Paleodischarge of the late Pleistocene Bonneville Flood, Snake River, Idaho, computed from new evidence
c' l^niT 1 US Geological Survey, Federal Center, Box 25046, Denver, Colorado 80225 ilAKULU L. MALL/IJ )
ABSTRACT The path followed by the Bonneville Flood down the Snake River was greatly modified by erosion and deposition. Particularly impressive The Bonneville Flood resulted from catastrophic outflow from are abandoned channels, areas of scabland, and gravel bars composed of Pleistocene Lake Bonneville about 15,000 yr ago, when the lake over- huge boulders and sand. Parts of the Snake River Canyon are now known topped its rim at Red Rock Pass in southeastern Idaho and discharged to have been flooded to depths greater than 130 m. a vast volume of water down the Snake River. This paper provides The altitudes of the erosional and depositional features produced by revised estimates of the paleodischarge, volume, and duration of the the Bonneville Flood indicate the maximum flood height and are used to Bonneville Flood, based on new evidence of its height and on current reconstruct the flood profile. When the profile is considered in the context understanding of the amount of lowering of Lake Bonneville. Evi- of the dimensions of the Snake River Canyon, particularly where the dence for the revised height of the flood is derived from the altitude of canyon is constricted, a peak discharge can be calculated. Based on new erosiona! features and flood deposits at the head of a constricted reach evidence of the flood height in a constricted reach of the Snake River at the of the Snake River Canyon at the mouth of Sinker Creek and from the mouth of Sinker Creek, 540 km downstream from Red Rock Pass, this altitudes of flood deposits at several places about 53 km upstream. paper describes such a calculation, using the step-backwater method Using the step-backwater method, we estimate that peak dis- (Chow, 1959). The constricted reach is cut in basalt between Mile 460 and charge for the Bonneville Flood through the constricted reach was Mile 462 (Figs. 1 and 2). Given certain assumptions, the peak discharge is 3 from 793,000 to 1,020,000 m3/s and most likely was 935,000 m3/s. estimated to have been from 793,000 to 1,020,000 m /s and most likely 3 This discharge is 2.2 times the discharge previously reported and is the was 935,000 m /s. This estimate is 2.2 times the discharge previously second largest flood known to have occurred in the world. At this rate calculated (Clifford T. Jenkins, in Malde, 1968, p. 11-12), using a flood of discharge, the shear stress for the flood would have been 2,500 height that recent study has shown to be erroneously too low. The esti- N/m2, and the unit stream power would have been 75,000 N/m/s, as mated discharge then is used to appraise the geomorphic significance of the compared with values of 6 to 10 N/m2 and 12 N/m/s for recent floods flood and to re-evaluate its duration, based on current understanding of the on the Mississippi and the Amazon. Other recent studies of the history amount of lowering of Lake Bonneville. of Lake Bonneville show that the volume of water released was 4,700 3 km , or about 3 times greater than the volume previously inferred. NEW OBSERVATIONS OF FLOOD HEIGHT Although this volume indicates a flood duration of 8 weeks at constant peak discharge, an accurate estimate of the duration would require Geologic mapping from 1979 to 1983 downstream from Bruneau dam-break modeling at Red Rock Pass. From a dam-break model, disclosed several features of the Bonneville Flood higher than any pre- flood hydrographs at Red Rock Pass and the hydraulics of the flood viously identified in this reach of the river. The flood features were missed wave along the Snake River could be computed. during earlier reconnaissance (Malde, 1968) because of their remoteness and difficulty of access. In the following discussion, altitudes of flood INTRODUCTION features are given in feet, in accord with contours on existing topographic maps, but other dimensions are given in metres. During the Pleistocene, closed basins in the Basin and Range prov- Erosional features on a basalt bench 133 m above the Snake River ince in the western United States contained lakes of considerable size (altitude 2,750 ft) at Mile 462 (Figs. 3 and 4) resemble those produced by (Williams and Bedinger, 1984). The largest of these Pleistocene lakes was the Bonneville Flood on basalt uplands elsewhere along the Snake River in Lake Bonneville (Fig. 1), estimated to have once covered 51,530 km2 southern Idaho (Malde, 1968, p. 20). Basalt has been partially stripped (Currey and Oviatt, 1985). The surviving remnant of Lake Bonneville is from the bench, leaving broad remnants that have irregular, embayed Great Salt Lake, which covers 4,700 km2. Lake Bonneville fluctuated outlines. The bench itself has a nearly vertical face 15 m high on the greatly in altitude, last reaching its highest level, the Bonneville shoreline, downstream side, and the erosional remnants have faces about 2 m high. about 15,000 yr ago (Scott and others, 1983). At that time, the lake The upstream faces have been smoothed and rounded, presumably by the overtopped its rim at Red Rock Pass and catastrophically discharged a vast abrasion of sediment-laden waters. The downstream faces end abruptly, volume of water down the Snake River in southern Idaho (Gilbert, 1878; without adjacent debris, and are thought to represent the effect of plucking Hunt, 1980)—an event known as the Bonneville Flood (Malde, 1968). of lava by the forceful passage of the flood. Scattered boulders of basalt on Red Rock Pass consequently was eroded to the level of the Provo shore- the downstream part of the bench are as much as 2 m in diameter. The line, and the flood ended. largest ones are angular, but some have rounded edges. Rounded smaller
Geological Society of America Bulletin, v. 99, p. 127-134, 7 figs., 1 table, July 1987.
127
Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/99/1/127/3952298/i0016-7606-99-1-127.pdf by guest on 02 October 2021 128 JARRETT AND MALDE
boulders and a few pebbles and cobbles also are present. The large Upstream from Mile 462, the valley of the Snake River broadens into boulders clearly a re lag deposits related to erosion of the bench. a wide basin that extends 53 km upstream to the mouth of the Bruneau A gravel bar in the lee of the basalt bench, on its northeastern side, is River (Fig. 1). The Bonneville Flood was hydraulically ponded in this thought to be a deposit of the Bonneville Flood (Figs. 3 and 4). The bar is basin because of the constricted canyon between Mile 460 and Mile 462 150 m long, 64 m wide, and about 15 m high. Angular and rounded (Malde, 1968, p. 39-42). Deposits of flood gravel in the basin occur only boulders as much as 1 m in diameter are scattered on the surface. The bar at the upstream end, mainly as deltas whose altitudes indicate the min- shows no evidence of post-flood erosion and apparently is preserved in its imum height of the ponded water. At site A in Figure 1 (SE'/4 sec. 5, T. 6 original form. The position of the bar in the path of the flood shows that it S., R. 5 E.), deltaic gravel at an altitude of 2,700 ft spilled into ponded was deposited where the velocity of flow decreased. Together, the eroded water in the Bruneau Valley. At site B a short distance west (SE'á sec. 11, basalt bench and the gravel bar indicate that the altitude of the Bonneville T. 6 S., R. 4 E.), floodwater that washed across a ridge at an altitude of Flood at Mile 462 was at least as high as 2,750 ft. about 2,750 ft deposited deltaic gravel in ponded water at an altitude of 2,725 ft. These altitudes are consistent with the altitude of flood features at
116°45' 116°30' 43° 30' Nampa 8 mi R3W
Walters Basin
f IDAHO
116°15' 116°00'
Murphy
I AREA I MAP I J UTAH
Mountain Home 8 mi fit 115°45' m
00'
EXPLANATION E Marginal or upland channel o
a Flood deposits E E Area of Bonneville Flood
470 Miles upstream from mouth of Snake River
@ Locality discussed in text 10 MILES
T 15 KM 42° 45'
Figure 1. Map of the area inundated at the maximum height of the Bonneville Flood between Mile 4.31 and Mile 519. Miles on the Snake River are measured upstream from its confluence with the Columbia River. Inset map shows the northern part of Lake Bonneville (shaded area) and the location of Red Rock Pass.
Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/99/1/127/3952298/i0016-7606-99-1-127.pdf by guest on 02 October 2021 PALEODISCHARGE OF BONNEVILLE FLOOD, IDAHO 129
Figure 2. View upstream of the constricted canyon extending from Mile 460 to Mile 462, where a peak discharge of 935,000 m3/s for the Bonneville Flood was calculated. At Mile 460, the canyon is 675 m wide at the rim and 200 m deep. During peak discharge, water in the constricted canyon was at least 133 m deep.
Mile 462, which marks the downstream end of the ponded water and the CALCULATION OF PALEODISCHARGE head of the constricted canyon. Downstream from Mile 460, the Snake River Canyon widens At Mile 460, the Bonneville Flood was confined by a rock-walled abruptly and contains scattered gravel bars deposited by the Bonneville gorge, cut in resistant, basaltic lava flows and tuff (Figs. 2 and 4). The Flood. The bars are littered with well-rounded basalt boulders, many of flood eroded the bedrock walls and talus but probably did not appreciably them more than 3 m in diameter. The altitudes of these gravel bars indicate deepen the canyon (Malde, 1968, p. 42). The altitudes of flood deposits the minimum height of the Bonneville Flood at its highest stage: an altitude downstream indicate that the flood was unimpeded and that a state of of 2,525 ft on the right bank at Mile 459.2; 2,550 ft on the left bank at critical flow existed at Mile 460. Under conditions of critical flow, the Mile 457.4; 2,560 ft on the right bank at Mile 455.6; and 2,480 ft on the peak discharge of the Bonneville Flood at Mile 460 can be calculated from left bank at Mile 451.7. The flood also overtopped the right rim of the the height of floodwater upstream and the dimensions of the canyon, given canyon at Mile 448 and deposited basaltic sand at an altitude of 2,545 ft. certain assumptions about pertinent hydraulic variables. The calculation of From these altitudes, a profile for the Bonneville Flood can be con- critical flow is one form of the step-backwater method (Barnes and David- structed, and a paleodischarge can be estimated, based on the dimensions of ian, 1978), which generally is recommended and applied for developing the channel and its slope (Chow, 1959). An estimate of the paleodischarge, stage-discharge relations in natural channels (Chow, 1959; Bailey and computed for the hydraulic conditions between Mile 460 and Mile 462, is Ray, 1966; Druse, 1982; Da vidian, 1984; Potter and Walker, 1987). described in the following section. Hence, we calculate the peak discharge of the Bonneville Flood by using
Basalt Bench
Figure 3. View upstream toward the canyon of the Snake River at Mile 462. The basalt bench at the left of center, which reaches a height 133 m above the river (alti- tude 2,750 ft), was scoured by the Bonneville Flood. Farther left, the light-colored mound in the lee of the bench is a grass-covered gravel bar deposited by the flood.
Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/99/1/127/3952298/i0016-7606-99-1-127.pdf by guest on 02 October 2021 1 16d22'30"
Figure 4. Map of the con- stricted reach of the Snake River Canyon between Mile 460 and Mile 462, showing locations of cross sections used to calculate peak discharge of the Bonneville Flood by the step-backwater method. The cross sections are numbered accoriiing to their dis- tances in metres upstream from Mile 460. The area covered by the Bonneville Flood is shown by pattern. Embayments to the left and right of the channel, al- though inundated by the flood, are considered to have had no measurable effect on the maxi- mum discharge. The diagonally ruled area on the right bank at Mile 462 represents a basalt bench at an altilude of 2,750 ft that was scoured by the flood. The stippled area is a gravel bar in the lee of the bench.
1 MILE
1 0 1 KILOMETRE CONTOUR INTERVAL 100 FEET
Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/99/1/127/3952298/i0016-7606-99-1-127.pdf by guest on 02 October 2021 PALEODISCHARGE OF BONNEVILLE FLOOD, IDAHO 131
the step-backwater method and by analyzing the effects of modified hy- 1:24,000). Because each cross section represents only a discrete sample of draulic factors. the geometry of the reach, the number of cross sections affects the accuracy of the computed discharge. Our analysis was based on the dimensions of Theory 15 cross sections between Mile 460 and the newly discovered flood marks at Mile 462 (Fig. 4). Spacing of the cross sections was based on the For conditions of uniform flow, discharge is computed from an equa- anticipated water-surface profile. For example, near Mile 460, where the tion that involves channel characteristics, water-surface elevations in the profile was expected to be comparatively steep, the cross sections were reach, and a roughness coefficient. The decrease in water level in a uniform closely spaced. Representative cross sections are shown in Figure 5. Em- channel represents losses caused by bed roughness. Any of the variations of bayments to the left and right of the canyon were considered to have had the Chezy equation can be used, such as the Manning equation (Chow, no effect on the flood, and these areas were not used to define the dimen- 1959). sions of the flooded channel. Rather, cross sections were drawn as if The Manning equation was developed for conditions of uniform flow vertical canyon walls existed at the sites of these embayments (Fig. 4). The in which the water-surface profile and energy gradient are parallel to the exact locations of these vertical walls have a negligible effect on the results stream bed, and in which the hydraulic radius and depth are constant of the hydraulic analysis. throughout the reach. The Manning equation further is presumed to be Barnes and Davidian (1978) suggest that the length of the reach for valid for nonuniform reaches in natural channels, if the energy gradient is step-backwater analysis be at least 2 Yi times the mean depth of the stream. modified to reflect only losses due to boundary friction and losses caused This condition is satisfied, in that features used to determine the flood by channel contractions and expansions. To evaluate these losses, resist- height at Mile 462 are about 3 km upstream from Mile 460. ance coefficients and coefficients for contraction and expansion of the channel are specified. This analysis is called the "step-backwater method." Step-Backwater Analysis of the Paleodischarge Cross sections that define the channel characteristics are obtained, and the standard-step method is used to balance the Bernoulli energy The U.S. Geological Survey step-backwater computer program J635 equation between them (Chow, 1959). In this way, water-surface altitudes (Shearman, 1976; J. O. Shearman, 1977, written commun.) was used for are determined for each cross section at a specified discharge. A range of the hydraulic analysis. This program computes the water-surface altitudes discharges is used to develop a stage-discharge relation at the cross section for a specified discharge when the initial calculation is based on an as- where the flood height is known. The estimated peak discharge then is sumed depth for the water surface. Cross-section dimensions, distances determined from the stage-discharge relation and the altitude of the flood along the channel from Mile 460, roughness coefficients, contraction and marks. expansion coefficients, and a range of selected discharges were entered into the computer program. Data Assembly After an initial run of the computer program, the hydraulic effect of possible differences from present conditions, such as changes in channel Channel and flood-plain cross sections were measured perpendicular dimensions and channel roughness, were evaluated. Initially, the cross to the flow path of the Bonneville Flood from the U.S. Geological Survey sections were assumed to be representative of conditions during the flood, Sinker Butte and Wild Horse Butte topographic quadrangles (scale a Manning's roughness coefficient of 0.040 was used, and contraction and
DISTANCE FROM LEFT BANK, IN METRES
Figure 5. Selected cross sections of the Snake River Canyon between Mile 460 and Mile 462, as seen looking downstream. The computed water-surface levels ( V), mean velocities (V), and widths of the water surface (B) are shown for a discharge of 935,000 m3/s.
Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/99/1/127/3952298/i0016-7606-99-1-127.pdf by guest on 02 October 2021 132 JARRETT AND MALDE
TABLE; I. PEAK DISCHARGES FOR THE BONNEVILLE FLOOD 0.5 for expanding reaches (Shearman, 1976), as was done in Run 1, although larger values are sometimes used. For the Snake River Canyon Run Assumed hydraulic characteristics Peak Manning's Contraction/ Changes in discharge between Mile 460 and Mile 462, we used upper limits of 0.3 and 0.7, roughness expansion width (m3/s) respectively, for these coefficients. coefficient coefficients (m) Finally, the effect of a more constricted channel that was enlarged by 1 0.040 0.0/0.5 0 1,005,000 flood erosion was considered. Although the canyon is cut in basaltic 2 0.040 0.3/0.7 0 863,000 3 0.030 0.0/0.5 0 1,020,000 bedrock, some erosion of the walls and talus almost surely occurred during 4 0.050 0.0/0.5 0 991,000 5 0.040 0.0/0.5 -15 963,000 the Bonneville Flood, particularly at constrictions where flow velocities 6 0.040 0.0/0.5 -30 920,000 were highest. The amount of flood erosion can be estimated by considering 7 0.050 0.3/0.7 -30 793,000 8 0.040 0.0/0.5 -23 935,000 the volume of flood gravel in the Snake River Canyon downstream from Mile 460. Most of the gravel was deposited downstream from Mile 448, in Note: these discharges a*e calculated on the basis of various sets of assumed hydraulic characteristics and a flood altitude of 2,750 ft at Mile 462. a broad expanding reach of the canyon known as "the Walters basin" (Fig. 1), which formed a sediment trap. The volume of flood gravel in the Walters basin has been estimated at about 2.1 x 108 m3 (Malde, 1968, p. 42). As described above, some bouldery gravel also was deposited in the expansion coefficients of 0.0 and 0.5 were specified. A stage-discharge canyon between Mile 460 and Mile 448, mostly as bars as much as 90 m relation in accord with the altitude of the flood marks at Mile 462 thus was in height. The distribution of the gravel bars indicates that the volume of computed, indicating a peak discharge of 1,005,000 m3/s (Run 1; see gravel from Mile 460 to Mile 44 8 is about a tenth of the amount in the Table 1 and Fig. 6). The hydraulic conditions used in the initial calcula- Walters basin. Thus, the estimated volume of flood gravel downstream 8 3 tions then were modified, using different roughness coefficients, contrac- from Mile 460 is about 2.3 x 10 m . The gravel must have been derived tion and expansion coefficients, and changes in channel width to evaluate by flood erosion of talus and bedrock downstream from Mile 462, because the paleodischarge (Table 1). ponded water upstream prevented, significant erosion (Malde, 1968, p. 42). Manning's roughness coefficient was varied between 0.030 and The area of the canyon walls arid the canyon floor vulnerable to flood 6 0.050, on the basis of guidelines given by Chow (1959) and verified by erosion downstream from Mile 462 is calculated to have been 25 x 10 2 Barnes (1967). Equations developed by Limerinos (1970) to estimate m . Hence, the volume of gravel can be accounted for by 9 m of uniform Manning's roughness coefficient yield a range from 0.035 to 0.040 for this erosion from the walls and floor in this reach of the canyon. Because study reach. erosion presumably was greater at constrictions where flow velocities The energy loss due to contraction or expansion of a channel between would have been greatest, hydraulic analyses were made by assuming that two of its cross sections is a function of the difference in the respective the width at sections 0 through 305 was 15 m, 23 m, and 30 m less than velocity heads. The loss varies with coefficients in the energy equation. The the present width at the time of peak discharge. Changes in width at wider values of the coefficients normally used are 0.0 for contracting reaches and sections upstream have an insignificant effect on the computed discharge.
Effect on the Paleodischarge of Modified Hydraulic Factors
The following results compare the paleodischarge obtained under the initial conditions (Run 1) with discharges computed by assuming different values of the hydraulic variables. The range of paleodischarge using these assumptions was from 793,000 m3/s to 1,020,000 m3/s (Runs 7 and 3; Table 1 and Fig. 6). Different values of Manning's roughness coefficient had the least effect on the computed discharge, primarily because of the influence of ponded water upstream from Mile 462. Compared to a roughness coeffi- cient of 0.040 (Run 1), a difference of 25% (0.030 and 0.050) changed the discharge only 1.5% (Runs 3 and 4). This finding is consistent with pre- vious experience that the step-backwater method is fairly insensitive to flow resistance for large depths of flow. Assumed changes in canyon width at sections 0 through 305 had a moderate effect on the calculated discharge. For a decrease of 30 m, which is a reasonable upper limit of erosion, peak discharge decreased about 8% (Run 6). Changing the contraction and expansion coefficients had the greatest effect on the calculated discharge, because of the variable width of the canyon above Mile 460. When very conservative values of 0.3 and 0.7 were assumed for the contraction and expansion coefficients, the discharge Figure 6. Stage-discharge relations at section 3020 for selected was decreased 14% (Run 2). hydraulic conditions. Run numbers refer to the numbered runs in In light of the effects of modifying the hydraulic variables, we judge Table 1. Altitudes are indicated for the flood marks used in the present that the most likely conditions for the Bonneville Flood are: Manning's calculation (2,750 ft, or 838 m) and for the flood height used in the roughness coefficient of 0.040, contraction and expansion coefficients of previous calculation (2,625 ft, or 800 m). 0.0 and 0.5, and a width 23 m less than the present width. The peak
Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/99/1/127/3952298/i0016-7606-99-1-127.pdf by guest on 02 October 2021 PALEODISCHARGE OF BONNEVILLE FLOOD, IDAHO 133
MILE 460 MILE 462 850
CO LU Altitude of Upstream Flood Deposits, CC 2700 Figure 7. Water-surface I— 2750 ft (838 m) profile and altitude of the ob- 800- served flood deposits of the 2600 Altitude of Downstream Water Bonneville Flood at its high- Surface, 2575 ft (785 m) est probable stage in the con- 2500 stricted canyon between Mile O D 750- 460 and Mile 462 (Run 8). I- 2400
Altitude of Riverbed, 2312 ft (705 m) 700 2300 500 1000 1500 2000 2500 3000 3500
DISTANCE UPSTREAM FROM MILE 460, IN METRES
discharge under these conditions was computed to be 935,000 m3/s (Run been as large as 21,300,000 m3/s (Baker, 1973), although this estimate 8). The computed water-surface profile for this discharge and the altitude probably is 30% too large (Baker and Bunker, 1985, p. 12). Independent of the observed flood deposits are shown on Figure 7. The profile shows paleohydrologic calculations (Clarke and others, 1984) and observations that the water level at Mile 460 decreased about 53 m, resulting in an of historic floods from ice-dammed lakes (Beget, 1986) suggest that altitude of 2,575 ft for the water surface just downstream. Under these Baker's estimate may be too large, but peak discharge from glacial Lake conditions, average flow velocities ranged from 26 m/s at Mile 460, to 9.5 Missoula undoubtedly represents the largest known flood in the world. m/s at midreach, and 13 m/s at Mile 462 (Fig. 5). The paleodischarge The Bonneville Flood, however, was the second largest. would have increased moderately with increasing depth of flow (Fig. 6). For the conditions used in Run 8, water-surface profiles higher than the GEOMORPHIC SIGNIFICANCE OF THE FLOOD flood marks at Mile 462 increase the computed discharge as follows: 1,000,000 m3/s at 5 m higher; 1,076,000 m3/s at 10 m; and 1,161,000 Wolman and Miller (1960) proposed that the relative geomorphic m3/s at 15 m. importance of extreme catastrophic events, or of more frequent smaller Clifford T. Jenkins (in Malde, 1968, p. 11-12), by assuming critical events, is indicated by the relative amounts of work done on the landscape. flow at Mile 460, and by using an altitude of 2,625 ft for the water The work done by streams can be expressed in terms of shear stress and surface—not the altitude of 2,750 ft described here—computed that the unit stream power, which indicate the potential rate of movement of Bonneville Flood discharge ranged from 346,000 m3/s to 462,000 m3/s. sediment. Shear stress is the product of the hydraulic radius, the energy He suggested that the most likely paleodischarge was 425,000 m3/s. Jen- slope, and the specific weight of the water. Thus, large values for the kins assumed that the dimensions of the canyon at the time of peak hydraulic radius or the energy slope indicate large amounts of shear stress. discharge were the same as at present, and he used a Manning's roughness Unit stream power is the product of the discharge, the energy slope, and coefficient of 0.040. For these conditions, the step-backwater method the specific weight of the water, divided by the water-surface width. yields a peak discharge of 503,000 m3/s (Fig. 6, Run 1). Because the For the Bonneville Flood at Mile 460, the hydraulic radius was 63 m, hydraulic assumptions made by Jenkins are incompletely known, reasons the energy slope was 0.0041, and the width was 500 m. A specific weight for the differences between his results and ours are uncertain. of 9,800 N/m3 for the water is assumed. From the calculated discharge of 935,000 m3/s, the computed shear stress was 2,500 N/m2, and the unit COMPARISON WITH OTHER FLOODS stream power was 75,000 N/m/s. For comparison, the Mississippi River and the Amazon River during floods have shear stress values of 6 to 10 The Bonneville Flood far exceeded the magnitude of historic floods N/m2 and unit stream powers of about 12 N/m/s (Costa, 1987). In the on the Snake River and on other rivers in the world. The maximum United States, some of the greatest known values of shear stress and unit recorded discharge on the Snake River near Mile 460 was 1,340 m3/s, stream power are for small basins. For example, Costa (1987) reported measured at a streamflow-gaging station 10.3 km downstream. The great- values of 858 N/m2 and 8,160 N/m/s for the 1973 flood on a tributary of est flood in recent times in the United States was the 1927 flood on the the Humboldt River, Nevada. Mississippi River, which had a peak discharge of 70,010 m3/s (Dalrym- As streams increase in size, their slopes typically decrease. Because ple, 1964). The largest known recent flood in the world occurred on the the slope decreases faster than does the hydraulic radius, large rivers are Amazon River during 1953, when the peak discharge was 385,000 m3/s characterized by small values of shear stress and unit stream power. The (Oilman, 1968). large values of shear stress and unit stream power for the Bonneville Flood Although the Bonneville Flood greatly exceeded the magnitude of at Mile 460, despite its size, are due partly to the large depth of flow, but any historic flood, it pales when compared with floods that resulted from mostly to the steep slope of the Snake River. The Snake River is one of the rapid outbreaks of glacial Lake Missoula, -14,000 yr ago. Using the steepest large rivers in the United States (Malde, 1968), and this factor slope-area method and a Manning's roughness coefficient of 0.040, peak would have helped to maintain voluminous flow, even at a considerable discharge at the breakout point for Lake Missoula is estimated to have distance downstream from Red Rock Pass.
Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/99/1/127/3952298/i0016-7606-99-1-127.pdf by guest on 02 October 2021 134 JARRETT AND MALDE
NEW EVIDENCE OF FLOOD VOLUME AND DURATION ACKNOWLEDGMENTS
The volume of water discharged during the Bonneville Flood equals We thank V. R. Baker, J. O. Shearman, and J. E. Costa for helpful the volume of water between the Bonneville and Provo shorelines, accord- insight and discussions about analyzing the hydraulic conditions of cata- ing to current understanding of the history of Lake Bonneville (Scott and strophic floods. A letter from D. R. Currey provided valuable perspective others, 1983; Currey and Oviatt, 1985). A previous estimate of the volume about the geologic history of Red Rock Pass. The manuscript was re- was attributed to downcutting only to the Bonneville shoreline, from a viewed by W. E. Scott and C. H. Swift. supposed level about 30 m higher (Malde, 1968, p. 12). The estimated
volume was therefore too small. REFERENCES CITED
Currey and Oviatt (1985, p. 13) reported that erosion of Red Rock Bailey, J. F„ and Ray, H. A., 1966, Definition of staff -discharge relation in natural channels by step-backwater analysis: Pass during the Bonneville Flood lowered the lake 108 m, from the U.S. Geological Survey Water-Supply Paper 1369-A, 24 p. Baker, V. R., 1973, Paleohydrology and sedimentology of Lake Missoula flooding in eastern Washington: Geological Bonneville shoreline (altitude 1,552 m) to the Provo shoreline (altitude Society of America Special paper 144, 79 p. Baker, V. R., and Bunker, R. C., 1985, Cataclysmic late Pleistocene flooding from glacial Lake Missoula; a review: 1,444 m). Because the area of Lake Bonneville was directly related to the Quaternary Science Reviews, v. 4, no. 1, p. 1-41. altitude of its surface, when the altitudes are adjusted for hydro-isostatic Barnes, H. H., Jr., 1967, Roughness characteristics of natural channels: U.S. Geological Survey Water-Supply Paper 1849, 213 p. rebound (Currey and Oviatt, 1985, p. 10), the volume released by down- Barnes, H. H., Jr., and Davidian, J., 1978, Indirect methods, in Herschy, R. W„ ed., Hydrometry: Principles and practices: New York, Wiley, p. 149-204. cutting of the outlet can be calculated as the frustum of a cone, in which Beget, J. E., 1986, Comment on 'Outburst floods from glacial Lake Missoula' by G.K.C. Clarke, W, H. Mathews, and the base is the area at the Provo shoreline (36,720 km2), the top is the area R. T. Pack: Quaternary Research, v. 25, no. I, p. 136-138. 2 Chow, V. T„ 1959, Open channel hydraulics: New York, McGraw-Hill, 680 p. at the Bonneville shoreline (51,530 km ), and the height is the difference Clarke, G.K.C., Mathews, W. H., and Pack, R. T., 1984, Outburst floods from glacial Lake Missoula: Quaternary Research, v. 22, no. 3, p. 289-299. in altitude between the shorelines (108 m). The volume of the Bonneville Costa, J. E„ 1987, Interpretation of the largest rainfall-runoff floods measured by indirect methods on small drainage Flood therefore was 4,700 km3. If outflow from Lake Bonneville is as- basins in the conterminous United States: Proceedings of the U-S.-China Bilateral Symposium on the Analysis of 3 Extraordinary Flood Events, October 1985, Nanjing, China (in press). sumed to have been at a constant discharge of 935,000 m /s, the min- Currey, D. R., and Oviatt, C. G., 1985, Durations, average rates, and probable causes of Lake Bonneville expansions, stillstands, and contractions during the last deep-lake cycle, 32,000 to 10,000 years ago, in Kay, P. A., and Diaz, imum flood duration would have been about 8 weeks. The flood H. F., eds., Problems of and prospects for predicting Great Salt Lake levels; papers from a conference held in Salt duration, however, surely was longer. Peak discharge would have been Lake City, March 26-28,1985: Salt Lake City, Utah, University of Utah, Center for Public Affairs and Adminis- tration, p. 9-24, achieved only after the outlet at Red Rock Pass had become enlarged by Dalrymple, T., 1964, Rood characteristics and flow determination, r>i Chow, V. T., ed.. Handbook of applied hydrology: New York, McGraw-Hill, p. 25-1-25-33. erosion, and the discharge probably waned as the level of Lake Bonneville Davidian, J., 1984, Computation of water-surface profiles in open channels: U.S. Geological Survey Techniques of decreased. An accurate estimate of the duration, however, would require Water-Resources Investigations, Book 3, chap. MS, 48 p. Druse, S. A., 1982, Verification of step-backwater confutations on ephemeral streams in northeastern Wyoming: U.S. dam-break modeling, as discussed in the following section. Geological Survey Water-Supply Paper 2199, 12 p. Fread, D. L„ 1977, The development and testing of a dam-break flood forecasting model, in Dam-break flood routing model workshop proceedings, Bethesda, Mary and: U.S. Water-Resource Council, Task Force on Dam-Break Rood Routing, p. 164-197. DISCUSSION Gilbert, G. K., 1878, The ancient outlet of Great Salt Lake: American Journal of Science, 3d ser., v. 15, p. 256-259, Hunt, c. B., 1980, G. K. Gilbert's Lake Bonneville studies, in Yochelson, E. L„ ed., The scientific ideas of G. K. Gilbert: Geological Society of America Special Paper li 3, p. 45-59. The constricted canyon between Mile 460 and Mile 462 is just one Jarrett, R. D., and Costa, J. E., 1985, Hydrology, geomorphology, and dam-break modeling of the July 15, 1982, Lawn Lake Dam and Cascade Lake Dam failures, Lar mer County, Colorado: U.S. Geological Survey Open-File Report place along the path of the Bonneville Flood, and it would be of geologic 84-612, 109 p. Land, L. F., 1980, Evaluation of selected dam-break flood-wave models by using field data: U.S. Geological Survey and hydrologic interest to evaluate the hydraulic characteristics of the Water-Resources Investigations 80-44, 54 p. flood throughout its length. Such an analysis needs to begin at Red Rock Limerinos, J. T., 1970, Determination of the Manning coefficient from measured bed roughness in natural channels: U.S. Geological Survey Water-Supply Paper 1898-B. 47 p. Pass, because outflow of Lake Bonneville from this breach would have Malde, H. E., 1968. The catastrophic late Pleistocene Bonneville Rood in the Snake River Plain, Idaho: U.S. Geological Survey Professional Paper 596, 52 p. been controlled by the rate of failure and by changes in the breach as it Oilman, R. E., 1968, Reconnaissance investigations :>f the discharge and water quality of the Amazon River: U.S. developed (Jarrett and Costa, 1985). Some aspects of the history of Red Geological Survey Circular 552, 16 p. Potter, K. W., and Walker, J. F., 1987, Case studies of Hood measurement error: Proceedings of the U.S.-China Bilateral Rock Pass are known (Sewell and Shroder, 1981), but further study of the Symposium on the Analysis of Extraordinary Rood Events, October 1985, Nanjing, China (in press). Scott, W. E., McCoy, W. D., Shroba, R. R., and Rubin, M., 1983, Reinterpretation of the exposed record of the last two pass is needed to obtain the chronologic and topographic data needed to cycles of Lake Bonneville, western United States: Quaternary Research, v. 20. no. 3, p. 261-285. model its failure (D. R. Currey, University of Utah, 1985, written com- Sewell, R. E., Jr., and Shroder, J. F.. Jr., 1981, Red Rock landslip; massive failure initiated by late Pleistocene Bonneville Rood: Geological Society of America Abstracts with Programs, v. 13, no. 4, p. 225. mun.). From the dam-break model, the following hydraulic characteristics Shearman, J. O., 1976, Computer applications for sxp-backwater and lioodway analyses: U.S. Geological Survey Open-File Report 76-499, 103 p. could be computed: the duration of flooding, the traveltime of the flood Williams, T. R., and Bedinger, M. S„ 1984, Selected geologic and hydrologic characteristics of the Basin and Range wave, and flood hydrographs at Red Rock Pass and along the Snake River province, western United States: Pleistocene lakes and marshes: U.S. Geological Survey Miscellaneous Investiga- tions Series Map 1-1522-D, scale 1:2,500,000. (Fread, 1977; Land, 1980; Jarrett and Costa, 1985). A sensitivity analysis, Wolman, M. G., and Miller, J. P., 1960, Magnitude and frequency offerees in geomorphic processes: Journal of Geology, for which one component would be the estimated paleodischarge between v. 68, no. 1, p. 54-74. Mile 460 and Mile 462, then would provide reasonable limits for the flow characteristics along the flood path. Such an analysis could be tested by
considering the nearly continuous geologic record of the Bonneville Flood MANUSCRIPT RECEIVED BY THF. SOCIETY SEPTEMBER 8, 1986 REVISED MANUSCRIPT RECEIVED FEBRUARY 12, 1987 throughout its length. MANUSCRIPT ACCEPTED FEBRUARY 12, 1987
Printed in U.S.A.
Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/99/1/127/3952298/i0016-7606-99-1-127.pdf by guest on 02 October 2021