Assessing the robustness and implications of econometric estimates of climate sensitivity

Abstract

Earth's transient climate response (TCR) quantifies the global mean surface air temperature change due to a doubling of atmospheric CO2 concentration after 70 years of a compounding 1% per year increase. TCR is highly correlated with near-term climate projections, and thus of relevance for climate policy, but remains poorly constrained in part due to uncertainties in the representation of key physical processes in Earth System Models (ESMs). Within state-of-the-art ESMs participating in the Coupled Model Intercomparison Project (CMIP6), the TCR range (1.1 ºC–2.9 ºC) is too wide to offer useful guidance to policymakers. Similarly, the sixth report of the Intergovernmental Panel on Climate Change, while not solely reliant on ESMs for its TCR assessment, produced a very likely range of 1.2 ºC–2.4 ºC. To complement earlier, ESM-based, estimates, we here present a new TCR estimate of 2.17 (1.72–2.77) ºC (95% confidence interval), derived based on a statistical relationship between surface air temperature and observational proxies for its main drivers, i.e. changes in atmospheric greenhouse gases and aerosols. We show that, within uncertainty, this method correctly diagnoses TCR from 20 CMIP6 ESMs if the same input variables are taken from the ESMs that are available from observations. This increases confidence in the new observation-based central estimate and range, which is respectively higher and narrower than the mean and spread of the estimates from the entire ensemble of CMIP6. Many ESM-based estimates tend to produce TCRs lower than the observational range reported here. Our findings suggest that a misrepresentation of the aerosol cooling effect could be the cause of this discrepancy. Further, the revised TCR estimate suggests a downward revision of the remaining carbon budgets aligned with the overarching goal of the Paris agreement.