[MITgcm-support] parameters in obcs sponge layer

Yangxin He y67he at uwaterloo.ca
Wed Mar 21 12:49:48 EDT 2018


Hi Jody,


Sorry, I wasn't clear. Yes, I did mean a mode-1 internal tide coded in obcs_calc.F.

For the obcs_sponge.F, I deleted the part that has the western/left boundary sponge, so this way I only have sponge on the eastern/right boundary.


I see in your data.obcs, you only got 10 points in the sponge. thats amazing. I am gonna try a much smaller time scale and 100 sponge points to see what happens.

If this doesn't work, I will try reducing the the grid size in the sponge. Just to clarify, I can read in a grid with stretched grid sizes near the right boundary and I don't need another package to do this, right?


Thanks


Yangxin

________________________________
From: MITgcm-support <mitgcm-support-bounces at mitgcm.org> on behalf of Jody Klymak <jklymak at uvic.ca>
Sent: Wednesday, March 21, 2018 12:35:09 PM
To: mitgcm-support at mitgcm.org
Subject: Re: [MITgcm-support] parameters in obcs sponge layer

Dear Yangxin,

Something like that should work.  Not sure what you mean by “internal waves”, but if you mean you force a mode-1 wave, you can do that via calculation in `obcs_calc.F`, but note that will apply *one* value of the boundary condition through the sponge.  Your sponge is 400-km long, so the wave won’t be very freely propagating.  For those sorts of problems, I suggest using `rbcs` package, and set the problem up with your ideal wave having spatial and temporal structure, so the wave is “freely” propagating through the incoming sponge.   You can find an example of this at https://github.com/jklymak/TwoDMode1Reflect

For the absorbing sponge, it seems to me your timescales are far too long to absorb fast mode-1 waves.

For these sorts of problems, I always stretch the grid near the bdys and far from my region of interest.  You don’t need a lot of resolution in the sponge region, so if you make it quite large it usually does a good job of absorbing everything.

Cheers,   Jody


On 21 Mar 2018, at 09:09, Yangxin He <y67he at uwaterloo.ca<mailto:y67he at uwaterloo.ca>> wrote:

Hey there,

Here is my toy problem,
I have no slip at the bottom, rigid lid at the top, I am sending internal waves of M2 tidal frequency using obcs_calc.F from the left boundary with max(U)=0.02 m/s, I want to impose a sponge layer using orcs_sponge.F on the right boundary.
1 grid cell in the y-direction, i.e. it is a 2d simulation.
50 points vertically with a uniform gird size of 40m.
2000 points horizontally with a uniform gird size of 500m.
An example of my data.obcs file looks like:
# Open-boundaries
 &OBCS_PARM01
 OB_Ieast=-1,
 OB_Iwest=1,
 useOBCSsponge=.TRUE.,
 &
 &OBCS_PARM03
 spongeThickness=800,
 Urelaxobcsinner=89424.,
 Vrelaxobcsinner=89424.,
 Urelaxobcsbound=4471.,
 Vrelaxobcsbound=4471.,

 &
What I want to achieve is this sponge layer shall work well so that all waves are being absorbed, no reflections back and it should work even after 30 or 40 tidal periods or even longer.
I have tried various values for the five parameters listed in &OBCS_PARM3, following the basic principle,
1) a fairly large U/Vrelaxobcsinner to make sure the forcing is small when the waves hit the sponge.
2) a fairly small U/Vrelaxobcsbound to make sure the forcing is still large enough so the waves can be fully absorbed.
1) the thicker the sponge layer is, the better it shall work.
The problem is none of the parameters I tried worked.

I assume somebody must have done similar work as this one. I would really appreciate some insights as to what values I shall assign those parameters.

Thanks

Yangxin


_______________________________________________
MITgcm-support mailing list
MITgcm-support at mitgcm.org<mailto:MITgcm-support at mitgcm.org>
http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support

--
Jody Klymak
http://web.uvic.ca/~jklymak/





-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20180321/64dc59f3/attachment.html>


More information about the MITgcm-support mailing list