3D spatial interpolation of soil heavy metals by combining kriging with depth function trend model

In this study, a hybrid method combining 3D kriging and depth function considering heterogeneity (3DK_DF) was proposed to enhance the accuracy and reliability of 3D spatial interpolation for soil heavy metals. Soil samples collected at varied depth intervals in a mining city in China were used as a case dataset. First, the parameters of a logarithmic depth function model for every horizontal soil sample site were fitted on the basis of the observed values of soil Cd collected at varied depth intervals. Second, the 3D trend of soil Cd was obtained on the basis of the spatial distributions of the parameters of the logarithmic depth function model. Third, 3D kriging was used to generate the 3D spatial distribution of residual Cd. Finally, the 3D spatial distribution of the soil Cd was obtained by combining the 3D trend and residual results. The interpolation accuracy of 3DK_DF improved by 29.71% and 48.9% compared with those of the 3D kriging without a trend analysis and the 3D kriging with a polynomial trend model, respectively. The proposed hybrid 3D interpolation method could be of great significance for the comprehensive assessment of soil heavy metal pollution.