Submitted to: Transactions of the ASAE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: June 3, 1997
Publication Date: N/A
Interpretive Summary: The Water Erosion Prediction Project (WEPP) model is used to predict erosion on hillslopes and small watersheds. It is used to help land users understand and evaluate the impacts of land management practices on soil loss and sediment yields from their land. It is also used by scientists and others to inventory the amount of erosion which is occurring across agricultural regions, which provides information for developing national and regional soil conservation policy. This study was undertaken to assess the accuracy of the WEPP model calculations of surface runoff and soil erosion on small agricultural watersheds. Fifteen such watersheds were selected from 6 locations around the United States, including sites in Georgia, Texas, Oklahoma, Ohio, and Mississippi. The study showed that the model performed well within the expected accuracy of natural resource models, and it pointed out some possibilities for model improvement for specific applications.
Technical Abstract: The Water Erosion Prediction Project (WEPP) watershed scale model was developed by the USDA for purposes of erosion assessment and conservation planning. Several sensitivity analyses and evaluation studies have been conducted for the WEPP hillslope version, but few efforts have been made to evaluate the accuracy of WEPP watershed model predictions by comparing predicted results to measured data. The purpose of this study was to evaluate WEPP watershed model applicability and prediction accuracy for small watersheds under different climates, topographies, soils, and managements. Data from 15 watersheds from 6 locations in the United States were compared to runoff and sediment yields estimates using WEPP95. The result shows that the WEPP predicted reasonable average annual and event runoff and sediment yield with default parameter estimation procedures for most of the sites and individual watersheds. Model efficiencies for storm by storm runoff and sediment yield were quite reasonable for general conditions. The distributions of measured and predicted events were similar. This study also pointed out the potential areas for model improvement, including better and more specific silage routines, addition of a weed component to the plant growth and management model, and better linkage of the sediment from contoured hillslopes to the watershed channel system.