Regridding/Mapping BRAN to ACCESS-OM2-01-CICE5 grid

COSIMA Ocean
1

Last topic creation for awhile for me (if anyone is getting concerned about my recent splurge of activity).

Anyhow, I’m not very hopeful that there will be any takers at helping solve this issue, but as I’ve been searching on and off for a little over a month on how to resolve this problem, I thought, I might as well put it up here.

Objective: generate re-gridding weights from BRAN to ACCESS-OM2-01-CICE5 grid

Stripped-down (i.e. just regular spherical coordinates in degrees) BRAN grid file:

Stripped-down (i.e. just regular spherical coordinates in degrees) CICE grid file. This file, in its original format can also be found ( /g/data/ik11/inputs/access-om2/input_20200530/cice_01deg/ ):

I’m then using the following command to attempt to regrid, but it either fails due to a segmentation fault (in the case of ‘bilinear’ or ‘conservative’), or it never finishes (after 48 hours of run-time!) with ‘patch’. I have tried many different formats of BRAN and CICE grids too! So I’d be very curious to hear anyone’s success and/or hints.

mpirun -np 32 ESMF_RegridWeightGen --64bit_offset -m patch -s ocean_tgrid.nc -d 0p1_cice_tgrid.nc -w map_BRAN_to_access-om2_cice_tgrid_patch.nc
2

Hey Dan,

I don’t see any particular reason why this wouldn’t work. Seems roughly inline with what we’ve done in the past.

This cookbook notebook is doing regridding to the 0.1° grid

https://cosima-recipes.readthedocs.io/en/latest/documented_examples/Regridding.html#gallery-documented-examples-regridding-ipynb

and this script is doing similar amounts of work with not dissimilar resources

What version of ESMF are you using? There was a bug with triple grids that Russ Fiedler fixed. You might be hitting that.

Cheers

Aidan

1 Like