Hi everyone I am new to iNaturalist and have recently started paying more attention to the plants, insects, and birds I come across during walks.I was wondering what approach experienced users recommend when trying to identify a species that the app isn’t very confident about. Do you rely mostly on the AI suggestions, field guides, or community identifications?Any tips for beginners on improving observation quality and getting more accurate IDs would be greatly appreciated.Thanks!
Well, I am not exactly an experienced iNat user, but I only identify taxa when I am really confident about it. If unsure, I use a top taxon which I am sure about. Like angiosperm or bird, or beetle.
Sometimes, in identifying, the problem is not so much to say if an organism resembles the suggested picture, but if there are other species with which it could be confused. Sometimes Wikipedia helps with that, sometimes not.
With unknown observations, I prefer to ID them to iconic taxa (plant, animal, fungi etc) that way I can quickly do a lot and leave finer ID’s up to more experienced users
Just FYI, the inadvertent choice of words in the title will probably confuse people (seems it already has) since a common activity of identifiers is to identify the so-called “unknown observations” of other users. As far as I can tell, you’re asking about identifying your own observations, which is a different process altogether. You may want to alter the title to reflect this.
Hi @neman, welcome to iNaturalist! You don’t necessarily need to worry about making species-level IDs for all your observations, you can just identify them to the level you feel confident (be it species, family, or just saying plant or animal) and let the community go further. The model can be useful, as sometimes it is easy to verify if the recommendation is accurate, but be careful, especially with obscure groups, where it can fail often.
If you do want to make more accurate IDs, there are many useful resources, which depend on the group you’re working with. For popular groups like birds, mammals, herps and such, field guides can be excelent and contain all necessary information. For more difficult groups, like insects, there might be no field guides available for your region, or they might be incomplete, covering only the most common species.
Very often, you can also find very good information online, including catalogs, guides, and keys. iNat itself is also a great source; beyond the model, you can see which species within a group can be found in your area, and you can browse observations with comments to find advice on how to recognize them. Something I find especially useful is the Similar tab. If you go to a taxon page, under the picture, you can find the Similar species, which shows species that can be mistaken for yours. There, if you click on the number of observations of a species, you’ll find the observations which have been identified as both, where very often you’ll find expert advice on how to tell them apart.
I hope this helps!
There is no easy answer to your question. When I look back on my early observations, it’s pretty clear I didn’t really know what I was doing. Most importantly, I was overly reliant on the algorithm. The one piece of advice that would have helped me most is: if you don’t know what it is, don’t guess a species-level ID. It’s okay to guess a genus-level ID based solely on the algorithm but then don’t routinely agree with any species-level ID offered by someone else.
Hear! Hear! This was exactly me. I am finding that I now sometimes err on the other side when it turns out I do actually know what the species is. I am learning to walk the line, but it has only come with experience both in observing and in venturing into making identifications for others. (I started with very common and easily identified local species and later moved into unknowns as well as a few trickyish local species that I learned how to key out for my own observations.)
My simple advice: look at CV, but never ID to species unless you are pretty sure from your own knowledge (which could be from a field guide or simply life experience). DO put some kind of an ID on though, whether to genus or something higher. Don’t leave it blank. You can always simply choose “birds,” “winged or once winged insects,” “flowering plants,” or something like that.
@neman Welcome to iNaturalist and have fun! It has become my favorite hobby and my knowledge has really grown with time!
Part of the learning curve is finding out where what you know fits as a taxon. My example was preying mantis – mantIs will be MantIdae ? No it is MantOdea. For my level of knowledge and interest. Displaying common names in up to 3 languages can also be helpful. Follow your notifications and use the response from taxon specialists for your learning curve.
When observing – try to capture field marks, 3 views. If plant then instead of dump it in the green waste – show us a wide view for scale / habitat. Then leaves, flowers, fruit. Or is it seaweed, moss, fern. Pine or palm tree ?
@neman I am so glad that iNat has gotten you more interested in the environment around you. It is amazing what you can notice, and the little dramas you will witness as you are more attuned to the world around you. For example - watching a tiny bird try to chase off a large squirrel that was showing interest in the birds nest (both cool animals to watch; and the drama of ‘who will win this battle?’ All interesting things to see, and that you would probably miss it you weren’t ‘tuned in’ to the life around you. BTW the bird won this round).
In terms of your learning to ID more of the things you see - I’d recommend learning from others and field guides. CV on iNat (or AI in other formats) can give you a clue, but don’t trust it totally. Tip - when using any AI - Make SURE you have the location entered before you even look at the suggestions or it will be pulling from a worldwide data-base of possibilities. If the location is entered, it will generally suggest more local possibilities, but double check, so you won’t be wasting your time on things only found on some place on the other side of the planet from where you are.
In terms of posting to iNat, as many have said - just ID to the level you are comfortable. If you know something is a ‘lizard’ but not what kind, it’s fine to just put ‘lizard.’ Assuming you have clear pic(s) others will help you identify it and then you will know what it is next time. iNat is a great learning tool. Please do put some ID (even if it’s just ‘animal’) rather than leaving something ‘unknown’ unless you really have no idea. This is very rare but can happen (I recently saw something on a rocky outcropping in the ocean, near a beach I was at, and really couldn’t tell if it was a seaweed, animal, or man-made object that washed up)
When taking pics, try to make sure they are clean and follow the suggestions from @DianaStuder above.
Most of all have fun.
As you use this site and pay more attention to the things around you, you will learn, likely much faster than you could even dream of.
If you click on each ID suggestion, it brings up a page with more info about it. I’ve found the taxonomy tab here to be really helpful for figuring out what higher ID to use when I’m unsure. So if the computer vision gives you 5 options but you have no idea which of those options to pick, you could go into each one and see which higher taxa fits all those options. For example, I know this is some sort of millipede or centipede but have no idea which kind but using the taxonomy tab, now I know that they both fall under the category of myriapods. The map feature can also be really helpful showing you what the expected range for each option is.
This is good advice, not just for plants. A lot of mollusk observations, especially bivalves, wind up in the unidentifiable pile because only one side of the shell is shown.
For the Other Kingdom ;~) Face, side view / profile, back / top – and if that ‘worm’ has legs, then how many ??
Lots of good advice from others already.
Following notifications is important, both for your own learning experience and for not wasting the time of identifiers. If you make an ID you weren’t real sure about, and someone more knowledgeable disagrees with it, you can always withdraw your ID (and should) and there is absolutely no shame in doing so. The occasional mistake is just part of the scientific process, and the withdraw feature is there for a reason.
The AI can be useful for putting some label on an observation, in order to help the right members of the community find it. For example if AI suggests sugar maple, red maple, and Norway maple, and you ID as maples, that helps identifiers who specialize in telling different maples apart find it much faster than if you had left it at “plants.” And if one of them says “No, this is a viburnum.” You can withdraw your ID of “maples” and let their ID of “viburnums” shuffle it over to the people who identify viburnums.
If you can afford good field guides and have the time to use them, I do think they are the best learning tool if your goal is to truly learn how to identify things for yourself. The combination of AI and then input from human identifiers can tell you what the organism in your photo is, but will you recognize the same species in a different place and at a different stage of growth? Or will you be able to tell when you find a different species that’s closely related to the first? Field guides force you to do the work of identification for yourself, which is a more effective method of actual learning, vs. just being given the answer.
You have already gotten a lot of useful replies but I think I can chime in with some of my experiences :)
As you were already told, it’s best to ID only to the level you’re confident (Even if that means “plant”. I wouldn’t even be confident on “plant” if I saw an algae!) but there are some ways to improve your guess if you want to put the time in:
-
For birds specifically, I bought an identification guide that contains all the birds that have been registered in my country, with full colour pictures of each and a small paragraph explaining how to distinguish them from the others. It has been insanely helpful. Birds and mammals are easy for this because there are very few species, insects and fungi are a whole other beast. These guides are pricey, and I would avoid any guide that says something like “100 xxx from your place” as they are useful to give you an idea of what’s out there, but not to ID an individual.
-
For any animal that I more or less know which kind of animal it is (Eg: A ladybird, a ray-finned fish, a mammal), I usually go to “explore”, choose an area quite large around my observation, and check all the species of that sort of thing that have been observed. If there are few observations I make the area even larger. Then I check observations of each species until I narrow down which ones seem to match, and I look at the taxonomy of these to see where do they all merge. This one has the advantage that you don’t actually need to know how to distinguish them from one another to be able to use the information to narrow down the search :)
-
There is a project titled “Observations with ID tips” that lists any observations where a person added a comment saying something useful for IDing. You can navigate the observations of that project to see if there’s anything added in the group of organisms that you’re trying to ID (Eg: check all swallows and open observations of the ones that look like your swallow) and see what people say about it
-
If you open any taxa and click on “See more” on the section with the default pictures, you can check all of the pictures from that taxa, and you will have a bunch of filters at the top that you can use to see only certain sex, age, kind of evidence, etc. I used that extensively, for example, to check all of the ladybird larvae from my city and figure out which one was the one I photographed, which would have been excruciating to do by looking at the observations of the species one by one. (The filters use the information MANUALLY ADDED to the “Observation fields”, so make sure to fill those whenever possible! Most people don’t, and it’s a very useful tool)
@neman I’ve altered it to “unfamiliar” species.
Welcome to iNaturalist!!
Wish I was newby again
- See what iNat suggests, and do so before you enter anything in the identification box.
- Compare your observation with other observations with suggested ID for general appearance and range. If it doesn’t look right, try another suggestion.
- Think of CV as another person that looks at things a different way and sometimes makes obvious mistakes.
Lots of good advice in this thread so far. I would just add that there are other similar threads on the forum that will give some other good community answers, like:
https://forum.inaturalist.org/t/im-no-specialist-how-should-i-proceed-with-my-observations/28639
iNat also has good Help resources for how IDs work and how users should approach IDing:
https://help.inaturalist.org/en/support/solutions/articles/151000170241-what-is-an-identification-
The two help articles cited above are for users who identify the observations of others. Are there similar help articles for observers? Since the two processes are not the same, advice is different in each case.
I disagree that
The guidance there applies to users who identify their own and others’ observations. In fact, some of the guidance is more targeted towards users IDing their own observations, e.g.:
“When you make an observation, identify the organism as best as you can, even if that is just “Plants” or “Birds”.”
“Don’t use the Agree button as a form of acknowledgement or gratitude, you can write a comment for that if you like.”