Spatial autogression model for heavy metals in cultivated soils of Beijing

In order to effectively reveal the effect of influencing factors on heavy metals and provide scientific basis for controlling heavy metals pollution, conventional linear regression model and spatial autogression model were applied to analyze the relationship between heavy metals and their influencing factors in Beijing cultivated soils. The results showed that the spatial autogression model for Cr, Ni, Zn, and Hg had a better goodness-of-fit than conventional linear regression model, and yielded residuals without spatial autocorrelation, indicating that the spatial autogression model could explain the relationship between heavy metals and their influencing factors excellently. Results showed that the important influencing factors for Cr and Ni were soil parent materials and land use intensity, and the main influencing factors for Zn and Hg were mining establishments, road and oil parent materials.