[MITgcm-devel] aside to negative H

[Jean-Michel Campin]

Hi Dimitris,

I don't want to change config_check.F, because it's better as it
is now. In particular for the special - undocumented - case 
where I use select_rStar = -1 with nonlinFreeSurf=0 to get
some of the r* feature but keeping Linear FreeSurf (and still
uses some r* arrays so needs to have NONLIN_FRSURF defined).

Cheers,
Jean-Michel

On Wed, May 30, 2012 at 10:25:57AM -0700, Dimitris Menemenlis wrote:
> Jean-Michel, we have not tried the "same" configuration with linear
> free surface,
> but Michael has several Southern Ocean set-ups that he runs with no problem
> using linear free surface.
> 
> Hong will try the configuration that gives negative values with
> linear free surface and get back.
> 
> (Sorry, as an aside, and I know that you don't like asides, wouldn't
> it be more accurate to add:
> 
>       IF (nonlinFreeSurf.EQ.0) THEN
>       IF (select_rStar .NE. 0) THEN
>         WRITE(msgBuf,'(A)')
> &   'CONFIG_CHECK: rStar is part of NonLin-FS '
>         CALL PRINT_ERROR( msgBuf, myThid )
>         WRITE(msgBuf,'(A)')
> &   'CONFIG_CHECK: ==> set nonlinFreeSurf .GT. 0 to use it'
>         CALL PRINT_ERROR( msgBuf, myThid )
>         errCount = errCount + 1
>       ENDIF
>       ENDIF
> 
> in config_check, rather than #ifndef NONLIN_FRSURF  ... #ENDIF
> since you already have checked for
> #ifndef NONLIN_FRSURF
>       IF (nonlinFreeSurf.NE.0) THEN
> 
> right before?)
> 
> The negative-H problem as I understand it is illustrated in this picture:
> http://ecco2.jpl.nasa.gov/data1/meeting/2012/05_23/NegativeH.JPG
> 
> Without understanding all implications,
> I think that best place to catch these errors is in ini_masks.F
> to make, so it's only done once.
> 
> Dimitris Menemenlis
> 
> 
> On 05/30/2012 05:32 AM, Jean-Michel Campin wrote:
> >Hi Dimitris,
> >
> >I will take a look at the r-star code to check pkg/shelfice issue.
> >
> >Just to clarify: you don't have negative H with linear free-surf,
> >only when you try to use r-star, right ? or did I missed something ?
> >
> >And OBCS does work with r-start, see verification/exp4/input.nlfs/
> >
> >Cheers,
> >Jean-Michel
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel