丘陵区水稻土Cu污染空间变异的协同克里格分析

针对丘陵红壤区铜冶炼厂周围水稻土污染区(1.40 km2),在景观尺度上,采用协同克里格方法,研究了影响表层土壤Cu含量空间分布预测的辅助因子.基于空间自相关性、间距、长轴方位角以及各种预测误差,评价了辅助变量(包括秸秆全Cu含量StrawCu、籽粒全Cu含量GrainCu、土壤全Cd含量Cd、土壤pH、土壤有机质OM、高程H)对表层土壤Cu含量分布预测精度的影响.结果表明,单辅助变量的协同克里格预测值与实测值相关系数的大小顺序为Cu/Cd＞Cu/H＞Cu/StrawCu＞Cu/GrainCu＞Cu/OM、Cu/pH,而多辅助变量协同克里格预测的相关系数大小顺序为Cu/(Cd,StrawCu)＞Cu/(Cd,StrawCu,H)＞Cu/(Cd,StrawCu,GrainCu)＞Cu/(StrawCu,GrainCu)＞Cu/(Cd,H).与土壤全Cu含量的普通克里格插值精度相比,利用表层土壤全Cd含量、水稻秸秆全Cu含量、高程作为辅助变量与水稻土表层全Cu含量进行协同克里格插值可以显著提高预测精度;但水稻籽粒全Cu含量作为辅助变量对预测精度影响不显著;而土壤有机质含量和土壤pH作为辅助变量反而降低了预测精度.在对表层土壤全Cu含量分布的多辅助变量协同克里格预测中,表层土壤全Cd含量和水稻秸秆全Cu含量的影响最大,其次是高程,水稻籽粒全Cu含量不能提高对表层土壤全Cu含量分布的预测精度.

Cokriging Analysis of Spatial Variability of Cu Pollution in Paddy Soils in a Hilly Region

The auxiliary variables to affect the prediction accuracy of Cu spatial variability in the paddy topsoils were evaluated using cokriging method. An area(1.40 km2)around the Jiangxi Copper Smelter was selected in a hill region of Guixi County, Jiangxi Province, China. Autocorrelation properties, range, azimuth angles in the long axis direction, and prediction errors were compared to determine the performance of cokriging with single-and multi-auxiliary variables. The relationship coefficient between the pred...