Sensitivity of soil loss equations to slope classifications

Abstract

Water erosion is a major cause of watershed degradation. Empirical erosion models are often used to estimate the soil erosion and sediment yield. Revised universal soil loss equation (RUSLE) predicts the annual average soil erosion. Modified universal soil loss equation (MUSLE) predicts sediment yield on a storm event base. The land attributes such as topographic features, soil types, vegetation covers and climatic factors have direct effects on soil loss. Slope gradient and slope length are two most important topographic features that affect soil loss. Classifying slope gradient with different criteria may lead to different model predictions. This thesis studied how the predictions of models (RUSLE and MUSLE) vary when different criteria are used to classified the slope gradients. Decomposing a watershed into a set of hydrological response units (HRU) can provide a better consideration of spatial variability of the soil erosion processes. HRU were used as analysis windows in previous watershed erosion studies. HRU were used as the experiment units in this study. Soil erosions were predicted based on each HRU. RUSLE and MUSLE models were integrated with a geographic information system (GIS). Slope was classified using three different schemes and inputted into GIS to decompose watershed into hydrologic response units. Model predictions were made for each classification scheme. The resulting three different sets of predictions were analyzed. This analysis indicates that a coaser classification scheme had higher soil loss prediction. Procedures to develop hydrologic response units, model input parameters and integrate soil loss equations with GIS are also developed to offer users a complete tool set. A new procedure using the flowlength function of GIS was also developed in this study. This procedure calculated the slope length of each HRU and could compensate the deficiencies of previous methods.