Optimizing vegetation restoration strategies based on pre-mining soil ecosystem quality in coal mining areas

Mining-induced subsidence areas require restoration strategies that support long-term ecosystem development. A key challenge is rebuilding self-sustaining soil–plant feedbacks. Vegetation configuration is critical because unsuitable assemblages may constrain microbial recovery, enzyme-mediated nutrient cycling, and the persistence of soil improvement below the surface layer. Based on 840 pre-mining vegetation–soil sampling plots, we examined the associations of eight vegetation configurations with soil physicochemical properties, enzyme activities, and microbial attributes in the 0–10 cm surface and 30–40 cm subsurface layers， aiming to identify vegetation configurations capable of sustaining soil quality improvement and supporting ecological restoration. Shrub-dominated mixed forests (VT6 and VT7) had a notably lower relative decrease from surface to subsurface rate of enzyme activity (48–52%) compared to pure shrub systems (VT2: 71%). Microbial community α-diversity exhibited a homogenization pattern, while fungal α-diversity showed significant positive correlations with nitrification enzyme activity and dehydrogenase activity. By integrating the Surface-Subsurface Response Index (SSRI) and microbial community α-diversity metrics through correlation analysis and principal component analysis, a modified soil quality index (SQI′) was developed. The large-sample dataset was then used to evaluate soil quality stability and the probability of high-quality soil occurrence among vegetation types, identifying VT6 as the optimal mixed-forest restoration configuration. The LEfSe analysis identified Pseudomonas as a genus-level bacterial indicator signal enriched in VT6. Structural equation modeling revealed that dehydrogenase activity (total effect: 0.36) and maximum water-holding capacity (0.23) were the most influential factors driving microbial diversity. Nitrification enzymes, urease, and dehydrogenase are commonly used as key indicators for assessing soil ecological functions. These findings provide a scientific basis for selecting effective vegetation types in ecological restoration efforts targeting arid, mining-induced subsidence areas.