As the largest terrestrial carbon (C) pool, soil carbon dynamics could largely determine the direction and strength of terrestrial C-climate feedback at the global scale. Owing to limitations in the knowledge of key processes concerning SOC inputs and outputs, predictions of global SOC stocks across different terrestrial C models vary almost 6-fold. Furthermore, it remains unknown whether and how global SOC stocks exhibit significant changes over the past decades, largely due to the sparse data over many areas of the Earth and the lack of appropriate methodology. In this study, using a newly developed Global Soil Database coupled with machine learning algorithm, we evaluated the changes in the global SOC stocks with quantified uncertainties over the past three decades.
Results/Conclusions
In contrast to the long-term standing view that soil C may be lost under continuous climate warming, we found no significant change in overall global SOC stock at the upper 30 cm depth, but a notable decrease occurred in soils of high-latitudes with cold and dry climates in the Northern Hemisphere and an increase in those of some tropical regions with warm and humid climates. The relatively stable global SOC stock over the past three decades suggests that the potential of global soil C sequestration is likely less important than previously expected. Due to this point, biomass C sequestration may be a major contributor of global terrestrial C sinks for atmospheric CO2 in the future. We believe that our paper would contribute greatly to global carbon cycle science, and be of broad interest to any scientists and policy makers who are involved in the areas of climate change, ecology and soil science.