Calibrating the WEPP model to predict soil loss for some calcareous soils

Accurate soil loss prediction at the regional scale is necessary for a better understanding of the soil erosion processes and conservation practices. This study aimed to calibrate the WEPP model for predicting soil loss under semi-arid region conditions in the northwest of Iran. Inter-rill and rill erosion were simulated at 59 points with three replications. Therefore, the ability of the regression equation in WEPP model and the derived regression and artificial neural network (ANN) models by Mirzaee et al. (2017) for predicting baseline soil erodibility parameters were evaluated to estimate soil loss. The results of the present study showed that the WEPP model that applied soil baseline erodibility predicted data by the regression equation in the WEPP model performed poorly in comparison to the derived aspatial models by Mirzaee et al. (2017). Additionally, the WEPP model that used soil baseline erodibility predicted data by the developed ANNs by Mirzaee et al. (2017) yielded the best results with the highest R2 (0.681) and the lowest RMSE (5.1 Mg ha−1) values for predicting soil loss rate. In general, the prediction map of soil erosion showed that soil erosion varied from 1.0 to 26.5 Mg ha−1.