Soil detachment by sediment-laden rill flow interpreted using three experimental design methods

Soil detachment by rill flow is a key sub-process of rill erosion on hillslope areas; however, many studies have focused on soil detachment by clear rather than by sediment-laden rill flow and most studies have used the time and labor-consuming full-factorial design. This study used three experimental design methods, including the full factorial design, orthogonal design, and Taguchi method for indoor flume experiments to explore the process of soil detachment by sediment-laden rill flow. We analyzed and quantified the effects of the slope, flow discharge, and sediment load on the soil detachment rate by sediment-laden rill flow, and compared the results of the three experimental design methods to evaluate the applicability of the orthogonal design and Taguchi method in indoor flume experiments. The results showed that the flow discharge and slope were positively correlated with the soil detachment rate, while the sediment load was negatively correlated with it. The main effects of slope, flow discharge, and sediment load on the soil detachment rate, which were obtained from the three methods, had the same trends. The contribution rates of the flow discharge, slope, and sediment load were 46.20-47.83, 34.42-40.88, and 7.45-12.87%, respectively. The contribution rates of each factor obtained from the three methods were slightly different, but the overall ranked order was the same. The Taguchi method and the orthogonal design can effectively reduce the number of tests and ensure low relative errors. The prediction of the soil detachment rate from the Taguchi method was closer to the full-factorial design than the orthogonal design. Therefore, the Taguchi method proved was most efficient for flume experiments on soil detachment processes by the sediment-laden rill flow.