Modeling soil erosion dynamic processes along hillslopes with vegetation impact across different land uses on the Loess Plateau of China

Developing a process-based soil erosion model that comprehensively considers the effects of vegetation is complex but crucial for evaluating sediment reduction during vegetation restoration. Current understanding of the effect of vegetation on sediment reduction along hillslopes across different land uses is limited. In this paper, we developed a dynamic model of hillslope erosion that integrates the effects of vegetation on soil detachability and hydrodynamics (VED) based on the feedback mechanism between soil detachment and sediment transport. The VED was calibrated and validated with runoff plot data for woodland, grassland, and farmland in the Wuding, Yan, Jing and Wei Rivers of the Loess Plateau in China. Unified baseline parameters and decay coefficients of vegetation were obtained. The model validation results indicated that the coefficient of determination, Nash-Sutcliffe simulation efficiency and relative error ranged from 0.59 to 0.99, 0.56 to 0.90, and -45.99 % to 46.36 %, respectively. The decay coefficients for soil detachment capacity (0), sediment transport capacity (Tc), and the sediment reduction effect of vegetation exhibited the following order: woodland > grassland > farmland. Compared with VED, existing process-based soil erosion models failed to effectively characterize the influence of vegetation on 0 and Tc on the Loess Plateau, with differences of several orders of magnitude. The individual contributions of runoff, soil detachability, and hydrodynamics with vegetation impact on sediment yield were quantified by VED. The contribution rates of vegetation to sediment reduction decreased gradually with increasing slope length because the soil detachment capacity of bare slopes exceeded that of vegetated slopes, resulting in a faster increase in sediment yield. The contribution rates of vegetation to sediment reduction for woodland, grassland, and farmland with 20 %-60 % vegetation covers were 66.75 %-99.95 %, 57.43 %-99.63 %, and 27.91 %-88.63 % in the boundary conditions of this study, respectively. The results reveal the potential for integrating VED into other distributed watershed hydrological and sediment models to assess the effects of vegetation restoration on sediment reduction at a watershed scale.