Problems with sightings being transferred to the Atlas of Living Australia

iNaturalist Australia and the partnership with the Atlas of Living Australia (ALA) has made the transfer of sightings between the two platforms easier than ever, but there are still many problems with sightings being transferred to the ALA (or not being transferred!). Some of these issues were touched on here and I plan on emailing the ALA Support Team directly to try to resolve them.

The point of this post is to A) gather everyone’s issues together in one place so we don’t end up sending them a hundred emails and B) gain some support from the community so that they know how many people would be helped by any solutions. Of course discussion of these issues would also be great to have here too!


As of right now I have 4 main issues but if anyone has any others I’ll add them in:

1. Incorrect mapping of iNaturalist taxa to ALA taxa
This seems to be the biggest issue - sightings on iNaturalist are sometimes not listed under the correct taxon on the ALA. iNaturalist taxa are rather fluid concepts that can be changed without too much effort, whereas ALA taxa are very much more rigid and are almost never changed. Any time iNaturalist changes to a more updated taxonomy or simply uses a different name to the ALA for whatever reason that may be, iNaturalist sightings generally do not map to the correct taxon. Usually they are matched with a higher-level taxon (undesirable but not strictly incorrect) but sometimes they are matched with related taxa (undesirable and incorrect). A good example of this is Trichonephila edulis, which is listed on the ALA under the old name Nephila edulis. iNaturalist sightings of T. edulis get mapped directly to Nephila rather than to N. edulis (e.g. here).

Other times, the taxa have identical names and yet are mapped to wildly different taxa on the ALA. I have no idea what causes this but it is a more problematic issue than the one above. For example, iNaturalist sightings of the genus Abantiades are inexplicably mapped to Diphyrrhynchus (Neoabantis) despite Abantiades existing in the ALA (e.g. here). This is extremely frustrating to say the least, especially when the automated ALA message claims that there were “No issues” with the taxonomic match.

A solution that was discussed with @thebeachcomber and @nicklambert was to have a place to log these problems where the relevant people could then go in and fix them. This enables us iNaturalist users to flag problems when we encounter them, rather than having someone from the ALA go through every taxon to check for issues, and is much easier than flagging every problematic ALA record (tedious, and as far as I can tell flags are rarely if ever checked). This list of problematic taxa would then be used by ALA staff to fix problems, and could easily be updated if a problem is fixed or if a new issue becomes apparent.

2. Updated iNaturalist IDs not updating on ALA
Incorrect initial IDs on iNaturalist sightings are common as we all know, and sightings are transferred to the ALA frequently enough that sometimes the community has not gotten around to reviewing them properly before they become visible on the ALA. This is unavoidable and is actually a testament to the efficiency of data transfer! If the ID is refined or changed to a completely different taxon, these changes are updated on the ALA as well. However, if the community ID is changed to a higher taxon, this is not updated on the ALA.

A good example is this sighting, which was initially incorrectly IDed as Tenebrio molitor and then approximately 3 months later was bumped back up to Pterygota. The sighting was initially transferred to the ALA as Tenebrio (back to the first issue again…), but when the ID was changed on iNaturalist it was not changed on the ALA, and is still under Tenebrio. This all happened about 9 months ago so it is not simply a matter of not having updated yet. Interestingly, the ALA changes the ‘supplied scientific name’ and ‘supplied rank’ of the sighting to the new iNaturalist ID, but does not then move the sighting to the relevant taxon.

3. Only sightings from Australia itself are transferred
Given the taxa it includes, the ALA appears to be designed for sightings of taxa in all Australian-owned territories, and this is certainly reflected in the records that it has from other data sources. Currently, however, only sightings from Australia in a very strict sense are transferred. Sightings from external territories such as Norfolk Island and Christmas Island are not transferred, and more relevant to most people, sightings from coastal waters are not transferred either. All of the places that should have their sightings transferred are here (the same as those used in the Bowerbird project), which comes out to approximately 12,000 sightings that are currently not on the ALA but should be.

4. Images shown in the ALA ‘Gallery’ tab seem to be chosen at random
When an iNaturalist sighting has multiple images, all of these images are transferred to the ALA on the page for the specific record, but only one is shown in the ‘Gallery’ tab for the taxon. The image chosen to be shown in the gallery seems to be chosen at random, and this is sometimes not the most desirable image to be shown.

A good example is this sighting. I think velvet worms are extremely cool, and these were the first photographs of this species, so I uploaded all of the photos I took of it including some blurry ones. It’s also perfectly reasonable to assume that sightings will include closeup images of certain regions needed for ID, habitat images, or other images along this sort of vein. On the sighting’s record page on the ALA, all 16 images are shown. However, on the taxon page, the single image shown seems to have been chosen randomly from the selection of photos - and of course it’s one of the blurry ones. Although not an enormous issue, this is undesirable and seems like it would be easy enough to fix. It would be much better if the image shown in the gallery was the first image on the iNaturalist sighting, as this is usually the best image for that purpose.


In addition to these four issues, I can’t seem to make head nor tail of which sightings are transferred to the ALA and which aren’t. For example this genus-level ‘Needs ID’ sighting was transferred whereas this species-level ‘Research Grade’ sighting was not transferred. Ideally, I think all sightings from Australia and its external territories should be transferred to the ALA.

One further thought I have had is that there is some data that the ALA has the capacity to deal with that is not yet being transferred to the ALA. The ALA has fields for sex, life stage, and geospatial uncertainty (listed as “Coordinate precision”). Currently, all iNaturalist sightings have no entry for sex and life stage, and the coordinate precision is recorded simply as “Unknown”. It seems a waste to not transfer this data along with the sighting and it would be fantastic if this could be incorporated in the future.


If anyone has any additional problems or suggestions please bring them up here, and equally let me know if you agree or disagree with any of the suggestions made here. The ALA is a fantastic resource in its own right, and it would be good to make our contributions to the resource as accurate as they can be. After everything’s been discussed I’ll contact the ALA and hopefully we can get these problems fixed!

6 Likes

Great post, covers a lot of issues I’ve been thinking about too. Some of my thoughts building on yours:

This is a big bugbear of mine. The ALA’s latest taxonomy update was February earlier this year; the latest before that was 2017. Updating taxonomy every 2-3 years isn’t good enough in my opinion when our knowledge of taxonomy is constantly changing as Matthew said. In my view this is one of the biggest problems with the ALA.

This irritates me to no end as well given the photo they pick seems completely random. I’m in the same boat, especially for a number of my rare invert observations. I’ve uploaded a series of ~10 photos for many showing small features; the first photo is the best one, but a random one got picked. This is especially bad for these rare species where my photos are the only ones available; why would the best one not be picked as representative of the taxon? Would a simple solution be to just always select the first image given that it’s the most likely to be in focus/be the best photo in the series.

There are some huge unexplainable discrepancies out there regarding this issue. A good example is the Bigbelly seahorse. On iNat currently, there are 806 observations of this species. Given the fact that the ALA accepts records from both iNat and elsewhere, e.g. from museums, and the fact that the ALA collected records before iNat existed, you would think that the ALA must have more than 806 records of this species, i.e. all the iNat ones + the museum ones/others. But this isn’t true! The ALA somehow only has 752 observations, of which only ~530 of these are from iNaturalist. Almost every single observation on iNat (804/806) is research grade, so why are there over 270 records of this species that have not been piped to the ALA?

5 Likes

Thanks for aiming to straighten out these data issues and discrepancies. I can’t comment with specificity on these topics but can offer suggestions to help delineate what requires action from ALA versus iNat staff.

Each week iNat automatically generates an archive for ALA to harvest, which can be found here (warning: will start downloading a big zip file). Then I believe they use the iNat API to update certain data on their end between these weekly archives. Checking this archive may help you get to the bottom of some discrepancies.

6 Likes

Thanks very much @carrieseltzer, I’ll have a look and see what information is getting sent to the ALA.

2 Likes

Agree with all the things you mentioned. How annoying is the random blurry photo thing? Does my head in!! :)
I have seen instances of all of these many times. Hope it gets better.

1 Like

i didn’t check in detail, but this is probably because of licensing. (see https://forum.inaturalist.org/t/how-often-does-data-get-updated-to-atlas-of-living-australia/12993/8. there’s not a further post to indicate that ALA made a fix to address licensing, but it seems like they must have.)

1 Like

Thanks for exploring this Matthew. I have nothing more to add but your points 2 and 4 have been my areas of frustration.

1 Like

perhaps this is a naive question/ignorance on my part, but surely licensing only applies to the photographs for each observation, and not also the datapoint (i.e. spatiotemporal coordinates) itself? So the record should still migrate across as a point on a map, but the photo does not.

there’s a license on the observation and a separate license for the photos. i would assume licensing that is too restrictive at the observation level would prevent the record from getting sent over. that’s the way i read ALA’s statement of how they get data from iNaturalist (https://support.ala.org.au/support/solutions/articles/6000223721-how-is-data-harvested-from-inaturalist-and-fed-into-the-ala-).

I’m wondering if the missing Bigbelly seahorse observations are related to the other point where original incorrect obs are kept on ALA (i.e. original ID might be at a higher level). I haven’t checked if both original incorrect and updated obs are transferred.

here are some examples of observations that have links to ALA records:

note that the ALA links for the all rights reserved records are now broken, and you can’t find similar observations in ALA if you search for them either. but the CC BY-NC observations link to valid ALA observations. if you look at the thread that i referenced earlier, it looks like ALA used to pull in records from iNaturalist regardless of license, and they said they were going to change it so that it would take only records with less restrictive licenses. so that would explain the now-broken links.

1 Like

A post was split to a new topic: Log in with Google on iNaturalistAU?

Here’s an odd one. I have a RG observation from 2 months ago (https://www.inaturalist.org/observations/53240865), but it hasn’t gone into the ALA (https://bie.ala.org.au/species/urn:lsid:biodiversity.org.au:afd.taxon:de5be57c-c42e-4287-bc45-e18bbd579f02). System definitely isn’t working properly