Machine learning reveals heavy metal migration pathways in Asia's largest Pb-Zn smelting region: Soil pollution simulation in Jiyuan

Abstract

Although heavy metal (HM) pollution caused by the non-ferrous metal smelting industry to the soil environment has been widely recognized, clarifying the migration pathways of HMs remains a critical scientific issue that urgently needs to be addressed. This study employs spatial autocorrelation and machine learning to uncover HM migration processes near polluting enterprises in a major Chinese Pb-Zn smelting base. Pb was the most enriched (17.48 × Henan background) and spatially heterogeneous(CV = 106 %) HM; its concentration, along with Zn and Cu, showed a pronounced decrease with increasing distance from the smelter. Source apportionment reveals that the contribution rates of anthropogenic sources to Pb, Cu, and Cr are as high as 89.23∼98.94 %. Further using the newly constructed spatial correlation model, showed that the spatial distribution of Pb and Zn exhibits significant heterogeneity, which is mainly influenced by the uneven spatial distribution of industrial enterprises. The migration pathways of Pb show distinct seasonal characteristics: within areas close to the plants, migration is mainly influenced by spring and summer wind directions, while in areas farther from the plants, migration pathways are associated with autumn and winter wind directions. Additionally, traffic activities have also been confirmed as an important factor affecting the migration and distribution of Pb. By revealing the distribution characteristics and migration patterns of HMs in the soil of the largest Pb-Zn smelting region, this study provides valuable scientific insights for subsequent soil remediation efforts.