[MITgcm-support] KPP Diagnostics

[Jean-Michel Campin]

Hi Chris,

I don't think there is an error in the diagnostic units of KPPg_SLT
(psu.m^3/s, as it appears in "available_diagnostics.log") and units are 
also consistent with other interior flux diagnostics from pkg/generic_advdiff 
(e.g., advective fluxes ADVx,y,r_SLT and diffusive fluxes DFx,y,rE,I_SLT)
where diagnostics KPPg_SLT is filled up (pkg/generic_advdiff/gad_calc_rhs.F,
line 677).

Regarding salinity units, you can interpret "psu" as beeing "g/kg":
this would be exact if using TEOS10 eq. of state; but even with other
EOS, the way salinity is treated in MITgcm corresponds to absolute salinity
and has nothing to do with conductivity (from which real "psu" is related).

An finally, regarding this:
> >From my reading of the documentation, it should be the case that
> KPPg_SLT = KPP_ghatK * surForcS
I don't know where this comes from (did not find it in the manual).
May be we should fix this part to avoid any units confusion.

Cheers,
Jean-Michel

On Wed, Dec 03, 2014 at 07:27:08PM -0500, Chris Horvat wrote:
> Hi all,
> 
> Trying to figure out about some of the KPP diagnostics. There is a
> diagnostic of the form
> 
> KPPg_XXX where XXX is SLT, TH, or a tracer.
> 
> In some earlier support emails, it was noted that this is the non-local
> flux of (salinity, say), however it has units of m^3 psu / s, instead of
> g/m^2/s as is the case for SFLUX or surForcS.
> 
> >From my reading of the documentation, it should be the case that
> 
> KPPg_SLT = KPP_ghatK * surForcS
> 
> but the units don't work here. My first thought is that to compare the
> units, you would take
> 
> KPPg_SLT --> KPPg_SLT / (rhoNil * dx * dy)
> 
> but this may be an incorrect assumption. Hoping someone may be able to shed
> some light on this issue.
> 
> I also am unable to find where this diagnostic is filled (!) as this is not
> in kpp_output.F nor does some greping show it to be in pkg/kpp or even
> pkg/* so I may be missing how it is filled.
> 
> Chris

> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support