2021 Clements Update and Continual Concerns

loarie · October 3, 2021, 8:01am

Ok - thanks again to rjq et al. for helping clarify the splits again. I’ve now made draft splits for each. The next steps are to create atlases for each split so we can move as many IDs of the input as possible to the proper output (as opposed to coarsening them to the common ancestor of the outputs).

It would also be nice to update (rather than loose) the taxon ranges, and I find that this really helps create the atlases. However, I realize updating the taxon ranges takes some GIS skills, so if its not possible, just creating atlases still helps.

I’m going to describe the steps I took for split#3 - Anthus similis to update the ranges, create atlases, and then check that most IDs were being properly reassigned to an output taxon. If you can help do these steps to any of the other splits I pasted at the bottom many many thanks for your help!

(also, while I usually prefer using a split input as an output I didn’t do that here because it doesn’t lend itself as well to keeping changes in draft mode for a while and then committing everything at once which is the plan here)

(also, these are just the split, it ignores the ones where a split was not deemed necessary above, and I also went ahead and manually made/committed the Glaucidium brodiei change because its complex involving a species split and a genus split)

Step 1. Editing the ranges
To download a taxon range to edit it, navigate to the taxon range page by clicking on the “link” symbol next to “Range” in the “Overlays” menu

Then clock on the kml file to download it. If the iNat taxon doesn’t have a range and there is one on the IUCN Red List you can download it directly from IUCN. Also, its worth checking the IUCN Red List because they may have already split the taxon and if thats the case they might already have narrowed ranges for the output taxa available saving the step of manually creating them.

Open the downloaded range in your favorite GIS software - I use QGIS which is free. jwidness made a great tutorial for editing ranges you can find here. I usually make a new temporary shape to create a clipping mask outline (blue shape) and then first do a Vector->Geoprocessing Tools->Intersect with this shape and the original shape to create one narrowed range (purple) and then a Vector->Geoprocessing Tools->Difference on the same two shapes to create the other narrowed range (green). Then I save each out as a kml.

Next, edit each of the inactive output taxa and add the ranges you created. Make sure to indicate that they are not IUCN Red List ranges. One is a narrowed version of the IUCN Red List range (ie the IUCN Range map is unsuitable - namely because its too broad)

And the other is not in the IUCN Red List (ie its lumped into the broad taxon)

Making atlases
Making atlases is easier if the output taxa have ranges, but if you really don’t want to make ranges you can skip to this step. If a taxon is split on iNaturalist without atlases, all IDs on the input taxon will be replaced with IDs of the common ancestor of the output (e.g. the genus Anthus). This can result in a lot of coarsened data and can and should be minimized with atlases. We’ve made a handy tool on splits called “Analyze IDs” where you can see what the destination of IDs on the input taxon will be based on how the split is set up. Without any atlases, all IDs of the input (in this case 254) will be replaced by IDs of Anthus. With atlases we can reduce this.

To create/edit an atlas click on where it says “Not Atlased” next to one of the output taxa. Your atlas should be as simple as possible while still doing a good job of capturing the distribution.

Analyzing the fate of the Split input taxon’s Identifications
Once all the output taxa have atlases (the red “Not Atlased” will be replaced by green “Atlased” labels) rerun the “Analyze IDs tool”. If you’re atlases don’t overlap most of the IDs should have distinct output taxa as destinations. If you still have lots of IDs outside of atlases or in overlapping atlases, you can click on them to inspect them, alter the atlases accordingly, and rerun the tool. In this example, the 2 outside of atlas IDs are misIDs on observations far outside of range and coarsening them is not a problem

Once you’ve added atlases (and ideally taxon ranges) to the extent that the Analyze IDs tool thinks most IDs are going to distinct outputs, please cross the split off the list below. Once they are all ready, we’ll be able to commit them all at once. Thanks again for your help!

Topic		Replies	Views
How to request Taxon addition -- iNat lacking many hybrids General question , recommendations , taxonomy	5	858	March 22, 2020
2020 AOS Supplement: Yay or nay Curators	8	1007	July 19, 2020
Questions Curators and Curating General	14	1642	October 8, 2019
Plans for 2019 Clements bird taxonomy updates Curators	59	5835	September 17, 2021
iNaturalist Updates for February 2023 News and Updates web , android-app , ios-classic-app , seek , monthly-update	10	1283	May 2, 2023

2021 Clements Update and Continual Concerns

Related topics