This talk will describe our efforts related to the formulation and development of a coupled ocean-sea ice-atmosphere-land surface model to be used for global climate simulations. All of the model sub-components are discretized using spherical geodesic grids. These grids discretize the surface of the sphere in a nearly uniform and isotropic manner and, therefore, avoid problematic grid-pole singularities.

Both the atmosphere model and the ocean model will utilize quasi-Lagrangian vertical coordinates in order to dramatically reduce spurious dispersion and diffusion in the vertical direction. The atmosphere model uses a hybrid sigma-theta vertical coordinate, while the ocean model uses an arbitrary Lagrangian Eulerian coordinate.

Our integrated and holistic approach allows the basic numerical modules to be used throughout the coupled model system. For instance, the same numerical discretization of basic operators, such as gradient, divergence, curl, Laplacian, and monotone advection, will used by all model components. In addition, all model sub-components use a single Message Passing Interface (MPI) module. This inter-operability will expedite the development and optimization phases. And finally, the use of geodesic grids for all model sub-components allows for the straight-forward development of a distributed flux-coupler.

We will present results from an extended (~20 year) climate simulation of the atmospheric model component. Preliminary results from the ocean model and sea-ice model may also be presented.