Identifying driving factors of soil heavy metal at the mining area scale: Methods and practice.

ABSTRACT

Identifying driving factors is of great significance for understanding the mechanisms of soil pollution. In this study, a data processing method for driving factors was analyzed to explore the genesis of Arsenic (As) pollution in mining areas. The wind field that affects the atmospheric diffusion of pollutants was simulated using the standard k-ε model. Machine learning and GeoDetector methods were used to identify the primary driving factors. The results showed that the prediction performances of the three machine learning models were improved after data processing. The R2 values of random forest (RF), support vector machine, and artificial neural network increased from 0.45, 0.69, and 0.24 to 0.55, 0.76, and 0.52, respectively. The importance of wind increased from 20.85% to 26.22%. The importance of distance to the smelter plant decreased from 43.26% to 33.19% in the RF model. The wind's driving force (q value) increased from 0.057 to 0.235 in GeoDetector. The average value of historical atmospheric dust reached 534.98 mg/kg, indicating that atmospheric deposition was an important pathway for As pollution. The outcome of this study can provide a direction to clarify the mechanisms responsible for soil pollution at the mining area scale.