ABSTRACT Soil organic carbon (SOC) and total nitrogen (TN) are fundamental indicators of soil fertility and ecosystem stability, especially in grasslands that are highly sensitive to climate change and grazing disturbances. In this study, we evaluated the responses of SOC and TN dynamics to future climate factors (temperature and precipitation) and grazing intensities across the Gannan grasslands. The DayCent model was used to simulate SOC and TN stocks for 2025–2075 under four CMIP6 scenarios (SSP126, SSP245, SSP370, and SSP585) and three grazing intensities (low, medium, and high). To capture nonlinear patterns and interaction effects, we coupled DayCent outputs with generalized additive models (GAM) to characterize smooth response trends and with XGBoost plus SHAP to quantify variable importance and interaction contributions. Beyond confirming climate as the primary control, we identified management‐relevant thresholds: SOC and TN losses accelerate once annual maximum temperature exceeds ~8.5°C, annual mean temperature exceeds ~10°C, or annual precipitation departs from the ~1200–1400 mm optimum, with the steepest declines under high grazing. SHAP further pinpointed two governing interaction pathways—maximum temperature × precipitation and grazing × mean temperature—providing actionable targets for adaptive grazing to dampen climate sensitivity and stabilize soil C and N under future warming.
Yang et al. (Wed,) studied this question.