Add EC-Earth3-ESM-1

[treerink]

Add the EC-Earth3-ESM-1 configuration.

See also ECE portal issue 1312-226.

[treerink]

Changes in the add-esm-configuration branch for both testing genecec & ece2cmor3.

[etiennesky]

Is there anything blocking the merge of this branch? It has been open for a few months and we are in the middle of production with this model version.

[etiennesky]

I think there is missing the required changes to modify ece2cmor3/scripts/modify-metadata-template.sh ?

[treerink]

I think there is missing the required changes to modify ece2cmor3/scripts/modify-metadata-template.sh ?

Yes, see 0bd5e26. Sorry this took so long.

[treerink]

I will start testing this branch with running genecec and see how well it goes and how many changes it gives.

For co2box I took: grid_label = gn From CMIP6_CVs /CMIP6_grid_label.json one could also consider grid_label = gm, but if I have the good understanding of the co2box model gn is most appropiate.

[treerink]

For the nominal resolution for the co2box model (see also #785) I took 10000 km according to the general description in this Appendix 2:
else nominal_resolution = “10000 km”

[treerink]

I have added the PISM component for the metadata part already here, though the PISM cmorisation component is not implemented yet.

[etiennesky]

I will start testing this branch with running genecec and see how well it goes and how many changes it gives.

For co2box I took: grid_label = gn From CMIP6_CVs /CMIP6_grid_label.json one could also consider grid_label = gm, but if I have the good understanding of the co2box model gn is most appropiate.

the grid of the co2box moel is a global (0D) one, so it could be gm, but that means a processing from some spatial data to a global average, so gn is fine!

[etiennesky]

For the nominal resolution for the co2box model (see also #785) I took 10000 km according to the general description in this Appendix 2: else nominal_resolution = “10000 km”

thats is not quite exact since it's a global grid, but I don't think there is something that fits well, but this works because

else nominal_resolution = “10000 km”

[klauswyser]

I processed many experiments for OptimESM, including output from the co2box model, always setting nominal_resolution = "10000 km". This worked well, no problem encountered.

[klauswyser]

I have added the PISM component for the metadata part already here, though the PISM cmorisation component is not implemented yet.

What do you think is missing for PISCES? The OptimESM runs were done with PISCES. I have run ece2cmor in spring (plain e5bc947f56c ) and the PISCES output was nicely processed. For example I can find chldiatos = "Surface Mass Concentration of Diatoms Expressed as Chlorophyll in Sea Water" in the cmorised output.

[treerink]

PISM <=> PISCES ?
Pisces should be fine, I am talking about the ISM PISM ... (?)

[treerink]

So far I witness only expected changes with running the full genecec, so with a bit of luck I will merge this branch at the end of the day into the master.

[treerink]

Then I have to change from:

./genecec-for-individual-experiments.sh ../resources/miscellaneous-data-requests/optimesm-request/optimesm-request-EC-EARTH-CC-varlist.json        CMIP        historical               EC-EARTH-CC      optimesm

to

./genecec-for-individual-experiments.sh ../resources/miscellaneous-data-requests/optimesm-request/optimesm-request-EC-EARTH-ESM-1-varlist.json        CMIP        historical               EC-EARTH-ESM-1      optimesm

and see whether updating the optimism ECE output control files will go smooth or current manual commits have to be sorted out.

Other experiments to be added there? Or is the experiment naming still under discussion - I lost the status on that.
Probably not so relevant and necessary because we use the same request everywhere. Only for the experiment metadata it matters, but one or the other way does not differ much in work (i.e. manually adjusting it in the current historical one).

[treerink]

I see Klaus that in r10143 you chose in metadata-*-co2box-template.json grid_label = gm and for grid you used box model where I used global grid now from another suggestion. What should I do, keep what I have and overwrite the optimesm branch or adjust one or both cases such that it matches the current optimesm branch?

[treerink]

(the comment attribute you added I will take over)

[treerink]

For co2box grid I used now global box model.

[treerink]

Branch merged into master with #831

[klauswyser]

For co2box grid I used now global box model.

This is a somewhat unlucky choice. The metadata for the co2box model were suggested in 1312-180, This suggestion was open for discussion but no comments were made so these changes have been comitted (see 1312-216). If you now change these metadata it means that all the data that have been cmorised will be different from those that will be cmorised with newer versions of the metadata template. Unless anybody thinks the old metadata are wrong for any reason I would suggest we revert the changes in the metadata to avoid confusion.

[treerink]

I just chose something which looked clear to me ;)

Same applies to grid_label = gm vs grid_label = gn

I am fine with any choice, I just want to finish it :)

[klauswyser]

Sure, no problem. However, changing metadata violates backward compatibility, something we should avoid if not necessary. Several experiments and members from OptimESM have already been cmorised, so I suggest we stick with the metadata that we used there.

[treerink]

Adjusted in e5d732b