[MITgcm-support] Questions about the OBCS Package

[Benjamin Ocampo]

Hi Matt:

The system I am simulating is a 120x120x80 curvilinear grid system with
heating from below.
Since I want to simulate convection with rotation and no boundary
conditions on the sides,
the OBCS package seems appropriate. The model spans roughly 40^o min. along
the
latitude (a square polar cap region of minimum length R_sphere*40^o).

I could try to simulate the entire sphere but then the efficiency of the
model drops off
significantly due to the fine grid (preferred method) needed to simulate
convection.
The regional model with the Orlanski configuration does indeed work so I
have no
problem with it.

As for i), to further clarify the question:

I) For the default OBCS, what mathematical model is used to simulate open
   boundary condition on the sides? Or how is it being simulated?

The reason why I ask this question is twofold:

 - To understand the mathematics and the algorithm in how the default
   open boundary condition is being simulated at sides.
 - To compare the mathematics of the default to the sponge and Orlanski
    algorithm.

This will help me further understand the mathematics behind regional
modelling in general.

Cheers,
Benjamin






On Sat, Jun 11, 2016 at 9:17 AM, Matthew Mazloff <mmazloff at ucsd.edu> wrote:

> Hi Benjamin
>
> i) The state will simply be prescribed on the open boundary.
>
> ii) sorry, I don’t use it
>
>
> I’m not sure your setup — you could provide more info. But most likely my
> advice will be to make a restoring region that is perhaps 10 cells thick
> and has a bathymetry with no gradient normal to the boundary. (That way you
> will not be forcing a normal flow that immediately has divergence from
> topographic gradients.) Prescribe the normal flow and T and S (and SSH if
> necessary for your setup) that you want. Set tangential flow to 0 and have
> a restoring on this tangential flow, which will damp the open boundary
> waveguide.
>
> These things I am suggested are intended to limit unphysical vertical
> flows.
>
> Open boundaries are no fun,
> Matt
>
>
>
> >
> > On Jun 10, 2016, at 9:45 AM, Benjamin Ocampo <rurik at ualberta.ca> wrote:
> >
> > Hi All:
> >
> > I have a couple questions about the OBCS package with regards to
> > regional modelling at the poles:
> >
> > i)   If I do not use Orlanski, Stevenson or Sponge configuration
> >      with the OBCS package, what method or algorithm does
> >      OBCS use to simulate Open Boundary Conditions (OBC)?
> >
> > ii)  Although I have read [Orlanski, 1976], I am still a little bit
> >      confused with the paper. As far as my understanding goes,
> >      the Orlanski OBC solves for the time dependent phase speed
> >      C of the fluid characteristic variable phi at grid points lying on
> >      (N-1) where N is the index locating the horizontal boundary
> >      sets the C at horizontal boundaries so long as C is b/w some
> >      range. Algorithm wise, what are the advantages and
> >      disadvantages of using the Orlanski configuration? When is
> >      the Orlanski configuration appropriate to use.
> >
> > Cheers,
> > Benjamin
> >
> > _______________________________________________
> > 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
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20160611/1dda58d2/attachment.htm>