2005 Salt Lake City Annual Meeting (October 16–19, 2005)

Paper No. 8
Presentation Time: 10:15 AM

PREDICTING 50-YEAR SEDIMENT VOLUMES WITH EXTREME-VALUE STATISTICS FOR SELECTED LOS ANGELES COUNTY DEBRIS BASINS


KEATON, Jeffrey R., MACTEC, 200 Citadel Drive, Los Angeles, CA 90040, jrkeaton@mactec.com

Los Angeles County Department of Public Works maintains 115 debris basins dating from 1927 to 1989. Fifty basins along the south margin of the San Gabriel Mountains were selected for study because uniform granitic and metamorphic rocks exist in the watersheds and some of the oldest basins are located there. Annual sediment-volume data from the basins were analyzed using extreme-value (Gumbel) statistics. These basins have south- to southwest-facing watersheds ranging in size from 0.1 to 1.7 mi2, with 12 to 62 years of sediment-accumulation history (up to 2000; data for the past 5 years, including the 2004-2005 storm season, were not available). Storms in 1934, 1938, and 1955 produced extreme volumes of sediment, and the 1934 storm fell in recently burned watersheds. Sediment accumulation occurred in 18% to 57% of the years of history for these basins. No trend in accumulated sediment volume can be discerned as a function of watershed area using conventional regression, suggesting that sediment-volume-rating curves may have limited applicability for these basins. Sediment volumes were predicted for each of the 50 basins using extreme-value statistics. The design debris-basin volume is defined by Los Angeles County as the quantity of sediment produced as a result of a 50-year, 24-hour storm in a typical watershed which has significantly recovered from the impact of a burn and where sufficient antecedent rainfall has produced soil saturation. Extreme-value power-law regression of the predicted 50-year sediment volumes produces Vs50 = 75552 Aw1.1099, where Vs50 is the predicted 50-year sediment volume in yd3, and Aw is watershed area in mi2 (n = 50; r2 = 0.4519). Basic units used by Los Angeles County (yd3 and mi2) were used in the statistical analysis. Predicted sediment volumes based on recorded sediment accumulation show that design debris volumes are reasonable for most debris basins. Predicted 50-year sediment volumes are less than design volumes for 43 of the 50 basins, and less than 50% of design volumes for 26 basins. Predicted 50-year sediment volumes exceed 110% of design for only 3 basins. Debate continues regarding the applicability of extreme-value statistics for this type of data and analysis.