Divergent environmental constraints shape the spatial patterns of annual gross primary productivity in China's terrestrial ecosystems.

Introduction: Annual Gross Primary Productivity (AGPP) is a key indicator of terrestrial ecosystem carbon sequestration. However, its spatial variability and biome-specific controls remain poorly understood. Methods: Here, we analyzed AGPP across forests, grasslands, croplands, and wetlands in China using data from 166 eddy covariance flux sites. We applied SHapley Additive exPlanations (SHAP) and structural equation modeling (SEM) to quantify the relative importance of environmental variables and to disentangle their direct and indirect effects on AGPP. Results: Our results revealed clear productivity patterns: forests exhibited the highest AGPP, followed by wetlands, croplands, and grasslands. In forests, AGPP was positively associated with temperature and precipitation but negatively related to high vapor pressure deficit (VPD). In grasslands, precipitation enhanced AGPP both directly and indirectly via vegetation structure (leaf area index), whereas excessive radiation constrained productivity. High soil organic carbon (SOC) was associated with lower AGPP in wetlands and croplands, likely reflecting nutrient immobilization under waterlogged conditions and the impacts of intensive agricultural management. Discussion: These findings highlight biome-specific responses of AGPP to climatic conditions, soil properties, and vegetation structure, and provide insights for improving carbon cycle models and guiding targeted ecosystem management.