[MITgcm-support] EXF period with interpolation

[Jean-Michel Campin]

Hi Renske,

I did check how exf_set_uv.F would work if allowing to have wind or stress period set to 0
(for constant in time forcing), and was surprised that it just works fine.
So I was wrong in assuming there would be a problem there.

I made these changes yesterday in the CVS repository (in exf_set_uv.F and exf_init.F,
removing other un-necessary stop) so that if you update your code today it should allow
you to use a zero period (and only 1 reccord in files) without stopping.

Cheers,
Jean-Michel

On Mon, Oct 12, 2015 at 12:09:16PM -0400, Christopher Pitt Wolfe wrote:
> Hi Renske:
> 
> I would recommend just replicating your force fields, say, once in your input files and picking your repeat time to be whatever you want it to be. Then the model will interpolate between your two copies of the forcing, but since they???re the same the interpolation won???t really do anything.
> 
> Christopher
> 
> > On Oct 12, 2015, at 11:57 AM, Renske Gelderloos <rgelder2 at jhu.edu> wrote:
> > 
> > Hi Jean-Michel,
> > 
> > Thanks, that clarifies things a bit. Unfortunately, I'm using a curvilinear grid, which is why on-the-fly interpolation seemed such a great idea in the first place. 
> > 
> > If I understand this correctly, the simplest work around will be to not use time-mean forcing but e.g. a mean seasonal cycle? 
> > 
> > Renske
> > 
> > 
> > ________________________________________
> > From: Jean-Michel Campin <jmc at ocean.mit.edu>
> > Sent: Monday, October 12, 2015 11:30 AM
> > To: mitgcm-support at mitgcm.org
> > Subject: Re: [MITgcm-support] EXF period with interpolation
> > 
> > Hi Renske,
> > 
> > This "STOP" in exf_set_uv.F has been added not too long ago (June 2014)
> > so I suspect it is necessary (does catch a real problem).
> > 
> > The issue has to do with the fact that the vector version (EXF_INIT_UV) that
> > would do the initial interpolation of a 2 compoinent vector field is missing (not yet
> > written), whereas, for a scalar field, we have an initialisation routine (EXF_INIT_GEN)
> > in addition to the in-time-step loop routine EXF_SET_GEN (both in same source file:
> > exf_set_gen.F).
> > 
> > There might be a simple way to go arround this but would need to check carefully.
> > 
> > In the mean time, if you are not in this case: usingCurvilinearGrid .OR. rotateGrid
> > you could try to set:
> >      uvInterp_stress = .FALSE.
> >      uvInterp_wind   = .FALSE.
> >      uvInterp_climstr= .FALSE.
> > by commenting out lines 517 & 521 in exf_init_fixed.F:
> > c # ifdef EXF_USE_OLD_INTERP_POLE
> > ....
> > c # endif
> > 
> > Cheers,
> > Jean-Michel
> > 
> > On Mon, Oct 12, 2015 at 02:29:24PM +0000, Renske Gelderloos wrote:
> >> Hi all,
> >> 
> >> 
> >> I'm trying to use the EXF package with on-the-fly interpolation and (initially) non-time-varying forcing. If I have understood this correctly, I set repeatPeriod to 0 and that should work. However, there is a safe guard built into exf_set_uv.F, that if USE_EXF_INTERPOLATION is set and uvecperiod .LE. 0 the model stops. Now, in my case uvecperiod is 0 because I have time-independent forcing. Did I misinterpret something or should this safe guard be .LT. and not .LE.?
> >> 
> >> 
> >> Renske
> > 
> >> _______________________________________________
> >> MITgcm-support mailing list
> >> MITgcm-support at mitgcm.org
> >> http://mitgcm.org/mailman/listinfo/mitgcm-support
> > 
> > 
> > _______________________________________________
> > MITgcm-support mailing list
> > MITgcm-support at mitgcm.org
> > http://mitgcm.org/mailman/listinfo/mitgcm-support
> > 
> > _______________________________________________
> > MITgcm-support mailing list
> > MITgcm-support at mitgcm.org
> > http://mitgcm.org/mailman/listinfo/mitgcm-support
> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support