Iron-mediated mineralogical control of organic matter accumulation in tropical soils

Tropical forest soils contribute disproportionately to the poorly-characterized and persistent deep soil carbon (C) pool. These soils, highly-weathered and often extending one to two meters in depth, may contain an abundance of iron-(Fe) bearing mineral phases. Short-range-order (SRO) minerals are of particular interest due to their high reactive surface areas and capacity for soil C stabilization through sorption or co-precipitation. We hypothesized that SRO minerals might serve as primary contributors to soil C accumulation and storage in surface (0-20 cm) and subsurface (50-80 cm) soils of the Luquillo Critical Zone Observatory (LCZO) in northeast Puerto Rico. Oxisol and Inceptisol soils obtained from 20 quantitative soil pits, stratified across quartz-dominated granodiorite and clay-rich volcaniclastic parent materials, were subjected to selective dissolution procedures to extract Fe-C associations: sodium pyrophosphate (PP) to isolate colloidal or dispersable Fe, HCl-hydroxylamine (HH) and ammonium oxalate (AO) to isolate SRO Fe, and inorganic dithionite-HCl (DH) to isolate more crystalline pedogenic Fe. Pyrophosphate extraction of colloidal or dispersable Fe also extracted the greatest concentrations of soluble C across all samples. Dissolved molar C:Fe ratios > 1 observed solely in the PP extracts indicated the presence of organic-rich non-sorptive associations, the stability of which may have stronger control on accumulation of total soil C in these soils than those of extractable SRO and pedogenic Fe. Pedogenic and SRO Fe phases were the dominant extractable minerals in both soil types, at the surface and at depth, and notably, correlated well with extracted C. This suggests that these phases are strongly associated with a smaller, but substantial, fraction of total soil C. Direct observations of the limited extractability of soil C ( < 50% in surface soils) during selective dissolution of Fe and the lack of correlations between extractable Fe minerals and total soil C concentrations did not support the common hypothesis that SRO mineral phases provide a dominant mechanism for soil C accumulation. Instead, SRO phases may control only a fraction of total soil C, and non extractable Fe-bearing minerals and non-sorptive mechanisms may play more important roles than previously thought.