PARENT SESSION
Oral Session #79: Conservation Ecology: Biodiversity Case Studies and Models. Presiding: T. Rooney.
Friday, August 10, 2001. 8:00 AM to 12:00 PM. Hall of Ideas J.
Inference about diversity estimates: Confidence intervals and statistical comparisons for sample species richness.
Garrett, K.1, 1
ABSTRACT- Measures of biodiversity such as sample species richness are of great importance in characterizing communities but are difficult to work with from a statistical standpoint. Estimates of expected sample species richness are often presented without any measure of confidence in the estimate. And comparisons of diversity between communities that depend on techniques such as rarefaction are limited by the fact that resampling from one sample always results in a new sample of the same or lower diversity. A new bootstrap method for constructing confidence intervals on expected sample species richness offers some improvements. This technique is based on weighting observations in samples by their contribution to the estimate of expected sample species richness. In a study of simulated sampling from large communities, confidence intervals constructed using the weighted bootstrap method typically gave coverage of the true expected sample species richness within a few percentage points of a desired 95% coverage.
KEY WORDS: biodiversity, bootstrap, rarefaction, species richness