Connect GO2DAT as an ODIS node

[jmckenna]

• GO2DAT (Global Ocean Oxygen Database and Atlas) is being developed within the UN Decade GOOD (Global Ocean Oxygen Programme)
• next steps:
  1. Agree on a common minimum metadata requirements form – this, GO2DAT has already achieved.
  2. Translate the common metadata file into JSON-LD/schema.org format which will be used to access and index all oxygen resources in OIH (or anywhere else) – this could be facilitated by the ODIS Team
  3. Each data provider/aggregator adds a JSON-LD snippet to the individual resource webpages - to be done by data managers with ODIS support
  4. A GO2DAT community is formed "inside" OIH as a "coalition of the willing" from among those who agree to connect existing databases. This could include mutual archival of resources - so essentially creating regional DACs with healthy redundancy. A basic agreement needs to be outlined – You (the ODIS Team) have some examples/templates from other communities that you could share with us.
  5. There should be no need to re-submit anything as done in SOCAT or GLODAP. All resources are accessible centrally from where they are.
  6. QC/QF - To be figured out by the GO2DAT community and part of the common metadata requirements sheet. We are finalizing this step.
  7. Once all oxygen resources can be pulled together through OIH, the visualization (the actual atlas) can be implemented. OIH plans to develop such features too, but otherwise, any GO2DAT partner could take that role on.

related to https://github.com/iodepo/odis-arch/issues/384

[veronique56]

metadata_instructions_oxygen_27102023.xlsx Metadata_submission__oxygen_27102023.xlsx

Hello Jeff, I just uploaded our 2 metadata files on oxygen Let me know if ok or not. We still need to refine what we want for final QC/QF. cheers veronique

[veronique56]

Hello Jeff I am going to update this week these 2 metadata Excel files since Nico spotted a few omissions from my side:

Suggestions for general additions (for machine readability): Fill value (e.g. -999 or NA or NaN or…) Abbreviations and units for longitude, latitude, date, time Some sort of ID for the sample itself (?)

Suggestions for oxygen specific additions would be: Temperature used for mass calculation (draw temperature vs in-situ temperature – the former is recommended) Primary standard normality assessed (e.g. using CSK)

And for uncertainties, I will need to have one separate entry for precision (based on replicate measurements, Vdub) and one for accuracy (based on a fit with golden standard? (e.g. GO-SHIP cruises). so stay tuned cheers veronique

[pbuttigieg]

At the IODE Management Meeting, it was noted that the mission and data associated with GO2DAT overlaps with those of the World Ocean Database.

ODIS is currently in the middle of this, which is appropriate as we're focused on the metadata, but we have the opportunity to make sure that these systems are interoperable or ideally (eventually) mirrored to optimise efforts and harmonise the innovation in GO2DAT with long-term infrastructure such as WOD.

We should note this aspect in our future coordination to promote a collaborative space and technical interoperability.

This is an scenario that is common with the Ocean Science division and is the expected overlap of the innovation and operations arenas. It is expected that many innovation actions will create their own local systems for nimbleness, but these risk either a) evaporating after the projects / funding is over and b) not being interoperable with infrastructural / operational elements such as WOD. The model that works is innovation with a path to integration.

CC @BoyerWOD

[veronique56]

Metadata_submission__oxygen_05022024.xlsx metadata_instructions_oxygen_05022024.xlsx

Hello Jeff and Pier Luigi, I have uploaded the metadata for oxygen as revised following Nico's advices. We will discuss with you the "235 DOXY Quality flag scheme" which still requires approval of our full GO2DAT steering Committee. Our goal is to have GO2DAT fully interoperable with WOD of course but also with EMODnet, CMEMS, Chinese DB,etc... Kind regards Veronique

[veronique56]

GO2DAT_Dataflow-figure 4-revised.pptx

Hello Jeff, Doug and Pier Luigi, I just uploaded a power point with 3 slides. Slide 1 is the one Marilaure was talking about during or last call a vertical data flow scheme published in Gregoire et al in 2021. On slide 3, I have illustrated the metadata flow/interoperability scheme, and on slide 2 the data flow, trying to learn the lessons from Pier Luigi's talk at the recent IODE DMT webinar. Due to our ignorance, unsure if this makes sense or not. In particular what ERRDAP can do or cannot do, in conjunction with ODIS. Feel free to edit/change/modify as you see would fit. Kind regards Veronique

[pbuttigieg]

@pbuttigieg and @NicoGEOMAR to draft JSON-LD templates for oxygen data exchange across ODIS

[veronique56]

GO2DAT_Dataflow-figure 4-revised.pptx I just added a slide 4 on my previous power point. Trying to summarize what Pier Luigi told us during our last visio call from March 5th. Feel free to edit/change/modify as you see would fit. Kind regards Veronique

[pbuttigieg]

@NicoGEOMAR @jmckenna and I worked on this here https://github.com/iodepo/odis-in/pull/7 and it's almost done.

The excel sheets are very overloaded: they attempt to capture multiple processes, datasets, people, organisations. In good data management, these would be separated out and linked together.

The result of that is that the JSON-LD is quite long, as it's opening up the complexity in the Excel sheet. In the future, the deconvolution of this will lead to better (meta)data exchange.

[pbuttigieg]

@NicoGEOMAR and I also brainstormed and tried to clarify the data flow and actors at play.

This is where we are, and the role of GO2DAT as an entity that patrols ODIS for O2 data, harvests it, uplifts/augments it, and then recirculates the uplifted/augmented data via ODIS to all other systems interested in O2 data (ensuring compatible data standards for frictionless import) is more clear to me.

@kevin-obrien of interest to you regarding the potential ERDDAP links

[veronique56]

Thank you so much to you for making the GO2DAT first steps progress under the GOOS overall data structure. To be honest, I am still trying to understand your diagram above. Will try to explain it in "non data scientist" words and will send it to you probably with questions. cheers veronique