Population Viability Analysis in the Classification of Threatened Species: Problems and Potentials

The new Red List system from the World Conservation Union (IUCN) constitutes five sets of quantitative criteria (denoted A-E) for identifying threatened species. All criteria are based on quantitative thresholds. A-D focus on a few commonly known risk factors while only E requires a full analysis of extinction risk. To date, nearly all Red List assessments have been based on criteria A-D. Criterion E is seldomly used, partly because of assessors' unfamiliarity with existing population viability analysis (PVA) methods, but also because data are usually inadequate for such analyses. However, to more rigorously examine the causes of the limited application, I compared estimated risk levels using PVAs to levels estimated using other criteria. I assessed risk levels for Swedish animal and plant populations and also properties of these analysed species that may influence the listing process. With very few exceptions, the threat category suggested according to the PVAs (criterion E) indicated a lower risk level than the category met by the criteria A-D. I also found that PVAs were performed on a biased set of species, i.e., most often for vertebrates with small populations that were threatened by factors affecting individuals in a way that can be described by a pure stochastic model. In contrast, PVAs were rarely performed for plants or invertebrates, species with large but decreasing populations, or species affected by deterministically negative changes in habitat or other resources. Additional reasons for the dominance of criteria A-D in the Red Lists may be a lack of congruence between the extinction risks expressed by the different criteria A-E. For instance, as a consequence of criterion E operating with fixed time windows, the threat category met may represent a too conservative view in cases where the PVA extinction risk trajectory has a sigmoid shape over time.