Disruptive taxon splits: a better system to manage the aftermath?

It was suggested that I consider bringing this issue up for discussion (to brainstorm potential improvements), as it’s not a glitch. The problem is cleaning up in the aftermath of a disruptive taxon split. Specifically, a situation such as this: when the first ID is incorrect, the 2nd ID bumps it up a taxonomic level with an explicit disagreement, and then a taxon split occurs which removes the explicit disagreement and the original incorrect ID is reinstated. It would be useful if there was an efficient way to get these back on track. Perhaps a modified Identify modal that can filter for these observations and allow one to mark them as “reviewed” a second time. Anything to make it easier to find what needs to be fixed and to keep track of the fixes as they’re made.

Here’s the original post: https://forum.inaturalist.org/t/inactive-taxon-taxon-split-glitch/9804

I’m not looking for an immediate fix, but as more and more observations accumulate, this issue will be magnified exponentially. Imagine 10 years into the future with thousands more observations that will all need to be, basically, re-identified from scratch following this sort of taxon split.

3 Likes

I wonder if it is possible honestly to just “edit” the existing IDs rather than post new ones following said splits. I’ve suggested having a “history” button before which could show taxon changes that affected it.

For instance, say a taxon is “swapped”. Instead of crossing out the current IDs and adding a new one with the “Added as part of a taxon swap” note, it would just automatically update the existing current ID and store that in the “history” of the ID. The “history” could be viewed with a button like the “compare” button that already occurs. Which would prevent the entire page being taken over in cases like this: https://www.inaturalist.org/observations/325405

2 Likes

That would be nice also–would look less messy. But my problem is going back and finding all the one’s that need to be re-identified. On my first round of IDs, I marked them “reviewed”, so now, after the taxon split, it’s next to impossible to find the one’s that need to be re-examined. It’s hard for me to wrap my mind around the issue and the solution, but a way to find observations (using the Identification modal) that were bumped up a taxonomic level and allowing one to mark them “re-reviewed” as one goes through the list of dozens or hundreds of observations might solve the problem.

2 Likes

Well in my opinion it is wrong that the explicit disagreement was removed in the taxon split, in the first place. … why do i think that is wrong?

If the disagreement is automatically preserved:
There is only one outcome: The taxon split results in a community taxon that still “needs ID”: This is not a critical error as identification with “needs ID” observations is an ongoing process.

Two things can happen if the disagreement is not automatically preserved:
1 The taxon split results in a community taxon that still “needs ID”: This is not a critical error as identification with “needs ID” observation is an ongoing process.
2 The removal of the disagreement during the taxon split results in a wrong species level Research Grade ID: This is a critical error as this produces wrong data!.

I see that in the Anemone example linked above, the automatic preservation of the disagreement might result in something, that was not totally intended by the person who disagreed. But if the disagreement is automatically removed, the result is quite the opposite of what the disagreeing person intended (as can be seen in the same Anemone example). In addition, in case the disagreement is not preserved the outcome can compromise data quality. While preserving the disagreement cant in any case compromise data quality. Therefore preserving the disagreement should be standard.

2 Likes

The explicit disagreement can’t be retained in the current system for the example he is referencing because the disagreement system doesn’t distinguish between specific taxon disagreements and entire branch disagreements.

If the disagreement is retained, then it says:
ID by pfau_tarleton: tribe Anemoneae (explicit disagreement with A. caroliniana), then that means he’s also disagreeing with Anemone (genus) – which definitely wouldn’t be what he intended. More info here https://forum.inaturalist.org/t/inactive-taxon-taxon-split-glitch/9804/8

Would be great to have more robust searching of one’s own identifications / observations so that they can be reviewed more easily. Something like review your IDs affected by this taxon change which brought you to a filtered Identify page so you could run through them quickly.

4 Likes

bouteloua I see that you are arguing process optimization … my argument is not about optimizing the process, but the outcome.

As you said: in the Anemone case mentioned above, retaining the disagreement has the following outcome: The disagreement is one taxonomic level to high, and the community taxon is one level higher than necessary. This is in my opinion a negligible error. As this is still a needs “ID observation”.

In other cases where there were already two agreeing species level IDs on an observation, and we have a single disagreeing ID not preserved in a taxon split, the outcome is the following: Wrong research grade observation. Data compromised! This is in my opinion not a negligible error.

Therefore preserving disagreement is much better for data quality than not preserving. … not preserving disagreement is better for data quantity, at the expense of quality. Which means that preserving or not preserving is not a technical, but a strategical question.