MPAS global meshes at various resolutions -- icosahedral vertices at the poles versus not at the poles

[mdtoy65]

I've plotted the 'areaCell' fields for some of the MPAS global meshes available at MPAS-Atmosphere mesh downloads. Looking at the arrangement of the 12 icosahedral vertices (corresponding to the areaCell minima), the north and south poles typically coincide with two of these vertices, with the exception of the 15km and 3km meshes (see attached plots). I’m curious as to why the 15km and 3km have a different orientation such that the vertices do not line up with the poles. Was this a conscious decision? Or did it just turn out that way for the set of offered grids. I understand that the user can rotate the stock grids to any orientation. Thank you.

 

[mgduda]

Interesting question! I don't think there was any specific reason for the different orientation in the 15-km and 3-km meshes. My best guess is that we may have started with an icosahedron generated using three mutually perpendicular golden rectangles, and subdivided those to get the 15-k and 3-km meshes, while the rest of the quasi-uniform meshes began with an icosahedron with two of its vertices at the poles (constructed using Euclid's method, perhaps, or through a rotation of the icosahedron generated using planes?).