373
7 Concluding remarks
This paper analyzes, in a systematic and within limits rigorous way, feedbacks between the thermohaline circulation and meridional transports in
the atmosphere that can be represented in a coupled ocean-atmosphere box
model. The approach is austere: The model can be analyzed to considerable depth, but many processes acting in more complex coupled models or
in the real climate are not represented at all, or possibly distorted in their
relative importance.
Chief among the processes in more realistic models of the thermohaline
circulation that are missing here is perhaps the feedback between convective activity, surface heat flux, and the strength of the deep circulation.
Lenderink and Haarsma (1994) first suggested a mechanism by which convection might be stopped, locally, by a local perturbation, without initially
affecting the vicinity much, but eventually changing the meridional overturning. Rahmstorf (1995b) later expanded this picture, showing that significant changes in overturning strength of an idealised global model can
be triggered by the cessation of convection in very few gridpoints. Rahmstorf and Willebrand (1995) discuss how the convective feedback depends
on the surface heat flux parameterization, and demonstrate a stabilising
effect of weaker SST restoring just like in the models with fixed E-P in
section 3 here. For a detailed discussion of the convective feedback and its
relationship with advective feedbacks, see the recent review by Rahmstorf
et al. (1996).
What then can be learned from this simple model, with its potential
to overestimate the importance of the processes it can represent (as, for
example, emphasised by Lohmann et al., 1996a)? A prime virtue of the
model is that it is fully coupled and allows a self-consistent representation
of extratropical, large-scale ocean-atmosphere interactions. It makes possible the investigation of a surprisingly rich variety of effects even beyond
the ones presented here (for example, one can analyze the heat transport
capability of the total climate system, depending on a few fundamental
parameters). The limits of applicability are as yet unclear; for example,
the difference between the two statistically steady states in the coupled
GCM of Manabe and Stouffer (1988) shows the important effects of differing thermohaline circulation patterns in the Atlantic, on meridional heat
7 Concluding remarks
This paper analyzes, in a systematic and within limits rigorous way, feedbacks between the thermohaline circulation and meridional transports in
the atmosphere that can be represented in a coupled ocean-atmosphere box
model. The approach is austere: The model can be analyzed to considerable depth, but many processes acting in more complex coupled models or
in the real climate are not represented at all, or possibly distorted in their
relative importance.
Chief among the processes in more realistic models of the thermohaline
circulation that are missing here is perhaps the feedback between convective activity, surface heat flux, and the strength of the deep circulation.
Lenderink and Haarsma (1994) first suggested a mechanism by which convection might be stopped, locally, by a local perturbation, without initially
affecting the vicinity much, but eventually changing the meridional overturning. Rahmstorf (1995b) later expanded this picture, showing that significant changes in overturning strength of an idealised global model can
be triggered by the cessation of convection in very few gridpoints. Rahmstorf and Willebrand (1995) discuss how the convective feedback depends
on the surface heat flux parameterization, and demonstrate a stabilising
effect of weaker SST restoring just like in the models with fixed E-P in
section 3 here. For a detailed discussion of the convective feedback and its
relationship with advective feedbacks, see the recent review by Rahmstorf
et al. (1996).
What then can be learned from this simple model, with its potential
to overestimate the importance of the processes it can represent (as, for
example, emphasised by Lohmann et al., 1996a)? A prime virtue of the
model is that it is fully coupled and allows a self-consistent representation
of extratropical, large-scale ocean-atmosphere interactions. It makes possible the investigation of a surprisingly rich variety of effects even beyond
the ones presented here (for example, one can analyze the heat transport
capability of the total climate system, depending on a few fundamental
parameters). The limits of applicability are as yet unclear; for example,
the difference between the two statistically steady states in the coupled
GCM of Manabe and Stouffer (1988) shows the important effects of differing thermohaline circulation patterns in the Atlantic, on meridional heat