Effects of physical soil crusts on infiltration and splash erosion in three typical Chinese soils

Physical soil crusts likely have significant effects on infiltration and soil erosion, however, little is known on whether the effects of the crusts change during a rainfall event. Further, there is a lack of discussions on the differences among the crusting effects of different soil types. The objectives of this study are as follows: (i) to study the effects of soil crusts on infiltration, runoff, and splash erosion using three typical soils in China, (ii) to distinguish the different effects on hydrology and erosion of the three soils and discuss the primary reasons for these differences, and (iii) to understand the variations in real soil shear strength of the three soils during rainfall events and mathematically model the effects of the crusts on soil erosion. This study showed that the soil crusts delayed the onset of infiltration by 5 to 15 min and reduced the total amount of infiltration by 42.9 to 53.4% during rainfall events. For a purple soil and a loess soil, the initial crust increased the runoff by 2.8% and 3.4%, respectively, and reduced the splash erosion by 3.1% and 8.9%, respectively. For a black soil, the soil crust increased the runoff by 42.9% and unexpectedly increased the splash erosion by 95.2%. In general, the effects of crusts on the purple and loess soils were similar and negligible, but the effects were significant for the black soil. The soil shear strength decreased dynamically and gradually during the rainfall events, and the values of crusted soils were higher than those of incrusted soils, especially during the early stage of the rainfall. Mathematical models were developed to describe the effects of soil crusts on the splash erosion for the three soils as follows: purple soil, F_c = 0.002t − 0.384; black soil, F_c = 0.002t + 3.060; and loess soil, F_c = 0.233 1nt − 1.239. Combined with the equation R_c = F_c · (R_uc − 1), the splash erosion of the crusted soil can be predicted over time.