Edaphic controls of soil organic carbon in tropical agricultural landscapes

Predicting soil organic carbon (SOC) is problematic in tropical soils because mechanisms of SOC (de)stabilization are not resolved. We aimed to identify such storage mechanisms in a tropical soil landscape constrained by 100 years of similar soil inputs and agricultural disturbance under the production of sugarcane, a C(4) grass and bioenergy feedstock. We measured soil physicochemical parameters, SOC concentration, and SOC dynamics by soil horizon to one meter to identify soil parameters that can predict SOC outcomes. Applying correlative analyses, linear mixed model (LMM) regression, model selection by AICc, and hierarchical clustering we found that slow moving SOC was related to many soil parameters, while the fastest moving SOC was only related to soil surface charge. Our models explained 78–79%, 51–57%, 7–8% of variance in SOC concentration, slow pool decay, and fast pool decay, respectively. Top SOC predictors were roots, the ratio of organo-complexed iron (Fe) to aluminum (Al), water stable aggregates (WS(agg)), and cation exchange capacity (CEC). Using hierarchical clustering we also assessed SOC predictors across gradients of depth and rainfall with strong reductions in Roots, SOC, and slow pool decay associated with increasing depth, while increased rainfall was associated with increased Clay and WS(agg) and reduced CEC in surface soils. Increased negative surface charge, water stable aggregation, organo-Fe complexation, and root inputs were key SOC protection mechanisms despite high soil disturbance. Further development of these relationships is expected to improve understanding of SOC storage mechanisms and outcomes in similar tropical agricultural soils globally.