P39 Regional climate simulations using variable-resolution meshes

Fowler, Laura D., William C. Skamarock, and Cindy Bruyere, National Center for Atmospheric Research

One goal for using the global non-hydrostatic Model for Prediction Across Scales (MPAS) for regional climate modeling is to bypass the need for nesting and nudging techniques at the edges of the computational domain, by providing a locally refined mesh within a much coarser mesh. Relative to traditional regional climate models, the horizontal discretization in MPAS based on unstructured centroidal Voronoi meshes leads to smoother transitions and two-way dynamics and physics interactions between the outer lower-resolution and inner higher-resolution meshes.

MPAS is currently being tested for seasonal-scale climate simulations on globally-uniform and regionally-refined meshes, using the same physics parameterizations as those in the NCAR Regional Climate Model (NRCM). MPAS is initialized with the same input data set as the NRCM. As in the NRCM, sea-surface temperatures are updated every 6-hours.

Initial multi-month simulations indicate that MPAS perfoms correctly for climate-type scale predictions for various combinations of horizontal resolutions over the coarser and refined meshes, especially considering that the MPAS physics has not been tuned to reproduce actual climate.

Results highlight seamless transitions for convection, cloud microphysics, radiation, and land-surface processes between the coarse and refined meshes, despite the fact that the physics parameterizations were not developed for variable resolution meshes.