Kathleen LoGiudice1, Shannon T. Duerr2, Michael J. Newhouse2, Kenneth A. Schmidt3, and Richard S. Ostfeld2. (1) Union College, (2) Institue of Ecosystem Studies, (3) Texas Tech University
Despite much study, the extreme temporal and spatial variation in Lyme disease risk remains unexplained. Here we further test a model based on the dilution effect hypothesis, which posits that Lyme disease (LD) risk is highest when the community of vertebrate hosts for the vector is species-poor and dominated by the most competent reservoir (the white-footed mouse), and lowest when the host community includes a diverse set of species with lower reservoir competence. The model was parameterized in a continuous forest tract at the Institute of Ecosystem Studies (IES) in New York State and tested by comparing observed values for nymphal tick infection prevalence (NIP) with those predicted by the model from 1995 to 2005. We further tested the model in 40 forest fragments in LD endemic areas of Connecticut, New Jersey and New York (the “tri-state study”) by measuring the species richness and relative densities of the major tick hosts and comparing corresponding model predictions to observed NIP in each fragment. In the longitudinal IES study, model predictions were significantly correlated with the observed NIP, accurate within 7 percentage points on average. In the tri-state study, model predictions were also significantly correlated with observed NIP although the relationship was less pronounced. Neither species richness nor the Shannon diversity index was strongly correlated with NIP, implying that a more complex definition of diversity such as that used in our model is necessary to describe this system. The results of these studies demonstrate that a simple empirically parameterized model based on the dilution effect principle can reasonably predict relative entomological risk across space and time.