[Source Apportionment of Heavy Metals in Agricultural Soils of the Metal Smelting Park Based on PMF and ML Model]

Exploring various methods for tracing heavy metals in soil and quantitatively understanding the characteristics of soil heavy metal pollution is of great significance for soil pollution prevention and control. Here， the soil around a metal smelting park in Hunan Province is taken as the research object. We analyze five heavy metal elements in the soil including Cd， As， Pb， Cu， and Zn； evaluate the degree of soil pollution using the geo-accumulation index； and perform source apportionment using spatial distribution， PMF， and SVR-SHAP. The evaluation results of soil heavy metal pollution indicate that all five heavy metals in the surrounding soil of the park had pollution risks， among which Cd and As had the highest risk. Based on the PMF model， two pollution sources were identified through source apportionment， namely anthropogenic sources and natural sources， with contribution rates of 53.16% and 46.84% to soil heavy metals， respectively. The PMF model determined that soil Cu and Zn were mainly influenced by natural sources， and Cd， As， and Pb were mainly influenced by anthropogenic sources. The research and SVR-SHAP model analysis results indicate that soil Cd pollution mainly came from smelting activities and agricultural activities； As pollution mainly came from smelting activities and fossil fuel combustion； and Pb pollution mainly came from smelting activities， waste disposal， and agricultural activities. There was a significant risk of heavy metal pollution in the farmland soil around the metal smelting park. Combining two models can more reasonably identify the main sources of heavy metal pollution in the soil.