On the relation between temporal variability and persistence time in animal populations

Abstract

1. The relationship between temporal variability, spectral redness and population persistence for a large number of long-term time series was investigated. Although both intuition and theory suggest that more variability in population abundance would mean greater probability of extinction, previous empirical support for this view has not been conclusive. Possible reasons are the shortage of long-term data and the difficulties of adequately characterizing temporal variability, two issues that are explicitly addressed in this paper. 2. We examined the relationship between population variability and quasi-extinction time (measured as the time required to observe a 90% decline of population abundance) for a large set of data comprising 554 populations for 123 species that were censused for more than 30 years. Two aspects of temporal variability were considered in relation with the quasi-extinction time: a baseline value (coefficient of variation over a fixed, 30-year, time scale), and a measure of the rate of increase of the population variability over time (spectral exponent). 3. The results show that the quasi-extinction time was shorter for populations having higher temporal variability and redder dynamics. The relation between persistence time and population variability was compared for different taxa, trophic levels, habitat type (aquatic and terrestrial) and body sizes and compared with theoretical expectations.