Optimal strategies for the surveillance and control of forest pathogens: A case study with oak wilt

Cost-effective strategies are needed to find and remove diseased trees in forests damaged by pathogens. We develop a model of cost-minimizing surveillance and control of forest pathogens across multiple sites where there is uncertainty about the extent of the infestation in each site and when the goal is to minimize the expected number of new infections. We allow for a heterogeneous landscape, where grid cells may be differentiated by the number of trees, the expected number of infected trees, rates of infection growth, and costs of surveillance and control. In our application to oak wilt in Anoka County, Minnesota, USA, we develop a cost curve associated with saving healthy trees from infection. Assuming an annual infection growth rate of 8%, a $1 million budget would save an expected 185 trees from infection for an average of $5400 per tree.We investigate how more precise prior estimates of disease and reduced detection sensitivity affect model performance. We evaluate rules of thumb, finding that prioritizing sites with high proportions of infected trees is best. Our model provides practical guidance about the spatial allocation of surveillance and control resources for well-studied forest pathogens when only modest information about their geographic distribution is available.