Chapter 2
Chapter 2
Chapter 2
Hydrologic A na lys is
2.1 The walcrslu~d is [he bll~ic unil llscd in mos t hydrologic cllicu lations rela ti ng
WATERSHED to the wate r ha lancc or computa ti on of rll.in fnll-ru noff a nd o lht'T losses. The
CONaPTS watershed boundary defines a cont iguous area, such tha t the OC t rainfa ll or
runoff over th aI area wi!! contribu te water to the ou tle t (Fig. l- \9(b». Ra in
that falls outside the wate rshed boundary will gene rate runoff to sollle olhcr
oullcl. A wat ershed boundHY cn n be drawn from II topographic mnp by
defining the rid gc line to the omlel. a nd ru noff will generally \Tavel from
higher 10 lowe r eleval ion in a d irect ion perpe nd ic ular to the eleva tion con·
to urs (see Fig. 1-I 9(b».
Hydrol ogists arc most ofte n concerned with the amount of surface
runo ff generated wi lhin a walershed th aI becomes strea mflow for a given
input rainfflll p1Htem . The main w<tt crshed characte ristics Ihat affect hydro·
logic response include the size, shape. slope. soil Iype, fi nd slowge wi lhi n a
watershed area. Th ~ issues arc described in some detail in this chapler and
form the basis fo r more detailed ana lyses in later chap ters. 'Tlle hydrognaph.
as prese nted ea rlier. is a plot of flow nne vs. lime for a given location withi n
a stream and represe nt s the main hyd rologic re sponse function. Several
74
"
example wate rsheds were depicted in Figu re 1- 19(a), and a typical water-
shed with subareas depicted is shown in Figure 1-22.
Wa tersheds a TC of len Ch l'lfach:rized by one main ch annel and by tribu -
taries Iha l drain in to a main chan nel aI one or more confluence points. Tile
subarea of II tribulflry can b\: de linea ted by starling :ll ihe con fl uence and
drawing a boundary (li ang the subarea ridgo:line. Large r watersheds !;an have
many subareas thaI contribu te runofr lo a single OUlle t. An impon nnl wa ter-
shed parameter is drainage area A. since it reflccB the volum!:) of wate r Ihnt
can be ge nerated from rainfall. WOler;;hcd length mca~u res are de picted for
s uba rcll E in Figure 1-22. Channd length L IS usuany measured (llong the
mai n channel from the ou tl el to the basin di\'ide. Length to cen troid L , is
measured Blong the mBin Ch1Ulnci to B point nea rest the c.::ntroid (center of
mass) of the watc rshed. These two length ~lrHmt'te rs hel p deter mi ne water-
shed ~hape and arc used ex tensively in unit hydrograph calculations dis-
cussed t:lter ($cction 2.3).
Anot her imponant physiographic parameter is ehllnnel slope S or
wlller~hed slope S •. which reflects the ratc of change of ckvatio n wi th di s-
tance along the main channel or wi thin Hli overland flow area. Both stope
measures are used in pcrfoTllli ug unil hydrograph. nood routing. and lime-
of-Ira\'cl calculations. Slope es tim ates arc used in scveral of the nni t hydro-
gra ph methods de!ltrihed laler in this chapt er. Fie ld su rvey s or topographic
maps. either paper or electronic. can be us.::L1 to mC,lsure elevation changes
so thaI wluershed or chan nel slopes can be determincd.
Soillypes in a watershed are critical. as they de term ine infil tralion
rales that can OCCll r for the art"I. Soil prnperties can vary significan tly across
a watershed area. and the USDA Nat urBI Resources Conse rvat ion Service
(NRCS) is responsible for developing soils maps to provide infomlntio n 00
soil type. soil texture. and hyd rologic soil groups. The three main soil classes
Bre chnractc riled by particle diameter d in mill. fur sa nd. sil t. and clay.
Typ ical va lues arc listed in T:,blc t\-2 in Chapte r 8. Soil text ure is im ponant
in delermining water-holding e:'lpaci ty and infiltmtion capacily of a soil
la)"er. Thus. sands generally infiltmte watcr at a gremer rme thBn do silts or
clays. Of course. there can be mixtures of $11.e5. which can complicate th e
oVCTBI1 soi l st ructure. The N ReS classified t housand~ of soils o n the ba~is
of runo ff pOlcn tial and gro uped the m into four hydrologic soil groups. A.
B. C. ,IIld D. T ype A infilt ra tes at the high es t rate. D :11 the lowest. The
relationship of soil t)'pe to in rihnn ion cap:lei!)' is presented in de tllil in Sec-
tions 2.8 !l nd 2.9.
LMnd use and hmd cover. in the foml of park ing lots and urban de"e lop-
ment. (".111 have profound effocts 00 watershed response. In fact. many of the
me thods dc!ltribcd later wt"'Te developed to address urba n devel opment
impacts in 11 wa tershed. For cxnmple. th e Rational Method [see Eq. (1-13)]
uses a coefficient C to r(' floct the runoff potential of a waters hed. T he value
of C for commercial (0.75) is gTeale r than rcsidential (0.3). which is greater
than forested (0.15). indicating thm mo re in tense de velopment generatcs
76 Chaptet 2
greater rates of runoff for the sam!: rainfall (see Sections 6.4 and 9.3). Urban
development is also characterized by the perce nt imper.'iousness, or paved
a rea. which can range from SO% 1090% for oolllll1ercial com pared to 20% to
40% (or residentia l areas (sec Ta ble 2-1), Several of the unit hydrograph
me thods (Sectio n 2.3) contain pa rameters thaI relale \0 urba n land use c«eels.
Tabl. 2MI Runoff C ~"'. Number! lor Selected AGncu/lurol, St.obutbon. ond I./rlxIn lond U.. l""locedent Moi...".,
Condition II; 10 • 0 251
lIydroiogic Soil
""
COII'_m...:ioI end bu.;ne .. 0,,,,,.]85'" imp-rvooulj 92
IndU$lIio1 di~ids 172% impeMooII
R... idenlioP
88
" 9J
"" "
85
1/ 4 0( J8 81 75 B7
1/3 0( 3.
1/2 ac
, ~
2S
2.
"" 7. 72
80
"
86
""
Poot'd "';!h curbs ond $Iorrn --..l
Gro.eI " " " ""
" "
70 85
Dif!
'FOfo_ ....lod doocripIianol~ bncI .... cwo.--'-'. _"'~E''II"_ i•.,Hao « " SodI"" 4, "H)dn>Iogy,"
72 87
"
C""",,", 9, Aus 1972 .
'Good..,...,." ",,,,,",,red ~"'" 9",,,1"11, ""d lit!.r otId bn.,h c..- ....
>c..._ .... mb.. "'" _pule<:! o .... ming tho! ... nJnoII &_ ..... """" IIf><I d.........~ I, ~~ t>wotd .......... ...It." mlnlm"m 01
,oJ _ direaed to Igw" ...;... odd-.ol ;"MorOlioo c~ o=N
'n. -""9 peroIcM Of_ fbwoII_ ~ '" b. I~ vood """"'0 ~ "" ' - CUJ'4 ..........
1~ ........... _ d, .... 01"'-"""""",. ° CVI"<e """,bot 0195......, bo vsed.
Hydrologic Analysis n
Main chan nel and lrih ulliry chan,cl eriSl i~ can ;,ffect stremnflow
response ill a variety of ways. As presented in Chapters 4 and 7. open-channe l
now factors such as slope. cross-sectional area. Manning's roughness cod·
ficieni ll. presence of obstruclioos. meander pallern. and channel condilion
can all contri bUIC. Effects of challnel geomorphology on now patterns 3re
cove red in more detai l in tex tbooks on geology and flu id mechanics. Flood·
plni n anu lysis an d floodplai n mapping. presented in Chapters 7. 10. alld 12.
are based la rgcl)' on an undcrslandingofthc nature of channel geomorphology
as it relates 10 O\'erland now processes. which produces runoff that must be
moved downstream by the channel. ·Tlte shape. slope. and character of a
floodplain will de t!;:rm ine Ihe vo lume and now rate of water th~t c;m be
snfel y hand led during e,;cessivc cvellls. The pict ure on th e ,,~,;tbook cover
shuWl> the effects of a major out o f bank flooding in I [ouston. TX. d uring
Tropical Siorm Allison in 2001. Rood problems occur in a watershed when
either too lIIuch water is general ed from a rainfall even t or the channel is
inadeq ua te or not properly maintai ned 10 hand le excessive overland flows.
Physiogra phic characteristics freque ntly used in hydrologic studies
have been com piled for the USGS-EPA Nlilional Urban Studies Program
(USGS. 1980) and are summarized below. La nd use characteristics should
be updated during the eour.oc of a hydro logic slUdy 10 ueeount for cha nges
occurring in Hwatershed. Such physiogrnphic i n for m~ tilJn eM be developed
from maps describing land use. soils. topogrAphy. and StOrm dra inage liS well
AS from acrial photograph y. Fonu na lcly. many advances in recent years in
the area of geographic i'lforma tion syslems (GIS). with the linkage of elec·
tronic maps and dala bases. havc allowed this process to be grea tly imp roved
in time req uiremenb and ovcrall accuracy. as dcscribctl in Chapter 10. A
selected list of para meters follows below:
I. Total dminage area. in squa re miles
2. Impervious arca in po::rccmage of drainage area
3. Average basin slope determined and main conveyance slope at
points 10% and 85% along the slTcum from th.: oUllet to the divide
4. Hydraulic conductivity of Ihe A hori7.olt of tit.: soil profile. in inehe~
per hour
S. Hydrologic soil group and wa lcr capacity for soils (A. B. C. or DJ
according to NRCS met hodology
6. Land us.c of the basins as a perccntage of drainnge Hrea. incl uding
(a) Rural an d pastu re
(b) Agricultural
(c) Low-de nsity res identi al (0.5 10 2 acres per dwctl ing)
(d) Mo.:diu m· dellsity rcsidc lllial (3 to 8 dwellings per acre)
(e) High·density residential (9 or more dwe llings per acre)
(f) Commercial
78 Chapter 2
(g) Induslri nl
(h) Under cOnst ruction (bare surfacc)
(i) Idle or ... ac~nll a (1 d
U) Wetland
(k ) Parkland
7. Detention ~ l orage , in ae-ft of 5torage, and detention storage in lit-ttl
Hcre of basin
8. I'c rcelll of wlI lershed ups tream from de ten tion storage
9. Pen.:cll l of area drained by a SlOrm sewer system
2.2 Sherman ( 1932) originally advanced the theory or the unit hydrograph (UlI).
....IT de fined 35 "basin o utn ow reSUl ting from 1.0 inch (1.0 mm) of direct runoff
HYOIIOGRAPH ge nera ted un iformly over the drainage: area at a uniform rainfa ll rAte during
niEORY II. specified period of rai nfall du ration." An import an t point here is that UH
is comp<>licd of 1.0 inch of di rect runoff, which is equi\'alenl to 1.0 ioch of net
ra infa ll for a gh'e n d uration, D. and therefore all los5t!s to inriltraliOIl must
be subtracted before computat ions, Scvcralnss umptions inherent in th e uni t
hydrosra ph approoch tend to limi t ils applic;llion for any give n watershed
(Johnstone and Cross, 1949):
1. Ra infall e ltceues o f eq ua l du ra tiou are ass um e d to produce
hydrogra phs with equivalent t illlc ba~es rega rd less of the inte ns it y
o f the rain,
2, Direct ronoff ordinates for a 5torlll of gi\'en duration arc assumed
di rec tly proportional to rainfa ll excess volumes, Th us, twice the ra in-
fall prod uces a doubling of hyd rogrol)h ordi nates,
J. The timc distributio n of d irect run off is assu med independent of
anteceden t precipitation.
4. Rain fall dist ri bution is assumed 10 be the same for all storms of equal
durat ion. bot h sp~ljally and temporully.
The classic statement of unit hydrograph theory Cll n be su mmarized
briefly: The hydrologic sys tem i5 linea r and time im'ariant ( Ooogc. 1973).
The linear assu mption implies th at complcl[ storm hyd rographs ca n be pro-
duced by ad ding up individual unit hydrographs. adj usted for rai nfall vol·
urnes and addcd and lagged in time. For example. a 2-inch r.linlall in X flou rs
will produce double the response of I inch falling in X hou rs o\'er a giwn
watershed. While th e as.<;umptions of linear ity li nd time invariance are not
strictly correct for every wa tershed. we adopt them as long as th ey are useful.
Nonlinea r cl[amplll$ el[ist in ope n-channel flow, laborutory runoff models.
and actual wa tersheds.
Hydrologic Analysis
figure 2- 1
~ ," Rsinfalllo5ks
Unit hydrograph de-
termina tion. lallolol
!, <S
RaIn fall eJ(Cft$
slOrm hydrogropf..
I--I,"~
Ib) Hyd rog ra ph
' 60 ', -1 m in u~ boseAow,
..
" '" --------------- lIa ~ 11..,...
t n n ~Ct
point
unit hydrogroph.
'"" , , 3 , , , , 6 8 10 11 12 13
Tome ( hf J
,.)
"..,
•
.5
",., Vol rainfoll - 2 ,n
1-<1' _ 4 h r ~
.
..:: '"
2.0 in.
Q,. - tOO d,
" "I , ,
'. (hr) '"1 "
6 8
Tome
.~
")
__________________--,
onQo-'_
>
..
"" ., -
~ 1 ;". d"c<.1
funolr
",f-C--- l',--'--~I
2 ~ 6 H 10 12
I'
,.,
Tim<: (hf!
80 Chapter 2
a risi ng limb. a CTes t segmenl. and a recession curve (see Fig. 1-26). T iming
para meter.; such a~ dllration D of rainfall exccss and lime to peak. Ip, arc
also illustrated. The main tim ing aspects of the hydrograph can be cha rac-
terized by the following para meters:
J . !)uratlon of rai nfall excess (D): the ti me from start to fini sh of rain-
fall Cl.cess.
2. Lug tim e (Ip): the time from the ce nter of mass of ra infa ll excess to
I he pe ak of the hydrogmph.
3. Time 01 TiS(' (TH): the lime from Ihe slarl of rai nfall excess 10 Ihe
peak of Ihe hydrograph.
4. Time ufconcenlration (Ie): the lime for II wave (of wa ter) to propa-
gale from the mo~ l diMan! point in the walcn.hcd 10 the oU llel. O ne
estimate is the lime from the end of net rainfall 10 Ihe inflection point
of the hydrograph.
5. Ti me base (Tb): the 10lal duration of the O RO hydrograph.
If rain rall~runorf lIata exist. it is use ful to devel op uni t hyllrographs
from measured ra infa ll even ts and streamnow (rom gages in a deri ned wate r-
shed. The following general ru les should be observed in developing uni t
hydrogra llhs from gauged watersheds:
1. Storms should be selected with a si mple struc ture wit h relatively
uniform spatial an d temporal distribu tions.
2. Wa tershed si7.es should ge nerall y fall betwee n 1.0 and 100 mi2 in
modern watershed ana lysis.
3. Direct runoff should range from 0.5 to 2.0 in.
4. Dura tion of ra infall excess D should be approximately 25% \0 30%
of lag time II'"
5. A numbe r of SlomlS of similar duration should be anal)'7.ed to ob tain
an ave rage uni t hydrograp h for thai du ration.
6. Step 5 should be repeated for st,'e ml rainfalls of differen t d umlion.
The following arc th e esse ntial ~teps for developing II unit hydrogrllp h from
single storm hydrogra ph (see Fig. 2~ 1 and Example 2-1):
It
(or th e basi n (Fig. 2-1 (c». The time base Tb is assumed oonstant for
Storrru; o f equal durat ion a nd thus it will no t change.
5. Check the volume of the unit hydrograph 10 make sure 11\;11 it is 1.0
in. (1 .0 mm). and graphically adjust ordinate s as required.
<
C"Q:g
.$
~
1.11
-;: OS
rl{jure E2-lo
l1unoff hydrogrcph
with ooseHow ond
rainfall
I 1 3 4
TII"~ (Ilr)
,..,
"C+=2~:3!=~.ii!i·,!lIi:i·,~:,;::,l=:'!=~'~O:::J"
Time Ih.)
First we find the net rainfall. After subtracting the infiltration. we plot SOLUTION
the raillfllil C)[CCSS hyetograph shown in Figure E2-1b. This represents
2.0 in. of rainfall, or I in J hr for 2 hr. Then we subtract base from nO"
all the now values. Finally, the hydrograph must tJ.e converted to 1 in. of
dirt:ct runoff over the watershed. or 0.5 in.lhr of rain for 2 hr. To do so,
figl.n f2-1 b
F on ~rlr;;:;;.,;::::::;---] Relultong
.~ 0.5 -lL_'._T,-+---i hyetogroph.
2 3 •
T""" (h.)
" (hopler 2
we I:lke each ordioMe minus its base flow and divide iI by 2. This en tire
proced ure is tabulated as follows.
• ••
'00
I
2
100
300 200
• '00
,,
3 700 000 300
1000
800
900
700
".
35.
0 000 '00 25.
7 ' 00 300 ".
,.
8
9
300
200
'00
100
200
100
••
100
'"•
•
"
The 2-h r U H g r<l phs as shown in Figure E2- 1c. T/;. the lime base o f the
staml. h q hT. and the lime 10 pc:,k,,. nlcas urcd from the celllcr o f mass
of rainfa ll is 2 hr.
Figure E2-1 c
2.nr UH. . ,
~
2-h r unit hydrO@.faph
"
..
.~
SOO
i '"' 200
" 100
, , , o , • 9 10
1 •
'.TIme{hr)
S<urve Method
A U H for a pa rtic ula r watershed is defined for a specific duration D of
rain fall excess (RF"",), T he li nea r prope rt y of the UII can be used to ge n·
craIe a UH of la rge r o r smalle r durati on. For example. gh'en a I· hr UH
for a particu lar watershed, a UH resulting from a 2-hr un it storm can be
dcvel opcd by add ing two t-It r UI-ls. th e second one lagged by I It r, addin g
together th e ordin at es. and dividing the resu lt by 2.l n this .... ay.t he I in.
of ra infa ll in 1 hr has been dist rib uted uniform ly on'r 2 hr in deri ving thc
2· hr UH (Fig. 2- 2), This laggi ng procedure is restricted to in teger multi-
ples of the origina l duratio n. In thc sa me w ~ y , the 3·hr UH results from
addi ng and laggi ng th ree logether and dividing Ihe ordi nates by 3 (see
Exa mple 2-3).
-' be S-curve met bod allows construction of a UH of any durati on.
Assume th at a UH of du ration IJ is known and Ihil t we wish to genera te a
UH for the same wutcrshcd wit h du rat ion D'. The Ii"st Stcp is to ge ne rate
the S·curve hydrograph by ad ding a series of UI-Is of duration D, each
lagged by time period D. This corresponds to the runoffhydrograph result-
ing from a co ntinuous rainfall excess inte nsity of l ID in.lhr. wherc D is
measured in ho urs. Note th at an equi librium hydrograp h (S-cur\'c) res utts
1.
liD '
I
R.i nf.lI cxCe.. for new hydrograph
+-
'/----D.- j
Figure 2-3
Graphical repffw:mlalion of !he 5curve hydrogmph method.
from this addition of many U Hs, each Jagged by D hr (see Fig. 2-3), lind is
eq uivalent 10 a continuous rain fall int ensity falling un til inflow a nd o ut flo w
are equ al.
By shifting th e S·c urve in li me by I) ' hr and SUbU3 c ling o rdinates
be tween the Iwo S-CUTVCS. the; resulting hyd rograph ( Fig. 2-3. c urve b)
must be d ue to rainfa ll of l I D in.lhr thll l occ urs for 0 ' hr. T hus. 10 con-
vert curve b 10 II UH , we must mul tipl y all thc hyd rograp h ordill Ples by
0 / 0 ' resulting in the U H of duration D ' . A homework problem explores
these calculations in detail. T he procedure is ideally suited for an Elcel
sp read shee t.
Tab&. E2-2
r.... (Iv) PrU. ~U. ~U. p.U. P$U. Q.
0 0 0
I
'"'""
SO 0
2 100 100 0
,,,
3 225 320 ISO 0
IS' "0 '"0 0 0 1115
125
'"370 0
" ""
100 ""
80
'"2" 0
,,
100 0 225 'OS
7
", J75
10
11
20
" ,
00
0
" 0
135
00
0
0
0
'"'25
80 .,
'"
300
"13
0
0 "20
0
"'",
Fi1fure 2-4
Grophital repre-
scnlalian of a unit
flydrOSroph Ie< a
multipoeriod ste<m
la)
u, u,
u,
u,
u,
u.
U,
02345678
Q,
Q"
o , >0
(,'
dc term innnl. T he motrix [PI dues not meet this condition, but by using the
transpose matrix (pT], a square symmet rical matrix (pTpJ can be ge ner-
ated. The ma trix [pTPI can then be used 10 obtain
( 2-4)
This equat ion ge nerales simultaneous nor mal equa tions for a least-squares
solution. Eq uation (2-4) Call be solved explkitly for rVI using
(2-5)
Equation (2-3) can be solved us ing the Gauss eliminatio n method of solv-
ing simultaneous eq uations o r Excel's mat rix solver tools. As with any nu-
merical procedure, unde r certain conditions the error in the values of I UJ
. ehople, 2
ca n gro w rapidly. and unrClllistic or oscill ating values can res ult . cspeci:tlly
fo r latcr o rdinates of the U H. Note that Ihe decomposition proccdLlrc is
rardy used. and it is usually beHeT to develo p UHs from si mple SlOrms,
if possible.
,• 80
60
,,
6
"
20
0
,• 80
60
"
100
80
25
"
100
,, " 60
6
2.
"
20
"
60
9
0
0 0 "
20
10
• 0
•
Art er we have lagged t he U I-I, we must add the thr ee un it hydro-
gnlphs toget her a nd divide by 3 ill ou r case to adjust to 1.0 in. of direct
runoff.
Hydrologil; Aoolysis .
.... "' .
l"'.UH
....UH
1(adOId 3 I-hr UHl/3J
0 0
25 '.3
.""
25.0
58.3
230 76.7
240 80.0
,"0 00.
'20 400
00 20.0
2. 67
• •
The new peak now is 80 cfs at I '"" 5 hr. To find the area o f the water-
shed. which is usuall y known, add up Ihe ordinates for each hou r of the
hydrograph and muhiply by the incremental hour.
In our case, the total flow is 375 ds ( I hr).
To find the area: (note I aere·inch "" 1.008 ds-hr)
375 cfs (I hr) = 372 acre-inch
A number of cquations have been developed o\'e r the years to predict many 2.3
oft hc above uni t hydrograph pa rameters liS n function of measurable wale r- SYNTHETIC UNIT
shed charactcristics. and seve ral of the more popular ,Lpproaehes arc pre- HYDROGRAPH
se nll;:d in this section. With the increasing use of computer models in DEVl:LOPMENT
hydrology and GIS met hods, the synthetic methods have generally replaced
the gauged-based mct hod for most watershed~. especinll y since many wa ter-
sheds a re not g~gcd.
Me thods for derll'ing UH for ungaged watersheds have evolved based
on theoretical or empirical formulas relating hydrograph peak now and tim -
ing to watershed cha racterist ics. These are usually referred to as synthetic
unit hydro;raplas (UH) nnd offer the hydrolog ist or engi ncer a mul titude of
mcthods for de\'c1oping a UH for a particular hasin. Mos\ UH met hods were
de\'eloped in the period from 1932 to 1970. and they still provide onc of the
most useful lmd accurate approaches for hydrologic prediction for a gh'en
f<.L infa ll event.
Synthet ic UHs. once dn'c1opcd for a watershed arCH. can be used I'o'; lh
historical or design rain falls (see ElI:anlple 2-2) to produce SlOrm hydro-
graphs lLt Ihe outlet of the watershed. As the waters hed changes ovc r time,
the UH ClLn be updAted to be ller represent land use and channel ~I terat ions.
Synthetic UHs deve loped along two main lines of Ihought: onc ass umed Ihat
each watershed had a unique UH related to specific watershed characteris-
tics, and the second assumed Ihal all Ul-b (ould be represen ted by a single
. Chapter 2
fam ily of curves o r a single equation. However, the fo rmulas all have ~nain
limi ting assu mptions and should be applied 10 new areas wil h ext reme ~au
lion. Some calibra tion \0 adjace n t wa tersheds whe re sl re am n o w gages e xist
shou ld be 1I1tcmplcd, if possible.
The first linc of dC\'elopment was based on th e rational melbod mod i-
fi ed to include the travel time 10 the o ut le t for a particular wu te rshe d. Oark
(1 945) ass ume d Ih al wate rs hed rC!;ponse wo u ld be given by ro ut ing the lim e-
a rea relaliol1 ship through an element of linear storage (sec Eq. 2- 23), which
tends \0 attenuate and lime-lag the bydrograph. T he Oark unit hydrograph
is contained in H EC -HM S and is prese nted in de tai l la le r in this stttion and
in Table S-1O, The second approach 10 U H developmen t ass um ed mat he-
mal ical repre sentations fo r the shape of thc U H. A use ful a pproach was
ad vanced by thc Soil Conse rvation SCT\"ice (SCS. 1964. 1986), bruied o n mea-
sure men ts from tlio uSlI nds of small watcrshed areas. which re presen ted a
dime nsio nl ess U H of discharge \IS. li me by a gam ma funct ion (discussed
late r). Since volum c is fixed, only o ne parameter is requ ired 10 det ermine
the e ntire U I-I. ci thcr the lp o r the peak flow rate, Qp.
In the late 195(Js, both approaches bega n 10 conve rge after the inl po r-
tan l contribut iollS of O' Kelly ( 1955) and Nash ( 1958. 1959). O' Kell y's work
was based on replacing the time-are a cu rve by a U H in the shape of an
isosceles tria ngle. E mpi rical cxprCSl;ions ..... ere neccssary to transfo rm the
mct hods of Clark (1945) into u~a bl e U I-I techniq ues for nct ua l basins. Jo hn-
stone and Cross (1 949) proposed one of the fi rst rela tio nshi ps for Ie- the time
o( concentration in ho urs, de fincd earl ie r in th e chap te r.
(2-6)
a nd fo r 5101"1\l8C delay lime. K. in hours_
K = 1.5 + 9O(A I LR), (2-7)
where L is the le ngth of the mai n sne am in mi, A is the area in mil, S is Ihe
slope in ftlm i, and R is an overl and slo pe factor.
Hydrologists ha ve realized that 11 number of pa ramete rs are important
in determ iniug the shape and liming of Ihe UH for a watcrshcd. T he lag o r
time to peak Ip and the time of concen tr ation I" arc o ft e n used (Fig. 2-1(a».
The time o f rise TN measured from time 1.ero to th e hydrog.ra ph pea k is
some times used. T he tim e base "fb of the hydro&ra ph is included to define
the duration of di rect ru noff. These ti ming para mete rs must be statis tically
or theore tically related to waters hed char acte ristics in developing II syn-
thetic U H.
The discharge pa ramc ter most ofte n tilled is the pea k disch<uge Q p. A
rouling pal"llmeter K is sometimes incl uded whe n the hydrograph has been
routed thro ug h a linear rese rvoir wi th storage delay lime K . This te nds to
delay and atte nuate the input hyd rograph. W at e r~ h ed parallleters of mos t
co ncern include area A . main channel lenglh L , length 10 wa tershed a:n trnid
L". and slope of main channel S (see Section 2.1 and Fig. 1-22).
Hydrologi<; Analysis
q,-- ---
" Q.
'. '.
P. rhally dc"clopod Fully !lt,..,1oped
(b) (e)
"J
Figu,. 2-5
Modifying Ioclon on unil hydrogrophl. (0) Nolurol watershed development, rep!".
sented by (Uf\'(! a in the bonom pori of the fig....-e. (b) Porliol development, repr.
so",ted by curve b. (c) Fully developed waten hed, represented by curve e.
92 Choprer 2
Figure 2-6
Changes in Iood u$8
(OUM! chong •• in 'f'll~
IO.<XX>
1959
o , 10
"
Tin, e( h' l
Most orthe methods for synthetic UHs re late lag time l p or time of r isc
Til of the hydrogrn ph to measur es of the length of the main challnt\] and
shape of the basin. Some met hods also re late ti ming to th e ilillerse of the
slope oC the mai n cha nnel or land. Thus, the longer the basin and the SIl1~l l1c r
the slope, the greater the lime of rise of the hyd rograph. as expected. Fig-
ure 1- 19(a) (Chapler 1) shows the general effect o f watershed shape and
land use (i.e .. elongated liS. concentrated) on U H respo nse.
A second rela tion is usually prese nted be' we en peak flow Qp and area
of basi n. and be twee n Qp and the inverse of the lp or Til of lhe hydrograph.
thus ind icating Iha l large r area s produce higher Qp. From continu ity, the
highe r the pea k now. the smaller Ip must be to keep the volume of the unit
Hydrologic Analysis 93
hydrogrll ph const;mt with 1.0 in. of direct runoff. This is ellsily seen by refer-
ring to Figure 2-5. where cu rves II. b. and (' nrc all unit hydrographs. Thc
following sediOll5 describe a few of the more popular sy nt hc tic unit hydro-
graph methods. Specific areas o f the Uni ted States have deye loped their own
empirical mcthods th nt are similar to the standard methods. but based on
IOCll1 wlltersht-d data.
Snyder'1 Method
Snydcr (1938) was the first 10 develop II synthe tic UH based on a study of
watersheds in the Appalach ia n High lands. In basins rangi ng from 10 to
10.000 mil. Snyder's relations are
(2-8)
where
= basin lag (h T).
Ip
L = length of the main st ream from the outlet to the divide (mi).
L~ = length along the main stream to a point nea rest the water,;hed
centroid (mi).
C, = coeffi ci crl1 usually ran ging from 1.8 to 2.2 (C, has been fou nd
to vary from (),4 in mountainous areas to 8.0 al ong the Gulf of
Mcxico).
(2-9)
where
Qr> = peak discharge of the unit hydrograph (cis).
A - dra inage arCll (mi 2).
Git = storage coefficient ra nging from 0.4 to O.8. where larger \'alucs of
Cp arc associated wit h smalle r values o f C,.
(2- 10)
where T~ is the time ba~e orlh e hydrograph, in daY!'_ Fo r small wa lC rslreds.
Equllt ion (2-10) should be replaced by multiplying II' by II val ue that varies
from 3 \0 5 as a be tt cr cst i m~te of T". Equntions (2-8), (2-9), <lnd (2-10)
define points for iI un it hydrograph produced by an excess rainfal l of dura-
tion f) c tp / S.5. For ot her rain fa ll excess duratio ns D'. nn adj usuxl formula
for Ip becomes
'i, = Ip + 0.25(/>' - Dj, (2-11)
where I; is th e adjusted log tim e (hr) for d uration D ' (hr). O nce the three
qu an tities Ip. Q p. and T" are known. the UH can be skctched so th aI the
area under the <:u rvc re presents 1.0 in. of direc t runoff from the watershed.
Snyder's method Ili a popular method because of its simplicity. CaUlion
should be used in applying Snyder's met hod 10 a new area ",ithout first dcriving
Cfogpler 2
"
coefficie nts for gaged streams in the gene ral vicinity of the proble m basin. T ht:
cocCficienlS C, and Cp have been found to vary considerably from o ne region
\0 anothe r. E'{ample 2-4 illustrates Snyder"s method. Width equations a\ 50%
and 75% of Q" arc presented in Example 2-4 \0 help sketch the Snyder UH.
SNYDER'S METHOD
Use Snyder's method to del'clop:I UH for the area of 100 mi 2 dCliCribcd
below. Sketch the appruximate sha pe. What dur.. tion rainfflll does this
correspond IO?
C, '= L8. L = 18mi,
C" = 0.6, Lc = IOmi
-""'""
indicated .
~Vn
jl
" ..
,
"""
9 12 15 18 21 24 27 W 33 36
nine ( hr)
3-1.4 hr
Area d .... "'n to
rcprtl/'nl I In. of rudolf
over tho ...·al.Bh<...J
Hydrologic Anolylil .s
Since this is 8 $111811 wll(cn.hed,
Tb ... 41,. = 4(R.6) hr,
Tb = 34.4 hr.
And the duration of rainfa ll
D = t,.iS.5 hr,
= 8.6/5.5 hr.
D = 1.6 hr.
FinalJ). the h)"d rogrnp h should be ~ lI1oot hed to represent 1.0 in. of di rect
runoff.
SCS Methocb
The methods deve l o~d by the Soil Conservation $cn 'ke (SCS. 1957. 1964)
arc b;1~d on a dimensionless h)'dtugrapb, developed from a large number of
UHs from gaged watel"5heds ranging in size li nd geograph k locaTion (Fig. 2-7).
The SCS is now called the Nll lural Resouf(:es Co nservation Service (NRCS).
More dcta ils and examph:s on SCS methods can be found in McCue n (2005).
T he C;lrl ieSI method assumed a hydrognl ph represenTed as a simple tri angle
(Fig. 2- 7). wit h rainfa ll du mlion D (hr). tim.:: of rise T R (hr), time of fall B (hr).
and pea k flow Q,. (d s). The "olume of direct run off is
QpTH QpB
Vo l = - - +- (2- 12)
2 2
0'
2Vol
Q,. = TN , B' (2- 13)
where
A = area of basi n (sq mil ,
Tf/, - lime of rise (hr).
.
Fig.... 2- 7
Chopre. 2
Q,
I
B
(. )
0.8 {--V. W
0.'
~ 0.6
"IS U.5
~
(S. 0.4
OJ
"
0.'
o :2) " 5
'''.
I')
Capece c\ al. ( 1984) foun d 1hal It fltCto r as lo w as 10-50 holds fo r flat, high-
w:lter-Iable watersheds ra lhe r than the val ue 484 present ed he re. Me C ue n
ind ica tes a possible ra nge from 300 10 600.
From Fig ure 2-7 i1 can be sho wn tha t
(2- 17)
wh e re
o= ra infall du ra tion (h, ),
I" = Jag time from ce ntroid of rainfall 10 Q" (hr).
Lag time Ip is eslimated from anyone of seve ral empirical e<lualio ns used
by the SCS. and thc one tha t is often reported is (note that un its must be
carefull y followed in the SCS melhod)
L o'~(S + Il,·7
Ip:: 1900Vy (2- 18)
where
Ip s lag time (hr).
L - length to divide (fl).
)' :: avcrage watershed slope (in percent).
S :: I{OO/CN - 10 (in.). (2-19)
CN - eur~ e number for various soi l/land usc (see Table 2- 1).
The SCS dimensionl ess UH can be used 10 dc~elop a curved hydrogra ph,
using the same Ip an d Qp as the triangular hydrograph in Figurc 2-7 (see
EJtample 2-6). The SCS also found that time of concen","ion ft: was equal
101.67 limes Ihe lag ti me above.
SCS (1964) runoff estimates IIssume a relationship betwee n accumu-
lated total storm rainfall P. ru noff Q. and infi lt rat ion plus init ial ~h!;tT3c tion
(F + ,~), Based 011 SCS wa tcrshed st udies
The SCS me thod uses thc ru noff cu rvc number CN. rela ted to pote ntial
abs traction 5 by CN = 100'1/ (5 + 10). or S (i n.) = (IOO) / CN - 10. Fig-
ure 2-8 prese nts SCS Eq uation (2- 22) in a gra phical form for a range of
CN va lues and rain falls.
Runoff eurve numbers for sekctcli land uses are presen ted in Ta ble 2-1.
where hydro logic soil group A is sa nd y and well drai ned. group 11 is sandy
loam. group C is clay loam or shallow s:mdy loom. and group 0 has a poo rl y
dnlined. he:lvy plastic clay tha t swells when wet. Group A has the highest
infilt ralion capaeit)' and Gro up 0 the lowest. The CN values in Figurc 2--8
assume norma l an tecedent moisture condition II. and ot her antecede nt
moist ure co ndi tio ns and effects of urbanizat ion can be deve loped using the
SCS report . Urbtlll Hy(/rology for Smnf/ Wmcrs/tn/l' (SCS. 1986). For a
98 Chapter 2
Figvre 2-8
Graphical iQllllion of loinlalkunoff eqoorion.
SOLUTION
""""~
woodod
"" """"
,
frocIion of Ar ... CN
C
,
0"'[0.5) ~O.2
0.4(0.5] -02 "
70
R...d.nt,ol
C
0 .6(O.5J - 0.3
0.6(0.5} .. 0.3 "
BJ
HydrologiC AnOlY1iJ ..
The weighted CN is
CN = 0.2(55) + 0.2(70) + 0.3(75) + 0.3(83) or
CN = 11 14 22.5 24.9 = 12.4
o r. using CN = 72. runoff volume is 3.9 in for the given rainfall (Fig. 2-8).
W hilc SCS mcthods are used e~ l cns ive l y in cngi nccring pract ice, ;! nd a
PC version for lI rbll n wa t ersh cd~ (TR-55) is availab le (SCS. 1986). th e
methods have some wca k nes.~es. as ])oim cd out by Capece et al. (1984).
II is difficu lt to mat ch measurcd hydrog raphs in areas with hig h water
tables. and the va rio us antecedent conditions (I. 1£. and [If) cannot han dle
the proble m accurat ely. The Strength of the SCS method is Ihe enormous
da tabase of soils in forma tio n. soils maps. and si te-spccifi(: rain fall- runoff
studies. M(:Cue n (2005) provides more details on SC:S assumpti ons and
design criteria. as compared 10 ot hcr opera tio nal comp uter models in hy-
drulogy. Ex ample 2-6 ill ustrates the SCS uni t hydrograph me thod. based
on using Eq uation (2- \8) for lag time.
Equlltion (2-18 ) gives the following rc\atio n:;hip for Ip: SOLUTION
LO.8(5 + It·]
J91X1v'Y
From Table 2-1, the SCS cur.·e number is found 10 be 78. Therefore.
S = lOOO;C\ - 10
- 1000/18 - 10.
S'" 2.82 in.
Con~erting L = 5 mi. or
L .." (5 mi)(52HO ftlmi) = 26.400 ft
100 Chapter 2
The slope is 100 rt{mL so conve rt [0 percent for Ihe equa tion
y = <100 ftlmi)(l mil5280 fl)(I00)
= 1.9
,,'
, ~ [ (26,400)0.8(2.82 + 1)0.;]
p 1900 Vl9
= 3.36 1Ir.
From Eq uation (2- 17) and wilh rai nfall du ration 0 = 2.0 hr,
TJ(= D/ 2 + II'
'0
B - 7.17 hr.
The trianguln r unit hydrogra ph is silo" 'n in Figure E2-6: not e Ihe lime
base of [IS hr. The ne),:' example demonstrates the usc of the dimcn-
siOlllcss SCS U H; for thi s example. T"blc E2-61ists the resul ting shaped
SCS hyd rograph.
Hydrologic Analysis 101
.
0 ,077 0967
"
JO 0055 0.977
0,989
12.21
13.1
14.B2
85
61
"
50
0029
0011
0
0.997
1.0
17.<14
2L8
J2
12
0
FiglJre E2-6
Center of mass ol raiDfall (I.o hr) Tfiongular unit
hyd'09rtlph for
D 2.0 ru- Example E2-6.
o , , o • "
Time (hr)
dan.. UH MeIhod
The Clark (1945) UH is based on the U5C of a watershed and is modeled n.~ a
linear channd in seril'S wi th a linear reservoi r 10 account for tra nslation lind
HHenuation. respectively. This conce ptual model defim:s the outflow from
the lincar clHt llne l as inflow Oi) to the lincar reservoi r. and the outflow
102 Chapter 2
(OJ) from the linea r re!ic rvoir as the instantaneous un it hyd rogrOll'h (IU H).
The linear channd uses an area-time relationship (i.e .. isochroncs of travel
lime 10 Ihc o utl et of watershed). and il is used 10 estimate the lime di stribu-
tion of ru noff [ro m th e basin. T ime of concentration is rcpn:sc nted by thc
time of runoff from the most remo tc part of the basin \0 t hc ou tlet. The
Jine ar reservoir signifies the co mbining effecls ofsloragc and resistance from
Ihc basin. Hoggan (\997) described the Clark method in details using Ihc
concept o f thc instantaneous unit hydrograph. D erivation sta rts ",ilh the
rela tionsh ip bet ..... een Ihe reservoi r outflow and the storage is ex pressed as
s, = ROj , (2-23)
where
51 '"' storage at en d of ~ri od i,
OJ = outflow d uring period i,
R - storolge coefficient.
The conti nuit y equa tio n can be modified as
_ Oj~ 1 + OJ Sj - SI~1
Ij - , =
A, (2-24)
whe re
1; '"
average inflow in period i , determined from the :lrea-t ime
met hod.
OJ - uutflo w during pe riod i.
Sj - slo rnge III Ihe end of pe ri od i .
Subslilllti ng Eq uati o n (2~23) into Eq uat ion (2- 24) yiel ds
T. _ Ol - t + 0 1 = ROj - ROj~1
(2~25)
! 2 tJ.1
C -
,.,
",,'-;''c: (2- 26)
2R + il t
T he outflow (0;) from the line ar rese rvoir can be ob taine d with Ih e flow
rates li nd d ura ti o n ( 6. t) by co mbining Equa tio ns (2-25 ) a nd (2- 26) as
fo ll o ws:
OJ = cT; + (I - C)Oj-l (2-27)
Nume rical solutio n) of Equation (2-27) can be used 10 plotl he insta nta-
neous un it hydrograph (IUH).
Two majo r parameters of the Oa rk uni t hyd rograph are Ihe time of oon-
centratio n T~ and Ihe SlOrage oocfficiell\ R. which is the slo pe of the sto rage-
outflowcutve for the linear reservoir. T~ and R values can usually be obtained
Hydrologic Analysis 103
from observed hydrogr.a phs for gllged basins: T~ is eslim3h:d 3S Ihe lime from
the end of a bo ["!;1 rdinfall excess 10 the inflec1 ion point on lhe receding limb; R
is estimaled by diyKJing the dirCCI-nmoff discharge at the inflection (X'lint by
the slop<: ofthecun"c althal point, The inOectioo point mcans where inflow 10
the li nea r reservoir is lcro. Atthc inO cctio n point. the oontinuilY eq uat ion can
be expressed as
(2-28)
By substitu ting Equation (2- 23). n new equa tion for R is der ived:
(OJ-! + 0;)/ 2
(2-29)
(0, 0, _1)/ 6 ,'
Equmi on (2-29) theo ret ically produces [he value of R as the mt io between
Ihe ave rage discharge al th e inflcctio n (X'lin l and the negative slope of the
hyd rograph allhis point (which is negative).
For ungaged basins. Tc- is usuall y de te rmi ned from 3n analysis of tra\'cl
limc in the basin or from regression relat io nshi ps that havc been dcvclo ped
(or the region. Even though the storage oocfficienl (R ) can not be obtai ned
fro m measurable wate n>hed character istics directly. many regress ion elj ua -
li ons h.we bee n de vel oped from so mc rcgio n~. Be,ause T,. an d Rare in te r-
dcpcnd e nl. lhe ,ombi ni ng va ri ables, Tc + Ra nd R/(T. + R) (ITC " Iwa ys
u ~ed in un il hyd rogra ph cornput mions. Thus. the C lark me thod is now e:l ll ed
the TC a nd R me thod. and is rout inely used in HEC-HMS models. as shown
in Table 5-10.
O nce a VB of give n d um tion based on kn own Siorms or synt hetic lII et hods 2.'
has beell developed for a part ic ular ha si n unde r a given set of ph ysiO- APPUCATIONS
gra phic ,onditions, it can th en be ut il ized for a num ber of hydrolog ic cal - OF UNIT
cul ation s (Fig_ 2-9), S uch hydrograp hs can be d~veloped for :llmost a ny HYDROGRAPHS
area ..... here VH methods hnve Deen d eveloped or by app lyi ng ge neral
me thods such as SCS's or Snyde r' s. In pa rticular. for a single h;lsin area.
the U H can be used ..... ith a given stor m eve nt 10 delermine th e storm
hydrog rap h through th e ad d· and -Iag proo:.:ed ure. For com ple,," watersheds
of Se \'eral subbasins:
I. Design storm hydrograph s for selected recu rrcncc-in ten'al sto rm s
( IO-yr.25-yr, IOO-yr) elln be de ... elopcd Ihrough convo lut io n (addi ng
and laggi ng) procedures for a givcn wate~h ed area (Fig. 2- 9(a».
2. Effects of la nd use changes, channel modificalio ns, storagt: additions,
a nd ol he r variables can be tcste d to determi ne chan ges in th e VII
(Fig. 2- 9).
104 Chepler 2
2-9
F"tg Ur1!
Unil hydrogmph •.,•
0 •;!•
0 Storm lIydrograph
opplicotions.
(oJ Development "
of c design storm
hydrog roph .
(bl De--",Iopmenl
of (] watershed ~ ~
hydrogroph.
,~
'nmc
Unit h)'drOlr. pl> Unit hydrogr~ph coovolU!1on
'"
•
•
'1
•.••,
" "
"
I
•I
'"
,,
A ,,
, c
.... -" -..... _."","'';--
, --:--;;=-:~
i
~
~ ..-s,~~
, '
De>"" oprncnt ,/
'"
3. Storm hyd rographs for each subbasi n can be simul ated by adding,
laggi ng, and ro uting the flows prod uced by unit hydrog ra phs
through channel reaches (Chapte r 4). Effects of vario us rainfall pal-
te rns and hind use distri butions I;s n be tested o n ove rall hydrologic
response of the large watershed (Fig. 2-9(b», as demonstrated in
Example 2-7.
4. Storage rouling methods (Chapler 4) ca n be used [0 translate inflow
Viis thro ugh a rescn'oir or detention basin of particular s;7.e to
aue nunte peak fl ow or lag lime to Ihe pcllk of the hydrograph. This
is used fur analysis of flood control options.
HydrologiC: Analysis IDS
~':mo=--,O'--,~__c"'-__CJ,-__C''-__~''-___ '=-__C' 8
•
UH, 0 100 300 450 350 250 150 100 50 0
UH2 0 tOO 4BO 720 560 400 240 160 BO 0
Figu... E2-7
. Suoorea 2 Suooroas I and 2 •
lde • .,topod) Subgre(l I is deveI-
aped, while subarea
.~uha..,a 1
2 i. und_1oped
(IInde"eloped)
By definition, the area under the U H curve (volume of direct runoff) SOUl11ON
divided hy !he area of the wate~I".:d equals one inch of ne! rainfall.
SUUl of UI [I ordinates = 175Ods-hr-1750ae-in. Sillcc this results in one
inch of ne! rainfllii. !he subHea I is 17.50 ae. (Note: I acre 0.00[5 sq mil
1750 ac = 2.73 sq mi.
Find the Io-yr dC!>ign stonn volumes for the 6-hr durallon and olher usc-
ful durations going down 10 I-hr from Figure I-IS. For a 6-)T Siorm. hav-
ing Ihe storm vo lumes for 1.2. and 3 hr will give a clear definition of the 3
hr of highest rainfall. After Ihat, the 6-)'r storm volume will do, as Ihere is
no need to be more specific thall to divide up the remuining low rainfall.
•J
,""
0.9
, ,"
" "
106 Cnoplef 2
Convert design rai nfall 10 a design l1et hye tograph (in J hr) by plnci ng
the highest rai nfa ll (t he I-h r val ue) al t hl: <:cn terof the Storm (inle n -al
2- 3 or 3-4). In this case, we will usc the in terva l J-4. The 2· hr I'n lne is
4.U, so 0.9 in . (4.0 in. - 3. 1 in.) fell between the pcnk I hr and the next
hi ghest hour (the earl ier in lcrv,ll of 2-3). The 3-hr va lue is 4.8. so 0.8
in. (4.8 in. - 3.1 in. -0.9 in.) fell from (-- 4-5 hr. This is the int e rval
direct ly aflc r the peak. so now the peak of the hyclograpb is lIcfincd.
Divide the rema in ing 0.6 in. of rai n ove r the remaining 3 hr. assum-
ing equal intensit y. a nd then subtrac ll hc infiltration losses each hour.
which yie lds.
, ; I ;.
11.,) [In. / hr) ~n'/hrI lin./hrl
0-1 02 0.15 0.05
1-2 02 0.1 0.'
2-3 09 0.' 0.8
3-. 3.1 0.1 3.0
"-5 0.8 o, 07
'-6 0.2
Toto l_ $.4 inc het
0.1 0 .1
Usc hydrogrnp h convolu tion (E~ccl spreadsheet) for each subarea. and
then combi ne the two storm hrdrogra ph~ to obtain the final sto rm hy-
drograph. Subarea I is shol\ n on the following page.
T_ Q
,
0
,
0
0 ,o
2
" 10 0
"
,,
3
"" JO 80 0 133
603
13 "
35
2<0
'60
300
900
0
70 0 1378
6
,
7 ,
8
"" 280 1350 210 10
JO
1883
1615
,
200 '050
'" 1173
9
3
0
10 120
80
"0
"0 '"'" "
35
'"
0 0
0
<0
0
300
150
10'
70 "
15
"0
'"",
0 35
0 ,
10
0 o
Hyd,oIogic Analysis 107
..,
r_
PIU. ~U. ~U. p.U. psU. p,U.
0
, "'" ,
,,
0 0 0
0
3 " " 0
"
•
J6
,,
7
20
'"
'"
'80
'"
'"
72
""
"
1876
1192
'"
378
0
" 0 ", 72
,
0 0
'1 he resu ltmg storm now IS calculated hy addmg thc flows together fr()m
sub~reas 1 and 2. Finnl peak now is 4895 ds at hour 6. ns ~hown below
in the shaded oolumn.
,-
.., . ,...,
0,
''''Q
' ' ', , '''
,,
0 0 0 0
., ,.,"
3
•5
,
'"" ,'"
803
1376
.. 1567
3582
85
1883 '"''
3012
."
4895
,,
7 16t5
1173
2~84
1676 3049
7" 1192 1937
10
'" '" 1222
""
235 J78
, '"'"
.." " 5
0
72
0 "0
The above computations show the final peak of 4895 ds 81 hour 6. They
can be easily done in Excel sprelldshccb. but for more complex wf\ ler-
$heds and more oomplcx rainfalls. where flood TOu li ng in slrclIms might
t>e needed. it is usually neCCs.~;lry to U!)C computer models such as IhO!le
presented in Chaplen; 5 and 6.
108 Chopte< 2
wh ere ;(T) is the rainfall e x~-c~s nt t ime r (Fig. 2- 10) and 1/(/ - T) elm be
viewed as a weighting functi on given to minfnll in te nsities th at lXcurred /It
lime (I - T) before. The int egml, kn own as the col1\'ol ution integral. gives
th e o ut put runoff ns a continuous function of lime f. Equa tion (2-1) from
Sectio n 2.2 is th e diSl;re tc form of Equation (2- 30).
The real usefulness of the lU H as a m~thentntica l concept can be see n
in relntion 10 the S-curve prese nted in Secti on 2.2. Considcr li n S-curve
formed by a con ti nuous Jainfall of uni t intensity. wh ich ca n be an infi nite
series of very smalltime un its, each sepa rated by T. Sumllli ng all the indi·
vidua l UHs of duration 1'. the S-cll rvc ord inatc5 become
SCI) "" 1'IU(l) + utI - r ) + u«( - 21') + .. . J (2-3 1)
since r also equals the depth for a ullit intensi ty. Then. as r ..... O in Ihe limi t
for the lUI-I,
Therefore. the S-curve is the integral of the lU H. and the IUH is the fi rst
deriva tive of the S-cur\·c. The slope o f the S-cu rve. ({Sl dl. is proportional to
th e ordinate of the rUH.
Hydrologic Ar.alysh 109
Figw. 2- 10
T Gropflkol r.prese,...
lotion 01how 10 .... se
on IUH 10 gencrote 0
-
'\
hydrogroph,
,(",IT
,
-"- t
,
'"
L. _ _
T he [UB has been widely use d. although it sufrc: rs the sanle disadvan-
tage as the UH in that rain fall- ru no ff respo nses mll y be nonlillcar and
de penden t on ante(:eden t conditions. Nonl inea r mode ls for the uni t h)'d ro-
graph hal'e bee n developed in the ge llenll literawre nnd nrc reviewed by
Chow ( 1964) and Raudkil'i (1979).
Nas h ( 1959) pre~nled a li m:llr UH mode l based on " re!;t: rvo ir,; in
series to represe nt a watershed. each wi t h a linear storage relo tion S - KQ.
that resu lted in II gllmm~ distributio n for shape. Figure 2- 11 ~ hoW5 various
shapes that resu lt for selecting Iwo pammcter,;11 and K in the model.
"8 ,, ,/
! ,,,
" , ,,, ••••
• ,, : :
, ,,. .J ,/
, ,-- .',
'0 • • " Tome (II,) " w
Figure 2-11
V<",ioos UH, 1><,,00 on rl. e Nosh lineor model.
water is simply balanced by the resistive forces of bed friction. A ll flows arc
assumed to obey the cqumion5 of contin uity and momentum (e.g.• Manning's
eq uation) as shown in Section 4,6. A complete de rivatio n of th e concept of
the kinematic ....8\'tl is prese nt ed there, and it is lI1 ent ioned here because the
kinematic wave mtlhod is a physics.based a pproach simi lar to Ih e unit hyd ro-
graph for ..... atershed areas. Modern distributed models solve these equations
nume rically over a fi nite grid that is used 10 rep resen t the va riation of eleva-
ti on and chan ne l configuration for 11 basin area (Chapte r II).
Solulion mel hods to a bove eq ull tions for th e ki n~ma tic wave using Ihe
meth od of characte ris ti cs are described in dClllil by Lighlh il1 and Whi tham
( 1955), Eagleson (1970). Ovenon and Meadows (1976). Ra udkivi (1979),
Ste phensou a nd Meadows (1986), and Singh ( 1996). Most pmetical applica-
t io ns of kinema tic wave met hods req uire the usc of nume rical methods
beca use of no nunifonn rainfall and va riable basi n charac teris tics and are
prese nted in Chapter 4 al ong wi th o the r num erical flood routi ng proced ures.
Modern distrib uted models arc based OIl th ese equa ti ons, and a n exam ple
model an d ap plica tion are presented in Chapte r 11.
Kinematic ( KW) wave routing can be used to deri ve overland flow
hydrographs, which C,IO be added to produce collecto r or chan nel hydro-
graphs and , event ually. can be flood ro uted as channel o r stream hyd rographs.
In the carly 1980s, the kinematic wave me thod was forma lly added to several
ava ilable compu ter models, such as the HEC HMS fro m the U.S. A rmy
Corps of E ngine ers (H ydrologic E ngineering Cente r, 2010). Figure 2- 12
Hydrologic Aool~i5 III
Q.
S~bbuin innow
------------------
, ,,
,, ,
,
, ,
,,
,, ,
,
L..
..
~<·-·~~-:::::c,...(~'I:no..S from
Main " coIle"I"" chan .... 1
----,, t ban .... 1
,,,
,
"
,,
,, , ,
,,
,
,, ,
,
,,
Q-
Sui>OOsin oUlfl;"..
wave flood routi ng in detai l, and Chapler 5 presen ts som~ applica tio ns for
the HEC· HMS models. The KW method also is a mai n drive r of the Storm
Water Manage ment Model. desc ri bed in C hap le r 6, lind th e d istributed
mode l VFlo. presented in C hupter 11 . Stephenson and Meadows (1986),
Singh (19%), and Vieux (2002) provide excellent relliews of kine mat ic wave
mod eli ng methods.
Mass transfer techn iques <I re based pri mar il y 0 11 the COll t"Cpt of tur bule nt
transfer of wate r vapor from a woter surface to the atmosphe re. Numerous
em pirical fann ulas ha\"(: been derived taexpressevaporation rale as a function
ofvnpor-prcssure differences and wind spt:ed above a lake or reservoir. Mall)'
such equations ca n be wrillcn in the fonn of a diffosive flux of water vapor.
whose conccpt da les back to Dalton's original theories (I3ru tsaen. 1982):
(2-34)
where
'5 = sal ura tion vapor pressu re at the T, of Ihe wate r surface.
c" = vapor pressure til some fi xed level above the wlltcr sur faee. the
prutl uct of relat ive hum idity limes sa tura tion vapo r pressure at
T~ of the air.
II - wi nd speed.
II. b = empi rical constanlS.
Chapter 2
'" An obstacle to comparing diffe rent evaporati on formu las is lhe vaTi·
ability in measurement heights for " and e". Reducing £l it form ulas \0 the
same measurement level of 2 m (6.5 (1) for wind speed and vapor pressure
and (liking in to acco unl lhc 30% difference between rese rlloir and pan evap-
ora tion. the scalie r be tween th e common formulas is considerably reduced.
The formula wi lh the best datubase is the Lake Hefner formula given by
Marcia no and H arbeck (1954), which also performed well al La ke Mead
(H arbeck , 1958). Some form ulas use a 7.ero val ue: of the CQnstant a in Eq ua -
tion (2- 34), due prob.1bly to sl11l1l1lo<:al air movemc llIS with velocities ins uf-
ficie nt \0 remove excess vapo r rrom above a pan surface. Harbeck and
Meyers ( 1970) presenl the formula
(2- 35)
where
£ ,. c\<l (JO ratiQn (em/day).
b = 0.012 for Lake Hdncr, O.01 18 for L.ake Mead
(em-d ay- J -m - J -s· m, ) ' ) .
~. - vapor pressure al wat er surface (mo).
~2 ; vapor pressure 2 m above water surface ( mb).
112 '" wind siJ(:ed 2 m above the water surfaee (mrs).
figure 2- 13 Q.
A lo ~e's _gy
budgct.
Q.
'" Q.
whe re
QN = net radia tion absorbed by t ll~ watcr body.
Qh - sensible heat tran sfer (conduction a nd oonv~ ction to the
atmosphere).
Q~ = energy used for cva)lQration .
Q6 = increase in energy stored in the water body.
Qy:O: advected energy of inflow and ou tflow.
Le tting L~ representlhe late nt hea t of va)lQri1.ation (Ctl.Vg) and R the ratio o f
lIeat Joss by conduction 10 heat loss by e\·aporation. Equa tion (2-36) becomes
QN + Q" - Q~
£ = pL~(1 + R) . (2- 37)
R :: 0 66
. (T, - T.)(-1000P)T
e, e.
' --'"
- T,
- - >-:-::-:- t, ta'
(2-38)
where
P "" at mosp heric press ure (mb).
T. - ai r tempe rature ("C).
T, = wate r surface tempe ratu re (QC),
e~ ... vapor pressure oC tile air (mb),
e, s sa tu rdtion vapor pressure al surface wa ter tc mpcrdture (mb),
"y = the psychometric consta nt O.66P/ looo (mbrC).
Pan EYClporation
Evaporat iQn ca n 00 measured from a sta ndard Weather Bureau class Aplin.
an open ga lvanized iron tank 4 rl in di ameter and 10 in. d;:ep mount ed 12 in.
above the gro und (~ee f ig. 1- 29). To estinHlte evaporation. the pan is fil led
t(l a depth or 8 in. and must be refilled when the dep th has fallen to 7 in. The
waler surface level is measured daily. and evaporatio n is computed as the
116 Chapler 2
Combined Methods
Pen mlm (1948) first used the best fea tures of t he mass tran sfer an d ene rgy
budge t met hods 10 derive a wat e r surface evapoT<1 tio n rel ation tha t is fairly
easy to .;:ompute. The Penman eq uatio n is
(2- 39)
wh e re
Ell = flllX of la ten t he al due to evaporlllion (e ne rgy/area-time)
'" pL~E wit h E in units o f LIT,
L . :z lIu ent he at of va poriJl;atio tl [Eq. (1-5)1, .;:u~ toma rily eval uated al
the tempera ture tlf the air (e ne rgy/ma ss).
d - slope of e,-vs.- T c urve, which is shown as ~ fy vs. T in Van Savel
(1966).
It is c us tomary to evaluate Ilfy or j ust d al the tem pe rlllUre of the ai r, no t
the te mperatu re of the waler surface. Alternati,·cly. since to re presents the
slope of sa turated vapor pres.~ure vs. temperature. it ca n be o btained by dif·
fe rentiating Equa tion ( 1-4), thus
Figu,. 2-14
Shallow Ia~" eY'Opo!'ation as (I lun<;rion of soIOI rodiorion, oi. lempcrolure, dew point, ond wind moY.,.
menl (Adopllld from Kgb).,. cr 01., 1995).
118 Chopt&r 2
SOWTlON r oin! 1 on Figure 2- 14 shows th e point whe re T" = 70"F an d the solar
radiatio n is 650 ly/d ay. Point 2 is drawn II I the intersecti on o f th e curve
o f T, = 5(fF and T~ = 70"F. A "crticallinc is drawn from poin l 2 10 the
c urve representing wind mo vement e qual 10 40 milday. 11lis is point 3.
Point 4 is fou nd at th e inte rsection of II horizontal line projec ted from
poin t 3 a nd n vertical li ne projected from point L The da ily la ke eva pora-
tion is then read from the graph as
E = 0.22 in.
Evapotron$pirotion
For II water budget o\'er the whole wate rshed. one is U$oally concerned wit h
the total evapora tion from all free-wa ter surfaces. plus transpiration, the l os~
of vapor th rough sma ll ope nings (stomat a) in plant tiss ue (e.g., lenves). For
most plmu s. transpira tion occurs only du ring daylight hoors du ring photo-
synthesis, wh ich ca n Icad to di urna l varia tions in the shallow ground wate r
ta ble in he avily vege tat ed arcas. The combi ned evaporation and transpira-
tio n loss is called evapotranspiration (ET) lind is a max imum if tile wat er
~u pp ly 10 bolh the pla nt lInd soil surface is unli mited. The max im um possible
loss is limi ted by mcteorologieal condi tions and is cnlled potential ET
(Thornthwaite. 1(48): potential ET is apprO)(i matcly eq ual to the evapora-
tion from II largc. free-wate r su rface such as a lak~ . Th us. methOlls discussed
in thc prcvious section for cvaporation ca n be used to predict pote nt ial ET.
Large surface area and high temperat ure of leaves ca n easily create
transpiration rates that eq ual o r e"en exceed potential ET. Exceeda ncc is
possible in ins tances of an ··oasis" of well-watered vegetation. such as a
crop or gu lf !;ourse. loca ted in a larger area in which actual ET is l es~ tha n
the potenti al. Ac tu al transpi ration an d. hence. actual ET arc usuall y li m-
ited by moisture supply to the plants. Ob\'ioIl5Iy the effect of li mited mois-
ture wi ll de pe nd on plant characteristics such as root depth and the abi lit y
of the $oil to tran sport water to th e roots. These effects have led to various
e mpirical factors that may be applied to potential ET. computed usi ng one
of the enpo ration methods discussed previously. For exa mple. Penman
(1950) dete rmined that ET from a vcgelilled land surface in Great Britain
was 60% to 80% of pote ntial ET computed using his met hod: Shih e t al.
( 1983) used n va lue of 70% for crops in southe rn Florida. On th e o ther
hand, Priestly and Taylo r ( 1972) multiplied Penma n equa tion c\'aporation
by 1.26 for well-watered crops 10 compu te ET. Ot her em pi rica l form ulas
120 Chapler 2
thll l incorpo rate radiat ion, ai r Ic mpcralUre. an d precipi ta tion are gi ven by
Thornth wlli tc ( 1948). Tu rt (1954). an d liamon ( 1961). and in agri cultu ral
practice the Blaney-Criddle meth od is widely used (Blaney and Crid dle,
1950: C riddle. 1958).
O nce again, pan evaporation can be used 10 eslimale ET if coefficie nts
are kn own for th e speci fi c vege tation. T here is II wide range of pan coe ffi -
cien t va lues, a s ind ic ate d in T nble 2- 2. Cautio n sho uld ~ used with nny pa n
coefficie n t mel hod for p re dict ion o f ET, si nce (he method is wea k to begin
with ( B rutsae rl , 1982) and coeffi ci e nts will e xperie nce a s trong seasona l
variat ion thu l follo ws vegetat ive growth [J3I1e rn s.
The comp lex II.lln ua l rela tio nship am ong potent ial ET , ra in fa ll, li nd soi l
moisture is illustra ted in Fig ure 2-15 fo r p la n ts with th ree d iffere n t rOOI
de pths. T he (ield OIpa city o f the so il is the mois tu re con tefllllbove wh ich
wal er will be d rai llcd b}' gravi ty. Th e willing poinl o f Ihe so il is the moist u!"e
content be low which pla nts can no t eXlrac! further water. As Ihe soil mois-
tu re conte nt is red uced below field capacity. act ua l ET becomes len th a n
po te ntia l ET. If th e soil wat e r con te nt is rcdu ced below the wilt ing poi nt, the
plan t mil y die. T hi s complica ted proeess dep e nds o n the plan t, soil type .
me teorOlogy. a nd seaSOn. A fu ll d iscussio n is beyond the scope o f this \e )[l;
fu rther dc tuils may be fo und in Brutsaen ( 1982). H illel (1 982). Salisb ury an d
Ross ( 1969). ASC E ( 2005), a nd Dingma n (2002 ).
F'ogure 2-15
Idealized onr.o.ool
~il moIslufe cycle
'i.I'ET >'i.R
for three vegetotion
types.
< PET
PET !'ET
Winler
I
,6
< /'It T AUlumn Spring PET
< 1'0 '
< PET I'ET < PET
:f.1'ET> 'i.R'L.::-'------'
Sum ..... '
W FiclJ capacil)' PET . t'olcnlial ."gpo lr~n,pita lion
for re ference crops of .llfalfa and wcU-clipp..:d f~scue grass (ASCE. 2005).
ASCE (2(X)5) recommends a standardized fonn of such an equation as the
basis fOf' ET estima tes from man y different vegetative surfaocs. 'f1le methods
work by com pu ting a reference ET es timate based on loeltl meteorology. and
then mult iplyi ng by crop coefficients th at incorporate crop-spe<:ific stomatal
resistance and surface roughness to convert the refe rence estima tes to ET
estimates for each crop. The crop coefficients vary by location. season. and
growt h sta ge of the crop. Rekrcnce ET is updated dai ly on the basis of local
meteorology.
An exa mple of dissemination of these very use ful data (e.g.. for irriga.
tion estima tes) is the: AgriMet seT"ice of the: Bureau of Reclama tion in the
U.S. Pacific No rth west and no rthe rn California and Nevada (hllp:/lwww.
u§br.govlpn/llgrilllt:t/gcne ral.html ). Daily rdere nce ET estima tes an:: com-
pu ted us ing the Ki mberly-Pemnan eq uation for II refere nce crop of a lfalfa
fo r more tha n 70 weather stations in the region. Crop coefficie nts li re site
specific and availa ble for almost all crops gro ....'Il in II locality. Real·time lind
'22 Chapter 2
historic weat her dal a may also be ex tracted {rom the weat her sta tions, mak -
ing AgriMct a val uable da ta resource for more than just ET estimates.
Application of rd erellce ET toward ET e~ti m atcs for "mlura l vegetation
(e.g .. forests. grasslands) depends nOi only on obvious location and vege ta-
tion-specifIC parameters but also on (lVailable moisture conte nt. However.
such esti ma tes can be att emp ted with additional information. 8S explained
by ASCE (1996). The adva ntage of applyi ng rd eTence ET estimates [0 nat-
lI ral watersheds is Lhe use of real -li me, si te-specific weat her dala for the
refe rence ET comput ation, wh ich arc much morc available than lire pan
measurements.
Some rai nfnll on vegetatio n is tra pped on leaves and ol her vege tative sur-
faces and subseq uently lost to the aHn05phere through evaporati on, The
trapped wnter, or inlerception, does nOI reach the ground surface and is a
loss of wate r Ihat might otherwise potenti311y run off from the catchment.
Rainfalli hat drops thl'Ough the vegetation is known as through.falL Intercep-
tion is an examp le of depression storage or ini tial abstraction I.h al is sub-
tracted from rain fall d uring the ini tial stage of a storm (Cha pte r.;; 2 and 6).
Because the ul timate fate of such "".Iter is evaporati on, inte rception is some-
times lumped wi th ove rall basin ET. However, estimates arc ava ilable as a
function of plant type. loclll ion. and season. acarly. leafy deciduous trees
may intercept conside rabl e water in the summ er an d aim05 t non e in the
wi nter. O n an annua l basis, Dingman (2002) repo rts annual interce ption as
a pe rcen t of annual preci pi tation on th e order of 24%-35% for Pacific
Northwest conifers and on the order of 10% for North Ca rolina hardwoods.
O n a storm eve nt basis, Shuttleworth (1993) repo rts canopy int erce ption on
the order of 1.2 mm for conife rous fo rCl!ts and 0.8 mm for broodleaf canopies
in full leaf. Interception capaci ty is recove red during in{ereve nt dry periods
as a funct ion of evaporation ra tes.
2.7 The process of infilt ra tion has bee n widely stud ied and represe nts an impor-
HYDROLOGIC tant mechanism for the movement of wat er int o the soil under grav ity and
lOSS- capilla rity forces. Infihration ,·olume is subtracted from a precipitation c"elll
INFILTRATION in order to determine the ne t volu me of ra infall, or rainfall excess, which is
equivalent 10 {he direct runoff from a w(ltershed area. Horton (1933) showed
tha t when Ihe rai nfall rat e i exceeds the infiltration rate f, water infi ltra tes
the su rface soils a! a rate tha t generall y decreases ..... ith lime. These rates are
usually re ported in inches per hour. For any gi ve n soil, II li mit ing curve
defines the max imUIl1 possi ble rates of infilt rat ion \'5, time. The ra te of infil -
tra tion depe nds in a complex wayan rai nfall intensity, soil type, surface
condition. and vegetal co,'e r.
123
' ., flin'/M
,,"'
1.43
," '"
, '"
0"'
0 .36
The Yo lumc (in inches o\'er Ihe watershed) Co1n be found by plotting the
curve given by the table of va lues and Ihen finding the area under the curve
bounded by I = 0 and I = 6 hr. The plot i~ shown in Figure E2-1O. The
curve is given by Ihe ('qualion below, and 10 find F, the ,·olume of infiltra-
tion and the area under the F curve. we mus t integrate Horton·s equa tion.
!';gu... E2-10
HOffOn infillrorion
wr-----------------------------,
curve. The Oleo
shod...:! undel Ihe
CUI"! Jepre~nts the
totol W)lume 01
inlilbation 0\'8/ (]
{).hI period.
Hydrolog ic Ano~is 125
Horton'S equation in the form of Equ ation (2-42) suffers from Ih~
fa ct Ihal infilt ration capacity d~crcas.es as a function of time: rcgard lcss
of Ihe aC lual amounl of water available for infiltration. Thai is. th e equa-
tion f1SS Um I;;S ponding on the surface lind II reduction of infiltration capac-
ity. regardless of whether or nul th e rainfa ll in tensity i cxce~ds th e
co mpu te d va lue of infiltTfltion capacity 10' Fur example. it is C(J mmon th at
t he infiltration capacity of sa nd y soils greatl y exceeds most rainfall inten·
sities. with values of 10 up to 23 in J hr (Table 2-3). Even inte nse rai nfall
pulses 8re seldom this high. wilh the co nseq uence tha t all rainfall may
in(i lmlte: that is. I;. i. T he infiltration cap~l city should be reduced in
proportion 10 the cum ulative infiltration vo lume. not in proportion to
time .
The cumula ti v~ infilt ration \'Olume is giv~n by the in tegral of Equa-
lion (2-42).
Horton para mete rs for nat ural soils (e. g., Table 2-3) arc ge nerally nol
appl icable to dis turbed soils fou nd everywhere in and nCllr constructed
(de\'cloped) areas. Pin el al. (1999, 2001) provide some da ta and guidance
o n infiltration parameters as a functio n of commonly measured soil compac-
tion paraffi cters. T his also ill ustrates th e importa nce of local infillromc\c r
measure me nts wherever possible.
O lher equations have been ueveloped uliJiTjng analyt ical ~o[utions to
the unsat urated flo w eq ulllio n [rom soil physics (see Section 2.8). Ph ili p
(1957), in a dassil.: sc t of pape rs , de \'cloped two equations of the fo rm
figure 2-17
Grophical rep'. " " Groos precipimiO)n ra le (;)
sentation of the phi
iOOcx method.
VQlu me o f
ruooff
V<)\u_uf
infiltnolion
Hydrologic Anolysis 127
n_ "'""'I
••
<>-2
f,n.{hrj
I.,
2-5 2.3
5-7 1.1
7_10 0.7
10-12 0.3
fhe (irst slep involves graphing the &i'·en data. liS in Figure E2- 1O. To SOlllOON
approach the problem, \\>e must first change Ihe area of Ihc watershed
into acres:
area(ae) = 0.875 sq mi (640acres/sq mil.
area" 560 acres.
We develop an equation for rainfall minus infiltration for each of the
time interval s 3S given:
2(1.4 - <b) + 3(2.3 <b) + 2(1.1 <b) + 3(0.7 - 4» + 2(0.3 - 4» "" 4.9
NOle that if <b is grea ter than the net rainfall for II ~pecilic lime period.
no negative rainfall is added inlo the runoff calculation.
The rate of infiltration can he found only by trial and error:
Assume <b = 1.S in.lhr. The runoff is [he volume of water abo'·e the
line al which y = 1.5 on Ihe j!raph in Figure 0-11. This 4> index would
then account for 3(2.3 - \.5) = 2.4 in. of runoff (neglecting negative
components). which is less Ihnn 4.9 in. Try again.
,
I I I .u Figure E2-11
Graph 01 ra infall
,
.
. .
intensify V$. time .
" >
1
44_4 ______________
" I
" -t--or
0 , • •
" "
128
More ad,'nnccd infilt rat ion methods based o n actua l soil types an d
solving the governing equa tions of now in porous media (i.e., Richa rd's
equation) are presen ted \\~th examples in Ihe next secti on. G reen and Ampt
(19 11) assu med II sharp welting front. separa ting initial moisture content
from saturatcd moisture con tent. which has pene trated to a dept h L in time
I since infil tra tion began. Water is ponded to a small depth on the surface.
and the Grecil and Ampt model can be used to predict cu mulat ive infilt ra-
tion as a function of time and soil type. The met hod requires e s t im ~t es of
hyd rau lic cond uctivity, po rosity. and wClling front soil suctio n head. This
mdhod has receivcd much attenl ion, since il is ba~d on readily measured
parameters for various soil classes.
2.' The aootraction tenn in the water-bal:mce equation usually includest hc loss of
GR EEN water thro ugh the interception of precipitation on aboveground vegetation.
ANO AMYf depression storage on the ground surface. and infi lt ra tion into und erlying soils.
INFILTRATION As previously disc ussed, infil tra tion Clln bel determined u si n g~imple empirical
MEIHOO methods or theoretically. The infillraliOI1 methods that rely on physical rat her
tha n empirical parameters providcsol ut ions to the govemingequalionsof flow
in porous media (i.e" Darcy's equation prese nt ed in C hapler 8) forspccific soil
types and gr.! in sucdistributioos. This SCC'Iion is a more adV'.!nced trcatment of
infilt rat ion. and it ca n be skipped without a loss of OOfItinuity in the lext.
inmtration is model ed as unsaturnted flow and differs fro m saturated
(aq uifer) flow because bot h a liqu id (watcr) :lIld a gas (air) phnse art: presen t
in the unsut urlltcu syste m. Gravity and frielion art< thc only forces at work
for sat ura ted fl ow; however. a suction ror~e as discus.~d in the next para-
grapb must be incluucd fur unsat urllled flow. Richard 's equat ion for unsat-
urated flow in porous media applies the equlltion of cOnlinuity to Darcy's
law in ordi: r to account for the presena: of a gas phase in the subsurface.
Richard 's eq untion takes the form:
I•
!
•
MoOS1U rt: C(>nlCnl0(% ) Hydrouioo cunduct"',ly k(cm/s)
rrr- .--.-----,
..0 "0,.-- - ----,
"" -1 """"
.I OJ
!
•< '"
'a,
~
'"
20
Figu,.. 2-18
Typicol reiationshipi In the unsaturated zone.
Hydrologic Anolysis 131
•
•
"• "• Fig....-e 2- 19
Moisture ond infiltra·
'~ 1 tion relationships .
V (a) Moi~ture prolile
.., ,1--",.----:.
"-
.~
WCl1iogl
fron! :
,
,
, 01 moment of surface
~turotion. (b) Mois·
ture profile 01 a later
,
=
< ,,
•
,•
AWl
./:
time. (c) Infiltrotion
beho~ior under
.. ,
• ,,
,••
different rO infoll .
(Aoopled from Moin
and lo,,,,,,.
1973)
,., ,'<
\,
I~ ~
__ __ ~,__'_'_O_~_~_O_O_o_o_=_=__
o f---_--'A'---_ _---1
,<)
Green nnd Am pt n:pfcscnts the typicnl relat ionshi p between moist ure
content and soil depth (Fig. 2- 19(a) as a shnrp wC11ing fron t scparating
sa turated moisture content at the top of the soil col umn from the initial
moisture conk" t at the column base. The Green and A mp! mode l is lIOW
widely accepted. since it C(tn be used to predict cum ulative infiltration as a
funct ion of lime and readily availa ble soil parameters. For e~ample. it "';IS
used as the basis for t he analysis of Houston ,lrea watersheds a fter T. S.
Allison in 2001 (TSARP. 2001).
The Grecn and Ampt modcl operates un de r five princip~1 assum ptions
(Ferguson. 1994):
I. The soil under consideration is ho mogeneous and stab le. implying
that macropores and prefe rential migr;;tion pathways should not be
considered.
2. The su pply of ponded water althe surface is not limited (an assump-
tion su bject to modification below).
132 COOpter 2
saluration I,. when Equat ion (2-52) beoo me5 valid. Th us, sellin g i "" f a nd
50h'ing for F, results in the following Green and Ampi equa tion,
(2-53)
We require i > K, in Equation (2-53) ~ nd remember Ih al capilla ry suctio n 1/1
is nega ti ve. The Green-Am pi infi ltration method will pred icl lhe following
results for various intensities of rainfall i:
I. If i :S K then f '" i (curve A in Fig. 2-19(c».
"
2. Ifi > K,.thc n f= ;unlil F = if, = F. _
3. Following surface saluTatio n.
Tabl. 2..... G'e<!n-Amp l ln li lwalion Paramor.,.. 10, VoriOOJ. Soil Tex!!... CIo....
Eff.oo.tMo Poro.iIy Welling Fronl Sudion HydrQUlic Conduc~viIy
Sail CIa..
So",
Poro.iry '\
'.
0.",17
HecxlI/1 I'eml K {tm/h.)
11.78
0.."
037"'-0.500
0 ... 37
035......0 ... 80
0401
'"
0.97-2530
,""",w", 6. 13 2.99
o 363-0..xl6 0 .329-0.473 1.35-279.
Sandy loam 0 . 53 0 . 12 11.01 1.09
0.351-0.555 0 .283-0.541 2.07-4.5 47
Loom 0 ."63 0.43 .. 8.89 0"
0.375-0551 0.33"'-0.53 .. 1.33-59.38
Sih loam 0.501 0.486 10.68 0.65
0.420-0,582 0.394-0.578 2.92-95.39
Sandy day loom 0.398 0.330 21 .85 0.1 5
0.332-0..464 0.235-0.425 4.42- 108.0
Cloy loam 0.464 0309 2088 0.10
0.0409-0.519 0 ,279-0.50 1 4.79_91 \0
Solly clo~ loam 0 ... 71 0.432 27.30 0.10
0 .418-0.52. 0347-0.517 567_131.50
Sanely cloy 0 .430 0321 23.90 0 06
Silly day
0370-0490
0 .",79
0207-0 .. 35
0 423 ,.,,,
".08-IAO.2
0.05
0 .• 25-0.533 0.33"'-0,5 12 6. 13_139."
C., 0475 0.385 31.63 0.03
0 .• 27-0523 0.269-0. 50 1 6.39- 156.5
s.....:., Rcwb Itt 01.• 1983
136 Chaptel 2
estimate for Horton's fin al infiltration capacity. Ie[Eq. (2-42)]. Ta ble 2-4
provid es valua ble estima tes for G reen-Am p! para mete rs, but local dal a
(e.g., Carlisle CI al. , 1981) a re preferable if available. MiS5i ng is th e ini t ia l
mo istu re conten!, since it depends on antecedent ra infall and mois ture
conditions. Ideall y, Md = n - 8; shou ld be mC iisurcd . but t his is rare ly
accomplis hed. A lower bound for maximum Md may be inferred from th e
" avai lable wate r ,apaeily, in.llo." fou nd on NRCS Soil Properties and
In terp re tation sheets. Available wate r ca pacily on these sheets is the di f·
ference between fic ill capacity a nd wilting poi nt, with typi~ a l values usu-
al ly range fro m 0.05 \00.30 in.!in. These are lower bounds on the max imum
M J , sincc they do nOI include wat er stored a bove field capacity. T he value
of Mil to use for a partic ular soil in ques tion is be$t de termined from a soil
te5t. Othe rwise:, <I co nservat ive (low) MJ value could be used for design
purposes (e.g., 0.10). Md "alues must be upda te d in contin uous simulation
hydro logic models tha t account for soil llloisture con te nt.
In areas of high water tables, there is a li mit to th e soil storage ca pac-
ity. an d infiltrat ion can not continue inde fi ni tel)' without comp le te satura -
tion of th e soil. In suc h cases, infilt ra ti on ceases. losses (ra infall
abstrac ti ons) become lero, an d rai nfall exceu intens ily eq uals rainfall
intensity. If site-speci fic informa tion i~ avail abl c. this capacity, S, CHn be
es ti mnted from soil moist ure dahl an d de pth 10 wllter table, L, as implied
in Figu.·e 2- 19b,
S = L(9. - 9;). (2-54)
where L is now the depth to the water ta ble. I n some localiti es. regional
informnl ion on available soil storage has bee n prepared (e.g.. South Florida
Water Management District, 1987). T he above value of S is esse ntially the
sa me a5 the storage ya lue in tht: SCS mt:lhods dcscribc:d ea rlier.
Th e examples below demonstrate the use of Green and Ampt equa-
tions with act ual soils data. An Excd spreads heet is avai lab le on the text-
book website for use in solving fo r infiltmtion f as a function of time.
Noting that Md = 9, - 0;, wc can solve Equation (2- 53) 10 obtain the SOLLJTION
volume of water that will infiltrate before surface salUration is reachcd:
.M,
p. = 7C-"';~~
(I ilKs)
-(9.37 in.)(0.518 0.318)
[(7.M in ..11r)/(1.97 in ..11r)]
P, = 0.625 in.
Until 0.625 in. has infiltrated. the rate of infiltration is equal to the rain-
fal! rate. After that point (surface saluTation). the rate of infiltration is
given by the equation [eq. (2-52)]
f = K.(1 - MJ'II/F).
Solving this equation for various values of F givc~ the graph shown in
Figure E2-12, whe re f decreases as F increases.
Figure E2-12
"• Infiltrotion rate ys .
1~
infiltration volume.
~
<
0
~
~
,,
,
0
0 , 3 , , • ,
F(in.) "
GREEN-AMPT TIME TO SURfACE SATURATION EXAMPLf 2-13
Silt loam has the following soil properties for use in the Green-Amp!
equation:
K. = 1.81 X 10 4 em/sec,
8, = 0.523,
I/! = 17.0em water.
For an initial moisture contenl of e, = 0.3. what is the ponding depth (f~)
and pondi ng time (I.) based on this rainfall intensity? What will be the
infiltration depth 20 minutes after the beginning of the rainfall"! (Assume
that the rainfall has a constanl intensity.)
i = 6Ks for 10 minutes.
138 Chapler 2
SOlUTlON The ini l ial moisture deficit M" ,., 0.523 - 0.300 '"' 0.223. In o rder 10
Ohlain the ponding time. "e comp ut e the ponding depth req uired 10
produce ~atu r3 li o n from Equatio n (2- 5.3):
F." (- 17.0cm)(O.223) / ( 1 - 6K./Ksl = O.76 cm.
Th e paneling l ime (r,) i s
2.9 A small percentage (about 13%) of precipitation over the en tire U.S_ falls as
SNOWFAU AND snow, but it can be n dom inant source of s\rcamnow. cspcdally in moun tain-
SNOWMELT ous areas during spring melt. This is cc rt ai nly true of the western United
Stotes. bu t significant perccntagcs of runofr in th e no rt heast and mi dwest
also originale as snow. Because the storagc and mclting of snow play an
importfln! role in the hydrologic cycle of some areas. hydrologists mLL~t be
able to re lia bl y predict the co ntribut ion of snowmel t 10 overall ru noff.
BeCfl use snowm el t begins ill Ihe spring and Ihe derived runoff is OUI of phase
with periods of grea tes t wate r demand. control schemes slIch as storage res-
ervoi rs ha ve been implemented in many areas for wate r supply. Under cer-
tain conditions. ~nowmel l can alsoconll;bute to nDOding problems. especially
in larger river basins with moun tai nous areas. In ad di lion to forecasting
nood si tuat ions. snowmelt predictions art also valuable to powcr com panics
that generate hydroelectricity and 10 irrigation districts_
Beca use of the range of uses for estimat ing th e con tri bution of snow-
melt to stream now and the variation of cond itions applicable to each case.
many met hods have been deve loped for com puting sno wmelt as it affects
Hydrologic: AnQly.i~
".
slreamnow. These methods ha\'c consistently found that the timing and
amoun t of runoff dcpend main ly on threc factors: the energy of a snowpack
available for mch. the areal extcn t of the mehing snowcove r. and the effects
of storage on the movement o f the meltwater (Male and Gmy. 1981).
Snow meas urements are obtained using standard and reco rding rain
gages. snow stakes. and snow boards. which measure accu mulation ove r a
period of li me. Snow surveys al a numbe r o f spaced poin ts lllong a snow
course arc nceemry 10 account for dri fti ng and blowing snow. The Nat ural
Resources Conservation Service (NRCS) uses devices call ed snow pillows.
which continuously record and transmit lhe pressure from snow at the bottom
of a snowpack. The NRCS h1l$ about 500 telemetry sites (~SNOTEL ~ sites)
througho ut th c western Unit ed StatestTllllsrnilling data from snow pi llows.
The distribution of mell n annual snowfa ll in the Un ited States is qu ite
variable. and errors may be large in mou ntai nous region s becuusc of the lack
of measuremen ts in remote areas. The re is a gradual increase in snowfall
with latitude and elevation, and from 400 to 500 in.fyr (- lOOO cmfyr) can
occur in the Sierra Nevada and Ca'!C..1de Range or the wC$tcrn United States.
Large snoW!; can also occur nea r the Great Lakes. wi lh average va lues O\'cr
100 in. (254 cm),
Ptr)'$iu of Snowmelt
The ene rgy exchanged betwee n the snowpaek. t he atmosphere. and the
earth's ~ u rrace is Ihe con trolli ng factor in rates of snowme lt. Other geo-
graphic.topoograp hie. and surface cover faclon; are also importa nt. Density
of snow is a major asp..-'Ct of predicting ru noff and is dete rmined by stweral
factors. Snow th aI is fo rmed unde r dry. cold condit ions will be less dense
than snow formed under we t and warmer con dition~. As snow HL'Cum LilatC$
on the gro un d over Il period or lime. its density increases wilh th e se tt ling
and oomP<'c tion due 10 factors such as wi nd. Win d increases th e dens it y of
snow in two ways: First. th e lurbulc nc.: break s icc crystals into smaller ones.
ma ki ng them se ttl e more compact ly. and second. dra~ fo rces an:: exerted on
the surface of snow due to the movi ng air an d particles. Newly [fl lle n snow
has a density abou t 10% that of liq uid water. but the density ca n increase to
50% wilh aging. The depth o r water that wou ld result ifall the snow melted
is called the snow wlter equivllent. SWt:. lind is gi\'e n by the equat ion
be found from eq uations for heat storage change (!J.H) (see Dingma n. 2002.
for examples), T he equations for each phase of snowpack development
(warming, ri pening, and melti ng) arc similar to IhOSll o f simple mel ting ice.
Fir.>t the snow warms from below freezing to O"C before it can begin melting.
<lccord ing to l he equation
H! = -c,·Pw·SWE-( T, - Tm} (2- 57)
where c, is the heal capaci ty o f ice (502.1 caVkg- K). p", is the densi tyofw8ler,
SWE is the water e q uiva le n t de pth o f the snow, T. is the te mpe rat ure of Ihe
SIIOW, and T ttl is the mel ting-point temperature (<fC). This brings the snow-
pack to the point wh ere it can begill phase change; snow turns to ret ained
water within while maintain ing a temperature of O"c. The pore spaces nil
unti] saturati on, and the energy inpul re quired is
(2- 58)
where Ottt is the maxim um vo lumetric wat eroon tent that the snow (;an retai n.
hs is the depth of the snow. and L is the la ten t heat of fusion (79.8 KJlkg).
Finally. ..... ater begins C1(iting the snowpack as th c phase change cuntin-
ues. The heat requi red for the snow to melt is
H] = (S WE - h .TCI)·P...· L (2-59)
whe re h ..w is the depth of the retained water th at is alrcud y melted (Ding-
man, 2002).
Add ing 1/ t. fi 2 , and H3 resulis in the change in heat Storage necessa ry
to achieve me lting. The approximation of 00 Iy (ca Vl;m ~) to melt I cm depth
of snow (from above) lakes into acco unt such equat ions. Assum ing the snow
is already ripe. the amoun t of mel t from II positive va lue of ~N (;3n be (ound
wit h a shortcu t
M = dHj 80 cmldny (2-60)
0'
M = dH j203.2 ill.lday.
Energy budget methods represent the best lcchniques for predicting snow-
mel t when all oomponcll ts can be estimated.
The Na tional Weather Servi ce (A ndc rron. 1973. 1976) has de\'eloped
an energy budge t model for runoff fo recasting. U nfort unatel y. so me e nergy
budget compone nts arc usua lly mi SSing. leading 10 the use o r simpler but
less accurate " degree-day methods ~ based solely on temper atu re da ta.
where
M = d:lily mel t as a deplh of wa ter eljuivaient (inJ day; mm/day ).
Ti = index air le mperature eF; 0c).
To = l);Ise melt te mpcralun;: (OF: · C).
DJ = degree-day me lt faClor. (in.ldil y.o F: mm/day·"C).
Empirical deg ree·day faciors range fro m 0.05 to 0.15 inJday-OF and can be
determined by several eq Ual ions. one of the sim plest being
DJ = 0.0111',' (2-62)
where Df is in un its of mm/day·· C above O"C and p, is the snow densi ty in kg!
m3. "Ib is factor is a ff~te d by location. time of year. and meteorological condi-
tions. and more complex equations IIlay lake in toacoounl vegetation transmis-
sion for radialion. solar radiation. and snow albedo (Ma le and G ray, 1981).
Equa tion (2-6 1) is bas ically a regression equa tion: ACOE (1956)
results are shown in Figure 2- 20. Either the max imum or ave rage dai ly tem-
pcralllre for Ihe ponion over O~C is usulllI)' used fo r Ihe indu ai r tempera·
ture T;. and C1'C is most commo nly used for Tn (Gray and Prowse. 1993).
However. bllsc mel t tempera tures ca n be below freezing. si ncc Equalion
(2-62) represents II li neari za tion abou t reference conditions well above
freezi ng. "nlus. the e q u~lions shown in Figure 2- 20 shou ld nOI be u$Cd for
T, < 34°F. The te mperature·in dex approach is most u~cfu l and acc urate
under nonnal cOll ditio ns and less so for extreme conditio ns. The ACOE
develo ped a numbe r of e(I U~lions ror gene ralized areas.
Open .. te
.II - Q06(T-2~J
Fu"'~"te
M O.05(T- 32)
0.'
The genera l equat ion used by the ACOE to calcul ate the smOlin! of
sno",mc lt accounts for the total amo unt of energy 8\'aila ble for melting the
sno"'p aek lit a point. The equation is given as:
M - Om334.9pw8 (2--63)
where
M •
snowmelt. mm per day of waler equi va lent
-
Q.. • algebraic sum of all heat components, kJ/m2/day
8 the rma l quality of the snow (e.g.. rBl ia o f he at re q uired 10
heal a unit weigh t of the snow 10 that of iec al O"C)
334.9 • latent heat of fusion of ice. kJ/kg
I'• • dens ity of wale r , kglm 3
The ge neral eq uat ion for snowm elt (2-63) call be combined wilh long-
wave radiat ion snow melt equntions described by ACOE. resulting in a non-
linear relaTionship between temperat ure and long-wave radia tion. These
relationships were simplified by filting linear approximat ions and shiftcd 10
Ibe Fahrenheit tcmpe rature scale to derive th e following long-wave rad ia-
tion me[t cq uat ion$_
a. For melt unde r d ea r skies:
M [=O.02 [2(T. - 32) - O.S4 (2-64)
h. For me[1 under a forest canopy or low cloud cove r:
M[-O.029(T,,- 32) (2-65)
where
M [ - long-wave radiat ion me h (inches/day)
T. _ ai r temperature. F
These arc general ized equll lio ns uscd by th e ACOE. but other factors
such as rai n, wind, and pe rce nt o f fores t cove r also affect t he ra te of snow-
melt. For [url he r informatio n regarding factors affecting snowmclt and
calcu[ating snowmelt rale. please rder to lite U.S. A rmy Corps of Engineers
Engineering and Design Manual for RU l'IOfffrom Snowmelt (1998 ).
United Slaies Anny Corps of Engineers. 1998, Engineeringand Desig.n
Man ual, R/UlOff From Snowmdt, Wash ington. DC.
--
00
-- •
• .~
••
, March Al'fil l ?7~
0J~-!,--,!--!,'-C,O'-"'J"-,",,-~,',-f,'9,-",",-,,,'-C,,!.c",,"C"!c-'J',_",oe",
1i_{d~I.J
Figure 2-21
Comporir.on of $flOW estimation modell. (from Andenon, 1976.1
when'. the sno ..... depth. losses. and snowmelt arc ass umed uni fonn in ellch.
In llIost praclic111 cases, snow melt is estima ted by the index meth ods pre-
sented in the prev ious section. The budget method itself is used in design
flood studies. Once inflow water has bee n calcul ated. storage routing tc,h·
niques are used to accommoo ate surface and gro und wa ter components.
each with a di ffe ren t storage time li nd number of roul ing un its. Storage roUI -
ing methods arc trea ted in det ai l in Chapler 4.
SUMMARY Chapler 2 presents the Ulnrxpl ofrail1r. n..rull()ff anal)l$is. or the OOfWcrsion of pre-
cipitation to runoff or strcamnow, a central problem of ensinccnng h~·drology. GrO!>S
rainf3 i! must b;: adjusted for losses 10 infiltration. e\'aporat ion, and depressioo 5tOI-
Age 10 obtain rai nf~1l c.ocss. wh ich equa ls direct runoff (ORO ).
The UH is defi ned as 1 in. ( 1 In.n) of direct ronnfr ge nerate d uniformly o~er 11
bIIsin for a specified period of rainfall nnd is a powerfu l method !llil1 used for m~ ny
wat"rshetl studies.. The concepl of the UH allo>o.'S for the conversion of r~infall eJI-
cess into a baSIn hydrograph, through a linear addmg an d lagging proced ure ailed
hydrogl1lph corovolution.
UHs can be derived from actua l storm data for gauged basi ns. Synthetic UHs
rcpresentthcorel kHI or empirical appro~chC$ that can be: Applied to ungaged ba-
sins, These are mOlit ofte n used for the ~nH I)'Si~ ofwater$l\c:ds where limited gBuging
might e~ist MO$t synthel ic UH formulas relate parameters such itS tune to peak or
la& time to measures of channel le ngth. slope, wlitersit«l siu, and wateMW shape,
A second reillt ion usually corre la tes peak flow to basin area and lag time. UHs are
grea tly affected by urban developme nt and channel iza tion. and UH tech niques have
bee n customi~ed fur loca l or regional condi tiuns using statisticul methods (Chapter3),
Sc.-eral of t he more popula r UH met hods arc presented with examl,l" s, incl udi ng
Snyder. SCS. and Clark (TC & R). all COI'1\ained within the HEC_ HMS oornputcr
model (Cha pter 5),
Ki nematic " 'a'oc methods developed in the 1970s fot overl and flow hydro-
graphs and channel hydrog", phs we re based o n 5Ol utlons of the continui ty and Man-
ning's equ ation for a gi\'e n geometry. Kine matic ",'a"" methods are now more useful
with the ,urre nt popu larity of nu mer ical methods and advanCl;d hydrologic models,
"'hich in d ud.:: HEC- HMS from tile U,S, ACO E Hyd rologic Engineerinll Ce ntcr
(2006. 2010) and distributed hydrologio; models (VieuJ[, 2'102),
Hydrologic 1OS5CS arc covered in detail in O!~ pter 2, including eVapora tlon
and ET for wat er bodi"" and ,ocgetated surfoces, The Penman equation is presented
along with a detailed ,,~~ mple of,15 use . Infiltra tion I~ is p",sented using both simple
empirical methods ~ U\: h a~ Hort on Mild more pbysics·ba.wd methods such as Green
and Amp!. Several c~Hl11ples and homework proble ms ale included to cover thell<'
methods in delail.
Chapter 2 also presents snowfall and 500"" melltheory 15;t relates to the rain.
fall-runotf proccu. The physics of ~1l()\\'me1t is briefly covered . and energy bodget and
e mpi rical degree.Jay e,!uations arc shown that can be used to estimate melt rat"" in
Hydrologic Analysis 147
inJday or em/day a5 a flJ ~tJon of air tcmp:ra lure, snow lemp: rature, and ather f(\(,
lors, Snowmelt can b.:: a signincant factor in flood flows and water supply est imates.
especiaUy OUt WCSt in the Rockies and in the SIO:n'll Nevada TlInge in Californ ia.
, r-
,.
l1
,
•• , I-
0
I , , , , ., , ,
o • "lime (Ii,)
Ts = 12h •.
TII - 4hr.
Q, = 200ds.
".-Iiere
Til - lime base of the U ~I ,
TR - time of ri!k!.
Qr · peak now,
de"clop II storm hydrogrnph [or Iloorly rainfaLl (in.) of I' = 10.1,0.5. 1.21.
(b) Rt:pca llh e I\oove problem for liourly rainfall (in.) of p,. [0.2. 1.0. 2.4J.
2.10. A ~k clch of the Buffalo Creek Wal en;hc(J is shown ill Figure P2- IO. Areal! A
Qnd () are identICal in ~i7.c. Sh3pe. slope, and channel length. UHs ( I liT) are
provi ded for nat ural and fully de,doped coodiuons roo bolh areas.
figure P1- 10
SkefCh of Buffalo
CteeIc walef~ed.
r ...... (hr)
UI-Y..ld.j
0
• .,, '"
, 2
290
3
• , • ,
". '", , • 90 ,. 8
.. "" ..
T1 .... (II.)
UH In(" )
•
• " " "
2 3
...
4
". 2"• • •
'" '" 80 53 "" ", •
<II) Assummg natural roodi liuns [or boll1 arc..... evaluate the peak uutnow /II
point 1 ;( 2..5 inJhr of ram ra lls for 2 hr. Assume a totlll inliltratiol1 l0S50f] in.
Hydrologic Ano~i' 149
t h ) AS$UlllC th m area B h ~s re ached full deve lo pmenl and area A has re-
m~incd in natunrl condil ions. DC lcnniroc the outn ... w h)-dmgrap/l&1 point
1 if a nel r.llnfall of 2 in A" fHlls for 1 hr
(c} Sketch (he uut flow hyd rognrph for the Buff:,1o Creek Watershed under com-
rlete dewlOl1rT1"nt (A and B Mth uroon i~etI) for the rainfall given in part (b).
2.11. The I-hr UH in lhe: accompa o)'mg lable ""$recorded lor a parlicular " 'ater-
'hed. Dt'tennin.. lhe ~7.c of the watershed in acres and Ih en oon\"en t he I. hr
UH inl O a 3-hr UH for the wnlcr<hed.
TI .... lh.) 0 2 , 4 , • 7
• • 10
Ufenl 0723.01475605030 " , o
2.12. A majorslorm evenl "'as reC(lrded [or Linle Bel" Creek. The rncrern" mal r~in
fall and measured hydrognrph data for Ihis stonn are pro>ilkd in T able P2- 12
in I ·h. increment s. The drainage area for the ba~n iii 3.25 nli~ . Assume thatlhe
base flow for Lrll ie Bear is h·m. (See Fig. 1- 26.)
-
Incremenlal
P...:ipilarion p...) Flow ""
Date/Ti"..
6/8/01 16:00 0 00
6/ 8/01 17:00 00' 0.'
6/8/0 1 18:00 0.0.01 0. 5
6/8/01 19:00 0 0.0
6/8/01 2000 0.08
6/ 8/ 01 2 1 00 0.05 "
J2
6/8/01 "00 0 ..012
"
.
6/8/01 23.00 0.' 78
0;00
6/9/01 0.3
'"
6/9/01
6/9/01 ,
1:00 0.1.01
0.1 9 '"
20'
6/9/0 1
6/9/01
3:00
4:00
0 ...5
0.01 '"
'"
6/9/01
6/9/ 01
' .00
600
0
'"
no
6/9/01
6/9/01
' ·00
8:00 '"
'"
6/9/0 1
6/9/01
9:00
10:00 '"
98
""
6/ 9/01 11:00
6/9/01 "00
6/9/01 13.00 27
6/9/01 14:00 0
(n) Using ~ I orm hydrograph. eslimate the "01 of runuff In inches m'er the
wHk ,.,;hcd.
150 Chapler 2
T_Ih<l 0 0.' 1.0 U 2.0 2.' 3.0 3.' '.0 ••• ' .0
I (cmI'N/
I (cm/hr/ ,,, I.,
0.75
1.25
'5 ,. I.'
2.'
'.3
,
U (mlls) 33 06 80
" " 3S 20 10 •
U, Vsmg . progra m suc h. as Exoc l. deveLop the S-c"rve for It..: "VC ll J(). mll) UB ,
ftnd Ihen dew lop Ibe IS-m in V ii from the 3O-m in V B.
Time (hrJ
U 1d>J
, "
0 0.25 0.'
70.9
0 .15
11 8 .6
1.0
109.4
1.25
81.6
...
60.'
T_""
u l.;hl
3.'
5. 8
3.15
H
'.0
33
4.25
2.<
...
1.8
4.7S
1.,
' .0
08 ,
5.lS
2.16. Usm g E xcel spreadsheet programs, de>'clop Ihe S-cu,ve for Ihe gIVen 3- hr UH .
and Ihen develop th e 2-hr UH from Ihe ) -hr UB.
o
o
2 3
•
28' •
• ", '80
, '", 210
Time (lit!
u 1d>J ' 30 00 30
"
,
10
",
1.17. O ivcnlhc folJcw.ing2-hrunit hydmgraph.eakulate the I· h, unit hydrograph. Then
back calculate and rtnd Ihc 2· hr unit hydroJITHph 10 prove Ihat the melhod of cal.
rolalioo is IlCC\lratc. Graph both unil hydrographs agamsltimeon the same plot.
Ti me !hrl
,
0 , 2 3
,sa • , •
Flow (cis)
" '" '" . 00
'"
r..... "rI , 8 , 10 12
, 13
flow (chi 350 300
'" 15' "
100
"
Hydrologic Analysis 151
2.18.. o..:vc lop 5\Orm lIydrograplls (rom UHs v( subareas 1 and 2 shown in Figure
.,,
1"2 18 for the gi •..,n rainfall and infiltratIOn.
i f
f,n./Iv) ron./hr)
I 0.5 0.'
2 0.2
,
3 "
3
0.'
0.2
0.2
"rI
Ti ....
UH \ Jeh)
a
a
I
200
2
<SO
3
.50
,
050
5
300
• a, • ,
ISO
UH2 Id.) a 150 300 500 3SO 250 125 100 50 a
Figwe 1"2-18
Wotershed with thrC(!
,, 3
suboreos and their
respective Io nd use
" ,
\ /~
- Meadovo"$
"""..
~~ " M~.<Jov,"$
'A
8
2.19. [k,~lop a oombmcd Storm hydrograph al poinl A in the watcnhed (Fig. PI- IS)
and lag route (shift in time only) Bs>um ing thlll travel lime from poim A to B is
exactly 2 hr.
l.lCI. o.",·elopa storm hydrograph (or subarea J from Ihe gi-c:n U H. <tdd 10 the rom·
binoo hyd rogrnph (rom PrOOlem 2.19. and p roduce a fin al Storm h)drogra ph
.,
~1 the o utlet of Ihe wBler.;hed. B .
5 '"
350
•
,,,
210
130
"a
' 52 Chapter 2
2.21. Redo E~ample 2-6 if tile watershed is soill~'pe 13 in good cover forest lan<J .
H ow doe~ the forested aTtn «mIpare to Ih" '1>(:11<10111 U H ?
2.22. Sketch tile SCS triangular arK! curvilinear UHs and the mass curve: fc-- a loo.mj~
WlIlenltcd ....h ich '" 60% good condition meadow a nd 40% gootI oo"er foresl land
The wa!C~he() l'()nsi~t$of70% iIOil group C and 30% soil group A. Tne 8'l::rage
Slope is 100 ftlmi. the rainfa ll dura tion is 3 hr, ond Ihe kllgth \0 di,-ide is 18 rni.
1.23. A small watershed has the characlcrislicsgivcn below. Find lhe peak diKhargc
Qp. t he basin lag lime I". and the time base of 11K: unil h)'drograph T". using
Sn~'uer's method.
3OO'L---;~~~
Unit hydrogroph for
Subbosin C. IPuk - 2fIO I
"".
"".
:g- l~).O
" 1((1.0
"'.0
..,!''-----~,------O.------.!-----~.!-----",",->---!,-,
Time (0.5 hr)
2..28. Using Figure 2-14. find the daily ev aporat ion from a shallow Jake ""illt the
following ehn racteristics:
Mean daily tempe rature - 2S.6"C.
Daily $Olar radiation = 550 callem 2,
Hydrologk Analysis 153
J
0.23
0 .50 '"
",
,• oos
0 -01 '"
7.76
,
6 0
0,02
7.56
7.J 3
,• 0.0 1
0
7.32
7.25
10 0 7.1 9
11 0 7,08·
."
0,01 7, 91
13 0 7.86
~ofihd '"
O.oJ
.,
r_ f
r.../hrj
,
1
'"
"• '"
2.50
1 20
l.J2. A j ·hr storm over I IS·ae basin prod uces a 5·in. rainfall: 1.2 in.lIl r for Ih e firsl
h<'lur.2.1 inJ hr for the second hour. 0.'>1 ,nl hr for the thi rd hour. and 0.4 in./h r
for the last 2 hr. Determi ne th e infiitraliun Ih at would rc~u l l from the Horton
mvdd ",ilh k " I.tlllr. ' .... 0.2 in.lhr. pnd/., - 0.9 in./hr. Plot t h ~ ove rland flow
for this condi tion 111 in./h r YiL r.
2.33. A plol oflhc infih ratoon eo ..·c oblain ed osing HOrloo 's e qua toon iS5hown in
1-1);ure 1-'2- 33. Prove ,har k _ (j" ; ,~) if F" is the area belwee n the cu rve ami
the Ieline. Find 1h ~ Hrca by In tegrat ion ove r t ;me . as 1ime al'prrn,chcs infi nity.
". Cfloplcf 2
figure P2- 33
Horton's i n~hrotiOf1
~l'>z
curve .
~ I- . . . . - --:-::
-_"":":
. --:::
--::-: -- ~- "
- -""
TIme (h.)
2.34. D elermi 'le th e tnckx of Figure 1'2- J.4 if Ihe runoff depth "''as <a) ."i.6 i n. of
rainfa ll over th e W8 lC,,;hcd area and ( b) 6.5 in.
0.8
.Iii
';:: 0.6
.1
f--I
- OA
j - I
0.' -
- -
, • 6 ~ 10 12 14 16 18 ~ n ~
"li me (h .)
us. A saody loam has an in ilial moisture content o f 0.18. hydraulic cond uct;,·;!),
o f 7.8 mmlh r. and ave ragc capill ary !!IJct iOil of 100 RIm . Rain fall5 al 2.9 cmJ
hr. and the fin al moist ure co nlc nt is meas ured to be 0.45. When docs surface
sa lunl1 ion occur? Plot the infiltra tion . 81<; "S. the infli mtion "olumc in 2-35.
2.36. Usc the panunelC~ glvcn to groph Ihe infihnllion note \'S. the infiJll1Iliun volume
for the SlIme SI01m for both IYI1,,:5 of !oOil. Prepare a graph lr>ing the Oreen-Amp!
mo.:thod. compo ring sillhe ~lIrves calcula ted Wilh bolh the lowe r-bound a nd upper-
bound porosi ty parameters. "Tht: rainfall inlC'nsily of the storm wM 1.5 emlhT for
5C\~raI hou rs,lInd the ini tial moisture OOflIC'nt of All [he toiLs ... as O 15.
0.42--0.58
'' '
16 .75
!<m/>;
065
Sondyday 0.3-0.49 23.95 0. 10
Hydrolog ic Anofy$is '55
2.31. The G recn nnd A mpl 'nfiltratiun eq uat ion is a loss funct H)n used IU compute
the eumulati,'e in filtration. F (em) for ~ giw n infi ltnlion ra te .f(cmlhr). ReCll Il
lilal. I. K,' ( ] - (M · wfl/ F). For Ihe given soil propc rlic$ hnd infill Ta lion
rat e. ans ....e r the following.
, , cm/hr
-, ,m
M 0.2"
,.., I l<mI" 'kml
0 0 0
001 15.3.4 0. 15
0.25 3.32 0.95
0.50 2." 2 1.55
0.75 2.07
(a) Compute Ihe mfil tr3 110n ra te./(un/hr) for F .. 2.07 em. and $ha w your
comp utat ions.
(b) Compllle the cumula ti,'c I nfil lr~ t i()f1 resulting from a cunsta nt rai n rale of
O.5emfh r for ] hr.
(c) At s;.luration. what is Ihe IIlr, lim tiun m tc in cmlllr ? Jusl;fy.
2.33. Please refer to the Green and Amptlnfi lt ra tion and Runoff Exa mple posted on
the te~lboo~ web, il ... a long wi lh the associated Excel spreadsheel. Com ple te
t he proble m III lhe example. and the n repea t Ihe pTOe" durc ",·.111 II conSlanl
rainfall raleorSO mmi hr for ].8 hn. DClerm;M: Ihe new cumulali, e f unoCfand
fUnQff eoefr.cicm.
M" I.E. O. H .• ami D M, GaAY. 1981. MSnOWCO\'Cr A hLat ion al1d RlInoff,M in D. M.
G ray and D. H. Ma le (Ells.). Humlbook ,,{SIlO"': PriIlClpl..s. "r"uu.
,111111118"",,,111
.{ U$", P~ rga mon Press, TorOfl tO. pp. 360-436.
MA RCtANO. J . 1.. a nd G. E. HAR Of.CK. 1954. '"Mas_ Tr;msfer StUdie s.'" IVllla Lou
1r1)'esrlg"'io,, ~' Lake f/~frra, USGS Prof. Pape r 269, Was hingtoll , D.C.
MCCU Er< . R. H .. 2005, I ly/ir%gle A I!(I/y<is (l J/{/ Deligrr. 2nd t d .. Pn:nt iC1! HHIL, Upper
Sadd le River, NJ.
Mm •. R . 0 .. and C. L L..5O", 1973. "Modc lin, lnfiltratiOfl Durins a Steady Rain ,"
IVar.. r RI'$Ollr. R~s.• 9(2), pp. JS4-J~.
Muu. A. F .. 190:4 (June). EI·lIpcr..,iOfl {Rim /.lIkes 1111(1 Re..",,·u,TS. Mi llllcsota Re·
wu"es Comm ission. 51. Paul.
NASIl. J. E. . 1')5k. "The Form o f the Ins tantaneous Unil Il yd rogroph," Grrrerlli A s·
s~lII bly of TorlJlllo, ' ''' fr llllli. A$.}Q('. Sri. "."",.01. (GelllbmKlit) (,,, /,, 42. Comp\.
Rend. 3, PI'. 114 1111.
NAsrr . J. E.. 1959. "SySlcma lic Dele n nina l;,)n of Unit liyd rograp h I'aramelers,- J.
CroP"y." Rn .. 64( I). PI'. I l l- l iS.
O' KELll . J . J ..
1955. 1lle Em ployment of Umt·Hyd rr:yaph.'l to Determine the flo,.-s 01
Irish Arterial Drainage Oumneb." InSl. Cn', E'I!. (I rt'/alld) I'roc., .((3). pp. J64-412.
Ovu mI'. D. E .. and M. E. M V,[J()"'5, 1976, $ron"'I"lun Modelirrg. Academic Press,
New York.
158 Choprer 2
Pt!I<)lAN. H. L. 1lJ..i8. "Natural Evaporation from Open Water. Bare Soil. and Grll!l:S, M
P,oe. R. Sac. London. seT. A , 2 7, pp. 77'1-787.
PEN ,'1A)<, H . L , 1950. "Evaporation o ve r the Br ili~h Isics," Qllart'- Roy. M'!t. Soc.,
70(330) , pp. 372.-383 .
.... IT. R. E. . S-E., Oi(.... C· K. 0>'(;. and S. C!.AIlK. 2001 . "Measu remcntsof Infilt ration
Rates in Compnc:ttd Urban Soihl," U"/dll8 SIQ'm "'Q/e~ 8MP OnignS lIm/ Perfo ,·
mll"ct 10 R«t!iving Wala Impact Mi/igll/ion, in B. R. Urbo!1l' s (Ed.), J'roc:eedingi
En gineeri ng Fou n,jul ion Conference, ASCE, Reston, VA. pp. 534- 538.
P lrr, R . E.. J. l.ANTlIP. R.H ANK. SO:>!, C. 1.. H F_"U. and O. X UE., 1999. Infiltration
IIlrough Disturbed Urban Soils ami Compo~l-A",mded 50;1 EfI«1S On Runoff
Qua/iry lllld QUQil/iry. EPAf6OO{R-OOKJ16, U.s. Environmental Prolea ion Atency.
Ci ncinnati. OH.
PtullP, J . R.. 1 ~ 1. "The Theory of InfiLt ration: I. T'hc Infiltrat ion Eq\l~ . ion an d Its
Sol \l tjon,~ Soil Sci., 83, pp. 345-357.
1'~lI'STLY.
C H. B.. and R. J. T",~I.O~. 1972. "On the I\ssess~ n' of Surfare Hell' Flux
and E~1Iporalion Using large Scak Pal1llT1ele n :' M(}j~ \I'M/liD" Rev., 100. pp. 81- 92.
R... UOKIV•. A . J .• 1919. Hydrology, Pergamon Press, Elmsford. NY.
RAw l.$. w . 1., O. 1_ BIlAKENSIH.• and N. MII.LU. 1983. - Grecn·Ampt lufil trat ion
Para meters from Soils Da.a," J. lIyd",ulic Eng. , ASC E. / 09( I ), pp. 62- 70.
RI,l ~IN . 1.. and
R. Sn:I:'< !tARDT, 1963, - Soil Water Rela tions During Rai n Infiltrntion.
I: Theory," SOli Sci. Soc. Amer. PrIX .. 1?, pp. 246-251.
RUIIN. J., R. StE.lNUAROT, and P. Rm< lCu, 1964. "Soil Wa ter Relations Duri ng Rll in
Infillra'ion , II: Moililure Conten. l'rofilts Ouring Rains o f Low Intensilies," Soil
Sci. $(J.c. Amer. Proc .. 28. pp. 1- 5.
S"'L Is 8un, F. n .. and C. Ross. 1969. Planl Physiology, Wadsworth Publishi ng.
Bel mon t. CA .
Sue R)'!"'/<. L. K.. ]932. "Strcamnow from Rainfall by the Uni' ·Graph Method." £llg.
NrNJ· Rec.. 108. pp. 50 ]- 505,
S,U H. S. F .. L H. A ~~tl<. 11 .. L. C. HAloi)lO!<O. J , W. JO .... F.5, 1. S. R OGUS, and A. G .
SWNsnu.. 1983, " Basinwide Wate r Require mem Esti mat ion in Southern Flor·
ida," Traru. A m. Soc. Agri. Ellg r, 26(3). Pl'. 700--766.
SmrrrLEwoRnl , W. J .. ]993, ~E vapormion ," Chapter " in IllIIufbook of Ilylfrolol/Y.
D. A. Ma idment (Ed.), McGmw Hill , Ne w York, 1993.
SI1'I()H, V. P .. 1996, KillemOlie WtJ"" M CHlclillg ill Wmer Ruou reu, John Wiley &
Son§., Inc" New York.
S~AG(.;S, R. W ., and R, KH"'lF.Et~ 1982, "Ch~ ptcr 4, [n lilt rltion ,~ in lI)'d mwgfc M CHI·
elillgofSmtJli Wmusheds, C. T. Ha nn , J . P. John!lRn, and D. L. Brakensick (Ed~.) .
America n Socie ty of Agricult ural Engineers, Monogr~ ph No.5, St. Joseph, MO.
SNY OU, F . F ., 1938. RSY nl helic Unit GrJphs," Trom. Am. Gcophys, UI1io ll, 19,
pp.447-454.
SoIL CONSf.J.VATIO!'I SUI' ICE, 1?S7. u~ of SlOrm alld Wale,shed CharaCfuislics ill
Sylllh,,'i( Ilydrogmph A llillysis mul Applieolloll. U.s. Department of Agricul1ure,
Washi ngton. D.C.
HydrologiC Anotysis 15.
SoiL CO:-SERVATIOI'I SERVICE. 1964. SCS NnriOlUl/ Enginu'mg /Iundbook. S(CliOlI
4' H}·tlr%K)'. updated 1972. US. Dep.mmcnt of Agriculture. Wa~hinl!ton, D.C.
Sot l. CO"~EI\'ATlON SE:.,~(£. 1986. Urball H,..I,oIo8,v fo' Smull WU/I'13hNls, 2nd ed"
Tech. Relea;;e No. 55 (I\'TIS 1'1387.)01580). U.S. Department of Aglieu llurc.
Was hington. D. C.
SouU I FLORIDA WAn. MA"AGF.." . i' I D ISI NIL'-. ]987. Mallagement 0"'/ SlOrug~ of
SlIrJacr Warus. Pam" luto,mll/jUlI MI",w,l. Vul. IV. Wesl Palm Beach. FL.
STErHH:so~. D.. and M. MEADO"'S. 19M. K",,,,,,,tltc lIydrology 111111 Modeli"g. EI§c-
vicr Science Publish ing Company. New York.
T UOI''''I HwAnE . C. W.. 19-'11. ~An Approach Toward a Ralional CI;usifical ion of
Clim(\1c." Grog"'plt. Rev.. J.!I. pp. 55-94.
Tropical Storm A!1iwn Recovery Pro~el (TSARP). 2001. hllpJ/www.I$IIrp.org/
TURC. L.. 1954. "Calcul du Bilan de L'cau E"alumion en Fonction des Prtcipit~tiorus
et des Tcmptralun:s.~ lASH Rome Symp. III Pub. No. 38. pp. 188-202.
UI<llU) STATU A.NV Cou'S Of E"GIN€US. NOlm PACi f IC DrvisioN. 1956. S"O'"
lIydm/ogy. Portland. OR.
U.S. GfOLOG ICAL SURVEY. 19110 ( Fcbrulory). USGS-EPA Urban Uydr%sy Stui/ies
ProS"'''' TechniclI' Coordination 1'11111. draft. 30 pp .. Reston. VA.
V... " HAVEL. C. H. M .. 1966. "P"le ntinl Evaporation: The Combin ation Conce pt and
ll! Expcrimcnlal Vcrification." H'(t/a RI!SOllr. Res .. 2(3). pp. 455-.j.(i7.
VI CS.~MAN. W.. J R.. and G. L. L I;WIS. 1996. Itllrtl(/UCliolllO 11>~lroJogy. 41h cd .. Hltrpcr_
Colli ns. New York.
VI~U"'. B. E .. 2002. Disl,ibme,' H}·tfmlogic Modeling in GIS. 2nd ed .. Klu we r.
Holland.
WV.VER. H. A .. and J. C. STt~IIESS. 1%3. ~Rclalion of E"3p<>f8\ion "f Potential
Evapotrnnspir.llio11.~ T",ns. Am. Soc-. Agr. E'lgs. 6. pp. 55-56.