Modelling the spatial spread risk of plant pests and pathogens for strategic management decisions

Abstract

Invasive pests and pathogens pose a grave and growing threat to agricultural systems, threatening food security and access to domestic and export trade, as well as having potentially serious environmental and social impacts. Management responses to pest or pathogen incursions include planning for conducting surveillance, implementing quarantine restrictions and treating infected farms. Smart and transparent decisions need to be made to control the spread of an incursion, often in the context of sparse information. Given the array of management options, modelling tools can help greatly in more targeted spatial prioritisation of surveillance and quarantine as the incursion response unfolds. We constructed a simulation model to allow managers of incursions to explore the possible effects of their decisions. The model operates at the farm-level of spatial resolution, with risk pathways parameterised via flexible networks of connections between farms. It models growth of the pest organism within farms, its spread between farms by natural or human-assisted means, and its detection and control by land-owners or managers. Quarantine options within the model can range from isolation of an individual property to region-wide restrictions. Surveillance options range from straightforward ranking by estimated risk to more sophisticated search prioritisation algorithms. The model continually updates risk estimates from previous surveillance results and ongoing prevailing risk propagated through the connection pathways. Initial risk estimates can incorporate tracing data, wherein response managers have documented previous high-risk pathways connected to properties with known infections. Simulated surveillance results can be replaced by actual surveillance history data, if available. Users such as biosecurity managers can explore model features and scenarios through the interactive model visualisation software “BioSpark”. Whereas the model is underpinned by simulation of within-farm growth and between-farm spread of the invading organism, the output for response program managers is primarily presented as visualised expected risk to farms of unknown pest status, and their ranked risk levels, to aid in guiding surveillance decisions. Risk is visualised as heat maps overlaid on the landscape, and can be viewed over each time step of the analysis. Sites recommended for surveillance are presented as a ranked list, with their locations flagged on the landscape. We have retained sufficient generality in the model, so that future developments can include considerations of different organisms, flexibility in number of risk networks to be entered, and application of the model to different jurisdictions. There is a need for model-based tools such as this to aid in framing, supporting and communicating agricultural management decisions in complex and challenging situations.