AGCD versions

[DamienIrving]

At the meeting on 17 March 2022 we discussed trying to figure out what version of the AGCD dataset CSIRO and BoM researchers are using to ensure we're using the latest (and same) version.

[chunhsusu]

I think this https://geonetwork.nci.org.au/geonetwork/srv/eng/catalog.search#/metadata/f7421_3667_8543_6351 is the formal publication (with doi) of the AGCD v1 and v2 data to the end of 2020, but it does not get updated frequently. It is sufficient for our historical assessment.

[hot007]

There has been some discussion around this in CSIRO in recent days after @ngben discovered a discrepancy in 2013 data between the commercially licenced collection in xv83 (copied from CSIRO's storage) and the research-only licenced official DOI version in zv2. Unfortunately because changes get occasionally applied to previous dates in AGCD and these changes aren't documented, we can't tell which is the 'correct' state easily.
CSIRO is requesting an updated back-catalogue from BoM in the commercially licenced package to ensure our data contains all latest corrections. Meanwhile as the file creation dates in zv2 are newer than those on NCI we can probably assume it is zv2 that contains a correction, though it is not extended up to current dates. I'm not sure what to think, to be honest, I hope we can get an updated copy to CSIRO soon to confirm which is right.
We ideally should also establish if it is acceptable for us to use a research-licenced product (zv2) in ACS - probably fine for BoM staff but slightly unclear for CSIRO, hence erring on the side of having a copy in xv83.
What do others think on the licence front?