Soil carbon sequestration potential of cultivated lands and its controlling factors in China

Understanding soil organic carbon (SOC) stocks and carbon sequestration potential in cultivated lands can have significant benefit for mitigating climate change and emission reduction. However, there is currently a lack of spatially explicit information on this topic in China, and our understanding of the factors that influence both saturated SOC level (SOC_S) and soil organic carbon density (SOC_D) remains limited. This study predicted SOC_S and SOC_D of cultivated lands across mainland China based on point SOC measurements, and mapped its spatial distribution using environmental variables as predictors. Based on the differentiation between SOC_S and SOC_D, the soil organic carbon sequestration potentials (SOC_P) of cultivated land were calculated. Boosted regression trees (BRT), random forest (RF), and support vector machine (SVM) were evaluated as prediction models, and the RF model presented the best performance in predicting SOC_S and SOC_D based on 10-fold cross-validation. A total of 991 topsoil (0-20 cm) SOC measurements and 12 environmental variables explaining topography, climate, organism, soil properties, and human activity were used as predictors in the model. Both SOC_S and SOC_D suggested higher SOC levels in northeast China and lower levels in central China. The cultivated lands in China had the potential to sequester about 2.13 ± 0.96 kg m^-2 (3.25 Pg) SOC in the top 20 cm soil depth. Northeastern China had the largest SOC_P followed by Northern China, and Southwestern China had the lowest SOC_P. The primary environmental variables that affected the spatial variation of SOC_S were mean annual temperature, followed by clay content and normalized difference vegetation index (NDVI). The assessment and mapping of SOC_P in China's cultivated lands holds significance importance as it can provide valuable insights to policymakers and researchers about SOC_P, and aid in formulating climate change mitigation strategies.