1. Prediction of Heavy Metal Pollution in Soil Based on SSA-XGBoost Model and 3D Geological Model.
- Author
-
Liu, Baoshun, Liu, Yingnan, and Zhang, Zijing
- Abstract
The problem of soil heavy metal pollution in decommissioned sites has become an environmental threat and challenge faced by countries around the world. Establishing a high-precision 3D model of contaminants is essential for pollution risk assessment and accurate monitoring of contaminated sites. In this study, 3D geological model and SSA-XGBoost model are proposed to predict heavy metal concentration in a site. These models can effectively improve the prediction accuracy of soil heavy metals, and the RMSE of the XGBoost model optimized by the SSA algorithm is reduced by 24.3%-34.3%. Compared with other machine learning models, the SSA-XGBoost model has optimal performance in improving the prediction accuracy of soil heavy metals. It is suitable for the areas with significant spatial heterogeneity of soil heavy metals. Using the SSA-XGBoost model and 3D geological model, the 3D spatial distribution characteristics of heavy metals in contaminated soil are determined. The concentration of soil pollutants ranked as As>Pb>Mo, and the overall pollution degree decreases gradually from top to bottom. The pollutants are mainly distributed in the production workshop area in the southwest of the site, and the miscellaneous fill layer is the main soil layer that needs to be remediated. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF