We examined, both experimentally and theoretically, whether rilling results in higher soil erosion than would have occurred without rilling. The possibility of rilling occurs when overland flow-driven erosion processes are dominant over erosion due to raindrop impact, and that is the situation assumed in this paper. Stream power (or a quantity related to stream power such as shear stress) is commonly used to describe the driving variable in flow-driven erosion. Five flume experiments were designed to investigate the relationship between stream power and sediment concentration and how this relationship is affected by the ratio of width to depth of flow (r), and the frequency or number of rills per metre width (N) of rectangular rills. This paper presents the results of these experiments and uses a physically based soil erosion theory to show that the results of the 5 flume experiments are in accord with this theory. This theory is used to investigate the effect of all possible rectangular rill geometries and frequencies on the maximum possible sediment concentration, i.e. the sediment concentration at the transport limit, by developing general relationships for the influence of r and N on sediment concentration.

It is shown that increased stream power, which can be due to rilling, does not necessarily result in higher sediment concentration.