GCM (and ESM) models divide the planet into a three-dimensional lattice of grid points. In the atmosphere, global climate models typically have a horizontal resolution of 124-186 mi. (200-300 km). and a vertical resolution from a few hundred yards (m.) to 0.6-1.2 mi. (1-2 km.). The vertical resolution of the models is generally finer at the surface (where it is needed to resolve important small-scale processes) and becomes coarser at higher altitudes. At the surface, the model levels are normally terrain-following (the levels run parallel to the surface topography), but at higher model levels the effect of the surface topography is less, and the model levels become flat. The ocean component of an ESM/GCM typically has a finer resolution than the atmosphere component because small-scale structures such as eddies are important for heat transport in the ocean. The ocean resolution is typically about 410-820 ft. (125-250 m.) in the horizontal and 6561,312 ft. (200-400 m.) in the vertical.
In the horizontal direction, the grid points typically run parallel to the lines of latitude and longitude, creating rectangular grid boxes that are grouped more closely together at the poles than at the equator. This is a waste of computational effort at the poles where there are lots of small grid boxes. In addition, this bunching of grid points at the poles can lead to computational instabilities. To avoid this problem, some researchers are developing models that run on non-rectangular grids such as Geodesic grids made up of hexagons and pentagons rather than rectangles, but this is still in the relatively early stages of development. In addition, tri-polar ocean grids have been developed; these grids have three regions where the grid points bunch together, each of which is located over land (so there is no need to treat the ocean in these areas). This removes the problem of numerical instabilities in the ocean model.
Was this article helpful?