Peatlands store one-third of Earth's soil carbon, the stability of which is uncertain due to climate change-driven shifts in hydrology and vegetation, and consequent impacts on microbial communities that mediate decomposition. Peatland carbon cycling varies over steep physicochemical gradients characterizing vertical peat profiles. However, it is unclear how drought-mediated changes in plant functional groups (PFGs) and water table (WT) levels affect microbial communities at different depths. We combined a multiyear mesocosm experiment with community sequencing across a 70-cm depth gradient, to test the hypotheses that vascular PFGs (Ericaceae vs. sedges) and WT (high vs. low) structure peatland microbial communities in depth-dependent ways. Several key results emerged. (i) Both fungal and prokaryote (bacteria and archaea) community structure shifted with WT and PFG manipulation, but fungi were much more sensitive to PFG whereas prokaryotes were much more sensitive to WT. (ii) PFG effects were largely driven by Ericaceae, although sedge effects were evident in specific cases (e.g., methanotrophs). (iii) Treatment effects varied with depth: the influence of PFG was strongest in shallow peat (0-10, 10-20 cm), whereas WT effects were strongest at the surface and middle depths (0-10, 30-40 cm), and all treatment effects waned in the deepest peat (60-70 cm). Our results underline the depth-dependent and taxon-specific ways that plant communities and hydrologic variability shape peatland microbial communities, pointing to the importance of understanding how these factors integrate across soil profiles when examining peatland responses to climate change.

Rising temperatures are expected to reduce global soil carbon (C) stocks, driving a positive feedback to climate change1-3. However, the mechanisms underlying this prediction are not well understood, including how temperature affects microbial enzyme kinetics, growth efficiency (MGE), and turnover4,5. Here, in a laboratory study, we show that microbial turnover accelerates with warming and, along with enzyme kinetics, determines the response of microbial respiration to temperature change. In contrast, MGE, which is generally thought to decline with warming6-8, showed no temperature sensitivity. A microbial-enzyme model suggests that such temperature sensitive microbial turnover would promote soil C accumulation with warming, in contrast to reduced soil C predicted by traditional biogeochemical models. Furthermore, the effect of increased microbial turnover differs from the effects of reduced MGE, causing larger increases in soil C stocks. Our results demonstrate that the response of soil C to warming is affected by changes in microbial turnover. This control should be included in the next generation of models to improve prediction of soil C feedbacks to warming.