[MITgcm-support] seaice

[mlosch at awi-bremerhaven.de]

I noticed that the seaice package doesn't use any of the grid 
information of the ocean code, but recomputes everything again. 
Some of this is probably necessary because the seaice code 
operates on a B-grid, but I assume if someone (I keep promising 
that, but never get around to doing it, maybe because at the 
moment I don't use it (o:) sorts these things out and copies the 
ocean grid information to the seaice grid variables properly, the 
package should have no more problems on cartesian grids.

Martin