Driving factors and spatial differentiation of soil organic carbon density in the Three-North Shelterbelt Engineering Region of China.

Against the backdrop of the Three-North Shelterbelt Engineering of China (TNSEC), identifying the drivers of soil organic carbon and its spatial variation is of great significance for supporting the national dual carbon strategy. The TNSEC has already carried out large-scale afforestation projects in many regions, including agricultural shelterbelts, water conservation forests, soil and water conservation forests, windbreak and sand fixation forests, and pasture protection forests. Based on 958 observed soil organic carbon density (SOCD) data points, this study combines structural equation modeling (SEM) with interpretable machine learning models (RF-SHAP) and validates the results using an independent dataset, aiming to reveal the driving factors and evolution patterns of SOCD in the Three-North Shelterbelt Engineering Region (TNSER) since 2000. The results indicated that SOCD in the Northeast China and North China was primarily driven by mean annual temperature (MAT), while in the Northwest China it was driven by elevation, with inflection points at 5.96°C, 4.37°C, and 1541 m, respectively. In addition, both the habitat quality index (HQI) and vegetation indices (VIs) exerted significant influences. From 2000 to 2023, soil organic carbon stock increased from 26.05 ± 4.74 Pg C to 26.91 ± 4.32 Pg C, showing an overall upward trend, and it may continue to maintain this growth trend in the future. SOCD exhibited significant spatial variation and an asymmetric relationship with HQI. The findings validate the positive contribution of large-scale ecological engineering projects to the national dual carbon goals and provide a basis for ecosystem carbon management.