Restructuring under GEOSocean_GridComp

[yvikhlya]

This PR addresses #447 and is contingent on GEOS-ESM/GEOSgcm#317.

[tclune]

cmake changes ok

[mathomp4]

I'm adding @GEOS-ESM/gcm-gatekeepers (aka @sdrabenh) as this PR, as @yvikhlya says, is one of those "must be done at the same time as a GEOSgcm update" things. Want to make sure he knows about this so we can coordinate it.

[mathomp4]

Note: This will never pass CI here.

It's a fun dependency issue, but the main thing is that by the time the CI system knows to checkout this branch, it's already done a mepo clone on the main of GEOSgcm. Thus, MOM5 and MOM6 are in the "wrong place" and so things go boom. You can see that the GEOSgcm CI in GEOS-ESM/GEOSgcm#317 did build because it put MOM5 and MOM6 in the "right place". (Well, AWS failed but that's wonky at the best of times).

So, I'll need to work with @sdrabenh when this goes in to turn off CI.

[yvikhlya]

Also note that I renamed GEOSocean_GridComp to GEOS_OceanGridComp. This way there is no need in "OVERRIDE" statement in CmakeLists.

[mathomp4]

Approving as superuser.