基于协同克里格插值法的年均温空间插值的参数选择研究

Study on Parameters Setting of Ordinary Cokriging Interpretation to Average Annual Temperature

为了确定可能的参数设定对协同克里格插值结果的影响,本文以福建省66个气象站点的2007年平均气温为例,在ArcGIS软件地统计模块的支持下,采用交叉检验方法对插值过程中半变异函数的最优理论模型、最佳分割距离、基台和块金以及空间数据变异方向性等参数和模型的选择进行了影响分析.结果表明,考虑高程影响因素的协同克里格(Cokriging)插值结果可以较真实的模拟气象的空间分布特征,而在插值时选用什么样的模型和参数对插值结果影响较大,为此,应综合数据和研究区的特征,对相关模型和参数进行合理的选择,进而获得最小的拟合误差和最优的插值结果.

In order to explore the effect of possible parameters setting on the kriging interpolation results,this paper took the average temperature in 2007 of 66 Fujian meteorological station as example,under the support of ArcGIS software,we did the impact analysis of the selection of models and some parameters such as the variogram’s optimal theoretical model in the process of interpolation,the best partition distance,base station and the nugget variance,the spatial data variation directivity,with the cross-validation.The results show that the ordinary cokriging interpolation results considing the impact of elevation can simulate the meteorological characteristics truly,However,choosing what models and parameters in the processing of interpolation has great impact on the interpolation results,therefore,considering the characteristics of comprehensive data and the study area,chosen the relevant models and parameters reasonably,so that to acquire the smallest fitting error and the optimal interpolation results.