Evaluation of Erosion Prediction Models for Forest Roads

Abstract

Forest roads can be a source of accelerated erosion, which can be detrimental to aquatic habitat, fish, and other aquatic biota. Erosion models are increasingly used to quantify sediment production from forest roads. This project evaluated the efficacy of these models to predict erosion from forest roads. Sediment production was measured from 44 road segments in the humid, temperate rain forests of Oregon and California. Sediment production from these road segments was estimated with four contemporary erosion models: the Washington Road Surface Erosion Model (WARSEM); Sediment Model 2 (SEDMODL2); WEPP:Road, an interface for the Water Erosion Prediction Project Model; and the revised universal soil loss equation (RUSLE). The erosion models consistently overestimated the amount of sediment produced by the road segments by 2 to 8 times. The results were highly variable, and there were considerable differences in erosion estimated by the four models, even for the same road segment. Further, the erosion models could not consistently identify the road segments that were the top sediment producers. It is hypothesized that the regionalized parameters used as inputs for the models do not adequately characterize the hydrology of the individual road segments. In the humid, temperate rain forests of the Pacific Northwest, surface erosion from forest roads is best predicted by the amount of runoff from the road during storms. Thus, research that will better quantify the hydrology of forest roads will provide better information to predict surface erosion from forest roads.