| description abstract | In this paper, the physical mechanism of the polar halocline catastrophe (PHC) is reexamined with emphasis on the role played by the surface heat flux. It is argued that, in a coupled ocean?atmosphere system, thermal changes in the atmospheric state in response to changes in heat flux from the ocean weaken the feedback responsible for the PHC. So far, the PHC has been observed in models that use mixed boundary conditions; that is, the freshwater flux is specified, but the surface temperature is relaxed to a specified value. Previous explanations of the PHC have focused on the role of the freshwater flux in establishing a freshwater cap and shutting off the deep convection. However, the establishment of a freshwater cap reduces the depth of the water column that is cooled by surface heat loss. As a consequence, the surface temperature is reduced. Since the difference between this and atmospheric restoring temperature is now less, there is a corresponding reduction in the surface heat loss to the atmosphere, and this acts to further stabilize the water column. We examine the importance of this reduction in surface heat loss by considering two numerical experiments that are identical except that one is run under mixed boundary conditions and the other under a flux boundary condition applied to temperature as well as salinity. In each case, the surface fluxes are diagnosed from an experiment run to equilibrium using restoring boundary conditions on both fields. This also provides the initial state for both experiments. A PHC is easily induced in the mixed boundary condition case but not in the case using flux boundary conditions. By reducing the magnitude of the heat flux but not its sign, a pool of fresh water appears at the surface, but its effect is weaker than that under mixed boundary conditions and, in particular, there is no collapse of the meridional overturning circulation. A pool of fresh water also appears in an experiment in which a small, positive heat flux is added at all latitudes, a situation of relevance to global warming. This leads to an initial cooling in a shallow layer at the surface of the polar oceans, before heating at lower latitudes leads to a collapse of this state. These experiments show that the reduction in the surface heat flux that occurs when the PHC develops under mixed boundary conditions is an essential feature of the PHC. The use of mixed boundary conditions assumes that the atmospheric state is fixed and does not respond to changes in heat flux from the ocean. If the atmosphere were allowed to adjust to changes in this heat flux, then a PHC would be less likely to occur. This has been demonstrated by coupling the ocean model to the zero-heat-capacity atmospheric model used by Schopf. This is justified, following Bretherton, because of the large horizontal scale of the sea surface temperature (SST) anomalies in the experiments. The authors were unable to induce a PHC with this model. In reality, the atmospheric boundary condition seen by the ocean lies somewhere between the two extremes of mixed boundary conditions, on the one hand, and Schopf's model on the other. We have investigated this intermediate region by conducting experiments in which SST anomalies are damped on successively shorter time scales. These show that if the damping time is reduced sufficiently, a PHC can again be induced. | |
