[MITgcm-devel] seaice+thsice update

[Jean-Michel Campin]

Hi folks,

I checked-in a new advection driver for thsice variable, 
pkg/thsice/thsice_advdiff.F 
and I started to evaluate how it works (in progress ...).

I wanted to compare the results with Martin's version 
in pkg/seaice (seaice_advdiff.F), so I tried it: 
- run global_ocean.cs32 for more than 400yrs with exf &
 CORE forcing, until the model stop because sea-level was 
 uniformly too low (with realFreshWater, seaice-AdvSchem=1).

- I also try with all the "C jmc: new version:" uncommented
 in pkg/seaice/seaice_advdiff.F ; I was not expecting 
 large differences, but in fact the mean sea-ice thickness
 in the Arctic is much larger (mean seasonal cycle: 
 from 3 to 3.5 m compared to 2 to 2.5 m before)
I checked the maximum ice fraction, goes up to ~ 110 % in the
previous version, and I think this is the reason why the 
ice was much thiner.

I am now trying without realFreshWater to run it longer.

I will check-in this modif ("C jmc: new version:" uncommented)
soon.

Cheers,
Jean-Michel