[Mitgcm-support] Re: Variable Az in MITgcm

[mitgcm-support at dev.mitgcm.org]

Thanks
/Petter
----- Original Message -----
From: "Alistair Adcroft" <adcroft at mit.edu>
To: "Petter Stenström" <petters at kth.se>
Cc: "MITgcm support" <support at mitgcm.org>
Sent: Wednesday, January 15, 2003 7:59 PM
Subject: Re: Variable Az in MITgcm


> Petter,
>
> You are right in all aspects of your message; calc_viscosity.F was
> created for this exact purpose. Be careful about lateral staggering
> of variables when calculating the sqrt(uzz+vzz) term; there is
> some four point averaging needed.
>
> Rgds,
>
> Alistair.
>
> > Petter Stenström wrote:
> >
> > Alistair,
> >
> > I need to use a variable vertical viscosity coefficient calculated from
local
> > shear and stratification according to:
> >
> > Az=const.*sqrt( (du/dz)^2 + (dv/dz)^2 - g/rho0*drho/dz )
> >
> > Browsing the code, I concluded that it should be fairly easy to do,
which
> > makes me suspicious... Can I just check with you that I haven't missed
> > anything important?
> >
> > It seems that I could just set viscAr to a small background value and
then add
> > a variable part in the sbrt calc_viscosity, based on local values of
du/dz,
> > dv/dz and drho/dz (based on the model variables uVel, vVel, rholoc and
> > recip_drF).
> >
> > It further seems that even though vertical dissipation is given in the
> > documentation as:
> >
> > Az*(d2u/dz2)
> >
> > it is actually calculated in the model as:
> >
> > d/dz(Az*du/dz)
> >
> > i.e. Av doesn't necessarily have to be a constant. Is that correct?
> >
> > Regards,
> >
> > Petter Stenström
> > Dept of Land and Water Res Eng
> > Royal Institute of Technology
> > SE-100 44 Stockholm
> > Phone: +46 8 7908657
> > Email: petters at kth.se
>