Kousa Dogwood (Cornus kousa) observations- wild or cultivated?

The question of whether the non-native Kousa dogwood should be considered invasive came up in an outside invasive plant group I belong to. The prompt was this article. Curious about this question, I queried the iNat database for all research-grade observations of the species in North America, looking for a trend over time. But randomly checking through individual observations, I noticed that most of the images don’t establish clearly whether or not these are wild. For example, many images have only a single close up, without the larger context of the location. In a few cases, the images were obviously captive (parking lots) or highly likely to be (arboretum, grassy or mulch-y background, etc.). I’m led to conclude that for this species, these observations aren’t reliable enough to estimate or track the wild/invasive population. It also makes me wonder, why is the default “wild”? Why not require a choice of wild vs captive upon upload?


In short, there is a choice of wild vs captive upon upload, both in the app and on the computer. It’s next to the geoprivacy settings on each observation. It would be really annoying for experienced users to have to get a pop up asking for confirmation that an organism is wild for each observation, so that’s probably not a solution.

So far, the solution I’ve found is manually filtering through each observation and asking the OP if it’s cultivated. If they don’t respond in a few weeks, I may ping them again, or I may mark it as cultivated and explain my reasoning in a comment. This only works if you’re willing and able to read your own messages and respond to a correction from the observer. For instance, a photo of a dogwood in a pot may be a wild dogwood that was dug and put in the pot; the observation could have the location and date of the wild observation, and merely use the potted plant as evidence that the species ID is correct. Tempers can fly when you mark something as cultivated, I’ve found. Experienced iNat users can get offended easily by the insinuation they don’t understand wild vs cultivated rules, and inexperienced iNat users may not understand iNat’s definitions. Anyway, after a few months of that, you should be able to generate a useful map of wild vs cultivated observations of Kousa dogwood.

Another thing I like to do is use the observation field “Invasive?” for plants that the poster claims were naturalizing. I don’t like to assume with that observation field, unless for instance it’s a population of invasive organisms that I have personally seen.

I think I saw the same Cornus kousa discussion. It was interesting.


The default is wild because iNaturalist’s focus is on wild organisms.


I pretty much agree with everything you said here-- I take a similar strategy when sorting cultivated observations-- but it is unfortunate that it can take a few months of work to come up with a general sense for where a species is escaping. It makes sense to ask if we can do better. I know it has been discussed on other threads before, but it would be nice if somewhere in the account creation process new users would have the captive/cultivated distinction explained more deliberately.


Yes, I know. However, many new users don’t get this right away. Or don’t care. But in just one species with a relatively small number of observations, I could quickly pick out research-grade observations that were captive, or likely to be -suggesting it may be the case for other species as well.

ok, I was just answering your question

Sometimes people add a note for plants that look cultivated but aren’t. That’s what I do, and I saw one like this for the dogwoods. Regarding messaging people, I’m not comfortable doing this, as a non-expert and irregular user.

Thanks for sharing your experience. I think this problem of wild-vs-cultivated plants is a major data quality issue for iNaturalist. The general disregard for logging establishment means was shocking to me when I first logged on to iNaturalist. iNaturalist could be a great tool for tracking biological invasions, but instead, as you’ve noted, it ends up being mostly useless in that regard because so many users upload cultivated plants without marking them as such.

But this concern is not universal. Some people think it’s just not a big deal to have a bunch of garden plants undifferentiated on the site and others oppose seemingly any change to the uploading process, even changes that don’t require extra steps (unlike e.g. the pop-up referenced above in this thread), preferring to shunt the workload of marking things cultivated onto others. To read some of these viewpoints see some responses to my proposal for separate “wild upload” and “not wild upload” buttons here:



Thanks for the link to that original article. I had no idea cornus kousa was becoming invasive. In my travels I only notice native cornus florida, especially along the edges of power line cuts and roadsides. I’m going to have to keep an eye out where people have them planted as yard trees in close proximity to natural areas. I noticed that’s where some introduced trees can easily get a foothold. I’ve found plenty of Callery pears, Norway and even Japanese maples in the woods several hundred feet away from the parent tree growing in someone’s yard. If I happen to find any Kousa dogwoods that are obvious escapes I’ll be sure to document them.

1 Like

Snarky planty ID alert: It would have been awesome if they had picked a picture of an actual Cornus kousa for the article…

There’s plenty of them planted around our campus, as well as lots of Cornus florida. I haven’t noticed it yet in locations where it wasn’t planted but I’ll keep an eye out for it.


This is the big issue in relation to ornamentals that could escape from cultivation. Especially if the position is imprecise with a rather large radius.
Ornamentals that most likely escape close to where are cultivated, in these cases, become hard to be evaluated as far as their wild status is concerned. As regards, if the observation is made in an urban area, also a 50 m radius could be too much to allow the evaluation.
Unfortunately many users seem to be attracted by small particulars of a large plant and to forget the importance of a photo of the whole plant.
Also the possibility to upload an observation with a single photo made in the moment, instead of more photos taken from the gallery, contributes to the belief that one photo can be enough.

I can only suggest this: if you know well the ecology of this species, you can hypothesize which observations refers to true wild invididuals and which ones depict a cultivated specimen depending on the position.
Moreover, often large, mature, healthy and flowering/fruiting plants growing very close to buildings are more likely to be non-wild.
In the case C. kousa has been also employed for reforestations/afforestations, this could make things more complicated since large plants growing in wild areas could be just planted.

If they provide a photo of the whole plant, they get told that it cannot be identified to species without such-and-such small particulars.

1 Like

Exactly, which is why I try to capture the entire plant, along with photos of the small identifying features. By themselves both are useful, but more so when taken into consideration together.

1 Like

That’s what I am writing since some time ago. Just one photo, either a detail or the whole plant photographed from afar in its environment often is not enough. Just two photos would be much better,
The difficult task is to convince users of this necessity.

1 Like

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.