Find unidentified trees?

[epiphyte78]

let me start off by saying that not being able to link to wrongly identified observations is extremely counterproductive. it makes it seem like it’s somehow shameful to guess wrongly. nothing could be further from the truth. inat should have a policy of zero-shaming for wrong id’s. this policy would allow us to link to wrongly id’d observations in order to learn from them. making mistakes isn’t shameful. failing to learn from mistakes is a different story.

i recently found an observation of a tree that the poster id’d as ficus sycomorus. i disagreed by submitting a different id, ficus racemosa. the poster replied…

This tree is within Fruit and Spice Park, where it is labeled as a sycamore fig.

i replied…

i’m green with envy that you’ve been to the fruit and spice park!! it’s really high up on my list of places that i want to visit since i collect rare fruit trees. didn’t know that they had this tree. if you search online for racemosa leaves and compare them to sycomorus leaves they are pretty different. if the fruit and spice park doesn’t actually have a sycomorus tree then i’d recommend ficus sycomorus ‘shikma balami’ since it ripens without pollination (parthenocarpic). while they are it they should also really try to get ficus timlada. i just published a report about it on the figfanatic forum.

the fruit and spice park has quite a few visitors each day so i figured i’d use the map function to search it for ficus trees to hopefully find other observations of the same tree. there were only 11 results and no other observations of the same tree. what are the chances that someone took a pic of the same ficus tree, and submitted to inat, but they really wrongly identified it?

out of curiosity, after being a member of inat for a decade, i submitted my 4th observation… a tree. except, i tried to do it from the perspective of someone more normal than myself.

my name is shirley and i went to a botanic garden and took a pic of a tree, but i forgot to look at the label. i’ll upload it to inat anyways.

there’s a textbox that says “species”, when i click on it, inat says “we’re not confident enough to make a recommendation, but here are our top suggestions:”… california bay, coast live oak, tree privet and so on. hmmm… what’s the difference between a recommendation and a suggestion? maybe i’ll just pick coast live oak?

all i know for sure is that it’s a tree. let me try typing in “tree”…

either shirley chooses a highly specific suggestion, all of which were very wrong, and will need to be corrected (but might not be for 6 years), or she chooses a way too general suggestion, that entirely wastes and squanders what she does know, that it’s a tree.

i’m probably not going to browse all the results for vascular plants in the fruit and spice park. will i browse all the trees at the fruit and spice park? maybe. but is there even a way to do this? my 1st choice would be to browse all the unidentified trees at the fruit and spice park. but it seems like inat really doesn’t want anyone to admit when they don’t know a tree’s identity. this would explain why there are so many wrongly id’d trees on inat.

someone is going to chime in that i haven’t participated on inat enough to offer any useful feedback. ignore the noobs, and the results should be obvious.

without a stitch in time…
without an ounce of prevention…

again, there’s absolutely no shame in making mistakes. but setting people up for failure is a different story.

and perhaps there’s a bit of bias against botany. there’s no category for trees but there’s a category for tree squirrels. or maybe it’s more of an issue with taxonomy, which was invented way before inat was invented.

[bouteloua]

Do you have an actionable suggested change? Trees aren’t a taxonomic group, so they can’t be incorporated as a taxon suggestion.

[natev]

@epiphyte78

let me start off by saying that not being able to link to wrongly identified observations is extremely counterproductive. it makes it seem like it’s somehow shameful to guess wrongly.

The forum was never intended to be used for the discussion of incorrectly identified observations. From " About the General category - please read before posting!"

The iNat Forum is also not a place to seek help with identification, promote projects, or to post unconstructive complaints/grievances.

I do appreciate your walkthrough of a typical user experience. It is true that identifying trees on iNaturalist can be difficult, especially for someone who is not well versed in plant ID or taxonomy, and this often results in users posting observations with overly broad IDs or incorrect species level IDs. However, when complaining about how difficult the system is, you have to compare it to the alternative. iNaturalist is complicated because identifying plants is inherently complicated, not because iNaturalist is bad at identifying plants. iNaturalist is better at identifying plants than any free plant ID app on the market, and it comes with the ability to ID many other types of organisms and it comes with a team of volunteers who correct the computer vision, something no other platform offers. Additionally, without an App, the new user would have to find a local plant identification book or a similar resource in order to have any hope of identification, which takes substantially more time and effort (and is substantially more prone to error).

this would explain why there are so many wrongly id’d trees on inat.

You may not be familiar with the data on this topic, so here’s some info that should change your mind. Well over 90% of the plants on iNaturalist are identified correctly. For context, iNaturalist is slightly more accurate than herbaria records (at least in well-studied regions). I’m sure you’re aware that herbaria are often considered the gold standard of botanical accuracy. In other words, if iNaturalist has too many incorrect IDs, then literally no resource on earth meets your standard, and you may be best off with your previously referenced plan:

i save up my pennies and start a site basically just like inat, but it’s a market. my creation would destroy inat. well, this is how evolution works. kinda.

again, there’s absolutely no shame in making mistakes. but setting people up for failure is a different story.

iNaturalist is not setting anyone up for failure just because it does not necessarily get inexperienced users to the correct answer every time. That’s like saying that a ladder is setting someone up for failure because it’s not a teleporter. iNaturalist is much better than any alternative, and the “solution” you’re describing is technically impossible as “tree” is not a taxonomic rank.

[paul_dennehy]

Yeah, “tree” isn’t a taxonomic category, just an informal description of some plants based on their growing habits. There’s no “tree” taxon for the same reason there’s no “protist” or “parasite” or “winged animal” taxon- these aren’t monophyletic taxonomic groupings. They can be useful designations, and information like “this is winged” or “this is a protist” or “this is a tree” can be entered in the observation fields if one has the desire to track observations based on these characteristics, but they aren’t going to show up as taxonomic categories, because they don’t form a single clade on the phylogeny of life. (i.e. in my yard, my apple trees are closer related to my rose bushes than they are to my spruce trees, so any taxonomic classification based on evolution which includes both the apple and the spruce trees must also include the rose bush, and hence is not exclusive to “trees”.)

As you can see in your screenshot though, typing in “tree” does indeed give you a suggestion for the most specific taxon which includes all trees- Vascular Plants. I disagree that if all Shirley knows is “this is a tree”, choosing Vascular Plants is “a way too general suggestion, that entirely wastes and squanders what she does know”. If she’s unable to determine which of the taxa of vascular plants the tree is in (Flowering Plant, Cycad, Ginkgo, Conifer, or Fern), then Vascular Plant is the best she can do taxonomically. This is just how taxonomy works. I have plenty of moths which I’ve placed at “Order Lepidoptera” (butterflies and moths) despite knowing that they are leaf miner moths. Why? Because “leaf miner moth” is an ecological strategy, not a taxonomic grouping. I may know something more than “it’s a moth”, but if I can’t place it even to Superfamily, then Order Lepidoptera is the best I can do.

@epiphyte78

it seems like inat really doesn’t want anyone to admit when they don’t know a tree’s identity

I’m not seeing where this conclusion comes from. The initial CV output said “we’re not certain enough to make a suggestion” (the reason being that in this case the photo is zoomed out to show the whole tree- CV does better with closeups of leaves and other features). If I were posting the photo and saw that, and I didn’t know what it was either, I’d type “tree” like you did, and then click the first thing that comes up, Vascular Plants, which is correct.

A feature request that’s pending, which I do support, is to change how the suggestions display when the CV can’t come up with a recommendation. Currently, it lists some “visually similar” species after saying that it can’t provide a confident recommendation, but I would prefer if it refrained from doing this, and simply recommended broad categories in these cases. If the CV can’t decide between an oak, a privet, or a bay, and has no confidence in an ID, then maybe it should just recommend “Angiosperms” or something super generic. So I agree this display can be confusing.

@epiphyte78

not being able to link to wrongly identified observations is extremely counterproductive

The point of this policy is not to shame misidentifications, it’s to keep the forum focused on meta-discussion of iNaturalist as a platform, not serve as a place to argue over specific IDs and observations- that’s what iNat itself is for. Without this policy, the forum would devolve into a deluge of posts saying “what’s this plant”, “someone correct this ID mistake”, “someone fix this incorrect annotation for me”, which is the purpose of iNat the website, not the platform discussion forum.
If you want to discuss a specific ID on a specific observation, do that on the observation. It would be extremely counterproductive for the discussion relevant to that observation to happen here on the forum- take that discussion to the actual website and tag people if you want to bring their attention to something.

@epiphyte78

inat should have a policy of zero-shaming for wrong id’s

Yep, it already does. If you see someone violating this policy, notify the mods.

[hawkparty]

and perhaps there’s a bit of bias against botany. there’s no category for trees but there’s a category for tree squirrels. or maybe it’s more of an issue with taxonomy, which was invented way before inat was invented.

Yes, taxonomical classification does predate iNaturalist by a couple hundred years. The catagories are not chosen randomly, they are based off actual scientific studies. There is no taxa that encompasses “trees”. Take this as an opportunity to learn more about classifications and how they work, instead of assuming it must be a problem with the website that can be “fixed” somehow. Just because you don’t understand something initially doesn’t mean it’s wrong.

[swampster]

@epiphyte78

someone is going to chime in that i haven’t participated on inat enough to offer any useful feedback. ignore the noobs,

iNaturalist tends to be a very welcoming place for new (and old) users. The reason you are getting less than enthusiastic feedback is because a) you’re proposing sweeping changes without demonstrating you understand the full implications, b) your multiple threads have been ripe with condescension and unconstructive criticism, and c) you aren’t open to considering any alternative viewpoints.

[iisips]

“Tree” is a botanical definition, not a systematic one: the morphology of a large, woody trunk with branches has convergently evolved in many groups. There are “trees” in Angiosperms and Gymnosperms. iNaturalist search is organized by a taxonomy. “Tree” would be as sensible as having a search for “red” or “small”,

Maybe you will find value in the Visual Language Demo, which can search in this way for you. You can isolate it to “Plants” and then search “tree” – I have linked it for you with this search entered. It is a demo, still. If you want to contribute financially to something on iNaturalist (as your other posts suggest), you could support this project.

On the Identify page, you can search for observations above/below certain taxonomic rank or ranges of specificity. I have linked a search for IDed at Kingdom level (and no more specific) in Miami-Dade county, FL (where the park you mentioned is). Custom places can be created on iNaturalist for searching within a defined area.

[duncanross]

@natev

identifying plants is inherently complicated

What? You mean you don’t just know it is what it is because of the way that it is? ;)

[dgwdoesthings]

My favorite saying: you can tell because of the way it is, except when you can’t :)

to be fair most of my identifications on inat are solely through my sight id skills and a relatively few have actually been keyed out or compared to a description of morphological features

[seblivia]

@iisips

“Tree” is a botanical definition, not a systematic one: the morphology of a large, woody trunk with branches has convergently evolved in many groups. There are “trees” in Angiosperms and Gymnosperms. iNaturalist search is organized by a taxonomy. “Tree” would be as sensible as having a search for “red” or “small”,

Not only is this the case, but several different definitions have been proposed for what counts as a “tree”, with most of them either including things people don’t typically count as trees, excluding things that would colloquially be called trees, or both

[dgwdoesthings]

In answer to the title question, you can find unidentified trees by searching the observation description for the word tree, if a user happened to type it.

@epiphyte78 use this url term &search_on=description&q=tree can be used in explore page or identify mode
example in explore: https://www.inaturalist.org/observations?place_id=54078&q=tree&search_on=description&iconic_taxa=Plantae,unknown
Example in identify: https://www.inaturalist.org/observations/identify?iconic_taxa=Plantae%2Cunknown&verifiable=true&place_id=21&q=tree&search_on=description (there is a ficus that is currently unidentified on the first page)
all of these observations have the word tree in their notes section

@epiphyte78

there’s no category for trees but there’s a category for tree squirrels.

As others have said tree is a botanical definition not a taxonomic one.
So since labeling something on iNat requires the use of existing taxonomic frameworks, that is not possible. This could be achieved through an observation note or observation field.
However your example of tree squirrels is not any better and has the same problem, as the common name “tree squirrels” is applied to the genus Sciurus but Sciurus does not include every single squirrel like animal that can be found in trees and likewise some squirrels in the genus Sciurus might be more commonly found on the ground and not in trees yet they are not in the taxa ground squirrels “Marmotini”.
This is an inherent linguistic clash between language and scientific terminology that is unavoidable. Common names will always be flawed.

I think your issue is with phylogenetic taxonomy or with the common names for things, and the inat forum is not the place for that.
You seem to be in favor of a taxonomy based on ecological characteristic?

@epiphyte78

either shirley chooses a highly specific suggestion, all of which were very wrong, and will need to be corrected (but might not be for 6 years), or she chooses a way too general suggestion, that entirely wastes and squanders what she does know, that it’s a tree

The issue is that knowing that the organism has a woody growth habit and is thus called a tree is not really helpful in getting it identified relative to other features. And yes, there is no expectation that a layperson would know this, and that’s why enthusiast volunteer their time on here to educate others.

There have been many prior discussion on science and especially taxonomy being daunting to lay people and yes that is an ongoing issue in citizen science and human interaction with nature but your discussion is not adding anything new.

also can use jeanphillipe’s https://www.inaturalist.org/posts/73398-phylogenetic-projects-for-unknown-observations and simply go through taxons that contain what would be called “trees”

[epiphyte78]

right now i’m working through the ficus trees of Uttarakhand, India (with the ‘verifiable’ box unchecked of course). there’s an unusual amount of id’s of ficus sycomorus (native to africa). so far all have been incorrect. i’m guessing that inat is suggesting sycomorus. but at least in this case it’s a bad suggestion. well, relatively bad. i suppose it could be worse.

[earthknight]

I kinda feel like this post and your previous 2 should be combined into one post as they keep coming back to the same issues with the same genus.

[jeanphilippeb]

Projects for trees to identfiy (or identified after their addition to the projects):

• Unknown / Trees of southern Africa (1)
• Unknown / Trees of southern Africa (2)

These projects are populated with observations whose Computer Vision suggestions match this list of trees taxa.

[iisips]

the “issue” you describe is the is the very point of iNaturalist. the machine-learning recommendations are only as good as the identifications knowledgeable users make. and any user can review and draw attention to observations, at any time, if a misidentification is suspected.

contrast this with private collections where only authorized individuals may access the specimens. as with the iNaturalist dataset, there are misidentifications and overlooked details throughout them. i have visited collections in person and on online databases and found many misidentified items… however. i cannot contribute my knowledge or alert them to correct these. on iNaturalist, i always can.

the solution is to participate actively and contribute the knowledge you have. the community ID system is not broken at all. misidentification is often a part of the identification process, iNaturalist or not.

[twr61]

I thought that I recalled an old feature request to allow annotations to taxa that would be helpful here, but was not able to find it in a quick search. The idea was that one could, for example, specify the life form of plants (tree, shrub, perennial, annual, etc) and then search for these characters.

[dendroleon]

ID it as a plant and let the community refine the ID. you aren’t ‘squandering’ anything by doing that.

or i guess you could just troll the forums and not read any of the responses.

[bouteloua]

Closing the topic as the questions raised have been answered many times and it no longer appears to be productive.