use ACCESS-OM2 grids, topography, etc

[aekiss]

The ACCESS-OM3 configurations currently use an offset-pole 1° horizontal grid using these grid, mesh, topography and initial condition files.

We should change these to the tripolar grids used at each resolution in the ACCESS-OM2 suite as defined in /g/data/ik11/inputs/access-om2/input_20201102.

[aekiss]

[UPDATE: actually we probably want to use grid files if we can - see below]

It looks like we'll need to generate unstructured mesh file representations of the structured ACCESS-OM2 grids.

@dougiesquire, @kieranricardo - I don't suppose you know how to generate unstructured mesh files like these?

/g/data/ik11/inputs/cime/inputdata/2023-03-10/share/meshes/*.nc
/g/data/ik11/inputs/cime/inputdata/2023-03-10/cpl/gridmaps/*/*.nc

A couple were generated by ESMF_RegridWeightGen. The rest don't say, but maybe they used ESMF_RegridWeightGen too?

[aekiss]

/g/data/ik11/inputs/cime/inputdata/2023-03-10/share/meshes/*.nc seem to use the ESMF Unstructured Grid File Format, but /g/data/ik11/inputs/cime/inputdata/2023-03-10/cpl/gridmaps/*/*.nc are different.

[aekiss]

but /g/data/ik11/inputs/cime/inputdata/2023-03-10/cpl/gridmaps/*/*.nc all have the same format and two of them used ESMF_RegridWeightGen:

$ for f in /g/data/ik11/inputs/cime/inputdata/2023-03-10/cpl/gridmaps/*/*.nc; do  ncdump -h $f | grep ESMF_RegridWeightGen && echo $f; done
		:run_command = "ESMF_RegridWeightGen --ignore_unmapped -m bilinear -w map_gx1v6_TO_ww3a_blin.150430.nc -s /glade/p/cesmdata/cseg/mapping/grids/gx1v6_090205.nc -d /glade/p/cesmdata/cseg/mapping/grids/ww3a_120222.nc" ;
/g/data/ik11/inputs/cime/inputdata/2023-03-10/cpl/gridmaps/gx1v6/map_gx1v6_TO_ww3a_splice_150428.nc
		:run_command = "ESMF_RegridWeightGen --ignore_unmapped -m bilinear -w map_ww3a_TO_gx1v6_blin.150430.nc -s /glade/p/cesmdata/cseg/mapping/grids/ww3a_120222.nc -d /glade/p/cesmdata/cseg/mapping/grids/gx1v6_090205.nc" ;
/g/data/ik11/inputs/cime/inputdata/2023-03-10/cpl/gridmaps/ww3a/map_ww3a_TO_gx1v6_splice_150428.nc

[aekiss]

/g/data/ik11/inputs/cime/inputdata/2023-03-10/cpl/gridmaps/*/*.nc are ESMF regrid weight files

[adele-morrison]

A couple of thoughts for perhaps slightly further down the track: 1. Perhaps it’s worth revisiting the ACCESS-OM2 grids if we plan to use these? e.g. in the 1/10th deg we currently use coarser than Mercator resolution around Antarctica, but given so much of our research focuses on this area, perhaps we should have finer res here instead? Likewise, is it possible or worth thinking about degrading resolution in northern hemisphere places we don’t care about? 2. I suggest we regenerate the bathymetry from scratch as we did for the panan, because we no longer have the B-grid advective restrictions that ACCESS-OM2 had.

…

On Tue, Jun 20, 2023 at 2:36 PM, Andrew Kiss ***@***.***> wrote: but /g/data/ik11/inputs/cime/inputdata/2023-03-10/cpl/gridmaps/*/*.nc all have the same format and two of them used ESMF_RegridWeightGen: $ for f in /g/data/ik11/inputs/cime/inputdata/2023-03-10/cpl/gridmaps/*/*.nc; do ncdump -h $f | grep ESMF_RegridWeightGen && echo $f; done :run_command = "ESMF_RegridWeightGen --ignore_unmapped -m bilinear -w map_gx1v6_TO_ww3a_blin.150430.nc -s /glade/p/cesmdata/cseg/mapping/grids/gx1v6_090205.nc -d /glade/p/cesmdata/cseg/mapping/grids/ww3a_120222.nc" ; /g/data/ik11/inputs/cime/inputdata/2023-03-10/cpl/gridmaps/gx1v6/map_gx1v6_TO_ww3a_splice_150428.nc :run_command = "ESMF_RegridWeightGen --ignore_unmapped -m bilinear -w map_ww3a_TO_gx1v6_blin.150430.nc -s /glade/p/cesmdata/cseg/mapping/grids/ww3a_120222.nc -d /glade/p/cesmdata/cseg/mapping/grids/gx1v6_090205.nc" ; /g/data/ik11/inputs/cime/inputdata/2023-03-10/cpl/gridmaps/ww3a/map_ww3a_TO_gx1v6_splice_150428.nc — Reply to this email directly, view it on GitHub <#36 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACA44U4NKI25P65ENE22PK3XMESDFANCNFSM6AAAAAAZMV4U2U> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[aekiss]

Yes, they would both be good to do further down the track. But first I'd like to have configs that are close to ACCESS-OM2 to reduce the number of things we need to debug and also so we can compare apples to apples.

[dougiesquire]

@dougiesquire, @kieranricardo - I don't suppose you know how to generate unstructured mesh files like these?

I don't off the top of my head, sorry. I downloaded all the inputs for those configurations from the CESM servers and many have little provenance about how they were generated.

[kieranricardo]

@aekiss I've created ESMF mesh objects (in Fortran) before but I've never attempted to save them to a file, sorry. I did a quick scan of the ESMF docs and I couldn't see a way to do this, which seems a bit weird to me so I probably just missed something

[dougiesquire]

@aekiss, I'm trying to work out where the CICE grid files in /g/data/ik11/inputs/access-om2 come from. Recent input versions all seem to point at /g/data/ik11/inputs/access-om2/input_20200530/cice_??deg/grid.nc but I can't find how these files were generated. Can you point me in the right direction?

[dougiesquire]

Note, this tools looks like it could be handy:

https://ufs-community.github.io/UFS_UTILS/cpld_gridgen/index.html

See also:

ufs-community/UFS_UTILS#647
https://github.com/DeniseWorthen/ufs-weather-model/wiki/Creating-new-resolutions-for-MOM6-and-CICE6-and-new-ICs-for-CICE6
https://github.com/DeniseWorthen/GridGen

[aekiss]

Hi @dougiesquire I actually don't know the origins of these files - they are part of the legacy configurations ACCESS-OM2 was based on. For consistency they should be derived from the mom supergrid ocean_hgrid, which is what https://ufs-community.github.io/UFS_UTILS/cpld_gridgen and https://github.com/DeniseWorthen/GridGen apparently do.

@micaeljtoliveira when you looked into this last year did you look at cice grids or just mom grids?

For the purpose of getting configurations together we could just reuse these files as-is, but it would also be nice to have some sort of provenance and workflow for them, so long as it isn't a rabbithole.

There's a slight weirdness in the 1° grid but we probably don't want to fix this just yet so we can compare apples with apples COSIMA/access-om2#271

[dougiesquire]

Thanks @aekiss. Yup I'm trying to push ahead with the existing OM2 grid, but keeping in mind that we probably want reproducible workflows for setting up the OM3 grids. Incidentally, are there reproducible workflows for generating the existing OM2 MOM grids? I could well be misunderstanding, but in the forum post you share above it sounds like folks have been having trouble recreating these grids using the FRE-NCtools make_hgrid tool.

[dougiesquire]

Ashley Barnes' Python tools look like they could be handy: https://github.com/COSIMA/regional-mom6/blob/main/regional_mom6/regional_mom6.py#L487

[micaeljtoliveira]

when you looked into this last year did you look at cice grids or just mom grids?

I looked only at the MOM6 supergrid.

in the forum post you share above it sounds like folks have been having trouble recreating these grids using the FRE-NCtools make_hgrid tool.

Yes, basically the make_hgrid tool does not have all the needed features. The one tool I found that seems to have all that we need, is this one:

https://github.com/NOAA-GFDL/ocean_model_grid_generator/

but I haven't had time to look into it in more detail.

[aekiss]

Rethinking my comment above re. meshfiles, @micaeljtoliveira has found that FMS2 doesn't support meshfiles, so we should change to grid files if at all possible. Otherwise we'd need to use FMS1 (as in the CIME-built CESM). Grid files are much easier to deal with too (e.g. plotting).

[dougiesquire]

Rethinking my comment above re. meshfiles, @micaeljtoliveira has found that FMS2 doesn't support meshfiles, so we should change to grid files if at all possible. Otherwise we'd need to use FMS1 (as in the CIME-built CESM). Grid files are much easier to deal with too (e.g. plotting).

I'm not sure this is going to be trivial. The MOM nuopc cap appears to allow internal generation of an ESMF grid, rather than passing an ESMF mesh, but I haven't been able to get this to work. I might focus for now on generating ESMF mesh/grid files for the OM2 grid (which also seems unnecessarily hard) and go from there

[dougiesquire]

I've moved the branch where I'm working on this from my fork into the main repo so that it's easier for others to see/contribute - https://github.com/COSIMA/MOM6-CICE6/tree/om2_grid_iss36

[aekiss]

Just for the record, this is a more actively updated version of the grid generator mentioned above
https://github.com/nikizadehgfdl/ocean_model_grid_generator
and make_hgrid (mentioned above) and other grid-related tools are available from
https://github.com/NOAA-GFDL/FRE-NCtools

[micaeljtoliveira]

Done in ACCESS-NRI/access-om3-configs#5