QUANTITATIVE ANALYSIS OF CHANNEL EROSION BY DEBRIS FLOWS, CENTRAL COAST RANGE, OREGON

In the central Oregon Coast Range, nearly all drainage basins periodically generate debris flows, and the flows often impact roads or structures at the mouths of basins. During 1996, at least two major storms produced widespread debris flows in the Coast Range. Two of these flows caused a total of 5 fatalities. Debris flows in the region typically mobilize from small slides and then increase in volume by erosion and entrainment of channel sediment. Estimates of volumes of sediment eroded and entrained, normalized either by basin area or channel length, are critical for accurately predicting debris-flow volumes expected at the mouths of drainage basins.

We measured erosion by 1996 debris flows in four clear-cut, first-order basins in the central Coast Range near Scottsburg: one basin in the Charlotte Creek drainage and three basins in the Mill Creek drainage. Basins ranged in size from 68,300 to 209,410 m². Hillside and channel slopes within the basins were greater than 10º. For each basin, we used 1:12,000-scale aerial photos taken in July 1992 and May 1997 to construct pre- and post-debris flow Digital Elevation Models (DEMs) with 2 m cell spacing in an analytical stereo plotter. Difference DEMs (1997 minus 1992) provided a quantitative measure of erosion and entrainment by debris-flows. All channels were eroded to bedrock in each of the basins. In the Charlotte Creek basin, a total of about 25,820 m³ of sediment was removed by debris flows. This removal resulted in an average basin lowering rate of 0.12 m³/m² and an average channel entrainment rate of 12.30 m³/m. In the Mill Creek basins, average basin lowering rates ranged from 0.12 to 0.20 m³/m² and average channel entrainment rates ranged from 17.38 to 26.24 m³/m.

We conducted a preliminary analysis of the magnitude of erosion with respect to two topographic variables (slope and total slope curvature) from the pre-flow DEMs, but found that neither topographic variable, either individually or combined, was an accurate predictor of erosion. Because all channels were eroded to bedrock, the best predictor of erosion was the depth of channel sediment prior to the debris flows. Combining our 1992 and 1997 DEM data with recently acquired bare-earth airborne LiDAR data may provide new opportunities for testing predictor variables and estimating channel refilling rates.