inconsistent DRS for `COSMO-CLM-6-0-clm3`

[larsbuntemeyer]

For example:

/mnt/CORDEX_CMIP6_tmp/sim_data/CORDEX-CMIP6/DD/EUR-12/CLMcom-Hereon/ERA5/evaluation/r1i1p1f1/COSMO-CLM-6-0-clm3/v1-r1/mon/tas/v20230222/tas_EUR-12_ERA5_evaluation_r1i1p1f1_CLMcom-Hereon-CCLM-6-0-clm3_nukleus-x0n1-v1_mon_196001-196012.nc

has inconsistent attribute values for version_realization: In the path it's v1-r1 while it's nukleus-x0n1-v1 in the filename.

[gnikulin]

There is another inconsistency: "the institution that is responsible for CORDEX simulations (institution_id) and model acronym (source_id) are 2 different DRS elements, i.e. separated by the underscore “_” in the file name", now they are separated by "-", should be CLMcom-Hereon_CCLM-6-0-clm3

[larsbuntemeyer]

This data does not show up in the catalog since parsing failed. NUKLEUS runs have been following some intermediate standard (long before CORDEX-CMIP6 archive specs were published). Also global attributes differ. I think for ESGF publication, these datasets need to be revisited. For the evaluation, i would be fine for now, if

• parsing DRS is successful, hence, data shows up in the catalog and
• dataset is CF compliant (quick check seems to be succesful for COSMO-CLM-6-0-clm3).

Otherwise, we can not guarantee reproducibilty if evaluation is based on intermediate datasets. Pinging @GeyerB.

[jesusff]

It would be nice to have the final name registered, though (WCRP-CORDEX/cordex-cmip6-cv#324). If we go on with a temporary name, we might end up referring in the evaluation to a source_id which cannot be found in ESGF later on.

[larsbuntemeyer]

Just a heads up here, since this source_id is not included in the catalog yet. If this source_id should go into the overview evaluation, it should provide fx variables sftlf for masking and orog for height temperature corrections. The directory

• /mnt/CORDEX_CMIP6_tmp/sim_data/CORDEX-CMIP6/DD/EUR-12/CLMcom-Hereon/ERA5/evaluation/r1i1p1f1/COSMO-CLM-6-0-clm3/v1-r1/fx/

is empty and probably not synced correctly.