EVALUATING THE EFFECT OF WEPP PREDICTIONS OF RUNOFF AND SOIL LOSSES IN THE LOESS PLATEAU REGION OF CHINA

Abstract

In this paper, we evaluated the Water Erosion Prediction Project (WEPP) model using 5-yr monitored runoff and sediment data from three different lands cover treatments in the Loess Plateau of China. The results showed that the WEPP model was able to predict 22 out of 37 recorded runoff producing events, and 20 out of 34 recorded sediment producing events. Prediction of annual runoff is more accurate than the prediction of sediment. The correlation coefficients between the annual monitored and simulated runoff ranged from 0.69 to 0.91, and 0.11 to 0.90 for sediment; for the three treatments (bare, corn and grass), the prediction for the grass plot is better than the other two. Soil moisture and storm intensity appeared to have greater effect on both predictability and accuracy of runoff events and sediment yields. For grass plot, Nash-Sutcliffe coefficient was 0.76 in runoff and 0.56 in sediment. WEPP predictions for the grass plot were in agreement with measured values, but no better than using the averaged measured values for the other two treatments (bare and corn plots). Application of the WEPP model should be decided with precautions in the Loess Plateau area