36. Trait Driver Models

Organised by Van Savage, Harvard Medical School to be held at Network Headquarters, Sydney First meeting was held 5-8 February 2008

PARTICIPANTS INCLUDE Van Savage - Harvard (Leader) Brian Enquist - University of Arizona Stephen Bonser - University of New South Wales Tony Dell - James Cook University Graham Farquhar - Australian National University Michael McCarthy - Australian Research Centre for Urban Ecology Jon Norberg - Stockholm University Melinda Smith - Yale University Colleen Webb - Colorado State University Greg Ames - Colorado State University Christine Lamanna - University of Arizona

Jan 08 >

OUTCOMES

Day 1-2: Talks given by everyone, and important points discussed. We all have similar visions that have the potential to bring together modeling and empirical data.

We discussed most exciting outstanding problem that we thought our group was well suited to tackle, and then broke into three groups that reflect this. We worked in these groups the rest of the time with occasional meetings between groups. The three groups are:

Group 1: (Ames, Dell, McCarthy, Savage) Goal is to compare empirical trait distributions to empirical data for the environment. This allows us to look for responses of traits distributions to environmental change. This should also help us identify which traits are more responsive to which environmental factors. We suspect environmental variance (rather than environmental mean) and trait variance (rather than trait mean) may be of special importance, and if we can demonstrate this, that could be exciting.

Group 2: (Lamanna, Smith, Webb) Goal is to compare temporal and spatial patterns, and look for types/mechanisms of change at different temporal and spatial scales, according to the type of scheme. To do this, we can perform the same analyses as for Group 1 but across space, and see what similarities and difference arise.

Group 3: (Bonser, Enquist, Farquhar, Norberg, Westoby) Goal is to write a “Forum paper” that discusses the importance of looking at trait distributions both empirically and theoretically and how this might be used to project the effects of environmental change on ecosystems into the future. This would allow “general” predictions about effects of climate change on ecosystems, and that ultimately is a major goal that the community should be working towards.

Groups 1 and 2 effectively joined together because they need similar data, code, etc., and these groups will continue to work in a combined fashion.

Day 3&4: Continued working in groups.

Groups 1-2: Compiled 5 data sets and wrote a Matlab code to compute the moments of the trait distribution, properties of the time series for the environment, and correlations between the trait moments and the environment. The code also automatically generates plots for all of these computed quantities to allow easy visual inspection and comparison.

Group 3: Developed an outline for the paper and drafted 6 pages of Introduction. They also constructed a figure that depicts the main ideas of a trait distribution changing in time in response to the environment.

Last updated September 2008