Abstract Precipitation is among the major drivers in hydrological modeling, and spatial rainfall distribution is essential for understanding ecohydrological processes. Inaccurate rainfall spatial surface is the main source of model uncertainty, especially where rainfall gauges are sparse in mountainous areas and, where the spatial precipitation distribution is closely related to elevation. This study proposes a daily rainfall interpolation method based on the Angular Distance Weighting (ADW) method considering the precipitation–elevation relationship and topographic effects derived from observed data. Furthermore, the method is used to generate daily rainfall in the Shiyang River basin from 1984 to 2012. Cross-validation results indicate that this method is rational, where the annual mean relative errors (MRE) are less than 30% at the majority of the stations. The rationality of the interpolated precipitation data at high altitude regions, where limited rainfall gauges exist, is also evaluated using a comparison of the observed and simulated runoff modeled by the Soil & Water Assessment Tool (SWAT). The SWAT model with interpolated rainfall could generate better-simulated discharges than that with observed rainfall data. The results demonstrate that the proposed method could provide an accurate spatial distribution of daily precipitation, which can sufficiently drive the hydrological model.

中文翻译：

---

基于修正ADW法的稀缺山区日降水空间插值——以石羊河流域为例

摘要 降水是水文建模的主要驱动因素之一，空间降水分布对于理解生态水文过程至关重要。降雨空间表面不准确是模式不确定性的主要来源，尤其是在山区雨量计稀少且空间降水分布与高程密切相关的地区。本研究提出了一种基于角距离加权（ADW）方法的日降雨量插值方法，考虑到降水-高程关系和来自观测数据的地形效应。此外，该方法用于生成石羊河流域1984-2012年的日降雨量。交叉验证结果表明该方法是合理的，其中大多数站点的年平均相对误差 (MRE) 小于 30%。还使用土壤和水评估工具 (SWAT) 模拟的观测和模拟径流的比较来评估高海拔地区内插降水数据的合理性，其中存在有限的雨量计。内插降雨的 SWAT 模型可以产生比观测降雨数据更好的模拟排放。结果表明，所提出的方法可以提供准确的日降水空间分布，足以驱动水文模型。还使用土壤和水评估工具 (SWAT) 模拟的观测和模拟径流的比较进行评估。内插降雨的 SWAT 模型可以产生比观测降雨数据更好的模拟排放。结果表明，所提出的方法可以提供准确的日降水空间分布，足以驱动水文模型。还使用土壤和水评估工具 (SWAT) 模拟的观测和模拟径流的比较进行评估。内插降雨的 SWAT 模型可以产生比观测降雨数据更好的模拟排放。结果表明，所提出的方法可以提供准确的日降水空间分布，足以驱动水文模型。