These calculations were made in order to test the code for symmetry breaking bugs. They turn out to be an elegant example of how the singularities on both the cubic and lat-lon grid are handled properly by the model.

The above solution is stable on the lat-lon grid because of the implicit-in-time formulation of the surface pressure equation. Had we used an explicit time-stepping method the convergence of meridians at the poles would have led to a restrictive CFL limit on time-step.