CISL Annual Report 2005 > CISL research plans

Numerical methods: HOMME plans

The following are the top priorities for

1. The next logical step in our efforts to develop a general circulation model (GCM) is to couple the comprehensive CAM physical parameterization package to the HOMME spectral element (SE) dynamical core. Once this is complete, we will validate using generally accepted test cases (e.g., baroclinic instability, aqua-planet, and AMIP) and compare the achieved integration rates with existing community climate models.

2. A discontinuous Galerkin (DG) spectral element shallow water model has been implemented within the HOMME framework and is currently being extended to a conservative formulation of the 3D governing equations. Once this conservative DG core is complete, we will leverage the CAM coupling and validation work done with the spectral element core to build a 3D scalable and conservative GCM.

3. Semi-implicit/OIFS time integration methods will be investigated for both the DG- and SE-based GCMs. These time integration methods require effective and scalable solvers that operate on the cubed sphere to achieve scientifically useful integration rates. We are investigating multilevel Schwarz techniques that have been shown to work and scale well on such problems.