Morphological development of rills and its relationship with hillslope erosion in the hilly and gully Loess Plateau

Abstract: Rill erosion has been widely recognized as one of the most important forms of soil erosion on hillslopes. A crucial impact of rill erosion can be also posed on other erosion processes in downslope areas (e.g. gully head retreat). Morphological parameters of rills can provide useful indicators for the initiation and development of rill erosion. Previous studies have investigated the rill morphology in the erosion-deposition processes of hillslopes. However, those experiments were mainly taken in the laboratory. The physiochemical properties of backfill soil used in laboratory experiments are rather different from those of the natural soil in the field. The representative experiments were largely confined to the field erosion processes. It is necessary to explore the rill morphology associated with the erosion processes in the field. In this study, a series of field scouring experiments were conducted to determine the morphological development of rills under the hillslope erosion in the hilly and gully Loess Plateau. Five erosion plots were established on a natural slope of a small catchment (i.e. Xindiangou catchment), particularly with the input flow of hillslopes of 25, 40, 55, 70, and 85 L/min. Terrestrial Laser Scanning (TLS) was employed to acquire the ultra-high terrain information prior to the test. The various morphological parameters of rills were then derived, including the geometric indicators (length, width, and depth of cross sections), derived indicators (the ratio of width to depth, rill density, rill cleavage, and average rill depth), fractal dimension, bifurcation ratio, and geomorphic information entropy. A systematic investigation was also made to determine the effects of indicative morphological parameters on the cumulative erosion and deposition mass, as well as the sediment yield in the hillslope erosion. Results showed that: 1) The width and depth of the cross-sectional rills, the average rill depth and rill cleavage increased as the experiment progressed under all the flow conditions. The width-depth ratio was greatly varied in the input flow rate. The rills were primarily wide and shallow under the low flow condition (25 L/min), while narrow and deep under the moderately low (40 L/min) and high flow (85 L/min) conditions. There was a great change between the narrow-deep and wide-shallow manner under the moderate (55 L/min) and moderately-high (70 L/min) flow conditions. 2) The fractal dimension of rills was found to change slightly under the low flow condition, whereas, there was a considerable change under the rest of the input flow condition. The bifurcation ratio of rills increased under the moderate flow condition, while decreasing under the rest flow condition. Furthermore, the geomorphic information entropy varied significantly under the different flow conditions. However, there was the same change trend of geomorphological information entropy and sediment yield, indicating the dynamic changes of soil erosion. 3) The average rill length, the average depth of cross sections, and the derived average rill depth served as better indicators for the cumulative deposition mass, erosion mass, and sediment yield under the moderately low flow condition. The derived average rill depth was also for the cumulative erosion mass and cumulative sediment yield under the low flow condition. The average width of cross sections, average depth of cross sections, and derived average rill depth better indicated the cumulative erosion mass under the moderately high flow condition. In addition, there was a less significant relationship between the rill morphological parameters and cumulative erosion mass, deposition mass, and sediment yield, as the input flow increased. The finding can provide a strong reference to enhance the current understanding of the processes and mechanisms of hillslope erosion.