Assessing eco-environmental risks in coal waste dump watersheds: An integrated spatial multi-criteria approach using RS, GIS, and the DPSIR framework

Coal-based solid waste dumps are among the most persistent and poorly quantified sources of ecological risk in coal-dependent regions, where steep terrain, fragile soils, and hydrologically sensitive landscapes exacerbate pollutant transport and ecological degradation. However, current environmental assessments rarely capture the fine-scale spatial heterogeneity of these risks or integrate them with socio-ecological dynamics, thereby constraining the development of targeted restoration and management strategies. Here we construct a spatial multi-criteria ecological risk assessment framework that integrates multi-temporal high-resolution remote sensing, geographic information systems, and the driver, pressure, state, impact, response (DPSIR) model. Applied to a representative gully-type coal-based solid waste watershed in Shanxi Province, China, the framework synthesizes 18 standardized indicators spanning climatic, anthropogenic, ecological, and governance dimensions. Indicator weights were determined using a hybrid analytic hierarchy process entropy weight method, and an eco-environmental risk index (EERI) was computed at 30 m resolution. Between 2010 and 2025, dump areas declined from 3.79 to 1.90 km², while reclamation areas expanded from 0.43 to 2.35 km². The resulting EERI indicated that high-risk watersheds are concentrated in areas with dense dump distributions, steep relief, and low vegetation cover. Grey Relational Analysis and Manhattan distance analysis consistently identified dump density and runoff volume as the primary drivers of spatial risk heterogeneity. By integrating fine-scale environmental data with socio-ecological drivers, this approach provides a transferable, spatially explicit tool for identifying and managing ecological risks at the watershed scale in coal mining areas, enabling precise intervention and informed decision-making for environmental risk mitigation.