Friday 3 May 2013

Bill Shipley: trait-based community assembly and other cool T-shirts (May 10)


Next Symposium speaker is Bill Shipley on trait-based community assembly.  He has developed a fascinating approach to modelling the role that traits play in determining the relative abundance of different species in a community, based on maximum entropy.  He will be speaking at the Symposium in the same session as Trevor Hastie, given their respective insights into the MAXENT method.

We’ll read his landmark paper on the topic in the Science rag in 2006
(He has since written a whole book on the subject, complete with helpful T-shirt suggestion.)

If you are extra keen, a couple of technical comments on the paper have also been published, which might be helpful to get an alternate view on the topic – see links to related resources on the Science website.

Hope to hear from you on the blog next Friday 2-3 Sydney time...

10 comments:

  1. First qn is about circularity - isn't it circular to use community-aggregated traits to predict spp via spp traits to predict community composition. This is a commonly raised issue and as I understand it the answer is no - at least, no more than the circularity in any regression where you use the y-values to predict y

    ReplyDelete
  2. And what does a community-aggregated trait mean?
    (the argument as I understand it is that if you can average over genotypes to get a spp average why not average over them in a community)

    ReplyDelete
  3. could you ramp this up to a global scale to look at how well traits explain community assembly at the global level?

    ReplyDelete
  4. next point is about how you measure abundance - a count of organisms or biomass-weighted?

    ReplyDelete
  5. we would have really loved a graph or diagram to visualise what the approach is doing!

    ReplyDelete
  6. inter- vs intra-specific variation: is it reasonable to give each spp a magic number? (Well everyone does it in other contexts anyway...)

    It would be great if there was an extension of this method which could handle intra-specific variation?

    ReplyDelete
  7. talking more broadly about the goal of analysis now - the example in the Science paper is succession but the method isn't just about succession it can be used to explain, in any community, why you get the spp you get in the quantities you get (via traits). "Trait-based community assembly"

    ReplyDelete
  8. A lot of discussion of fitness and what it means and how you measure it - the motivation behind this model is related to fitness of spp (and how traits vary fitness in different conditions) but how do you measure it? e.g. is abundance a good measure?

    ReplyDelete
  9. I'd love to see site traits as well as spp traits being included directly in the model! "Three-table analysis"

    ReplyDelete
  10. General positive feedback about the paper. It presents a significant step up from previous population ecology approaches.

    ReplyDelete