Interdisciplinary approaches, such as statistical ecology, are increasingly needed to tackle pressing environmental challenges. Uniting the disciplines of biology and statistics can help us to better understand and ultimately conserve the environment. For example, monitoring the abundance of animal populations over time is important for effective conservation and management, including sustainable resource acquisition from the environment. However, estimating abundance is difficult for many species that are not always observable (e.g., when at sea or migrating). This is an area of active research and development within the field of statistical ecology.
The grey seal population in the UK presents an ideal opportunity for the development of statistical methods for abundance estimation using a comprehensive long-term data set. This PhD project will lead to innovations in statistical ecology and make real-world contributions to the management of grey seals in the UK.
The UK hosts approximately 40% of the global population of grey seals and the population is protected under both national and international legislation. The Sea Mammal Research Unit (SMRU) at the University of St Andrews has monitored the UK grey seal population for over 30 years. Their findings feed into the NERC Special Committee on Seals (SCOS) reports which are used by UK and devolved governments to inform sustainable management of seal populations and marine spatial planning. However, accurate estimates of population size and trends (e.g., Thomas et al. 2019) depend on reliable estimates of grey seal pup production (i.e., the number of pups born each year). Multiple counts of breeding colonies are conducted over a season, and are combined with information on life history parameters to derive a birth curve and estimate pup production (Russell et al. 2019). With modern statistical methods and computational capabilities, the student, with support from their supervisors, will develop a new pup production model that can account for recent changes in survey methods and sources of observational uncertainty, and ultimately provide more robust estimates of pup production.
The project will be tailored to suit the student’s specific skills and interests, within the overall research topic. We envision that the student will build on a preliminary pup production model developed by the supervisory team to: 1) explore Bayesian approaches to model fitting that incorporate different sources of information (e.g., data, information from previous studies, expert opinion) to improve inference; 2) create a hierarchical multi-year, multi-colony models so that available information from data-rich colonies and years is effectively shared with data-poor colonies and years; 3) develop statistical methods for estimating uncertainty at a sub-population (e.g., management area, regional units) level; and 4) conduct sensitivity analyses via simulation to determine the timing and number of surveys that would maximize the robustness of pup production estimates given available resources. Statistical modelling will be conducted in the statistical software R with model development through packages such as TMB, nimble, and RStan.
King’s College London
King's College Hospital, Denmark Hill, London, UK
October 01, 2022
Imperial College London
September 01, 2022
Lancaster University, Bailrigg, Lancaster, UK
October 01, 2022