Fast-decaying plant litter enhances soil carbon in temperate forests but not through microbial physiological traits

Conceptual and empirical advances in soil biogeochemistry have challenged long-held assumptions about the role of soil micro-organisms in soil organic carbon (SOC) dynamics; yet, rigorous tests of emerging concepts remain sparse. Recent hypotheses suggest that microbial necromass production links plant inputs to SOC accumulation, with high-quality (i.e., rapidly decomposing) plant litter promoting microbial carbon use efficiency, growth, and turnover leading to more mineral stabilization of necromass. We test this hypothesis experimentally and with observations across six eastern US forests, using stable isotopes to measure microbial traits and SOC dynamics. Here we show, in both studies, that microbial growth, efficiency, and turnover are negatively (not positively) related to mineral-associated SOC. In the experiment, stimulation of microbial growth by high-quality litter enhances SOC decomposition, offsetting the positive effect of litter quality on SOC stabilization. We suggest that microbial necromass production is not the primary driver of SOC persistence in temperate forests. Factors such as microbial necromass origin, alternative SOC formation pathways, priming effects, and soil abiotic properties can strongly decouple microbial growth, efficiency, and turnover from mineral-associated SOC.


Table S4
Mean values (± 1 SD; n = 9) for microbial biomass carbon (MBC), microbial growth rate (MGR), microbial carbon use efficiency (CUE), microbial turnover rate (MTR), and microbial necromass concentrations at the study sites. Necromass is reported on a "per mass soil" and "per mass soil C" basis and fungal-to-bacterial ratios (F:B) are also shown. See Table S3

Table S5
Climatic and mean edaphic properties (± 1 SD; n = 9) at the study sites including mean annual temperature (MAT) and precipitation (MAP), oxalate-extractable iron (Feox) and aluminum (Alox), total carbon (Tot-C) and mineral-associated organic carbon (MA-SOC) concentrations, and the proportion of SOC stored in MA-SOC (MA-SOC/Tot-C). See Table S3 Table S1 for abbreviations.  Indirect effects of litter quality are mediated through the microbial physiological trait index. Numbers above and below paths represent standardized coefficients during early-and intermediate-stage decomposition, respectively, with significance levels indicated (*p < 0.1, **p < 0.05, and ***p < 0.01). The early and intermediate incubations were harvested on days 30 and 185, respectively, for soil C and microbial physiological traits were measured on days 15 and 100 (i.e. the approximate midpoints). Thickness and color of lines correspond to coefficient magnitude and direction, respectively. Total and indirect effects of litter quality on soil C formation are also summarized with standardized coefficients.

Fig. S6
Path analysis showing the direct and indirect effects of the litter quality index (Litter quality) on soil-derived (i.e. pre-existing) particulate SOC (Soil-derived Particulate SOC). Indirect effects of litter quality are mediated through the microbial physiological trait index. Numbers above and below paths represent standardized coefficients during early-and intermediate-stage decomposition, respectively, with significance levels indicated (*p < 0.1, **p < 0.05, and ***p < 0.01). Thickness and color of lines correspond to coefficient magnitude and direction, respectively. Total and indirect effects of litter quality on soil C formation are also summarized with standardized coefficients.

Fig. S7
Linear mixed model coefficients relating the microbial physiological trait index to soil C:N, fine root biomass (Root mass), soil pH, potential net nitrogen mineralization (Nmin), the litter quality index (Litter quality), and extractable dissolved organic carbon (DOC). Plot shows standard error (n = 54; inner bold lines) and 95 % confidence intervals (outer lines). Coefficients were centered and standardized to show the relative importance of each predictor despite the different scales on which the variables were measured.

Fig. S8
Bivariate relationship between soil C:N and the soil microbial physiological trait index across 54 plots and six forests. Data are fit with a second-order polynomial (Adjusted R 2 = 0.21; P < 0.01). Site abbreviations are defined in Table S3.  Fig S1D), fine root biomass, soil pH, oxalateextractable iron (Fe-ox), and soil clay content. Plot shows standard error (n = 54; inner bold lines) and 95 % confidence intervals (outer lines).  Fig S1D), fine root biomass, soil pH, oxalate-extractable iron (Fe-ox), and soil clay content. Plot shows standard error (n = 54; inner bold lines) and 95 % confidence intervals (outer lines).