I’ve launched on a personal mission to curtail the over-identification of Plains Fleabane (Erigeron modestus). According to the Flora of North America Association, it is usually found in Ariz., Kans., N.Mex., Okla., Tex. in the USA, and Coahuila, Chihuahua, Nuevo Leon in Mexico. Until recently, iNaturalist had identifications all over Canada and the East Coast of the USA. I’ve been re-identifying most of them as Asters and Allies (Tribe Astereae) since there are a lot and I haven’t put the time toward making a more precise identification.
I could use your help!
You can help in two ways:
By looking through pictures of Plains Fleabane and checking that they are correct. I suspect that there are false identifications in iNaturalist, leading to more false identifications in a vicious cycle. If I’m right, the taxon could be flagged for curation, but I’d like someone else to check if this is an issue first.
By confirming or re-identifying observations near the edges of Plains Fleabane’s range. It’s easy to re-identify observations far outside Plains Fleabane’s range, but I’m not exactly sure of the limits of its range. If someone more knowledgeable can examine observations where you think the limits are and clarify these limits, then someone less knowledgeable can follow behind and re-identify observations further afield as I have been doing. You can use the map of observations of Plains Fleabane.
It’s less important, but you can also find some of the observations that others had identified as Plains Fleabane mixed in with this search query.
Even after correcting, it will be a while (forever?) until the CV can distinguish Erigeron species. Now, users could leave their observations at genus level, but people love to click the first suggested species. Since I also struggle with Erigeron, all I can say is good luck Daniel.
Personal example: it took me about 6 months to correct Penstemon strictus and this is what the “Similar Species” tab looks like (plus another partial row out of view):
Setting them to tribe, while perhaps the correct thing to do, can make them difficult to track down later if you are hoping to add a more specific identification. I don’t know what the numbers are like on Erigeron modestus, but I often keep a personal spreadsheet to track ones I want to go back to later if I am setting them to a higher level taxon that is hard to filter on in iNat.
Don’t we first need to know how to identify a plant as not being Plains Fleabane? I’m sure most are misidentified, but maybe they’re being established outside their original range? If we assume they don’t occur outside their range, and unidentify all of them based on that alone (without even knowing how to identify them) that would be unfortunate also.
I don’t know anything about Plains Fleabane (we have fleabanes in my area and I only ever identify them to the genus level or higher because I have no idea how to tell them apart, lol), but I do work with database maintenance and clean-up in my day job. My first thought is that it might be a good idea to utilize iNat’s project tools for this. You could start one to collect Plains Fleabane ID’s outside its known range that haven’t been reviewed yet and another for those that have been reviewed and either found to be accurate or re-categorized. The second could require the person moving the observation to note or select characteristics they used to ID (like how Found Feathers requires information regarding where on the body the feather was likely located).
This approach or something like it would take advantage of data sorting features already built into iNat, allow for both crowd sourcing/public participation and some level of quality control, and avoid superfluous re-analysis of already checked observations. The project description could link to an identification key with notable features to look for, meaning even folks with little previous experience might be able to learn through participating. Also relevant since range is such a factor is that location data might be better, as project participants can opt to reveal hidden or obscured coordinates for included observations. Plus, having a permanent archive of sorts would record the action as well as the outcomes, meaning that future interested parties would be able to make informed decisions about including observations in data sets without having to look through comments on individual posts.
Data clean-up is an onerous task at the best of times, but having a solid process before tackling it is half the battle. Hopefully my brain dump of stray thoughts is useful!
We wouldn’t want to flag taxa because they’re being mis-identified. Only if there is a taxonomic change that needs to be made.
What you’re describing here is pretty much every species on iNat–not unique to this group. Some are worse than others, and it will always be that way. Some taxa will always have high rates of misidentification simply because they’re difficult to identify and have many similar species. The best we can do is focus on taxa that we’re passionate about, know how to identify, and see value in devoting our limited time.
The most frequent comment I see on disputed identifications is that the observation is outside the range of that species. The person providing such a comment seldom says anything about identifying features.