根据全国2010年607个气象站点的日降水量资料以及各站点平面分布与高程信息，构建了多维线性空间插值方法。在不同插值控制站点数目( n = 4,8,12 )情景下，分析了多维线性空间插值方法与反距离权重插值、样条插值和克里格插值法（球面模型和高斯模型）在我国年降水量空间插值中的应用情况。对比发现，不同的控制站点情境下，各种插值方法的插值结果差异显著，但随着控制站点数目的增加，各插值结果的差异性减小；多维线性插值方法在多种情景下都表现出较好的插值精度，该方法将是今后空间插值的一个切实可靠的手段。

Spatial interpolation is an important approach to investigate the spatial distribution of meteorological and hydrological variables, and the suitable interpolation methods and reasonable control stations are beneficial to improve the precision of interpolation results. In this paper, a multidimensional liner spatial interpolation method was developed based on the daily precipitation at each of 607 stations in 2010 and the horizontal distribution and elevation of each station. Under different scenarios with varying control stations (n=4, 8, and 12), The applications of multidimensional linear interpolation, inverse distance weighting interpolation (IDW), spline interpolation, and kriging interpolation (Spherical model and Gaussian model) to calculate the annual precipitation of 30 test stations were analyzed. The results showed that the interpolation results from different methods are significantly different, and the number of control stations has a significant impact on the precision of results with the existence of more control stations resulting in less difference between interpolation results. The multidimensional linear interpolation showed good interpolation precision under all scenarios and it can be an effective and reasonable solution for the future research of spatial interpolation.

国家重点基础研究发展计划（973计划）项目(2010CB951102)；国家自然科学基金创新研究群体基金项目（51021066）；国家自然科学基金(51209225)；水文水资源与水利工程科学国家重点实验室开放研究基金（2011490511）