[MITgcm-support] basic questions about periodicity in MITgcm domains

[Jean-Michel Campin]

Hi Jeff,

The type of grid (Cartesian, spherical polar grid or curvilinear) does not influence
the way the halo/overlapping-region of each tile is filled up by the exchanges.
And since the only working option with default exch1 routines is to assume doubly periodic
domain, it applies to any type of grid.

Now, once a wall or an open-boundary is specified, the halo region beyond the wall or OB
is not relevant anymore, so that it does not matter if exch S/R assume a periodic domain.

There are more options within pkg/exch2 that allow to remove some domain connectivity
to make it non periodic, but this does not save much time to disconnect one periodicity rule.

Hope this answer your question,

Cheers,
Jean-Michel

On Tue, Aug 04, 2020 at 08:34:09PM +0000, Jeffery R Scott wrote:
> Hi all,
> 
> The manual describes MITgcm domains as doubly periodic in x,y unless walls exist at least one of the top-bottom left-right boundaries.
> 
> For Cartesian and spherical polar grids, this is straightforward. However, can someone confirm if using OBCS for a boundary 
> effectively eliminates all periodicity for this boundary? (all variables, regardless of which variables used in obcs)
> 
> Unfortunately the manual is lacking much documentation about curvilinear grids (and, the doc on EXCH2 is incomplete).
> I would think there is no periodicity assumed using a curvilinear grid (?)
> 
> Thanks,
> Jeff
> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support