[MITgcm-devel] possible problem with seaice-ocean stress

[Martin Losch]

Hi Dimitris,

I have the feeling that since the ice-ocean stress works properly  
(that is the stress on the ocean is affected by the ice) the  
followling if-statement in seaice_model.F is wrong
C solve ice momentum equations and calculate ocean surface stress
       IF (
      &  DIFFERENT_MULTIPLE(SEAICE_deltaTdyn,myTime,SEAICE_deltaTtherm)
      &   ) THEN
          CALL dynamic solver
       ENDIF

This if statement makes it possible to have along seaice model step  
and not call the model each ocean time step. However, the stress for  
the ocean model is updated at each ocean timestep and then modified  
at each seaice timestep. For example you have  
SEAICE_deltaTtherm=deltaT=1200, and SEAICE_deltaTdyn=3600, then you  
would have ice-affected surface stress fu/fv only at every third time  
step, and for two time steps, the ocean feels the atmospheric stress,  
as if there were no ice. Am I right?

The solution would be to call seaice_ocean_stress (or ostres) at each  
time step, but this will not work for the B-grid (because ostres  
require cor_ice, which is only computed in dynsolver).
I don't know anyone who is using this feature, so I would suggest  
that we turn it off permanently until we have found a good solution  
for this.

Martin