Spatiotemporal analysis of the quantitative attribution of soil water erosion in the upper reaches of the Yellow River Basin based on the RUSLE-TLSD model

Soil erosion leads to land fertility reduction and land degradation, which threaten food and ecological security. Therefore, it is fundamental for soil erosion risk control to accurately assess the dynamics of water soil erosion and reveal its driving mechanism. This paper aims to integrate and calibrate the transport limited sediment delivery (TLSD) function with the revised universal soil loss equation (RUSLE) model, and then to simulate the long-term soil erosion by water in the upper reaches of the Yellow River Basin (UYRB) and estimate the influence of each factor on soil loss using the GeoDetector method. The results reveal that the water erosion dynamics in the UYRB were reliably simulated by the RUSLE-TLSD model (with the Nash-Sutcliffe efficiency coefficient surpassing 0.809). From 1982 to 2019, the net soil loss rates ranged from 1.63 to 5.41 t⋅ha−1⋅a−1 in the UYRB, with the most intense erosion occurring from June to August. Steep hillside grasslands, especially in the canyon zone with fragmented topography, suffered from intense erosion, while river valleys and the feet of steep slopes experienced sediment deposition. Major factors that impact the distribution of water erosion in the UYRB include the slope followed by vegetation (soil erosion risk should be monitored on steep grasslands with slopes greater than 15°), and the combination of the two represents the core interaction factor, which explained at least 46% of the water erosion distribution. The annual net water erosion in the UYRB showed a slight increasing trend from 1982 to 2000, which was mainly affected by natural factors, but significantly decreased after 2000, which was mainly caused by vegetation restoration. In the UYRB, the implementation of the “Grain to Green Program” has had significant benefits in terms of water and soil conservation, and forests have a greater ability to control erosion than grasslands.