Developing a spatial epidemiological model to estimate Xylella fastidiosa dispersal and spread

Understanding the dispersal of Xylella fastidiosa is essential for effective management of the disease. In Puglia, Italy, surveillance is focused on buffer and containment zones established at the edge of the infected region with the aim of containing further spread. Success of this strategy will strongly depend on whether these zones are wide enough to form a barrier to long distance dispersal of the bacterium. In this presentation, I will describe our progress towards estimating the dispersal range of Xylella in Puglia using a generic spatial epidemiological model adapted to the biology of the pathosystem. The model simulates the spread of the disease across a heterogeneous landscape depending on the location and timing of introduction, the distribution of host plants, the rate of infection growth in infected olive groves and both short and long distance dispersal. Long distance dispersal seems to be a crucial feature of the Xylella epidemic, causing rapid spread of the disease over large areas but in an unpredictable manner. To calibrate the model, we used Approximate Bayesian Computation to compare model simulations to Xylella surveillance data and remote sensing of severe damage. This allows us to contrast a simple spread scenario with more complex scenarios such as anisotropic dispersal in the direction of prevailing winds and spatial variation in disease transmission. In doing so we characterise the spread and estimate the year of introduction. Finally, I will discuss potential for using the model to simulate management strategies and new outbreaks in other regions, using the UK as a case study.