[MITgcm-support] is Orlanski radiation condition working in the rotating rank of MITgcm?

[Martin Losch]

Yi-Chih,

I am sorry. I don’t understand what you are doing (and why, but that’s maybe not so important).

Does you config have anything to do with verification/rotating_tank?
This experiment mimics a small tank on a turntable in a lab. It’s not clear to me how to upscale that to the Arctic without fundamentally changing the parameters of the model. 
Thus, I cannot follow your conclusions that you can fix anything with open boundaries. You need to make sure that your flow parameters are set correctly (eg. horizonal and vertical viscosity, Coriolis parameter, etc ….)

Using the grid data from arctic40km or cs_4km confuses me even more in this context. If you do things like this, the coriolis parameter should be automatically set according to the latitudes in the domain. You cannot change the radius of the earth without adjusting the grid information (because grid spacing ~ radius of the earth).

I don’t think anyone can help you with this unless you are clearer about you configuration. Maybe it would be a good idea to 
(1) post your configuration somewhere, namelist files, “code” directory, “input files" (please no large files on this list)
(2) post a plot of the “disturbances” that you observe, with and without orlanski obcs

Martin


> On 14. Dec 2018, at 19:03, Yi-Chih Huang <dscpln at gmail.com> wrote:
> 
> Hello Martin,
> 
>     I am using the curvilinear coordinate on the rotating_tank for latitudes 75N-90N by applying the old grids on MITgcm_contrib/arctic40km/ and MITgcm_contrib/arctic/cs_4km.  Generally, the numerical disturbances occurred around the lateral boundary and moved back to the center of the domain.  Thus, the default lateral boundary condition does not seem to be good enough to handle these disturbances.  My experiences with Orlanski radiation condition were really good.  Therefore, I updated b.c. by adding obcs in code/packages.conf and having a run/data.obcs as I showed earlier.  Although no error messages about boundary condition are in the logfiles of execution, the output patterns are almost exactly the same as the original one.  If my updating processes for b.c. was wrong, what else should I do to make it working?  
> 
>     By the way, I will change the radius of the planet.  Those old gids on MITgcm_contrib/arctic40km/ and MITgcm_contrib/arctic/cs_4km will not be suitable for my work on curvilinear coordinates accordingly.  Can I change the radius of the planet and focus on latitudes 75N-90N by working on the grids of CS or LLC?  I know very little about them.
> 
>     Thanks,
> 
>           Yi-Chih
> 
> Date: Fri, 14 Dec 2018 13:50:11 +0100
> From: Martin Losch <Martin.Losch at awi.de>
> To: MITgcm Support <mitgcm-support at mitgcm.org>
> Subject: Re: [MITgcm-support] is Orlanski radiation condition working
>         in the rotating rank of MITgcm?
> Message-ID: <CABCD918-87C9-43E4-9B04-053E15DE67EF at awi.de>
> Content-Type: text/plain; charset="utf-8"
> 
> Hi Yi-Chih,
> 
> I can only guess what you are trying to do: You use the verification/rotating_tank setup, you probably change f0 from 0.5 to something more ?polar?, e.g. 1.45e-4 for a latitude of 85.5 degN, but then your grid spacing needs to be adjusted (from the lab-scales in the original rotating_tank experiment), etc ? There are many places, where you may not satify stability criteria. See here <https://mitgcm.readthedocs.io/en/latest/examples/examples.html> for examples of stability criteria in other tutorial experiments.
> 
> Introducing open boundary conditions to get rid of grid scale noise is not a good idea. Normally open boundar condition tend to be more of a problem than a solution.
> 
> Martin  
> 
> On Fri, Dec 14, 2018 at 8:27 AM Yi-Chih Huang <dscpln at gmail.com> wrote:
> Dear MITgcm experts,
> 
>     I ran into numerical disturbances around the boundary working on the rotating rank of MITgcm in the Arctic regions.  I tried to apply Orlanski radiation condition to the rotating rank in data.obcs below.  However, the patterns with the data.obcs are almost the same as the original one.  There is no error message in the log files.  Do you think Orlanski radiation condition is working in the rotating rank of MITgcm?  How should I update the boundary conditions on MITgcm?
> 
> # Open-boundaries
>  &OBCS_PARM01
>  OB_Ieast=102*-1,
>  OB_Iwest=102*1,
>  OB_Jnorth=130*-1,
>  OB_Jsouth=130*1,
>  useOrlanskiEast=.TRUE.,
>  useOrlanskiWest=.TRUE.,
>  useOrlanskiNorth=.TRUE.,
>  useOrlanskiSouth=.TRUE.,
> # useOBCSbalance=.TRUE.,
> # OBCS_balanceFacW= 1.,
> # OBCS_balanceFacN= 0.,
>  OBCS_monitorFreq=3000.,
>  &
> 
>  &OBCS_PARM02
>  CMAX=0.5,
>  cvelTimeScale=10000.,
>  useFixedCEast=.FALSE.,
>  useFixedCWest=.FALSE.,
>  &
> 
>     Thanks much,
> 
>                Yi-Chih
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support