[MITgcm-devel] offline_exf_seaice experiment update

[Jean-Michel Campin]

Hi,

Few updates regarding offline_exf_seaice experiment, forward test
(so that everyone is on the same page):

1) set all {field}period to zero for constant-in-time forcing.
  Since, to my knowledge, it's not tested anywhere else
  ({field}period=0 with non empty {field}File), I though it
  would be interesting to try this since none of the current forcing
  have time dependence.
2) regarding thermodynamics-only exp. (input.thermo & input.thsice),
  will adopt same channel-like set-up, starting with lower ice-conc
  at Northern and Southern side (2 row on each side, with 0. & 10.%
  on S. side, 0. & 1.% on N. side), similar to current input.thermo.
  Will try to have, when possible, same params in the 2 experiments.
  No plan to changed the other test-exp. with dynamics.

And just a comment:
 It turns out that for these 2 thermodynamics-only seaice-params setting, 
 pkg/thsice is more strait-forward than pkg/seaice (not surprising from 
 my point of view, since I know better the former, but there was few
 traps that did not seems obvious neither to Gael, who knows better than
 me pkg/seaice). Here are few points that one need to be aware of:
 a) default initial condition for snow (not the most natural).
 b) the default value (=F) of SEAICErestoreUnderIce is not what we 
    were expecting. This will be fixed in the next update of
    offline_exf_seaice test-exp (BTW, it also affects tests that 
    don't use pkg/seaice thermodynamics but use seaice dynamics).
 c) results should not depend whether or not SHORTWAVE_HEATING is
    defined (and I checked, this is the case with pkg/thsice);
    since this is a 1 level ocean set-up, adding all the SW heating
    in the surface level or at depth is not making any differences
    (there are some machine truncation differences, which I can explain
    given how the code is written in exf_getforcing.F). 
    However, with pkg/seaice, results change (because SWFracB is not
    zero if #define SHORTWAVE_HEATING).

Cheers,
Jean-Michel