Arctic soils are among the largest stores of organic carbon globally. Because rapid arctic climate warming is predicted to promote decomposition while also stimulating plant growth and increasing woody plant dominance, there is great interest in developing mechanistic descriptions of the system's carbon dynamics as it responds to warming. I used the Toolik Long Term Ecological Research site moist acidic tundra greenhouse experiment, which was established in 1989 and is the longest continuously running tundra warming experiment in existence, to explore the consequences of long-term warming on tundra biogeochemical dynamics. Warming increased plant biomass and woody plant dominance; however, the tundra soil system proved resistant to carbon loss, which may reflect counterbalancing abiotic and biotic system feedbacks that developed over the course of the treatment.

Surprisingly, the greatest biogeochemical changes in the greenhouse experiment were detected at depth. Warming increased the mineral horizon soil carbon stock, thaw depth, decomposer activity, and trophic complexity. By stratifying sampling across season and depth, I found that decomposer activity in the greenhouse deep organic and mineral horizon soils were stimulated during the late winter. This effect was not apparent during the summer and was reversed in the early winter. Therefore, a shift from direct summer warming to indirect winter warming may profoundly affect carbon storage in deeper tundra soils by destabilizing the large permafrost mineral soil carbon pools. Warming simulations using a stoichiometrically coupled, acclimating microbe-plant-soil (SCAMPS) model further supports the possibility that winter warming may more strongly promote tundra soil carbon mineralization than summer warming.

Over two decades of experimental summer warming increased ecosystem tundra carbon stocks, because woody plant biomass increased while soil C stocks remained stable. This suggests that warming-driven acceleration of carbon loss from tundra systems may be a transient response that is stabilized over time by a restructuring of the plant and soil communities. At an annual scale, warming-driven, seasonally-linked changes in extracellular enzyme substrate and product availability may create stabilizing feedbacks to warming-driven increases in tundra decomposer activity. As such, these stoichiometrically-constrained feedbacks may ultimately limit warming-driven soil C loss in arctic tundra systems.