[MITgcm-support] seaice dynamics stability

[dmenemenlis at gmail.com]

Matt, could you make your setup available somewhere and I will take a look.  I have never seen the problem you describe (except in EVP solver for small sub-cycling).  I wonder if it is caused by geometry of your grid, i.e., the long skinny grid cells near 80S?  If you send a figure of grid scale noise, I can look in 1/8th-deg integration to see if same happens.  D.

Dimitris Menemenlis
mobile: 818-625-6498

-----Original Message-----
From: Matthew Mazloff <mmazloff at mit.edu>
Date: Friday, Dec 19, 2008 3:57 pm
Subject: [MITgcm-support] seaice dynamics stability
To: mitgcm-support at mitgcm.orgReply-To: mitgcm-support at mitgcm.org

Hi JMC,

As you suggested I am going to decouple ice stress from ocean stress  
in attempt to find a more stable seaice solution.
Is all I need to do is comment these lines at the end (lines 216 and  
217) of seaice_ocean_stress.F?

   fu(I,J,bi,bj)=(ONE-areaW)*fu(I,J,bi,bj)+areaW*fuIceLoc
   fv(I,J,bi,bj)=(ONE-areaS)*fv(I,J,bi,bj)+areaS*fvIceLoc


-Matt



And dear others,

I find better (closer to observed) seaice coverage with ice dynamics  
on.  Unfortunately with my 1/6 solution I have only been able to run  
~650 days and I want a 3 year simulation.  My ice area fields show  
significant grid scale noise at most times, and the model finally  
crashes where ice is being converged into the coasts.

I have played with many parameters, from drag coefficients to  
advection schemes....and the crashes are really pretty random...its  
frustrating.  If anyone has any suggestions please let me know

Thanks
Matt


_______________________________________________
MITgcm-support mailing list
MITgcm-support at mitgcm.org
http://mitgcm.org/mailman/listinfo/mitgcm-support