[MITgcm-devel] other celsius2K in seaice_solve4temp.F

[Martin Losch]

Hi Jean-Michel,

this entire thermodynamics part is still misterious to me. These different constants 273.16 and 273.15 sometimes seemed to be there on purpose, although the purpose evades me (use a different freezing/melting temperature of freshwater when coming from above or below freezing?), so that I never dared to touch them.
If you think that these constants are only there for unphysical (ie. historic) reasons, than changing everything to celsius2K consistently is more than called for. The same for TB(I,J)=271.2

Martin

BTW. I am using #ifndef SEAICE_EXTERNAL_FLUXES (o: Quite useful, when you want to have your own private bulk formulae ...


On Jan 30, 2012, at 3:22 PM, Jean-Michel Campin wrote:

> Hi,
> 
> I realized when editing seaice_solve4temp.F that I may have missed few
> things when changing the default celsius2K from 273.16 to 273.15:
> since there are few hard-coded 273.16 value in seaice_solve4temp.F
> that (may be ?) should have been changed to celsius2K before 
> changing the default value of celsius2K.
> Although I am not completely sure since some 273.16 value was 
> there in pkg/seaice (in former growth.F, groatb.F, budget.F)
> before celsius2K was added (Jun 2003).
> 
> So I have 2 questions:
> 1) Should I replace those 273.16 in seaice_solve4temp.F with celsius2K ?
> (I already did this in seaice_budget_ocean.F but it does not matter too
>  much there since nobody use this #ifndef SEAICE_EXTERNAL_FLUXES code)
> It would be just logical since celsius2K was just 273.16 before
> my recent switch; plus celsius2K is already used in few places,
> in seaice_solve4temp.F and in seaice_growth.F.
> 2) In a similar move, I think it would make sense also to change 
> TB(I,J) = 271.2 _d 0 (ifdef SOLVE4TEMP_LEGACY):
> to:
> TB(I,J) = celsius2K + SEAICE_freeze (presently ifndef SOLVE4TEMP_LEGACY)
> since the default SEAICE_freeze is -1.96 which, providing the older
> default of celsius2K=273.16 is used, gives TB = 271.2 in both cases.
> This later modif does change the truncation error, but most tested 
> exp. still "pass" with at least 13 digits, except AD tests lab_sea.evp
> and lab_sea.noseaicedyn, not clear why.
> 
> Cheers,
> Jean-Michel
> 
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel