Problem loading zg for 6hrPlevPt

[sloosvel]

Describe the bug

When trying to work with the geopotential height for mip: 6hrPlevPt, the variable on BADC (and on ESGF) is called zg. But the tool gives the following error:

esmvalcore._recipe_checks.RecipeError: Unable to load CMOR table 'CMIP6' for variable 'zg' with mip '6hrPlevPt'

Which is right because there is no such variable called zg in this table. Is this an issue with the way the files are named? Or is this something the tool should be able to handle?

I tried to attach an example recipe but apparently attaching yml files is not supported, here is an example dataset:

• {project: CMIP6, mip: 6hrPlevPt, dataset: CNRM-CM6-1, exp: historical, ensemble: r1i1p1f2, grid: gr, start_year: 1982, end_year: 1982}

[zklaus]

Maybe it has something to do with the fact that in this table there are two variables (zg7h and zg27) both with the same out_name of zg that offer essentially the same information but on different levels.

Have a look at PCMDI/cmor#409 and https://github.com/PCMDI/cmor/blob/master/LibCV/PrePARE/out_names_tests.json to see how CMOR/PrePARE deals with that.

[sloosvel]

Yes I have noticed that in this table there are several cases in which the variable name differs from the out_name. I think that the problem comes from assuming that var_name = out_name when loading all the cmor information for the variables? In the VariableInfo class the out_name is not something that gets read.

[thomascrocker]

Just to add a +1 to this issue from me. One of the diagnostics I have been asked to implement in ESMValTool for the LaunchPAD project https://www.geog.ox.ac.uk/news/2019/0708-launchPAD.html requires geopotential height (zg) data from the zg27 CMIP6 table. It looks like #1099 should address the problem once it is merged.

Also tagging @ellendyer and @rjames-at-rwjames for future reference

[schlunma]

Moving this to v2.6 to follow the corresponding PR.

[sloosvel]

I don't think I will have time to take care of this for 2.6, sorry. I will remove it from the milestone.

[valeriupredoi]

sorry, no way this can be sorted out for M2.7 😢

[valeriupredoi]

hi @sloosvel how realistic would be a solution for this in v2.9 given code freeze is kicking in on Monday? 🍺

[sloosvel]

I already said a long time ago that I don't have much time to focus on this!

[valeriupredoi]

fair enough, will bump to 2.10 then, cheers 🍺