|
<br />~-
<br />
<br />.---~..
<br />
<br />-
<br />
<br />.
<br />
<br />SCOUR AND FILL IN STEEP, SAND.BED EPHEMERAL STREAMS
<br />
<br />565
<br />
<br />Simulated Flood Hydrograph. laborarory experiments were
<br />simulated floods that used an ephemeral stream or "flash flood"
<br />hydragraph of ,he form
<br />
<br />O<t<lOs;
<br />
<br />w:lter turned on at inlet, Q rises almost linearly from 0 to Qmax
<br />tOs<t ~to; Q Q......
<br />Q [_ 3(1 - t,) ]
<br />to < t < tmu; Q mn; ~p .
<br />(tmu - to)
<br />
<br />where Q = water discharge. t = time from beginning of simulated
<br />tIood, to = O.lt mu, t max = 1 h = durarion of run. and Qmu =
<br />7,500 cm.l/s.
<br />The exponential rail of this hydrograph is representative of the
<br />taHing limbs of surface runoff hydrographs of ephemeral streams in
<br />California (Gupra and Moin, 1974). The flartop peak and short,
<br />nearly linear rise are similar to those of the Alamogordo Creek
<br />watershed (Renard and Keppel, 1966). This very sharp rise ra
<br />maximum discharge, requiring less than 10 s, along with the an-
<br />tecedent wet ~d condition, produced an abrupt translatory wave
<br />JS rhe leading edge of che laborarary flash flood, The single "wall of
<br />water" translatory wave is thought by Renard and Keppel (1966)
<br />[0 be relatively uncommon in the field; a series of translatory waves
<br />,of small amplirude building ro full flood depth is the general be.
<br />havior. The laboratory flood wave increased in height as it passed
<br />through the inlet section and the upstream pan of the test section;
<br />rhis indicates behavior qualitatively similar ro that generally ob-
<br />~erved in me field, but of such a small scale and-&horr duration that
<br />Individual translatory waves did not occur.
<br />Sediment-lnput Relaoon. Since rates of water discharge and
<br />)~diment input rate were independent variables in these openM
<br />,:Ircuit experiments, any sediment-input relation was possible. For
<br />Jiffering water-SOUrce circumstances in the field., such as snowmelt,
<br />rhunderstonn, or steady rain, mere will be greatly varying relations
<br />between rate of water input and rate of sediment input. Modeling
<br />[his process was beyond the scope of this research. A relevant
<br />'iediment-transporr relation between sediment-cnput rate Gd and
<br />water discharge can be derived horn field observations (Bennett
<br />Jnd Sabol, 1973):
<br />
<br />CII = cxQ-',
<br />
<br />.~.here cr: and f3 are constants. For Rio Grande data (Nordin, 1964)
<br />.vich sand grain-size similar to that in these experiments, a fit of all
<br />J.1Ca in Nordin's Figures 16 and 17 gives 1.8 < 13 < 2.4. Bennett
<br />'nd Sabol (1973) found that the best fit to field data is obtained by
<br />I zero.intercept quadratic relation of the form
<br />
<br />G. ~ yQ' + 6\2,
<br />
<br />,vhere y and 0 are constants. For these experiments, a compromise
<br />;edimenr-input rdation was chosen of the form
<br />
<br />GsI. =aQ2,
<br />
<br />.vith Gma:('d = 41.1 cm3/s at Qmu = 7,500 cm3/s.
<br />Establishment of a Graded Scream. Simulated floods were run
<br />Ising the above input conditions. The initial channel bed was ar-
<br />lticially leveled and wetted so that no trapped air would interfere
<br />"l(h pressure-transducer measuremencs. After each flood, mean
<br />:Jevation and slope of the channel bed was detennined and com.
<br />oJred with the preflood values. If initial and final values differed,
<br />')ume slope was adjusted to the final channel bed slope before the
<br />lext run. A series of runs permitted the flume slope to be set so that
<br />lIirial and final bed elevations were the same within the resolution
<br />li pointMgage measurementS. [nitial and final slopes differed by no
<br />
<br />more than 3%, and bed conditions differed only in that the initial
<br />bed was flat and the final bed was rippleMcovered. Since it was
<br />necessary to level the bed to determine mean elevation and slope,
<br />this difference was unavoidable. However, it had no significant c:f.
<br />fect on experimental results since the ripples would have been de-
<br />stroyed in the..first few seconds of the next simulated flood. With
<br />final mean-bed elevation and slope essentially constant from one
<br />run to the next, the channel satisfied Mackin's (1948) definition of
<br />a graded stream and allowed measurement of changes in mean-bed
<br />elevation during simulated floods without net changes.
<br />Mean-Bed Elevation Determination during Simulated Floods.
<br />No satisfactory method was found ro measure mean.bed elevations
<br />directly while water was running. For this reason, mean-bed eleva.
<br />tion for the flume as a whole was determined at several rimes dur.
<br />ing a flood by dividing rhe known volume of bed sand in ,he flume
<br />by me bed area. The volume of sand in the flume was determined
<br />by sampling rotal sediment discharge at the outlet, subrracting it
<br />from the known rate of sediment input, and integrating the differ-
<br />ence as a function of rime. A basic assumption of this method is
<br />that there were no persistent or propagating discontinuities in mean
<br />bed elevation, such as a single scour or fill "wave:' No such disM
<br />conrinuiries were visible during simulated floods; merefore. this as.
<br />sumption appears valid. Discontinuities smaller man about 5-mm
<br />amplirude probably would nor have been deteaable on a ripple.
<br />covered bed. but ripples formed only late in each simulated flood
<br />when sediment-transport ratcs were very low and changes in
<br />mean-bed elevation effectively over.
<br />
<br />Experimental Resules
<br />
<br />A graded channel was established fOr the chosen hydrograph and
<br />sediment-input curve at a slope of 0.0089. Six simulated floods
<br />were 111n at this slope, and the mean.bed behavior was calculated.
<br />Calculated mean-bed behavior was sensitive to error in sedimene-
<br />discharge measurements at the outler, since the calculation used
<br />measured discharge as the average discharge in the interval berween
<br />samples. Thus, in aU cases me calculated change in mean.bcd e1e4
<br />varian for a complete simulated flood differed from that measured
<br />by point gage after the flood.
<br />Figure 6a is a plot of the deviation of mean.bed elevation from its
<br />initial value against rime for run F-l-l1. The calculated change in
<br />mean-bed elevation has been adjusted in twO ways co match the end
<br />result measured by poine gage. The openMcircle pointS in Figure 6a
<br />represent an adjusted curve in which the mean.bed elevation
<br />
<br />
<br />~ .06
<br />> D'
<br />~
<br />~
<br />~ D2
<br />~
<br />~ 0
<br />~ ..06
<br />~a
<br />o
<br />e: .50
<br />z
<br />o 0
<br />:i
<br />~ -.~o
<br />o
<br />~...oo
<br />~
<br />~
<br />o .c5Q
<br />~
<br />m
<br />~-l.OO
<br />~
<br />~
<br />b 0
<br />
<br />AREA OF
<br />!':;!lI..~r+1I wm.. BOOY-SMIFT AQ.AJSTlolEN
<br />VALUE F'-HI WITH LWEAR I
<br />, : Al).JUSTMENT--,
<br />i 1 SAND~N
<br />
<br />so
<br />
<br />60
<br />
<br />'" '0
<br />ELAPSED TIME IMINI
<br />
<br />'0
<br />
<br />20
<br />
<br />SAND
<br />GRAIN
<br />
<br />MEAN-8ED ELEV BEHAVIOR
<br />,
<br />
<br />I,IIAx.ANnOUNE
<br />TROUGH OEPTH
<br />(AGAINST WINDOWI
<br />
<br />
<br />-+--
<br />
<br />I,IIAx...R1Pf>1.E
<br />TROUGM DEPTH
<br />1ACA,IN$ T W1N09W t
<br />
<br />,
<br />'0
<br />
<br />'0
<br />
<br />60
<br />
<br />20
<br />
<br />lO '"
<br />ELAPSED TIME {I.UNt
<br />
<br />Figure: 6. Calculated mean-bed devacion change during run F-l-l1; (OIl
<br />adjusted curve:; (b) unadjusted curve with bed-fonn reworking added.
<br />
<br />i
<br />r'
<br />
<br />II'
<br />L,
<br />.1-,
<br />!
<br />
<br />j.'
<br />,
<br />
<br />I
<br />i
<br />
<br />:1
<br />:1
<br />,j
<br />1';'1'
<br />"
<br />:il'/'
<br />,"
<br />':1
<br />"
<br />:1
<br />il'
<br />','
<br />,I
<br />"
<br />II.
<br />,Ii
<br />l
<br />,j:!,!'
<br />"
<br />I
<br />l! ,
<br />Iii.
<br />1],!:,.
<br />:i1
<br />
<br />~I':
<br />;: I ~j:
<br />Ii ! ~
<br />:' i:
<br />'I'
<br />, ,
<br />, ,
<br />I'
<br />II
<br />.:111'1.",:
<br />~ ' ,
<br />'"Ii
<br />:,.:':I:i '.'
<br />1" ~
<br />, l'
<br />,II:
<br />
<br />, 'I
<br />
<br />:jlil:
<br />I,,' I
<br />Iii
<br />llr
<br />
|