Introduction

Climate feedbacks, which either amplify or dampen the response of the climate system to radiative forcing, are fundamental in determining the extent of global warming caused by anthropogenic greenhouse gas emissions. While many feedbacks, such as water vapour and ocean carbon uptake, are well-represented in the Earth system models (ESMs) utilised for Intergovernmental Panel on Climate Change (IPCC) reports, warming-induced emissions (WIE) from natural sources remain largely absent. WIE refers to the additional greenhouse gases released from natural environments as temperatures rise. Because these emissions amplify warming, their omission from current models risks systematically underestimating future temperature projections and overestimating the remaining carbon budgets.

In the IPCC Sixth Assessment Report (AR6), none of the 11 ESMs evaluated included WIE from wetlands or freshwaters. Only two incorporated permafrost ecosystems, and five included wildfires. This research addresses this gap by deriving functional relationships between global surface air temperatures and WIE rates for major natural sources, projecting their cumulative warming impact through to 2100.

Methods

The study quantified WIE and their subsequent temperature impacts using MAGICC version 7.5.3, a reduced-complexity climate model calibrated to output from the sixth Coupled Model Intercomparison Project (CMIP6). The analysis focused on three Shared Socioeconomic Pathways: SSP1-2.6, SSP2-4.5, and SSP4-6.0. These scenarios represent a range of future trajectories from strong mitigation to continued growth in radiative forcing.

Six major WIE sources were considered: carbon dioxide and methane from permafrost and wildfires, and methane from wetlands and freshwaters. WIE was defined as the increase in emissions relative to the 2020 baseline. The researchers performed linear regression to establish relationships between global surface air temperature and emission rates. For permafrost, the study used SSP-specific time series projections that account for gradual top-down, abrupt, and post-fire thaw. The resulting WIE were added to original SSP emissions to calculate the additional warming attributable to each source individually and in combination.

Results

The research found that each warming-induced methane source scales linearly with temperature. The combined global warming-induced methane emission rate is estimated at 97 ± 6 Tg CH4 yr−1 °C−1 across all three SSPs. This rate is equivalent to approximately 2.6 ± 0.2 Pg CO2e yr−1 °C−1 using the IPCC AR6 GWP100 value. Both permafrost and wildfire carbon dioxide sources also scale linearly in SSP2-4.5 and SSP4-6.0, yielding a combined emission rate of 7 ± 1 Pg CO2 yr−1 °C−1. In the SSP1-2.6 scenario, the relationship for permafrost carbon dioxide is non-linear, as emissions decline more rapidly than they increase once temperatures peak.

By 2100, the total WIE from the sources considered could cause additional warming of 0.2 ± 0.2 °C in SSP1-2.6, 0.3 ± 0.2 °C in SSP2-4.5, and 0.4 ± 0.2 °C in SSP4-6.0 relative to 2020. This effectively amplifies post-2020 anthropogenic warming by approximately 20% to 30%. Permafrost carbon dioxide is identified as the single largest contributor to this warming, followed by wetland methane. Notably, the uncertainty range for total WIE-induced warming remains entirely above zero for all SSPs by 2100, indicating a robust net warming effect despite inherent model uncertainties.

Discussion

The projected WIE-induced warming of 0.2 °C to 0.4 °C represents a significant amplification of anthropogenic warming. For context, the projected methane emission rate from WIE in 2100 spans the current anthropogenic emissions of North America and Asia combined. These emissions are currently absent from most ESMs, unaccounted for in national climate pledges, and not addressed by existing mitigation strategies. The study notes that the linear relationships found emerge through global integration and distinct source mechanisms, despite the exponential nature of microbial decomposition.

The findings have critical implications for policy frameworks, particularly the transient climate response to cumulative emissions (TCRE). The report suggests that omitting WIE from ESMs biases TCRE low. Current IPCC AR6 corrections use a fixed mass of carbon per degree of warming, but the researchers argue WIE are better characterised as an enhanced emission rate. It is recommended that the IPCC AR7 adopts scenario-dependent corrections to reflect the temporal dynamics of these feedbacks. Until climate models comprehensively incorporate WIE through improved process understanding, their projections are likely to underestimate warming and overestimate remaining carbon budgets.