When Vortex was born, the most promising modeling framework was the Weather Research and Forecasting model: our good friend WRF. For almost 15 years, every simulation run in the Vortex system has used WRF and we have become experts on it. Meanwhile, we have kept our eyes open for interesting competitors to WRF and we are currently getting to know a promising candidate: MPAS.
MPAS stands for Model for Prediction Across Scales and it is a “collaborative project for developing atmosphere, ocean and other earth-system simulation components for use in climate, regional climate and weather studies”1. The atmospheric dynamics and physics are essentially borrowed from WRF but the most basic and far-reaching difference between WRF and MPAS is the structure of the grid.
In WRF, space is divided into rectangular lat-lon cells with specific grid spacing and the refinement for higher resolutions is possible thanks to nested domains. There are several disadvantages in this setup and further research led MPAS developers to decide that the best structure was an unstructured grid: a Spherical Centroidal Voronoi Tesselation. This is a mathematical partition of a sphere that can have any given number of regions and adapt to varying densities over the globe, which effectively means one can control the area of the cells and make smooth variations from high-resolution to low-resolution areas (see picture above).
This new “un-structure” allows for good scaling on massively parallel computers, avoids problems at the poles and abrupt mesh transitions in nested domains, and increases accuracy and flexibility for variable resolution simulations. Regional simulations are able to smoothly transition from reanalysis resolution at the borders to the desired high-resolution of the central cells.
In the same way, WRF replaced MM5 not so long ago, maybe MPAS will replace WRF in the near future. MPAS is getting strong, and here at Vortex, we are working on making it stronger.