Comparison of spatial interpolation methods for rainfall erosivity in a typical basin in the Hengduan Mountain region, Southwest China

作者：Jinru Xie, Lin Ding, Xiangdong Wang, Wei Qin, Hai‐Chao Xu, Minghao Zhang · 发表于：Ecological Indicators · 年份：2025 · DOI：10.1016/j.ecolind.2025.113451 · 被引用次数：6 · 研究领域：Soil erosion and sediment transport、Hydrology and Watershed Management Studies、Hydrology and Sediment Transport Processes

• Elevation should be an auxiliary variable to aid the mapping of rainfall erosivity. • The relationship between R-factor and elevation is non-stationary. • The number of gauges has an impact on spatial interpolation result. • Fewer samples amplify accuracy variance with distribution shifts. The rainfall erosivity factor (R-factor) is an important parameter in the universal soil loss equation (USLE), the revised universal soil loss equation (RUSLE) and several other soil erosion prediction models. It is necessary to choose suitable methods to map the R-factor at the basin and regional scales to effectively apply erosion prediction models and establish precise soil and water conservation measures. When rain gauges are sparsely and unevenly distributed on high and steep terrain, it is often difficult to obtain ideal results from traditional spatial interpolation methods. To optimize the interpolation method for the R-factor in mountainous areas and explore the impact of the number and distribution pattern of rain gauges on the interpolation results, the Longchuan River Basin in the Hengduan Mountain region in Southwest China was selected as the study area. Four methods, namely, inverse distance weighting (IDW), ordinary kriging (OK), global regression kriging (GRK) and geographically weighted regression kriging (GWRK), were selected to conduct a comparative analysis under various scenarios of gauge number and distribution. Compared with traditional univariate methods, GRK and G...