Sources of soil carbon loss during soil density fractionation: Laboratory loss or seasonally variable soluble pools?

Abstract To model global C budgets more accurately, we must better understand the dynamics and turnover of the many functional C pools that exist in soil. While soil density fractionation (SDF) is widely used to separate soil C pools based on the degree of stabilization by soil minerals, several studies have noted substantial losses of soil mass and soil C following SDF. As the source of these losses is unknown, they may lead to erroneous conclusions about SOM dynamics and inaccuracies in models of soil processes and global C cycling. For example, lab handling techniques such as air-drying soil could have an impact on soluble C losses. Alternatively, the observed C losses could represent pools of potentially soluble C that can vary by ecosystem and season, and thus might represent ephemeral, not well stabilized, and generally overlooked pools of soil C that are missing from SOM budgets and models. For example, in summer-dry climates such as the Pacific Northwest, soluble organic products of decomposition could accumulate over the summer when temperatures and microbial activities are high and precipitation is low, and subsequently leach during the fall and winter with precipitation. To address these divergent possibilities, soils were collected seasonally from a summer-dry forest in Oregon, and subsamples were subjected to 4 different laboratory handling procedures prior to fractionation: 1) air-drying and 2) oven-drying to simulate common laboratory drying techniques; 3) leaching to evaluate if potentially soluble C represented a pool that could be removed with simulated precipitation; and 4) immediate fractionation of fresh soil to determine if drying of soil caused artifacts in pools of potentially soluble soil C. Contrary to initial hypotheses, there were no seasonal trends to soluble DOC pools or total C loss during fractionation. Average mass loss during fractionation was 6% of initial dry weight and total C loss was 9% of total soil C. Soluble losses represented only 9% of total soil C loss. Particulate, or non-soluble mass loss was dominated by the high C:N free particulate organic matter, or LF. Thus, most C loss in this system was due to laboratory losses and that loss was biased towards an underestimation of LF matter in soil. DOC was a small enough component of loss that soil preparation prior to SDF was not significant. These results imply that in temperate soils, even in seasonally extreme ecosystems, relative proportions of mineral-associated organic matter and free particulate organic matter are seasonally robust, air drying of soil does not introduce error, and that immediate fractionation after field collection is not critical.