中文摘要：

本文选取江西省86个气象站点2008年的年降水量数据，基于地统计学理论对数据进行探索性空间数据分析，得出该组气象数据插值的理论最优模型。采用不同全局趋势及变差函数模型进行插值，通过交叉性检验证实了该理论模型的有效性。最终结果显示，采用3次函数剔除全局趋势并以球状模型作为变差函数模型的考虑高程的Co—Knging方法插值结果最优。

英文摘要：

Based on geostatistics methods and the serial climate data of 86 meteorological stations of Jiangxi in 2008, this paper discussed the optimization model of the precipitation data through Exploratory Spatial Data Analysis. The data were interpolated with different trend functions and variational function models, result of which proved the effectivity of this optimization model. The cross validation result showed that altitude-considered Co-Kriging interpolation with cubies function as trend removal and Spherical model as variational functions had a higher precision.