Spatial and epidemiological modelling for wildlife and agricultural health

点击翻译按钮获取中文摘要

This project will develop spatial and epidemiological models for wildlife and agricultural diseases. It will use two distinct host-parasite systems where there is a need for complex models to address fundamental applied issues. Although the two applied problem areas (trichomonosis in UK finches, and uptake of Integrated Pest Management by UK farmers) involve different stakeholders, different data issues and different epidemiology, they are linked by a commonality of approach and modelling challenges. Key questions applied to both systems include: using model-fitting to understand the disease; assessing the outcomes of disease control policy; and assessing feedback across scales. The first two chapters will concern the use of mathematical modelling in understanding the finch trichomonosis outbreak, which has been ongoing since 2005, and has resulted in drastic declines in British greenfinch and chaffinch populations. The availability of population and disease incidence data allows forthe development and fitting of detailed mathematical models for this disease system. The core research questions which we seek to address are: (1) can we infer the spatial patterns in the disease dynamics from routinely collected wildlife data sets, (2) what mechanisms are driving seasonality. Initially, a deterministic time-dependent disease model will be developed. A key aspect of the model development methodology will be working with the BTO and IoZ experts in constructing a functional representation of the relevant biological parameters, such as contact rates, density dependence of disease, and population birth and survival rates. This model will be fitted to the population and disease reporting data using likelihood optimisation methods (maximum likelihood and MCMC), allowing us to infer mechanisms and routes of infection. This model and corresponding fitting will then be extended to include the spatial dimension of the data. The next chapters will focus on agricultural diseases. This will involve the integration of farmer-behaviour models with crop disease models, in order to investigate strategies to reduce dependency on chemical pesticides and increase the uptake of Integrated Pest Management (IPM) by farmers. Currently the use of pesticides continues to rise; in large part due to the perception by farmers that the alternative disease management strategies, such as Integrated PestManagement (IPM), are difficult and costly. In order to encourage the uptake of IPM by farmers, the UK government offers a scheme of incentives, primarily as payments per-year. Behavioural modellingcan be used to guide incentive schemes by evaluating the impact of various IPM adoption outcomes on the disease system. Two key questions which this project will aim to answer are (1) how successful would varying levels of uptake of IPM be at controlling disease, and (2) what fraction of farmers would need to initially take up IPM in order for it to be successful at controlling disease. This problem can be approached using a deterministic ODE model initially; investigating the outcomes from different initial conditions, and different parameter values dictating the interactions between the disease system and the farmer-behaviour system. Later methodological approaches can then be expanded to include stochastic frameworks, which explore localised or individual-based dynamics.The intention is to begin this work by looking at yellow rust in cereal crops. Models already exist for this disease system, and there are a number of well defined control strategies which are known to work. One such strategy involves the use of resistant crop varieties, which are important for yellow rust control, but which farmers find challenging because of the need to change varieties frequently since resistance can be quickly overcome.