Depth-dependent changes in the amount and stability of newly formed mineral-associated organic matter in temperate soils

The largest and most persistent portion of soil organic carbon (OC) is stored with mineral-associated organic matter (MAOM). Advancing our understanding of the processes and drivers involved in the formation and stabilization of MAOM  is thus key to improving predictions of the mitigation potential of soils and their response to global change. While it is known that MAOM content and stability change with soil depth, our mechanistic understanding of the factors driving these changes is still incomplete. We expected that mineral type, as a determinant of a soil’s OC sorption and stabilization capability, and vegetation cover, as a modifier of organic inputs into the soil, would shape depth patterns in the content and stability of MAOM. We exposed pristine goethite (iron oxide) or illite (phyllosilicate clay) for five years in 24 forests (15 deciduous and 9 coniferous) and 23 grasslands (14 on mineral soils and 9 on organic soils) in three regions across Germany. Minerals were placed at 5 cm and 30 cm depths at all 47 sites, and additionally at 0 cm depth (i.e., at the boundary between the organic surface layers and the mineral soil) at forest sites. After recovery, the OC content of the field-exposed minerals was determined by dry combustion and the composition of the newly formed MAOM was determined using X-ray photoelectron spectroscopy (XPS). Stability of MAOM was indicated by the mineralizability of OM associated with the field-exposed mineral samples, by measuring the release of carbon dioxide per gram OC in laboratory incubations. Results show, on average, three times more MAOM formation in topsoils than subsoils at sites on mineral soils but we did not find any effect of depth at sites on organic soils. Changes in MAOM formation across depth reflected organic inputs from the overlying soil and were more substantial for coniferous forests than other vegetation covers, especially directly beneath the organic surface layer. We observed more substantial depth changes in MAOM content for goethite than illite. There was a consistent decrease in MAOM mineralizability (i.e., increase in stability) with soil depth for illite but not for goethite. Interestingly, the mineralizability of goethite-associated OM from forests was higher in subsoils than in topsoils. Mineralizability of MAOM was negatively correlated with the share of highly oxidized compounds (i.e., carboxylic/carbonyl C) across depth for goethite but not illite, suggesting different mechanisms underlie the depth-dependent changes in the stability of OM associated with the two minerals.  Overall, our study evidence that depth patterns in the amount and stability of MAOM in soils are shaped by mineral type and vegetation cover.  This insight can help guide process-oriented grouping of soils for improved prediction of soil OC content and stability across depth at larger scales.