I added Narcissus pseudonarcissus. There are too many hybrid and variety forms, and also some other species, misidentified onto this taxon.
2 Likes
I’ve added Panorpa communis (Common European Scorpionfly) which the computer vision seems to be identifying all Panorpa sp. as this in the UK (and I’m guessing other countries too). Of the UK species males can be told apart from a good macro photo showing ventral surface of the genital capsule. Some sources suggest females can be differentiated with a microscope but others suggest dissection is required. It appears the CV is alos suggesting this species in america where it isn’t present as there is the (closed) flag from last year with the following description:
“the AI keeps suggesting American observers Panorpa communis although the species doesn’t occur in America resulting in many wrong IDs. I don’t know if this can be fixed this way but I’ll try”
I wasn’t sure if I should create a new flag or not.
Panorpa can be ided by wing pattern too.
http://sea-entomologia.org/IDE@/revista_60.pdf
https://das-neue-naturforum.de/lexicon/index.php?entry/59-skorpionsfliegen-bestimmungshilfe
https://www.jungledragon.com/image/62384/panorpa_-_details_to_photograph_for_id.html
https://www.jungledragon.com/tag/46848/panorpa_id_help.html
2 Likes
I added Aralia spinosa, which doesn’t occur north of Delaware along the east coast (the trees there are A. elata instead). I’m not sure if Philadelphia gets spinosa, though, so I left it alone
I added the genus Jasminum, it is a hot mess.
I have scrolled through it for less than 15 min and found 15 different genus of plants identified as this (some in research grade), including Rosa, Citrus, Petunia, Oenothera, Primula, Philadelphus, Forsythia etc.
4 Likes
I’ve added Clogmia. I usually get it to come up as a suggestion for any Moth-flies I post up…
But Psychodidae are notoriously difficult to take beyond family.
2 Likes
Can I add something which is the inverse of most of the above?
Anthomyiidae desperately need including in a CV update, as a large % of diptera spotted in UK are this family. But also very difficult to go beyond family, so we are unable to reach enough genus / RG observations for them to be counted. This has a roll-on affect of them being placed in incorrect families.
1 Like
Research grade doesn’t play into which taxa are included in the model anymore. For the last few trainings it’s been this: https://www.inaturalist.org/pages/help#cv-taxa
3 Likes
Oh! Great. Yes, I think I was already told that but couldn’t quite remember, hence the
“genus/RG” …I figured this issue likely stems from family level IDs not being taken into account then, right?
Or if not, how does iNaturalist define “community ID”?
Do I just need to tick the “No, it’s as good as it can be” in 50 observations for it to be counted? (even though this makes it “casual”?)
Worldwide there are 3080 observations that are at a family level, many of which have more than one agreement… but I’ve never(or very rarely) had it suggest Anthomyiidae for the grey unidentifiable Anthomyiidae of which there are so many…
( sorry, I realise this is kind of abusing point 3 in the what not to do on this thread list! - its just one of the biggest issues I think for the CV with UK diptera, would be amazing to fix :) )
1 Like
I’m curious what taxa are allowed to show up in the suggestions and how they’re prioritized. I know family, genus, and species can, but perhaps family should show up more.
2 Likes
Added Sphinx chersis, specifically the pupal stage. Basically every single generic looking brown moth pupa (and sometimes even non-Lepidoptera pupae) gets computer ID’ed as this species, or sometimes other similar Sphingid species. I have no idea why it keeps choosing this species in particular (maybe the pupal pattern somehow looks similar to the adult wing pattern?). In reality 99% of these generic brown pupae cannot (and should not) be ID’ed to species level unless the observer kept it until the adult eclosed. Just took a quick scroll through and virtually every single pupa currently ID’ed as Sphinx chersis is not a Sphinx chersis.
2 Likes
I added Harpaphe haydeniana. This striking black and yellow millipede is common in Pacific coast states & province (southeast Alaska to California). There are at least two very similar pantropical species. Also, Harpaphe doesn’t live in the eastern U.S., but some similar species do. The family name Xystodesmidae is a better choice for the tropical and s.e. U.S. observations.
A smaller problem is that at least two other species of Harpaphe live in limited areas of California. Since the range of H. haydeniana completely overlaps the range of the other two, at least these misidentifications won’t be misleading. If that were the only problem, this wouldn’t be worth worrying about.
I also added Prunus americana. This common thicket-forming shrub is widespread in North America, but doesn’t occur in California or Nevada and barely reaches eastern Washington and Oregon. Most of the erroneously identified shrubs are the related Prunus cersifera, but a diversity of other Prunus and Malus and even more distantly related things are misidentified. And who wouldn’t agree with an identification for Prunus america in North America?? Many observations have three or four or five agreeing, erroneous identifications.
2 Likes
I remember adding Pacific Black Duck x Mallard hybrid to the wiki but I think somebody removed it.
The computer vision continues to recommend this taxon for observations of mallards and nothing has been done about it
2 Likes
https://www.inaturalist.org/taxa/54359-Amanita-magniverrucata is said to be endemic to California, I cleaned Old World records, but someone with more knowlede should check New World ones.
1 Like
Both the computer and humans are identifying American Holly (Ilex opaca) all over the world. Some are clearly Ilex aquifolium. Others are oaks. But a lot of the time I don’t know enough about Ilex opaca to be sure.
1 Like
Aloe vera – both humans and the computer are identifying as this species a diversity of Aloe and Agave, as well as the occasional bromeliad and I think Kinophia. Of course, many of the records are cultivated but not marked as such.
1 Like
Looking through some observations for Indonesia, quite a number of plant observations have had the observer assign ‘Dysithamnus mentalis’ as an id - which is an antbird - on August 10 2019.
Identifiers have valiantly reassigned verifiable ones to plants on the whole, and I am just going through the casual ones - because they seem to have been enough to persuade the computer vision that plants actually could be identified as Dysithamnus in this part of the world [edit - possibly generically, as the remaining observations have often been missing location].
3 Likes