Connecting to the data citation service

[durack1]

From @MartinaSt

From: Martina Stockhause
Date: Friday, May 19, 2023 at 2:52 AM
To: Durack, Paul J.
Subject: CMIP6Plus - input4MIPs
Hi Paul,
I have thought about CMIP6Plus and input4MIPs a bit more:

1. DOI granularity/-ies: For CMIP6-input4MIPs DOIs on two granularities were registered. The easiest is to do it for CMIP6Plus in the same way and have two new DOIs for your AMIP-1-1-9 data. However, if the CMIP6Plus period is to be short, it might be better to have only a DOI on the finer granularity (with the possibility to relate it to the coarser one for CMIP6). I guess an example makes these options clearer:

• 2 new DOIs on data collections 'input4MIPs.CMIP6Plus.CMIP.PCMDI' and 'input4MIPs.CMIP6Plus.CMIP.PCMDI.PCMDI-AMIP-1-1-9'
• 1 new DOI on data collection 'input4MIPs.CMIP6Plus.CMIP.PCMDI.PCMDI-AMIP-1-1-9' with the possibility to relate it to 'input4MIPs.CMIP6.CMIP.PCMDI'

2. Data Access Links: I will use links into the CoG and change them together with all the others for CMIP6 when ESGF moves to MetaGrid.

3. CV/citation service workflow: Currently, the ESGF publication of the input4MIPs data is first, then the insert into the citation DB and the DOI publication. For some cases, this leads to DOI registrations before the citation manager could adjust the citation metadata. I don't want to change this for a rather short period of CMIP6Plus but rather live with the inconvenience: writing an email to me in case this has happened.

4. Errata/Versions: Yes, that is something we should improve. There will be a delay for displaying this information on the DOI landing pages.

Which option under 1. would you prefer?
Best wishes,
Martina

[durack1]

Thanks @MartinaSt, just moving this to github so it's out of inboxes (you might also note that I have migrated the #2 issue from the input4MIPs-cmor-tables, as we will have to ascertain a format that serves at projects as source_id and other content is migrated here - I intend to reproduce all ESGF existing datasets in this way, so that the CVs and the ESGF project are synchronized.

Regarding your question above, the finer granularity (e.g., input4MIPs.CMIP6Plus.CMIP.PCMDI.PCMDI-AMIP-1-1-9) makes more sense, as I am fully expecting that we will be generating multiple version releases of datasets as problems are identified and fixed, and having an explicit (per version) DOI would be far preferable that not explicitly defining these down to the release/version information.

[durack1]

DKRZ will not support DOI minting for this project, so will convert to a discussion to get this out of the issue list