Data merging: are we moving forward or dealing with Frankenstein's monster

I’m sitting in the Sydney airport waiting for my delayed flight –which gives me some time to ruminate about the mini-conference I am leaving. The conference, hosted by the Centre for Biodiversity Analysis (CBA) and CSIRO in Australia, on "Understanding biodiversity dynamics using diverse data sources", brought together several fascinating thinkers working on disparate areas including ecology, macroecology, evolution, genomics, and computer science. The goal of the conference was to see if merging different forms of data could lead to greater insights into biodiversity patterns and processes. 

Happy integration

On the surface, it seems uncontroversial to say that bringing together different forms of data really does promote new insights into nature. However, this only really works if the data we combine meaningfully complement one another. When researchers bring together data, there are under-appreciated risks, and the resulting effort could result in trying to combine data that make weird bedfellows.

Weird bedfellows

The risks include data that mismatch in the scale of observation, resulting in meaningful variation being missed. Data are often generated according to certain models with specific assumptions, and these data-generation steps can be misunderstood by end-users, resulting in inappropriate uses of data. Further, different data may be combined in standard statistical models, but the linkages between data types is much more subtle and nuanced, requiring alternative models.

Why these are issues stems from the fact that researchers now have an unprecedented access to numerous large data sets. Whether these are large trait data sets, spatial locations, spatial environmental data, genomes, or historical data, they are all built with specific underlying uses, limitations and assumptions.  

Regardless of these issues of concern, the opportunity and power to address new questions is greatly enhanced by multiple types of data. One thing I gained from this meeting is that there is a new world of biodiversity analysis and understanding emerging by smart people doing smart things with multiple data. We will soon live in a world where the data and analytical tools allow research to truly combine multiple processes to predict species' distributions, or to move from evolutionary events in deep history to modern day ecological patterns.

NCEAS is dead; long live NCEAS. A view towards NCEAS 2.0.

"is this a wake or revival?" Jim Brown

March 21-22, 2012, Santa Barbara, CA. National Center for Ecological Analysis and Synthesis (NCEAS) symposium.

A special invitation- only symposium marking the end of NCEAS as we know it, saw a number of interesting talks and retrospectives about where NCEAS has been and where it is going. 170 people attended, including some former postdocs, working group participants and leaders in ecology. The reason for this introspective meeting is that NCEAS's core NSF funding is about to end, without renewel. Jim Brown's quote from his talk, whether we were here for a wake or a revival really captured the spirit of the meeting.

The goals were twofold. First was to look back and celebrate the accomplishments of NCEAS. University of California at Santa Barbara is globally one the top influential research institutions in the world, and this success has been driven in large part by the success of NCEAS. More than 5000 people have come to NCEAS and their efforts have resulted in thousands of publications, and many citation classics. The early visions of NCEAS were broad and fuzzy and by all accounts NCEAS has exceeded all expectations.

The second motivation for ts meeting was to think about the future. What can NCEAS be under different funding regimes, and how should it move forward? The is no doubt that it will be fundamentally different, but can there be a successful continuation of the NCEAS model, will it die, or will it give birth to a new enerprise, NCEAS 2.0?

The symposium saw great talks, from people like Jim Brown and Jane Lubchenco, and interesting panel discussions on numerous topics (see #treas2012 in twitter for synopsis of the meeting). There were a lot of past tense statements.

However, it was clear that there was much to celebrate. NCEAS clearly impacted ecology. Did its success simply coincide with cultural changes in the field or did it drive changes? The consensus was that it drove changes. It fostered large collaborations. Dave Tilman said that before NCEAS, ecology was largely local and lab-driven, but NCEAS offered a way to get people together to ask bigger questions. The postdoctoral fellows have been extremely successful, with the vast majority ending up in faculty positions in top institutions. It was acknowledged that many sub fields were created or coalesced at NCEAS, including disease ecology and metacommunity dynamics.

Why has it been so successful? NCEAS is a special inclusive place where people want to come, away from their responsibilities. The technical help here and expertise that made anything possible, any data challenges were overcome and analytical difficulties solved. Postdocs were given complete independence and were allowed to pursue collaboration and networking. Jim Brown remarked that NCEAS is the single greatest event in the history of ecology. Subfields now talk, lab projects are now geared towards collaboration and linkages with other work in ways that did not exist before.

So then, what will the future hold for NCEAS? The answer to this was left vague and uncertain. People argued for what NCEAS 2.0 should look like. For example, it was argued that NCEAS 2.0 should resurface something like science 2.0, making the focus data and data sharing, changing methods and philosophy of how science is done. Massive anonymous collaboration requires assumed standards and altruism. Other arguments focused on the need for NCEAS to reach out to new partners and to go global.

Peter Karieva said it well. NCEAS 2.0 should be interacting with major corporations, since they represent the drastic impacts on ecological systems around the world. 1.0 was about data accessability, 2.0 should about applicability and tools to affect change.

Whatever NCEAS 2.0 looks like, it will be different. There seems to be two ways forward. One is that it struggles to maintain its past activities or one that like the Phoenix rises from the ashes and boldly goes forward to again push the ecology in new directions.

ESA Austin: Day 3

Another great day of talks, and surprisingly so. I went to several talks that were more or less randomly chosen, and was impressed by some of the science that graduate students and younger scientists are doing. While others see a potential decline in ecology, I see a very bright future.

M. Duffy examined how virulence susceptibility and the cost of disease resistance in daphnia and its pathogen (yeast). This talk had one of the best setups I seen, and is based on a trade off between r and resistance. Epidemics result in increased resistance, as one would expect with evolution, but equal number of lakes showed evolution for increased susceptibility because of these tradeoffs. Small epidemics should result in increased susceptibility because of a greater fitness cost from reducing r than from mortality. Large outbreaks should result in greater susceptibility as mortality is high. She also showed that environmental context could alter expectation (e.g., productivity or predation).

In the next talk N. Loeuille examined the heterogeneity of resource dispersal. Classic models assume homogeneity in resources availability for competition. But with different diffusion rates, niche competition may be decoupled from tradeoffs needed for coexistence. He used a model with differential dispersal of resources. Depending on tradeoffs, the model will produce the evolution of diverse strategies of disperal. There will be specialization on single resource if dispersal is symmetric. If dispersal rate is too high or too low, but equal, then the resources support lower diversity. If two resources disperse differentially, creates heterogeneity at different scales and will support higher consumer diversity.

I ran over to Tad Fukami’s talk. He examined phylogenetic patterns in priority effects in the assemblages of yeast that colonize flower nectar. He hypothesized that there should be a strong priority effect with close relatives, since they tend to occupy similar niches. He tested this by Introducing species in different orders and assessed relatedness effects using genetic sequences. The experiment was directed by natural history of the system, like time length of flowers, microbial population dynamics in flowers. If one species colonizes first, he showed that it can reach carrying capacity regardless of the presence of other species. If it arrives second, there are major effects on abundance, but differs between which is first species. Closely related species result in strong priority effect, but weak with distant relatives. Result robust even if you control for differential ability to deal with abiotic conditions. The potential mechanisms include differentially reducing amino acids, and different growth rates on sugars.

Kevin smith gave a great talk on extinctions. He used several large, recently assembled datasets to examine how range size correlates with extinction risk under different scenarios of habitat destruction. Randomly, you would expect that broad species have a low probability of extinction overall and endemics have a high probability. Across the datasets, widespread species are going extinct at higher probabilities then a random model. Land snails conform to random model based on species range. However, for bird and amphibian datasets, the rare species bahave as expected with high extinction risk, but the middle ranged species have higher than expected extinction risk

L. Prevost examined how the theory of Island Biogeography (IBG) explained diversity patterns in fragmented habitats in mid to high elevation habitats in Costa Rica. The short answer is not very well, there were not distance or area effects on plant diversity. Communities have low similarity, no relationship with distance, but are similar according to elevation. It seems as though species responses to heterogeneity drives the system, so she recommends that many small reserves could be valuable.

In a very interesting and stimulating talk, A. Rominger examined fluctuations in evolutionary history. He showed that there are more fluctuations than predicted by various models including random walks (which conforms to a Gaussian distribution). Gaussian is observed in small time slices, but variances change over time. Fluctuations within orders fit Gaussian very well, but different from one another. Volatility itself evolves by gamma distribution. He shows that volatility is inherited within orders, and fascinating and controversial conclusion.

John Parker testing the often assumed, but understudied assumption that exotics differ in advented populations versus native. Basically, is there an away-field advantage? Examined home and away for 1000 worst invaders, across many taxa. Looked at size, reproduction, population growth. None of these were particularly enlightening. For example they are not bigger in away sites, size at home predicts invader size 1:1. There is some variation, but no consistent trend. Fecundity, also no consistent trend, with noninvasive just as likely to increased fecundity. Abundance not consistent but slight trend to be bigger away. Survival, growth, same thing. Overall slightly better away, but not greatly. We need to reexamine our hypotheses.

Day 2 in Austin

The second day of ESA got off to a good start in Austin, with a day full of more community ecology talks than one person could attend. I split my day between Community Assembly and Neutral Theory II& III (it's interesting to note that ten years in neutral theory is now included in so many eponymous sessions) and Biodiversity I, and regretted missing talks in many other sessions.



The Community Assembly and Neutral Theory II covered a diverse range of systems (from microcosms to primates), scales, and methods of study. Lin Jiang presented an experiment examining the relationship between diversity and invasibility, in particular testing whether priority effects reduce the oft-seen negative relationship between diversity and invisibility. Most manipulative experiments "assemble" communities instantaneously rather than continuously and stochastically as in natural systems, and so more realistic assembly may weaken the sampling effect and niche complementarity, which are suggested to drive the negative relationship. Using protist-based microcosms of 5 resident species and one invading species, Jiang examined how more realistic assembly of communities affected the diversity-invasibility relationship. Under these conditions, there was still evidence of a negative invasion diversity effect. His most interesting result however were that in fact the presence of a close relative had the strongest influence on the success of the invasive species, in line with other theoretical and empirical results (although not conclusive given the small number of species).

In the Community Assembly and Neutral Theory II session common topics in this session included the widely-used framework of hierarchical filters (i.e. abiotic, biotic, dispersal limitation) determining local species composition and tests of the predictions of the neutral theory (with a focus on non-SAD predictions). For example Wang et al. looked at the patterns of clade age versus abundance that were predicted by neutral theory, in comparison with empirical data from the BCI dataset. There was a clear divergence between the observed data, which included old clades with high or medium abundances compared to the neutral theory prediction that old species should have low abundances. Wang examined how relaxing the assumption of equal rates of speciation among species affected the age-abundance patterns, but concluded that different rates of speciation among species wouldn't produce the observed pattern without then failing the SAD predictions. However, one astute commenter noted that it might also be important to model the possibility of changing rates of speciation through time.

There were many other interesting talks. For example, in Biodiversity I, Matthew Leibold provided a master class on resilience to human disturbance, focusing on concept of communities and ecosystems as complex sets of coupled oscillators. Finally, Angela Brandt presented the results of a 7-year experiment in California grasslands (which no doubt represents many hours of hard work), in which she examined the relationship between invasion success, resource availability, and disturbance. In particular, she framed the question from a phylogenetic context, and discovered evidence that disturbed communities tended to be both more species-rich and phylogenetically diverse, and also less phylogenetically clustered, compared to non-disturbed communities. However, if communities received nutrient enrichment and disturbance, invasion was greater, and diversity lower, than in communities that received disturbance treatments only.

Looking forward to day 3!

***Addendum by Marc:
There were two additional talks that were particularly interesting. First was Andy Gonzalez’s talk on Evolutionary rescue, which is when a population is in demographic decline, heading towards extinction and adaptation saves it from extinction. This is particularly important in changing environments. He has previously shown this using yeast growing under salt stress. His question now was whether migration in a metapopulation with heterogeneous affects this evolutionary rescue. This is an interesting question because too little dispersal means that genetic variation or beneficial mutations do not get to other patches, and too high means that suboptimal genes are maintained in patches where they are maladapted. Not surprisingly he showed that in a constant environment, dispersal is not very important. In heterogeneous metapopulations, patches at the edge of salt tolerance thresholds increase in yield with dispersal.

The other very interesting talk was from K. Anderson on niche-based sorting in highly diverse palm assemblages. She looked at soil type and resource availability, as well as herbivore damage. She experimentally planted 13 species across a soil gradient with and without herbivore exclosures. In low nutrient soils, the palms invested more in roots while in high resources soils, there was a greater investment in above ground biomass and an increase in photosynthetic rate. Leaf toughness also increased in poor sites, meaning that they were more resistant to herbivory and plants growing in the high resource sites experience more herbivore damage. She mentioned that there were differential responses from the different species, and I am very interested to see more about this neat system.

ESA Austin: Day 1

We are now through a great first day at the ESA Austin meeting, and have been having a great time both at talks and out on the town in Austin (see photos). Looking over the program, it was obvious that the day had too many good talks, and that it was impossible to see them all. Considering that I was giving a talk, I decided to spend my entire time in my session on biodiversity and ecosystem function. It may seem lazy, but there were a bunch of talks that sounded great. Here are short summaries of all the talks in the session (excluding mine of course).

The Decemberists playing at Stubbs. Fantastic show (maybe the highlight of day 1, if not for the many interesting talks)

Darwin's pub, great name, OK pub.

The first talk by Nicolas Mouquet was probably the best. It was on the relationship between species diversity and ecosystem function, asking how we can move from the question of how many species to which species. The ultimate answer, according Mouquet, comes from evolution. By understanding the evolution of specialization, one can discern the importance of niche complementarity in the additive contributions to ecosystem function. Using simulations, he showed that the relationship between richness and function is dependent on whether species are specialist or generalist and the strength of tradeoffs in resource use. He then told us about fantastic experiments that evolve bacteria on differing resources, creating specialists and generalists. Positive diversity-function relationships were higher but not stronger in assemblages of generalists, because they deal with heterogeneity better. He manipulated the amount of evolutionary history in assemblages and found that the relationship between evolutionary diversity and function was stronger with groups of specialists. This research goes beyond most others in that it explicitly links coexistence to ecosystem function.

Next was a talk by J. Passari, looking at ecosystem multifunctionality in a long term plant experiment. He examined eight different functions and examined how local, large scale and among site diversity influenced ecosystem function. He found that multifunctionality increased with increasing local diversity but less so with diversity at larger scales.

Greg Crutsinger, showed how genotypic and phenotypic differences in coyote shrub morphs resulted in differences in arthropod abundance and richness, and changes in litter communities.

M. Striebel examined how phytoplankton diversity affect function. She showed that total pigment diversity (representing photosynthesis) increased with phytoplankton diversity. Also she examine how this diversity affected zooplankton diversity and found positive relationships in oligotrophic and mesotrophic systems, but not eutrophic ones.

J. Mclaren manipulated the functional group richness in desert and arctic plant communities and examine the community and functional responses. There was some compensation by other functional groups, but only a weak overall affect on function.

J. Petermann manipulated basal resource diversity and predator richness in bromeliad aquatic communities. She measured four functions and found only weak effects, it seems as though bromeliad leaf complexity may drive some of these relationships.

E. Harvey showed how multiple extinctions in complex food webs can have important cascading effects on ecosystem function. He measured multiple functions in freshwater and marine communities, and that different extinctions had differential effects and some where non-linear.

JJ Weis gave a very interesting talk where he used a model to assess how intra- and interspecific diversity affect function. He found that high complementarity resulted when species had high genotypic variation but low genotypic breadth.

Finally, T. Hanley, who is also a student in the same lab as JJ Weis, examined how intraspecific variation affected population dynamics of daphnia and their algal prey. There wasn’t any effect of daphnia genotypes on algal or daphnia dynamics, but daphnia genotypic diversity increased through time.

What is interesting about this group of talks is the diversity of organisms, systems, scales and functions being considered. These talks are a great signal that biodiversity-ecosystem function research transcend locales and is now a broad, mature field of study.

Beware of the fake conference

I don't know about other researchers, but I get inundated with e-mails about upcoming conferences from organizations I've never heard of, on topics that, at best, only tangentially relate to my work. I think that most of these are put on by for-profit groups that try to cash in on hot topics. But now there are truly fake conferences that are fronts designed to get your financial information. In a new post by Bob Grant on The Scientist news blog, he relays a detailed example of such a scam conference. He went so far as to contact the venues and speakers listed by the scammers, and of course none of those listed had ever heard of the conference. The ultimate enticement was the offer to pay for travel expenses, and the presumption is that they would offer to reimburse you, but need your bank account information. The evolvability of internet scammers is truly impressive.

BES day two...

It was a day full of talks, and I really enjoyed being able to spend the whole day just absorbing the presentations. Here are some highlights from today's events:

• Ian Wright gave an interesting talk about the history of functional plant ecology. Basically, covering where trait ecology has been and where we are now. It is really amazing to see the truly large scale analysis and collaborations currently driving modern trait analysis.
• In another overview type talk, Gerlinde De Deyn, talked about carbon sequestration in soils. I'm sure to many ecosystem ecologists this maybe well known, but I found it fascinating. Did you know that tundra ecosystem have as much soil carbon as tropical rain forests? The reason is that tundra has very slow process rates (cold) while rain forests have fast production rates. In fertilization experiments, soil carbon is reduced, so in order to manipulate soil carbon stores one must understand the interaction between plant traits and soil organisms.
• Finally, Kyle Dexter, in a great field survey of Inga tree species in a region of Peru, showed that different functional diversity metrics show differing patterns of over- and under-dispersion. For example, phylogenetic diversity tends to be under-dispersed for Inga assemblages, while chemical and anti-herbivory traits are over dispersed and leaf size measures are under-dispersed.
  

Also, there was the annual general meeting, which was rather somber as three obituaries were read aloud. The three deceased, John Harper, Bob Jefferies and Simon Thirgood, were all superb ecologists who absences were obviously felt. Harper (my Master's advisor's advisor) and Jefferies (my colleague at Toronto) were both eminent ecologists with long and distinguished careers, while Thirgood (a fellow Senior Editor at the Journal of Applied Ecology) was in the prime of his very successful career.

Exploring ecology through GMOs

This year's Tansley Lecture at the BES meeting was a superb presentation given by Ian Baldwin from the Max Planck Institute for Chemical Ecology. He was enjoyable to watch as his folksy, mid-western American style disarmed the listener and leaving them unprepared for his ascorbic wit and, at times, controversial message. Prof. Baldwin* is a chemical ecologist who studies plant biochemical and physiological processes and their interaction with herbivores. Through his use of molecular tools and superb natural history, he has gained new insights into how and why plants respond to herbivory. He has discovered the pathways allowing wild tobacco, Nicotiana attenuata, to detect chemicals in tobacco hornworm spit and the resulting chemical defense response. More than this though, part of his talk was about the use of transformed plants to study this plant defense response. Using genetic tools, his group was able to knockout certain segments of these biochemical pathways in order to determine how various chemicals affected hornworms. He showed chemical responses involved signaling hornworm predators whereas other responses directly targeted wornworm's ability to digest plant material.

I think that ecologists are often wary of GMOs and his talk was a convincing case for their use in basic research, and he advocated for a more reasoned approach to their use.




*Note: He has run into trouble with German authorities over using the title 'Dr.' -see here.

An ecological meeting, British style.

I'm in Hatfield, outside of London, for the British Ecological Society (BES) meeting, and I'll be blogging thoughts, happenings, etc. I posted several observations from the American equivalent (ESA) last month, and one thing is readily apparent, even though talks do not start until tomorrow. While the ESA meeting is a great chance to meet a lot of people and see a lot of talks, at a few thousand conference goers, it can be a bit overwhelming. At any given time the ESA meeting had up to 26 concurrent sessions, which makes for a lot of running around if there are talks you want to see in different sessions. At the BES there is a maximum of 7 concurrent session. The BES is definitely smaller than ESA and the meeting and venue feel compact and intimate. I'm looking forward to experiencing the BES meeting this week.

Check back for more.

Macroecology is dead, long live macroecology!

I went to a session on a macroecology yesterday, which featured some wonderful speakers, and came away with an unsure feeling about this field. The Session started off with a fantastic talk by Rob Dunn on how macroecologists differ on what the main mechanisms are for explaining diversity patterns. He argued that perhaps the complexity of natural and human-altered systems make simple generalizations not very fulfilling. Next Lauren Buckley showed how species turnover had complex relationships with broad environmental changes and that species turnover patterns are better correlated with other species turnover then with the environmental variables we think drive the patterns. Next Brian McGill tried to make the case for a truly unified theory. He walked through several general models of random species packing, showing that some models fit observed data very well. I was impressed by the data/model fits but am skeptical of a general theory which lacks biological mechanisms. My view of a scientific theory is that it ought to contain basic mechanisms and that a unified theory should explain patterns and processes at multiple scales. That said, I also think McGill has done more to forward the field than almost any other younger ecologist. In Allen Hurlbert's talk, he nicely showed how independently accounting for energy and area can provide a better basis to constructing and understanding species-area relationships. The basic reason is that area and energy availability differentially affect the number of individuals.

Back to my real life tomorrow!

Pleasant invasions surprise

Normally I run around ESA looking for talks that have the best potential to inform or entertain me. This time around I decided to go to a session on invasions and communities and settle in for the long haul. Am I glad I did. I was afraid the session would be dominated by similar sounding talks, but instead each talk was wildly interesting and different. Talks included looking at the genetic variability of the dominant native resident as a proxy for niche preemption. Another good one looked at the role of propagule pressure for an understory invasion into tropical dry forests -I seldomly hear about invasions into these ecosystems. Next was a look at how invaders behave over long term successional trajectories and they by and large appear to follow native trends. Next was a great modeling talk where individual-based models and riverine networks were used to assess the role of distrubance and trait differences in invasion dynamics. The final one I saw was on how to potentially restore Californian serpetine plant communities using little more that gravel and a few chemicals, with the goal of reintroducing extirpated butterflies, which have not been able to cope with the shift to exotic-dominated grasslands.

I am looking forward to more great talks!

Species interactions & evolution

Hi from ESA Albuquerque!

I've been in the organized session on species interactions and evolution all morning and there were some great talks (e.g., Silvertown, Ackerly, Cavender-Bares, etc.). But I think what really got me excited were some of the questions after each talk. Following Jonathan Silvertown's talk, Steve Hubbell asked some questions that get to the heart of addressing what phylogenies mean for community assembly. Silvertown showed that within plots, species of a large South African family of plants in the Fynbos seemed to spatially segragate according to hydrological niches and that within these plots there was a lack of phylogenetic signal in this niche. Hubbell then asked two critical questions: How many other species (in other families) co-occur in these niches and if related species have similar niches at a larger scale. To me this is at the core of uderstanding how phylogenies inform our understanding of community assembly. Basically, what haven't we measured? If we include all sister species into a phylogeny, do we change our understanding of the processes structuring communities?

More later!