Assessment of Land Reclamation Effectiveness and Driving Mechanisms in Typical Metal Mining Areas in China Using Remote Sensing and Explainable Machine Learning

ABSTRACT Although large‐scale land reclamation (LR) has been implemented in open‐pit metal mining areas, long‐term ecological restoration effects remain unsystematically revealed due to insufficient continuous monitoring, hindering the accurate achievement of mining area ecosystem resilience and carbon neutrality goals. This study proposed an Iron Mine Eco‐Quality Index (IM‐EQI) to better reflect the Malan Iron Mine's ecological quality (1990–2024), with multiple methods exploring IM‐EQI's long‐term temporal evolution, spatial pattern changes, and nonlinear driving mechanisms. The results illustrated that: (1) IM‐EQI had high consistency with the Remote Sensing Ecological Index (RSEI) and the Mine‐Specific Eco‐Environment Index (MSEEI) ( R 2 = 0.90, p < 0.01) and better characterized the information richness of the iron mining ecosystem; (2) After 2010 reclamation, most areas' ecological environment quality (EEQ) improved sustainably (61.45% mild/significant improvement) with continuous H–H clustering; (3) XGBoost‐SHAP revealed nonlinear relationships/threshold effects between driving factors and IM‐EQI. Single‐factor importance and inter‐factor interaction analyses consistently showed land use dominated IM‐EQI spatial distribution—mining land exacerbated ecological risks and reduced land sustainability, while land use's synergies with precipitation/temperature amplified open‐pit mining's negative ecological impacts. This study's findings provide quantitative support for targeted metal mine reclamation optimization and long‐term ecological management and offer practical paradigms references for “ecology first, green development” in resource‐based regions.