Spatial Distribution Changes and Factor Analysis of Topsoil Organic Carbon Across Different Forest Types on Hainan Island

Topsoil organic carbon (SOC, 0–20 cm) is crucial for terrestrial carbon stocks and the global carbon cycle. This study integrated field survey data, re-analysis climatic data, and remote sensing-derived environmental factors to examine SOC distribution and its drivers across forest types on Hainan Island using machine learning models and statistical analysis. The results showed that univariate analysis had limited explanatory power for forest SOC, with terrestrial plantations exhibiting significantly lower SOC than mangroves and natural forests. For mangroves, vapor pressure deficit (VPD) was the most influential factor, followed by precipitation (PRE), the normalized difference vegetation index (NDVI), and forest age; meanwhile, for terrestrial forests, VPD, altitude, PRE, and NDVI were vital drivers. The optimal models demonstrated relatively stronger predictive performance (R2 = 0.71 for mangroves; R2 = 0.81 for terrestrial forests). Mangroves showed higher average SOC (27.91 g/kg) than terrestrial forests (15.82 g/kg), while higher concentrations in the central–western region were attributed to natural terrestrial forests. This study reveals the spatial variation patterns of forest SOC and its environmental regulation mechanisms on Hainan Island, providing important references for forest carbon stock management and environmental protection.