不同插值方法下的土壤重金属污染三维空间分析

Three-Dimensional Spatial Analysis of Soil Heavy Metal Pollution under Different Interpolation Methods

摘要: 为确定某地块区域内土壤污染物的空间变异与三维分布状况，选择退役某合金制造工厂内土壤重金属Cu为研究对象，对比Kriging、IDW和Nearest Neighbor 3种三维空间插值方法对污染预测的不确定性影响。结果表明，不同插值模型计算结果差异较大。交叉验证结果表明：Kriging模型插值精度最高，其预测结果能较真实地反映实际污染情况；三维可视化方法相较于传统的二维空间分析可以解决二维平面不能直观展示污染物垂直空间变化的问题，可以多方式多角度展示土壤信息，反映土壤内部污染物形态。

Abstract: In order to determine the spatial variability and three-dimensional distribution of soil pollutants in contaminated sites, the soil heavy metal Cu in a certain alloy manufacturing plant was selected as the research object, and the uncertainty effects of three 3D spatial interpolation methods on pollution prediction, Kriging, IDW and Nearest Neighbor were compared. The results show that the results of different difference models are large. The cross-validation results show that the Kriging model has the highest interpolation accuracy, and the prediction results can reflect the actual pollution situation. Compared with the traditional two-dimensional spatial analysis, the three-dimensional visualization method can solve the problem that the two-dimensional plane cannot visually display the vertical spatial changes of pollutants, and can display soil information in multiple ways and at multiple angles, reflecting the form of pollutants in the soil.