|
<br />,~.a ~ I~ "... . , . ' I . ~
<br />. ~,,\too ~ _ ._, " I ,.','.~ .>-,'~'.'~'.~f'"
<br />
<br /><
<br />
<br />MA~-28-2000 rUE 10:42 AM
<br />
<br />FAX NO,
<br />
<br />P 01
<br />
<br />'70",., B;e.dWI"HrJ <S 30:'5 S61...4-4-71\ ~~ ::r;;ee..G-)~I-
<br />(F::X , '3 0 :> 1./~-6 - b t; ~ 7-
<br />JOURNAl. OF THE AMERICANWA'l'Kll RESOURCES ASSOCIA'l"ION C! Ii)
<br />\"11. :W,NO, 1 "MERlCANWATEltRESOUIICESASSOCIATION FE8RUAIlY20UO ./
<br />
<br />-
<br />
<br />-
<br />
<br />}~STIMATION OF Tm~ HEel LOSS PARAMETERS FOR
<br />ROUTING 'l'HE PROBABLE MAXlMUM FLOOD I
<br />
<br />Michel C. llou{Ctclel2
<br />
<br />n
<br />,
<br />l
<br />I
<br />I
<br />,
<br />!
<br />!
<br />,
<br />I
<br />j
<br />l
<br />!
<br />
<br />AUS'l'H^C1~ Twenty etol'TLl events wc-re wood to iclc(;t design valueii
<br />.r Lh. IlI!:Cllo," paromcL.rs S'I'TtTL ami eNS'I'I, in ortler '" route
<br />Lhi,! 1.l\llJabl., m;lximum nood, I'MF, thJ'ouith the Enlllewood water.
<br />~I'l.d, Ohio, 'l'ho1'I.JnJ.JJlctet' Srrn1'L rcpn:scl1ts the initiol volume or
<br />""hLor JOlSl dUD 1.0 illtCl'CGJ)tion UllU incompl"te ,olul'l.1r.iOll or the soil
<br />priUl' to ~he SLOI'n1, "hllil pilornmcter CNSTL. rcp11:!scl\ts a i:t1ntinuous
<br />1o~" ruta and depl4mlll only on thlil' w;:J.l.criihcd, When opLhnizet\ !l'oln
<br />each litonn event., S'l'H'\'l.. varied between 0.0 anu. 3.4 inches wiLh
<br />lln ~\'ien"Q or 1.0 ineh; CNS'l'I4 v:uilied botween. 0.02 ~mu. 0,20
<br />inch/bout, anu it. fCJllowcd a nol'1nal proD:\hiHty disLribut\ol\ with. a
<br />I1lCan of nbout 0,1 inchlboul', 1'M abs.enec or con-elation between
<br />oplimum CNSTL vi,11l,l.tHij anc.l each or totnl ninfi\ll, ~ot:1110ili, D.c<1
<br />rUl\tJtl' dUl''lllinn lliUI)pol'lOd the lit!l~c:tion or the mc..tn eNS'l'(. as u
<br />dl.:~iltn villuc. PMI;' nutinll through thu Englewood watershed
<br />rf.lvc.\!l;d t.hnL Lhe l'MIi' ll~ ihul,hlllot is not iCIUiiLivc to STft1'L. bu~
<br />hlflhly Dflec:wd uy eNS'l').. variiltions. 'J'he in$ctl:S:iitivity to St'RTi.
<br />"';1:1 L1Ull! ..u ..liD rru.I.JI\C111 gr b 1l~\Ul tAl. "hll ouUul. ur lht waLar&h~d
<br />LIM\ Cllustld the buUl.lop or watc-r in the wa~nhc:d, ~hcreby mD.lik.
<br />tllS Lhco l!Itorage errect of STR'rt.. The peak PM}'" Increalicd by Ll.bout
<br />27 purcent. whl2n. the design eNS1'J", walil decreased to 0,015 inchl
<br />}IOUl'. gOl.I dccJ'ensed by abol,l:t 18 percc-nt whe-n the deljiif1"l. eNS'!'I..
<br />'Willii illcI'elU5eQ lfJ 0,16 ;ncMlDur.
<br />ll'l~Y TERMS: rainfull I...: 10.. parumeters: PMF; b..u1Iow; opti-
<br />mum,: UI~Cl.)
<br />
<br />.
<br />i
<br />i
<br />I
<br />1
<br />i
<br />i
<br />j
<br />1
<br />,
<br />!
<br />
<br />INTRODUCTION
<br />
<br />Many federnl and stnte agencies hnve adopted the
<br />prubable maximum precipitation, PM?, (WMO, 1986)
<br />lIod tho probable mnximum flood, PMI;', concepts as
<br />design and operating criteria for engineered dams,
<br />'I'ho routing of the PMF through ll. watershed requires
<br />knowledge of the rainfall loss (due to infiltration),
<br />O~ercstimation of Tainfnl1 loss in the design of dam
<br />elll"ankmcnt.~ cnn rc~ult in overtopping of the dam in
<br />the event of a PM? On the other hand, huj.(c sums of
<br />
<br />money would be spent in vain to elevate the embank-
<br />Incnts if rainfall loss were underestimated,
<br />Tn most rainfall-runoff models, rainfall absLraction
<br />is represented by a conceptual model that depends on
<br />n few parameters, A widely useel rainfall loss model is
<br />the two-parameter Initial and Uniform option in the
<br />IIEC1 (USCOE, 1980) package, This loss Inodel
<br />lISSllmes that all rainfall is lost until a volume of ini-
<br />tia I loss STR'l"L (measured in inches) is satisfied.
<br />Afterward, rainfall is lost at a constant rate, CNSTL
<br />(meas\lrcd Cn inchesJhour), The parameter S'l'R'l'L
<br />varies between storm events because it depend. on
<br />Lhe existing soil moisture befoTe the stOl'ID. 'l'he
<br />pRO'lnnct.r CNSTL doponds on watci...hod charactoris.
<br />tics such as topography, soil covel'age, and soil perme-
<br />ability. This parameter is commonly assumcd
<br />invariant froIl! one storm event to another in a water-
<br />shod. Because these parameters are watershed-avor.
<br />aged quanti ties, they cannot be obtained by p,hysicnl
<br />mC;lsurement, Consequently, they are estimated by
<br />litting computed (or simulated) flood hydrographs to
<br />observed ones for a set of rainfall events, However,
<br />each event leads to a different value of CNSTL. The
<br />varit\tlon ofCNSTL estimates between storm events
<br />is due to the inability of the model to represent nceu.
<br />rotely the dynamics of the watershed, and stochastic
<br />and/or sampling errors (Sorooshian and Dracup, 1980;
<br />Dotes and Townley, 1988; Bates, 1988; Kuczcra,
<br />1990).
<br />The estimation of the loss p;lrameters from ;I stonn
<br />event requll'es npal'ation of the baset10w (due to
<br />I:roundwnter flow) from the total runolf hyclrograph,
<br />'rho com mOll approach assumes that baseflow Nccdo.
<br />at ,m exponential rate (Horton, 1933; Werner llllll
<br />
<br />IPnll(-~' No. !)~OA:3 oC\.hlJ JIJuflltll arrha AmltriC:'1/1 \Vnt~r l~t:jQ!,m!!f'A"isur:inljva, llhiCU.!irtions ar~ opon unUt October 1, 2000.
<br />~A::!llIh;~ullt l'f'orf):lRor, Ol.:'pt, of Civi\ tlnrll~nviJ'Qn.mental EnGiMtwing. 'r\![l1(lt~ Ull\versit.;r, 19.17 North 12th S~., PhiladelphlLl. Penz,sylvl1ni"
<br />J ~J 1':2 t1':, Mail: bDuCi.ulclo-">>h'O,tl.!'\")llo...,du).
<br />
|