An assessment of the role of the ocean circulation in governing SST variability outside the tropical Pacific

It has been proposed that the ocean outside of the tropical Pacific region, plays an important role in climate variability, in particular when decadal time scales are concerned. The prevailing argument is that the upper ocean (gyre) circulation responds to atmospheric forcing, and reverses SST anomalies forced by the same atmospheric conditions through more local, thermodynamic interaction. Because the circulation response lags the thermodynamic one by a few years, the interaction can lead to oscillations with decadal time scales. If this is the case, there is potential for predictability in the coupled system.

Using a set of forced ocean GCM (OGCM) experiments, we explore the role of the ocean circulation in the maintenance of SST anomalies in the North Pacific and Atlantic Oceans, as well as in the tropical Atlantic. These anomalies are known to be associated with prominent climate variability patterns, the so-called Pacific Decadal Oscillation, North Atlantic Oscillation, and the variability of rainfall in northeast Brazil and sub-Saharan Africa. All our ocean models are forced by observed surface winds which drive a thermodynamic model of the atmospheric sub-cloud layer, particularly suitable for forcing ocean models and studying SST variability. A hierarchy of ocean models from simple one-dimensional to two dimensional, thermodynamic slab models and fully dynamical OGCMs are used. We find that ocean heat transports is important for obtaining a superior simulation of the observed SST variability since 1950 or so, including decadal time scale anomalies observed in the three ocean regions mentioned above. In all these cases the circulation provides instantaneous reinforcement or damping (depending on the geographical location) of the heat flux generated anomalies, through the wind-forced Ekman circulation. There is no evidence however, of a gyre-circulation response to wind-stress variations that could lead to a negative feedback and give rise to decadal oscillation. Inevitably we conclude that there is little hope to glean decadal predictability from monitoring and modeling the coupled atmosphere-ocean system outside the tropical Pacific.