Invasive tree pests pose a significant threat to UK forests and robust tools are required to accurately and promptly quantify their spread. In this talk, I'll introduce a framework for modelling the spread of invasive tree pests using a spatiotemporal reaction-diffusion equation. From this equation, we can determine the pest's expansion rate directly from model parameters, which are inferred in the Bayesian paradigm. An adaptive Metropolis-within-Gibbs Markov chain Monte Carlo (MCMC) computational scheme is employed to obtain posteriori estimates of the parameters, accounting for uncertainty in the observational data. The framework is applied to the ongoing spread of oak processionary moth (OPM) and the effectiveness of the model at capturing the apparent biphasic spread of OPM is demonstrated, providing evidence that the expansion rate has increased since 2019. Thus, the proposed framework is a powerful tool for quantifying tree pest spread, could underpin future prediction and management approaches, and demonstrates the practicality of Bayesian inference methodologies for training and validating models of invasive pests. I'll finish the talk with an outlook for developing the model further, including by introducing host heterogeneity and environmental influences such as wind-driven transport.
Jamie McKeown (Newcastle University)