Decoupling effects of driving factors on sediment yield in the Chinese Loess Plateau

Investigations of runoff and sediment yield changes and their relationships with potential driving factors provide good insights for understanding the mechanisms of hydrological processes. This study attempted to present a comprehensive investigation on the spatiotemporal variations of sediment yield in the Loess Plateau using continuous observed data at 46 hydrological stations during 1961–2016, and its responses to changes of precipitation, land use/cover and vegetation cover were analyzed by using the Partial Least Squares-Structural Equation Model (PLS-SEM). The results indicated that sediment yield reduced pronouncedly during 1961–2016 in the Loess Plateau, and 77.9% of this variation was explained by the combined effects of precipitation, land-use change, vegetation dynamics and runoff reduction. Indirect effects of precipitation, land-use change, and vegetation cover on sediment yield were 0.242, −0.528 and −0.630 (P < 0.05), respectively, and direct effect of runoff on sediment yield was 0.833 (P < 0.05). According to the Pearson Correlation Coefficient, the strongest positive correlation existed between annual sediment yield and runoff (r = 0.88, P < 0.05), followed by vegetation cover (r = −0.47, P < 0.05) and land-use change (i.e. forest land and grassland) suggesting their significant trapping effects on soil erosion. However, lower correlations were examined between sediment yield and precipitation indices (−0.14<r < 0.34), and a relatively higher relationship was examined between sediment yield and heavy rainfall (P₂₅) (r = 0.34). Overall, changes in runoff and land-use/vegetation cover well explained variations in sediment yield in the Loess Plateau. The findings are expected to provide scientific and technical support for future soil and water conservation planning in the Loess Plateau, and are valuable for sustainable water resources and sediment load management in the Yellow River Basin.