Battisti, D. S., U. Bhatt, and M. A. Alexander, 1995: A modeling study of the interannual variability of sea surface temperature in the wintertime N. Atlantic Ocean. J. Climate, 8, 2897-2927.
A new model for the upper North Atlantic Ocean is presented and used to hindcast the SST from 1950 to 1988. The model consists of a matrix of one-dimensional (independent) columns in which a variable-depth, bulk mixed layer overlies a diffusive convective thermocline. The climatological annual cycle of heat flux convergence by the oceanic circulation is implicitly included in the formulation of the forcing. The 39-yr control integration of the model includes as surface forcing the shortwave and net longwave radiation from a control integration of the community climate model. Sensible and latent heat fluxes are determined from instantaneous values of surface temperature, humidity, and wind speed from the atmospheric model, and the SST simulated by the ocean model using the bulk formulae. The hindcast is performed by repeating the control integration, adding the observed, monthly mean surface anomalies in surface temperature, humidity, and wind speed for the period 1950-88. Thus, the simulated SST anomalies are generated explicitly by anomalies in the latent and sensible heat fluxes. A separate hindcast integration is presented, using as forcing the "observed" sensible plus latent beat flux anomalies rather than the surface atmospheric field anomalies to demonstrate that the major results are not predetermined by the formulation of the coupling.
The ability of the model to hindcast the wintertime interannual variations in SST is demonstrated by simple correlations with observed anomalies and by comparing the composite of warm and cold events observed with those simulated by the model. There is a good quantitative agreement between simulated and observed SST anomalies throughout most of the North Atlantic Ocean. Since the model formulation explicitly excludes any effects due to anomalies in the ocean advection, our results confirm the hypothesis that wintertime interannual to subdecadal variability in SST is mainly due to local anomalies in the air-sea flux of sensible and latent heat and not to anomalies in oceanic advection. Significant disagreement between hindcast and simulated SST anomalies is limited to a small region extending from Cape Hatteras to Nova Scotia along the U.S. coast. Here, the observed surface flux anomalies are anticorrelated with the SST anomalies, implicating important changes in oceanic advection in the generation of interannual wintertime SST and surface flux anomalies.
Both the sensible and latent heat flux anomalies are shown to contribute substantially to the wintertime anomalies in SST in the subpolar Atlantic, while the heat flux anomalies are predominantly determined by the latent heat flux in the subtropics. Entrainment anomalies contribute to a lesser extent to the mixed layer temperature anomalies throughout the basin. Sensitivity studies are performed to highlight the atmospheric processes and variability that account for the surface heat flux anomalies.