Monday, 4 February 2013


I spent last week in the company of approximately 450 waterfowl researchers, managers, and scientists. If that doesn't speak for itself, here's my recap.

The Peabody Ducks
The hotel's duck theme, explained here, was prevalent from duck-shaped soap and butter to duck-printed towels to a chocolate duck for dessert. It was easy to see why the conference organizers had chosen the Peabody Memphis Hotel for this conference!

I learned about WETLANDSCAPE through talks by Adam Janke and Brett Werner, and with subsequent chatting with Brett and Carter Johnson. In simple terms: they've created a model which takes precipitation and temperature as inputs and predicts various characteristics of wetlands on the landscape. Although it's aimed at the Prairie Pothole Region at the moment, they told me it shouldn't be too difficult to expand out to the Boreal region. I told them about Ying Fan and Gonzalo Miguez-Macho's Depth to Water Table model I encountered via this article on waterfowl distribution, since they said they'd had some trouble with ground water. I'll be keeping an eye on this project, since my current work is limited to climatic variables, which are somewhat indirect in terms of influencing waterfowl populations. It would be nice to have a time series of ponds and other water bodies instead.

Habitat Selection Discussions
I was pretty excited  to hear discussions on habitat selection by waterfowl during the first plenary.  At least three people mentioned Johnson's hierarchical framework for habitat selection.  However, I got the impression that people feel that research on this topic has been exhausted, since much of the research in the 70s was aimed at this question. Moreover, although people mentioned habitat selection at the first order (geographic range or distribution), no one discussed which features ducks might be cuing in on to determine their distribution. This question has haunted me since I started looking for remotely sensed environmental variables to include in my models: What environmental factors determine or limit each species' distribution?  I'm starting to think that this is an unreasonable, unanswerable, and perhaps esoteric dissertation question, since it delves into evolutionary history, species divergence, behavioural aspects such as philopatry and dispersal, behavioural responses to land use change, and various other sub-fields of waterfowl ecology. In other words: it's not something I can tackle in one chapter of my PhD. I chatted with Pat Devers about Black Ducks specifically, and we were at a bit of a loss to deal with such a large question. He did mention, however, that the Black Duck Joint Venture has compiled a list of hypotheses on more fine-scale Black Duck habitat selection, and he told me about a PhD thesis on the topic, so I'll be following up with him shortly. Lastly, I appreciated Mike Eichholz's point in the synthesis section where he mentioned that we need to treat individual species separately rather than manage for "waterfowl" because we currently work under the (false) assumption that what's good for mallards is good for all ducks. One of my thesis chapters tests a question of species vs. group using distribution models, so it was encouraging to see that the question is quite relevant.

Dynamic Wetlands = Prairie Pothole Region?
Mike Anteau gave an interesting talk about Gadwall which suggested that wetland dynamics (measured by a wetland change index) are important for Gadwalls (a prairie-pothole species). I wondered if this "wetland dynamics" attribute might explain the apparent importance of variability in Autumn precipitation in my species distribution models. Additionally, I wondered if this feature is what makes the PPR the PPR. I'm finding that "% cropland" is acting as a surrogate for "Prairie Pothole Region", but what is the underlying feature that defines the region? Could it be this wetland dynamics? Or perhaps a combination of high density of small wetlands in addition to strong dynamics? I'd like to quantify what really defines the region and use that in my models. Unfortunately, I don't have time series of wetlands in the boreal region, so I'm not really able to quantify wetland dynamics in the boreal at the moment (but see WETLANDSCAPE, above).

Waterfowl Breeding Population and Habitat Survey (BPOP survey) Side Meeting
A smattering of BPOP survey-using researchers met on Tuesday night to discuss what we're doing and chat about the survey in general. Ken Richkus gave us an update on the review that's been going on for a couple of years, which was great to hear. Three of the pilot biologists were there as well, to chat with us about how some of us are using the data. Tom Nudds suggested that we should have a directory or database of all work that is written on or uses the database, include unpublished stuff, reports, grey papers, etc. John-Michel DeVink also suggested the counter-point: some way of documenting what we've tried that didn't work out. Tom and I brainstormed afterwards and started thinking that a forum or listserv might be very useful. Some central location to see everyone who is working with the database, as well as a forum or email list to post questions about "who has tried ___?" or "is anyone working on ____?" and get feedback. We could also use such a forum for project updates and to keep on top of the literature.  I had started reviewing the database early in my PhD, so I have a large list of papers that use it already - I'll be cleaning that up and making a BibTeX version shortly so I can pass it on to whoever compiles the database.

waterfowl conservation planning & WBPHS
A poster I presented  in 2010 on using the BPOP survey data to inform conservation planning.

Tracking Waterfowl
The radio-tracking and telemetry work was fascinating. To see ducks that were all tagged in relatively similar areas end up all over the breeding range was astounding and reignited my interest in behavioural ecology: How do individual ducks decide where to go? How strong is natal philopatry? And of course I saw the direct implications for my species distribution models - how do my predictions match up with their inferred breeding sites? I briefly spoke with Kurt Anderson, who gave the talk on Black Ducks, and I hope that we'll speak more in the future. Mike Schummer gave a talk about how well the BPOP survey tracks Scaup. I missed the talk due to a conflict with one of my own talks, but I got the gist of it via discussions: several of the tagged Scaup were either missed, double-counted, or settled outside of the BPOP range. I'm not yet sure of the direct implications of these results on our interpretation of BPOP survey-based Scaup research (for that, I'll need to read the paper), but it's worth keeping in mind.

Quantitative Methods
It was fun to hang out with the people from Utah State University, including Dave Koons` grad students.  We remarked on how much of an increase we've seen in hierarchical modelling since the last NADS conference in 2009. People seemed very interested to talk about the quantitative methods I presented in both a talk and a poster (talk co-authored with Chris Roy, another PhD student at U.Laval). I've never been so busy at at poster session before! I enjoy talking about statistics, so it was a great experience for me.

My poster at ECNAW - sadly I forgot to get one of me standing in front of it. Some of the Koons' crew

Ideas vs. Science, and the Whole is Greater than the Sum of its Parts
Over the course of the conference, I tried to find my place among these waterfowl experts. As a young, female, non-hunting, mostly vegetarian student, I was a bit of a minority.  Research-wise, however, I had a bit of help from Aaron Pearse and Courtney Amundson, on the first and last days, respectively. Aaron used an interesting framework for his plenary talk, based on Dyson's 2012 Science paper on whether ideas or tools drive the progression of science. I can't remember if he said this or if I thought it on my own, but I recall thinking that each individual researcher's interests will fall somewhere between the two extremes: idea-driven vs. tool-driven. On the last day, Courtney mentioned something I'd be thinking over the duration of the conference: it takes all types. Like any field of biology, what starts as simple life history diverges into more and more specific questions. We might be fiercely proud of our own little niche (quantitative methods, bottom-up or top-down effects, migration patterns, harvest, etc), but at the end of the day we really do need every little piece to form a coherent picture. Following these two perspectives, I'd say that my research is mostly tool-based: computers and statistics can handle these huge questions and datasets, so let's address questions using methods we couldn't use before. Although it's not yet super apparent to everyone why these fancy hierarchical Bayesian methods are important, I think us quantitative people have an important role too.  If we can communicate the methods clearly and use them to answer the right questions, we'll see them being used more and more frequently to answer both theoretical ecology and applied conservation/management questions.

I was able to attend this conference because of an Excellence Award from the Qu├ębec Centre for Biodiversity Science. Supplementary funding was provided through the Centre for Forestry Science.

No comments:

Post a Comment