[MITgcm-devel] seaice

[Dimitris Menemenlis]

Hi Martin, just to be sure that they are properly attributed and that Jinlun 
does not get any undeserved blame, I am fully responsible for the horrible 
useCubedSphereExchange hacks in seaice_init.F.  At the time that I wrote them, 
they were meant to be a very temporary fix to get us going with high-resolution 
cubed-sphere integrations.  My hacks are by no means rigorous or satisfactory in 
any way:

1) They assume that there is no sea ice anywhere except in faces 3 and 6, thus 
avoiding having to deal with cube corners.  This is done by setting the UVM mask 
to zero on all faces except 3 and 6.

2) In lieu of generalized curvilinear coordinates, which I did not know how to 
properly implement, I assumed that for purposes of sea-ice dynamics the cubed 
sphere grid is approximately rectangular on faces 1 and 3, hence setting CSTICE 
and CSUICE to one and TNGTICE and TNGICE to zero.

Dimitris