[MITgcm-devel] Re: obcs_external_fields_load and OBCS exchanges

[Patrick Heimbach]

Martin,

Quoting Martin Losch <mlosch at awi-bremerhaven.de>:

> Patrick,
> that was it.
> I use now in obcs_init_variables after applying the boundary conditions:
>         CALL EXCH_UV_XYZ_RL(uVel,vVel,.TRUE.,myThid)
>         _EXCH_XYZ_R8( theta, myThid )
>         _EXCH_XYZ_R8( salt , myThid )
> and obcs_prescribe_exchanges remains unchanged.
> I get identical results.

Great. Can you check this in.

> Do we need an IF(useOBCSprescribe) for the exchanges in 
> obcs_init_variables? Probably not.

The IF(useOBCSprescribe) should still be there, I think,
since no exchanges are needed if you don't prescribe. 
obcs_calc, orlanski, obcs_apply_... already include the full halo, 
so exch in these cases would be redundant.
Or was I missing some corner issues?

Cheers
-Patrick



> On Sep 24, 2004, at 3:26 PM, Patrick Heimbach wrote:
> 
> > Martin,
> >
> > I quick thing to check:
> > The obcs_prescribe_exchanges routine was conceived
> > to just exchange velocity fields at the end of dynamics.
> > Now you're probably using the same routine after
> > obcs_init_variables. The tracer fields are thus not
> > exchanged (which they didn't have to in the original plan).
> > Can you try?
> >
> > -Patrick
> >
> >
> >
> > Quoting Martin Losch <mlosch at awi-bremerhaven.de>:
> >
> >> Patrick,
> >>
> >> I now have a configuration with four oben boudnaries on which I
> >> prescribe boundary conditions from a file. I run on one tile and on
> >> 2x2 tiles and unfortunately the results are different.
> >> 1. It gets a little better, when I include the call
> >> obcs_prescribe_exchanges in obcs_init_variables
> >> 2. after 1. only the east-west boundaries are affected (see attached
> >> figure), and I absolutely do not see why.
> >>
> >> Martin
> >>
> >>
> >
> >
> > --------------------------------------------------------
> > Patrick Heimbach   Massachusetts Institute of Technology
> > FON: +1/617/253-5259                  EAPS, Room 54-1518
> > FAX: +1/617/253-4464             77 Massachusetts Avenue
> > mailto:heimbach at mit.edu               Cambridge MA 02139
> > http://www.mit.edu/~heimbach/                        USA
> >
> 
> 


--------------------------------------------------------
Patrick Heimbach   Massachusetts Institute of Technology
FON: +1/617/253-5259                  EAPS, Room 54-1518
FAX: +1/617/253-4464             77 Massachusetts Avenue
mailto:heimbach at mit.edu               Cambridge MA 02139
http://www.mit.edu/~heimbach/                        USA