|
<br />.'
<br />
<br />Study Area
<br />
<br />The community of Buffalo Creek is located in the foothills of the Colorado Rocky Mountains of .
<br />unincorporated Jefferson County within Pike National Forest near the confluence of Buffalo Creek
<br />and North Fork (NF) South Platte River (fig. 1). The community, at an elevation of about 2 012 m
<br />consists of several hundred homes within a montane forest (predominantly lodgepole and' ,
<br />ponderosa pine, do~glas fir, and aspen). Topography is rugged and soils are shallow,
<br />moderately well drained, and composed of coarse sandy gravel (Colorado Water ConseNation
<br />Board, 1997). Bedrock on average is about 1 m below land surface, but is exposed on many
<br />steeper hillslopes. Accumulation of organic litter (duff) on unburned forested areas averages
<br />about 60 mrn. The climate is semiarid and mean annual precipitation is about 400 mm, much of
<br />which falls as snow from about October through April. Many streams are ephemeral in the study
<br />area and flow into Buffalo Creek and the NF South Platte and South Platte Rivers, which primarily
<br />are fed by melting snowpack and trans-basin flow diversions. Flood flows in the vicinity of
<br />Buffalo Creek can result from intense, localized thunderstorms, generalized rainstorms, and spring
<br />snowmelt (Jarrett, 1990). Long-time residents reported no significant flooding in the Buffalo Creek
<br />area in at least 70 years. The 1996 wildfire, driven by strong winds, burned most vegetation and
<br />produced hydrophobic soils in much of the bumed area, making the area more susceptible to
<br />flooding (Colorado Water ConseNation Board, 1997).
<br />
<br />Methods
<br />
<br />No systematic precipitation, streamflow, and sediment monitoring networks existed in Buffalo
<br />Creek prior to the wildfire. Therefore, data collection consisted of determining rainfall, peak flow,
<br />and sedimentologic data for most streams in the Buffalo Creek area. Rainfall-bucket survey data
<br />were obtained throughout the Buffalo Creek area for each rainstorm. In addition, paleoflood and
<br />sedimentologic data were used to estimate rainlall amounts and spatial variability (Jarrett, 1990.
<br />In 1997, four recording precipitation and three streamflow gages were placed in the Buffalo,
<br />Morrison, and Spring Creek basins by the USGS in cooperation with the Denver Water
<br />Department (John Moody, USGS, written commun., 1997). In early 1997, the NWS provided 13
<br />Buffalo Creek residents with non-recording precipitation gages to assess the spatial variability of
<br />rainfall, monitor flood potential, and assess/refine rainfall estimates from Doppler radar (NWS-
<br />WSR-88). Because few gages were located in the burned area and very localized nature of
<br />convective storms, basin-wide monitoring continued for each runoff event.
<br />
<br />Paleoflood hydrology is the study of flood-transported sediments and botanic information from
<br />past lIoods preserved in stream channels and is particularly useful in providing hydrometeorologic
<br />data for ungaged basins (Jarrett, 1990, 1991). Most studies involve prehistoric floods, but the
<br />methodology also is applicable to studying modern floods. Rainfall amounts on sparsely
<br />vegetated hillslopes (burned or unburned) can be inferred from the amount of hillslope erosion
<br />(rills, gullies, and headcuts), maximum size of sediments transported, and depositional
<br />characteristics. The paleoflood rainfall estimates can be compared with other sources of rainfall
<br />data (gaged, bucket sUNey, and radar) or they can provide rainfall data when no other source
<br />exists.
<br />
<br />Because of the hazards and flashy nature of floods in burned areas, peak discharges were
<br />estimated using the slope-conveyance method (Barnes and Davidian, 1978) for streams in the
<br />Buffalo Creek area after each rainstorm. Sites were selected primarily where bedrock is exposed
<br />across the channel. A few sites were located in alluvial channels; reaches selected were
<br />relatively straight and uniform, which reduces potential errors due to channel change. In such
<br />reaches, net change in total-lIow area probably are small for one flood, although sediment loads
<br />may be large. Cross section data were collected for channels in burned and adjacent unburned
<br />basins. Monitored basins range from about 0.1 km2 to the total bumed area of about 50 km2. The
<br />burned area is located just upstream from the South Platte River at South Platte streamllow-
<br />gaging station (06707500) shown on figure 1, which has a total drainage area of 6,680 km2 The
<br />flood of record is 53 m3 /s, which resulted from snowmelt, since the gage was installed in 1904.
<br />Peak-discharge data were estimated for 75 sub-basins in the study area having different basin
<br />characteristics such as vegetation cover, bum intensity, watershed aspect and slope, sediment
<br />
<br />Draft 3/30/98
<br />
<br />3
<br />
|