Interview with Dr. Michael Orr

 

Dr. Michael Orr is an assistant professor at the Institute of Zoology, Chinese Academy of Sciences in Beijing, China.

Hi Michael! I want to mostly dig into your recent paper on global bee distributions that was published in Current Biology. But first, can you tell us a bit about yourself? Where do you work and what is your day-to-day life as a bee researcher like?

Thanks for having me. Always happy to chat about bees. Right now I’m a President’s International Fellowship Initiative (PIFI) funded postdoctoral researcher working on bees in Beijing at the Institute of Zoology, Chinese Academy of Sciences. I did most of my bee work in North America, especially the southwestern deserts, but have been in China for about three and a half years now. It’s probably very different from what most people think of China as, and it’s more similar to life in the USA than one might expect. I live in a bit of a shoebox but it’s an expensive area, though still very cheap when compared to western megacities. I live about a kilometer from the institute so I just walk to work. As the institute is located near Olympic Park on the north side of Beijing there’s actually a decent bit of green space nearby and, though urban, it’s not too urban. Sometimes I’m able to escape to the park for collecting or peacetime missions to visit my pet Anthophora villosula (sister species to the better-known European Anthophora plumipes) nest aggregation under a very heavily trafficked bridge located over a busy thoroughfare. Happily, I have not yet been detained for loitering in a fenced-off area under that bridge.

It gets pretty busy when I’m in the office. Things happen fast in China and you always have to be ready for whatever might come your way. Recently, most of my days involve jumping from task to task trying to gradually push several projects toward completion. It’s kind of like that carnival game where you compete with others to get a horse or something to the end of a race by shooting water at it, except I have to shoot all of them at the same time and there’s no one else with a water gun. This might sound scarily hectic to some but it’s working out alright for me so far. There are downsides, a lot of the paperwork is of course in Chinese and hard to navigate for me, but people have always been helpful with that and it’s really something you have to expect when living in another country. There are a lot of short deadlines, though, and it’s not uncommon to be asked for something like two pages of writing due basically immediately to be able to apply for a grant (it actually happened last week). But the funding is very good and there are plenty of resources foreigners can apply for even as a postdoc. Practically, I’d rather have short deadlines on a grant I have a chance at than months to prepare for a grant with a 5-10% success rate. I certainly can’t complain. I’m happy to chat with anyone who’s thinking about making the move to China, we could definitely use more bee researchers here.

How did you end up in bee world, and what are the research questions you find most interesting?

I always knew I wanted to work on insects from an early age, so I majored in entomology for my undergraduate study at Cornell. I must admit that I was initially more interested in sociality research than in taxonomy and biodiversity, and I was much more interested in ants early on (don’t tell anyone). Funny enough, it was honey bees that initially got me interested in bees, via conversations and lectures from Tom Seeley and Nick Calderone. Come my junior year, it was time for me to settle down on a subject, having dabbled previously in spiders, katydids, and invasives, and I was lucky enough to join Bryan Danforth’s lab, working closely with Mia Park on bee biodiversity in apple orchards (a fortunate sidestep from honey bees, I now believe). That’s when I first got into specimen databasing and bee identification, along with some basic molecular work. From Cornell, I’ve worked in the labs of a number of other bee people, including Sean Brady, Sam Droege, Terry Griswold, and now Chao-Dong Zhu. I’ve worked on taxonomy and systematics, various bits of evolution, some behavior, recent dabbling in macroecology… the unifying theme behind all of this is the bees, and knowing more about them than most people, but I’ve also recently got more involved with conservation work in collaboration with Alice Hughes, and that kind of applied work is definitely very rewarding as well.

These days, the questions I’ve been asking in my research seem to change a lot depending on the day and the specific manuscript I’m working on. If I had to sum it up in a sentence, as the reader might prefer I do by now: I want to know why bees live where they do, how they do so, and what we can do to help them. More broadly, I’m also interested in how we can make taxonomy (the discovery and description of new species) sustainable in the current academic climate, and the proper usage of biodiversity data for understanding distribution patterns.


Dr. Michael Orr conducting fieldwork

Dr. Michael Orr conducting fieldwork


Can you tell us the story behind your recent study on global bee distributions? What motivated you to do the work and how did you assemble the team of researchers who co-authored the study?

There’s a lot to unpack there. I’ve always been very interested in the distribution of bees and drivers thereof, and I know many of the others on this project are too. John Ascher and I had been talking for years about the importance of basic distributional and life history data, and about how few others seemed to recognize the value of data generation and sharing. In general, beyond bees, there are just far too many studies out there that uncritically leverage public distributional data or “expert” range polygons and then perform “global” analyses in journals that start with Sci* or Nat*. The best-available-data arguments I’m sure authors would mount to defend themselves simply amount to BAD arguments. By working with lacking, incomplete data, people are obscuring the fact that there’s so much still unknown in the world. There is still so much we don’t know, and the big issue here is that the people most able to provide this information are those other researchers pay the least attention to, and when everyone is pretending we already know everything we need to then there’s also no reason to fund further knowledge generation. Collaboration is foundational in science but somehow everyone seems to have forgotten the people actually generating the data, so we wanted to do something big that would showcase what you can do when combining real knowledge with cutting-edge methods, to show what science should optimally be like. This is a bit of a lofty goal and I doubt we accomplished much on that track but I hope we made it at least halfway, and that it makes people think at least a few seconds more before they run their models on random, unvalidated data that they just downloaded. Because this kind of work has real consequences, the ranges and richness projections people make might be the only information governments have for prioritizing protected areas or other policies, so we owe it to the organisms we study that we at least do our due diligence and at least try to get it right rather than taking shortcuts for “global” studies in high-impact journals.

The actual study took about a year to put together, but that’s on the back of an immense amount of work that others really did to set the stage for this paper. First, there’s decades of work John did documenting bee species richness worldwide, country by country, in many places even state by state or province by province. What he’s done is something special, I simply cannot overstate how important that work was for this paper. At the same time, it would have also been impossible without the technical expertise (GIS wizardry) of Alice Hughes. I could not be more impressed or prouder of how good she is at what she does (which is a rather staggering number of things). Everyone contributed but this project simply could not have happened without those two, and certainly it would have never worked out without the support from Chao-Dong Zhu and his lab.

A lot of the initial planning for this paper was done remotely, but it really all came together when John and I discussed matters in person (the third time we’d met, I think it was) at the 2018 International Society of Hymenopterists meeting in Matsuyama, Japan (over sushi, bee collecting, but sadly not karaoke). Alice and I were totally in sync at this point, following many detailed discussions, but it was really hard to convey the whole plan through email or voice calls, so I’m not honestly sure if it would’ve moved past planning if John and I didn’t have that chat in the flower garden of Ehime University (we were frequently distracted by bees so maybe somewhere else would’ve actually been wiser, oh well). What matters is that it worked and that was when I really knew we had the team fully together, when it was really going to happen.

Anyways, as for the paper itself, there’s still a lot to unpack (most I’ll drop into the next question!).

Can you summarize some of the “big picture” findings of your study?

I’ve emphasized this before but it warrants repeating. I think that our biggest message might be just how lacking biodiversity data are at present. People focus on the patterns themselves, but these could still be greatly refined both at smaller scales and for individual species (rather than for species richness as a whole). Bees are one of the hottest topics in the invertebrate realm, but we still found 1700+ species in the wrong hemisphere (E v W) in public databases. In the countries investigated, up to ~30% of documented species were incorrect. This is incredible. If someone took these data uncritically and tried to map out richness, even super coarse resolution analyses (100x100km, etc.) couldn’t account for some of these errors. This shows that not only do we need more resources for identification and digitization, we need better data stewardship. The pipeline cannot end once new specimen records are online! Methods for correction, and incentives for people to do this are so critically important to getting things right. In an ideal world, taxonomists would be readily employed for their irreplaceable subject knowledge and they could spend some of their time systematically checking and correcting these kinds of data. GBIF and others are doing the best that they can but they can’t be expected to have experts on-hand for all taxa. They are using various top-down methods for error checking and flagging, but these approaches cannot always correct the data and make them usable, often expert knowledge is needed. If there were proper hiring and career incentives to provide that knowledge, this issue would disappear. For instance, it should mean something when someone says in a personal statement that they personally verified and digitized tens of thousands of specimen records, but right now I honestly don’t think it does. In part this is because it’s a complicated matter; how should data be valued? Certainly there’s a qualitative difference between 10,000 Eurasian tree sparrow records from northwestern Europe and 10,000 records of mixed birds from Papua. There are ways these things could be weighted and properly valued but I’ve yet to see real efforts at this. As soon as such a system rolls out, though, we’d have to worry about data ownership much more….

Anyways! For the other aspects of the study, I think that this is probably best done with bullet points so the audience doesn’t get impatient. Hopefully this makes it easier for people to cite because I know no one likes having to read past the abstract to cite papers these days.

  • Even in North America where >60% of records exist, sampling is very uneven and a small number of extremely well-sampled areas dominate the total.

  • There are more bee species in xeric and temperate areas than in tropical and other relatively humid places. This might be due to higher turnover in xeric areas, but this warrants further testing.

  • Global hotspots include the SW USA, Mediterranean Europe and northern Africa through into the Middle East, South Africa, the xeric parts of South America (especially Chile and Argentina), Central Asia, and Australia.

  • Bee species richness is strongly related to solar radiation, dryness, relatively low seasonality, relatively low wind, and a number of others in our final 18-variable model. But keep in mind that these are at the global scale! So this takes into account a lot of extreme values, and the optimums will often be different from the most extreme values such that something like low humidity is good but not the absolute lowest possible value. Some precipitation is still needed for plants, after all.

  • Plant species richness and net primary productivity were important for bees too, but only when you downweight areas with lots of tree cover; when not downweighting trees, net primary productivity was actually bad for bee richness.

  • There should still be MANY undescribed bees in China, Chile/Argentina, Australia (enough that our richness reprojections may not be accurate enough), and maybe north Africa but a lot of people have done work there historically so it is likely better known.


Of course, there’s a lot more than that in there, so please read the paper if this is all interesting to you.


The importance of taxonomy and identification is immeasurable. This is one of the biggest challenges for any inventory or monitoring initiative. If you have thousands of specimens, who will put the correct names on all of them?
— Dr. Michael Orr

Moving forward, what are the data that we need to better assess global distributions?

There’s still a lot of work to do. The most effective conservation work is typically done at local levels, by local stakeholders, and it takes an immense amount of data to properly plan fine-scale conservation management. (And there’s the obvious need to properly communicate out science and often incentivize conservation for stakeholders.) There are so many factors that can impact bees and we only have accurate data across continents and scales for a minority of these abiotic and biotic factors (aridity and other climate factors, plant resources, etc.). The data for bees themselves are, of course, even worse off, as we’ve shown in the paper. North America is well documented, though there’s always room for improvement, but Asia accounted for only 1% of total unique-locality records. Keep in mind that it took quite a bit of work by many different researchers in North America to bring us to this place, not just in terms of data accumulation but building up the taxonomic foundation on which all of that relies. Many other areas are not so lucky and still an immense amount of species remained undescribed or virtually unidentifiable because of lacking identification resources and reference material. This just makes things even more difficult. That’s not to say there aren’t undescribed species in North America, though, as there are quite a few in the more difficult groups that haven’t seen enough prior taxonomic work. For example, Zach Portman described nine new species of the Perdita subgenus Heteroperdita bringing it up to 22 species and I described seven new Anthophora (Micranthophora) of now 26 total. But, for comparison’s sake, Remko Leijs recently described 26 new Australian species for the subgenus Leioproctus (Colletellus), which previously had only one described species. There’s a lot of work to do everywhere, basically, and some places more than others.

This is why monitoring is so important. Right now, we need no-regret solutions that we can take to avoid the decline of bees and other vital pollinators. Some of these are easy and mesh well with other initiatives with bigger pockets (area protection, etc.), but some may conflict with the economy and there it gets more difficult. We know that we should reduce pesticide use and preserve natural areas, but by how much should we reduce pesticides and where, exactly, should we protect? We have to keep in mind that everything is controlled to some degree by economic considerations, that we will always need to find a balance between the conservation-ideal and the conservation-reality.

Right now, I think we should focus on defining a set of indicator species that are widespread enough to be compared across sites to better understand change over time; we cannot use every species because many things are too rare, but at the same time, in the USA where >60% of public unique bee locality-species records exist, we are at the point where we can use individual species rather than just total species numbers. Best case, these species can also be identified easily, which could even enable non-lethal sampling for abundance metrics perhaps with non-toxic paint marking to prevent recounts, but for many groups this will be impossible and that’s why we really do need taxonomists going forward. The importance of taxonomy and identification is immeasurable. This is one of the biggest challenges for any inventory or monitoring initiative. If you have thousands of specimens, who will put the correct names on all of them? Who will make sure all the specimens from all the sites are using the same species concepts, basically applying the same names to the same things? We can’t just bin things as “little black bees” or “big bees, maybe Xylocopa idk,” different bees have different life cycles and activity periods, as well as local habitat requirements and many other characteristics. As such, this set of representative pollinators should also encompass a diversity of life histories, so none go unaccounted for. Beyond this, we can still use total metrics such as richness, abundance, or diversity, but if we want to really dig into how species are impacted then we certainly want these types of representative species as well, to help us pull out across-species differences in reactions to anthropogenic pressures. Honestly, anyone can make a statistical test come up significant through various permutations, subsets, etc., and although we know it’s all not what you’re supposed to do, it still happens. And even when I or others call it out as a reviewer, sometimes editors ignore it. By relying on both metrics like richness and actual indicator species, we can get a much clearer view of what’s going on. Where we sample will also be critical, and in general rather than trying to sample everywhere I’m much more in favor of targeting specific ecoregions, habitat types, and maybe even in some cases microhabitats under varying management regimes from natural to agricultural to see how these different pressures impact bees in different areas. In doing so, using these various metrics, we’ll have a much better idea of what areas are most susceptible to anthropogenic impacts.

Why are bees of the genus Anthophora, and the subfamily Anthophorinae in general, so much better than bumble bees?

I’m glad that you asked that. I don’t think that there’s any one reason why Anthophora are so great, I’d actually say it’s more that they’re simply better in every possible way?

Should I regret inviting you to add extra questions?

Probably.

Dr. Michael Orr searching for bee nests

Dr. Michael Orr searching for bee nests

References for further reading:

If you’re interested in the publication this is based roughly on, the paper is available open access here:
Orr, M. C., Hughes, A. C., Chesters, D., Pickering, J., Zhu, C. D., & Ascher, J. S. (2021). Global patterns and drivers of bee distribution. Current Biology, 31(3), 451-458.

For more information on data shortfalls globally, both for point data from specimen records and expert range maps, you can look at this paper:
Hughes, A. C., Orr, M. C., Yang, Q., & Qiao, H. (2021). Effectively and accurately mapping global biodiversity patterns for different regions and taxa. Global Ecology and Biogeography.

Here’s also a very relevant preprint on sampling and accessibility biases in what point data do currently exist:
Hughes, A. C., Orr, M. C., Ma, K., Costello, M., Waller, J., Provoost, P., Zhu, C., & Qiao, H. Sampling biases shape our view of the natural world.

I’d be remiss not to end by advocating greater appreciation of taxonomy. If we didn’t have taxonomists, none of the above papers would have been possible. More information on the challenges taxonomy faces and the need for greater career recognition for generating vital biodiversity data can be found here:
Orr, M. C., Ascher, J. S., Bai, M., Chesters, D., & Zhu, C. D. (2020). Three questions: How can taxonomists survive and thrive worldwide? Megataxa, 1(1), 19-27.

Orr, M. C., Ferrari, R. R., Hughes, A. C., Chen, J., Ascher, J. S., Yan, Y. H., Williams, P. H., Zhou, X., Bai, M., Rudoy, A., Zhang, F., Ma, K.-P., & Zhu, C.-D. (2021). Taxonomy must engage with new technologies and evolve to face future challenges. Nature Ecology & Evolution, 5(1), 3-4.