Quantifying risk of environmental exposures using Bayesian statistics. (2019)
Type of ContentElectronic Thesis or Dissertation
Degree NameDoctor of Philosophy
PublisherUniversity of Canterbury
AuthorsJaksons, Rodelynshow all
In statistics, the use of observational data is key in understanding what factors are associated with a change in risk. Often, the data also contains a temporal and spatial structure and needs to be accounted for in the modelling. By doing so, one can understand how risk changes over space and time, and to evaluate areas of increased risk. To adequately deal with the complexities of spatio-temporal and observational data, Bayesian hierarchical models can be used. In this thesis, we use Bayesian statistics in the four case studies to quantify risk to environmental exposures and to identify which variables are associated with the greatest change in risk. The applications deal with predictive modelling for water quality management, spatio-temporal analysis of campylobacteriosis disease risk, estimating the extent of underreporting in epidemiological data, and modelling the emergence dynamics of the western corn rootworm for pest management. The models for each application are explained in detail, and the results discussed in depth. Additionally, we also discuss how the methods used in the applications are relevant to other disciplines.