Coal gangue-based soil amendments shift microbial nutrient limitation and support sandy-land ecosystem recovery

Severe C–N–P disequilibrium in sandy soils constrains microbial nutrient cycling and ecosystem recovery. Coal gangue-based soil amendments offer a scalable route to rebuild resource supply, yet their capacity to shift microbial C versus N/P limitation, as diagnosed by ecoenzymatic stoichiometry, remains unclear, particularly in relation to potential heavy-metal risks. Here, we evaluated five amendment rates, 0, 15, 30, 60, and 120 t·ha −1 , in a multi-year semiarid sandy-land field experiment using soil physicochemical analysis, microbial biomass stoichiometry, ecoenzymatic vector analysis, plant traits, and ecological-risk assessment. All treatments remained within a low overall ecological-risk category (RI < 150), although the highest rate strengthened risk signals. The 60 t·ha −1 rate increased soil organic carbon by 4.46–23.01%, total nitrogen by 4.13–15.14%, microbial biomass carbon by 11.99–59.40%, and Leymus chinensis yield by 27.69–134.04% relative to the other treatments. Vector analysis showed that carbon was not the dominant constraint, while amendment addition alleviated microbial phosphorus limitation; at 60 t·ha −1 , the inferred limitation shifted from P toward N. Partial least squares path modeling identified soil nutrients and microbial biomass as the main drivers of microbial resource limitation. Overall, moderate inputs (30–60 t·ha −1 ) provided the best risk–benefit trade-off, with 30 t·ha −1 favoring forage quality and 60 t·ha −1 maximizing biomass recovery. These findings define a practical, low-risk amendment window for improving degraded sandy agroecosystems and supporting circular use of coal-based solid wastes.