Description of need:
iNaturalist automatically marks an observation as ‘captive/cultivated’ if there are at least 10 observations of that taxon (at genus and below? higher taxa?) in the area and 80% of those observations are ‘captive/cultivated.’ I feel that this is a good feature. However, in genera in which some species are overwhelmingly ‘captive/cultivated’ and others are mostly not, it would be helpful if the automatic vote were revoked when an observation at genus is refined to one of the species that is mostly not ‘captive/cultivated’.
Feature request details:
Case in point: the genus Magnolia in Washington, D.C. There are tons of magnolias (especially Magnolia grandiflora and the various common magnolia cultivars) planted around the city as street trees, in yards, etc., that are (thankfully) automatically marked as ‘not wild’. However, there are established (and self-seeding) wild populations of Magnolia macrophylla and Magnolia tripetala in the wooded parks of the city. Here’s the current breakdown:
Magnolia cultivars (M. stellata, M. × soulangeana, etc.): 495 observations, 100% casual
Magnolia (at genus): 438 observations, 97% casual
Magnolia virginiana: 153 observations, 56% casual (should likely be higher)
Magnolia macrophylla: 89 observations, 18% casual
Magnolia tripetala: 84 observations, 4% casual
Magnolia acuminata: 21 observations, 95% casual
Now, if you’re walking in a woodland park in D.C. and you observe a wild magnolia but are unsure of the species and suggest the genus Magnolia, that observation will be automatically marked ‘captive/cultivated’ by iNat (unfortunate, but a good tradeoff versus having hundreds of street-tree magnolias not auto-marked as ‘captive/cultivated’, in my opinion). However, if an identifier finds that observation and refines the ID to a species that is not 80%+ casual (in this case, M. macrophylla or M. tripetala), the iNat ‘captive/cultivated’ auto-vote remains (Some examples here, here, and here; ‘wild’ votes added by me). I suggest it should be automatically removed in such cases (when there are at least 10+ verifiable/RG observations of that species in the area or something of that sort).
Counterpoints: if an identifier knows enough about the iNat system to go looking for mislabeled ‘casual’ observations to mark as ‘wild’, they can just take the extra step to add that vote themselves. Plus, the ‘captive/cultivated’ mark does get automatically removed if the observation gets to ‘research grade’, though that does require one more person to find a ‘casual’ record if the first refiner did not vote against the iNat auto-vote.
Thanks for reading. Perhaps others can chime in with better suggestions (or other counterpoints), or examples of other taxa for which this is an issue.
I’m not familiar with that; do you mean native wild roses put at genus Rosa get auto-marked as ‘captive/cultivated’? In my area there’s so much invasive Rosa multiflora that even if cultivated roses were all properly marked as ‘captive/cultivated’ I think it would still be below 80% for the genus.
No it’s the opposite, where people uploading cultivated roses don’t have them automatically set as cultivated because there are enough legitimately wild rose observations to where it doesn’t hit the 80%
Yes then my request is about the opposite of the rose situation, where a genus is at 80%+ casual (so all observations at genus get auto-marked ‘cultivated’) but there are species within that genus that are mostly wild. I don’t have a problem with that, but I think if someone adds an ID moving the genus observation to one of those mostly wild (<80% casual) species, the ‘cultivated’ auto-vote from iNat should get removed.
Not sure! Looks like six of those are on the national mall itself, and those at least are almost certainly planted. I think it’s more properly native farther west in Appalachia. There are quite a few ‘native-ish’ (native to somewhere in eastern North America) trees planted in the city (as street trees, in the native sections of arboretums, botanic gardens, etc.). Actually, I think the only ‘properly’ native magnolia in the area is M. virginiana, and even those are probably more planted in yards and gardens than actually growing wild. The most popular magnolia in the area, the Southern Magnolia (M. grandiflora), has a native range much farther south, yet is not marked as ‘introduced’ (in fact, it has that green ‘endemic’ star next to it’s name!).
I suppose this would be done when the observation is reindexed after a new ID is added. I don’t know how computationally difficult it would be, but it would add at least a little load.
Have a feature request too. Stop automatic iNat ‘captive/cultivated’ tag within the natural range of an organism.
It happens to often that organisms are automatically tagged as ‘captive/cultivated’ by iNat within their natural range. Last example i found was: https://www.inaturalist.org/observations/250827809
I would be against this because of the counterpoint outlined in the initial request. Dropping iNat’s cultivated vote after a (likely somewhat) expert IDer has interacted with the observation is going to lead to a counterintuitive outcome for them - the identifier left an observation with a species ID and cultivated definition to the best of their knowledge, but they will have effectively (and silently) removed the vote for cultivated. If someone goes through to improve some IDs on cultivated plants, they may be effectively and erroneously adding a bunch of cultivated observations to the Needs ID pool. I understand the idea/motivation behind this, but I am generally against systems that lead to results that the users are unaware of/counterintuitive. IDers would probably catch on in time, I guess, but I feel like the existing “safety” mechanism of removing the automatic cultivated vote if/when the observation reaches RG at a species that doesn’t trigger the automatic vote is a good enough balance in this situation.
You know, that’s a fair point, and it’s not something I thought of when I made the feature request. It may also be that the case I outlined—species that are mostly wild within genera that are mostly cultivated—is not actually that common, and may not warrant a change to the current system.
I do think many users are unaware of the iNat autovote, and are probably hesistant to vote against it (if iNaturalist says this is captive/cultivated, it must be so!). Just yesterday I ran across an observation of a (non-native, common street tree) elm sapling growing out of a crack in the sidewalk—it triggered the iNat autovote, and then the user added a second ‘not wild’ vote. Again, probably something that doesn’t happen too often, and maybe just vote ‘wild’ and leave a comment (what I did) is a good enough solution all told, but I feel it’s good to bring up these cases where the current system might not be working as well as it could (I may also be seeing it a lot more than others, as I mostly ID plants in a highly urban area).
