Issue 18450

New datasets: mixed msgs about basis of record

Reporter: kylecopas
Assignee: ahahn
Type: Feedback
Summary: New datasets: mixed msgs about basis of record
Resolution: Done
Status: Closed
Created: 2016-04-29 14:54:34.107
Updated: 2017-10-06 11:00:03.931
Resolved: 2017-10-06 11:00:03.911
Description: New Norwegian publisher with three new marine datasets. Metadata describes the location of specimens, but basis of record is declared as 'human observation'.

Will flag for Dag.]]>

Comment: [~kylecopas] Not sure whether this still needs action, Kyle - had you contacted Dag, as you say in the issue? Can we close this?
Created: 2017-10-06 10:40:27.166
Updated: 2017-10-06 10:40:27.166

Author: kylecopas
Created: 2017-10-06 10:59:46.632
Updated: 2017-10-06 10:59:46.632
All clear. From Dag via email:

Regarding event core - these datasets are published by the Institute for Marine Research (IMR), and as Jan Legind will remember, Arnfinn Morvik of IMR participated at our sample-based dataset workshop in Trondheim last year. I am off for one week of holidays today, but will mention this possibility of converting to event-core datasets to Arnfinn when back in office. However, the primary purpose for IMR of using the IPT is actually NOT publishing to GBIF, but for publishing to their own international community marine data portals (as I believe Kyle Braak knows more details about than me).

Many thanks for noticing and for looking into the new datasets from Norway!!

Regarding the occurrences and their basis of record - I think that it would be useful if the GBIF portal presentation for data users would care less about the "occurrences" (dwc:Occurrence) with their dwc:occurrenceID, and care more about specimens with their dwc:materialSampleID and observations with their dwc:organismID - and the eventID connecting them. dwc:occurrenceID is more the technical link back to the evidence of these specimens or observations - the link back to the datasets providing information of the observations and specimens. While occurrenceIDs are important for the data publishing and data indexing purposes - but actually one of the most important obstacles for users accessing the data for scientific analysis.

The very same animal can be observed (basisOfRecord=HumanObservation) by the IMR and later deposited at the museum in Bergen where the basisOfRecord will be PreservedSpecimen. Both the observation record and the specimen data record could preferably share the same eventID and organismID (if declared), but should each have different occurrenceID. The observation and the specimen each provide (respectively different type of) evidence for the presence of the animal at the respective location and time.