The dynamics of the coffee rust disease: an epidemiological approach using network theory

Dynamic modeling of plant pathogens has usually been accomplished with either a mean field or spatially explicit approach, searching generally for either broad generalization or precise prediction. In search of a qualitative intermediate that is able to query spatial particulars of transmission, we take an approximate approach using network theory. Some elements of network theory are applied to a specific case of the early spread of the coffee rust disease (agent = Hemileia vastatrix) on a single large shaded coffee farm in Chiapas, Mexico. We find that infection rates within connected components are more homogeneous than infection rates among components, suggesting that the initial stages of this disease show pattern that can be detected using simple ideas from network theory.