Measurement and Modeling of Erosion from Four Rural Unpaved Road Segments in the Stillwater Creek Watershed

Abstract

The purpose of this study was to measure sediment yield from four rural unpaved road segments in the Stillwater Creek Watershed in central Oklahoma and compare those measurements to predictions generated by the Water Erosion Prediction Project (WEPP) erosion model using four different sets of erodibility parameters. The four road segments ranged from 160-250 m in length, were crowned, and had bar ditches on both sides that drained directly into streams. Sediment traps were connected to each bar ditch and consisted of a settling trough, an H-flume and pressure transducer to measure discharge and a pumping sampler. A data logger controlled data retrieval and storage. Each sediment trap collected erosion from one half of the road area and the associated bar ditch and cut slope. Data from 26 storms were collected during June-November, 2004. The total precipitation from individual storms ranged from 3 mm to 56 mm. Maximum five-minute storm intensities ranged from 3 mm/hr to 100 mm/hr. The total erosion for individual storms from the four segments ranged from 1 kg to 3,230 kg, with a mean of 370 kg across all four segments The cumulative total erosion from the segments through the study period was 5,340 kg for 19 NE, 5,900 kg for 32 NW, 6,880 kg for 32 NE, and 14,250 kg for 19 NW. Assuming that all of the 479 km of rural unpaved roads in the Stillwater Creek watershed eroded at the observed mean rate of 135,300 kg/ha, the total estimated quantity of sediment eroded for the 22 storms from rural unpaved roads in the watershed was 152 Mg/km/yr. Using the selected uncalibrated parameter sets, WEPP predicted overall total erosion ranging from -49 to -31 % of the observed total erosion; and overall total flow ranging from -19 to -8 % of the observed total flow.