Taxonomic Swap 91524 (Tehty 16-04-2021)

A well defined species, not even very close to Coleataenia longifolia in characters.

Weakley (2020) - Flora of the Southea... (Viittaus)
Lisännyt edwinbridges huhtikuu 16, 2021 12:49 IP. | Tallentanut edwinbridges huhtikuu 16, 2021
korvattu seuraavalla:

Kommentit

@edwinbridges please don't curate in a direction away from POWO without first opening a flag and discussing why you propose we deviate from our reference for plants. POWO considers this a spp

Lähettänyt loarie melkein 3 vuotta sitten

If you insist on following the antiquated source POWO you are going to lose me and many other curators. It is piece of junk.

Lähettänyt edwinbridges melkein 3 vuotta sitten

I am very angry about this

Lähettänyt edwinbridges melkein 3 vuotta sitten

For all plants in the southeastern United States the latest authority is Weakley (2020), which is fully referenced and documented, and with a discussion of reasons for taxonomic and nomenclatural decisions given in every case where there is controversy. I will continue to follow it except for newer updated references on particular groups.

Lähettänyt edwinbridges melkein 3 vuotta sitten

@edwinbridges, I'd ask you to please keep conversations civil. No need to get angry. Looping in Alan (@whiteoak) who is onboard with iNaturalist's policies which is to follow POWO unless we explicitly deviate which requires (a) opening a flag on the taxon in question, (b) describing the deviation you are proposing with regards to how it maps to POWO, (c) waiting a week or so to let others comment on the deviation proposal, and then (d) making the taxon change and (e) the deviation

This policy is meant to (1) ensure curators know what direction to curate in by having deviations explicitly documented. Without a deviation indicating our intent, there's nothing to signal curators not to curate in the direction of POWO by re-lumping Coleataenia abscissa into Coleataenia longifolia. (2) make sure we have a globally consistent taxonomy that works across the globe and doesn't create situations where we have multiple incompatible taxa existing side by side (e.g. Coleataenia longifolia sensu lato Coleataenia longifolia sensu stricto)

Its not a big deal that you didn't do (a)-(c) but please try to do so in the future when you want to curate away from POWO (its easy to check by looking at the green 'match' or yellow 'deviation' decorations above). Also POWO isn't antiquated they update weekly and usually if you email them their really good about incorporating changes. It would definitely be preferable to invest time in getting POWO up to date with Weakly by corresponding with POWO than making and maintaining a very elaborate system of deviations. But making deviations is fine too as long as the community is on board, you've though through the global implications of the deviation, and the deviation is properly documented/mapped.

re: (b) Also - I call this a "swing split" when we carve off a ssp and elevate it as a sibling. One thing this could do is leave behind lingering IDs of Coleataenia longifolia that referred to the old sensu lato concept that are now interpreted as disagreements. Can you look for these and if <10 such mismatched identifications exist just leave a new ID with a comment and or mention signalling to the IDer that what we mean by Coleataenia longifolia has narrowed and encouraging them to replace their ID or looping in others to outvote their ID if necessary. If there are >10 such mismatches involved, sometimes its worth acompanying a taxon change like this with a split (Coleataenia longifolia s.l. -> Coleataenia longifolia s.s. + Coleataenia abscissa) which will replace existing IDs of Coleataenia longifolia with IDs of Coleataenia. this is disruptive but better than having lots of wrong IDs of Coleataenia longifolia that refer to a broader concept

Thanks for starting making the deviation (e) I made a few modifications:
https://www.inaturalist.org/taxon_framework_relationships/472753
the deviation should be a single mapping of all taxa involved on the iNat side and the POWO side that either have different meanings (e.g. Coleataenia longifolia) or different names (e.g. C. abscissa <-> C. longifolia abscissa). I also deleted the lingering 'match' on the now inactive Coleataenia longifolia abscissa 1234579 and added a link to the deviation to this taxon swap (ideally we'd link to the flag (a) where the discussion around the deviation is documented happened but this discussion after the fact here will do )

Thanks for your help trying to keep the iNaturalist plant taxonomy well curated and making sure everyone is on the same page about what taxonomy we're intending to follow and how it maps to POWO through explicit deviations

Lähettänyt loarie melkein 3 vuotta sitten

Hi @loarie and @edwinbridges -- I've been pondering the issue of POWO and its relevance as a "standard" for a while now, and have been frustrated with all the discrepancies relative to eastern North America, many of which I do regard as "antiquated" on the part of POWO. How to deal with this? Because of the numbers (and the time involved in the high-quality curation of a swap, especially a split), I've been reluctant to start committing taxon swaps on an ad hoc basis. But, the status quo is not really acceptable for many of the purposes of iNat.

Part of it comes down to the simple reality that a global reference curated by a couple of people a continent away (mainly, I understand, Rafael Govaerts) is always going to lag behind detailed local knowledge of the flora of a region. But I do also appreciate that iNaturalist has a difficult balancing act and can't follow 493 "standards" for vascular plants and 29 for mosses, etc., with different taxonomies for each country or state or region of a country. That creates a temptation to go lumpy (which is what POWO has done), which does create a "dumbing down" that creates tension against regional experts who recognize finer scale species (in some cases).

I like the idea of opening a dialogue with POWO and by that means ideally be able to deal with many discrepancies 1) in a more comprehensive and efficient way, and 2) at the root (as it were) of the "standard".

If that does not prove efficient and effective, then the a-e sequence does seem like an efficient, non-laborious path that creates a level of documentation that (One hopes) will not lead to a lot of flipflopping (Edwin changes Coleataenia longifolia abscissa to Coleataenia abscissa, someone a few months later notices that POWO has Coleataenia longifolia abscissa and swaps it back).

@bouteloua and I talked about this a year or more ago, and we ran a "crosswalk" between POWO and Weakley (2015), which she did some further analysis and assessment of -- but then I didn't follow through on the findings (I think she has been nibbling away at some of the needed changes that were apparent). Maybe we should rerun that with Weakley (2020) and look at the full extent and develop a plan to work through issues more comprehensively...

I'd really like to get these fixes done. The most vexing ones are where taxa are completely submerged in POWO and therefore observations cannot be attributed to them -- and quite a few of these are G1, G2, or G3 taxa with the conservation need to accrue observations. I can tolerate where ranks differ -- though this particular example is an egregious one. As Edwin mentions, no-one would mistake the two taxa, and very likely nearly all observations made to just Coleataenia longifolia refer to C. longifolia sensu stricto.

A related topic maybe worth discussing (Scott) is the AI algorithm IDing only to species rank. I get that this makes the AI easier, and gives the casual user what seems to be a positive ID (albeit at a degraded level of identification). This interacts to a degree (in my mind) with taxonomy, standards, "apparency" of ID characters, etc. But it tends to perpetuate the idea that infrataxa don't matter, creates a pile of IDs that aren't to the finest level of ID possible, and interacts with and exacerbates the frustration for Edwin of the wrongful reduction of the "abscissum entity" to infraspecific status.

Lähettänyt whiteoak melkein 3 vuotta sitten

Thanks very much Alan for all your comments.

For better or for worse the basic philosophy we're going with for vascular plants is: (1) use a global external reference and (2) explicitly map deviations from this reference
iNaturalist has enough complexity and incurred disruption from taxonomic changes (e.g. moving around observations, changing ancestries and the implications on IDs - not to mention community confusion) that I think the benefits of being able to explicitly map to a global reference outweigh the costs. But it would be good to discuss this further. Butterflies are a good example/comparison on the iNaturalist of an 'observose' group (like vascular plants) where we don't have/aren't using a global reference. Lets call this topic "pros-cons of using a global reference"

But assuming we are using a global reference like POWO, and there's an issue with it there are, as you say, two options:
a) create and maintain deviations from POWO
b) work with POWO to make the desired change

re: a - Deviations work well, like in this Coleataenia longifolia case when few taxa are involved and the mappings are simple. The mappings get unusably complex when lots of taxa are involved (e.g. often the case when genera are split and lots of epithets change and lots of taxa in the genus don't have 1-to-1 relationships). Also if there's just too many deviations making/maintaining them is just too much work. This is the case currently with Cactaceae where those on the site who want to use Lode vs whats on POWO that it doesn't seem worth the time building and maintaining so many deviations. Which leads too....

re: b - ...I think this approach sounds awesome in theory. It would be great to develop how to do this more efficiently. I don't think I'm the right person to lead trying to develop this. But I'd love to see it move forward and would love to help in a supporting role.

Alan - which do you think is more realistic for the issues with POWO you're most concerned with - a or b? Do you have any ideas how to improve either of these?

RE: places on the site where iNat doesn't give infraspecies as big a platform as species (e.g. the species counts and various species searches and the computer vision classes we train on where we roll infraspecies to species). I agree that this isn't perfect. There's advocates on both sides for and against infraspecies so to a certain extent we can't make everyone happy. From purely a data quality and curation standpoint though, things are pretty rough as is just trying to deal with species and these problems get a lot harder and blurrier if we were to try to offer this kind of functionality at the ssp level, but doesn't mean we couldn't get there. I think the former point is the more difficult one....

Lähettänyt loarie melkein 3 vuotta sitten

Lisää kommentti

Kirjaudu sisään tai Rekisteröidy lisätäksesi kommentteja