Planting trees, i.e., afforestation, has been widely recommended as a nature-based climate solution, primarily due to its carbon sequestration potential. However, it remains less clear how large this potential carbon sink is. Furthermore, it is an even more challenging question where we should plant trees for maximal carbon sink potential. Importantly, the terrestrial ecosystem carbon pool includes both organic and inorganic carbons, with the latter accounting for about one third of the pool. While traditionally the inorganic part (soil inorganic carbon, SIC) is usually considered stable, recent local-scale studies showed considerable SIC responses to agricultural management and land use changes. Hence, large-scale assessments of afforestation impacts on carbon cycle need to consider both organic carbon and SIC.
As organic carbon in tree biomass is relatively easier to estimate with biomass growth functions, here we focus on soil organic carbon (SOC) and SIC. Using field data from 619 afforested plots and 163 control plots across semi-arid northern China, we provide a comprehensive assessment on afforestation-induced changes in both SOC and SIC. We find context-dependent effects of afforestation on both SOC and SIC. Afforestation increases SOC density (SOCD) in C-poor soils but decreases SOCD in C-rich soils, especially in deeper soil. Similarly, afforestation increases SIC in acidic soils, while decreases SIC in alkaline soils. Fitting a linear mixed model and a structure equation model further, we find that afforestation-induced soil pH change is the most significant factor regulating SIC responses. Our findings suggest that previous studies may overestimate the carbon sink benefit of afforestation, and provide a data-based understanding on the comprehensive soil carbon dynamics following afforestation and its underlying mechanisms. Our work highlights the importance of including pre-afforestation soil properties in models of soil carbon dynamics and carbon sink projections.