IPSL-CM6A-LR ssp370 r1i1p1f1 forcing data missing?

[doblerone]

Hi @jesusff
Can you double check the availability of the IPSL-CM6A-LR ssp370 r1i1p1f1 forcing data?
I got a comment that some variables are only available up to the year 2055 on ESGF.
Checking on ESGF with
https://esgf-data.dkrz.de/search/cmip6-dkrz/?source_id=IPSL-CM6A-LR&experiment_id=ssp370&variant_label=r1i1p1f1&frequency=6hrPt&variable_id=ta%2Cua%2Cva
I can see that ua and va are only available up to 2055 while ta up to 2100.

[jesusff]

Hi Andreas,
Thank you for reporting. Yes, Ole reported this problem short ago and IPSL is looking into how to solve it (SSP370 was not in the original CORDEX data request). These data were not saved and the OS of the machine changed in the meantime, so the bit-by-bit reproducibility of the r1i1p1f1 member is not clear. Latest news I'm aware of is that they might be able to roll back the OS to produce the missing fields, but we need to wait. Otherwise, a new member could be produced, initialized from the same historical member and using SSP3-7.0 forcing. Were you planning simulations nested into IPSL-CM6A-LR as well?

Unfortunately, the automatic tracking of available boundaries from ESGF in this site does not verify that the full period is available or the integrity of the corresponding files. Any finding in this respect from the community is really useful to adapt the availability information. E.g. I will drop IPSL-CM6A-LR ssp370 from the available boundaries until this is solved.

[doblerone]

Hi Jesus,
thanks for the confirmation, I was not aware that you are already know !
It was actually Ole who mentioned it :-)
DMI does the IPSL-CM6A-LR SSP370 run, I don't have any plans to do so.