How does North Atlantic sea surface temperature variability differ between fully coupled and slab ocean models?

Abstract:

We compare North Atlantic sea surface temperature (SST) variability between coupled models with a dynamical ocean (henceforth fully coupled model or FCM) and models where the ocean is represented by a motionless slab (henceforth slab ocean model of SOM). Currently, there is disagreement within the community regarding if/how SST in FCMs and SOMs differ. We implement a rigorous statistical method which tests whether SST in FCMs and SOMs are different and, if so, determines the spatial pattern that explains the differences as well as the associated variate time series. This method indicates that one major difference is that the variance of the SST tripole pattern is larger in the SOM compared to the FCM. We speculate that this is the result of diminished oceanic damping in the SOM. In addition, the FCM contains a mode of low frequency variability over the subpolar North Atlantic that is absent in the SOM. We determine the driver of this variability by computing regression maps between the variate time series and oceanic fields of interest, including wintertime mixed layer depths and the Atlantic Meridional Overturning Circulation. In summary, we conclude that ocean dynamics are more important for setting SST in the subpolar North Atlantic, in line with previous studies.