Continuous time models of animal movement, applied to animal-borne tags and spatial capture recapture

Most animals must move to forage, avoid predators and locate mates. Hence understanding animal movement is crucial to a full understanding of individual animal success and, ultimately, whole-population dynamics. Ongoing technological developments mean that increasing numbers of animal-borne tags are being applied; depending on the model these may yield noisy observations of position (via ARGOS satellite), or precise positions (via GPS), and other data such as height or depth (for diving animals), direction, acceleration, etc. Although movement occurs in continuous time, analytically tractable continuous time models tend to be overly simplistic (e.g., assuming random movement (Brownian motion)); more complex modelling tends to involve discrete time approximations. This project will investigate the potential for fitting more complex continuous time models using analysis where possible, but simulation-based inference where not. An additional application of such models is in fitting spatial capture-recapture models – these are used to estimate population size from time series of “captures” of marked individuals in a population. The capture does not have to be physical, for example identifying a uniquely-marked tiger on a series of infra-red-triggered camera traps throughout a forest. This PhD will suit students with an interest in (and prior training in) statistics and applied mathematics.