Improved USLE-K factor prediction: A case study on water erosion areas in China

Soil erodibility (K-factor) is an essential factor in soil erosion prediction and conservation practises. The major obstacles to any accurate, large-scale soil erodibility estimation are the lack of necessary data on soil characteristics and the misuse of variable K-factor calculators. In this study, we assessed the performance of available erodibility estimators Universal Soil Loss Equation (USLE), Revised Universal Soil Loss Equation (RUSLE), Erosion Productivity Impact Calculator (EPIC) and the Geometric Mean Diameter based (Dg) model for different geographic regions based on the Chinese soil erodibility database (CSED). Results showed that previous estimators overestimated almost all K-values. Furthermore, only the USLE and Dg approaches could be directly and reliably applicable to black and loess soil regions. Based on the nonlinear best fitting techniques, we improved soil erodibility prediction by combining Dg and soil organic matter (SOM). The NSE, R2 and RE values were 0.94, 0.67 and 9.5% after calibrating the results independently; similar model performance was showed for the validation process. The results obtained via the proposed approach were more accurate that the former K-value predictions. Moreover, those improvements allowed us to effectively establish a regional soil erodibility map (1:250,000 scale) of water erosion areas in China. The mean K-value of Chinese water erosion regions was 0.0321 (t ha h)·(ha MJ mm)−1 with a standard deviation of 0.0107 (t ha h)·(ha MJ mm)−1; K-values present a decreasing trend from North to South in water erosion areas in China. The yield soil erodibility dataset also satisfactorily corresponded to former K-values from different scales (local, regional, and national).