土壤属性预测中Kriging参数对插值精度的影响研究  

Effects of Parameter Condition on Prediction Precision of Soil attributes by Kriging Interpolation

作　　者：巫振富[1] 赵彦锋[1] 陈杰[1] 

机构地区：[1]郑州大学水利与环境学院,自然资源与生态环境研究所,河南郑州450001

出　　处：《土壤通报》2012年第3期551-558,共8页Chinese Journal of Soil Science

基　　金：国家自然科学青年基金项目(40801080);国家自然科学基金项目(40971128)资助

摘　　要：为了有效提高土壤属性预测中的Kriging插值精度,研究借助GS+5.1和ArcGIS 9.3等软件来分析河南省西平县宋集乡土壤全磷和有机质在不同参数条件下的Kriging插值精度,结果表明:决定系数R2最大的变异函数模型,插值结果误差不一定最低、平滑效应不一定最弱,需比较分析插值结果以选择合适的模型;近距离实验变异函数值r#(h)的拟合对提高Kriging插值精度具有重要意义,拟合变异函数理论模型时,除综合考虑各个r(h)#的拟合程度外,同时对近距离r(h)#应给予足够的重视;插值的最佳邻域样本量一般在内层和次内层邻域样本权重达到平稳时取得,对于样本点空间分布相同的不同土壤属性,空间相关性较弱(块金系数较大)的土壤属性,其Kriging插值的最佳邻域样本量较大。To compare precision of soil attribute prediction with the method of Kriging interpolation under the different parameter conditions,the contents of the total P and organic matter of topsoil in the studied area located in Xiping County,Henan Province were predicted with the aid of the software GS ＋ 5.1 and ArcGIS 9.3.The results indicated that the variogram with maximum determination R2 did not necessarily mean the minimum interpolation inaccuracy and the weakest smoothing effect,and choice of variogram model should be figured out through comprehensive analysis of interpolation output.It was shown that the fitting of closer experimental semivariance r#（h） was of significance in precision promotion of Kriging interpolation,therefore more attention should be paid to the closer r#（h） when fitting a theoretic variogram model.For interpolation,the optimum numbers of neighborhood samples could be obtained only when the weights of the innermost layer and the sub-innermost layer neighborhood samples reached a balance.As for the target soil attributes with the same spatial distribution of the sampled sites,the weaker spatial interdependency of the attributes was（bigger Nugget/sill）,the bigger optimum number of neighborhood samples should be taken,for a better interpolation.

关 键 词：土壤属性 预测精度 KRIGING插值 变异函数模型 最佳邻域样本量 

分 类 号：S159-3[农业科学—土壤学]