Contrasting Soil Organic Carbon Stocks and Bacterial Communities Among Croplands Reclaimed for 25 Years in a High-Groundwater-Table Coal-Mining Subsidence Area

Understanding differences in soil organic carbon (SOC) stocks among reclamation modes is essential for evaluating ecological recovery in reclaimed croplands. However, the factors associated with these differences remain poorly understood in coal-mining subsidence areas with high groundwater tables. We investigated croplands reclaimed 25 years before sampling under three reclamation modes in eastern China: coal gangue backfilling reclamation (CGBR), fly ash backfilling reclamation (FABR), and deep excavation–shallow filling reclamation (DESFR), with unaffected cropland serving as the reference (CK). Soil samples were collected from three layers (0–15, 15–30, and 30–45 cm) to assess SOC stocks, labile carbon fractions, water-stable aggregates, hydrolytic enzyme activities, and bacterial communities. Correlation analysis, random forest modeling, and exploratory partial least squares path modeling were used to examine associations between SOC stocks and soil properties. Soil carbon pools and bacterial communities varied among reclamation modes and soil layers. SOC stocks in the 0–45 cm profile were 70.35, 65.05, and 51.78 Mg C ha−1 under CGBR, FABR, and DESFR, respectively. Among the evaluated reclamation modes, CGBR had the highest mean SOC stock, equivalent to approximately 98.2% of the reference cropland (CK) level and 8.15% and 35.86% higher than those under FABR and DESFR, respectively. DESFR had the lowest SOC stock despite relatively high aggregate stability. Exploratory analyses indicated associations of SOC stocks with carbon and nitrogen availability, hydrolytic enzyme activities, bacterial diversity, and aggregate properties. These findings support considering soil ecological recovery alongside engineering performance when evaluating reclamation effectiveness in the studied subsidence area. Long-term monitoring and multisite observations are needed to characterize SOC recovery trajectories and assess the broader applicability of these findings.