Demographic buffering of life histories? Implications of the choice of measurement scale
Bjørkvoll, Eirin Marie; Lee, Aline Magdalena; Grøtan, Vidar; Sæther, Bernt-Erik; Stien, Audun; Engen, Steinar; Albon, Steve D.; Loe, Leif Egil; Hansen, Brage Bremset
Abstract
Life-history theory predicts that the vital rates that influence population growth the most should be buffered against environmental fluctuations due to selection for reduced variation. However, it remains unclear whether populations actually are influenced by such “demographic buffering,” because variation in vital rates can be compared on different measurement scales, and there has been little attempt to investigate whether the choice of scale influences the chance of detecting demographic buffering. We compared two statistical approaches to examine whether demographic buffering has influenced vital rates in wild Svalbard reindeer. To account for statistical variance constraints on vital rates limited between 0 and 1 in analyses of demographic buffering, one approach is to scale observed variation by the maximum possible variation on the arithmetic scale. When applying this approach, the results suggested that demographic buffering was occurring. However, when we applied an alternative approach that identified statistical variance constraints on the logit scale, there was no evidence for demographic buffering. Thus, the choice of measurement scale must be carefully considered before one can fully understand whether demographic buffering influences life histories. Defining the appropriate scale may require an understanding of the mechanisms through which demographic buffering may have evolved.