[MITgcm-support] How to choose obc balance coefficients?

[Martin Losch]

Hi Camille,

I think using the same factor for both boundaries is OK in most situations.
Sometimes, if you know that there is little flux across a certain boundary, you wouldn’t want this flux to be “corrected” and then would use a smaller value (or even zero). Again, here some intiution (or experience) is required. Keep in mind that the code balances the flow at each time step. There are also situations where you would like to have a fluctuation in you inflow (e.g. tides) and the balancing needs to be done over longer time scales. In that case you’d have to do the balancing offline.

Martin

> On 14 Mar 2017, at 16:03, Camille Mazoyer <mazoyer at univ-tln.fr> wrote:
> 
> Hi Martin,
> Thanks for your reply.
> 
> Le 14/03/2017 à 15:11, Martin Losch a écrit :
>> Hi Camille,
>> 
>> did you see this:
>> <http://mitgcm.org/public/r2_manual/latest/online_documents/node236.html#SECTION00731560000000000000>
> Yes I read this, I understand it, and I checked on the code source how it is written also.
>> ? In your situation, you only need to worry about OBCS_balanceFacW and OBCS_balanceFacS. Your idea about the relative factor values is correct. If you have an extra volume flux “FLX” that needs to be distributed between your two boundaries, it will happen like this (provided the factors are >=0):
>> extraFluxW = FLX * OBCS_balanceFacW/(OBCS_balanceFacW+OBCS_balanceFacS)
>> extraFluxS = FLX * OBCS_balanceFacS/(OBCS_balanceFacW+OBCS_balanceFacS)
> 
> I agree that I only have to worry about OBCS_balanceFacW and OBCS_balanceFacS. your explanations correspond to what I understand.
>> 
>> How you actually choose the factors depends on your intuition and the (relative) quality of the data that you prescribe.
> => Ok. Thanks. Intuition and quality is not very easy to relate to a factor for me.
> Is it quite common to use a coeff equals to 1 everywhere if the datas come from a bigger-scale run ?
> My intuition is to use =1 everywhere, because the OBC datas quality is the same for each OBC.
> 
> Thank you,
> Camille
>> 
>> Martin
>> 
>>> On 14 Mar 2017, at 13:42, Camille Mazoyer <mazoyer at univ-tln.fr> wrote:
>>> 
>>> Dear all,
>>> 
>>> I have questions about the OBC balance term, and maybe somebody can answer me here. I have other questions, but I think I'm going to write other mails, to seperate questions.
>>> 
>>> I'm not sure about the balance coefficient I should use for my OBC: I have 2 OBC (south and West). Other boundaries are land. I'm working on a small scale area (200 Nx, 100 Ny, 75 Nz with a grid of 100m scale ). Could you correct me if I'm wrong, but what I understand about the balance coefficient is:
>>> 
>>> - the higher it is, the bigger is the correction
>>> 
>>> - so if I am confident about the South OBC values I give,  and not confident about the West values I give: I should give a small coeff like 1 for south, and a bigger coeff like 2, 3, 4 or more (?) for west OBC.  That means OBC South will be less changed than West OBC to ensure zero net transport. Usually, we give a small coeff for an OBC where we we know the values, and we want to set the values, and a big coeff for an OBC where we don't care if the flow is adjusted. Is that correct?
>>> 
>>> So if I'm not sure about one OBC more than the other,  I can have coeff=1 for both.
>>> 
>>> My last question: to sum up, how to choose coefficients?
>>> 
>>> Thanks a lot,
>>> 
>>> Camille
>>> 
>>> -- 
>>> ------------------------------------------
>>> Camille Mazoyer
>>> Phd Student
>>> Mediterranean Institute of Oceanography (MIO)
>>> Institut de Mathématiques de Toulon (IMATH)
>>> Université de TOULON
>>> Bat X - CS 60584
>>> 83041 TOULON cedex 9
>>> France
>>> tel: +33.4.94.14.24.50
>>> Email: mazoyer at univ-tln.fr
>>> http://mio.pytheas.univ-amu.fr/
>>> http://imath.fr/
>>> 
>>> 
>>> _______________________________________________
>>> 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