Coupling Genetic and Demographic Data to Reveal Dispersal Processes in Emperor Penguins.

Dispersal is a ubiquitous phenomenon that affects the dynamics of the population and the evolution of natural populations; however, it is challenging to measure in most species. Furthermore, the influence of informed dispersal behaviors, referring to the nonrandom selection of breeding habitats by individuals, on species' responses to rapid global change is substantial but difficult to comprehend. Here, we present a modeling framework to assess the dispersal characteristics and behaviors of a metapopulation when observations provide information on its neutral genetic structure for a restricted sampling of locations. Our mechanistic-statistical model couples a deterministic model capturing the spatio-temporal dynamics of four genetic clusters across all breeding colonies by integrating demographic processes with genetic projections, with a probabilistic observation model describing the probability to sample an individual from a given genetic cluster. We apply this new framework to the emperor penguin, a species living in Antarctica and currently experiencing habitat loss. The model estimates the species' dispersal distance, rates of emigration, and behaviors associated with dispersal (informed or random). By incorporating these estimations with satellite censuses of breeding colonies, we can identify environmental and demographic factors that influence the dispersal of emperor penguins. Finally, we provide new global population forecasts for emperor penguins that can inform conservation actions in Antarctica.