[MITgcm-devel] ocean_emissivity consistency

[Menemenlis, Dimitris (3248)]

I am about to check in this check.  It will require changes to
pkg/seaice/seaice_readparms.F
verification/lab_sea/input.salt_plume/data.seaice
verification/seaice_obcs/input/data.seaice
Let me know if this is a problem.

>>>> NOTE <<<<
If you have data.seaice files that include SEAICE_waterAlbedo
"and" you are using SEAICE_EXTERNAL_FLUXES
you will need to remove the SEAICE_waterAlbedo line
or your code will stop and issue a warning.

When the free-drift package is fixed, I will put the same check
for OCEAN_drag.

Dimitris Menemenlis

On Dec 21, 2011, at 10:55 AM, Jean-Michel Campin wrote:

> Hi Dimitris,
> 
> If you can take care of this, would be nice.
> I am about to check-in changes in pkg/exf (for the interpolation
> parameters) but it will be confined to exf.
> 
> Cheers,
> Jean-Michel
> 
> On Mon, Dec 19, 2011 at 12:29:18PM -0800, Menemenlis, Dimitris (3248) wrote:
>> Makes sense ... and less disruption the way you suggest.
>> 
>> Do you want me to add the SEAICE_waterAlbedo check,
>> or will you take care of that?
>> 
>> Cheers
>> 
>> Dimitris Menemenlis
>> 
>> On Dec 19, 2011, at 12:25 PM, Jean-Michel Campin wrote:
>> 
>>> Hi Dimitris,
>>> 
>>> I think some constants are "constant" enough that having them set
>>> in model/src make sense. I tried a long time ago to put some
>>> in COMMON /PARM_ATM/ (PARAMS.h), like celsius2K, but it did not
>>> propagate as default for other pkg (exf is using cen2kel, set to
>>> a different value), may be because it was not possible to change
>>> it in file "data" ?
>>> In the same way, stefanBoltzmann, Latent heat (fusion, evap), 
>>> air heat capacity (this one is already there: atm_Cp) could be 
>>> defined in PARAMS.h
>>> 
>>> But regarding albedo, since it can be computed in many different
>>> ways (fct of temp/age of snow, zenit angle ...), it's less clear to me 
>>> what benefit we would get from setting some default in model/src.
>>> The drag coeffs are not so universal neither (are they ?) but
>>> some surface emissivities might be. My point here is that if
>>> one constant is used only once to computed a derived quantity 
>>> (e.g., a flux), it's probably better to keep it in the pkg
>>> which compute this flux.
>>> 
>>> The case of SEAICE_waterAlbedo is little bit different, and
>>> would be better to have it set to UNSET_RL in case we use
>>> pkg/exf, and stop if it set (in data.seaice) to something else.
>>> 
>>> Cheers,
>>> Jean-Michel