ESA 2012 preview: all the good talks are on Friday

Ok, the post title is not literally true*. But it sure seems that way! So if you’re fortunate enough to be able to stay through Friday (sadly, I’m not–family commitments…), you’re in for a treat. Here are some Friday highlights:

Barreiro et al., 8 am, Ballroom 258. Ecological stoichiometry has been hot for years, and while it’s not an area I follow that closely, I have the vague impression that people haven’t really moved on that much from the ideas that made this area hot in the first place. For instance, I have the impression that lots of people are still really into the “growth rate hypothesis”, basically the idea that P limitation = slower growth because making RNA requires lots of P. But growth rate is not a fixed constant, and so there’s surely more to the stoichiometry-population dynamics link than just the “growth rate hypothesis”. As Barreiro et al. have shown, by feeding a rotifer on algae manipulated to have different N:P ratios, following the resulting long term population dynamics, and using the data to parameterize a dynamical model. They find differences in extinction rate, cycle period and amplitude, and other features of population dynamics under N- vs. P-limitation, and the model reproduces the results.

Gilbert et al., 9 am, C123. Stochastic drift, aka demographic stochasticity, is a hot topic in ecology lately, and is going to be a focus of mine at the meeting because it’s something I’m currently working on. It’s precisely analogous to genetic drift in evolution. A major shortcoming of most recent work on drift in ecology is that it’s highly indirect. People are trying to detect or quantify drift by measuring things that are putative “far downstream” indirect consequences of drift, which creates all sorts of inferential and interpretive difficulties. So I love the quite direct experimental manipulation Gilbert et al. did: Set up 123 annual plant communities, comprising monocultures and various mixtures of species, with a gradient of total absolute population size. Because just as in evolution, the “strength” of drift relative to other processes should depend on population size. They not only found the expected effect of population size, they also found interesting effects of species interactions. Basically, species interactions not only alter species’ population sizes, thereby altering the strength of drift, they also generate different patterns and strengths of “selection”, thereby mediating the ability of drift to create among-replicate variation in relative abundance. Gilbert et al. also seem to have detailed individual-level demographic data and so perhaps can say something about the amount of demographic stochasticity during different life history stages.

Adler et al., 9 am, Ballroom 256. “Trait-based ecology” is hot right now. But there’s actually not much good data showing that the putatively-key traits that people often measure (usually because they’re easy to measure) actually drive variation in demographic rates. And it’s demographic rates that matter for things like life history evolution, population dynamics, species coexistence, etc. Adler et al. have compiled a bunch of data on plant “functional traits” (things like wood density and specific leaf area) and demography (from matrix models), and asked how they’re related. The upshot is that traits matter, but the most oft-measured traits leave a lot of demographic variation unexplained. Sounds like a useful reality check for “trait-based” ecology, and perhaps a prod to get people to worry more about what traits we ought to be measuring.

Snyder et al., 10:10 am, Ballroom 256. Ecologists routinely study the conditions under which species coexist. They rarely stop to ask why we should expect those conditions to occur. In particular, it’s not like evolution by natural selection cares whether species coexist or not. So why should “eco-evolutionary dynamics” exhibit any tendency to evolve towards coexistence-promoting regions of parameter space? Snyder et al. approach this broad issue in the context of eco-evolutionary models with demographic variability (e.g., temporal variation in fecundity). Under what conditions can you select for greater demographic variability (e.g., by selecting for traits that make a species more sensitive to environmental fluctuations)? This issue is interesting in its own right (basic evolutionary theory says the answer is “never”). It’s also interesting for its implications for coexistence in variable environments. You have to have demographic variability for variability-based coexistence mechanisms like the storage effect to operate, so insofar as basic evolutionary theory is right, selection will tend to undermine the operation of variability-based coexistence mechanisms. Synder et al. use a combination of theoretical models and real data to identify conditions selecting for increased demographic variability. They find them, but the effect is weak. Which raises a puzzle: we observe variability-based coexistence mechanisms in nature, even though selection seems to either oppose, or at least not strongly promote, their operation.

Chesson, 10:30 am, D136. Peter Chesson, the man who’s done more than anyone to develop modern coexistence theory in its most general form, will be talking about the key conceptual limitation of that theory: the assumption of stationarity. Basically, what this means is that even though environmental variables and species’ densities may fluctuate, even quite dramatically, those fluctuations have unchanging long-run properties (i.e. the mean, variance, and other statistical moments don’t exhibit any long-term directional trends). But of course real environments aren’t stationary. Which raises not just empirical problems, but more importantly, conceptual problems: how does one even define “coexistence” in a non-arbitrary way in a non-stationary environment? Because in the long run, a non-stationary environment can doom every species to extinction! Peter is, as far as I know, the first person to think about this very important problem, and it looks like he has some quite interesting and general results, based in part on some new concepts like “quasistationarity”. This is the hardest sort of theoretical work to do well–work where you need to define new concepts just to be able to fully articulate the question, much less answer it. Once I find out from someone what Peter had to say, I’ll have to add an additional post to my series on coexistence theory in variable environments.

Other cool-looking Friday talks (listed by speaker): Hart, Bieger, Gounand, Meisner, Simonis, Godoy, Haynes, Cortez.

As an aside, I’ve always wondered why ESA ends with a half-day morning session on Friday, as the vast majority of attendees leave before that. Yes, I know some folks will always leave a multi-day meeting early, no matter if it ends with a half- or full-day session. But for instance, if memory serves the recent Evolution 2012 meeting ended with a full-day session, and the final day was much better attended than the final day of ESA. Which I think stands to reason: I mean, wouldn’t you think people would be more likely to stay for a full-day session than for a half day? Am I just totally off-base on this? What am I missing? And the answer can’t be that you need that extra half-day on Friday to fit all the talks in, because there are no talks on Monday morning. My suggestion: hold talks on Monday morning, rearranging the workshops and awards ceremony and whatnot if necessary, and drop the Friday morning talks and posters. There’s partial precedent for this: back when I was a grad student, talks at the ESA meeting started on Monday morning, though even then the meeting still ended with a half-day on Friday.

Hope to get around to a preview post for the other days soon. And I’ll be blogging from the meeting as usual.

*Although it is at least as true as some putative generalizations in ecology. 😉