
Photo from the first field site we sampled.
I have a new paper out (coauthored with John Stella) in Forest Ecology and Management, looking at beaver impacts on the landscapes surrounding their ponds. It’s a pretty cool paper, if I dare say so myself — we collected a huge amount of data from a massive study area and found both good evidence for the impacts beavers have on their surrounding environments and some cool hints as to how they pick what trees they’ll harvest for their dams and for their food stores. I might turn that into a post later on — for now, you can check out the paper (full text linked above) or my tweet thread on our basic findings:

What I want to talk about today is how this paper came to be, and all the human context that just can’t make it into a journal article about beavers. I think it’s worthwhile to understand the forces that shape what research gets done and what research gets shared, and take a moment to think about what that means for what we think we know about our world.
The Who
I cannot fully bring myself to believe that this whole project kicked off in 2017. In about two months, I will be three years older than I was when I started this one singular paper. At the time, I was a junior at the State University of New York College of Environmental Science and Forestry (ESF), very callously looking for a lab to work with to brush up my resume. This is something that always takes people by surprise — not only was I not particularly interested in beaver when I started this project, I had never actually seen one in real life! But I had taken a class with John and enjoyed it, and he had mentioned being open to undergrad researchers, so I sent him an email when the semester started up.
I’m not saying this is the way most research projects start, and I’m not saying this is a particularly good way for them to start — I’m still mortified by our kickoff meeting where John asked me what topics in his lab I was interested in for almost an hour while I just stammered. But at the end of the meeting we’d figured out that John had a student working on a project about beaver in the Adirondacks that I could contribute to.
And so we locked down the first thing every research project needs: someone to work on it. If no one wants to research something, no one will.
The How
Now that I was on the project, we were free to move onto the second need: a way to pay for it. Particularly with students (and especially with undergrads), a huge amount of the work involved in research is done via unpaid labor — I spent hundreds of hours on study design, background research, and planning out the field season in exchange for three college credits. But gas ain’t cheap, and we were planning a large field season that would require plenty of it — not to mention the ancillary costs I’d pick up from using my own car to get between our study sites and the opportunity cost of not having a full-time job or internship that summer.
So we got creative to try and find funding. The project morphed from an undergrad research project into an honors thesis, to try and unlock a line of funding there. I applied for fellowships and even got one. But mostly, we got lucky when John met a scientist from the DEC and pitched him on our project. This scientist — someone in the division of fish and wildlife — offered to create an internship for me to work through, paying me salary plus expenses to study how beaver create habitat for other species through their dams and foraging for trees.
Then that fell through, and the DEC stopped answering our emails — until John ran into a different DEC scientist, Jerry Carlson with the division of lands and forests, and pitched him a project studying how beaver change the forests around them through their dams and foraging for trees. Jerry agreed to fund the project and we were fully established for the field season — albeit with a different project than the one we had originally envisioned, having altered our original question to better match the interests of our funders.
This is the second thing you need for research: someone to pay the bills. If no one is willing to pay for what you’re interested in, well, it might just be time to be interested in something else.
The Where and The When
The field season itself was one of the more demanding things I’ve done so far. It’s been a while since I counted, but over 13 weeks I wound up driving over 2,000 miles, hiking over 250 miles, and measuring over 10,000 trees at 19 different sites across the entire Adirondacks. I’d work 60 hour weeks made up of 4 15 hour days, living out of my car and driving back to Syracuse for a shower on the weekend.

Distribution of field sites across the Adirondacks. Darker patches represent public land we could sample on.
Anyone who’s done field work knows how taxing it is. But you also get to see places almost no one else will ever go to. In June I borrowed a rowboat that was left at an empty campsite to cut across a pond I needed to get around, so that I could bushwhack a mile further into the woods.

In July I got to see one of the largest beaver dam failures I’ve ever seen, a mud patch maybe half a mile long that we started referring to as “beaver armageddon”.

That name was coined by my field assistant, Beth Newkirk, who was working with me for three weeks later on in the season. Beth was a fantastic tech, always fun to work with and thrilled to be out sampling.

Pulling together the field season — figuring out where we would sample, how to get there, where to sleep, and then actually doing it — was a massive undertaking. Even our little three month sampling trip, where I was the only full-time, full-season employee, took up about six months of human life to coordinate. This is the third necessity: to do research, you need the time to actually do the work.
The Why
The next month and a half were spent on data entry and analysis, then two months getting it into a passable thesis in December 2018. Then we all walked away for a while. I went and graduated, got a new job, got a better job, and completely forgot about the project. It wasn’t until John asked if I’d want to turn the thesis into a paper in October 2019 that I actually started working on the project again.
I mention this mostly because I want to talk about just how common it is for research to languish for years without any progress. Back in 2018, Andrew Gelman wound up publishing a paper that he had completed in 2012 and almost entirely forgotten about. One of my most-cited papers in this article was Raffel 2009, a study describing sampling that had happened almost a decade earlier in 2000. People get busy, other projects are more interesting or just moving faster, things get lost in the mix. But this means that there is a lot of publishable research out there that no one ever learns about, or at least takes a while to hit the collective conscious. A large amount of effort went into generating that knowledge that we might never build upon.
Our paper was ready to get sent around by March 2020, a full five months after we started editing what was already a completed thesis (though both John and I were extremely busy during this period, which likely made things move a little slower). We submitted to our first journal on March 6th and began the full peer review process.
Peer review is messy, slow, and byzantine, but most of that is irrelevant here. The thing I want to highlight is that after submitting to our first journal on March 6th and being highly responsive to editors and reviewers, our paper — which was rejected once without review and then accepted with only minor revisions — was finally available online July 24th. Just like before, to get through this process, you need a person who’s willing to lead the manuscript through to the finish line. You need someone who has the time to spend a full month exchanging emails with the managing editor of the journal going back and forth on image resolution requirements (the journal-specified 300 dpi still came through as blurry on our third round of proofs). Otherwise it won’t happen.
The What
So that’s this paper. Three years, five apartments, two states later, it’s time to close the doors on this project. But I want to talk about what this timeline means for what we understand about our world.
First, the only topics which get researched are the ones which have people willing to spend the time and the money on them. While this sometimes works to prioritize things we need to be working on — see for instance the massive influx of COVID-19 research during the global pandemic — if no one wants to put in the funding or the time to research important issues then they’ll go understudied.
The issues with funding are well-known and broadly speaking easy to understand. Research funding is generally a political process, at the whims of whatever group is in charge. And when that group has incentives other than increasing the sum of human knowledge, research funding is often one of the first things to get cut. This means that funding only gets doled out to projects that are politically popular, resulting in a huge gap in our understanding of sustainability science and other politicized issues. Similarly, issues more important to marginalized groups get sidelined in favor of studies that support whoever pays the bills.
But in ecology, the issues with time availability can possibly be even more important. We ran our study off a single field season, going out to sites that beaver had already established themselves at and measuring stems they’d cut over the years. But if we wanted to prove that it was beavers who were causing those impacts — run an experimental study, instead of an observational one — we’d have to release beavers into an unimpacted river (or ideally, multiple rivers), track them as they set up dams and colonies, and then spend years tracking what trees they cut where. It’d be an order of magnitude more effort — and require not only an order of magnitude more money, but time, too.
This is a common problem in ecology, particularly at larger spatial scales; ecological processes don’t map neatly to funding cycles or publishing pressures on researchers. If you’re curious how some long-lived organism will be impacted by something, you have to spend a long time monitoring it.
There’s ways around this. You can publish multiple papers on the same system over time — report your findings on impacts in a young forest, and then a middle aged forest, and so on. You can investigate every aspect of your system under the same treatment — this is a specialty of the Long Term Ecological Research Network, who will generally run a single experiment on a site for years (“How do these fields respond under increased temperatures and CO2, like we’d expect from climate change?”) and then do deep dives into every aspect of their broader question (“What species live in these fields? How fast do they grow? What species are more likely to survive?”). But any solution that still involves studying a system for a long period of time will require a lot of money to see to completion.
More commonly, ecologists just study questions that won’t require these multi-year efforts and million-dollar grants. If you study shorter-lived species or smaller systems, focus on flies or bacteria or plants you can fit in a lab garden, you can still generate knowledge without needing to dedicate nearly the same amount of time or money to the problem. You’re more free to alter your study to match what funders want to see than someone who’s locked into a project for ten or more years. Even our study — which was originally focusing on beavers as a way to create habitat diversity — got shifted to be something that our funders would be interested in.
But there’s a problem with designing small, discrete experiments that can pivot to be what funders want them to be. We all live in a beautiful, complex, wonderfully unfathomable natural world made up of overlapping systems which amplify and intensify each other, creating a whole that is much larger than the sum of its parts. And intensely studying the atoms of this larger universe doesn’t shed much light on the larger forces we individually interact with every day.
There’s a place for this sort of basic research; a lot of lab work can feed back into applied research and help us make better predictions about how systems will react to our management. But the combined pressures of limited funding and limited time means that some immediately important questions — such as how ecosystems will respond to coming global change, or how we can shore up our agriculture systems to prepare for imminent water shortages — have been broadly understudied. We aren’t quite flying blind into a terrifying future, but we certainly could have had better glasses on by now.
The easy fix to this problem would be increasing funding for these issues, so ecologists are able to establish more of these long term research projects and get a better understanding of how we might preserve some of the diversity of life we depend upon through the next century. At least in the USA, we’re yet to see it. NSF funding hasn’t increased since 2004.

And so for now, we continue onward, leaving these larger scale questions to a generation we hope will have more foresight than our own.
One last thing. Beth Newkirk, my assistant for the field season, passed in 2019. I just want to repeat how grateful I am to have known her and to have had her help on this project. Beth was a bundle of energy, absolutely hilarious, and fantastic to work with. She’d ID every bird she heard calling as we were tromping through swamps. Thank you, Beth. We miss you.