How many cultivated plants are marked as such on iNaturalist?

Thought I'd try to address this question with a semi-scientific approach:

For the past year (1/12/2020 - 1/12/2019) in a 25 square-mile square centered on Blacksburg, VA:

To the best of my ability to determine, there have been:

131 certainly cultivated, marked wild (plants in orderly rows in flower beds, landscape plantings, hanging baskets, houseplants, exotic trees not known to spread from cultivation, etc.)
90 probably cultivated, marked wild (mostly street trees and commonly cultivated plants that could conceivably be escapes or otherwise wild, but are almost certainly planted, just lacking in enough photo context clues to be certain)
1150 probably wild, marked wild (when in doubt, or totally confused, I chose this category)
131 certainly cultivated, marked cultivated
5 probably wild, marked casual (for some reason other than "captive/cultivated")
2 probably wild, marked cultivated (both were chickory, which were growing as weeds in a lawn) for 1509 tracheophytes in the Blacksburg area:

76%-83% observations were of wild plants;
17%-23% observations were of cultivated plants;

of the cultivated plants:
37% were definitely incorrectly marked as wild;
62% were likely incorrectly marked as wild.
37% were definitely correctly marked as wild.

Posted on January 13, 2020 02:09 AM by ddennism ddennism


This is interesting, thanks.

I'm sure you know this but where you choose to sample matters greatly. That is, the spatial distribution of cultivated plants is not random.

Posted by trscavo about 3 years ago (Flag)

So by extension, the iNaturalist servers are hosting a relatively enormous number of inappropriate listings at the national or continental level, which slows data searches, adds operational/storage expense and contributes to global warming via use of fossil fuels to run the mechanical hard drives and CPUs processing all those unnecessary records. What remedy can you recommend? An AI system that flags inappropriate (cultivated) records for possible deletion by an editorial board of some kind?

Posted by kgivens about 3 years ago (Flag)

Yeah, @trscavo - I chose an area similar to where I live. I'm sure it's different in wilder places. The loose plan is to compare USA state universities with and without biology classes that use iNat for bio-blitzes.

Posted by ddennism about 3 years ago (Flag)

Thanks the interest, @kgivens . The context for this journal entry was basically a sanity-check for myself after reading some responses to my proposal here:

I wanted to see if my local environment was unusually inundated with observations of cultivated plants that are mis-classified as wild. I've also been frustrated recently by the large number of Trillium observations that I suspect are not wild, but can't really prove that they aren't escapees/far-flung native occurrences. I share your worry about the consequences of iNat being a host for everyone's garden photography, but my main concern is that downstream data-users will ignore iNat in the future because a cursory glance at the quality of the plant data will show it to be a poor mix of cultivated and wild observations.

Posted by ddennism about 3 years ago (Flag)

Add a Comment

Sign In or Sign Up to add comments