The dimensionality of individual niche variation

The inherently multidimensional nature of the niche has not yet been integrated into the investigation of individual niche specialization within populations. We propose a framework for modeling the between- and within-individual components of the population niche as a set of variance-covariance matrices, which can be visualized with ellipses or ellipsoids. These niche components can be inferred using multiple response mixed models, and can incorporate diverse types of data, including diet composition, stable isotopes, spatial location, and other continuous measures of niche dimensions. We outline how considering both individual and population niches in multiple dimensions may enhance our understanding of key concepts in ecology and evolution. Considering multiple dimensions as well as the within-individual component of variation can lead to more meaningful measures of niche overlap between species. The impact of a population on its food web or ecosystem can depend on the degree of individual variation (via Jensen's inequality), and we suggest how the dimensionality of individual specialization could amplify this effect. Finally, we draw from concepts in quantitative genetics and the study of animal personalities to propose new hypotheses about the ecological and evolutionary basis of niche shifts in multiple dimensions. We illustrate key ideas using empirical data from sea otters, wetland frogs, and threespine stickleback, and discuss outstanding questions about the consequences of multidimensional niche variation. Setting variation among individuals in an explicitly multivariate framework has the potential to transform our understanding of a range of ecological and evolutionary processes.