid

down. Wave

action

athigh

water

levels

willvery

quickly form

a

beach,

eroding

finematerials

and leaving

adeposit

of sand

and

gravel

along

the

streambank.

Signs

of beach

erosion

can

remain

visible

for

severalyears after

a

major

flood.

How

old are

the

plants

which

are

growing

ovdr

the

sanddeposits

and eroded

areas?

After

getting

some

idea

of

how

high

the

recent

water

levels

have

teen,

talkwith

local

residents

abouthigh

flows,

as

outlined

in

aection

G2.

Ifyou

have

some

information

from

the plants

and

soil

deposits

near

the stream,

it

will

be

easier

to evaluate

the

recollections

ofresidents

about

high flows. Keep

in

mind,

however,

that

youmay

not

have

seen signs

of

any

flood

which

occurred

more

than

2or3

years ago,

while

longtime

residents

mc"

remember

floods

from

over50

years ago.

You

may

learn

of

floods

which

were

considerably

larger

thanwould

be indicated

by the

evidence

youhave seen

near the

stream.

2. Interviews

with

residents

andhistorical

accounts

The

next

step

isto

use the

knowledge

ofthe stream

obtained

from

your

field

survey to

interview

local

residents

and to

draw

conclusions

about

past

flood

levels.

People

who

have lived

for

a

long

time

neara

stream

mayremember

large

floods

quite

well.

If

ahistoric

flood

was

large

and

if watermarks

remained

visible

for

some

time,

residents

will

know

thehighest

water

level

andwhen

the

flood occurred.

Asample

of

the interview

notes

youmight

take

isshown

in

figure

6.

By

talking

with

severalpeople

and

by

comparing

their

recollections,

you

may

get

accurate

information

on

the largest

floods

in the

past

50

years.

Start

an interview

with

ageneral

discussion

of the

stream

and

ofhow

it

behaves

throughout

the

year, and

evaluate

whether

or

not

the person

being

interviewed

is

familiar

with

the river.

Next,

mention

physical

evidence

of past

floods

near

thestream.

Finally,

ask

the person

being

interviewed

to point

out

marks

where

the

highwater

has

been

recorded

orwhere

he

remembers

high

water.

14 Estimating

peak

flow

at a

site


20/54

Figure6

An

illustration

of

peak-flow

field

observations

10/27/81-Started

field

study,

looking for

high-water

marks

on

east

bank, 500

metersupstream

fromroad

crossing.

Found

large

mass of

decayed

drygrass

and

branches

5mabove

waterline

in

atree.

Debris

appeared

to

be several years old.

Observed

other

smaller

bunches of

debris

at

slightly

lower

elevations.

Found

sand

deposit

25

meters

from

bank.

Significant

plant

growth

indicates

ithas

been

there

for

at least

5years.

Elevation

is 4.5

mabove

current

water

level.

10/28/81--Talked

to

Paul

Jones,

oldest

resident

(75)

living

near

the

stream.

He

showed

me

a'mark

on

his

door

where

he said

thewater level

was

in storm

of

December

1953.

He rememberedhis father tellinghim

about

a

storm

around

1890

which

had

beenabout

3meters

higher.

Talked toJohn

Smith

(62).

He

didn't

know

anything

about

flood in

1890.

Highest

waterhe

remembered

was

marked on

alarge tree

nearhis

house.

There

v: sa

date

(December

16, 1953)

written

on

the

tree.

Level

agrees

well with

flood

reportedby Paul

Jones.

Smith

also said

he

only

remembers

2

times

in

the

last

50

yearswhen

stream

has

driedup.

It

is important

to

tell

the

people

being

interviewed

that

accurate

data

will

help

you

design

asuccessful

and

reliable

project and

that

manydata

sources are

being

obtained for

comparison.

Be

careful

not

to lead

the

person

beinginterviewed.

If

aperson

thinks

that

describing

high flood levels

will

help

get

a

hydroelectric

powerplant, he

may tend

to

exaggerate

the level

of

flooding

he remembers.

Or

if the

person

being

interviewed

thinks

that

you do not

want to

know

aboutvery

high past floods,

he

may

describe

flooding

as having

been

less than

he actually

remembers

it.

To

avoid

problems

ofbiased

answers,

talk

to several

people

and

get

them

to

pointout physical evidenceof flood levels,

like

high-water

marks

on

buildings.

Ask

for any

available

written

documentation

or photographs

to confirm

their

reports.

Dateswhen

the

floods

occurred

and

asearch

forhigh-water

levels

at a

number

, nearby

pointsmay

alsohelp

confirm

reports of floods.

Other

ys of

verifying

residents'

reports

of

high-water

levels

include

searching

old local

newspapers,

church

records,

or

local

government

records of

road or

bridge

repairs.

15 Estimating

peak flow

at asite


21/54

Figure

7

Stream

cross section

at flood

stage

Mrk on

tree

Wterline

reported

by

John

Smith

with date

Current

waterline

Horizontal

scale

I

cm 4

meters

Vertical

scale

I

cm

2

meters

2

13.6

boxes

x

4

meters/box

x

2meters/box

-

108.8

meters

Area

-

Width

-

72meters

Average

depth

-

area/width

- 1.5

meters

Current

depth

-

0.8

meters

Measured velocity-

1.7

meters/second

Flood

velocity

-

mas.

velocity

x averae

depth

-

2.6

meters/second

~curren

aptH)

Once

you

have

finished

your

field

inspection

and

your

interviews,

examine

the

results

as

recorded

in

your notes.

Askyourself

the

following

questions:

(1) Have

youfound

a

reliable

water

level

for

the largest

flood

which

anyone

remembers?

(2)

Do you

know

when

the flood

occurred?

(3)

Is

there

physical

proof

of thehigh-water

level?

If

the

answer

to

all three

ofthese

questions

is

yes,

you

are

ready,

to

goon

to the

next

step

of

calculating

the

flood

discharge

from

tie water

level.

If

youhave

not been

ableto

accurately

document

the

largest

flood,

it may

be

necessary

touse

aslightly

smaller,

more

recent

high-water

level

that

isbetter

remembered

and

documented.

3. Calculating

discharge

fromwater

level

The

final

step

in afield

study

for

peak

flow

is

to

transform

the

flocd

level

you

have

determined

into

an estimated

peak

discharge.

To

calculate

the

flood

discharge,

you

need

to know

the

cross

sectional

area

and thevelocity

of

the

stream

at the

peak

water

surface

elevation.

Thecross

sectional

areaof

thestream

may

be

calculated

by

first

plotting

the

height

above the

ground

of

the

peak

water

surface

along

alineperpendicular

tothe

stream

channel.

Plot

peakwater

depth measurements

at regular

intervals

from

the point

where

the

high-water

line

touches

the

ground

onone

side

of the stream,

all

the

way

over

tothe

same

point

on the

other

side.

Figure

7shows

an

example

of

astream

cross

section

plot

at flood

stage.

16

Estimating

peak

flow

at

a

site


22/54

The

next

stepis

to

count thenumber

of

boxes between

the ground

surfaceprofile and thehigh-water

line.

This numberisthen

multiplied

by the scale factorsonyour plot to yieldthe

peak

flood's crosssectional

area.

Figure

7shows

how

this

calculation

isdone.

After

youhave

determined

the

areaof

thepeak

flood,youmust

determine

the

velocity

of

the

flow. Thiscanbest

be doneby

measuring

offastretchof 50meters

along the channel,

then

timinga

floating

object as the streamcarries

it

through

the

50-meter stretch. The

velocityof the

flowisthen--

Velocity50

(Eqn

)

average

time to float

through

(stretch

seconds)

Throw

theobject

into

the

center

of the

streamandmake

sureit

doesn't

strike anything

as it

moves

through the reach. Time

f"e

object

through

the

reachat leastfive times.

Thisvelocity

needs to be adjusted

to

take

into account thehigher

velocities

of

flood

flows. To do this, find

or estimate thewater

depthnear thecenter

of

the

channel

and the average

waterdepth

at floodlevels from

yourcross

section

plot. Use these numbers

as

shown

infigure

8to

find

the factor

by

which

you

mustmultiply

yourmeasured

velocityinorder

to getflood

velocity.

Onceyou

know

the flood

velocity

and the

cross

sectional

area,multiplythe

twonumbers togetherto

obtain

thepeak

flood discharge.

Q- VxA (Eqn

2)

H.Calculation

of

peak

flows from

watershed da'a

This section

describes

how

to calculate

the

peak

floodflow ata

possiblehydroelectric

siteusingthe

available

data

on the

watershed

and the

tables

and

figures presentedhere. The

things

that

you

willneed

are

a

contourmapof

the watershed and

regional

data onrainfall

intensityfor different storm

or rainfall

durations.

Rainfall

intensityis

given

inunits

ofmillimeters

per

hour. If45 mm of rain

is

measuredin

2hours, therainfall

intensity

is22.5

mm per

hour. Rainfall intensity

willdecrease

as

storm

duration increases.

InBoston,Massachusetts, for

example,

arainfallintensity

of

150

mn per

hour

willoccur for

storms of 10 minutesduration,but

a

rainfall

intensityof

30

mm

perhour

will occur for storms

of 4hoursduration.

The

steps

involved

in

calculating

the peakdischarge area-

(1)

Using channel

and

watershed

characteristics, determine the

flow time

for waterto

move through

the watershed.

17 Estimatingpeakflow

ata

site


23/54

Figure

8

Finding

floodvelocity

Flood

velocity

-

measured

velocity*

(depth

of

flood/measured

depth)

2

/3

Flood

velocity/measured

velocity

4.0

/

3.5

_

3.0 _

2.5 _

2.0 _

1.5

1.0 -

2.0

3.0 4.0

Depth

of

flood/measureddepth

5.0

6.0

7.0

8.0

(2)

From

regionalmeteorologicdata,

determinerain

intensity

for

the

design

storm.

The

durationof

thedesign storm

is

assumedto

be

equal tothe

flow timefor

water tomove

through thewatershed.

(3)

Determine

watershed losses,

the

"excessrains"

or theamount

of

rainfallthatbecomes

surface

runoff,

andthe

peakdischarge.

18 Estimatingpeakflow

at

a

site


24/54

This calculation

method

assumes

that

apeak

discharge will

occur

when the

storm

duration

is equal

to

the flow

time

for

water to

move through

the watershed.

This

is

an

idealization

of the actual

interactions

between

stormraiifallduration

and peakflow,

but

it

gives

peak

flow estimates

of

reasonable

accuracy.

Each step

in the

calculation

is

explained

in

the

following

sections

and

easy to

follow examples

have

been included.

For each

step, yourcomputations

should

be

written

down like the

examples

so

that

you

may

checkyour

work.

1.

Calculating

the flowtime

for

water

to

move

through

the

watershed. To compute

the flowtime

for

water

to move

through the

watershed,

you

willneed todetermine--

L =

channel length

in

kilometers

ER

- changein elevation

between

the

highest

point

in

the

watershed

and

the site, inmeters

AREA

- watershed area, upstream

of

the

site

in

square

kilometers

These

data

canbe found from

acontourmap

of thewatershed.

and ER

are

used in Eqn

3

to calculate

the flowtime

(TF)

through

the

watershed.

This flowtime

(TF) is assumed

to equal the

duration

of the

storm

rainfall

that

will

cause apeakflow.

The

duration

of the storm rainfall

is used to find the

rainfall

intensity

during

the storm

(RI),

using

local data similar

to that

shown

in figure9. The storm

rainfall intensity

(RI)

less a

loss

rate

(LR)

is

used to

find the

rate

of

runoff

or

rain

excess

(XR)

during

the storm (Eqn

4). Finally,

theexcess rain

is converted

to flood peakdischarge

by

multiplying

by

a

necessaryconstant

and

the watershed

area

in

Eqn

5.

The

flowtime

in hours

(TF) canbecalculated

from

L

andERusing

Eqn

3,or it canbe

read

from

figure 10.

0.95

x

(L

3

/ER)

"385

TF

-

(Eqn

3)

In

Eqn

3,

L

is in

kilometers,

ER

is in

meters,

and

TF is in

hours.

19 Estimating

peak

flowat

asite

L


25/54

Figure

9

Anexample

of rainfall

intensity

vs.

duration

ofrainfall

for

a50-year

return

period

Rainfall

intensity

(millimeters

per

hour)

140

120

100

80

lot.

40

20

4

3

12

16

20 24

Duration

of rainfall (hour)

2.

Estimating

rain

intensity

Once

youhave

calculated

the

TF from

Eqn

3,use

dataon

duration

of rainfall

versus

rain

intensity

to

find the

rainfall

intensity

on

thewatershed.

The

relationship

between

rainfall

duration

and

rainfall

intensity

isusually

displayedgraphically,

as shownin

figure

9.

It

maybe

possible

to

make

agraph

similar

to

figure

9

with

your own

data.

Graphical

plots

ofrainfall

intensity

versus

duration

of

rain

areoften

made for

different

statistical

return

periods.

A

rainfall

intensity

of 60

,m/hour

for

4hours

duration

would

be

equalled

orexceeded

once

every50

years,

as

shown

on

figure

9.

To usefigure

9,find the

duration

of

rainfall

inhours

that

equals

TF

on the

bottom

scale,intersect

the line,

andread

the

rainfall

intensity

(RI),

from

thevertical

scale.

20 Estimating

peak

flow

at

a

site


26/54

Figure10

Flowtime

forwater

to

move through

the

watershed

from

watershed

characteristics

Channel

length (kilometers)

TF- 2

0

200

400

600

800

1000

1200

Elevation

(meters)

3.

Estimating

excess

rain

anddetermining

peak

flow

Rainfall

will

be

lost during

a storm

due

to infiltration.

Table

3

describes

various

soil

types

and

gives

a

loss rate

for

each.

Losses

are

greater

in

watersheds

which

have

very

heavy

vegetation.

A

correction

factor

forvegetation

density

can

be

selected

that is

multiplied

by the

loss

rate

for

the

watershed

soils

to

give

the

total

loss rate

for the

watershed,

LR.

Now

that

you

have

determined

the

rate of

rainfall

loss,

youmay

calculate

excess

rain

by subtracting

your

losses

from

the

rainfall

intensity.

The total

excess

rainfall

rate

is

then--

XRR

I

-LR

(Eqn

4)

where

XR

RI,

and

LR are

all

in units

of

irm/hour.

21

Estimating

peak

flow

at

asite


27/54

Table

3

Watershed

loss

rate

and

correction

factor

for

predominant

vegetation

Predominant

soil

type

Loss

rate

(mm/hr)

Impervious

rock

1

Tight

clay

1

Clay

and

silt

3

Silt

and

sand

5

Sand

and

gravel

10

Correction

to

loss

rate

Predominant

vegetation

(multiply

by)

Sparse-Little

vegetation,

bare

soil,

scrub

brush

0.5

Moderate--Grassland,

cropland,

mixed

forest

1.0

Heavy--Dense

forest,

tropical

forest

2.0

Example:

A

watershed

with

clay

silt

soils

anddense

forest

cover

would

have

aloss

rate

of

3.0-,/hour

anda

correction

factor

of

2.0 so

its

total

loss

rate

would

be 3.0

x

2.0

- 6.0

un/hour.

The

final

stepis

todetermine

the

peak

discharge

at

the

site from

Eqn

5.

Peak

flow (cubic

meters/second)

- 0.28

x

XR

x

watershed

area

(ki2) (Eqn

5)

where

XR

is

exccss

rain

in

un/hour

and

peak flow

is

in

cubic

meters/second.

An

example

of a

peak

flow calculation

is

given

in

appendix

A.

I.

Determining

final

peak

flow

value

If

youhave

beenable

to

follow

both

the field

survey

and

the

calculation

methods

for

determiningpeak

flow,

youwill have

two

values

for

peak

flow

which

may

bequite

different.

This

isto

be

expected.

The remaining

task

is

to evaluate

the

L

values

in

orderto

arrive

at

onebest

value.

To

do

this,

you

will need

to

review

theprocess

by

which

you

calculated

each

number

and

determine

wherethere

was

likely

tohave

beena

large

amount

of

uncertainty

in

your

study.

22

Estimating

peak

flowat

a

site


28/54

Which

method

gave

youa

higher

value?

Wasone

value

more

than

twice

as

muchas

the

other? If

the

calculation

method

value

was

much

larger,what

was

the return

period

for the

rainfall

intensity

versus

duration

of

rainfall

data

that

youused?

If

itwas

more

than

50

years,

it

should

have

givenyou

alarger

peak flow,

because the field

study

interviews

will document

floods

which

have

occurredwithin

thepast

50

years.

If

the field

study

gave

youa

lower

peak-flow

value,

compare

the

velocity

figure

youused in

the field

study to

the

average

flood

velocity,

which is

channel

lengthdivided

by

theflowtime

through

the

watershed(TF).

Thetwo

velocities

should

be fairly

similar,

although

the

fieldstudy

velocityis

likely

to be lower.

Try

calculating

the

field

study

peak

flowusing

the Chezy-Manning

equation,

V

=

1/n

D

2

/

3

S

1

/

2

where

nis

resistance

to

flow,Dis

mean

depth in

meters

at

flood

stage,

andSis

the slope

along

thechannel.

In

natural

streams,

meandepth (D

in

meters)

can

be assumed

to

beequal

to the

hydraulic

radius and

isthe

cross

sectional

area

ofthe

flow

dividedby

its

wettedperimeter.

Slope (S)

isinunits

of

meters

of

elevation

loss per

meterof

length

along

the channel,

so itis

dimensionless.

Velocity

(V)is

inmeters

per

second.

For

Manning's

equation,

typical

values

ofn

follow:

Clean,straight

channel

withsand

bed 0.035

Winding

channel,

sand

orgravel

bed

0.045

Winding

channel, graveland

stone

bed,

some

streambank

vegetation

0.06

Rocky,

winding

channel

with

pools

and

obstructions,

streambank

vegetation

0.08

Do the

peak

flows agree

more

closely

now? Ifso,

use the

new

value for the

field

study

result.

In

general,

if the

field

study

and calculation

methodresults

for

peak flow

agree

within30

percent,

you

might

average

themand

use

this as your

final

result.

If

you

have reason

to

believe

that

one

method

was better

than

the

otheron

your

watershed,

your final

result could

fall

between

the two

results,

but

closer

to

theone

on

whichyou

place the

most confidence.

23

Estimating

peak

flow

at asite


29/54

Section4

Estimating the flow-durationcurve

at a

site

This

section gives methods

forestimating

the flow-duration

curve

at a

hydroelectric site.

SectionJisa

summary

of field

investigations

forestimating flow-duration

curves,

and

section

K

describes calculationmethods

for

flow-duration

curvesbased

on

meteorologic

information.

The flow-duration

curve

isamost

useful

tool

for evaluating

low

flows at ahydroelectric site. The

flow-durationcurve in

figure

5

is

repeated

as figure 11. These

data

canbeused

to

estimate

the flow

that

canbeusedfor

powerproduction90 or

95

percent of the time.

Thisgives an

immediate

indicationof the

reliability

ofpower

production

at asite. Forexample,

ifa

20-meterhead

wereavailable,

the

flow-duration

curve in

figure 11

would

indicate

that

as

119-kWpower

source could;be developed

at

85-percentreliability,

assuming an 80-percentoverall

efficiency

fortheproject.

J.

Field investigationfor estimation

of flow-duration

curves

at a

project site

Peoplewho

livenear

ariver

orwho divertwater

froma

river

are

likely

to remember the typical stageat

different timesof the

year. Highand low

stages aswell asnormalstages throughout

the

yearwillbeknown to local residents, even

though theywill not

know

thecorresponding

discharge. Knowledge

of

stage

canbe

converted

to

discharge using

astage-discharge rating

curve. A

stage-discharge

ratingcurve can

beconstructedusing

measured

cross

sections

at

two

ormore

stages

and calculating

flow

velocities

at

each

stage, as was

described

insectionG3.

Dischargeat

a

given

stage isthe flowvelocitytimes

the cross

sectional

area

of

the flow at that stage.

When stage-discharge

ratings

are

plotted on log

scales, linear or near-linearcurves

are

found. Figure 12 is

an

example of

astage-discharge

rating

curve.

To construct aflow-durationcurve,

the

following sources of

informationmight be used.

(1)

Are

there lakes or

reservoirs on the stream? Stagerecords

over

aperiod

of years

at a

lake or

reservoir

can

beused to

calculate

streamflow

volumes

entering the reservoir.

Whenmonthly

stage

records

exist at

areservoir, locateareservoir stage

versusreservoirvolume chart

for

the

reservoir.

The

streamflow

volume entering

the

reservoir eachmonthwillbe

the reservoir

volume at the end ofthemonth, less the

reservoir

volume

at the

first of themonth,plus

the volume of water

released

from

the

24 Estimating the

flow-duration

curve

at

asite


30/54

Figure

11

Flow-duration

curve formonthly

flowvolumes

Monthly

flowvolume

(million

cubic meters)

3.5

3.0

2.5

2.0

0

1.5

0

_

1.0

__ _

_

_0

0

0.5

0

20

40

60

80 100

Percent

of timemonthly flow

volume

are

equalled

or exceeded

reservoir during

the

month.

If

astage versus

reservoirvolume

chart

is

not

available,

one can

be

made from

maps of

the reservoir

area atdifferent

elevations,

sincethe

increase in

reservoir

volumebetween

two elevations

is

the

elevation

incrementtimes

the

reservoir surface

area.

(2)

Correlation

of

low flowson

aregional

basis ispossible.

Do

streamflow

recordsexist

on

nearby,

similar

streams? TheU.S.

Geological

Survey

has summarized

commonlyusedmethods

for

regional

correlation

offlows inWater

Supply

Paper

No.

1975,

titled

"Generalization

of

Streamflow

Characteristics from

Drainage-Basin

Characteristics,"

by

D.M.

Thomas

and

M.

A.

Benson

(55 p..).

25

Estimating the

flow-duration

curve at asite


31/54

Figure 12

Stage-discharge

rating curve

Stage

(meters)

8.0

_

_

4.0

2.0

/

pInt of zoro flow

10

20 30

Discharge

(cubic

meters

per

second)

(3)

Review

any

existingdata

on

flow

durationcompiled in

the

region

in

previous

studies.

(4)Question

local

residents, particularly

people

who are

using

the

stream

in

someway: ferry

operators, irrigation

farmers,or

people

who divertwater

for

water

supplies.

What

stages

are

expected

by season

ofthe year?

If

rainfall

in the region

is

seasonal,

what stages

are typical

of

the

lowest flow

month

of the

year?

What stages

are

expected

during

high-flow

months? What

is

the lowest

stage

that residents

remember? How

does

the lowest

stage

that residents

remember

comparewith

the typical

low

stage

that isobserved

eachyear?

(5)

In

somewatersheds, low-flowconditions

are

associated

with

poor

water

quality.

Is there

sometimes

unacceptable

water

quality,such

as

veryhigh

salt content?

Whendoes

this

condition

occur?

(6)Are

therewater

intakes along

the river?

Have

thesewater

intakes ever

been relocated

because of

low stage

in the stream?

Where

monthlystreamflow

canbe calculated

from

lake

or

reservoir

26

Estimating

the flow-duration

curve

at

asite


32/54

Wheremonthly

streamflow

can

be

calculated fromlakeor reservoir

levels,

as

in (1)above,

or if regional studies

to relate

flow

at

asite

to nearby

gaged streams are

done as in

(2)above,

the

data

can

be

used

to

construct aflow duration as was done

in

section

F.

If

regional studies

of

flow

duration

already

exist

from(3)

above,

flow-durationcurves

will be

available.

Wheninterviews are

used

to establish seasonal stages,

as in

(4)above,useManning's

equation described in

sectionIor

float

&!asurements to

establish

a

stage-dischargerelationship,

as was

described

insection

G3. Convert

stages

to discharge to estimate

monthly flows

foratypical

year. These

meanmonthly flows canbe

used to

construct aflow-duration

curve as

in

sectionF. If the

field

interviews, (5)and

(6)

abovi,

give clear

evidence of

occasional

low flows

that are subscantially less

than

the

minimum

expected annual

flow, amultiyear flow-durationplot

might be

made.

To make

amultiyear

flow-duration plot,

estimatemonthly

flows for

additional years and include

the

historic

low flows.

If

your

field information shows

theminimumflow in

an 8-year period,

8

full years of estimated

monthly flows should

be used to construct

the

flow-duration

curve. These

additional onthly flow

estimates

ensure that the

lowesthistoric

flow

correctly

influences the

flow-duration

curve:

the lowest

monthly

flow

in

8

years is

a

99-percent condition;

in otherwords, in

95months outof 96

this

minimumflowwouldbeequalled or exceeded.

K.Calculations

ofaflow-durationcurve

from

meteorologic

data

Streamflow results from

precipitation.

Flows ofwater from

the

land surfaceduring and immediately following

precipitation

create

floodhydrographs

and peak

flows. Water that

is absorbedby the

soil duringrainstorms

moves as subsurface

flow

into

stream

channels

and

provides

low flows

inperiodswhenrain does

not

occur.

Flow-durationcurves are

based

on

continuous

streamflow

data. On

ungaged streams where streamflowmeasurements

are not

available, precipitation and

potential evapotranspiration

records

canbe

used to calculate

continuous

flows.

Thecalculations

mimic

keyhydrologicprocesses: infiltration of

water

into the

soil

profile,

surface

runoff,

and

flow

along subsurfaceflowpaths into

the

stream.

The

calculation

method

described in

this

section uses

monthly

precipitation and

potential evapotranspiration data

to

calculate

monthly streamflow.

When

the

calculatedtrinthly streamflows are found, they

areused

like observed flows

to

calculate

the flow-durationcurve.

A

27

Estimating

the

flow-durationcurve at

asite


33/54

Figure

13

A

sketch

of

monthly

runoff

calculations

from

rainfall

and

potential

evapotranspLation

data

PRECIP,

MOISTURE

STORAGE

EXCESS

MOISTURE

DIRECT

LL4W

RECHG

TO

FIGURE 4.3

A SKEWC1 OF

MONTHLY

RUNOFF

CALCULATIONS FROM

RAINFALL

AND

9DTENTIAL

EVAPOTRANSPIRATION

DATA

GROU

ATER

STORAGE

GROUNDWATER

FLOW

TOTAL

DISCHARGE

sketch

of

the

calculations

isshown

in figure

13.

The

time

of

flow

in

strem

channels

in small

watersheds

isusually

less

than

1day

and

this

time

of

flow

can

be

neglected

when

monthly

interval

runoff

volumes

are

calculated.

Calculations

of

monthly

flows

from

meteorologic

data

are

based

on

the

water

balance

in the

watershed.

The

water

balance

equation

is-

Precipitation

- actual

evapotranspiration

+ storage

- runoff.

The

waterbalanceequation

applies

to

the

watershed

over

any

time

interval.

Where

precipitation,

actual

evapotranspiration,

and

runoff

are

the

volumes

of

water

entering

and

leaving

the

watershed

in

the

time

interval,

and

storage

isthe

change

in

soil

moisture

and

groundwater

storage

in the

time

interval,

the

initial

storages

less

the

final

storages.

Water

is held

in

storage

in

the

soil,

in

groundwater

aquifers,

and

in

lakes

and

snowpacks.

Allwater

flows

28

Estimating

the

flow-duration

curve

at

asite


34/54

into

orout

of

the

watershed

are

assumed

to

beincluded

in

the

runoff.

Thefollowing

steps

canbe

used

tocalculate

monthly

runoff

and

a

flow-duration

curve

from

meteorologic

data;

theyare

described

in

the following

sections.

(1)

Assemble

5ormore

years

ofconcurrent

rainfall

and

potential

evapotranspiration

data.

(2)

Estimate

the

watershed

characteristics

of thebasin.

(3)Calculate,

using

a

tabular

form,

themonthly

streamflows

for

5

ormore

years

based

on

these

rainfall

and

potential

evapotranspiration

data.

(4)

Calculate

the

flow-duration

curve

using

the

calculated

monthly

streamflows,

aswould

be

done

ifobserved

datawere

available

(section

F).

1. Rainfall

and

potential

evapotranspiration

(PET)data.

Streamflow

calculations

requiremonthly

rainfall

on

thewatershed

andmonthly

potential

evapotranspiration

data.

Rainfall

data that

is

observed

in

or

near the

watershed

must

be

found.

National

meteorologic

services

nreca water balance

Documents