中文摘要：

时空地统计是传统地统计学结合时间维的扩展,其行业应用非常广阔,然而目前时空地统计模型的计算实现研究较少,时空变异函数是时空地统计的核心,研究时空变异函数理论模型的实现是时空地统计技术突破的关键。该文探讨了时空变异函数的积和模型理论,讨论了时空变异函数的两种实现方法和数据时空平稳化的预处理方法。以东北三省月平均气温数据为实验数据,检验了两种时空变异函数构造算法的效果,分析获知东北三省月平均气温时空结构性显著,空间结构性比时间结构性更为显著。基于现有实验数据,发现在空间48km内和时间3个月内东北月均值气温是相关的,超出这个范围则相关性逐步消失。该文工作提供了一条实验可行的时空地统计计算的技术路线。

英文摘要：

Currently, spatial-temporal geostatistics still stays in a theoretical and experimental exploration stage, without a kind of commercial software and computational research reports about its model. The computer implementation of spatial-temporal variation function is the breakthrough of the technique in spatial-temporal geostatistics. This paper shows the theoretical about Product-Sum model of spatial-temporal variation function, and gives out two kinds of methods to get the stationary data both in time and in space. The monthly average temperature data of the northeastern provinces is used as experimental data in the paper. Through the test used by two kinds of spatial-temporal variation function, the result shows that the space-time has a significant structural and the structure of space is stronger than that of time. By the experiment, the result shows that the monthly average temperature in space within 48 km and time within 3 months in northeastern provinces is relevant and disappear gradually when it goes out of the range. In a word, this paper provides an experimental feasible technical route to do the space-time statistical calculation as a foundation.