Abstract: Preserving and increasing soil organic carbon pool is an effective natural way to mitigate climate change. However, the sensitivity of soil organic carbon to climate change in different ecosystems in mainland of China is still unclear. To investigate the response of soil organic carbon to climate change, it is important to classify it into particulate (POC) and mineral-associated organic carbon (MAOC) components. In this study, we assessed the spatial distributions of POC and MAOC in mainland of China and simulated their responses to future climate change using machine learning methods. The results showed that: 1) the stocks of soil organic carbon, POC, and MAOC in the top 20 cm soils of mainland China were 45.3, 20.7, and 24.6 Pg, respectively. 2) Soil organic carbon components were positively correlated with altitude and negatively correlated with air temperature. 3) Under the SSP585 scenario, the stocks of POC and MAOC in China would decrease by 4.80 and 2.13 Pg, from 2020 to 2100, respectively. The contents of soil organic carbon components were higher in Northeast China and Qinghai-Tibet Plateau, but lower in the Inner Mongolia Plateau, Sichuan Basin, North China, and Northwest China Plain. The sensitivity of POC to climate change was higher than MAOC. Climate warming would cause the greatest loss of soil organic carbon in the meadow ecosystem.

Key words: soil organic carbon, particulate organic carbon, mineral-associated organic carbon, spatial distribution, sensitivity