Large areas of farmland have been established in China’s Loess Plateau region by backfilling forest and grasslands in gullies with topsoil from hillslopes. High amounts of plant litter are incorporated into the topsoil during such processes; however, their effects on soil erosion resistance under natural conditions are not fully understood. Therefore, the present study investigated seasonal variation in soil erosion resistance to overland flow and the underlying factors in the gully-filled farmland. Two typical plant litter species were selected in this study, i.e., Artemisia sacrorum Ledeb. (ASL) and Bothriochloa ischaemum (L.) Keng. (BIL). Overland water flow was tested under six shear stress levels (5.66–22.11 Pa) in natural soil samples incorporated with plant litter and without plant litter to determine soil erosion resistance parameters (rill erodibility Kr and critical shear stress τc). The study was conducted over 524 days (through May 6, 2017, to October 11, 2018), covering two rainy seasons and one winter. Soil properties changed significantly over time, with physical crust thickness and cohesion increasing from 0 to 10.7 mm and 2.9–13.4 kPa, respectively. In contrast, the mass density of incorporated plant litter decreased from 0.7 to 0.2 kg m−2. Soil erosion resistance increased with time, which showed that Kr decreased (from 0.516 to 0.021 s m−1) and τc increased (from 0.49 to 7.98 Pa). Compared with the initial state of new farmland (Kr = 0.507 s m−1,τc= 0.48 Pa), Kr decreased by 77–96% and τc increased 8–14 times at the end of the study. Plant litter incorporation markedly enhanced soil erosion resistance in a time- and species-dependent manner. Compared with the bare soil without plant litter, the Kr values of ASL and BIL treatments were 3–64% and 2–83% lower, and their τc values were 0–81% and 0–55% higher, respectively. Kr was significantly and positively correlated with cumulative precipitation, cumulative rainfall kinetic energy, soil cohesion, and physical soil crust thickness, while τc was negatively correlated with the influencing factors. Further, both Kr and τc could be well predicted by physical soil crust thickness and cohesion (R2 = 0.79 and 0.87, p < 0.01, n = 27). According to the results, seasonal patterns of soil erosion resistance in the gully-filled farmland established under plant litter incorporation were mainly controlled by rainfall and physical soil crust structure. This study provides a reference resource for soil erosion prediction in farmland ecosystems on the Loess Plateau.
Read full abstract