Correlating check dam sedimentation and rainstorm characteristics on the Loess Plateau, China

•Layered sediment volume V has prominent power relation to rainfall erosivity R.•A new scale method to reconstruct R was presented.•Distance of dams has greater influence on R reconstruction than hysteresis pattern.•The scale method performs well and has good applicability in adjacent regions.•Sedimentation in dam deposition wedges is a feasible indicator to reconstruct R.

On the Loess Plateau, China, check dams are necessary for soil and water conservation. In this study, the relationships between check dam sedimentation and storm characteristics were investigated, and rainstorm events for an area of data scarcity were reconstructed using optimal regression models. Four typical check dams of the Loess Plateau (the Shipanmao, Hualiang, No. 3 Guandigou, and No. 4 Guandigou dams) were selected for case studies. Soil profiles behind the four dams were divided into 21, 25, 31, and 31 layers, respectively. The one-to-one link between sediment layers and corresponding storm characteristics were considered based on the peak fallout of 137Cs and on the principle that higher flow corresponds to higher sediment deposition volume. Analysis showed that the layered sediment volume (V) was closely related to rainfall erosivity (R), and to the maximum rainfall intensity over 30 min (I30). A statistically significant power regression model between V and R was observed, along with an exponential regression model between V and I30. Based on the power regression model between V and R, a new approach was developed in order to reconstruct rainfall erosivity in rainfall data-scarcity areas. Different hysteresis patterns (counterclockwise, clockwise, and hybrid) at the single storm scale were observed between V and R for the Shipanmao dam deposition wedge (depositional area behind a dam), and these impacted differently on the scale reconstruction results. However, the distance between the reconstructed dams and the known dams had an even stronger influence. The scale method developed in this study was shown to perform well and was able to reconstruct rainstorm events for adjacent regions with similar climatic and geomorphological conditions. Furthermore, the results of this study have improved our understanding of single rainstorm dynamics and soil erosion mechanisms on the Loess Plateau.