Sea surface temperature anomalies, planetary waves, and air‐sea feedback in the middle latitudes

The mechanisms that contribute to the generation and damping of large‐scale mid‐latitude sea surface temperature (SST) anomalies are discussed. The SST anomalies reflect primarily the response of the upper ocean to the changes in air‐sea fluxes that are associated with daily weather fluctuations. Heat flux forcing is dominant in the lower middle latitudes, while wind‐driven entrainment may be most effective in the high latitudes; advection by anomalous Ekman current is generally less important, and Ekman pumping is negligible. The SST anomalies decay in part because of entrainment effects associated with mixed‐layer deepening and oceanic mixing and in part because of heat exchanges with the atmosphere. The three approaches commonly used to model the evolution of SST anomalies are reviewed: case studies based on monthly or seasonal anomaly maps of the large‐scale SST and atmospheric anomalies, numerical simulations with one‐dimensional mixed‐layer models, and stochastic forcing models. We stress the similarities in the different approaches and discuss their main advantages and limitations. The response of the atmosphere to mid‐latitude SST anomalies is considered. First, we discuss the poorly known relationship between SST anomalies and diabatic heating. Using a crude assumption for the air‐sea coupling, we consider a two‐layer quasi‐geostrophic channel model and discuss the stationary wave response to SST anomaly forcing and the resulting air‐sea feedback. It is found that the back interaction of the SST anomalies onto the atmosphere causes a weak SST anomaly damping at large scales and a strong one at small scales; the air‐sea coupling should also act as an eastward propagator for the SST anomalies. The response of more realistic linear wave models to prescribed diabatic heating is then reviewed, and it is suggested that realistic mid‐latitude SST anomalies have a weak influence on the atmospheric circulation, corresponding to changes in the geopotential height of 10–30 m at most. This order of magnitude is consistent with the results of general circulation model experiments and with the limited climate predictability associated with mid‐latitude SST anomalies.