A null model of community disassembly effects on vector-borne disease risk

Community structure is heterogeneous at a variety of spatial and temporal scales, and this variation has been shown to influence the risk of zoonotic diseases such as West Nile Virus and Lyme disease. Theoretical models and most empirical evidence suggest that the greatest influence of host diversity occurs when transmission is frequency-dependent (i.e., the rate of contact is constant). These theoretical models are generally based on ordinary differential equations and become intractable when considering more than a few species. This makes it particularly difficult to predict how we might expect the transmission of infectious diseases to change as community structure changes in space or in time. Here we develop a model in which we construct a network of interactions between hosts and vectors to quantify the change in risk under different scenarios of community disassembly. Decreased vector biodiversity always reduced mean risk, while a change in host community structure led to increased or decreased mean risk depending on the manner in which community disassembly altered mean competence of the "new" community. These trends in mean risk can be generalized across a multitude of natural systems because they do not depend on the distribution of host quality, though simulation suggests that variation around the mean can be very high. The primary value of model is that it can be used to establish upper and lower bounds on the expected change in disease risk with decreasing biodiversity.