Document: GAR-3-81-16

Reliability of time series/diffusion model approaches to PVA?

BELOVSKY, G.E.* and T.A.CROWL

Utah State University Logan UT 84322 USA ¹

Abstract:
Population viability analysis (PVA) has become an important part of assessing the likelihood of threatened/endangered species persisting under various conservation scenarios. One common PVA approach assesses population growth rates and their variance using time series analysis and then incorporates these values in diffusion models for population growth to forecast the likelihood of extinction over a chosen time period. An experimental data set on extinction dynamics developed using brine shrimp, Artemia franciscana, in replicated laboratory populations varying in initial numbers, carrying capacity and environmental variability was used to test the validity of this PVA approach. Two questions were addressed: 1) How well does this PVA approach predict extinction probabilities using a complete extinction data set (i.e., replicated populations followed until all become extinct); 2) how well should this PVA approach work when the time series for threatened/endangered species are of much poorer quality than the experimental data. The findings for both questions suggest that this PVA approach is subject to errors reducing its utility in conservation planning.

Keywords:

This abstract is being presented at: 4:15 PM in session:
Oral Session #47: Zooplankton Ecology.