[MITgcm-support] viscAh & viscAz!

[Baylor Fox-Kemper]

Hi Goran,
   I'm not sure how the flags are set for generic_advdiff...  Perhaps 
using mom_vecinv flips it? (JMC can comment?).
   You are in a similar situtation to the one that inspired me to add 
viscC2leithD, high-res and wanting to use Leith.  The Leith scheme only 
detects vortical motions (i.e., it is proportional to grad(horizontal 
relative vorticity)).  However, once your grid is smaller than the 
deformation radius, divergent motions can be important, and in fact you 
can have divergent instabilties.  Thus, I added viscC2leithD, which 
squashes grad(div v) the same way that ordinary Leith squashes 
grad(curl v).  Leith originally proposed the scheme for purely 2-d 
calculations (where there is no div v), so this is just a little tweak.
    I'd propose to use

viscc2leith=0.2,
viscc2leithd=0.2,

instead of
viscc4leith=0.2,

See if that works.  You might crank up the values to 1 if it doesn't 
work initially.
   Cheers,
    -Baylor


On Jul 7, 2005, at 1:20 PM, Göran Broström wrote:

> Hi Baylor
>
> I have some computer problems at the moment (and I am starting my 
> vacation pretty soon) so I have not been able to test what is wrong 
> with my runs.
>
> I saw that you use "generic_advdiff" in packages.conf. Should this 
> flag be set (besides using mom_vecinv
> which I assume is essential)?
>
> About what scheme I use. Frankly I do not know what scheme to use but 
> trying to follow Dimitris I started with
>  viscAh=0.,
> # viscAhMax=unset
>  viscAhGrid=0.,
>  viscC2Leith=0.,
>  viscA4=0.,
> #  viscA4Max=0.,
>  viscA4Grid=0.,
>  viscA4GridMin=6.e-4,
>  viscA4GridMax=8.e-3,
>  viscC4Leith=0.2,
>
> But I have tried different versions. If you have better suggestions I 
> will be happy to try these settings.
>
> Sincerely
> Göran
> ps. Dimitris, misu.su.se is Meteorological Institute, Stockholm 
> University.
>
>
>
> ----- Original Message ----- From: "Baylor Fox-Kemper" <baylor at MIT.EDU>
> To: <mitgcm-support at mitgcm.org>
> Sent: Wednesday, July 06, 2005 4:15 PM
> Subject: Re: [MITgcm-support] viscAh & viscAz!
>
>
>> Are you useing ViscC2LeithD as well as ViscC2Leith? or C4 versions of 
>> both?
>>   Cheers,
>>
>>    -Baylor
>>
>>
>>
>> On Jul 2, 2005, at 3:01 PM, Göran Broström wrote:
>>
>>> Hi
>>>
>>> Reading the anwers in this thread I decided to try the Leith scheme. 
>>> Unfortenately the model seems to blow up after a few timesteps 
>>> (using values on parameters that I found on the web page). My 
>>> questions is therefore if anyone can help me with the settings to be 
>>> used with this scheme. The model I am working has a resolution of 
>>> about 0.1 degree and I use spherical coordinates.
>>>
>>>
>>> Sincerely
>>> Göran Broström
>>>
>>> _______________________________________________
>>> MITgcm-support mailing list
>>> MITgcm-support at mitgcm.org
>>> http://mitgcm.org/mailman/listinfo/mitgcm-support
>>
>>
>> _______________________________________________
>> MITgcm-support mailing list
>> MITgcm-support at mitgcm.org
>> http://mitgcm.org/mailman/listinfo/mitgcm-support
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support