Will we cross 1.5C?

The timing and likelihood of crossing 1.5 °C has major implications for policy decisions, yet the uncertainties contributing to the threshold-crossing time are not well quantified. We introduce a new framework that allows a comprehensive separation and quantification of the uncertainties due to model response uncertainty, internal variability, and uncertainty in observed warming since the preindustrial period. It is well established that several of the recently published CMIP6 models have a higher climate sensitivity than CMIP5 models, with some studies arguing that the highest values shown by CMIP6 models are not plausible. Using the CMIP6 ensemble would result in an overestimation of the response uncertainty. Therefore, we estimate the response uncertainty by using an emulator tuned to assessed climate sensitivities from the published literature. Quantifying the effect of internal variability is challenging because internal variability can not be separated cleanly from the response uncertainty. We use single-model initial condition large ensembles to provide a clean estimate of the effect of internal variability. Lastly, the timing of crossing 1.5 °C depends on how much the Earth has warmed to date. We use the comprehensive uncertainty estimate from the HadCRUT5 dataset to estimate the effect of historical warming uncertainty. We find that the largest contribution to the uncertainty in when we might cross 1.5 °C comes from the response uncertainty, followed by the historical warming uncertainty. The earliest time for crossing 1.5 °C, here defined as 5% likelihood of the combined distribution of uncertainties, does not depend on the scenario. In every scenario except SSP5- 8.5, the 20-year-averaged period that crosses 1.5 °C with a 5% likelihood is 20132032. The central estimate (50%), on the other hand, does depend on the scenario, with SSP5-8.5 crossing 1.5 C in 20182037 and SSP1-1.9 in 20262045. All scenarios except SSP1-1.9 have a likelihood close to 100% to cross 1.5 C before 2100. Even in SSP1-1.9, the scenario with the strongest emission reductions, there is a 58% likelihood to cross 1.5 C by the end of this century. This implies that even in SSP1-1.9, the world may stay below 1.5 C only if both climate sensitivity and historical warming are near the lower end of their respective distributions.