A Spatiotemporal Framework for Quantifying and Projecting Soil Organic Carbon in Grasslands of China

Abstract

Accurate estimation of grassland soil organic carbon (SOC) stocks is essential for understanding carbon dynamics, yet estimates of soil organic carbon density (SOCD) in China remain uncertain due to differences in data sources and methods. To address this issue, we developed a spatiotemporal model of grassland SOCD using a SOCD database, five categories of drivers, and a random forest (RF) model. The RF model estimated grassland SOCD in the upper 1 m with an R2 of 0.62. From 2001 to 2020, mean SOCD increased from 9.25 to 11.64 kg C m−2, and total SOC stocks increased from 42.33 ± 0.47 to 50.71 ± 0.48 Pg C. In the grasslands without land-use change, SOCD accumulated at an average rate of 145.8 g C m−2 year−1. Climate change is projected to affect grassland SOC stock, with both warming and drying expected to reduce SOC stocks. Under a low-carbon-emission scenario, China’s grassland SOC stocks are projected to remain relatively stable through 2100. However, under a high-carbon-emission scenario, SOC stocks are expected to decline by approximately 11.9% by 2100. By improving the sample base, estimation method, and grassland distribution data, we provide a more robust assessment of the spatial and temporal dynamics of grassland SOC stocks in China.