search

ACCESS-NRI / accessdev-Trac-archive

Archive accessdev Trac contents as issues
Apache License 2.0
0 stars 0 forks source link

Generate remapping weights for 0.25 ocean #382

Open penguian opened 3 years ago

penguian commented 3 years ago

type_non-CMIP6 | by aph502@nci.org.au

Generate remapping weights files for 0.25 degree ocean to N96 atmosphere grid

Issue migrated from trac:382 at 2024-01-31 18:36:11 +1100

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au changed status from assigned to accepted

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au set owner to aph502

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au commented

Martin,

I am in the process of doing this, but am having trouble deciding which atmosphere grid is appropriate.

I started with this one

/g/data/fs38/publications/CMIP6/CMIP/CSIRO-ARCCSS/ACCESS-CM2/historical/r1i1p1f1/fx/areacella/gn/latest/areacella_fx_ACCESS-CM2_historical_r1i1p1f1_gn.nc

but it seems to be on a staggered longitude grid 

 lon = 0.9375, 2.8125, 4.6875, 6.5625, 8.4375, 10.3125, 12.1875, 14.0625,

compared with some of the data variables, e.g.

/g/data/fs38/publications/CMIP6/CMIP/CSIRO-ARCCSS/ACCESS-CM2/historical/r1i1p1f1/3hr/uas/gn/latest/uas_3hr_ACCESS-CM2_historical_r1i1p1f1_gn_196001010300-197001010000.nc
 lon = 0, 1.875, 3.75, 5.625, 7.5, 9.375, 11.25, 13.125, 15, 16.875, 18.75,

Ideally the grid file would contain cell bounds and cell areas, but that is not essential.

Cheers

Aidan

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au commented

Hi Martin,

Scott explained the Arakawa C-grid was the explanation for the offset in the grids.

What remapping weights files are required? All three?

Cheers

Aidan

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au changed _comment0 which not transferred by tractive

penguian commented 3 years ago

@martin.dix@anu.edu.au commented

For CM2 we have all three.

rmp_cice_to_um1t_CONSERV_FRACNNEI.nc rmp_cice_to_um1u_CONSERV_FRACNNEI.nc rmp_cice_to_um1v_CONSERV_FRACNNEI.nc rmp_um1t_to_cice_CONSERV_DESTAREA.nc rmp_um1t_to_cice_CONSERV_FRACNNEI.nc rmp_um1u_to_cice_BILINEAR.nc rmp_um1u_to_cice_CONSERV_FRACNNEI.nc rmp_um1v_to_cice_BILINEAR.nc rmp_um1v_to_cice_CONSERV_FRACNNEI.nc

I'm not sure if your version of the 0.25 degree grid is the same as the one used for the files in /g/data/access/access-cm2/input_O025/cpl_n216?

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au _uploaded file lon_diff.png (230.0 KiB)_

Difference between longitudes

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au _uploaded file lat_diff.png (215.1 KiB)_

Difference between latitudes

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au commented

It looks like the grids are the same, with a maximum difference of 1.5e-5 in longitude, and 3.5e-6 in latitude

lon_diff.png

lat_diff.png

Is that close enough to use those remapping weights, or would you like me to generate them? It is no big deal to do so.

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au commented

I have just checked if the CICE grid (/g/data/ik11/inputs/access-om2/input_20201022_old_topog/cice_025deg/grid.nc) is the same as the ocean grid (/g/data/ik11/inputs/access-om2/input_20201022_old_topog/mom_025deg/ocean_hgrid.nc).

They are the same within 5e-14.

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au commented

Martin,

Scott tells me the horizontal grid for the N216 config is not the same as that used for CM2, so I'll go ahead and regenerate the remapping weights files.

I see there is some duplication with multiple remapping weights files for atmosphere to cice:

rmp_um1t_to_cice_CONSERV_DESTAREA.nc rmp_um1t_to_cice_CONSERV_FRACNNEI.nc rmp_um1u_to_cice_BILINEAR.nc rmp_um1u_to_cice_CONSERV_FRACNNEI.nc rmp_um1v_to_cice_BILINEAR.nc rmp_um1v_to_cice_CONSERV_FRACNNEI.nc

Can you let me know what ones are currently used for CM2 so I know what to create, i.e. conservative v bilinear, that sort of thing.

Cheers

Aidan

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au commented

Martin,

Which 1 degree CICE grid should I be using?

I found this one, but it doesn't have the v grid

/g/data/access/access-cm2/input_O1/cice_360x300/cice_grid_20101208.nc

Cheers

Aidan

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au commented

Hi Martin,

Sorry for my confusion, of course I need the 1 degree atmosphere grid. Sorry.

I guess it is this file?

/g/data/access/access-cm2/input_O1/cpl_n96/oasis3_grids_N96_06032014.nc

and the um1u.lat, um1u.lon, um1v.lat, um1v.lat, um1t.lat, um1t.lon variables.

Which of the namcouple files in /g/data/access/access-cm2/input_O1/cpl_n96 is the one currently used in CM2?

Cheers

Aidan

penguian commented 3 years ago

@martin.dix@anu.edu.au commented

That's the correct grid file.

namcouple_112fields_360_bil_GA7.1 is the one we're using.

penguian commented 3 years ago

@aidan.heerdegen@anu.edu.au commented

A first iteration of the remapping weights is documented in this repository

https://github.com/aidanheerdegen/access_cm2_025_remap_weights

I did not have time to check round-tripping of fields.

As explained here

http://climate-cms.wikis.unsw.edu.au/Interpolating_NetCDF_files_to_different_grids

that is very simple to do with the NCO tool ncks.

It is simple to regenerate these at any time if different options are required. Note for example that the option -i was used (--ignore_unmapped), which ignore unmapped destination points. In this case, the output matrix won't contain entries for the unmapped destination points.I believe this is not the same as the remapping weights used in the CM2.