[MITgcm-devel] [MITgcm-cvs] MITgcm/doc CVS Commit

Jean-Michel Campin jmc at ocean.mit.edu
Sat Jun 25 18:37:57 EDT 2011


Ian,

I did try on cartesian grid, and it is still wrong on beta plane
(and I can see why, since fCori is not periodic in Y-direction),
but it's working on f-plane.

I think that at the end, it will be better to change the current set-up 
(either to use cartesian-grid and fplane, or to use lat-lon grid with 2 column),
because right now, the problem is more than just the restart which fails.

Cheers,
Jean-Michel

On Sat, Jun 25, 2011 at 05:16:51PM -0400, Jean-Michel Campin wrote:
> Hi Ian,
> 
> Thanks for the  update. 
> I think I know what could be the problem (or at least one of the problem):
> The 1 column set-up has a problem if usingSphericalPolarGrid=.TRUE.,
> (lat-lon grid), except if the column is centered at the equator,
> because it requires the same dxG at j=1 and j=2 (symetric).
> Could you give it a try switching to Cartesian grid on Beta plane ? 
> There are few parameter to change (delX,delY,f0,beta ...) but 
> it would be usefull to check that the restart works on cartesian grid
> single column even if we decide to go back to several grid points on 
> Lat-Lon grid (the minimum would be 2, in the Y direction, with 1 grid 
> point on land).
> Sorry, I sould have though about this earlier.
> 
> Thanks,
> Jean-Michel
> 
> PS: I cc to devel
> 
> On Fri, Jun 24, 2011 at 09:09:23PM -0700, Ian Fenty wrote:
> > J-M,
> > I was able to run the tst_2+2 script sequence as you described but I
> > am unable to isolate the problem.
> > 
> > I tried alternative versions of the setup including:
> > 1) commenting out all of my data.seaice parameters
> > 2) running without the seaice package at all
> > 3) starting at temperatures well above freezing ( in case the values
> > in the pickup are being capped after reading )
> > 
> > What should I try next?  Go back to 5x5x23?
> > 
> > -Ian
> > 
> > 
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel



More information about the MITgcm-devel mailing list