Variable-Resolution Global Ensemble Analysis Using Model for Prediction Across Scales (MPAS)

Practical predictability is inherently limited by resolvable scales depending on the grid spacing in discretized models. The Model for Prediction Across Scales (MPAS) that has been recently developed at National Center for Atmospheric Research (NCAR) solves the fully compressible nonhydrostatic equations using variable resolutions to allow multi-scale global atmospheric simulations.

Using the Ensemble Kalman filter data assimilation, we discuss the challenges of the global analysis over the MPAS unstructured grids. Included are forward operators in the unstructured grid mesh, the effective noise control, and the scale-dependent analysis over the refine mesh area. Through the successful data assimilation of real observations, we demonstrate the benefit of the variable-resolution mesh in the cycling context.