Submitted to: Transactions of the ASAE
Publication Type: Peer reviewed journal
Publication Acceptance Date: 6/3/1997
Publication Date: N/A
Citation: 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. In order to use the model, one must divide a land area into units of hillslopes and channels. That is called ¿watershed discretization¿. This study was made to evaluate how the discretization should be done. In this study we also analyzed the model¿s calculations of erosion to determine which pieces of information about the land area are most important to measure. These results help users decide how to spend limited time and money in collecting information for the land area to obtain the best possible soil erosion information.
Technical Abstract: The Water Erosion Prediction Project (WEPP) model is a continuous simulation tool for predicting erosion and deposition due to rainfall, snowmelt and irrigation. It can be used on a hillslope or on a watershed defined by several hillslopes and impoundments connected by channels. To verify that the watershed model behaves rationally and consistently over a wide range of input we analyzed the sensitivity of the model to watershed discretization and channel input parameters. Effects of watershed discretization were evaluated for selected events within a one- year continuous simulation by comparing results for two watersheds under various discretization schemes. Impacts of channel input parameters were assessed by comparing the value of a linear sensitivity coefficient for all user-specified parameters (slope, Manning coefficient...) and for the same selected events as well as for annual averages. We found the hillslope length to be a key parameter in the prediction of watershed sediment yields. Estimates of total sediment yield at the watershed outlet were highly sensitive to the total Manning coefficient and the Manning coefficient for bare soil as well as the channel slope. They were less sensitive to the channel bank slope. Erodibility and critical shear stress were found to be important for events causing channel erosion. The results were sensitive to the hydraulic conductivity for events with small runoff and small sediment contributions from hillslopes. Results provide guidance to the model users for the discretization of the watershed and for the estimation of channel parameters.