Multiple surface runoff and soil loss responses by sandstone morphologies to land-use and precipitation regimes changes in the Loess Plateau, China

•Pisha sandstone reliefs in the Loess Plateau are considered a highly vulnerable relief.•50 erosive precipitation events were considered during seven years.•Surface runoff (R) and soil loss (SL) per different land-use (LU) type was observed.•Precipitation regimes combined with LU changes influenced R and SL significantly.•The optimization the of vegetation patterns according to the precipitation regimes is key. In this research, a long-term (2014–2020) approach to quantify surface runoff and soil loss generated by different land use and precipitation regimes on Pisha sandstone hillslopes from the Loess Plateau, China was carried out. Using the K-means clustering algorithm, 50 precipitation events were classified into three different regime types based on depth, duration, and maximum 30-minute intensity. Permutational multivariate analysis of variance (PERMANOVA) and Principal component analysis (PCA) were used to test clustering rationality. Our results suggest that runoff coefficient (RC) and soil loss (SL) exhibited significant differences depending on precipitation regimes: heavy storms (precipitation regime II) were found to induce the largest surface runoff and soil loss, followed by precipitation regimes I and III. RC and SL were also found with significant differences (p cropland > artificial grassland > native grassland > shrubland > forestland. Results of this study suggest more attention should be paid to vegetation selection and land-use type depending on precipitation regimes in Pisha sandstone morphologies. Accordingly, forestland and shrubland should be the first choices to control soil erosion when land-use conversion is implemented, whereas bare land, croplands, or artificial grassland (e.g., alfalfa) should be carefully considered and future vegetation restoration policies should assess the proper re-vegetation type before any action may be initiated.