Permafrost, soil that remains below 0 ?C for two or more years, currently stores more than a fourth of global soil carbon. A warming climate makes this carbon increasingly vulnerable to decomposition and release into the atmosphere in the form of greenhouse gases. The resulting climate feedback can be estimated using Earth system models (ESMs), but the high complexity and computational cost of these models make it challenging to use them for estimating uncertainty, exploring novel scenarios, and coupling with other models. We have added a representation of permafrost to the simple, open-source global carbon-climate model Hector, calibrated to be consistent with both historical data and 21st century ESM projections of permafrost thaw. We include permafrost as a separate land carbon pool that becomes available for decomposition into both CH4 and CO2 once thawed; the thaw rate is controlled by region-specific air temperature increases from a pre-industrial baseline. We found that by 2100 thawed permafrost carbon emissions increased Hector’s atmospheric CO2 concentration by 10-15% and the atmospheric CH4 concentration by 10-20%, depending on the future scenario. This resulted in around 0.5 °C of additional warming over the 21st century. The fraction of thawed permafrost carbon available for decomposition was the most significant parameter controlling the end-of-century temperature change and atmospheric CO2 concentration in the model and became increasingly significant over even longer timescales. The addition of permafrost in Hector provides a basis for the exploration of a suite of science questions, as Hector can be cheaply run over a wide range of parameter values to explore uncertainty and easily coupled with integrated assessment models to explore the economic consequences of warming from this feedback.
Published: September 25, 2021
Citation
Woodard D.L., A.N. Shiklomanov, B.S. Kravitz, C.A. Hartin, and B. Bond-Lamberty. 2021.A permafrost implementation in the simple carbon–climate model Hector v.2.3pf.Geoscientific Model Development 14, no. 7:4751-4767.PNNL-SA-158216.doi:10.5194/gmd-14-4751-2021