We have constructed an idealized model of the Atlantic basin and described the steady circulations that resulted from forcing under symmetric restoring boundary conditions (Experiment 1). A small reduction in restoring temperature in the Southern Hemisphere (Experiment 2) led to a rather significant change in the intensity of the vertical circulations. Upon a switch to mixed boundary conditions, the state produced by Experiment 1 progressed to a final steady state with only transient variability, while the state resulting from Experiment 2 settled into a regular oscillation with a period of approximately 30 years.
It has become clear that several different kinds of variability are possible under constant mixed boundary conditions. Each region of deep sinking in the ocean can be capped by fresh-water and can therefore be variable. Weaver (1995, in this section) shows that variability arising from the variability of the sinking region in the Labrador Sea involves processes different from variability arising from the variability that involves the sinking region in the Nordic Seas. Similarly, this paper shows variability involving the sinking region poleward of the Antarctic Circumpolar Current.
We conclude with some questions and qualifications of the above work.
The simple type of coarse-resolution sector model presented here has a number of features that resemble the real ocean, and therefore proves useful in isolating these processes for detailed study. In particular, we saw that a small change in the restoring temperature led to rather large changes in the thermohaline circulation, even at high northern latitudes, and to even larger changes in behavior under a switch to mixed boundary conditions. While we cannot yet explain this sensitivity, we note that such unexpected sensitivities, if shown to be realistic in more complex models, may have important climatological consequences.
The basic question that arises from this and all other