[MITgcm-devel] Add sea ice surface forcing to pkg/seaice

[Menemenlis, Dimitris (3248)]

Jean-Michel, I tried to subtract snowPrecip from precip in a not-too-inelegant way,
but it may change some test results.

On my laptop, all the forward verification experiments that involve pkg/seaice
remain unchanged.  I have also tested the adjoint  global_ocean.cs32x15
verification experiments on laptop and they too remain unchanged.

On a linux box, there is some changes though:

before:
Y Y Y Y>16<16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 22 16 16 16 22 16 16 16  .  .  .  .  .  .  .  .  .  .  .  . pass  global_ocean.cs32x15.seaice
Y Y Y Y>16<16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 22 16 16 16 22 16 16 16  .  .  .  .  .  .  .  .  .  .  .  . pass  lab_sea.hb87
Y Y Y Y 16>16<16 pass  global_ocean.cs32x15.seaice_dynmix

after:
Y Y Y Y>11<16 16 16 16 16 16 16 16 16 13 14 16 16 16 14 16 22 16 16 16 22 16 16 16  .  .  .  .  .  .  .  .  .  .  .  . pass  global_ocean.cs32x15.seaice
Y Y Y Y> 2< 8  9  8  7 13 12 10  7  5  7  2  4  6  6  3  4 22  4  6  6 22  4  6  5  .  .  .  .  .  .  .  .  .  .  .  . FAIL  lab_sea.hb87
Y Y Y Y 16>16< 9 pass  global_ocean.cs32x15.seaice_dynmix

I will wait for verification tests tonight on baudelaire before deciding how to proceed.

I can force results to be bit-identical in a brute force way using "IF ( snowPrecipFile .NE. ' ' )"
tests and repeating code every time "PRECIP*(1-AREA)" is needed, but it would not be very elegant.

Dimitris Menemenlis

On Apr 22, 2012, at 5:57 AM, Jean-Michel Campin wrote:

> Hi Dimitris,
> 
> regarding (1): I don't know if my version of staf is up-to-date,
> but it works since it's used everyday for daily test on baudelaire:
> in: tools/example_scripts/csail/test_baudelaire, line 10:
>> if [ -d ~/bin ]; then export PATH=$PATH:~/bin ; fi
> so that I can use ~jmc/bin/staf
> -rwxr-xr-x. 1 jmc 14107 2010-11-26  /home/jmc/bin/staf*
> but anyone should also be allowed to use it (x permission for everyone).
> 
> (2) got few fails last night on iblade & solasrv (the one Martin run)
> for the forward global_ocean.cs32x15.seaice test. Might be that 
> the code was downloaded in the middle of your check-in.
> 
> (3) - nothing to say.
> (4) there is 1 thing to clarify: In exf, precip is "Total Precipitation",
>    and snowprecip (same units) is therefore a fraction of "precip".
>    This is how it's described in EXF_FIELDS.h and how it's used 
>    in thsice with exf.
>    I see in your modif that snowprecip is added to snow-thickness but I
>    don't see it removed from precip (+ rain is still converted to snow ?).
> I would prefer to have things clarified before turning on snowprecip in 
> these experiments.
> 
> Cheers,
> Jean-Michel
> 
> On Sun, Apr 22, 2012 at 01:51:09AM +0000, Menemenlis, Dimitris (3248) wrote:
>> Jean-Michel, a couple of comments and questions about verification experiments:
>> 
>> 1. How do I get permission to testreport adjoint verification experiments on baudelaire?
>> 
>> 2. I was not able to test the adjoint seaice verification experiments.
>>    (I think that I am using an outdated version of staf and have contacted Thomas about it.)
>>    So there is a small possiblity that some adjoint seaice experiments will fail tonight.
>> 
>> 3. I commented out snowprecipfile definition in
>>   global_ocean.cs32x15/input.seaice/data.exf
>>   global_ocean.cs32x15/input_ad.seaice/data.exf
>>   global_ocean.cs32x15/input_ad.seaice_dynmix/data.exf
>> because snowprecip is now active in pkg/seaice and this will change results.
>> 
>> 4. Would it be OK to modify global_ocean.cs32x15/input.seaice
>>    as a test for snowprecip in pkg/seaice ?