[FEAT] mag grid message passing in the N/S direction

[aaronjridley]

This is critical for working. There are three "regions" that need to work:

1. at the pole (n or s) - we are making a super-cell here with the weighted average with the N/S velocity being the weight.
2. at the equator - No message passing. In the N/S direction, the N/S ghostcells should become vertical (lower) boundary conditions.
3. other places - need an interpolation to get the values along the field-line (or above the field-line - if you are looking towards the pole, the field-lines are shorter, so the messaged passed value will be "above" the field-line.)