Temperature and mineralogical effects on decadally cycling mineral associated soil organic matter

Climate and parent material interact to form mineral assemblages that contribute to soil organic matter persistence across a range of time scales. Mineral associated soil organic matter (MAOM, the heavy soil component separated by density fractionation) generally contains more C and persists in soils longer than free or occluded light material. Yet while some MAOM persists for centuries, other forms of MAOM turnover on annual to decadal timescales. In order to predict the response of soil C pools to changes in inputs and decomposition rates under climate change we must be able to distinguish the relatively labile component of this mineral-associated soil C pool from the relatively passive component.

We collected samples in 2001, 2009, and 2019 from 9 sites along a combined gradient of parent material (granite, andesite, basalt) and mean annual temperature (MAT) (6.7°C, 9.1°C, 13.6°C). Mean annual precipitation was similar across all sites (80-130 mm yr^-1). We measured the radiocarbon (¹⁴C) values of bulk soils, respired CO₂, density fractions, and thermal fractions. We used selective dissolution and quantitative x-ray diffraction to determine mineral assemblages. We modeled turnover rates for bulk soil and MAOM using SoilR with C stocks and ¹⁴C data as constraints.

Using the difference between respired ¹⁴C and bulk ¹⁴C as a proxy for soil C protection, we observed a strong negative correlation with poorly crystalline mineral content at all time points, suggesting these secondary minerals play a key role in protecting soil C from decomposition. Poorly crystalline mineral content was greatest in the andesite soils, followed by basalt, then granite soils. Temperature also affected poorly crystalline mineral content, with greater abundances in sites with MAT of 9.1°C than in warmer or colder sites across lithologies.

Mineral assemblages were also related to the change in bulk ¹⁴C over time. Between 2001 and 2019, bulk ¹⁴C declined 4-5‰ yr^-1 faster (p < 0.05) in granite and basalt versus andesite soils at the 9.1°C MAT sites. Within the andesite soils, bulk ¹⁴C declined 6‰ yr^-1 faster at 13.6°C than 9.1°C (p < 0.05). Overall, slower rates of bulk ¹⁴C change were correlated with older mean C ages in the models. When compared within each MAT regime, our models revealed andesite soils to have older mean soil C ages than the basalt or granite soils. Respiration fluxes from these soils were more enriched in ¹⁴C than the fluxes from the basaltic or granitic soils, and were also enriched relative to the atmosphere. This indicates active decomposition of older decadally cycling soil C derived from mid-20^th century nuclear weapons testing in the andesitic soils but not in the basalt or granite soils.

Measurements of ¹⁴C in MAOM and associated thermal fractions (currently underway) will enable us to quantify the relative amounts of MAOM cycling at time scales relevant for improving near-term C budgets not only at our sites, but with implications for improving future models of soil C cycling at broader scales as well.