Adam Marcus Headshot Hipporeads Samuel Fey is an ecologist broadly interested in how environmental variation influences the composition, structure, and functioning of interactive biological communities. He considers himself a community ecologist, but his research extends into physiological, population, and ecosystem ecology as well. Most of his research takes place in aquatic ecosystems and he enjoys working with lake and pond pelagic food webs. He incorporates field observations, short-term lab experiments, and mathematical models to address his research questions. Sam received his Ph.D. in Ecology and Evolutionary Biology from Dartmouth College in 2014, and is currently a postdoctoral fellow in the Department of Ecology and Evolutionary Biology at Yale University.


Interview by Wudan Yan

At a first glance, the following events may seem random, mysterious, and disconnected: a 75% decline in moose over North America in the span of a year, seabirds dropping dead along the Pacific Coast, and a massive coral die-off all throughout the Caribbean.

Samuel Fey, a postdoctoral fellow of Ecology and Evolutionary Biology at Yale University, thought so, too, until he and a group of researchers quantified animal deaths in the last 70 years documented in the scientific literature and found something striking: the frequency and magnitude of these animal deaths (referred to in the paper as “mass mortality events”) in this time frame have been on the rise. Furthermore, it’s possible that humans are playing a role in these animal deaths as well. These findings were published last December in PNAS.

Our science editor, Wudan Yan, spoke with Samuel about his involvement in this study, the challenges that came with it, and how we as a society can use this information as we continue to live in a highly anthropogenic world.

WY: How did you get involved in the PNAS paper? Was this a primary project for you, or something on the side?

SF: My research interests focus on how fluctuations in the environment affect populations and communities. Much of my work to date is on how temperature affects organisms and how small increases in temperature — or very rare but episodic increases in temperature events — affect the ecology of natural systems. This project fits in nicely because it accounted for factors that animals have to contend with daily and how the role of rare events on those factors further impacted the organisms. The project also contextualized the importance of temperature on ecological communities against the backdrop of many other stressors that animals face.

While the story fit in with my interests, it actually came out of conversations I had with Adam Siepielski, who was a postdoc while I was a graduate student at Dartmouth. Adam and I shared office space and — as scientists tend to do — we spent a ton of time discussing interesting science questions. There was a period of time where there were a few large animal mass mortality events reported in popular media.

WY: Which ones?

SF: I think there was a huge fish die-off that happened somewhere in the southern US, which got picked up by various media outlets. One day Adam came into work and was telling me about this story that he had heard on the radio and I was like, “That’s crazy. I had read an article a few weeks ago about these large bird die-offs that were very unexplained.”

We realized these events occur with some regularity, so while each individual event looks a little idiosyncratic when they’re reported by the media — these events are rarely put in any greater context. When these events get reported, they look strange and catastrophic and morbidly fascinating, but at the end of the story, the reporters kind of just say “…and this happened.” Whereas we said, “This is really strange, maybe someone in academia had done a pretty rigorous, quantitative analysis of these animal mass mortality events to see what mechanisms or patterns underlie them.” We were really surprised to see that nothing had been done across animal taxa on animal die-offs, so we thought this was a really good opportunity to try and basically answer a question we had.

WY: The paper talks about patterns in mass mortality events (MMEs) throughout animal taxa. How are MMEs related to extinctions?

SF: We spent some time coming to the definition and ultimately defined animal mass mortality events as a large proportion of a population dying off across all life stages relative to the generation time of the organism in question over a short time period. It’s a little bit verbose but there’s an important distinction between a extinction and a MME. An extinction is a global extirpation of a species, so there are none of those organisms left. In contrast, animal MMEs can be very large and kill up to 99% of a population, resulting in the death of billions of individuals and producing lots of dead biomass — in some cases, millions of tons. But the important distinctions here is that a species can still recover following an MME.

For instance, a really catastrophic disease can sweep through an area: one of the events in our database is a 1983 MME of sea urchins in the Caribbean. There was a waterborne pathogen, which caused 99% of these organisms in the Caribbean to die off. Although this pathogen almost completely led to the extirpation of sea urchins in this area, the sea urchins were able to recover following this event.

We can also think about the scale of death on a continuum. On one end of the spectrum, we have a low percentage of death that occurs at a varied background level. On the other end of the continuum is an extinction where all members of a species are eradicated. What our study focuses on is some grey area in between that continuum. As ecologists and evolutionary biologists, we have a reasonably good grasp on the ecological and evolutionary importance of these background-level mortality events that come from predation, resource limitation, or senescence. We have a pretty good understanding of extinctions and how they’ve changed through time, but what was lacking here — and what we tried to address in the paper — is an understanding of these mid-level mortality events.

WY: Do we have any idea how many MMEs have led to an extinction? I can imagine that it would be difficult to work with the data sets to go back in deep time and suss those things out.

SF: Everyone we’ve talked to about this story with is really interested to learn about what causes extinctions and if there is any foresight into predicting when they’re going to happen based on MMEs. I think that would be dream and very interesting science to do — that’s ten steps ahead of where we are right now. In the future, you can imagine such a case. One of the big points we make in this paper is that we need to improve how MMEs are monitored and how data is shared among those who collect information on such events. The papers in our database were from peer-reviewed academic journals. In the future, if we pooled together data collected over time, we can start asking: is there a relationship between large MMEs and extinctions? At this point, we don’t know with the data that we have.

WY: In the discussion of the PNAS paper, you note that the MMEs in the database may not be representative of what’s happening out in the world. Why do you think there’s insufficient documentation?

SF: We definitely only studied a subsets of the events that are happening since the data we used came from peer-reviewed journals. Many events are reported — in the media for instance — that don’t make their way into the scientific literature. If you work for the USGS, it may not be part of your job to document these types of events. I think the data should be collected in a way that can be shared and accessible to the scientific community; there doesn’t have to be a paper written about every event as long as there’s some way to verify the validity of the measures people are making in the field.

There’s also the idea that these are very rare and seemingly random events, so it’s possible that no one witnesses these things happening. Even if a MME happens in a remote part of the world, it could potentially go unnoticed. If there’s a large die-off of marine mammals and the carcasses sink to the ocean floor, they are not as likely to be observed than if there’s a mass stranding that leaves marine mammals dead on a beach in plain sight.

WY: That’s interesting because another point you discuss in the paper is that MMEs are occurring in organisms that seem to be a critical part of ecosystems. So wouldn’t there be a ripple effect eventually even in the events you may not be able to immediately see?

SF: Yes: MMEs have the potential to have cascading ecological, evolutionary implications. MMEs can produce millions of tons of dead biomass, so organisms that live off of eating dead biomass will profit. Alternatively, if an animal that is normally a predator undergoes a MME, that’s good news for the prey. These are some types of ripple effects that could propagate through food webs.

WY: We spoke about the insufficient documentation of MMEs as a challenge of the study. What other challenges were there?

SF: The three main things we looked in regard to MMEs was [1] how the frequency of these events have changed through time; [2] how the magnitude of these events have changed through time; and [3] how the causes of these events have changed through time.

In regards to frequency, we saw that the number of MMEs increased through time among all the taxa we looked at. But then the question becomes: is this is a true increase or is this an increase to publication bias? We know that there’s been more scientific productivity through time. So all things being equal, the likelihood of an event happening may stay the same, but the likelihood of it being reported increases because there are more scientists out there looking for these events. Any time you raise awareness of an issue in science, other people tend to look for it. This is a problem pervasive in a lot of academic fields: it eventually becomes a question of detecting an epidemic versus an epidemic of awareness.

WY: Were there any challenges or surprises that came about when you quantified the magnitude or cause of MMEs over time?

SF: We reported the causes as they are and were able to say that events caused by multiple stressors, biotoxicity, and disease have increased during the duration we looked at these events. At the same time, there’s potential for why these causes have increased in time. For example, there are advances in how disease are detected because of advancements in technology over time.

We were surprised to see that the magnitude of MMEs to increased over time. We thought that maybe the magnitude would decrease through time. If scientists look to find these events, they may pick up on the smaller events, so I thought this would create a bias in the magnitude to transmit over time. It was a bit of a headscratcher, then, to see increases in magnitude of these events for birds, fish, amphibians, and vertebrates.

WY: At the same time, one could argue that those increases in magnitude aren’t that surprising. The angle that I’m thinking about this is that the animals you just listed are more susceptible to environmental stressors (e.g.: toxins in the water or air).

SF: Many of these animals across all taxa do live in freshwater ecosystems, and freshwater ecosystems were the most commonly reported habitat in which an event occurred. One way to look at that information is that these environments are particularly susceptible to these kinds of catastrophic events. Since aquatic ecosystems occupy the lowest point of the landscape, it’s possible that effects in terrestrial ecosystems can cascade into them.

WY: What’s the follow-up to this study?

SF: It would be great to do a follow-up study with a larger database that represents not just peer-reviewed studies but also federal databases and a way to track down cases that have been reported in the media. Moving forward, we want scientists to do a more thorough job documenting the proportion of a population that dies during an event. That’s the kind of gold standard for reporting these MMEs because it enables us to compare how catastrophic an event is among taxa. For instance, 10,000 small fish dying vs. 100 monk seals dying differ in magnitude, but those respective MMEs can affect the populations differently.

WY: Towards the end of your paper, you discuss human activity influencing MMEs. I was wondering if you could weigh in on the ways you hope this paper can influence human action in the years to come.

SF: Human perturbation — such as source pollution, weapon testing, spills — was the second largest contributor of MMEs in our database. It’s also difficult to know how human activity might interact with events in our database but aren’t necessarily directly connected to human activity. For example, events caused by biotoxicity are naturally occurring (due to harmful algal blooms), but changes in how humans use land are also connected to biotoxicity. It’s difficult to say humans caused this event, but there’s a mechanistic understanding in what underlies algal blooms, so the overall contribution of humans is likely higher than what’s reported in the direct human perturbation category.

In terms of how human behavior and activities can change in the future, it’s hard to imagine anything that we’re doing now not changing in the future. The world has changed dramatically in the last 70 years of the study, and we hope that as we collect better data, it will become more apparent what the greatest cause of such events are, so we can better address and control those issues.

WY: There seems to be an increased conversation about the role of emotions in climate and environmental science. How does your research impact you at an emotional level?

SF: This is by far the most depressing project I’ve ever worked on. When you look at the figures of the paper and see data points on a graph, it’s easy to lose track that each one of those data points represents some catastrophic event that led to hundreds — sometimes even hundreds of thousands or millions — of animals dying in a short period of time. We were reading paper after paper about horrible things that animal populations have to contend with so it becomes very depressing after a while. Then, to compound all that, we borrowed from the field of the statistics of rare events, that has been developed to predict events such as school shootings, financial panics, and plane hijackings. There were many depressing elements behind this study.

WY: Is the emotional impact of your work discussed among your colleagues?

SF: It’s something we talk about in passing. Many of us got into this field because we have an admiration and appreciation of the organisms we work with, so it’s hard to hear and read about these catastrophic die-offs either due to natural or human-induced causes. The flip side is knowing that by doing work in an area that’s important, it’s uplifting to know that you have the potential to shed light on something that really needs attention called to it. The depressing nature of these individual events also underlie the importance of research in such areas.


Further Reading


Image credit: Thomas Quine via Flickr