Implications of permafrost carbon cycle feedbacks for TCRE: evidence from Earth system modeling

TCRE – the linearity between global warming and cumulative anthropogenic CO2 emissions – underpins the concept of remaining carbon budgets and is critical for designing mitigation policies in line with the Paris Agreement. The future response of carbon sinks to anthropogenic perturbations is a major source of uncertainty in estimates of future TCRE. In particular, the strong Arctic warming is expected to lead to permafrost thaw, exposing the large amounts of soil organic carbon stored in permafrost to decomposition, and eventually releasing CO2 and CH4 to the atmosphere in a positive climate-carbon feedback. On the other hand, CO2 fertilisation and permafrost nitrogen release are likely to enhance vegetation carbon uptake by counteracting negative feedbacks. However, both the amplitude and the timing of the resulting future net carbon balance in permafrost regions remain highly uncertain. In particular, previous studies, using either land surface or intermediate complexity models, have shown no consensus on the strength of the nitrogen-mediated feedback. In addition, future TCRE estimates are based on (fully coupled) Earth system model (ESM) projections. However, only two ESMs in the CMIP6 ensemble represent permafrost carbon and the last IPCC assessment of TCRE used external estimates of permafrost carbon cycle feedbacks. The inclusion of permafrost carbon cycle processes in ESMs is therefore necessary to improve the reliability of future projections and inform policy decisions.Based on the CMIP6 version of the Institut Pierre-Simon Laplace ESM, we developed IPSL-Perm-LandN, a new ESM that includes an explicit land nitrogen cycle and key permafrost physical and biogeochemical processes. Under future increasing atmospheric CO2 concentrations, the permafrost region remains a carbon sink in IPSL-Perm-LandN despite significant soil carbon losses due to permafrost thaw. In particular, we show a strong negative feedback arising from permafrost nitrogen release, which supports a large land carbon uptake and prevents the carbon-climate feedback parameter γ from increasing (negatively) by more than 10 PgC.°C-1. However, this is likely to be overestimated by our representation of soil nitrogen dynamics and plant nitrogen uptake. Our findings highlight the importance of better constraining the nitrogen cycle in permafrost regions and better representing permafrost carbon processes in ESMs to reduce the uncertainty in TCRE and remaining carbon budgets.