Soil properties play an important role in rill development and erosion. In this investigation, rill morphology developmental processes under sandy loam (SL), light loam (LL), medium loam (ML) and heavy loam (HL) soils on the Loess Plateau, China, were compared using laboratory experiments. Experimental analysis included two rainfall intensities (90 and 120 mm/h) and four slope treatments (0°, 15°, 20° and 25%). Results indicate that HL is the most prone to rill development, and SL, LL and ML are prone to rill development under heavy rain, with SL rill erosion being the most sensitive to heavy rain. The development of rills in SL are mainly characterized by an increase in rill width and merging nodes; rills in HL were mainly characterized by an increase in rill length, merging nodes and rill number. LL and ML rill development indices were between SL and HL. Differences in runoff collection caused by rill morphology differences further promoted differences in soil erosion. Rainfall intensity has a positive effect on rill shape parameters of all soils; slope has a positive and negative double effect on SL, LL and ML rill shape parameters, and only a positive effect on HL rill shape parameters. The sensitivity of rill parameters to rainfall intensity and slope angle depends on soil infiltration performance, surface soil stability and soil structure stability. Based on soil characteristic factors and rill morphological parameters, an empirical model of slope erosion in the loess region was established.

Slope length is an important topographic factor for controlling soil erosion and pivotal parameters in the soil erosion model. The impact of slope length on soil erosion was studied under different grassland and different rainfall intensity through simulated rainfall experiments. The experiment included five rainfall intensity treatments (1, 1.5, 2, 2.5 and 3 mm h−1), four grass cover treatments (0, 30%, 60% and 90%) and five slope length treatments (2, 4, 6, 8 and 10 m). The results show that the rate of soil loss increased exponentially with increasing slope length under 0 and 30% grass cover. Under high grass covers (60% and 90%), the slope length increased sedimentation from runoff and reduced slope erosion.The increase of slope length led to enhancement of soil loss as rainfall intensity increased. At 1 mm min−1 rainfall intensity, natural grass slopes (60%) controlled soil loss very well and were not affected by slope length. At 1.5–2 mm min−1rainfall intensity, the soil erosion increased with slope length, but the overall soil erosion amount was small. This indicates that grass cover at 10 m slope length had a good impact on soil erosion. When the rainfall intensity exceeded 2 mm min−1, soil loss increased with slope length. Regression analysis showed that soil erosion was more strongly related to grass cover and rainfall intensity than to slope length.