HOME     SCHEDULE     AUTHOR INDEX     SUBJECT INDEX    

PARENT SESSION
Oral Session 4: Biodiversity I: Communities and Function.
Presiding: C Holzapfel and D Ellsworth
Monday, August 2, 8:00 AM to 11:30 AM, Meeting Room B 110.

On the generality of biodiversity – ecosystem function patterns observed in random assembly experiments.

Gross, Kevin*,1, Cardinale, Bradley2, 1 North Carolina State University, Raleigh, NC2 University of Wisconsin - Madison, Madison, WI

ABSTRACT- As concern about global loss of biodiversity has intensified, the number of studies examining how species richness influences ecosystem functioning has increased dramatically. In the first round of experiments, studies have attempted to elucidate causal relationships between biodiversity and ecosystem functioning (BEF) by (i) selecting an experimental pool of species from the regional pool of all species that live in an ecosystem, and (ii) randomly drawing species from the experimental pool and placing them together in an experimental unit to see how they influence a nutrient uptake, biomass production, etc. Often, we want to draw inference not just about the experimental pool, but about the "true" BEF relationship that characterizes the regional pool. Here, we ask whether such generalizations are safe. Using a model, we show that BEF patterns can vary remarkably among experiments conducted with different experimental pools drawn from the same regional pool. This variability itself is not surprising; however, the frequency with which random assembly experiments suggest incorrect conclusions about the regional BEF relationship is striking. The problem is exacerbated by the fact that customary statistical analyses ignore this variability. We outline the limited set of circumstances where BEF patterns from an experimental species pool can be generalized to larger collections of randomly assembled communities. We then introduce resampling based statistical methods for estimating the variability in BEF patterns among experiments, and show how these methods may enable broader inference about the effects of biodiversity on ecosystem processes.

Key words: species richness, resampling methods, biodiversity, ecosystem function

All materials copyright The Ecological Society of America (ESA), and may not be used without written permission.