changes in table formats for CORDEX-CMIP7

[larsbuntemeyer]

Here, I'll try to summarize my understanding of changes in variable definitions for CMIP7:

• variable metadata becomes more independent from frequency and realm, table, etc using the branding.
• frequency attribute is not part of the variable metadata (cmor tables) but is defined in the dataset table.
• simple example notebook: https://github.com/WCRP-CMIP/cmip7-cmor-tables/blob/main/cmor_demo.ipynb

A related issue that documents changes: WCRP-CMIP/cmip7-guidance#35

[larsbuntemeyer]

There is the option "_cmip7_option": 1 which seems to "activate" the new formatting features. Let's see how far this will be compatible with other CV's. Maybe we can just keep the old table format but there might be some features with respect to CMIP7 we might miss.

[larsbuntemeyer]

Here is how the CV can be created from esgvoc:

• https://github.com/WCRP-CMIP/cmip7-cmor-tables/tree/main/tables-cvs

Let's check if we can

• create CORDEX-CMIP6 CV from esgvoc ([CORDEX-CMIP6] TypeError : Pydantic annotation aren't classes ESGF/esgf-vocab#193)
• use that with latest cmor and _cmip7_option

I fear hat we will run in some issues then...

[LluisFB]

I would suggest as soon we can adapt new CMIP7 tables and CVs the better. Otherwise we will have to do it later on, when simulations will we started to be running / done...

Sorry, but I am still quite overwhelmed with all these IT infrastructures. Do you have any idea in how we could test it, provide an example and see if we can assist with it?

[larsbuntemeyer]

I think, at least for the Fast Track, our CORDEX-CMIP7 CV and tables will probably be a subset or extension of the current tables and CV.