search

CICE-Consortium / CICE

Development repository for the CICE sea-ice model
Other
58 stars 132 forks source link

Add tripolar netcdf grid to input data #977

Open anton-seaice opened 1 month ago

anton-seaice commented 1 month ago

I am looking at writing tests for the grid load routines, and there doesn't seem to be any test coverage for a tripolar grid in netcdf format? Have I missed something?

p.s. We should add one.

apcraig commented 1 month ago

That's probably correct. The grid files used in Consortium testing were generated in the user community. It's possible we don't have an example of a tripole netcdf grid file.

anton-seaice commented 1 month ago

Shall we add a tx1nc ? Is this a CESM grid? Someone make a bathymetry netcdf for it already, so we could create a netcdf if needed.

dabail10 commented 1 month ago

The tx1 is never used. It is just meant to be a test for tripole grids. I think that we should stick with standard tripole grids that are actually used with MOM. We have a 2/3 degree tripole. NOAA/EMC uses a 0.5-degree and 0.25 degree. Can we also couple this with reading MOM supergrids?

anton-seaice commented 1 month ago

I suggest we add both a "normal/CICE/POP" grid now, and a supergrid when that gets implemented ? The lower the resolution the better in my view - they are to demonstrate the tripole works and tests run faster on a lower resolution

(ill also suggest we just drop the binary grid formats, but thats a different discussion :-))

anton-seaice commented 1 month ago

I realised that the forcing data in our test cases is offline regridded to the grid format used. As we already have forcing data in "tx1" format, its a strong argument to re-use "tx1" in a netcdf and a mom-mosaic-netcdf.

anton-seaice commented 1 month ago

Does tx1 exist already in a tripole netcdf and mom supergrid formats ?

Or should we just create a new tripole grid (at say 2 or 4 degrees nominal) + a new set of forcing data and use that instead?

dabail10 commented 1 month ago

Definitely not in MOM supergrid format. I'm not sure we even have a netcdf version. Though, this could be created easily enough.