COMPLEXITY: Physics of Life

Jennifer Dunne on Reconstructing Ancient Food Webs

Episode Notes

Looking back through time, the fossil record shows a remarkable diversity of forms, creatures unfamiliar to today’s Earth, suggesting ecosystems alien enough to challenge any sense of continuity. But reconstructed trophic networks — maps of who’s eating whom — reveal a hidden order that has been conserved since the first complex animals of half a billion years ago. These network models offer scientists an armature on which to hang new unifying theories of ecology, a way to answer questions about how energy moves through living systems, how evolution keeps producing creatures to refill specific niches, how mass extinctions happen, what minimal viable ecosystems are and why.  Untangling this deep structure of food webs may also shed light on technology and economics, and guide interventions to ensure sustainability in agriculture, conservation efforts, even venture capital investment.

This week’s guest is Jennifer Dunne, SFI’s Vice President for Science and Fellow at the Ecological Society of America. Dunne got her PhD in Energy and Resources from UC Berkeley, joined SFI’s faculty in 2007, and sits on the advisory board for Nautilus Magazine.  In this second half of a two-part conversation, we discuss her work on reconstructing ancient food webs, and the implications of this research for our understanding of ecologies, extinctions, sustainability, and technological innovation.

Visit our website for more information or to support our science and communication efforts.

Join our Facebook discussion group to meet like minds and talk about each episode.

Jennifer Dunne’s Website.

Related Reading:

Modern Lessons from Ancient Food Webs

Parasites Affect Food Web Structure Primarily through Increased Diversity and Complexity

Highly resolved early Eocene food webs show development of modern trophic structure after the end-Cretaceous extinction

The roles and impacts of human hunter-gatherers in North Pacific marine food webs

A primer on the history of food web ecology: Fundamental contributions of fourteen researchers

Quanta Magazine features Dunne on humans in food webs.

Jennifer on This Week in Science at InterPlanetary Festival 2019.

Learn more about The ArchaeoEcology Project.

Follow us on social media:

Twitter • YouTube • Facebook • Instagram • LinkedIn

Episode Transcription

Michael: So there is something really deep in general and profound here that I want to put a pin in and circle back around to because historian William Irwin Thompson who, associated with various SFI professors over the years through the Lindisfarne Association, was the one who introduced me to the idea that we destroy forest in order to create libraries, and that is in someway like a transubstantiation of the Amazon, you know, and so I'm even looking up in your office at this article on the dwindling web, and this question of, “Can we actually reconstruct this to show that a technological ecosystem is literally replacing biological trophic networks?”

Let's put a pin in that. In order to get into that depth, it's important to go into your more paleontological work and the conservation of the structure of these trophic networks over millions of years. Your research suggest that something really consistent and universal, like a thermodynamics kind of level is going on here, so, I'd love to hear you talk about the work that you did with the Smithsonian on the reconstruction of the Burgess Shale. There's a great line from the piece that you and Justin Yeakel wrote for American Scientist, saying, "Reconstructing the feeding interactions with the Burgess shale is no easy task, particularly with species with whose strange morphologies inspire names such as Anomalocaris and Hallucigenia." I mean, these are bizarre organisms how on earth did you actually put together a trophic network that you were confident enough about to actually draw these associations.

Jennifer: Right. Yeah, so that's really fun, fabulous work also, and I obviously over the last 10 or 15 years in my career, I've become obviously fascinated with deep time, both on human timescales and on geologic timescales and so there's actually a really good Santa Fe Institute story behind the Burgess shale work and the first paleo food we work that I was involved with. And I was at SFI as a young post doc in the early 2000s and I gave a talk on my early food web work, and some person I didn't know sat down next to me at lunch, he had been at my talk and without much preamble he said, "So do you think that we can do this for really ancient ecosystems?" and I said, "How old?" and he said, "You know, 500 million years, give or take." And I laughed and that was the start of the paleobiology food web work.

That person was Doug Erwin, who is a very well known paleobiologist at the Smithsonian, he's also the curator of the Burgess shale collection at the Smithsonian. And so, he and I basically just kind of brainstormed about could we do this kind of stuff for really old systems and how would we do that? What would it look like? The very questions you asked. So we brought together a group of paleobiologists we thought might be friendly to this idea and might potentially have the right kinds of data and some food web and had an ongoing working group. And one of the things we did as a group and this is sort of similar to what the first thing my humans in food web or humans in ecological networks group did. The humans in food webs or humans in ecological networks people had to figure out what are the main categories of use and interaction with other species?

What the paleo people had to figure out was what are the different line of inference that can be used to hypothesize a feeding link between taxa that are half a billion years old or older? Or you know are less for some of the data sets we were looking at. And like how do you know that are Anomalocaris is feeding on trilobites? You know, you can't watch it, can't see, can't observe it, can't do a feeding trial, it can't extract gut contents.

Michael: It's unscientific to just assume from the palps on that thing that it’s…

Jennifer: Well no, so that's not unscientific, that’s a piece of the science. So I like to joke that I locked all the paleobiologists into our medium conference room and I told them they wouldn't get any green chile or margaritas until they came up with a set of inferences so they did, we did. So basically there's a lot known about the morphology and even the philogyny that kind of evolutionary history or organisms and fossil assemblages. And in the case of certain fossil assemblages like the Burgess shale, they're called Lagerstätten, so they are very well preserved fossil assemblages' they're very unusual. So these are not normal but it's like a sudden event like the whole ecosystem got smothered or it all died all at once for different kinds of reasons.

Michael: Yeah, the Solnhofen limestone, Liaoning in China, like that kind of stuff.

Jennifer: Yeah, so and just some kind of event or confluence of events means that all of the species including the soft bodied bits get preserved very faithfully in the fossil record altogether. Most of the time you only get the hard shelly boney bits that get preserved but that's only a piece of the ecosystem but in the case of the Burgess Shale and some of the Cambrian shales and other throughout time, there have been this exquisite preservation from primary producers all the way up to predators and parasites and things. So you know species, you know what all the possible pairs are, you know, how do you decided who's actually eating whom. So, every feeding link is a hypothesis. In the fossil record every link is a hypothesis because you don't know it for sure, for sure, for sure, so it's a hypothesis but it's annotated with one or more lines of inference.

The best line of inference actually is gut contents. So there actually are fossilized gut contents, and the experts know how to distinguish gut contents from things that just died on top of each other, that actually can be distinguished.

Michael: Or pregnancy, which is another thing.

Jennifer: Right, yeah. But finding fossilized gut contents is not very frequent. And you occasionally, actually get things fossilized in the moment of feeding, like where a big fish had its mouth around a little fish and there's examples of that, yeah not very common, it's the smoking gun, but there's many other lines of evidence. So, the functional morphology of the organism tells you a lot about what it can and can't eat so that's where you get to the palps, or what are the teeth like? The teeth will tell you something about the kinds of things it could and could not eat. There's also bite marks, so a lot of things end up with bite marks that get preserved when they die.

Jennifer: So for example, one of the reasons AnomaloCARis, or AnomALAcaris, I never could get a paleobiologist to commit to one of those two pronunciations…

Michael: Well it's etymologically it's “CARis” right, but then you have like Archaeopteryx with the "P" pronounced or whatever, yeah.

Jennifer: Yeah, so yeah. Anomalocaris, Anomalocaris, anyway, they feed on trilobites in our food web, a big part of why we hypothesized...that's the royal we, hypothesized that they feed on trilobites because many trilobites have been found with tooth marks that match the Anomalocaris mouth parts and teeth.

Michael: So that's a pretty confident hypothesis?

Jennifer: Yeah. So many things are fairly confident so, but you may have... I mean and then there's the body size tells you something about what you can or can't eat or what can and cannot eat you, so there's like 10 or 12 of these kind of different...and we number them one through 12 or whatever. And then, so every link has one or more lines of evidence, you know, it's like, mouth part, body size, gut content assigned to it. And then what we do is we assign a certainty level to every link. How certain are we that this link actually happened? And we do that...and we have a very simple way of doing that. Certain things are high certainty and then of themselves, gut contents are high certainty.

But then if something only has one line of evidence, then we say that's low certainty. If it has two or more lines of evidence, or no, two lines of evidence then its medium certainty and then if three or more, gut contents, then it's high certainty. And that's important to do, I mean you have to some way of dealing with uncertainty and this is certainly not the only way you can deal with uncertainty in the fossil record and constructing a food web. There's other people who do it in a different kind of way using kind or more probabilistic statistics. They actually tend to work with more impoverished kinds of data than what we were working with, so.

In any case, out of this eventually pops a food web, or multiple food webs because we looked at a couple of systems, and then where every link is a hypothesis with one or more lines of inference and a certainty level assigned to it then we can start to do the questions of like, okay, the simplest question is, is this organized differently or similarly to modern food webs. And the role of this certainty assignment allows us to do a sensitivity analysis. So we can look at the structure of that food web with all of the links included but then we can systematically knock out links, we can knock out low certainty links, we can knock out random links and we can see how much does that change if at all our understanding of the structure? I mean eventually it will but you can actually monkey with it quite a bit before you change your final conclusion but it gives us a way to deal with the uncertainty which is really important when you're working with that kind of data.

And what we found basically, long story short is that these Cambrian food webs, the Burgess Shale and the Chengjiang Shale food webs are structured fundamentally similarly to modern food webs regardless of the habitat. So modern food webs regardless of habitat are structure similarly, and these more than a half a billion year old food webs are also structured similarly and it's nonrandom structure. And so that suggest and again you mentioned this at the beginning, that's in spite of the fact that many of the species in the Cambrian or in the Burgess shale, they don't have any modern descendants, they had really crazy body plans that don't exist anymore and they're all like these marine creatures and so a lot of them kind of were evolutionary dead ends, I mean that's not a great term but it just, you know, everything eventually becomes an evolutionary dead end.

But yeah, but often they're just body plans that were never seen again so they couldn't be more different in terms of multicellular organisms to modern creatures but yet they're still coming together in food webs in fundamentally similar kinds of ways. I mean that suggest to me that it's web structure's not something that evolved, it didn't evolve over time from the Cambrian to now, there's other things that are constraining it to be like that and we don't really know that yet, that's still an open question; like I'm sure it has something to do with thermodynamic constraints of how you distribute resources in a complex network, it could have something to do with either way the different kinds of necessary nutrients and chemicals balance across a whole system, I mean it could have something to do with many things and that's an open frontier of research frankly.

Michael: So you're not as confident as I am here in my arm chair to speculate that this has strong implications for astrobiological research, for example, that we're not necessarily going to observe the same kind of trophic networks in a methane-based ecosystem?

Jennifer: Oh I actually think we probably would.

Michael: Yeah?

Jennifer: Perhaps. I mean because I don't think it's about evolving a structure because I think it's about other kinds of constraints, it depends on the nature of those constraints, I mean I've often wanted to be able to construct really detailed food webs for like chemoautotrophic-based food, I haven't got that data, or I don't have access to it yet, maybe someone out there is putting that together.

Michael: You're working on hydrothermal vents and you're listening to this conversation! Who are you? Email us.

Jennifer: So, yeah, so I mean, you know my base hypothesis at this point because of the work I've done will always be that I expect the structure to be similar and if it deviates from that then there's going to be some interesting kind of things to unpack, to understand and explain what the differences are, so.

Michael: Well, so let's talk a little bit about this structure that's conserved.

Jennifer: Mmmhmm.

Michael: You mentioned again that there is a fixedness in the pattern of species interactions independent of species identity, habitat and time. You mentioned also in a piece that you wrote for the Layman et al. 2015 article that food webs, they've very similar to small-world networks, meaning short path links between all of the different nodes, but that they'll not scale-free.

Jennifer: No, I mean some of the features that we tend to see, or that we see in food webs, I already mentioned one actually in talking about humans, but that was long time ago. So you think of a food web as all the species that co-occur in a particular habitat, so over some kind of space in time you kind of integrate across that. You know if species never see each other, you know then they aren't really a part of the same food web, they don't have the opportunity to feed on each other. In any case, within the food webs that we've looked at from across time in space and habitat and everything else, one very strong pattern is that most things are fairly specialized in what they feed on, so they feed on a fairly small fraction of the species that are available to them in the food web.

But then you have this fairly long tail of the distribution where a few things are very highly generalized in what they feed on. And so, in the systems we've looked at so far for example, humans have always been highly generalized. You know, I don't think that'll necessarily always be the case for humans but so far that's what we've seen; and that pattern is not a scale-free pattern, it's not a power law, that you know power laws kind of exploded in popularity but were kind of over applied to everything for a long time, and the world's more interested and complicated than that. But in the case of food webs, they tend most cases they tend to be more of an exponential distribution so it's a skewed distribution still, instead of like the data following a line on the log, log plot like they would if it was a power law, they follow a line on a log linear plot. So skewed but not as highly skewed as a power law.

So, many, many specialists, a few generalists, and in an exponential kind of way. There are some food webs that skew more toward a uniform distribution and there's some that skew more towards a power law distribution but, you can do a scale collapse where you show data from a bunch of different food webs collapsing onto a universal curve, that's a nice little trick from statistical physics that's very handy for trying to find generalities or universalities across different kinds of data sets. But there's actually some more commonalities beyond that type of exponential type of distribution. They're a little hard to describe in simple language but I know short version is, there are some very simple models of network structure and in particular food web structure that basically they assumed that exponential type distribution but then they place some other constraints on how the nodes, how the species are linked together. And one of those, the niche model for food webs actually does a really good job of describing kind of various kinds of properties of empirical food webs.

And, we use that model actually as a way to compare across food webs, because you have to take into account network structure changes systematically with the size of the network and also with how many links are in the network. And this is actually a really, really important point because you'll see or hear too often someone comparing a small food web and its properties, like how many omnivores are there? Or how many top predators are there, or what's the average trophic level, and then they'll compare this small food web to this much bigger more detailed food web. And they'll go "Look, see they have different average trophic levels. Or they have different percentages of omnivores, so they're different.” But actually the big one is just the big version of the small one once you scale up, that's actually you would expect it to have a greater percentage of omnivores or a higher main trophic level.

So, this using kind of null models or models like the niche model, which is a little bit more than a null model, that allows us to understand how food webs of different sizes, how they compare to each other, and do it in a more rigorous way. And this is something that I had a lot of conversations about with my network theory and statistical physics colleagues when I first came to SFI, and who continue to acknowledge that this is indeed a very hard problem. How you rigorously and systematically compare the structure of different-sized networks and it's something that I grapple with all the time in food web research, so.

Michael: One of the things that comes up in this comparative analysis of trophic networks is that food webs with lower connectivity may be more sensitive to species extinctions. And you've published numerous papers on looking at networks as they are thinned out before a mass extinction, thinning the network and making it sort of more vulnerable to perturbances.

Jennifer: Right, yeah. Yeah, so some of my work and some of my very earliest work and my most well cited work is actually about using food webs to look at the robustness of ecological networks or ecosystems to species loss, but doing it from the lens of species in an ecosystem are interconnected through their interactions, feeding interaction being a very important one. And as you lose species, you're losing all the links to them also and all the links from them to other species, so it's not just them, their embedded with a set of feeding interactions and those feeding interactions are embedded within a broader network of feeding interactions. So, what I've done and colleagues of mine have done and I mean a fairly simple thing but it's just to do little simulations because you can't really do it in the real world. You can do it in really small microcosms but you can't go out into the real world and just pull a species of out it and make it go extinct.

Michael: Well, maybe researchers in a hundred years will be able to look at our middens and…

Jennifer: Well yeah, you can kind of look back and try to recreate the extinction sequence and people are starting to do some of that. But yeah, you can start plucking species out, you can do it different ways, you can do them randomly, you can pick the very highly-connected species, the hubs, you can pick low-connected species. But you start thinning the food web out and you're basically asking a question, you'll get a secondary extinction just looking at network structure if something loses all of its potential prey items; and so that's when you get a secondary extinction. And then if something depends on that then if that goes extinct then you can get a cascading extinction.

And one of the things we found is, and this is no big surprise now, but a densely connected network is going to be more robust. I mean it's got more redundancy, it's got more pathways, so things seem to have more than one option or more than two options; and so it takes a while before you knock out all of the prey of any particular species. But as you thin and thin webs of species and the links attached to them, they come less and less robust to further disruption. In the simulations we did using really data sets you could actually knock out about 20% of the species before you got significant secondary extinctions but then once you start getting secondary extinctions it proceeds very rapidly, you start getting cascading extinctions.

So you go from like, you can lose species, lose species, lose species, and then all of a sudden you really start getting these cascades. And I was involved in a study of the Adriatic Sea and different kinds of ecological indicators through time for the Adriatic Sea region, so this was a human-impacts-on-ecosystem kind of paper that I did not...I was a minor co-author on but there was a food web and robustness part of it. So they looked at Adriatic Sea food web across tens of thousands of years of human habitation and basically it's not until the last 200 years that you really see relative abundance of many species going way down and other kind of environmental indicators becoming worse.

And, they also put together simple food webs and showed a big change you don't see the food web fundamentally changing much until 200 years ago. And then it gets reduced quite a bit and you can do...another thing you can do with robustness studies is you can take one food web and like in the earlier state of the Adriatic Sea and show this is how robust it is to sequential extinctions. And then you can do that same thing for the reduced web of the present day and you can show there's huge reduction in robustness, in potential robustness of the two systems.

And Justin Yeakel's work in the Egyptian system also kind of looked at this contraction of a food web or in his case a piece of a food web. So the medium- and large-bodied mammals, the carnivores and herbivores of ancient Egypt over the last 13,000 years since the late Pleistocene. And there you have a huge contraction of that piece of the food web just like in the Adriatic Sea, you'll have a significant contraction of the whole food web. Yeah.

Michael: I just want to suggest that you go into a little bit about the methodology of that one. Because that's super fascinating.

Jennifer: Yes, it is really cool and I'll get to that in a second. But I mean, the kind of punch line and then I'll go back to the Egyptian study, because it is super cool, is that they did a different kind of analysis on their contracting food webs through time, they did what's called a stability analysis. So we did this robustness analysis on our contracting food web. They did, a stability analysis on their contracting food web. We both came up with the same answer lately which is like, as you contract these food webs and you impoverish them, they become less and less stable and resilient and robust to further change. Not a huge surprise, but important to show.

So, the Egyptian study, I wish I had been involved in that study but Justin Yeakel, former SFI postdoc who is now a professor at UC Merced, and he and I wrote a less technical kind of review of old food web research.

Michael: Can we link to that in the show notes?

Jennifer: Yeah, absolutely, that's a good thing for people to take a look at, it's a nice review. It's already kind of out of date and you know things are moving so quickly, but everything that's in it is accurate up until a couple of years ago. Anyway, Justin led this study of a piece of a food web of ancient Egypt. So, I mean, he looked at Egypt through the ages since the late Pleistocene, through kind of the classic ancient Egypt time of the pyramids and everything else, up until the present day. And so, what they did which is really hard to do across that 13,000 year period and I forget how many time slices but quite a few; they were able to reconstruct this piece of the food web with high fidelity, so the mammal predators and the mammal herbivores that they fed on.

And, so at the end of the late Pleistocene before humans were on the scene but they were kind of low level hunter gatherers, they weren't having huge impact yet or developing everything; I think there were like 38 species in that little teeny piece of a food web. And you see the snapshots through time and it contracts and contracts and contracts until you only have eight species left in today's food web, I mean just a massive contraction. But, how did they create these snapshots of the food web? Well, over the course of ancient Egypt, there's a ton of art, a lot of it funerary art but a variety of different...not just funerary art but a lot of it was, some of it was makeup palettes and other things.

They just did a lot of artwork and they depicted a lot of the animals that were present during that whatever particular period of time was on the artwork. So there's this Heirakonpolis palette, which depicts a bunch of these mammals, both predators and prey. It also has some fantastic animals that aren't actual real animals.

Michael: The snake neck jaguar, that kind of thing?

Jennifer: Right, yeah. So, you don't include those in the food web but most of the animals we know existed from other records also. And so, they were able to pull all of this interesting archaeological data off of this artwork at different points in time and basically what you're seeing is fewer and fewer species represented in the artwork over thousand of years because a lot of them just went extinct for various reasons, you know, climate change is happening. There was a wet period for several thousand years after the late Pleistocene and it went into a very dry period and there were three desertification events and you can actually see the predator prey ratios through time for these little food webs change significantly in idiosyncratic ways at the points at which you get super droughts.

You have preferential extinctions of the herbivores over time, you can see this in the change in the predator-prey ratio over time and then they also, like I said, they did the stability analysis for the food webs in each of the different time slices showing that they were becoming less and less stable through time and now you've ended up with the incredibly impoverished food web that's really at risk for even further reduction.

Michael: Yeah, you know, listening to this, you mentioned earlier that there's this possibility of a negative correlation in the richness of cosmology and the richness of the ecology and again, I wonder...I've read somewhere someone making the argument that as we lose species of living organisms, that there's an extended notion of the species, the way that we might think of a species of mineral, and that the Anthropocene is defined by an unprecedented proliferation in the kinds of things on our planet. You know in the number of new mineral species, the number of new configurations of matter; that we're going through what in one sense is a sixth mass extinction and in another sense would appear to be an enormous evolutionary radiation, but one that requires us to expand our definition of what qualifies in that picture, right.

So, if we can just look at this somewhat less cynically than we are used to, the question opens up into the applicability of these trophic network models to a study to evoke W. Brian Arthur in this, and Kevin Kelly and other people that have writing meaningful interesting works about technology as a form of life…do you think that we can use these trophic network models do predict economic and technological niches? And do you think that we may be able to use this to guide...I mean, I hate thinking about, like robot pollinator bees, like that’s just such a horrible scenario to contemplate.

Jennifer: Techno bees.

Michael: Yeah. But do you think that we are sort of being driven into a thermodynamic outcome here? And that we can use this in a predictive way, not just to look back and understand this deep generality but to look forward and actually anticipate and move into positions that we can then use to proactively keep these networks in place and allow them to continue to proliferate?

Jennifer: Right. [Laughs.]

Michael: I mean, what a mess. Feel free to jump off into the speculation here.

Jennifer: No, as a good scientist I don't love speculating in really giant ways so, I like to stay closer to home. I mean some of the things that we want to do with the ArchaeoEcology Project and thinking about human use networks, and this is a very early stage project but we are doing it with the notion of thinking about sustainability of socioecological environmental systems in mind. And whenever I talk about this stuff I usually get asked can we do this for modern systems? Would it be useful? What would it tell us? And yeah, ultimately...certainty I would like to pull it up into thinking about...using it as a different lens or a different framework for thinking about how humans interact with their ecosystems and environments in the modern world and using it to think...I mean you know even like the ideas of food deserts or people who have access to really a rich diversity of foods and the impoverishment of biodiversity by agriculture, because they only pick one particular species, one type of potato and there's like a hundred of them in Peru, which I got to see, which is really cool.

I mean in terms of also thinking about the relationship between technology and innovation and resources of different kinds, I mean I'm coming at it through this biodiversity ecosystem lens. That's not to say you couldn't apply it to non-biotic resources and we've talked a little bit about how to incorporate non-biotic resources into how we characterize these networks. We're trying to get this right for simple systems first. I mean I think in terms of like, one of the things that we can ask of the prior systems and then use to think about some of the questions you brought up for current and future systems is like, what is the possible landscape of ways that people could create technologies or interact with other species given the kind of opportunities they had in any given system? And what made them explore or exploit a large part of that landscape versus only exploit or explore a small piece of it?

And if they're only exploiting or a exploring fairly small piece of it, why? And what would it mean to look beyond that? And so, yeah, I think there could be some kernels in the kind of thing that we're doing if you try to think of the whole space of ways to put together different resources to create new technologies. Of course, you know, this know, I'm not a huge techno-optimist, I mean humans are very good at creating technologies; we’re very bad understanding the outcomes and the negative repercussions of our technologies, we've proven that time and time again and I think we'll continue proving that. But I think that's also where our network perspective can be very handy, it's like, okay as you develop this technology let’s try to really embed with within the socioecological, environmental network and try to understand what the ramifications are.

Humans are all too good at thinking like, "Oh, I'm going to make this better for humans by introducing this thing." And, I mean you know, we have a long history or introducing species both intentionally and unintentionally, our initial introductions often go badly awry; cane toads and…

Michael: Beavers.

Jennifer: I mean, yeah. I mean beavers are great ecosystem engineers, I mean…

Michael: They've completely transformed parts of South America, but it’s bizarre.

Jennifer: Yeah, absolutely, but they also have their indigenous habitat where they enhance hugely the biodiversity of many montane systems. You know, but yeah, introducing like fur-bearing animals into some place where they didn't use to exist — and ones they happen to be really fabulous ecosystem engineers the way that humans are — yeah, that's going to cause problems.

Michael: Well, here is a sort of more focused or specific questions when it gets into the application or the translation of this science into street-level human activity, which is, it seems like you could use this as a lens to identify where there is a missing company in a particular economic system that you can say, Oh, that these people in this particular region are under-served through some missing piece of infrastructure, and that there's sort of deep biomimicry that we could use to start identifying how we've allowed this more piecemeal approach to technological innovation to bring things out of balance and how we could bring them back into balance.

Jennifer: Yeah, I think many things are possible. But I mean I always kind of come back to why I got into food webs in the first place which is wanting to grapple with the whole system, but wanting to grapple with it in a quantitative kind of, simplified but quantitative way. Humans are very good at being reductionists and looking at little pieces of things and really kind of drilling down into that little piece, I mean it underlies a lot of our huge successes of physics and associated engineering in the 20th Century. That's great to a point but it can also end up leading you astray and we see this time and again in an ecological context. And so, for me yeah it's always that, what is the whole system? And what piece of it are you occupying and why? And if you want to alter it by taking something out or putting something in, what are the likely implications of that given a whole system look versus like this little teeny piecemeal look?

And that sounds all very ethereal, but I'll give one ecological example. There was an eminent theoretical ecologist name Peter Yodzis who died of ALS a few years ago. He did some really cool food web modeling to kind of get at this point. So there's a number of systems where human fishers feel that they compete with seals or sea lions for fish and it still happens in various places around the world where there is a big cry to extirpate or exterminate the seals or sea lions because hey, they're eating our fish. And so if you think about it from a food web perspective here’s, humans are one node, seal or sea lions are another node, they both have links to some fish species of interest to both, and so that's like the simplest possible local little look at a little piece of a food web.

And so human thinking goes, "Oh, we just get rid of the seal or the sea lion from this thing and all of a sudden that allows the population of the fish to grow because one their predators is gone, so they grow and then humans get more fish. Yay.” Great, however Peter Yodzis took this thing on and he basically said, "Let's put this into a food web" and he did this actually for a fishery called The Benguela Fishery off of South Africa. And so he took real data and he did some dynamical modeling, some fairly simple dynamical modeling and he did this thing where he didn't include humans exclusively, but he had a little food web and seals or in that case sea lions were a part of it, so in these little toy models and simulations he would pull those out completely or he would do a press perturbation where he would reduce the numbers of the sea lions and then he would run the population dynamics on this food web network and there we like three different fish species that the sea lions fed on that were also of interest to humans that humans want more of.

So, pull the sea lions out or drastically reduce their numbers and, this is a probabilistic study, about half the time those fish species, their population would increase. But guess what? Half the time they went down because of all the indirect interactions, because it's not just the sea lions and the fish and the humans. There’s all of these other players and there's all these other pathways and all these loops and all these indirect effects. And because of all the complexity in the network, it was definitely not a sure thing that removing this other predator that is a competitor to humans was going to actually have the effect that humans want. And a lot of the times, it actually screws it up for humans and gives them fewer fish. So, that to me is just...he wrote a series of papers a while back where he kind of looked at this issue and I think it was very illuminating and really important.

Michael: That is interesting. So, I'd like to wrap this up, you've given me an enormous amount of your time, I appreciate it.

Jennifer: And I'm sure you've got a lot of material, you're going to have to weed through.

Michael: Yeah but it'll be fine. I'd like to tie a bow on this by starting with a right angle, right, it’s how you have to tie a bow. And, I guess the question is, you've got two kinds of people listening to this show and some of them are science-track and some of them aren't, so what do you imagine to be the take away for people who are not pursuing scientific research? What do you hope that people carry out of this?

Jennifer: Yeah, I mean I think what I hope people carry out of this is that we can use these approaches, these scientific approaches drawn from ecology and physics and archeology and that it’s really important and really useful to study humans as a part of their ecosystems and that we're just starting to really dive into a lot of detail about the variety of ways that humans are interacting with other species and how they're using technology to access resources and what that means for their interactions with other species. And what we're hoping and what we're staring to see is that, that's providing us a new way to think about sustainability, of coupled natural human systems, it really underlines how important it is to realize that human systems are not separate from environmental systems, they’re not separate from ecological systems, and that we gain a lot by studying them in an integrated transdisciplinary way.

And the other thing just to underline is that we hear too often the doom and gloom of “humans are having negative impacts on the world in different kinds of ways and at different scales” and that's certainty going on but we also have very beautiful examples of humans as just kind of fitting into, and in some cases being really critical towards the kind of diversity and stabilization of, ecosystems. And there's very interesting lessons we can learn from that, that can hopefully help us to think about current sustainability and future sustainability.

Michael: Awesome. The last question I have for you is a little bit more personal. I mean we've heard in this your story of being a kid playing in the creek to being the VP for Science at SFI and so you have an excellent position from which you can advise or council young people who are interested in pursuing a career in science. What advice would you give someone in 2019 who is interested in pursuing this life path?

Jennifer: So I actually want to speak to the kids who think they're not interested in pursuing a career in science, because that's what I was, actually. I mean although I had my roots in being a little kid naturalist or whatever, I thought ultimately I didn't want to be a scientist. I mean I think the kids who want to be scientists will find ways to be scientists, which is great. To them I would say, avoid the silos. I mean the future of science and I think the most interesting and important questions are really the crossroads of traditional disciplines, and the fun part of science is working with people who have expertise that's totally different from yours.

I mean, when I was a kid I wanted to be an archaeologist at some point, I wanted to be a paleobiologist at some point. I didn't become those things but now I grew up and I play with those people, I work with those people, I play with those people. So that's for the kids who already think they're on a science track, which is great. For the kids who think they aren't on a science track, or the people who think they aren't on a science track, I would say, science is way more fun and cool then you can even imagine and it's a living breathing kind of creating new understanding and forging new frontiers at its best. And it's a place where you can be incredibly curious and incredibly creative. I think a lot of people who aren't scientists don't understand how creative science is and how delightful that is. And it's a community, and we're standing on the shoulders of giants, and also we're standing on the shoulders of midgets. And most of us are midgets, but that's okay because a lot of midgets makes you a giant.

Michael: Right, it's like an 80/20 giants/midgets kind of…well the other way, yeah.

Jennifer: So yeah, so I would really...I think unfortunately because of misperceptions about what science is and how it operates and what it does, misperceptions I had for a long time, there's a lot who turn away from science who actually could have a lot to contribute to science especially because they have different brain set then someone who just comes up thinking I'm going to be a scientist or a mathematician and who misses, I think interesting opportunities to be creative and to contribute in very important ways to the world, and do so in a way that's very fulfilling personally.

Michael: So maybe the capstone on this is that the meta is that there is a trophic network of knowledge that requires some measure of super-generalists, and a lot of specialists, and there's room in the scientific program for everyone.

Jennifer: Yes, I think that's a really good way to use that metaphor drawn off of my and others’ food web research but yeah, there's room for many different minds and many different kinds of intelligence and many different kinds of creativity in science.

Michael: Awesome. Well this has been a super inspiring conversation, Jennifer. Thanks a lot.

Jennifer: Thank you. That was fun.

Michael: Yeah.

Jennifer: Yup, excellent.