Order from disorder: do soil organic matter composition and turnover co-vary with iron phase crystallinity?

Soil organic matter (SOM) often increases with the abundance of short-range-ordered iron (SRO Fe) mineral phases at local to global scales, implying a protective role for SRO Fe. However, less is known about how Fe phase composition and crystal order relate to SOM composition and turnover, which could be linked to redox alteration of Fe phases. We tested the hypothesis that the composition and turnover of mineral-associated SOM co-varied with Fe phase crystallinity and abundance across a well-characterized catena in the Luquillo Experimental Forest, Puerto Rico, using dense fractions from 30 A and B horizon soil samples. The δ13C and δ15N values of dense fractions were strongly and positively correlated (R2 = 0.75), indicating microbial transformation of plant residues with lower δ13C and δ15N values. However, comparisons of dense fraction isotope ratios with roots and particulate matter suggested a greater contribution of plant versus microbial biomass to dense fraction SOM in valleys than ridges. Similarly, diffuse reflectance infrared Fourier transform spectroscopy indicated that SOM functional groups varied significantly along the catena. These trends in dense fraction SOM composition, as well as ∆14C values indicative of turnover rates, were significantly related to Fe phase crystallinity and abundance quantified with selective extractions. Mössbauer spectroscopy conducted on independent bulk soil samples indicated that nanoscale ordered Fe oxyhydroxide phases (nano-goethite, ferrihydrite, and/or very-SRO Fe with high substitutions) dominated (66–94%) total Fe at all positions and depths, with minor additional contributions from hematite, silicate and adsorbed FeII, and ilmenite. An additional phase that could represent organic-FeIII complexes or aluminosilicate-bearing FeIII was most abundant in valley soils (17–26% of total Fe). Overall, dense fraction samples with increasingly disordered Fe phases were significantly associated with increasingly plant-derived and faster-cycling SOM, while samples with relatively more-crystalline Fe phases tended towards slower-cycling SOM with a greater microbial component. Our data suggest that counter to prevailing thought, increased SRO Fe phase abundance in dynamic redox environments could facilitate transient accumulation of litter derivatives while not necessarily promoting long-term C stabilization.