[MITgcm-support] Modelling at the Poles

Mon Apr 18 17:52:15 EDT 2016

Hi Jean-Michel:

Thank you for your advice. I will probably not use exch2 for now since I am
currently just
trying to get a serial run to work.

I have looked into the arctic examples that Dimitris has suggested and I
have decided to try to
mimic the arctic40km example except I want to generate the grid such that
it is located at the
North pole.

However, based on the /MITgcm_contrib/high_res_cube/matlab-grid-converter
and
/MITgcm_contrib/high_res_cube/matlab-grid-generator files, I am not familiar
to how gengrid_fn.m sets the latitude and longitude ranges as well as the
resolution for a
single face (so as to avoid the singularity at the North Pole).

Is there a way to modify gengrid_fn.m to set the longitude and latitude
ranges and to output
for a single face?

I also understand that I need to run convertMITgrid.m to output the
appropriate MITgrid files.
Using MATLAB, I executed the program file gen32grid.m, generating the
directory file
cs-conf-c-32-512/. Then, I used convertMITgrid.m in the directory,
outputting
the following error:

convertMITgrid
ans =
Not enough data was available to be read: off EOF?
ans =
Not enough data was available to be read: off EOF?
ans =
Not enough data was available to be read: off EOF?
ans =
Not enough data was available to be read: off EOF?
ans =
Not enough data was available to be read: off EOF?
ans =
Not enough data was available to be read: off EOF?
ans =
Not enough data was available to be read: off EOF?
ans =
Not enough data was available to be read: off EOF?
ans =
Not enough data was available to be read: off EOF?
ans =
Not enough data was available to be read: off EOF?
Attempted to access yg(1,1,3); index out of bounds because size(yg)=[2,2,1].
Error in convertMITgrid (line 24)
yg(1,end,[1 3 5])=yg(1,1,3);

Did I do something wrong running with the MITgrid generation process? I
should mention that
I did not edit any of the files.

Cheers,
Benjamin

On Sat, Apr 16, 2016 at 10:10 AM, Jean-Michel Campin <jmc at mit.edu> wrote:

> Hi Benjamin,
>
> trying to clarify few things regarding pkg/exch2 and grids:
> 1) If you use a non-obvious subdomain topology, e.g., cubed-sphere type
> topology
>    (in this case: useCubedSphereExchange=.TRUE., in eedata), you need to
> use pkg/exch2.
>   If it's a regular cube (squared faces) without blank tile, default
> setting works
>   fine and you don't need parameter file "data.exch2"; Otherwise, e.g.
> Lat-Lon cap
>   grid or with blank tiles, you will need to set few parameters in
> "data.exch2"
>   There are examples of data.exch2 for lat-lon cap grids in
> MITgcm_contrib, e.g. here:
>
>  MITgcm_contrib/gael/verification/global_oce_llc90/input.ecco_v4/data.exch2
>   or here:
>    MITgcm_contrib/llc_hires/llc_270/input/data.exch2
>   and some example also (with a README file there) here:
>    MITgcm/utils/exch2/input/data.exch2.*
> 2) If you are using a simple grid topology (e.g., cartesian grid,
> SphericalPolarGrid,
>   or curvilinear grid all in 1 piece, i.e, single face), you can use
> pkg/exch2
>   (and without useCubedSphereExchange=T and without blank tile, the
> default pkg/exch2
>   setting works) but you don't need to except if you have blank tiles.
>   The verification experiment: verification/global_ocean.90x40x15 is an
> example of
>   a simple grid set-up using pkg/exch2:
>    a) without blank tile if non-MPI built & run (code/SIZE.h, dont'use any
> data.exch2)
>    b) with blank tiles if building and running with MPI (code/SIZE.h_mpi,
> input/data.exch2.mpi)
> 3) If you want to use a curvilinear grid (usingCurvilinearGrid=.TRUE.,),
> you need to specify
>   file(s) containing all the grid information (horizGridFile= ...).
>   It could be usefull to start by using an already existing grid as
> Dimitris suggested.
>   And from there, you could try to generate your own grid file(s). One
> thing to keep
>   in mind is that, for now, MITgcm only supports locally orthogonal grids.
>   Again, you can decide not to use pkg/exch2 if the grid is all in one
> piece (and no blank
>   tile is used).
>
> Cheers,
> Jean-Michel
>
> On Sat, Apr 16, 2016 at 09:17:13AM -0600, Benjamin Ocampo wrote:
> > Hi Dimitris:
> >
> > Thank you for the response. I will run cs_4km code and determine how I
> can
> > set the latitude and
> > longitude ranges including the resolution.
> >
> > Cheers,
> > Benjamin
> >
> > On Sat, Apr 16, 2016 at 5:53 AM, Menemenlis, Dimitris (329C) <
> > Dimitris.Menemenlis at jpl.nasa.gov> wrote:
> >
> > > There is a couple example MITgcm Arctic Ocean configurations here:
> > > http://mitgcm.org/viewvc/MITgcm/MITgcm_contrib/arctic/
> > > and here: http://mitgcm.org/viewvc/MITgcm/MITgcm_contrib/arctic40km/
> > >
> > > On Apr 15, 2016, at 9:59 PM, Benjamin Ocampo <rurik at ualberta.ca>
> wrote:
> > >
> > > Hi All:
> > >
> > > My apologies if I have been littering the support with unanswerable
> > > questions but
> > > I appear to be at a conundrum.
> > >
> > > My current objective is to try to get a spherical regional model to
> > > successfully work
> > > at the North pole but I am at a loss. As far as I know, there exists a
> > > singularity
> > > at latitude = 90N due to the spherical laplacian term in the
> Navier-Stokes
> > > equation and hence convenctional lat-lon grids do not work.
> > >
> > > Thus, I have read into using the exch2 package to determine if it is
> > > feasible to
> > > get a regional model working at the North Pole. However, I am at a
> loss as
> > > I am
> > > not familiar in setting up the data.exch2 file (for a single face)
> with a
> > > fixed lat-lon
> > > domains (eg. 0E to 360E, 80N to 90N or 0E to 360E, -90N to -80N) with
> > > dlat and dlon resolution (eg. dlat = 0.1,dlon = 0.1 resolution).
> > >
> > > Does using the package exch2 provide a feasible method for regional
> > > modelling at the poles? If so, how do I correctly set up the data.exch2
> > > file?
> > >
> > > 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
> > >
> > >
>
> > _______________________________________________
> > 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/20160418/ddccf70f/attachment.htm>