Statistical distributions of soil loss from runoff plots and WEPP model simulations

Soil conservation measures may be better designed with the knowledge of daily distributions of erosion. Collecting reliable data to determine daily erosion distributions is costly; however, new process-based soil erosion models have the potential to simulate extended records. The objectives of this study were to analyze frequency distributions of measured daily soil loss values and to determine if the Water Erosion Prediction Project (WEPP) model accurately reproduced statistical distributions of the measured daily erosion series. A Log-Pearson Type III (LP III) distribution was fitted to measured and WEPP-predicted soil loss values from six sites for periods ranging from 6 to 10 yr. Cumulative soil loss as a function of storm recurrence interval was used to show the relative contributions of large and small storms to total soil loss at each site. Results showed that both measured and predicted frequency curves fell within the 95% confidence range of the LP III distributions. This was true both using weather data from the site for the period of monitoring as well as when using synthetic weather series generated with the CLIGEN model. Thus the results were encouraging in terms of using WEPP in conjunction with CLIGEN to generate long term daily soil loss frequency distributions, which can contribute toward alleviating the problem of having only a short monitoring period for measured erosion. Cumulative soil loss results indicated that large storms contributed a major portion of the erosion under conditions where cover was high, but not necessarily under conditions of low cover.