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 IPCC report found large regions of low model agreement in the change 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 use 11 SMILEs with CMIP5 and CMIP6 forcing and consider 9 ocean regions and 24 land regions. We then assess, for both DJF and JJA, whether temperature variability projections are constrained by selecting for models capture observed variability in individual regions and seasons. We consider projected changes at various warming levels to account for differences in warming between models and the use of different future scenarios across CMIP5 and 6. We identify MPI-GE and CESM2 as the SMILEs that capture observed variability sufficiently. across most regions (29 & 30 out of 33 in DJF and 28 and 26 in JJA respectively). Whether temperature variability projections are constrained depends on both season and region. For example, in DJF over South East Asia the constraint does not change the already large spread of projections. Conversely, over the Amazon the constraint tells us temperature variability will increase in DJF whereas the entire model archive does not agree on the sign of the change. This method can be used to better constrain our uncertainty in temperature variability projections by selecting SMILEs that best represent observed variability.