Improving early detection surveillance for Xylella fastidiosa in Apulia

The emergence of Xylella fastidiosa in Europe has highlighted the importance of surveillance for protection of host plants against exotic pathogens. However, surveillance for such pathogens is resource intensive and must therefore be efficient and sustainable. Using the example of the Apulian outbreak, we consider here how best to conduct surveillance in pest-free areas which are at a high risk of infection due to spread from adjacent outbreaks. To do this, we consider where inspection and survey sites should be, which diagnostic test to use during surveillance, and whether to sample hosts or insect vectors. We identify the optimal locations of surveillance sites by linking together an epidemiological model of pathogen spread and a statistically-informed sampling model with a computational optimisation routine. This approach allows us to demonstrate the considerable differences in survey deployment in order to maximise the number of infected hosts detected compared with maximising the overall probability of detection. We go on to consider the impact of the ""detection lag"" before infected hosts become detectable on both the detection probability and the expected prevalence of infection at the time of first detection. Finally, we consider whether sampling of hosts or sampling of vectors would be expected to facilitate earlier detection, and discuss some of the challenges associated with these strategies. As well as providing specific recommendations for improving surveillance activities for X. fastidiosa in Apulia, our methods are generic and can be applied in other areas and for a range of pests and pathogens