Modeling landscape-scale pathogen spillover between domesticated and wild hosts: Asian soybean rust and kudzu

Many emerging pathogens infect both domesticated and wild host species, creating the potential for pathogen transmission between domesticated and wild populations. This common situation raises the question of whether managing negative impacts of disease on a focal host population (whether domesticated, endangered, or pest) requires management of only the domesticated host, only the wild host, or both. To evaluate the roles of domesticated and wild hosts in the dynamics of shared pathogens, we developed a spatially implicit model of a pathogen transmitted by airborne spores between two host species restricted to two different landscape patch types. As well as exploring the general dynamics and implications of the model, we fully parameterized our model for Asian soybean rust, a multihost infectious disease that emerged in the United States in 2004. The rust fungus Phakopsora pachyrhizi infects many legume species, including soybeans (Glycine max) and the nonnative invasive species kudzu (Pueraria montana var. lobata). Our model predicts that epidemics are driven by the host species that is more abundant in the landscape. In managed landscapes, this will generally be the domesticated host. However, many pathogens overwinter on a wild host, which acts as the source of initial inoculum at the start of the growing season. Our model predicts that very low local densities of infected wild hosts, surviving in landscape patches separate from the domesticated host, are sufficient to initiate epidemics in the domesticated host, such that managing epidemics by reducing wild host local density may not be feasible. In contrast, managing to reduce pathogen infection of a domesticated host can reduce disease impacts on wild host populations.