Hidden Markov models for spatially structured populations

Hidden Markov models (HMMs) offer a very powerful, flexible, and efficient structure for likelihood computation. They’re a popular tool for problems in statistical ecology since the structure often has an associated, and insightful, ecological interpretation for the system being modelled. This PhD will look to develop a general and unified framework for a collection of models that have yet to be expressed as an HMM formulation. One such group of methods are dynamic occupancy models that can be applied to spatially structured populations, surveys with spatially structured sampling, or the combination of both.  In particular, this project will focus on approaches that can be applied to metapopulations with well-defined patches that experience colonisation-extinction dynamics, continuously distributed populations where data are collected through spatially structured sampling, and multi-state/hierarchical extensions that allow for inference on the spread and prevalence of disease and pathogen transfer within spatially structured populations. The methods will be explored through a combination of simulation and application to real-world case studies.