Ocean modelling efforts in the global climate system

The climate subsystems consisting of ocean, atmosphere, cryosphere (ice-sheets and sea-ice), marine and land biosphere, interact on time-scales from days to millennia and on horizontal scales from a few metres (or less) to global dimensions. Since current computational resources prohibit integrating a model incorporating all subsystems on these spatial and temporal scales, climate researchers have used different combinations of the subsystems in a varying degree of sophistication. In this paper I will try to give an overview of ocean models which have been used to explain and elucidate the role of the oceans in different frequency ranges of the climate variability spectrum. Modelling of such a diverse range of scales in space and time can only be performed today with restrictions in the spatial resolution or integration length. Coarse resolution ocean general circulation models (O-GCM) can be integrated for O(5000 years), coupled atmosphere-ocean GCMs (A/O-GCMs) for O(100 years) and eddy-resolving ocean GCMs for O(10 years) with the current computer technology. Equilibrium solutions of coarse-resolution O-GCMs are sensitive to the parametrisation of the atmospheric forcing, which is in most of these cases a rather crude coupling to climatological surface forcing fields. Multiple equilibrium circulations are possible depending on the boundary conditions. Coupling to an active atmospheric model introduces further difficulties, like a drift of the coupled system into a new equilibrium, which differs from the equilibrium solutions of the uncoupled sub-components. Because the deep ocean adjusts to different surface forcing characteristics on time-scales of millennia, an equilibrium cannot be reached with current computer resources. The importance of the mesoscale eddy field in the dynamical and thermodynamical balance of the Southern Ocean and its effect on property transports in this region has recently been the subject of much controversial discussion. Because the Southern Ocean is believed to be one of the most sensitive regions of the globe to greenhouse warming, it is important to incorporate the effects of eddies correctly into coarse resolution models. Multi-level primitive equation modelling of the global ocean circulation in eddy-resolution is a major computational task which has just been started. The principal differences and problems encountered in all these approaches to study the role of the oceans in climate dynamics will be highlighted and discussed.