The MILLENNIA peat cohort model: predicting past, present and future soil carbon budgets and fluxes under changing climates in peatlands

ABSTRACT: Globally, considerable carbon (C) is stored in soils, particularly in peatlands. These stores play a potentially significant role in atmospheric C-cycle feedbacks, and thus need to be accounted for in global models. However, present global soil models do not accurately represent peat C-stocks and -dynamics; thus, their climate–soil C feedback predictions are questionable. A major shortcoming of current models that are based on the decomposition of soil C pools is the lack of representation of long-term (non-equilibrium) soil organic carbon (SOC) accumulation, as peat cohorts, with cohort age information. Whereas C-pool models are commonly ‘spun up’ to equilibrium over several hundred years using an average climate, in nature, soils actually evolve over many thousands of years with associated changes in litter amounts and quality, which affect SOC accumulation, and hence peat formation. Secondly, peat soils have a unique hydrology, and changes in the water table depth (WTD) of peat are important in regulating SOC turnover, yet current non-cohort C pool models fail to include such dynamic hydrological processes. We have developed an improved peat age-cohort model called MILLENNIA, with a variable WTD driving C-dynamics during Holocene peat accumulation, allowing validation with peat age data and the testing of a realistic WTD-driven peat SOC stock response to climate-change scenarios. Model C-dynamics showed particular sensitivity to water table dynamics through precipitation and runoff, as well as to litter quality and decomposition rates. We show that predicted SOC accumulation and peat ages compare well with observations from a UK peatland site, which is currently (on average) a weak net C source with strong climate sensitivity.