Linking hydrological and landscape characteristics to suspended sediment-discharge hysteresis in Wudinghe River Basin on the Loess Plateau, China

Abstract

This study aims to quantitatively evaluate the effects of hydrological and landscape characteristics on suspended sediment concentration and discharge (SSC-Q) hysteresis. Runoff and sediment data of 405 flood events were collected from 10 basins in Wudinghe River basin on the Loess Plateau of China. Hysteresis analysis is a key tool for the interrogation of transport of sediment, which have been shown to be important within catchments. It can provide insight into catchment function, allowing the development and testing of process-based hypotheses. The principal components and corresponding explanatory power were extracted quantitatively from hydrological and landscape characteristics via principal component analysis (PCA) and redundancy analysis (RDA), respectively. The most common hysteresis type was anticlockwise. The proportions of the total event sediment loads for different hysteresis types were not always generally correlated to the frequency of the corresponding events. Furthermore, the composition and explanatory power of controlling principal components varied among clockwise, anticlockwise, figure-of-eight, and complex hysteresis types. For hydrological principal components, the antecedent precipitation, intraday precipitation, and sediment principal components controlled clockwise, anticlockwise, and figure-of-eight hysteresis types, which had explained 47.03%, 58.82% and 52.05% of the variation of hysteresis index, respectively; the antecedent precipitation and time lag principal component controlled complex hysteresis type and accounted for 28.22% and 34.65% of the variation of hysteresis index, respectively. For landscape principal components, grassland principal component dominated clockwise, anticlockwise, and figure-of-eight hysteresis types, and accounted for 95.00%, 69.70%, and 92.00% of the variation of hysteresis index, respectively. Largest patch index principal component controlled complex hysteresis type and accounted for 100% of the variation of hysteresis index. Therefore, within this catchment, whereas hysteresis patterns varied due to all of the hydrological and landscape variables to some extent, precipitation, suspended sediment, area of grassland, and the largest patch area had stronger impact on hysteresis dynamics.