Lev Ginzburg Fest: celebration of a “retirement”

“Lev Ginzburg has retired”. For anyone who knows Lev, this combination of words does not make a whole lot of sense. Is it possible that such a lively and active scientist would hang up his yellow pad and pencil in order to put his feet up in some retirement community far away from the world of academic science? Not likely, but Lev Ginzburg has retired from Stony Brook University, sliding into Emeritus status while remaining highly unlikely to let up on his usual level of frenetic scientific activity. Lev won’t be teaching courses or advising students at Stony Brook any more, and if that is what we call “retirement”, then that’s worthy of a gathering to celebrate Lev’s many accomplishments as an academic and advisor. On Saturday, December 12th, 2015, friends, family, colleagues, and former students all gathered at the Wang Center to acknowledge Lev’s accomplishments thus far… even as those of us who know him expect a “retirement” with little let up in scientific output.

The day started off with a series of fun and interesting talks inspired by and about Lev’s work. Jessica Gurevitch, who co-organized the academic part of the event with Dianna Padilla, gave a short introduction of Lev’s career trajectory from Leningrad, Russia to becoming a refusnik emigrating to the United States and eventually landing at Stony Brook. She discussed the impressive number of Ph.D. students that Lev advised: ten. She also acknowledged Lev’s particular support of women in science.

Resit Akçakaya was the earliest of Lev’s Ph.D. students (#3) to speak (Lev’s first and second students, Carlos Braumann and Sinjae Yoo, both sent tributes but were unable to attend). Lev’s student during the years that ratio-dependent predation theory was first getting some serious attention, Resit described how he transitioned from work on theoretical ecology to work on conservation ecology after discovering how roughly theoretical ecologists (present company excluded, of course) often treated each other. One of Lev’s big insights early on was that extinction risk could be represented by quasiextinction curves, and Resit extended this approach by developing the metapopulation models that would become a critical feature of the RAMAS Metapop software produced by Lev’s company Applied Biomathematics. Resit talked about some of his more recent work, which focuses on the issue of how much warning time is needed in order to prevent a species discovered to be at risk of extinction from actually going extinct. While such “warning time” can be deduced from populations models, this approach only works for species whose major population growth parameters can be reliably estimated. A lot of species under threat haven’t been studied sufficiently to allow parameterization, so Resit wanted to know if there were any other ways of approximating warning time. Using historical data on species that actually went extinct, Resit and his collaborators discovered that IUCN criteria can be reliably used to provide adequate warning times (~60 years on average), but only if the lowest threat categories are used as the initial “warning” to which policy-makers respond. Resit also discussed how in one facet his work has come full circle: his dissertation project analyzed lynx-hare population cycle data using ratio-dependent predator-prey theory, and one of his most recent conservation papers looked at extinct risk in Iberian Lynx using an assumption that is ratio-dependent in nature.

Jon Borrelli discusses the citation trajectory of AG1989 (photo: Paul Distenfeld)

Jonathan Borrelli, Lev’s latest and last student, spoke next. He gave a really fun and creative talk that looked at what it meant to be an “againstnik”, Lev’s term for being willing to go against the status quo of ideas. Very cleverly, Jon came up with a way to analyze the impact of Roger Arditi and Lev’s first joint paper on ratio dependence (1989): using the paper as the fundamental unit of “selection” and citations as a proxy for “fitness”, he analyzed how the idea of ratio dependence evolved over time. Applying his expertise at analyzing networks, Jon created a network by looking at the interconnectivity between papers citing Arditi-Ginzburg-1989 (AG1989) and papers citing papers that cite AG1989. He was able to find four distinct “modules” that he characterized as separate approaches to citing AG1989. One module was devoted to discussing the value of ratio dependence as an idea, whereas another was devoted to the theoretical implications of the theory. The two other modules encompassed experimental and applied clusters that were citing AG1989. Although it’s interesting that these distinct clusters emerged, it is not necessarily indicative of healthy science: separate domains may have been looking at the idea of ratio dependence in relative isolation. Jonathan also showed that “againstnik” ideas like ratio dependence take a while to catch on but eventually can have large overall impact.

Roger Arditi appears via video (photo: Paul Distenfeld)

Some of Lev’s most important collaborators were not able to make it to Lev Ginzburg Fest, including Roger Arditi. But via video Roger did provide an interesting talk that explored an idea that Lev has advocated for years: that the most commonly-used form of the Logistic Equation creates bizarre and unrealistic behaviors if used for more than one population. Looking at a model depicting a pair of populations linked by migration, Roger showed that the logistic equation produces counter-intuitive results. In contrast, using the original equation proposed by Verhurlst in 1838 remedies the problems with the Logistic but removes the term “K”, which is commonly interpreted as representing the connection of maximum sustainable population size to ecological resources.

Scott Ferson considers our intuition about incertitude (photo: Paul Distenfeld)

Scott Ferson — Applied Biomathematics’ longest-standing employee and a major contributor to the literature on risk — spoke next on the two kinds of uncertainty encountered whenever we try to make predictions about the future of anything. The first form of uncertainty is stochastic in nature: it emerges from the inherent variability of a system. Not surprisingly, we call this kind of uncertainty “variability”; figures such as quasiextinction risk curves capture this variability by looking at the probability that a certain event will happen over a certain interval of time. Even if you have perfect estimates of the parameters that describe a stochastic system, you still will be uncertain of future outcomes due to the variability of the system. But what if — as is almost always going to be the case — you do not have a perfect estimate of all the parameters (or even the mechanisms) that describe a stochastic system? Not knowing with accuracy how a system works introduces an additional form of uncertainty, an uncertainty that makes the predicted probabilities produced by variability uncertain. This additional source of uncertainty is called “incertitude”, and differs from “variability” in that incertitude can be reduced at least partially by efforts to better understand and parameterize the system. Scott pointed out that humans are famously bad at understanding risk and probability, highlighting a number of paradoxical experimental results that suggest that humans process variability and incertitude in different manners. These paradoxical results can be obtained cross-culturally and therefore appear to indicate something about how the human brain evolved to assess risk; in fact, fMRI scans reveal different regions of the human brain that appear to be separate “probability sense” and “ambiguity detector” mental calculators.

Kevin T. Shoemaker spoke next via video. Kevin was a post-doc at Applied Biomathematics before landing a job as Assistant Professor at the University of Nevada, Reno. He spoke about how GIS-based risk assessment can be used to both estimate and mitigate the impact of wind-power turbines on migratory bat populations. He focused specifically on a model of the Indiana Bat. Because very little is known about both the actual incidence rates of Indiana Bat fatalities due to collision with wind turbines and the migratory patterns of this species, Kevin used GIS data to assess likely habitat and migration patterns across the American Midwest. Interestingly, he was able to utilize electrical circuit theory to estimate the “current” (density) of migrating bats based on their “sources” (where the bats roost in the summer), the “resistance” (less preferred migratory route characteristics) of their migratory landscape, and the “grounds” to which they migrate. This map of different migration densities could then be compared to the current siting of wind turbines to estimate population risk. Kevin discovered that the current set of wind farms do not pose a threat to Indiana Bat populations, but that failure to account for bat migration patterns as more wind farms are added to the landscape could lead to population collapse. Kevin would like to work towards a tool that would use his techniques to provide wind farm designers the ability to find ideal locations that balance wind power generation potential with lower risk to migrating bat populations.

Junhyong Kim points out just how massive our genetic code is (photo: Paul Distenfeld)

Junhyong Kim, Lev’s fourth Ph.D. student, was the next speaker. Junhyong began with the statement that “every year I realize how ignorant I was last year”, a nice way of summing up the joy of a scientific career. Although the work that he now does is far from that of Lev, he explained how valuable his time working with Lev was, in particular for allowing him to appreciate the structure of ideas and how to differentiate kinetic versus dynamic processes. He also pointed out how remarkable it was that Lev has been working on some of the same ideas, some for as many as forty-five years, especially given how contemporary science focuses on the kind of short-term projects that ensure a continual funding supply. Junhyong presented some of his work on how cells function together to form an individual. Specifically, he discussed how single-cell transcriptome analysis has demonstrated that there is significant variation in what genes are being transcribed even within a particular cell type. Although cell types do produce distinct sets of proteins, the variation in transcriptome between cells within a certain cell type can — for some cell types — rival the level of variation observed between different cell types. It may be that rather than each cell producing the exact proteins required to function, different cells may respond to available concentrations of proteins found in their tissue type. Junhyong suggested that this means that there is a kind of “ecological dynamics of cells”, that cells respond to each other metabolically much the way that birds in a flock respond to each other behaviorally.

Pablo Inchausti introduces his talk on a model of tropical tree diversity (photo: Paul Distenfeld)

Pablo Inchausti was the last of Lev’s Ph.D. students to give a scientific talk. He talked about an individual-based model (IBM) that he had developed to explain the high levels of tree diversity observed in the tropics. What’s amazing about tropical forests is how well they have been characterized: there are data sets for tropical forests across the world that chronicle diversity for huge swaths of forest over several decades. Using the traditional equilibrium theory of niches to understand tropical forests is impossible, as analyzing diversity patterns for the approximately 300 tree species in the tropics would require well over 900,000 parameters. There’s also good reason to suspect that traditional niche approaches might not work for tropical rainforest, which is what inspired the Neutral Theory of Stephen Hubbell. Pablo wanted to see if a model that was mechanistic — rather than neutral — could capture patterns of tropical tree diversity. He created a very simple IBM with three replacement rules, and it did a good job of accounting for existing diversity patterns. Interestingly, the diversity patterns of tropical forests across the globe were all best captured by the same approximate niche width, suggesting some kind of community-level convergent evolution (perhaps an example of non-adaptive evolution?). I also thought that it was interesting that the mean-field model required so many more parameters than the IBM: apparently not all IBM’s are over-parameterized in comparison with their mean-field equivalents.

One of the slides I presented in my “Lev Ginzburg, an analysis” presentation

The remaining talks were tributes to Lev. I volunteered to give a brief overview of Lev’s career as well as funnel in ideas from folks who could not attend the event. My “analysis” was a faux principle components analysis (PCA) in which I claimed to analyze Lev’s publications over five decades; shockingly one of the PCA axes was the “axis of insanity” along which Lev’s publications varied from “sane(-ish)” to “blazing mad”. I wanted to highlight not just that Lev had produced some somewhat controversial (some say “crazy”) ecological ideas but that his career was far more diverse: he has published a lot of papers with evolutionary implications, and much of what he has published has been quickly embraced by his scientific peers.

It was my honor to pull the ideas of former students Carlos Braumann, Sinjae Yoo, Jeffrey Yule, Janos Hajagos, Paul Bordeau, Jessica Stanton, and Toni Lyn Morelli into a brief overview of Lev’s many wonderful qualities as a teacher and mentor. I spoke of the vision that Lev had for his students, visions that were often beyond what we as his students could see: many people spoke of Lev’s optimism about and encouragement of their career prospects. I also discussed how Lev’s encouragement — sometimes in the form of persistent pushing — led all of us who were his Ph.D. students to publish our work early and often; as a result, Lev can — and often does — say “all of my students have good jobs”. Many of the featured speakers discussed just how much time Lev was willing to spend with his students, an amount of time that sometimes inspired the incredulous jealousy of our fellow graduate students. Everyone who worked with Lev mentioned the role of his obsession in his process, how he would think about something, ruminate on it, obsess some more, and then suddenly have something ready to publish. Many also mentioned Lev’s humor, which could just as easily be used to diffuse the sting of a research setback as it could be to subtly dispatch competing theories. Perhaps the most inspiring aspects about Lev that were consistently mentioned are those that relate the least to science. Lev inspired his students by being resilient in his diverse pursuits, not just in the way he tackled a variety of scientific challenges but also in the way that he dealt with the challenges that life threw his way. Parties after colloquia at Lev’s were legendary because of the warmth he created in his home. And all of us who were his students felt a part of his family, as Lev enthusiastically looked after our interests as a parent might.

Mark Burgman pays tribute to Lev via video (photo: Paul Distenfeld)

Mark Burgman was not able to make the trip from Australia, but he did send a really nice video tribute that summarized Lev’s substantial impact on his career. Mark does some pretty high-level risk assessment work for the Australian government, and he explained how his work at Applied Biomathematics and working with Lev was instrumental to building the rich career he now enjoys.

Pablo Inchausti returned to the stage to give a six-slide tribute to Lev; he insisted that he had traveled the greatest distance of any attendee — over 7000 kilometers — just to deliver these slides. Pablo talked about first meeting Lev at the time that he was — in Pablo’s word — stricken with “allometric fever”. Pablo then discussed a lot of memories that were familiar to us as Lev’s former students: the smoke-filled office, the yellow pads used to scrawl ideas, and the long discussion sessions in that office with those yellow pads. Pablo also discussed a connection he had with Lev: a past in the communist party (Pablo’s via family, Lev’s a matter of survival in the U.S.S.R.). Pablo finished his tribute by quoting Dizzy Gillespie’s depiction of his mentor Louis Armstrong: “no him, no me”. Pablo’s emotional tribute was a fitting end to the day’s talks.

The full roster of live speakers at Lev Fest (photo: Paul Distenfeld)

There’s no way that any one-day set of talks can do justice to the career of an accomplished academic, but this day’s talks did as well as possible. Dianna and Jessica did a great job organizing the event, which was a very nice mix of different kinds of tributes to Lev. Jonathan Borrelli also deserves credit for making a great little program for the event, as does Emily Rollinson for performing the academic cat-herding required to get our diverse audio-visual presentations to run flawlessly.

Pablo Inchausti listens as Lev narrates some of the visual history on display

After the talks were over, we all headed to the Simons Center for a nice dinner. The festivities then — most appropriately — headed back to Lev’s house. Lev’s wife Lara did a fantastic job of setting up a great party, including a lot of wonderful images from Lev’s past. One of my favorite moments of the night found Lev explaining many of the images featured on a large photoboard chronicling his youth and career. There were these incredible photos of Lev as a kid, images of a wacky youth that made me smile because they made so much sense. Lev talked extensively about his experience emigrating from the USSR, including a long stint in Italy during which he became close with the son of Vito Volterra, who was a hero of the Italian resistance to Mussolini and a member of the Italian Supreme Court. With nothing to his name, Lev reached out to Volterra’s son, who became interested in how his father’s work in population ecology had fared in the biological sciences. Not surprisingly, Lev charmed Volterra’s son and ended up hob-nobbing with the scientific elite across Italy. I loved this story because it reinforced both how little Lev started out with and how he used a combination of personal charm and scientific wit to become successful.

Beyond getting the chance to celebrate Lev, I really appreciated the opportunity to get to know some of Lev’s students — Pablo Inchausti, Junhyong Kim, and Jonathan Borrelli — whom I had never met. If the ten of us who Lev mentored represent a population, it is a population very representative of Lev’s diversity. We all work on different challenges in very different contexts, but there’s a shared pioneering spirit in all of our work that very much reflects Lev’s own willingness to explore uncharted territory.

As he was my Ph.D. advisor, perhaps it is self-evident that I wouldn’t be where I am today without Lev. But what Lev provided me with goes well beyond shepherding me towards the completion of a dissertation and the earning of a degree. In fact, I would say that helping me (greatly) to earn my degree was among Lev’s smaller contributions to my career. What Lev has really given me is an approach to scientific problems and an example of how to live a rich life of ideas. Although I may not think about the same things as Lev, his way of thinking has greatly affected the way I think about the problems that matter to me.

Right now I am also in the process of editing a short tribute video to Lev, which I will post soon.

Congratulations Lev on completing the first stages of your already-accomplished career. I am excited to see what you do next in “retirement”.

Below is an image gallery from the non-academic portions of Lev Fest: