Same Binding Agent with Different Stabilization Capacities Regulates Soil Organic Carbon Across pH Gradients in Tropical and Subtropical Soils

Management of soil organic carbon (SOC) has attracted attention because of its critical role in maintaining soil health and mitigating global climate change. In tropical and subtropical soils, where SOC decomposes quickly, long-term SOC accumulation depends on stabilizing organic carbon (OC) as mineral-associated organic carbon (MAOC). Previous studies suggest that binding agents of MAOC vary with soil pH: clay content and exchangeable Ca/Mg are key controls in alkaline soils, whereas active Al/Fe (hydr)oxides in acidic soils. Meanwhile, MAOC is regulated not only by the abundance of these binding agents but also by their OC-stabilization capacity, both of which are likely pH-dependent yet remain poorly quantified. Thus, this study aims to (1) clarify the controls (e.g., precipitation, net primary production (NPP), binding agents) on MAOC, (2) quantify OC-stabilization capacity of main binding agents, with the investigation of their OC-stabilization mechanism across soil pH gradients.

A total 72 soil samples spanning strongly acidic (pH ≤ 5.5), weakly acidic (5.5 < pH ≤ 7), and alkaline (pH > 7) conditions were collected from Indonesia (n=14), Japan (n=12), Cameroon (n=16), Tanzania (n=13), and India (n=17). MAOC was quantified using density and particle-size fractionation (density > 1.7 g cm⁻³, particle size < 53 μm). Correlation analyses and structural equation models (SEM) were used to identify the primary controls on MAOC contents, incorporating NPP, precipitation, soil pH, exchangeable Ca/Mg, clay content and active Al/Fe as candidate explanatory variables for each pH class. Based on the unstandardized SEM coefficients, the OC-stabilization capacity of main binding agents was quantified. To assess the OC-stabilization mechanism, necromass C and non-necromass C (i.e., MAOC − necromass) were quantified and investigated relationship with the binding agents using correlation/regression analyses.

MAOC contents were 78 % of total SOC across all samples, with alkaline soils showing lower MAOC and active Al/Fe than strongly acidic and weakly acidic (42 vs. 181 vs. 238 cmol kg⁻¹, respectively) (4.6 vs. 9.2 vs. 17 cmol kg⁻¹, respectively). Correlation and SEM analysis identified active Al/Fe content as the primary control of MAOC content rather than clay content and NPP, across all pH classes. Exchangeable Ca/Mg showed no significant contribution even in alkaline conditions. These results indicate that lower MAOC in alkaline soils is due to the lower active Al/Fe content. Furthermore, the SOC stabilization capacity of active Al/Fe was also 28% lower in alkaline soil than in weakly acidic and strongly acidic. In strongly and weakly acidic soils, active Al/Fe was positively correlated with both necromass and non-necromass C, whereas active Al/Fe was positively correlated with non-necromass C but showed no relationship with necromass C in alkaline soils. These results suggest that in alkaline soils, active Al/Fe stabilizes non-necromass C but not necromass C, causing the low OC-stabilization capacity. Thus, in tropical and subtropical regions, active Al/Fe are the primary control on MAOC binding in all pH class, and their low abundance and low OC-stabilization capacity in alkaline soils partly explain the low MAOC and SOC contents.