A modified CSLE for slope-scale soil loss prediction under different vegetation types in China

Vegetation can significantly reduce soil erosion by intercepting rainfall, reducing surface runoff, and improving soil properties. Increased vegetation coverage strengthens these effects by enhancing canopy density, root development, and litter accumulation. To quantify these effects, multiple coverage-based equations have been proposed for estimating vegetation-related factors in soil erosion prediction models such as the Universal Soil Loss Equation (USLE) and the Chinese Soil Loss Equation (CSLE). However, these equations often neglect understory vegetation, inadequately incorporate vegetation types, and rely on region-specific data, limiting their broader applicability across China’s diverse ecosystems. To address these limitations, this study developed three equations applicable nationwide for estimating the biological control factor (B) based on vegetation coverage for grassland, shrubland, and woodland. The method underwent calibration and validation using data from 54 sites, with further evaluation at 8 independent sites. Results revealed substantial enhancements in sediment yield prediction accuracy relative to the storm-based CSLE. During calibration, Nash-Sutcliffe Efficiency (NSE) values increased from 55.76% to 70.25% for grassland, from 44.08% to 81.95% for shrubland, and from -61.18% to 73.12% for woodland. Validation results exhibited parallel improvements, with NSE increasing from 61.07% to 74.55% for grassland, 48.44% to 77.41% for shrubland, and -77.00% to 68.23% for woodland. These findings collectively underscore the superior predictive performance and expanded geographical applicability of the proposed method across China's diverse ecosystems.