Range maps for vascular plants

I’ve been thinking that we should start adding range maps for vascular plants. For example, USGS has range maps available in ESRI shapefile format based on the Atlas of United States Trees (Little, 1981). They are not 100% accurate (especially for Canadian and Mexican ranges of certain tree species) but they provide a good starting point.

Would any botanists be interested in initiating this? Any hesitations or concerns?

3 Likes

Hi–housekeeping note that I moved this to its own topic since it’s only tangentially related to a tutorial about creating range maps.

On iNat, I find them less useful than atlases, and there’s no history of tracking changes or a good area for documentation/collaboration, which is why I haven’t really focused on them.

1 Like

Range maps on iNat should be global, so if said tree is endemic to North America, then the Little map is appropriate, but if its not it wouldn’t be appropriate

1 Like

Yeah, i think a push to fill out more atlases based on said maps might be more valuable.

1 Like

You are probably aware that the first few volumes of Flora of North America North of Mexico included detailed range maps, instead of just state-level dot maps…They were small, though, and I don’t know if they were GIS digitized or were just hand-drawn. Could probably find out. Another potential source of range info…

1 Like

Another thought, if we really wanted to make a push to populate atlases, would be to mine the online specimen data portals of the world. Or has GBIF has already done that for us??

Most online specimen records should have at least two administrative levels in their data (Country/State-equivalent) if not a third (County-equivalent).

Could we develop a one-off batch process to match those data up to iNaturalist Atlas places, and populate atlases for taxa that don’t yet have them?

To avoid problems with misidentified specimens, maybe when a taxon has at least two independent specimen records from the same administrative unit, that unit gets added to the Atlas.

For taxa with existing atlases, a more circumspect reconciliation process might be needed.

I guess the big question for iNat staff would be whether the value of the resulting atlases would exceed the time and effort to do something like this.

3 Likes

How does one access the atlases? Just today I was searching in vain for a plant range map I could trust.

Atlases aren’t a fully fleshed out feature. Right now only curators have a link to create and view atlases on each taxon page. It’s in the Curation drop-down. Unfortunately I don’t think there’s a way for you to search atlases to find out if one exists yet for a particular taxon…

image

1 Like

Thanks. I used to just rely on calflora.org. But now that the taxonomy of iNaturalist diverges more from calflora.org, I use BONAP. Neither had a map of the plant I’m looking for. So I look forward to when correct maps are somehow on iNaturalist instead of just maps of right and wrong identifications.

1 Like

I have a concern that there are quite a lot of observations of cultivated/captive specimens not marked as such which would skew maps.
It is quite common for these even to be ‘Research Grade’.

3 Likes

i spend a fair bit of time looking at range maps and checking the outliers, it’s pretty easy to clean them up with either fixing the IDs or marking as captive, it’s a quick way to do a lot of high value data quality control.

6 Likes

Same here. When I am investigating a particular taxon, first thing I do is go to the map tab on the Taxon page, and visit the pins that seem to be out of place. Most often it’s a mis-ID, but sometimes it can be a captive observation.

Next thing I do is check the observations for “similar species” to find mis-IDs that are really my taxon of interest. Curation is a never-ending process for everyone on iNaturalist. Just the nature of the beast, so to speak.

1 Like

Reverse populating from GBIF is an interesting idea. I’m not sure it has to be an either/or, but I would actually focus on using it to populate the checklists first before the atlases.

Checklists are much more user available, and I’m not sure the site has quite figured out the use case for atlases yet, Right now they have a kind of ‘solution in search of a problem’ feel to them.

One concern I would have is that sites like GBIF dont tend to update their records when taxonomy is updated. For example, the Northern and Hen Harriers of North America and Europe respectively used to be considered a single species under the name Circus cyaneus.

If you go look at the GBIF page for C. cyaneus https://www.gbif.org/species/2480487

It has thousands of records for North America that are no longer valid, which would get imported. Cleaning up stuff like that, and any potential impact on if they add geographic intelligence into the computer vision system could be a nightmare. And then making sure if cleaned, it does not simply return should the process ever be run a 2nd time.

1 Like

Yes, a taxonomic cross-walk to current iNat taxonomy would have to be part of the process, no matter how up-to-date (or not) GBIF taxonomy is. Hopefully we curate the synonymy well enough at our end that most of that could be automated. But some would have to be manually resolved no doubt.

The main use-case I have in mind is to inform better geo-awareness of the Computer Vision suggestions in iNat. I don’t know whether atlases would make that easier, or if drawing directly from checklist data would be just as effective. But I would certainly want to at least tap in to existing atlases as appropriate.

Works well for lumping or straight swaps, not so well for splitting.

1 Like

Fair point, and maybe that’s where Atlases have their best (though still imperfect) use case – just as with internal iNat splits.

but the checklists include species not really there due to huge accuracy circles: https://forum.inaturalist.org/t/geoprivacy-obscuring-and-auto-obscure-discussion/457/31

That might be another case for populating atlases directly from something like GBIF, and considering those to be the curated source for range data in iNaturalist. Any conflicts with checklist data would be resolved in favor of the atlas data.

Also, while I don’t understand all the nuances of how large (in)accuracy circles interact with places and their checklists, something sticks in my mind that the behavior is somewhat different for iNat Standard places vs. community curated places. For the record, here I am only contemplating iNat Standard Places (and/or their checklists).

My example in the cited thread was for Placer County, CA, which I think is a standard place. As an amateur just trying to contribute as I can to the identification process, I think both the maps and the checklists should be based on more widely accepted ranges than the iNaturalist maps and checklists. Otherwise, errors are perpetuated.

1 Like


This is the potentially (definitely?) erroneous data point @sgene used as an example, which has been imported into GBIF: pink squares.

3 Likes