Quantifying the Drivers and Constraints of Runoff and Sediment Yields: Insights From the Xiangjiang River Basin

Abstract

The spatiotemporal heterogeneity of runoff and sediment yields significantly influences both human water resource utilization patterns and watershed ecosystem health. However, the constraining effects of environmental factors on these processes remain understudied. Here, based on accurate simulations of hydrological processes, machine learning algorithms and structural equation modeling were employed to systematically analyze the nonlinear driving effects and influence pathways of factors impacting runoff and sediment yield. We elucidated the constraints imposed by factors on the spatiotemporal variation in runoff and sediment yield. The findings revealed that precipitation (69% contribution rate) and evapotranspiration (23%) served as the primary direct drivers of runoff variation, with other factors playing minimal roles. Slope, clay content, precipitation, and forest cover (each with ≥ 10% contribution) exerted direct control on sediment yield, whereas precipitation indirectly influenced it through runoff and evapotranspiration pathways. The majority of the factors exhibited substantial constraining effects on runoff and sediment yields, although some factors had a negligible driving role. A comprehensive evaluation of the driving and constraining effects of environmental factors on runoff and sediment yields and a quantitative assessment of their potential is essential to formulate and implement watershed management strategies for runoff and sediment yields.