England M.H.
J. Phys. Oceanogr., 23, 1523-1552, 1993.

KeyWords
World Ocean modelling. Annual cycle. Weddell sea. Southern-hemisphere. North-atlantic. Bottom water. Fresh-water. Climate. Ocean flow.

Abstract
A hierarchy of coarse-resolution World Ocean experiments were integrated with a view to determining the most appropriate representation of the global-scale water masses in ocean general circulation models. The largest-scale response of the simulated ocean to the prescribed forcing in each model run is described. The World Ocean model eventually has a realistic approximation of continental outlines and bottom bathymetry. The model forcing at the sea surface is derived from climatological fields of temperature, salinity, and wind stress. The first experiment begins with a quite unrealistic and idealized World Ocean. Subsequent experiments then employ more realistic surface boundary conditions, model geometry, and internal physical processes. In all, 16 changes to the model configuration are investigated.

A fundamental dynamical constraint in the Drake Passage gap appears to limit the outflow rate of bottom water in the Antarctic region. This constraint acts to decouple the extreme Antarctic waters from the rest of the World Ocean. In a similar manner, including a surface wind stress acts to decouple the two hemispheres by limiting near-surface meridional flows across the equator. In the Atlantic basin, this decoupling becomes negligible when North Atlantic Deep Water (NADW) production is simulated. It is found that the representation of low salinity Antarctic Intermediate Water (AAIW) is sensitive to the level of horizontal diffusion employed by the model, as well as the chosen geometry of the Drake Passage gap and the amount of buoyancy provided by the model's deep water. For example, provided that lateral diffusion rates are not too excessive, a fresh tongue of AAIW is simulated if either sufficiently dense bottom water is formed off Antarctica. or if enough NADW outflows into the Southern Ocean. The inclusion of an isopycnal mixing scheme is shown to improve the representation of AAIW in coarse-resolution models.

The rate of horizontal diffusion and the relative location of the Drake Passage gap to the polar westerlies determines the shape and strength of an intense meridional overturning cell in the Southern Ocean. The inclusion of an isopycnal mixing scheme does not affect this circulation pattern significantly. On the other hand, the intensification of NADW production can substantially weaken the downwelling component of this cell by drawing more water of Southern Ocean origin northward. Accurately simulating NADW production and outflow requires a complete seasonal cycle in thermohaline forcing in the North Atlantic. The return path of NADW is primarily via a ''cold water route'' (i.e., the Drake Passage), although sufficiently strong NADW formation sees some return flow via the Agulhas leakage (i.e., the ''warm water route''). By the last experiment of the present study, the model reproduces the subtle vertical layering of deep and intermediate water masses quite accurately. This represents a major success for the coarse-resolution multilevel ocean model.