FLOOD01775 - Laserfiche WebLink

Home Browse Search

flood damages. Sediments moved on hillslopes ranged from silt to cobble-sized material, and 2.5- m diameter boulders were transported in some channels (fig. 8). The distinct black, bum boundary on undisturbed trees and rocks in channels (fig. 9) were used to define pre-fire channel- bed levels and estimate the amount of channel change. Many tributaries degraded during the July 12th flood (fig. 9), although locally agradation (1-2 m) occurred along Buffalo, Spring, and Shinglemill Creeks and on alluvial fans at tributary junctions in burned areas. Many new fans had dimensions of about 100 m x 30 m x 1.5 m (fig. 7). September 1996 reservoir releases increased flows in the NF South Platte River that were capable of transporting sediments from the Buffalo Creek fan to South Platte, which produced about a meter of agradation in pools. For many burned basins, the location of exposed bedrock or firm ground used to estimate discharge changed for each storm (within +/- -100 m of other sites) due to shifting channels. Therefore, a fixed streamflow-gaging station may not produce reliable records without the use of a costly control (e.g., weir, flume) or may need to be moved. Flood-transported sediments and debris in Spring Creek dammed the South Platte River to a depth of about 3 m, which caused backwater (ponding) and reduced the peak flow from Spring Creek. The peak flow was 240 mats (+/- 20%) in the South Platte River downstream from Spring Creek (fig. 1, site 5), about 270 m3/s less than the Spring Creek peak flow. Without this ponding and attenuation of the Spring Creek flood, the peak flow at the South Platte River gage could have been as large as 595 m3/s. - ----- Effects of the Buffalo Creek Wildfire Two approaches were used to estimate the effect of the fire on flood runoff. First, July 12, 1996 flood data for severely burned and unburned basins, which had similar rainfall amounts, were plotted against contributing drainage area (fig. 10). Peak discharge from severely burned basins was 20-40 times larger than for unburned basins. Unburned basins within areas of maximum rainfall had minimal runoff, which likely reflects rainfall interception by the duff in unburned areas; in addition, basin slopes in unburned areas generally are 20 percent or less. Slightly burned areas, which might have similar runoff as prescribed-bum watersheds, had substantial flood and sediment runoff, but less than moderately- and severely-burned basins. A number of severely- burned basins in areas near maximum rainfall had unit discharges (peak discharge divided by drainage area) of about 60 m3/s/km2; the maximum unit discharge is about 40 m3/slkm2 for all previous Colorado floods (Jarrett, 1990). Since the fire, rainstorms have produced 9 floods (5 in 1996 and 4 in 1997) larger than the estimated 1 OO-year (pre-fire) flood (FEMA, 1986); most storms were preceded by 5 to 10 mm of rainfall. Continued flooding from small rainstorms and having similar runoff as in 1996 indicate that little watershed recovery has occurred by the fall of 1997. Flood-frequency relations for Sand Draw and the South Platte River at South Platte with corresponding July 12, 1996, peak discharges are shown on figure 11. Trans-basin diversions were assumed to be negligible on large flood peaks and hence the upper end of the frequency relation for the South Platte River. On July 12, 1996, peak discharges in small, burned basins such as Sand Draw, Spring Creek, Spring Gulch, and Shinglemill Creek were about 10 times larger than the 1 OO-year flood. On Sand Draw, and other small tributaries in burned basins near the center of the storm, peak discharges exceeded the 1,OOO-yr flood. The second approach compared flood data for burned areas in Buffalo Creek with other Colorado Front Range foothill extreme floods resulting from about 150 to 175 mm of rain in about an hour in similar basins (slopes, soils, vegetation) that were unburned (fig. 10). Maximum rainfall on July 12, 1996 for Buffalo Creek in burned areas is similar to large rainfall amounts for unburned basins in Colorado. Maximum peak discharges for the July 12, 1996 rainstorm produced flood peaks about 1.5 times larger than floods in unburned basins elsewhere in the Colorado Front Range (fig. 10). A number of severely-burned basins in areas near maximum rainfall had unit discharges (peak discharge divided by drainage area) of about 60 m3/slkm2; the maximum unit discharge is about 40 m3/slkm2 for all previous Colorado floods (Jarrett, 1990). The combined peak discharge for Buffalo Creek and Spring Creek is about 960 m3/s (-50 km2 contributing burned area), which is greater than the 1976 flood of 883 m3/s in the Big Thompson River (McCain et aI., 1979) from a contributing area of about 250 km2. Clearly, the wildfire had a major role in the severity of flooding Draft 3/30/98 6