Checks to add to ACCESS-OM3 Payu driver

[dougiesquire]

We've discussed adding checks to the Payu driver to be run at setup to help ensure the configuration is valid. This issue is to list/discuss the checks we might want to include.

• Check that the naming of MOM outputs in diag_table is consistent with the case_name set in nuopc_runconfig - see Still using CESM compset names in output filenames #106
• Check for consistency in physical constants - see Inconsistent physical constants #53
• Check for consistency across parameters set in multiple places - see Configuration parameters set in multiple places #108

[anton-seaice]

There are various grid files which are shared across models ..., I don't know if checking for consistency there makes sense (or is possible, the caps are already doing some checks).

e.g. from ice_ic:

  bathymetry_file = "./input/topog.nc"
  grid_atm = "A"
  grid_file = "./input/grid.nc"
  grid_ice = "B"
  grid_ocn = "A"
  grid_type = "tripole"
  kmt_file = "./input/kmt.nc"

These files names also need to match those in manifests ... although the models already have some error handling for that

[dougiesquire]

Possibly check that MOM outputs will not have missing values in coordinates - see COSIMA/cosima-recipes#325 (comment)

[anton-seaice]

Check that the metadata.yaml file has been updated for the experiment?

[dougiesquire]

Check consistency between stop frequency in nuopc.runconfig and MOM diag_table - see #94

[anton-seaice]

Check consistency between WW3 switches and WW3 config

(p.s. some things in this issue might make more sense as check in the nuopc driver, I haven't thought about it a lot)

[aekiss]

Linking to related discussion: payu-org/payu#421

[dougiesquire]

Check that nprocs in ice_in matches ice_ntasks in the nuopc.runconfig - see ACCESS-NRI/access-om3-configs#48 (comment)

[anton-seaice]

Check that nprocs in ice_in matches ice_ntasks in the nuopc.runconfig - see COSIMA/MOM6-CICE6#48 (comment)

I will investigate how this is set in the nuopc driver. We might be able to put a code fix in instead.

[dougiesquire]

I will investigate how this is set in the nuopc driver. We might be able to put a code fix in instead.

That would definitely be preferred

[anton-seaice]

Check that nprocs in ice_in matches ice_ntasks in the nuopc.runconfig - see COSIMA/MOM6-CICE6#48 (comment)

I will investigate how this is set in the nuopc driver. We might be able to put a code fix in instead.

I asked CICE-Consortium/CICE#945

[CodeGat]

Coming here from ACCESS-NRI/model-config-tests#52 - will give implementing these as QA checks a go!

[anton-seaice]

Consolidated list from above suggestions:

• Check that the naming of MOM outputs in diag_table is consistent with the case_name set in nuopc_runconfig - see Still using CESM compset names in output filenames #106
• Check for consistency in physical constants - see Inconsistent physical constants #53 and Not using mesh areas #240 (comment)
• Check for consistency across parameters set in multiple places - see Configuration parameters set in multiple places #108
• Check that nprocs in ice_in matches ice_ntasks in the nuopc.runconfig - done in CICE main CMEPS driver
• Check consistency between WW3 switches and WW3 config
• Check consistency between stop frequency in nuopc.runconfig and MOM diag_table - see Datebug in MOM cap #94
• Possibly check that MOM outputs will not have missing values in coordinates - see More pint + cf-xarray usage in examples cosima-recipes#325 (comment)