[MITgcm-support] boundary

Martin Losch Martin.Losch at awi.de
Mon Jul 30 06:12:35 EDT 2012


Mohammad,

I am not quite sure what you plan to do, but I have a few suggestions about your setup:
1. you use a cartesian grid for a 520 x 720 km domain (why did you turn it sideways? the top.mat file contains bathymetry where the geographically western opening towards the Persian Gulf is along the x-axis), but you also need to specify the coriolis parameter. I think for lat=25degN you could use something like f0=6.6e-5, you might also want to use a beta=4.8e-12. Alternatively you can use spherical coordinates, and then you wouldn't have to worry about the coriolis parameter (but it would require to rotate your topography).
2. you can use deltaT=100 (or whatever is required) instead of using more time stepping parameters (these are used for asynchronous time stepping)
3. If you want to simulate tides, you might be better of with an energy conserving time stepping scheme (the default AB2 scheme tends to damp fast waves), see internal_wave example for Crank-Nicolson parameters or the documentation: <http://mitgcm.org/public/r2_manual/latest/online_documents/node41.html>
4. your (vertical) diffusion/viscosity is very large. Typical values are diffKzT/S = 1.e-5 and viscAz = 1.e-3 or less. Lateral friction for a 2km grid also appears large to me. 
5. data.obs: with your domain orientation i think you should use these values:
  OB_Ieast(253:360)= 108*1, # geographically this is towards the south, but because of your model grid orientation it's the "eastern" edge of your grid
  OB_Jnorth(115:360) = 246*-1, # see above
  OB_Jsourth(29:118) = 90*1, #
If you want to prescribe periodic tides then you can modify obcs_calc.F (probably easiest) or prepare files and read them like this:
 useOBCSprescribe = .TRUE.,
 OBNuFile = 'OBmeridU.bin',
 OBSuFile = 'OBmeridU.bin',
 .... see documenation and code for possible files
<http://mitgcm.org/public/r2_manual/latest/online_documents/node236.html>, OBCS.h etc.

6. SIZE.h: why do you subdivide your tiles per processor? If you have 4cpu I would use this:
      PARAMETER (
     &           sNx = 130,
     &           sNy = 180,
     &           OLx =   2,
     &           OLy =   2,
     &           nSx =   1,
     &           nSy =   1,
     &           nPx =   2,
     &           nPy =   2,
     &           Nx  = sNx*nSx*nPx,
     &           Ny  = sNy*nSy*nPy,
     &           Nr  =   25)

Hope that helps

M.

On Jul 26, 2012, at 11:37 AM, mohammad akbari wrote:

> 
> 
> 
>     Thank you for help me.
>     Topgraphy is nx=260,ny=360 and nr=25. please see file size.h and
>     Topogrphy of  picture. In the left of picture the grid points 29:118,
>     in the right of picture the grid points 115:360 and  south 253:360 are
>     my open boundaries.
>     1-      I use from example tutrioal-global-oce-latlon. This example I add
>     obcs. Is presentation open boundary follwing  is correct?
>     2-      How can apply tide to west
> 
> - Done.
> 
> 
> <untitled folder.rar>_______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support




More information about the MITgcm-support mailing list