[Source Apportionment of Heavy Metals in Soils Based on Machine Learning Algorithms and Receptor Model]

Area of Science:

• Environmental Science
• Geochemistry
• Data Science

Context:

• Coal gangue heaps are significant sources of soil heavy metal pollution.
• Understanding heavy metal source apportionment is crucial for environmental risk assessment.
• Previous studies often lack comprehensive analysis of multiple pollution sources.

Purpose:

• To analyze the source apportionment and influence factors of heavy metals in soils surrounding a coal gangue heap.
• To compare the effectiveness of machine learning algorithms (Decision Tree, Random Forest, Support Vector Machine) and the Absolute Principal Component Scores-Multiple Linear Regression (APCS-MLR) receptor model.
• To identify key driving factors of human activities influencing heavy metal contamination.

Summary:

• Heavy metal concentrations (Cd, Hg, As, Pb, Cr, Cu, Ni, Zn) in surface and profile soils were analyzed.
• Random Forest model demonstrated superior performance (R² values ranging from 0.528 to 0.853) compared to Decision Tree and Support Vector Machine.
• Key human activity drivers identified include the number of coal gangue units, vertical height difference, and distance from the heap.
• APCS-MLR model attributed soil pollution to natural (42.5%), mining (37.1%), and mixed sources (20.4%).

Impact:

• The study provides a more comprehensive and accurate assessment of heavy metal sources near coal gangue heaps.
• Integration of machine learning and receptor models enhances the reliability of source apportionment results.
• Findings can inform targeted remediation strategies and environmental management policies for mining-affected areas.