New paper on insect monitoring and the BIML dataset
Zach Portman
Hi Beeple,
We have a new paper out in the Annals of the Entomological Society of America that addresses an important topic in bee monitoring — the need to preserve specimens. It’s an unfortunate fact that hundreds of thousands of publicly-available bee specimen records for the US are based on specimens that have been destroyed. This is because the USGS Bee Inventory and Monitoring Lab has destroyed most of their specimens, though they are still reported to places like GBIF as “preserved specimens.” We refer to records like this as “ghost records” and it seems like this may be an increasingly widespread practice, particularly with the decline of traditional museums and natural history collections.
As we explore in the paper, the destroyed specimens, particularly when combined with taxonomic changes, create a number of issues, particularly taxonomic artifacts that make it look like species are increasing or decreasing.
We focused on the BIML data because it is the single largest source of publicly-available recent bee data in the US and it is increasingly used by many scientists for monitoring and inventory purposes. In addition, Sam Droege had previously told me that they destroyed approximately 75% of specimens, which equates to hundreds of thousands of specimens. We invited Sam to peer review the manuscript in order to be as accurate as possible and to foster scientific dialogue on this issue, and Sam provided a peer review. Unfortunately, in his review, Sam denied all permission to cite him as a personal communication and we had to cut out a number of details about the extent of specimen destruction and dispersal. Since we could not get an accounting of which specimens have been destroyed or dispersed, right now the safe bet is to treat all records from BIML (or specimen records from any source that aren’t in a museum) as potentially unverifiable unless the physical specimen can be located.
Sam – it would be nice if you could provide an accounting of which specimens have been destroyed or dispersed in a reply to the article or here on this listserv.
At the end of the paper we include a number of recommendations for researchers and funders who are working with specimen-based data. I hope that this will shed light and stimulate some scientific dialogue on this important issue.
Here is a free link to the paper.
Zach
Hollis Woodard
This is incredibly poorly timed in terms of what I think it is trying to accomplish, but it is an excellent argument for why we need more investment in infrastructures like the USGS Bee Inventory and Monitoring Lab.
Now is a time that we should be advocating for our country's most important resources, infrastructures, and people that support bee research and conservation, such as the BIML and its employees. Greater support for the BIML is particularly timely given that it is currently under attack by the administration, which reflects the broader attack on science happening right now.
Just a note on the Darwin Core terminology: “preserved specimen” is accurately used by the BIML as it does not specify that a specimen is stored in perpetuity (I would hate for someone who read this post to think otherwise).
Hollis
Neil S Cobb
I am still trying to understand what the authors were trying to accomplish beyond a "gotcha" pub that borders on being captious. I think every section has either mediocre analyses or conclusions that are not supported by their results.
It is true that the type of records produced by BIML does not have an adequate category in GBIF, but it should be a simple request to have the records listed under a category such as "occurrences". I assume that at some point, Sam or someone at the USGS made an executive decision regarding how to allocate a very limited budget. There might be a few other people out there who have contributed as much to our knowledge of bee biodiversity, but they are few. Identifying some problems with a handful of taxa does not negate all the amazing, credible data BIML has provided on over 1,000 species.
One specific example where I was searching for merit is when they compare the three BIML datasets with 2 "control" species and conclude the two control species "demonstrate the pattern we would predict for most bee species: they have been recorded relatively consistently over time (though with a high degree of temporal and geographic variation) and their patterns of relative abundance show a general similarity across the 3 datasets (Fig. 2)." They contradict themselves in the same sentence, how can you claim "have been recorded relatively consistently over time" and then state "though with a high degree of temporal and geographic variation" . They provide no statistics and the difference in patterns does not look that different. And why would you use this data for monitoring purposes? Inventory purpose yes, and regardless the end user should screen data for the specific question(s) they are addressing.
The eight generic suggestions for the future are sanctimonious crap.
Need to workout, so much to complain about.......this pub needs a response to the editor.
Neil S. Cobb,
Biodiversity Outreach Network Office: 11 W Silver Spruce Ave, Flagstaff, AZ 86001-3541 ID 84-2609936
Mobile Office - Text & WhatsApp: 928-607-4075
Zoom Office: https://us02web.zoom.us/j/81630476460
ORCID https://orcid.org/0000-0002-6155-9444
--
You received this message because you are subscribed to the Google Groups "beemonitoring" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beemonitorin...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/beemonitoring/13e0ce0a-f8e7-452b-82bb-f179d4e204ddn%40googlegroups.com.
Brianne Du Clos
Zach Portman
Dear Hollis,
This is an issue that has been ongoing for many years, and while it’s unfortunate that it came out in the current political climate, I think we can both advocate for support for science while at the same time advocate for robust scientific practices. In the paper we advocate for funding museums and natural history collections as a solution to this problem.
In addition, the threat that the Bee Inventory and Monitoring Lab could be shut down means it’s especially important to properly account for the specimens to ensure this information is available for posterity.
Finally, the definition of a “preserved specimen” is an important point in the paper, since most scientists believe that a “preserved specimen” is a specimen permanently deposited in a collection or natural history museum (discussed in the 4th paragraph of the discussion). However, GBIF and BIML use a definition whereby a record is a “preserved specimen” if it was preserved when it was identified even if it was subsequently deliberately destroyed. In my experience most scientists are surprised to learn that this is the definition being used, I know I certainly was when I learned about it.
Zach
--
Neil S Cobb
Neil S. Cobb,
Biodiversity Outreach Network Office: 11 W Silver Spruce Ave, Flagstaff, AZ 86001-3541 ID 84-2609936
Mobile Office - Text & WhatsApp: 928-607-4075
Zoom Office: https://us02web.zoom.us/j/81630476460
ORCID https://orcid.org/0000-0002-6155-9444
To view this discussion visit https://groups.google.com/d/msgid/beemonitoring/5dc73d72-1618-4134-aeb0-fb35b5eced06n%40googlegroups.com.
Taliga, Chris - FPAC-NRCS, CO
Please see the Darwin Core Standards: https://dwc.tdwg.org/terms/
For referencing the correct ontology.
Christine Taliga
Plant and Ethno Ecologist
National Plant Data Team | Soil and Plant Science Division
Natural Resources Conservation Service
Denver Federal Center, Denver, CO 80225
| c: (303) 880-9681
From: beemon...@googlegroups.com <beemon...@googlegroups.com>
On Behalf Of Neil S Cobb
Sent: Tuesday, July 15, 2025 10:50 AM
To: bdu...@lumcon.edu
Cc: beemonitoring <beemon...@googlegroups.com>
Subject: [External Email]Re: [Beemonitoring] Re: New paper on insect monitoring and the BIML dataset
[External Email]
If this message comes from an unexpected sender or references a vague/unexpected topic;
Use caution before clicking links or opening attachments.
Please send any concerns or suspicious messages to:
Spam....@usda.gov
To view this discussion visit https://groups.google.com/d/msgid/beemonitoring/CACmApPj2p%3D6%2Bso908KN0jWO6ay4A553TsO3Nh1fhO8JjoXZZDA%40mail.gmail.com.
This electronic message contains information generated by the USDA solely for the intended recipients. Any unauthorized interception of this message or the use or disclosure of the information it contains may violate the law and subject the violator to civil or criminal penalties. If you believe you have received this message in error, please notify the sender and delete the email immediately.
Hollis Woodard
Hey Zach,
Thank you for giving more context. Sounds like our common ground is: 1. we need more investment and support for bee research and conservation infrastructures, especially right now; 2. we need to store specimens in appropriate ways when we can for the reasons outlined in your paper (but to repeat point 1, we need more investment to be able to do this in the best ways; and I’d add that we need to look for ways to minimize physical specimen collection - again, when we can); 3. we need to improve education around data standards so everyone is clearer on what terms do and do not mean.
I’ll be watching the group dialogue around your post carefully and if you (or anyone) feel that anything gets posted that does not follow the group guideline’s, please email me to let me know (hollis....@ucr.edu).
Everyone: please let’s keep this constructive as I think we mostly have; there are a lot of good points in everyone’s posts around this issue. You can refer to the group guidelines on the main google group page.
Hollis
La Gringa
Neil S Cobb
Neil S. Cobb,
Biodiversity Outreach Network Office: 11 W Silver Spruce Ave, Flagstaff, AZ 86001-3541 ID 84-2609936
Mobile Office - Text & WhatsApp: 928-607-4075
Zoom Office: https://us02web.zoom.us/j/81630476460
ORCID https://orcid.org/0000-0002-6155-9444
To view this discussion visit https://groups.google.com/d/msgid/beemonitoring/47e211d8-a9cd-444e-8999-d96700417874n%40googlegroups.com.
Woodrow, Woody
Sent: Tuesday, July 15, 2025 2:16 PM
To: beemonitoring <beemon...@googlegroups.com>
Subject: [EXTERNAL] Re: [External Email]Re: [Beemonitoring] Re: New paper on insect monitoring and the BIML dataset
|
This email has been received from outside of DOI - Use caution before clicking on links, opening attachments, or responding. |
Sam Droege
Ah, this is why we have a beemonitoring listserv to share results and discuss. And what a delightful topic filled with taxonomy, monitoring, philosophy, nuance, and some sharp edges to cause us to pay attention.
Context. The Bee Lab has been around since 2002 or so (shout out to Harold Ikerd, Laura Moore, and Kim Russell for being our first collaborators). We were investigating the nuances of bee bowls (still are) and Kim was doing a study of vegetation management under powerlines. That blossomed into many projects, and we have now labeled over 800k specimens and identified over 600k. We had no money for drawers or cabinets initially (really only this past year have we been able to actually curate all the specimens we still have using donated drawers and homemade drawer racks). Frugality led to keeping rare species (often at the Smithsonian), examples of each species, and largely giving away specimens to others for DNA, reference collections, education or recycling them for their pins. Not optimal from a data verification point of view as the paper points out, but useful from both a budgetary point of view and beneficial to the greater community of people learning about bees. We function much more like the museums did 100 years ago. They swapped specimens constantly and you can see the same collector names show up in multiple museums across the continent. For many smaller institutions and individuals (including many whose specimens we identified) specimens ultimately become dermestid food over time due to neglect, staff changes, absence of institutional curated collections and priorities.
The targets set in the paper for curation are exemplary, but impossible for most of us to meet giving our constraints. So it goes. For the Bee Lab, we certainly would spend more of our limited time and staff working with the database and implementing some of the ideas within the paper if we were a monitoring program, but we are not, never have been, and never claimed to be one (just so you know). We just have a lot of data and share that data and specimens freely as we continue to expand and curate that dataset. We depend on people to send us corrections, and they (the corrections) get incorporated.
So now. What is the BIML Database? The way to think of the database is not as a project or even a goal in and of itself but as a tool. Our database is there to capture and organize data from separate, independent, and diverse bee projects (over 100 of them) so they can be extracted for analysis or given away. The database contains data we collected, collaborators collected, and sometimes data from groups we are just helping out. We almost always do all the identifications or double check them. We use the data constantly, it gets uploaded to Discoverlife every 2 weeks or so and we update records almost every day. Every couple of years the data are pushed to GBIF, but you can always ask us for the latest version (the latest version is attached to this email). Note that many of the issues spoken about in the paper were actually resolved months ago in our database. Our database is a living thing in a way, constantly (though often very belatedly) getting its taxonomy retread and identifications changed when we find mistakes.
Want some examples of projects? Some of the categories of projects and individual datasets are listed as follows: Testing and comparisons of new traps, inventories of a national parks and refuges, data on timed floral use, scavenging bees from another group's malaise sample detritus, happenstance local collecting of bees in MidAtlantic region to understand the distribution and abundance of species, large scale projects across the U.S. with the National Park Service and USDA National Experimental Forests and Ranges, and on and on. Lots of moderate and sometimes very small student and ag projects in there too a crazy number of trapping techniques.
We almost always (can't think of any exceptions) try to pin up and identify all the specimens. We are grossly behind on some projects that are not a priority. The specimens are there, but some have been waiting for an id for decades. Most of the western species we have shipped off to other collections since our fragile state of existence indicates that we are not going to get to those specimens. Our extensive wasp bycatch largely goes to Matthias Buck in Alberta, and he generously sends dets back which we enter (eventually) into the database, and he then keeps the specimens for the Alberta collection.
In the old days we had aspirations to set up a national monitoring project like we have done for other species groups here at USGS but could never get enough money to make that happen (as the RCN can speak to, the statistics and sample sizes surrounding a national program are also tricky tricky, kudos to RCN for continuing to pursue that end).
Use of our dataset and all bee datasets on all the platforms comes with inherent problems. Data are collected for different reasons, collection techniques vary, effort varies, skills capturing and identifying vary. No dataset is immune. Analyses need take this into account. Our data has issues. Often, we wish we could find or had kept certain specimens so we can double check the identification. Even with specimens we retain there are times when we hold up our hands and say that we don't know what the species name is. You can see some of that information in our dataset with "?" behind a name, "near" in front of a name, "complexes", "species", and "groups" added to a name to acknowledge our uncertainty. Sometimes these are updated later, particularly when revisions call into question past identifications. Sometimes having the specimens would resolve things, sometimes not.
If you have the time and personnel to track all specimens then you could update everything but inevitably specimens wander off and can't be found or are on a one way loan trip etc. and you are then in the same boat as everyone else... can't find the specimen to resolve a name change. This is not a new problem. All the older bee literature has unresolvable groups in light of new revisions. In the world of vertebrates this is even worse. Trail's Flycatcher was split (in the 80s?) into Willow and Alder Flycatchers. Without adjustment the data change over time appears as if Traill's Flycatcher populations had crashed and Willow and Alder and increased. Any birder would know not to analyze their data without trimming the dates and that is what we would do way back in the day when I was running the North American Breeding Bird Survey (I have had bird roots since I was about 7). The solution with bees is the same and should be obvious, trim dates or lump species together if you want to analyze change for groups involved with taxonomic issues. Drop species from datasets if it is a hot complex mess.
Recently we have been writing up species accounts for our region, the continent, and involved with a number of assessments of rarity, status, decline, and vulnerability. In general, the bulk of the data out there in GBIF land and other repositories (thank you everyone for entering those data) are well conformed, distributions and dates line up. But for each species there are outliers. Most of the time we resolve those problems by inspecting the data. We rarely need to contact the museum for a double check. Why? Because the distributional problems are rare, resolvable (because we can adjust for known taxonomic issues), or can simply be left as question marks for future results that don't have much actual impact on decisions about status and rarity. It’s a judgment call process that improves the more you know about taxonomy and natural history.
Sometimes you do need to be conservative, for example when building lists for countries, counties, provinces etc. It pays for such a list to be conservative. We have done that with our list (as you can see by the attached file on species we have decided to leave off the list until more information comes in). A good recent example is the lovely publication of the Bees of Massachusetts. They (the authors) were going through our data and noticed some records for Lasioglossum floridanum from Cape Cod generated from a large project looking at bees in barrens and sand sites in the Northeast. Appropriate habitat but well north of East Coast records. Rightly so the authors contacted us for a double check. We could remember finding the specimens, thinking the records were very cool, and setting them aside because of their regional rarity. Well, it turns out that they we sent so far aside that we could not find them again for a double check so suggested they drop those records. A few weeks ago in our general effort to put everything to bed in the lab we found them. I double checked them and then sent them to Michael Veit for his collection and future addenda to the state list.
So, yes, one must use everyone's data with caution and an understanding of taxonomic changes. Important issues can often be resolved by contacting the owners of the data. Sometimes parts of the data must be curated prior to analysis for many of the reasons stated above and other times you may not want to use the dataset at all if it doesn't suit your needs. Our data could fall into any of those categories and we are fine if you want to use the data, modify the date, extract parts of the data, or don't use it at all. Mostly this email response is simply to give people some perspective
Coda: We are not shy about mentioning that we have specimens to give away and that many common species have their pins recycled. We don't keep track of how many are given away or destroyed because we are not a monitoring program or museum and don't see the need. We keep an extensive reference collection but otherwise every specimen will eventually leave. We had asked the authors to not use our personal communications not because we thought those statements were inaccurate or did not want that information published, we asked that they not be used because we felt they would be misrepresented.
sam
--
You received this message because you are subscribed to the Google Groups "beemonitoring" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beemonitorin...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/beemonitoring/CAMZwbbNevVJxCtEB-KR0M29w2DVzNRyezwVthZe87aczs6D1%2BA%40mail.gmail.com.
Zach Portman
Dear Sam and all,
Thank you for your thorough response.
I would like to apologize to you and the broader listserv for my original email coming off as overly aggressive and accusatory. I was a bit too overly focused on providing transparency to make up for what I see as a hole in the paper in terms of not knowing the disposition of the specimens.
Sam, I appreciate that you provided a review and that you have continued to participate in the conversation.
I do think there is a strong need for a broader conversation about monitoring bees, how this is best done, and especially what can be saved or not. For me, these collected bee specimens represent our biological legacy, and if they are not available to scientists and future generations to benefit from, that is a serious loss that I feel deeply.
I can assure you that our paper is more measured than my original email and I encourage people to read it. If people wish to write a response to it in the journal, I would encourage that.
Sincerely,
Zach
To view this discussion visit https://groups.google.com/d/msgid/beemonitoring/CADH2cDAngO-Z1pXpHWBpknTVM9Dr0RkakuxTaKDfR1gqGdSw4A%40mail.gmail.com.
Hollis Woodard
I just want to say that am incredibly moved by both of your posts and how willing you are to reflect and not just react. If only the world outside of the beemonitoring listserv could be more like that…
Tensions are so high right now as so many of the people, infrastructures, values, etc, that are completely intertwined with bee research, monitoring & conservation are under direct assault. Many of us are also mourning the recent loss of a beloved colleague, Sheila Colla. It is a rough time. I personally get a lot of strength from how committed everyone here is to doing good work in support of the bees. I also really like working with you all even if we disagree about things sometimes.
And now, with that, I will fade back into the shadows where I spend way more time quietly reading your productive conversations about bee monitoring rather than posting.
Also - welcome to the many new members who joined in the last few days.