The problem is that a lot of the researchers using this kind of data are not taxonomic experts themselves, so they come in and download big blocks of the database without understanding the issues. For example, people have published country lists for bees based on GBIF data. They might do some checking and based on that assume a certain error rate, say 5-10%, without recognizing that in certain groups it can be >90%.
3 Likes
Not downplaying the issue, but 500 or so is pretty manageable and could be easily ‘fixed’ (reidentified, downgraded, whatever is needed) in a couple of hours. Myself and others have been doing a lot of similar clean ups recently for misidentified Australian inverts. We’ve had plenty of cases with similar numbers, and they get fixed pretty easily. And once you do a massive downgrade/reID, most of the community cottons on to the change and becomes self-correcting.
14 Likes
also as a minor interesting point, these two statements together imply that either the collections are also misidentified, or that the Alabama expert is wrong
4 Likes
. . . . or that it’s just another one of those taxonomic disagreements : )
2 Likes
On the east coast (and namely New Jersey) Eastern Mosquitofish are native, but Westerns are far more common. 140 observations of eastern mosquitofish in New Jersey alone, likely less than 10 are actually easterns, they’re very uncommon, but everyone assumes they’re what’s around since they’re native. I’m working on collecting truly verified observation data now so I can finally put all these mis-IDs to bed with a proven range map.
6 Likes
For what it’s worth, the Midland County observations of Lemna minor I have on iNat were keyed out from specimens (at least one of them, and I think using FNA). Doesn’t mean I keyed them correctly, but I remember them keying relatively cleanly.
Also, you might check the BONAP distribution map as well. They get things wrong too, so take it with a grain of salt, but it’s possible your sources are outdated or underrepresent the actual distribution of the species. Not to take away from the main point that this group needs curation, but just something to keep in mind.
10 Likes
Can I ask if you have any sources for this information?
Great idea!
In addition, I think it would be helpful to have a “tool” (e.g., “critical ID comment box” beside the normal comment box) forcing us “normal users” to provide more information on why we consider the identification to a higher level as valid and a warning sign, e.g., exclamation mark, indicating that this identification is higher than the “typical” identification.
I agree with everything everyone has said, happens a lot with moths. …I have gone back and made changes to my identifications many times to reflect this issue. ex. Genus Eupithecia. there are hundreds of RG observations of Common Eupithecia that should really be confirmed further with genitalia examination. But I have noticed, some identifiers have been more cautious about offering identification, myself included, which helps. It’s a learning curve for sure.
2 Likes
I would also look at the top identifiers of the species for Texas and try to contact them to explain the situation. They could be unaware of the status of the species in your area.
Yep, I’m aware of how ignorant some researchers are and that this sort of thing is happening. But that’s not a problem for iNaturalist…it’s a problem for the greater research community. Genbank, the DNA database, is chock full of errors–unfortunately, there’s no way the community can address them or even comment to their existence. This sort of problem (ignorant researchers using data repositories that obviously contain errors) is true for any repository of data contributed to by many individuals.
The problem is that PhD programs are pumping out ignorant researchers.
1 Like
Maybe just for Lemna, because that hasn’t been my experience.
CV often gives its top suggestion as just genus for lots of organisms (both on my own observations and when identifying others’ observations).
3 Likes
Unfortunately a big problem on here with no end in sight. I would support more restrictions on CV suggestions for taxa that are unlikely to occur in that given area. Sure, a plant endemic to a small area of New Zealand could appear in the Bronx, but it’s highly unlikely. With a species like Lemna, or anything that might require microscopy or DNA barcoding to identify to species there should be some sort of way an expert can limit the CV suggestions to genus (or whatever taxon level) is possible. If someone has done the necessary work they could always search for the taxon they need.
1 Like
I recently learned about the taxonomic mess that is our treatment of Taraxacum in the US. Probably all of them are misidentified, ha! 
1 Like
I find that after the most recent update the CV seems more likely to suggest non-local species. My vague memory is that it was doing better with geography before that.
Mostly iNat offers me A Species. I haven’t found a good trick to force it to offer the genus.
I click the offered species, delete the species word - and that forces the genus to display.
1 Like
well and also all the problems with taxonomy, such that for many taxa ‘species’ can’t even be identified unless you are one of a few experts due to taxonomic splitting and constant name changes. The field of taxonomy has completely diverged from everyone else who wants to identify organisms, including other branches of biology. For Lemna, being so tiny and needing underside pictures, there are probably legitimate species that are getting missed for these reasons and the computer vision is being overactive… but the problem isn’t only on the end of the ‘non experts’ here.
@ydobon Sure (apologies to the group for the digression): Here are a few range maps for Solidago canadensis based on the taxonomy that inat, ITIS, and POWO follow:
http://www.efloras.org/object_page.aspx?object_id=54142&flora_id=1
https://uwaterloo.ca/astereae-lab/research/goldenrods/classification-and-illustrations/solidago-canadensis
No entry for Solidago canadensis in flora of WA (currently 218 inat observations):
https://burkeherbarium.org/imagecollection/results.php?Terms=solidago&x=0&y=0&Type=Names
No entry for Solidago canadensis in flora of OR (currently 216 inat observations):
https://oregonflora.org/taxa/search.php?search=solidago
Discussion of non-native status in BC and scarcity of confirmed records here under Introduction (currently 575 inat observations):
https://linnet.geog.ubc.ca/Atlas/Atlas.aspx?
sciname=Solidago%20canadensis&redblue=Both&lifeform=7
3 Likes
Thanks for accumulating that information.
It probably depends on the photo and/or the taxon/location for whether it offers genus first or species.
1 Like