(Not sure if this is the right section, or even if it’s something inat has control over)
I have noticed that misidentified inat observations which were formerly research grade but were later corrected are still shown in GBIF. Is this something inat has any control over, or is it something GBIF would have to fix themselves?
How long has it been since they were corrected? If an ID is changed on the iNat end, it isn’t reflected instantly on the GBIF end
Last ingestion history was 2026 Aug 13 for data up to 2026 Aug 10?
I’ve actually seen this happening on several websites, including many of the top resource platforms available for IDing plant species.
I’m facilitating a project on Commelinaceae in southern Africa and have spent considerable time navigating to online resources to double-check accepted names, synonyms, images and descriptions - only to find that there is an image being used from an iNat observation that i) has been identified to RG based on an incorrect initial suggestion, or ii) has since been re-identified.
I’m not sure how it works with GBIF (I know that some websites automatically draw from iNat) but in other cases I’ve actually sent emails to the website administrators pointing out the error - and in all cases I received a reply and a correction.
I would be interested to know what the turn-around time is on ‘auto-corrections’ on GBIF, for instance.
for many (and I assume almost all) of these other sites and platforms, they would have drawn in the pics/IDs once, and there is no automated feed to update details; it has to be done manually. So not too surprising that things get out of date
the typical schedule is that GBIF reingests iNat data once every week, but sometimes there are delays
Thanks for the response. Once a week is actually not bad. That gives me incentive to get keep identifying - and looking for misidentifications - wherever possible!
What exactly do you mean by “still shown in GBIF”?
If you mean that the GBIF icon is still shown in the observation’s DQA section, that is a known glitch in the iNaturalist interface. There are still hundreds of thousands of observations like that which were added years ago (see here). If you tick and untick a DQA item, the data will refresh and the GBIF icon should disappear.
However, if you don’t mean that, please provide some example GBIF links that demonstrate the problem. These should be for iNaturalist observations added more than a few weeks ago, since it takes at least that long for the data to sync properly.
I mean if you browse occurrences in GBIF, these former research grade obs are still listed. Here is one for example, the one for which I first noticed this phenomenon:
https://www.gbif.org/occurrence/6195613676
I updated the ID on the corresponding inat observation less than a week ago though, will wait and see if it gets corrected within the week
From the iNat Help site:
Which iNaturalist observations are exported for GBIF, and how often does this export happen?
Research-grade observations with CC0, CC BY, or CC-BY NC licenses. Note that this refers to the observation data license, not the license (or lack thereof) for photos or sounds.
iNaturalist generates data for GBIF once a week, and they attempt to ingest it once a week. You can see GBIF’s ingestion history of iNaturalist data, and the dataset itself.
and
If an observation’s ID or data quality grade changes, will that change be updated on GBIF?
Changes like this will be reflected in presence or absence of the observation in the DwC-A file iNaturalist generates for them. These changes should show up on GBIF when they import that archive.
Not to the original question but to @ bazwal highlighting of “hundreds of thousands of observations like that which were added years ago”
My perspective on that using catchphrase from a media series of recent past would be these words “…It is known” . Is it? Seems i missed that memo. Sounds like something some coding could resolve in a few minutes, or even the volunteer curators could manualy resolve together in a couple of hours. Seems not the problem in hand, but i ask why a bunch of observations sit with that error since then if “It is known” ?
There are some exports like Calflora which will not delete sightings that are taken from iNaturalist. This has been a problem in cases where planted/cultivated plants are getting taken from iNat, fixed on iNat’s end so they turn casual, but not updated on Calflora.
ALA used to have issues too but I don’t think they still do.
the ALA updates/re-ingests everything from iNat every single week
The issue comes up quite regularly on the forum, with users either seeing the GBIF icons appear when they shouldn’t, or not appearing when they should. For example:
These all mention the same work-around I suggested.
Given how often this is reported, a simple fix may not be possible - otherwise you’d think it would have been dealt with long ago? But then again, I haven’t found any official response explaining exactly what causes the problem and why it cannot be solved (although I didn’t read everything and didn’t search that hard).
Interesting, thanks for context on the issue. I focus on flagged issues for curation and mostly on arthropods, this is all new to me. I don’t remember this issue as raised for curator issues but interesting to hear different context on forum.
As for the “otherwise you’d think it would have been dealt with long ago”. Well no, i would not. But i comment no further on that to avoid possible moderation about staff actions that could better support the recognised problems of infrastructure highlighted by the community of forum members, moderators or curatiors etc
It was intended to be a diplomatically rhetorical statement 