The Link Between Marine Boundary Layer Clouds and Interannual to Interdecadal SST Variability over the Atlantic and Pacific

Marine boundary layer clouds may amplify modes of climate variability found over the Atlantic and Pacific oceans via their strong coupling to sea-surface temperature (SST). However, climate models poorly simulate the relationship between subtropical boundary layer clouds and SST. Is there a detectable influence of this relationship on model-to-model differences in the representation of basin-wide patterns of SST variability? Over the North Atlantic, the dominant mode of summer-to-summer SST variability exhibits a horseshoe-like pattern, while over the North Pacific, the dominant mode also exhibits a horseshoe-like pattern associated with the Pacific Decadal Oscillation. Over each ocean basin, subtropical and midlatitude anomalies of net cloud radiative effect (CRE) and low-level cloud fraction associated with the dominant mode are co-located with anomalies of SST, consistent with low-level clouds’ acting as a positive feedback on SST. Generally, models participating in the Coupled Model Intercomparison Project phase 5 (CMIP5) underestimate the magnitude of SST and CRE changes associated with these modes. Models that simulate the strongest coupling between subtropical SST and shortwave CRE produce a higher amplitude of SST and net CRE variability over the subtropical northeast Atlantic, subtropical northeast Pacific, and northern midlatitude Pacific than that produced by models that simulate the weakest coupling. Therefore, the amplitude of interannual to interdecadal SST variability over the northern oceans during summer in coupled climate models may be sensitive to their representation of marine boundary layer clouds.  Results from a similar analysis of modes of variability over the South Atlantic and Pacific will also be presented.