Variation in the Size of Animal Populations: Patterns, Problems and Artefacts

Abstract

We investigate problems associated with the measurement and interpretation of population density variability that have confounded most, if not all, previous studies of the subject. The most commonly used measure of variability, S.D. (log[N + 1]), is subject to potentially very large bias, and other measures based on untransformed densities can be affected by the long-tailed frequency distributions of population data. The fact that variability often depends on the mean density is usually ignored, yet it affects the generality of any statement based on a single estimate of variation. When comparing variability between species, the generality of any conclusions is also limited by the choice of spatial and temporal scale of sampling. This potentially confounds many studies. We discuss the patterns that emerge from considering spatial and temporal Taylor power plots, and suggest that combinations of each plots for individual species allow a classification according to markedly different types of dynamics; the underlying mechanisms of which are not well understood. Finally, we consider ways of making meaningful comparisons of spatial and temporal variability across species.