The impact of SST biases on projections of anthropogenic climate change: A greater role for atmosphere-only models?

There is large uncertainty in the model simulation of regional climate change from anthropogenic forcing. Recent studies have tried to link such uncertainty to intermodel differences in the pattern of sea surface temperature (SST) change. On the other hand, coupled climate models also contain systematic biases in their climatology, largely due to drift in SSTs. To the extent that the projected changes depend on the mean state, biases in the present-day climatology also contribute to the intermodel spread in climate change projections. By comparing atmospheric general circulation model (AGCM) simulations using the climatological SSTs from different coupled models, we show that biases in the climatological SST generally have a larger impact on regional projections over land than do intermodel differences in the pattern of SST change. These results advocate for a greater application of AGCM simulations with observed SSTs or flux-adjusted coupled models to improve regional projections of anthropogenic climate change.