I agree with some of Miriam Richards' comments. I have a few observations about Zach's Medium post, which are also about his past related Medium post and co-authored publication that recommended against using bowl traps in bee monitoring studies.
They mentioned in the earlier post that no authors had expressed disagreement with his recommendation against using bowl traps. At the time I read that, I realized that some and probably more than a few authors must disagree even if they didn't publish responding articles or replies. The recommendation against using bowl traps arguments in large part seemed to portray bowl traps as being used alone in studies, vs. many studies not only combine bowl trapping with netting, but also combine those with taking photographic records, visual observations, mark recapture, etc. Although, the posts did also make a few mentions of combined collection methods and how each method has a set of biases. They frequently mentioned bumblebee-specific monitoring, e.g. in the context of mark recapture, or collecting photographic records (e.g. uploaded to bugguide or inaturalist), yet it's well known that any Bombus-focused study shouldn't extensively use bowl traps, because bowls mostly catch different bees with small or medium body sizes. For that reason, I consider some of the arguments about Bombus studies to be separate from studies that focus on all bees or only on small halictid bees, and so not necessarily a refutation of the latter.
I agree that the described "taxonomic bottleneck" is real, although that seems to partly be a different domain of consideration, e.g. related more to practicality. Certainly, if authors weren't able to identify or store their Dialictus specimens that would be a problem. Although, many studies identify those specimens as far as they can but leave some at subgenus rank or use "c.f." It's been said that there are only a handful of Dialictus identification experts. So, it may be good advice that most authors including most beginners could be better off not dealing with that bottleneck, but I'd make an exception for studies conducted by Dialictus experts or by other authors who Dialictus experts identify for. Some of the arguments seemed to come close to saying bowl traps don't work to use in monitoring at all, and/or would only return biased results. It's well known that every existing collection method including non lethal methods have certain biases. Which seems to be part of the basis for many studies using combined methods, and for noting biases and limitations and trying to correct for them in their analyses. If anyone were claiming bowl traps can't tell us anything in monitoring, that seems insufficiently proven by examples/evidence, despite that some articles were cited.
They also seemed to argue that the entire community should take this view of turning away from bowl traps for monitoring. By contrast, it may be informative and useful for a diversity of different approaches to be used by different authors in studies, or at least for a small percent of authors who are informed of bowl biases and the taxonomic bottleneck to still use trapping, including to fully identifying the Dialictus specimens, etc. I agree that it may be best for the whole community to become better informed about these matters and to consider not using bowl traps, but at the same time it shouldn't be a problem if select experts who do identify their specimens do use bowls, to some extent, e.g. in combined collection method studies. Re: lethality considerations, these like the taxonomic bottleneck consideration seem to be in a different domain of consideration from the rest of the arguments, but are still worth (mostly separately) considering. It would seem that nearly everyone would agree with the recommendation to reduce the number of specimens collected where possible, to only collect the number of specimens necessary. One possible conflict is that some of the mentioned studies did intentionally use high sample sizes, yet regarded them as necessary or useful.
The arguments suggest that maybe some of those studies killed too many bees or even the declines those authors found. That concern is real, although is also very difficult to prove has occurred, so it may be ideal to emphasize that as more of a possible concern than a mistake certain authors definitely committed. Overall, I at least partly agree with most of their arguments. But, I regard such arguments to be more like optional recommendations of what methods would be ideal for more (but not all) authors and students to shift to use, and where part of the main argument is using practicality considerations (e.g. avoiding the taxonomic bottleneck). I also agree that minimizing or in some study contexts avoiding lethal collection is ideal where/if possible. I have another survey example that could be considered. I helped the 2017 to 2022 NYS pollinator survey (ESNPS), collecting specimens in the early years with bowls and netting (also as part of a thesis) but shifting to mostly only non lethal photographic records in the final years. The study was designed by bringing in multiple bee experts and had a large number of participants and occurrence records. The methods included bowl trapping and netting (unsure if if vane traps were used) and photographic records, and many participants chose to only take photographic records. That survey wasn't only a monitoring study, it also was based on enlarging the previously known state checklist of bee species, although monitoring was also part of their goals.
I didn't conduct their monitoring analyses, and like a few others I know wasn't offered authorship on that survey's final report. As one side note, when the report was published I noticed a few out-of-range species, which John Ascher and Joel Neylon agreed seem to be errors. Leaving those few species aside, I don't know of any publications that criticized the ESNPS monitoring results. The following link has links to their report and to additional appendix documents:
https://www.nynhp.org/projects/pollinators/. I'd be interested to learn whether Zach and his co-authors or others would consider that survey to have produced accurate results "despite" that it partly used bowls. In addition, I'd recommend giving a similar reading and review of other recently published or ongoing US bee surveys that used bowl traps as one of multiple collection methods.
Brian