Patterns of influenza seasonality are likely driven by a combination of seasonal patterns in human behaviour, meteorological influences on viral survival, and the evolution of cellular and antibody-mediated immune responses following infection and vaccination. In temperate regions these effects have (at least pre-SARS-CoV-2) combined to yield annual winter outbreaks. Patterns of influenza disease in tropical and sub-tropical regions are often biannual or exhibit more complex patterns. Developing influenza transmission models that can explain transmission patterns across a range of geographies may help disentangle the role of immunity, behaviour and climate to improve forecasting and the timing of vaccination campaigns against influenza.

Often seasonal influenza is modelled using ODEs that capture seasonally forced multi-strain SEIRS infection dynamics and demographic turnover, with strain coupling determined by assumptions about co-infection and cross-strain immunity. However, statistical inference is challenging for such complex mechanistic models, even in temperate regions with good surveillance data, and is not well explored for non-temperate regions.

In this project we will explore the potential to use Approximate Bayesian Computation (ABC) to identify the parameters of seasonal influenza transmission models in a variety of geographical regions, using commonly available microbiological and epidemiological surveillance data. We will first consider how details of the model structure and ABC implementation influence the posterior estimation and its computational cost, informing the utility of such modelling to guide public health policy in different settings.

Extensions may include exploring the degree to which enhanced data is required to address questions such as:

· forecasting the timing and burden of influenza seasons in temperate, tropical and subtropical regions,

· understanding the role of seasonal vaccination on the maintenance of influenza seasonality,

· untangling the role of climate, behaviour and viral survival in epidemic models for endemic influenza.

This project lies at the intersection of applied mathematics and statistics. We are looking for an applicant with excellent analytic and critical thinking skills, particularly as relevant for statistical analysis, inference, and modelling of epidemic processes, with capability to interpret and communicate results. Ability to code in a scientific programming language (e.g. C/C++/python/R) is also required. Previous research experience and/or knowledge of epidemic modelling are desirable.