[MITgcm-devel] pkg/frazil

[Jean-Michel Campin]

Hi Dimitris,

reagarding (1), I would propose one of these 3 experiments:
 global_with_exf (standard input)
 global_with_exf (input.yearly)
 global_ocean_ebm 
with a little preference for global_with_exf/input.yearly
Can also think of one of the 2 global_ocean.90x40x15 (standard or dwnslp),
but these 2 start from a pickup and it might be better to just start
from initial condition.
But right now, all of them uses allowFreezing=.TRUE. Is it a problem ?

Regarding 2 , it's not very clear to me since did not have time to look 
at pkg/frazil. But for (3), I don't think it's just an historic reason,
since r_unit*Kelvin/s has the same form for any tracer ( r_unit * tracer_unit / s)
and is used internally whereas W/m2 fit better when exchanging fluxes 
with other components/interfaces/files.

Cheers,
Jean-Michel

On Sat, Mar 03, 2012 at 08:20:16AM -0800, Menemenlis, Dimitris (3248) wrote:
> Jean-Michel and Gael, as requested I have changed the pkg/frazil forcing terms
> from K/s to W/m^2.  Three questions:
> 
> 1) Which non-sea-ice verification experiment should I modify to test pkg/frazil
> 
> 2) Since K/s units can be problematic, as you explained, because of time-varying drF,
> shouldn't we replace K/s with W/m^2 in mypackage so as to discourage the use
> of K/s when folks use mypackage as a template to build new packages?
> 
> 3) For temperature forcing units we also at times use r_unit*Kelvin/s rather that W/m^2.
> Is there a reason, other than historic, to use r_unit*Kelvin/s rather that W/m^2?
> 
> Cheers, Dimitris
> 
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel