An agent-based modeling approach to represent infestation dynamics of the emerald ash borer beetle

Agent-based modeling (ABM) is a bottom-up approach capable of operationalizing complex systems. The approach can be used to reproduce the spatio-temporal patterns in ecological processes such as insect infestation by representing individual dynamics and interactions between “agents” and their environment from which complex behavior emerges. The emerald ash borer (Agrilus planipennis; EAB) is an invasive species native to south-east Asia which has infested and killed millions of ash trees (Fraxinus sp.) across the eastern United States as well as Ontario and Quebec in Canada. Efforts to model the insect's behavior are ongoing, but current models are limited to approaches that do not address the complexity that emerges from the dynamics between individual beetles and their varying spatial environments. The objective of this study is to develop an ABM to represent the interactions of the EAB and the emerging spatio-temporal pattern of the insect spread. The model is implemented on real datasets from the Town of Oakville, Ontario, Canada from 2008 to 2010. Tree inventory and land use data acquired from the Town of Oakville were used to represent the spatial environment of the EAB agents. The EAB interactions are implemented in the model as subroutines, each representing a stage in the EAB life cycle using a temporal resolution of one day. Model verification was performed based on the literature documenting the life cycle processes of the EAB to represent EAB behavior. The model is calibrated using the rate of spread observed in the Town of Oakville from 2008 to 2009 and is validated using datasets delimiting the spatial extent and severity of EAB infestation in 2009. When comparing simulated and observed data, there is a 72% agreement for the locations of the infestation. This indicates that the developed ABM approach offers a model able to capture the complex behavior of EAB where both the spatial extent and severity of infestation are simulated realistically. The model generates insights about the underlying processes governing EAB behavior, highlights areas of uncertainty in modeling the complex spatio-temporal patterns of EAB infestation, and is a useful tool for forest and pest management.