[MITgcm-support] MITgcm-support Digest, Vol 128, Issue 38

[Martin Losch]

Neil,

please have a look at <http://mitgcm.org/public/r2_manual/latest/online_documents/node129.html>, more specifically <http://mitgcm.org/public/r2_manual/latest/online_documents/node131.html#SECTION004122100000000000000> where a couple of stability criteria are described. The CFL criterion for advection is only one of them (usually the most important one, but at high resolution the one for viscosity often hits you and that’s why the Leith-scheme and the normalized parameters viscAhGrid etc are so useful).

Martin

On Mar 20, 2014, at 12:28 AM, Neil Patel <nigellius at gmail.com> wrote:

> Thanks. Btw, it looked like my issue was my time step was too big. It was stable for a long time, but eventually crashed with enough model time. I tried using the same time step you described (600 second). Is it possible at 1/6° and Leith Viscosity a larger time step could be stable? Is the time step determined by the CFL criterion or something else?
> 
> Thanks,
> 
> Neil
> 
>> 
>> 
>> Message: 3
>> Date: Thu, 27 Feb 2014 17:25:17 -0800
>> From: Dimitris Menemenlis <dmenemenlis at gmail.com>
>> To: MITgcm Support <mitgcm-support at mitgcm.org>
>> Subject: Re: [MITgcm-support] Viscosity parameters
>> Message-ID: <2F879B4C-B263-4645-A0A8-F3E4B2B0C23C at gmail.com>
>> Content-Type: text/plain; charset=windows-1252
>> 
>> If you want to use Leith, you could start by trying:
>> 
>> viscC4Leith=1.5,
>> viscC4Leithd=1.5,
>> viscA4GridMax=0.5,
>> 
>> These are the values we use for the CS510 configuration,
>> which is close to 1/6-deg grid spacing around Equator.
>> A description of this Leith+ scheme is here:
>> http://ecco2.org/manuscripts/2008/FoxKemperMenemenlis08.pdf
>> 
>> On Feb 27, 2014, at 4:44 PM, Neil Patel <nigellius at gmail.com> wrote:
>> 
>>> I've read the details of the Leith / Smagorinsky parameters. I presume they're a replacement to setting viscAhGrid/viscAhGridMax parameters, but not sure what the advantages or disadvantages of either are.
>> 
>> 
>> 
>> 
>> ------------------------------
>> 
>> Message: 4
>> Date: Fri, 28 Feb 2014 09:28:40 +0100
>> From: Martin Losch <Martin.Losch at awi.de>
>> To: MITgcm Support <mitgcm-support at mitgcm.org>
>> Subject: Re: [MITgcm-support] Viscosity parameters
>> Message-ID: <51D9A937-3899-48A8-8D39-340D3301FAF6 at awi.de>
>> Content-Type: text/plain; charset="windows-1252"
>> 
>> Neil,
>> 
>> I think that your viscAhGrid is too small. I?d try something like 0.1
>> Also you might want to try biharmonic viscA4Grid.
>> 
>> The flow dependent viscosity parameterization (especially Leith, see Dimitris suggestion) are usually the most stable choices.
>> 
>> Martin
>> 
>> On Feb 28, 2014, at 1:44 AM, Neil Patel <nigellius at gmail.com> wrote:
>> 
>>> Hi,
>>> 
>>> I'm having trouble understanding the different viscosity parameters. Read the mitgcm manual many time over but still am confused. From what I can tell:
>>> 
>>> 1) Vertical eddy viscosity doesn't affect the numerical stability, so choose the value closest to the observations for ViscAr
>>> 2) Horizontal eddy viscosity does affect the numerical stability and the number chosen should be similar to various energy scaling / grid parameters.
>>> 
>>> currently, I'm running a 1/6? resolution simulation. I set 
>>> 
>>> viscAr=5.6614E-04,
>>> 
>>> and 
>>> 
>>> viscAhGrid=0.01,
>>> viscAhReMax =0.1,     
>>> viscAhGridMax=0.1,  
>>> 
>>> I'm not sure what the difference between the last two. With the current setup I have, will the viscosity always be within the CFL condition? And so, If my model crashes  ? "SOLUTION IS HEADING OUT OF BOUNDS" is what I've been getting, with the parameters above could the viscosity parameters be the issue or is it soley due to too large a time step. What would be the negative of setting the viscAh parameters to a very low number? Would it be an unphysical ocean or just a poor parametrization of small flows or neither?
>>> 
>>> I've read the details of the Leith / Smagorinsky parameters. I presume they're a replacement to setting viscAhGrid/viscAhGridMax parameters, but not sure what the advantages or disadvantages of either are.
>>> 
>>> Thanks for any explanation you can give.
>>> 
>>> Thanks,
>>> 
>>> Neil
>>> _______________________________________________
>>> 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
>> 
>> 
>> End of MITgcm-support Digest, Vol 128, Issue 38
>> ***********************************************
>>> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support