The category of "cultivated" is problematic for plants in urban landscapes

I think the original intention of the cultivated flag was to just sort of weed out the house plant type observations. But, when many of the plant experts on the site also reference your second example… that’s where the confusion comes in for us non-expert users. Especially when the recommendation is to just use your “best judgment” on a case by case basis. Problem is that we don’t know how much we really don’t know…if that makes any sense! You can have 100 people submit an observation of the same tree in a forest preserve. One person knows the location well and that it was farmland 50 years ago and all planted recently. The other 99 people think it must be wild because it looks wild and natural.

1 Like

I don’t think anyone is suggesting that it’s a perfect system or filter, and I’ve been among those previously bringing up some of the same problem scenarios in the discussion above that you are now and suggesting possibilities for more granular DQA.

What I am suggesting is that spinning off into questions like “what is natural?” isn’t likely to get us anywhere in improving the issue of captive/cultivated tags. There are functional reasons why it’s useful (moreso for some than for others) to have observations pre-sorted by that filter, and there are certainly ways for that functionality to be improved.

2 Likes

We are discussing this, as iNaturalist is structurally discriminating against so called “casual” observations. E.g. by not giving them “needs ID” status, filtering them out of standard searches and by not showing them in the “compare” feature. The consequences of this discrimination are different for different taxa.
For birds for example, it makes perfect sense to filter out the occasional escaped cage bird - most bird observations in urban landscapes will still be easy to ID with the iNaturalist compare feature.
This is very different for plants. Most plants in urban landscapes are planted, and the most abundant taxa will therefore not be displayed in standard searches and will be filtered out in the compare feature. The discrimination of casual observations is therefore detrimental to plant identification in urban landscapes. Especially beginners, but also experienced botanists on a visit to another continent, will find that this filter as a major obstacle when trying to ID plants.

It also seems that some of the experienced users you mentioned, did not realize until now, that we are in the middle of an extinction crisis. And they seem to lack the foresight that data on planted and cultivated taxa is urgently needed, e.g. to locate the last survivors of species that are extinct in the wild.
I would think that the structural discrimination within iNaturalist against “casual” observations, especially in the case of cultivated plants, is a not very thought through feature.

hope that helps

4 Likes

Hey @mreith, I think you are having a separate argument than what the original poster asked. He wanted to know where to draw the line between a wild and not wild plant.

The discussion has been hijacked into a separate one on the validity or not of iNat’s policy on how it deals with casual observations. IMO we should have this discussion in a separate thread.

5 Likes

Great discussion(s)! I’d like to touch on three points.
First, and perhaps I missed a comment about this, though it’s super basic in my mind. If a plant is being watered through human means, it’s “cultivated.” For my part, I simply make that distinction and no others, sorry if that goes wrong in other people’s opinions. To me, a naturalized plant or a volunteer outside of yard, garden, or active restoration site seems a candidate for wild in my opinion. Just because its not native or usually occurring, doesn’t truly suggested it was planted. After all, dropped seeds, pollinators, wind, and animal droppings can also explain the presence of specimens outside expected ranges, beyond human intention.

Second, I have been guilty of marking observations as casual without providing an ID when it obviously lies within a yard, botanical garden, college campus, etc. I’m with the group of folks here who love iNat for its mission of observing wild nature and would rather spend my time trying to ID those observations. If people really just want to know what plant they are looking at they can use the plant ID app or just see what iNat suggests and not upload it, IMHO.

Third, users being creative with how to use iNat is just a fact (as previously pointed out), but having them try to change the site and mission to suit their purposes, seems a little too presumptuous and I hope iNat doesn’t alter the mission much to accommodate. I do see value in “cultivated plants” as hosts for other life and I am actively recreating my own yard to attract pollinators and other urban life, but I do not post observations of plants in my yard. Perhaps researchers, students interested in those kinds of uses could post the fauna in those environments and adjust the filters/tag to highlight the host plants…? Just my thoughts.

1 Like

I imagine there is some gray areas. But generally, I would think even a prairie restoration area should be marked as ‘wild’ at some point. Particularly, when there is no way to discern what was planted and what has seeded since the planting. I know I’m part of a group studying a Prairie/Fen which has many areas that have been managed with seeding and burning. One area has an entire ‘hill prairie’ that was transported from a different location. I would think all of those examples should be marked as ‘wild’ and that would be the most helpful for those studying the biodiversity of the particular Prairie and Fen location.

The other gray area is volunteers that pop up in vacant lots and gardens and yards. These I would also presume to be ‘wild’ if they were not planted and not intended to be there. But if you just planted a bunch of pale coneflowers in your front lawn, that should be ‘cultivated’ even if it is a naturally occurring plant in your area.

Wow…this is the first time I’ve checked this thread since initiating it. Thanks for a great conversation, everyone.

Just one quick thought: this debate seems to hinge on whether one comes to the question as a ecologist or a biogeographer. For the former, the key question is whether an organism is being “propped up”. For the latter, the key question is whether an organism has been introduced. These are very different meanings, and neither is unambiguously conveyed by the word “cultivated”.

3 Likes

this is unnecessarily hostile, and also untrue. Please refrain from personal attacks.

yep, let’s keep that in a different thread. It isn’t relevant to this separate question.

To me, it is very simple… when you join iNat you read the guidelines which spell out site policy regarding captive/cultivated, and then you apply the captive flag according to how you personally see it fitting the policy. If everyone does that, and even if we differ, the majority viewpoint will reflect in what gets applied by the system.

For me, the PROBLEM is that cultivated observations drop out of the identify pool. I respect that the emphasis on the site is for wild organisms, but I think it is not too much of an ask to have them as part of the identify pool, and have a sticky option (via settings) for the identifier to choose to have them filtered out, in much the same way as locations can be preset and appear as a default filter (for me New Zealand by virtue of accessing through the .nz domain)

2 Likes

its at least not untrue, sorry if that came over as hostile.

It appears you’re aware of the part of the post I was referring to since you edited it out of my quote. I’m glad you recognize it was inappropriate. Did you know you can edit it out of your post too?

I’ve thought at length about this; I published this blog post on bplant.org, which is a site I’m developing that shares some features in common with iNaturalist, including the reporting of plant observations, but that is strictly plant-focus. The post outlines a different, somewhat more detailed way of handling and reporting this stuff.

I’m not sure if iNaturalist is going to consider adding more fields or categories to deal with this stuff, but the way we’re handling it on bplant is to break this into two different categories: organism type (wild type vs. cultivar) and origin (Planted, Seeded, or Volunteer). There is also the capacity to leave both fields unspecified, as well as the separate capacity to report both fields as unknown.

Unlike iNaturalist which adopts the “wild by default” and then uses the data quality assessment to flag observations as “not wild”, our approach is “unspecified by default”, and then allowing conscientious users.

We also have online help popups on the form itself, that clearly explain exactly what each of these categories mean.

This contrasts with iNaturalist, where there is considerable uncertainty about what it means for an organism to be “wild”.

I’m not sure if or how much the approach we’ve taken would generalize to iNaturalist…bplant is strictly plant-oriented. The distinction of wild-type vs. a cultivated or domestic variety is certainly useful with animals though; eBird has such categories (they treat them as separate taxa even though scientists usually treat them as the same species) for many species of birds for which domestic varieties exist and are sometimes observed in the wild, including mallard, graylag goose, rock doves, and several others.

Lastly, and this isn’t discussed in the blog post I link to, some of these things are not about the plants themselves, but rather, about the area in which the plants occur. This includes a management regime, which may include specific behaviors like mowing, trimming, hand-pulling/uprooting of plants, herbicide applications, watering or irrigation, mulching or fertilization…and it may include more general practices like management for natives, management for aesthetics, management to remove specific species, or other considerations. We haven’t yet implemented a system to track these things on bplant, but we are planning to do so. It will be done by area and we will likely have it be managed or reported by individual users who are somehow involved in or knowledgeable about the management of these plots of land.

The information about the planting and seeding of plants into an area will already be tracked by user observations, but management policies will not be.

We have also considered a field for marking plants as killed / removed. For example, if I go out into an area, and see some garlic mustard but then pull it up, I could mark it as removed and specify the method I used to remove it. This could be useful to track and monitor efforts to control invasive species. We have not yet implemented anything like this…it’s just an idea right now, but one that our small userbase seems to think is a good idea (our users nearly all are involved in some sort of invasive plant control in some way shape or form), and we may implement it soon.

I hope this can spark some constructive dialogue!

2 Likes

To an extent I think you are conflating ‘believe the iNaturalist definition is wrong’ with ‘uncertainty as to what the definition is’

I think the definition is clear, yes newcomers may be uncertain and seek advice, but that is not an iNat specific issue, it happens on all similar sites. And any definition here or on any other community will face the same core question of ‘how do I know the source of this plant?’

The bigger issue is there is a cohort of community members who disagree with the definition.

3 Likes

It’s really not that hard, at least for plants. Did someone plant this individual where it is, or did it come up on its own?

I agree with @cmcheatle, it’s not the definition that’s a problem. IMO the problems are only:

  1. deciding what to do with cases where working out if someone planted an individual is difficult to determine; and
  2. deciding how to treat specimens marked as ‘not wild’.

Both of those are different (though related) discussions to “What constitutes a ‘wild’ or ‘not wild’ organism?” which is the gist of the OP’s post.

4 Likes

This is not at all clear from the iNaturalist documentation.

Nor does this cutoff you give, the way you have worded it, even correspond fully to the notion of “wildness” as people usually understand it. To use several examples of border cases:

  • What if the plant was seeded by people, but germinated and survived on its own? If this is the case, does it matter whether or not it was in an otherwise wild or unmaintained area, vs. a flower bed?
  • What if the plant came up on its own, but would not have survived except for the care of humans, such as by watering it, or removing competing vegetation?
  • What if the plant came up on its own, but is the obvious descendent of a cultivar, and is of a species that is unable to form a sustaining population in the wild? I.e. there are species that are not naturalized in North America and are not recorded by BONAP or similar authorities as occurring in the wild in North America, yet come up from seed in a garden environment.
  • What if the plant came up from its own, from a wild population, but someone transplanted it into another wild area in a habitat similar to where it might naturally occur? Or what if they transplanted it into a flower bed?
  • What if a person puts down a specific medium, such as gravel or sand, that did not exist previously, and this causes a plant to germinate that would not have germinated in the medium that naturally existed, and the plant seeded in on its own, and then people leave the plant to grow there on its own?

These are not theoretical examples. These are all examples I have personally observed.

And the point here is that there are numerous border cases. These are only a few and I’m sure we could think up more if it came down to it.

Going back to my first point though, it is also not at all apparent from reading the iNaturalist documentation that they intend to use the definition (planted or not) that you gave. I.e. I am quoting the documentation directly, here it says:

“If you observed something that is not wild, like a garden plant or a lion in the zoo, make sure to mark it as captive/cultivated to prevent it from becoming research quality.”

This is a vague statement, not a clear definition, and in the context of that statement it doesn’t link to anything clearer or more concrete. There is no system or list of criteria to determine where to draw the line in the difficult border cases.

I wouldn’t have wasted my time writing that blog post and coming here to share it and talk about it if there hadn’t been some sort of deficiency, and I’d suggest taking some time to reflect a little bit more carefully about it, because I think it’s pretty obvious that there is some room for improvement here.

I’m not even seeing anything resembling a definition in the iNaturalist documentation. Instead I see this, which references the concept of “wild” without defining it and without linking to anything giving a clear definition or list of criteria with which to determine what is wild. It gives a few examples, not definitions. And the examples are egregious, overtly obvious cases, not difficult border cases which are the cases which would need to be clarified if users were to have a reasonable sense of where the cutoff actuually lies.

Furthermore, in the data quality assessment, which specifically includes an option to mark whether or not an organism is wild, there is no clarification, link, or explanation of what “wild” means.

How are we supposed to have integrity to the data if the term “wild” is not defined, and if there is a clear definition somewhere, how is it that I’ve been searching for it for weeks and haven’t found anything more concrete, and instead stumbled upon this discussion where the OP was raising concerns about the cutoff being unclear?

Your response makes little sense and I am absolutely not convinced by what you’re saying.

If there is any “iNaturalist definition”, it is either vague and implicit, or hidden from view, possibly buried in documentation that no one reads. I’m someone who has actively sought out as much documentation as I can find, and read it with a critical eye. If I find it vague or ill-defined, the average user isn’t even going to have thought about it at all.

Your explanation and reasoning here I think illustrates how the concept of wildness as construed by iNaturualist is ill-defined. You start by saying:

But then in your reasoning, you go on to say:

Growing on its own, in competition with other plants, is not the same as whether or not a plant is planted.

I can plant a plant in a wild area, then leave it and never intervene, and it’ll be left to compete on its own with other wild plants.

Conversely, I can go into an area, and cut back or uproot the plants in competition with a plant that seeded into an area on its own. This will be giving the one plant an “artificial” boost. Perhaps the plant would not have survived on its own. Perhaps it would have, but now, with my help, it grows faster, bigger, and produces more seed.

The point I’m trying to make here is that wildness is more complex than just whether or not a plant is planted. You begin by stating that this is just about a plant was planted by people, but your reasoning and logic demonstrate that you are actually using a different, broader definition in your head. You are thinking about wildness in a way that implicitly assumes a lot more than just whether or not a plant was planted. People who use the iNaturalist data to do scence may also make similar assumptions. One of the biggest problems with trying to do science is becoming aware of your implicit assumptions…when I read peer-reviewed papers, I’d say it’s the #1 flaw I find in people’s arguments. It seems highly likely to me that people using the iNaturalist data might make these assumptions, and thus derive flawed conclusions from their research. However, even if they don’t, the lack of more descriptive fields and lack of integrity of the iNaturalist data (because let’s be honest, most records in the system, including ones marked “research quality” don’t have people putting care into checking that field, because it’s buried down in the data quality assessment that gets very little use relative to the amount of energy people put into submitting new observations.) is going to severely limit what people can do with it.

And thus, the simplicity and limitations with which iNaturalist currently records this data, are going to show up in research people try to do using their data. And this is going to limit its usefulness.

In the blog post I shared, I gave numerous concrete examples of ways in which the additional information we record on bplant.org could be useful for tracking plant populations, especially in cases that are relevant for conservation purposes.

In my experience, wildness encompasses not only whether or not a plant was planted (or seeded, which is an intermediate case your definition doesn’t mention), whether or not the area it is growing in is maintained and other plants are being weeded, trimmed, or otherwise removed or inhibited, whether or not the plant is a cultivar (and may have altered physical characteristics as a result) vs a genetically-unique wild-type plant, and possibly even other things.

In urban areas, as the OP raises concerns about, an overwhelming majority of plants are going to be “border cases” like this. And of course, more people live and work in these areas, so people reporting the plants closest to where they are naturally spending the most time, they’re going to be mostly running up against these border cases.

I think this illustrates why, in order for the field determining wild vs. non-wild to have more integrity, it’s going to need to be more explicitly spelled out. And I personally think that it’s probably not going to be as simple as making it a single either-or field.

The way we are handling this on bplant.org is to break it into two separate fields. I’m not sure whether or not this is the best way to handle it, but, after considerable thought and discussion with others, including seeing the concerns people have been raising here, I think it is probably at least a slightly better way than how iNaturalist is doing it.

iNaturalist currently has the problem of there not being a clear definition that is well-advertised and easily accessible and clearly displayed to all users before they specify whether or not organisms are wild. And then there is the whole “wild by default”…which leads to huge swaths of observations (just look at Manhattan, NYC) of plants that are clearly not wild, but that haven’t been marked as not wild because only a tiny portion of users ever completes the data quality assessment.

This could be very easily solved by just making the field default to an “unspecified” state and requiring people to explicitly mark organisms as wild.

I think little would be lost by this too because there are many species that are rarely or never cultivated or domesticated and for these, the field could be ignored. It seems like a win-win, a no-brainer to me.

1 Like

Don’t all these fall under:

1 Like

I agree, there are border cases, or “grey area” cases as I call them… but iNat has a DQA flag that peiople can vote with, so assuming everyone has read the guidelines, considered them and allowed their own personal view to be influenced by it (and there are many users that have such strong feelings one way or the other that they can read and NOT alter their perspectives), then the aggregate result of that vote should reflect an accurate view of where the organism sits in the captive/cultivated spectrum. ANY individual looking at an individual case is always going to be able to pick holes in how it sits with the guidelines.

For me, my feelings are most elevated around the issue of re-vegetation efforts (habitat restoration), and how the successive generations of those plantings should be treated…

2 Likes

Here is the definition right in the open on the help page
https://www.inaturalist.org/pages/help#captive

I don’t know about the app, but there absolutely is a link in the dqa 3 of the 4 first words in the section are a link.

You can argue that cultivated is a poor threshold or an unclear term, but that help page makes it very clear, if it was yesterday or a hundred years ago, if a human planted it, to iNat it is cultivated.

As others have said I fail to see how your examples are an iNat specific issue and not a general issue for users of any biodiversity platform. None of your 5 examples can not be validated by asking the simple question iNat’s criteria lays out - did a human plant it or not.

For several of yours the answer to that is no, therefore it is considered wild. As I stated, disagreeing with that threshold is fine and saying no that is a cultivated plant, and many users here do, that does not make it unclear.

3 Likes