[MITgcm-support] sponge package in OBCS in ocean modeling

[Jody Klymak]

Dear Yangxin,

You don’t *need* `obcs_sponge.F` in your model code if you don’t modify it.  However, if you use it, the sponge will be applied to your mode-1 internal wave forcing on the west side as well your absorber on the RHS.  That means that a thick area on the LHS is being sponged to the same phase of the internal tide whereas if the wave were propagating through that thick area its phase would change.  Whether you care about this depends on the level of fidelity you want in your incoming wave versus what you specify in the code.  

I’ve done this in many two-d simulations, and it works fine, you juts need to diagnose the incoming wave (and make the sponge be far away from your region of interest.)  The advantage of doing this versus having an unsponged internal wave source is that any back-reflected waves from your region of interest are absorbed and don’t reflect back into the domain.  Without the sponge they definitely will back reflect.  

I’ve always used Carter et al 2008 (JPO) as a useful reference for the benefits of setting sponges.  I did exactly what you are doing in Klymak et al 2013 (JPO) looking at iternal tide scattering using the method above.

For Klymak et al 2015 (JPO) we had to be more careful about the incoming tide, so instead of using obcs, I used rbcs. That lets you set the time constant (strength of the sponge) everywhere in the domain, and have a time-propagating signal where you want there to be one and an absorber on the downstream side.  I prefer this methodology, but it is more expensive.  

Cheers,   Jody




> On Nov 9, 2017, at  10:50 AM, Yangxin He <y67he at uwaterloo.ca> wrote:
> 
> Hi,
> 
> I am using OBCS package to send in a mode-1 internal wave from the left/western boundary, and I want to add a sponge layer only at the right/eastern boundary. I have used rigid lid for my surface/northern boundary and no slip for the bottom/southern boundary. I have set useOBCSsponge=.TRUE. in data.obcs.
> My questions are about setting the sponge layer.
> 1) 
> There is a routine "obcs_sponge.F" under pkg/obcs. Is this the only routine I need to add in to my model code for the sponge layer?
> 2)
> Since I only need sponge layer for the eastern boundary, if I erase everything about the other 3 boundaries in the obcs_sponge.F, will it work?
> 3)
> I would like to understand the theory behind this sponge layer. Can you provide a reference/paper etc. that the code in obcs_sponge.F is based on?
> 
> Thanks a lot
> Yangxin
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org <mailto:MITgcm-support at mitgcm.org>
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support <http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20171109/793dd409/attachment.html>