[MITgcm-support] seaice_salinity

[Matthew Mazloff]

Hi Dimitris,

Right, it may not be the seaice model itself, the seaice dynamics may  
be causing the ocean model to blow up.  And it might be the aspect  
ratio.  Unfortunately I don't really know when and where the  
instability is occurring.  The problem is the crash always takes at  
least 14 days to occur.  So I really don't get that output....maybe I  
can code something so that when the stop is called it writes the  
final state.  I'll look into that.

So do you have any hunches as to what is happening at what may cause  
this.  Earlier you wrote,
  "The CS510 solution also has some very saline water in narrow  
fjords around Antarctica.  I think what might be happening is that  
ice is acting as a salinity pump, taking freshwater out of the fjord,  
when ice dynamics are turned on, but there is no mechanism for  
flushing the salty water out in the model ocean.  Let me know what  
you find.  D."
The model is more stable with #undef SEAICE_SALINITY so perhaps this  
is the case.  Is there anything I can further do to prevent this  
"salinity pump".

-Matt



On Nov 18, 2008, at 9:45 AM, Dimitris Menemenlis wrote:

> Matt, are you sure that it is the sea ice model that is unstable  
> and blows up?
> Admittedly the CS510 resolution is a little coarser than 1/6 degree  
> around Antarctica but we have not had problems with sea ice model  
> stability for quite some time.
>
> Also looking back at 1/8 and 1/16 degree integrations, the sea ice  
> was stable, although the code has changed since.
>
> I wonder if possibly the aspect ratio of your grid is what causes  
> the problem?  For all the integrations mentioned above the grid  
> aspect ratio is approximately square.
>
> Do you have any additional information on the specific way in which  
> your model blows up?  Do NaNs start creeping in the surface flux  
> fields produced by the sea ice model.
>
> What might help would be to diagnose the following:
>
>  249 |SIfu    |  1 |UU      U1      |N/m^2           |SEAICE zonal  
> surface wind stress, >0 increases uVel
>  250 |SIfv    |  1 |VV      U1      |N/m^2           |SEAICE merid.  
> surface wind stress, >0 increases vVel
>  251 |SIempmr |  1 |SM      U1      |kg/m^2/s        |SEAICE upward  
> freshwater flux, > 0 increases salt
>  252 |SIqnet  |  1 |SM      U1      |W/m^2           |SEAICE upward  
> heatflux, turb+rad, >0 decreases theta
>  253 |SIqsw   |  1 |SM      U1      |W/m^2           |SEAICE upward  
> shortwave radiat., >0 decreases theta
>
> in the time step right before and right after the crash.
>
> D.
>
> Dimitris Menemenlis <menemenlis at jpl.nasa.gov>
> Jet Propulsion Lab, California Institute of Technology
> MS 300-323, 4800 Oak Grove Dr, Pasadena CA 91109-8099, USA
> tel: 818-354-1656;  cell: 818-625-6498;  fax: 818-393-6720
>
> On Nov 18, 2008, at 9:23 AM, Matthew Mazloff wrote:
>
>> My biggest concern is how unstable the seaice model is!  Do you  
>> know any tricks?  Perhaps turning down max vel. or shortening the  
>> time step?
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support