Constraining Temperature Variability Projections Using Smiles That Best Represent Observed Variability

Projecting how temperature variability is likely to change in the future is important for understanding future extreme events. This comes from the fact that such extremes can change due to both changes in the mean climate and its variability. The recent Intergovernmental Panel on Climate Change (IPCC) report found large regions of low model agreement in projections of temperature variability in both December, January, February (DJF) and June, July, August (JJA) when considering 7 Single Model Initial-Condition Large Ensembles (SMILEs). In this study we use the framework described by Suarez-Gutierrez et al, (2021) to constrain future projections of temperature variability by selecting the SMILEs that best represent observed variability.

We consider 11 SMILEs with CMIP5 and CMIP6 forcing over 9 ocean regions and 24 land regions. First, we find that CESM2-LE, GFDL-SPEAR-MED and MPI-GE-CMIP6 perform best in DJF and CESM2-LE, CESM1-LE and GFDL-SPEAR-MED perform best in JJA. Second, our results show that out of the 11 models there are regions where only 1-2 models capture historical observations adequately. Here we find that even the handful of models that simulate today's climate adequately show diverging behaviours in variability change under future warming, showing little improvement in model agreement compared to the IPCC. Third, we find that models that are very widely used show a rather poor performance in capturing observed temperatures and should be used with care. Lastly, and possibly most importantly, we highlight that the climate science community cannot yet afford to move away from having several independent climate models to sufficiently sample model uncertainty. Our attempt to constrain projections of future temperature variability change is only one example of change in the climate system that cannot yet be foreseen with the best available knowledge.