[MITgcm-devel] seaice

[Martin Losch]

Well, I don't think that there is a bug in the seaice-model (although  
there is always one more bugs, isn't is?). However, I feel (and it's  
not more than a feeling) that the handling of the ice-ocean stress  
needs looking into. Shouldn't the stress between ocean and seaice be  
the same for ice and ocean (except for the sign of course)? Can we  
expect the system to work if that's not the case?

A more general question: After staring at the seaice code all day, is  
there a reason (except convenience) that the seaice model has it's  
own grid parameters (instead of using the ocean model parameters such  
as dxG, dxU, etc.)? This culminates in having the variable  
RADIUS=6370km defined several times in separate subroutines instead  
of making it a global variable, the grid-lengths delX2,delXY, etc are  
recomputed everytime, lsr.F is called, etc. Would it make sense to  
clean that all up and introduce the parameters in model/inc/GRID.h in  
the seaice-model routines? Or is this a waste of time?

So many questions ...
Martin


On Feb 12, 2006, at 8:28 PM, Dimitris Menemenlis wrote:

>> as a matter of fact I use exf_bulkformulae, and thus CORE-winds in  
>> two cases:
>> one with seaice, one without. In a third case I use trenberth wind  
>> stress and
>> ncep heat fluxes.
>
> OK so either ice-modified heat and salt surface fluxes matter ...  
> but that is not likely, as you said, since stratification is  
> reasonable.  Or there is a bug.  One candidate is the asymmetric  
> time stepping.  I have never used pkg/seaice with asymmetric time  
> stepping.
>
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel