[MITgcm-support] cosPower with GMRedi and Leith viscosity

[Martin Losch]

Hi Kaitlin,

I haven’t looked into this in a long time, but as far as I remember all of the ISOTROPIC_COS_SCALING business is related lat-long grids and has been superseded by scaled horizontal viscosity parameters. In the current model/inc/CPP_OPTIONS.h these CPP flags are marked as “Retired”. Here’s the comment for ISOTROPIC_COS_SCALING in CPP_OPTIONS.h
C o ALLOW isotropic scaling of harmonic and bi-harmonic terms when
C   using an locally isotropic spherical grid with (dlambda) x (dphi*cos(phi))
C *only for use on a lat-lon grid*
C   Setting this flag here affects both momentum and tracer equation unless
C   it is set/unset again in other header fields (e.g., GAD_OPTIONS.h).
C   The definition of the flag is commented to avoid interference with
C   such other header files.
C   The preferred method is specifying a value for viscAhGrid or viscA4Grid
C   in data which is then automatically scaled by the grid size;
C   the old method of specifying viscAh/viscA4 and this flag is provided
C   for completeness only (and for use with the adjoint).

So you are doing fine for viscosity (with Leith, which scales by grid size just as viscAhGrid). At higher resolution I would use the biharmonic versions. Be aware that Leith will sometimes give you very small viscosities so that you may want to add some minimum viscosities on top of that. See the headder of pkg/mom_common/mom_calc_visc.F for a quick summary of the options.

For tracers, this option can still be used, but with GMREDI you probably have diffKhT and diff KhS = 0 anyway. In the gmredi-code, the CPP-flags and cosFacU/V never appear.

I think that your vertical diffusivities are maybe a little large. I’d use 1e-5.

Martin


> On 26. Apr 2018, at 12:28, Naughten, Kaitlin A. <kaight at bas.ac.uk> wrote:
> 
> To whom it may concern,
> 
> My name is Kaitlin Naughten and I have just started a postdoc at the British Antarctic Survey. I am setting up a Weddell Sea configuration of MITgcm, using an older configuration from a colleague as a starting point. The old configuration used version c62r which I have now updated to c67, so I expect many options and "best practices" have changed between these versions.
> 
> Right now I'm trying to make sure I'm doing something sensible with mixing. GMRedi and KPP are switched on, and the relevant namelist parameters are as follows:
> 
> viscAr=1.E-4
> viscC2leith=3.0
> viscC2leithD=3.0
> diffKrT=1.E-4
> diffKrS=1.E-4
> cosPower=2.
> 
> From old messages here, it looks like my choices of mixing schemes are reasonable. But I am confused about whether cosPower is needed. ISOTROPIC_COS_SCALING is defined in GAD_OPTIONS.h, but not in CPP_OPTIONS.h, so I understand cosPower only affects the tracers. But since GMRedi is on, is cosPower actually necessary and does it do anything? Furthermore, should I have cosPower scaling the momentum terms, or does the Leith parameterisation already scale by grid size? I've looked at some old messages and at the code but I can't quite figure it out.
> 
> If it matters, the grid is polar spherical, quarter-degree scaled by cos of latitude, and I can see some eddy activity resolved. Later we plan to increase resolution.
> 
> Many thanks,
> Kaitlin Naughten
> This message (and any attachments) is for the recipient only. NERC is subject to the Freedom of Information Act 2000 and the contents of this email and any reply you make may be disclosed by NERC unless it is exempt from release under the Act. Any material supplied to NERC may be stored in an electronic records management system.
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support