[MITgcm-devel] description of seaice parameters

[Martin Losch]

Hi,

I'll try to add documentation for these parameters: 
SEAICE_sensHeat/latentWater/latentIce to SEAICE_PARAMS.h and 
seaice_readparms.F.

pstar = SEAICE_strength=2.75e4 is believed to be very high by our ice 
modelers (Ralph Timmermann in particular). He said that he wouldn't 
question anything between 1.5e4 and 2e4 (or 2.5e4), based on his 
experience. Also, this value in particular seems to be very different 
between Artic and Antartic applications.
Also, a freezing point of Tf=-1.86 seems to make more sense. In the 
end, having Tf=f(S,p) does make most sense.
HO (or H0=h naught, as they say here) is effectively 0.5m in 
pkg/seaice, Ralph uses 1.0m (derived from personal experience).
I personally have no idea about these constants and report this just 
for the sake of collecting parameter ranges.

About these large Weddell polynyas, I still feel that ocean heat flux 
(too warm surface waters) are the most plausible explanation, but 
again, it's only a feeling. I want to try this with a coarse 
experiement to play with this, but haven't gotten around to it, yet.

Martin

On Sep 14, 2005, at 5:23 AM, Dimitris Menemenlis wrote:

> Hi Martin, thanks for sweeping through pkg/seaice.  That's really 
> helpful.  A few more comments in addition to Jinlun's:
>
>> The biggest difference is in SEAICE_freeze, which is -1.96degC in
>> seaice_readparms.F, but -1.86degC in his code. Maybe still another
>> reason for the thin ice around Antarctica and the large erroneous
>> polynyas in the high-res cubed sphere experiments.
>
> I have run a sensitivity experiment with -1.8 deg C freezing point, 
> which gives slightly thicker ice around Antarctica, as expected. 
> Eventually, I assume we will want this to depend on salinity.
>
>> Also he uses a smaller value of pstart=1.5e4 (instead of
>> SEAICE_strength=2.75e4 in seaice_readparms.F). Only FYI.
>
> I have also run a sensitivity experiment for this parameter.  Slightly 
> larger, rather than lower, pstar seems to help reduce the Polynya 
> problem.
>
>> However, we found that the parameters 
>> SEAICE_sensHeat/latentWater/latentIce are not really the coefficients
>> as described in SEAICE_READPARMS.h but (density of air)*(heat 
>> capacity of air/latent heat of water/latent heat of ice)*(transfer 
>> coefficient)
>> At least, that's what we found by comparing the code and default 
>> values. If you want me to do it I'll change the description in 
>> SEAICE_READPARAMS.h to reflect this, so that future seaice-novices do
>> not stumble over this?
>
> I remember working out what these constants actually were at some 
> point in time and that there were some scaling factors.  But I forgot 
> to add the scaling factors to descriptions in SEAICE_PARAMS.h as I 
> should have.  It would be great if you can update the definitions.
>
> Dimitris