Mucking Through Multi-factor Experiments: Design and Analysis of Multi-factor Studies in Global Change Research

At the Ecological Society of America 91^st Annual Meeting

August 6-11, 2006

Memphis, Tennessee

 

Sponsor:

Terrestrial Ecosystem Response to Atmospheric and Climatic Change (TERACC)

 

Organizer:

Aimée T. Classen, Oak Ridge National Laboratory, Environmental Sciences Division, Oak Ridge, TN 37831-6422, Tel (865) 574-7848; email: classenat@ornl.gov

 

Abstract:

 

Progress has been made in the past decade to understand terrestrial ecosystem responses to climatic change using numerous single-factor and a limited number of mutli-factor experiments. Model based analyses have also been widely used to speculate on future ecosystem responses, but their predictions remain largely untested. There is an increasing awareness that multiple, and often confounded, environmental variables may dictate the structure and function of ecosystems. Therefore, in order to answer the question “What are the potential consequences of global environmental change for ecological systems?” multi-factor experiments are needed. Multi-factor experiments are complicated by design, however, and they demand a concomitant increase in the conceptual and analytical complexity of statistical analyses for their interpretation. This symposium brought together researchers from a variety of projects, backgrounds, and expertise to discuss our collective ability to understand and interpret the results from multi-factor experiments.

 

The symposium included talks that tease apart some of the complexities in multi-factor experiments using current projects as examples, discussed creative ways to analyze complex results using mechanistic models, and discussed how to fully integrate models into future experimental designs. The progression of talks discussed (1) lessons learned from multi-factor modeling, (2) results from established multi-factor experiments and the challenge of interpreting those results, and (3) new and creative analytical approaches for de-convoluting complicated experimental results. The symposium concluded with an open discussion by panel members and the audience on future multi-factor experimental design and synthesis.

 

 

 

Introduction

 

A multifactor world, a multifactor problem.  Claus Beier (Risø National Laboratory, Denmark)

 

Lessons learned from multifactor modeling:

 

Potential Long Term Impacts of Global Change on C and N cycling in Forest and Grassland Ecosystems.  Steve Del Grosso (USDA Fort Collins, CO), Bill Parton and Dennis Ojima (CO State University).

 

Modeling multi-factor interactions in CO₂-enrichment experiments.  Ross McMurtrie and Belinda Medlyn (University of South Wales, Australia).  

 

Using multifactor global change experiments to answer big science questions: linking science to monitoring and policy.  Kevin Percy (Canadian Forest Service) and David Karnosky (Michigan Technological University).

 

An overview of existing multifactor studies:

 

Lessons learned from almost 10 years of multifactor work at BioCON.  P. Reich (University of Minnesota).

 

Linking above- and belowground processes in a multifactor world: analyzing and interpreting multifactor experiments.  Aimée Classen, Rich Norby (Oak Ridge National Lab, Tennessee) and Jake Weltzin (University of Tennessee).

 

When do interactions matter? Continuing lessons from Jasper Ridge.  Jeff Dukes (University of Massachusetts), H. Henry, Rebecca Shaw, N. Chiariello (Stanford University) and C. Field (Carnegie Institution).

 

New analytical approaches:

 

    Using single factor experiments to answer multifactor questions.  Paul Hansen, S. Wullschleger, T. Tschaplinski, Rich Norby, and Carla Gunderson (Oak Ridge National Lab, Tennessee).

 

    Multifactor experiments as a model selection problem: An application of Bayesian Reversible Jump MCMC.  B. Beckage (University of Vermont) and James Clark (Duke University, North Carolina).

 

Future design and synthesis discussion  

 

Meeting Publication

 

Dermody, O.  2006.   Mucking through multifactor experiments; design and analysis of multifactor studies in global change research.  New Phytologist  172:600-604