Projected increase in soil erosion risk in the karst landscape of the Pearl River Basin under high carbon emission scenarios

Abstract

Soil erosion (SE) risk in watersheds is highly uncertain under the coupled effects of climate change and land use change (LUC). Accurate assessment of SE risk dynamics under future environmental conditions is essential for effective ecological protection and sustainable watershed management. This study highlights the critical role of SSP-RCP scenario input on model predictions and emphasizes the need for consistent integration of climate, natural, and socio-economic factors. We present an innovative integrated downscaling model framework that improves upon traditional downscaling approaches by linking large-scale watershed land use and climate change. In addition, we have optimized the Revised Universal Soil Loss Equation (RUSLE) model by incorporating the rock outcrop factor and integrated key socio-economic drivers to achieve precise assessment of SE risk. Our results show a consistent increase in SE risk under all three SSP-RCP scenarios, with high-risk areas predominantly located in the karst regions of the upper Xi Jiang River Basin. Applying GeoDetector and Geographically and Temporally Weighted Regression models further indicates that subsurface characteristics, natural factors, and anthropogenic activities are the primary drivers of SE risk. Specifically, economic growth and in LUC (cropland and grassland) are the key factors driving the escalation of erosion risk in the upper watershed. Our study presents a novel large-scale watershed downscaling model and an optimized RUSLE framework, enabling an accurate spatiotemporal assessment of SE risk under climate and LUC in the Karst Landscape of the Pearl River Basin. These results provide valuable insights for targeted soil erosion prevention and management in other karst regions.