|
<br />TABLE 2. Hydraulic, energy, and geomorphic data for ten well-documented floods that demonstrate different kinds of stream-
<br />power graphs
<br />
<br />Peak stream Mean Duration Energy expended Geomorphic
<br />power stream power (s x IOJ) per unit area impact
<br />(Jilm') (Jilm') Goules x 10')
<br />
<br />52 GEOMORPHlCALL Y EFFECTIVE FLOODS
<br />
<br />Flood
<br />
<br />Kind of
<br />power-
<br />graph
<br />
<br />Reference
<br />
<br />Centralia, Wash. 3300 1650 0.38 620 Small C Costa. 1994
<br />Porter Hill, Oreg. 2900 1450 1.0 1500 Small C This report
<br />Plum Creek, Colo. 630 110 68 3900 Extreme B Osterkamp and
<br /> Costa, 1987
<br />Roaring River, Colo_ 4300 1200 7.2 8500 Extreme B Jarrett and Costa,
<br /> 1986
<br />Rubicon River. Calif. 6100 3600 22 29,000 EXlreme B Scan and Gravlee,
<br /> 1968
<br />Teton Dam, rd. 17,200 3400 29 109,000 Extreme B Ray and Kjelstrom.
<br /> 1978
<br />Mississippi River. Ark. 12 6 1200 21,600 Small A Baker and Costa,
<br /> 1987
<br />Bonneville Flood, 300 150 11,200 1,700.000 Small A O'Connor, 1993
<br />Burley Basin, 1d.
<br />Bonneville Flood, 90,000 20,000 11,200 220.000,000 Extreme B 0' Connor, 1993
<br />Rock Creek, rd.
<br />Missoula Flood. 60.000 8100 430 3,500,000 Extreme B Benito and
<br />COlumbia River Gorge, O'Connor, 1991
<br />Oreg. and Wash_
<br />
<br />-~-
<br />
<br />previously, nor to the Mississippi River floodplain or the
<br />Snake River alluvial floodplain at the Burley Basin in
<br />Idaho. The Mississippi River and Bonneville paleoflood in
<br />the Burley Basin were similar in that wide alluvial
<br />floodplains and flat channel gradients prevented peak or
<br />average stream power per unit area from exceeding erosion
<br />thresholds. The other floods all caused severe and
<br />widespread channel and floodplain erosion, channel
<br />modifications, and erosion of bedrock, where present.
<br />These floods exceeded alluvia) or bedrock erosion
<br />thresholds, and were clearly effective geomorphic agents.
<br />
<br />3.2 Calculations of lotal energy expenditure using time-
<br />integrated stream power per unit area
<br />
<br />The average energy per unit area (0) that is expended
<br />over the duration of a flood can be represenled by:
<br />
<br />o = hQS/w dt
<br />
<br />where 'Y is specific weight of the fluid (9800 N/mJ for
<br />clear water), Q is discharge in mJ/s. S is energy slope,
<br />
<br />w is water-surface width, and t is time in seconds. We
<br />have numerically calculated [J for seven large, well-
<br />documented historical floods, and two paleofloods (Table
<br />2), by evaluating reported measurements of valley cross-
<br />sections, the flood -hydrograph, and a stage-discharge
<br />curve. Limitation of the data sources are discussed
<br />below.
<br />Following floods, hydro graphs are constructed in a
<br />variety of ways. The ideal situation is to have a stream
<br />gage properly operating throughout the flow. In other
<br />situations hydrographs can be constructed from peak-
<br />discharge measurements, observations of duration, and
<br />assumptions about hydrograph shape [e.g. Costa, 1994]
<br />(Figure 7). For dam-failure floods, downstream hydro-
<br />graphs can be constructed from reservoir draw-down rates,
<br />or dam-break models [e.g. Jarrett and Costa, 1986].
<br />Cross-sections of channels and floodplains are nearly
<br />always made during surveys following floods [Williams and
<br />Costa, 1988] (Figure 8). They are required for determining
<br />the hydraulic variables necessary to calculate discharge.
<br />The primary problem with cross-section accuracy results
<br />from possible scour or deposition during the flood, and the
<br />
<br />--
<br />
|