Request for OBIS Support: Draft Term Change for occurrenceStatus

Data publishing
1

We’re drafting a Darwin Core term change request to update the default vocabulary for occurrenceStatus — and we’d like to ask for OBIS’ support and feedback.

Motivation

The current default values "present" and "absent" is confusing. A taxon not found does not necessarily mean it is absent.

We’re a group of volunteers who have been working on the definitions for concepts related to absences and are now drafting a term change request for occurrenceStatus.

Concerns related specifically to this change request that were raised during the DCG meetings

Implementation complexity

We suggest keeping "present" and "absent" as non-preferred synonyms of "detected" and "notDetected" under Stability Justification.

What about modellers who do mean present/absent with the term occurrenceStatus?

Some modelling outputs use "present"/"absent" as estimates, but this reinforces the need to distinguish between detection/non-detection and inferred presence/absence. Modellers could still use extended vocabularies where justified — the proposed change doesn’t prevent that. (please see usage comments)

Is there demand for this change?

There’s strong literature support (listed in the term change request) as well as example from GitHub issues and personal experience amongst the members.

Follow up meeting

The OBIS DCG suggested holding a follow-up meeting to continue the discussion.

2 Likes
2

@ymgan I’m not able to commit to the meeting right now, but here are some thoughts in response to your comment in the OBIS Slack.

can someone please provide example of occupancy model which uses occurrenceStatus = present/absent in the output? thanks a lot!

I looked at a variety of papers, including those that are referenced in your google doc. I couldn’t find a case where outputs were described as presence/absence, but I did find many cases where model results were described casually as presence/absence, albeit generally with a caveat or about the probability of false detection/nonDetection.

To reiterate my comments in the DCG the other day, this feels largely semantic to me. I don’t feel a strong need for this change, perhaps because this characteristic is inherently noisy and I’ve grown comfortable fooling myself into gauging presence/absence through interpretation of other contextual metadata.

I’m willing to support the change based on my respect for the opinions of those who have opened the issue. Please keep me informed when the term change request is submitted, and I’ll voice my support.

2 Likes
3

Hi @sformel

I have posted the term change request: Change term - occurrenceStatus · Issue #546 · tdwg/dwc · GitHub
Thanks a lot!

1 Like