Predators may strongly suppress or spread epidemics, so it is important to understand when and how each outcome arises. For example, predators can suppress epidemics by culling prey populations and hence inhibiting their density-dependent transmission of parasites. Then again, predators can also spread disease through several other mechanisms. These mechanisms need to be evaluated to prevent major mistakes in disease control efforts (i.e., manipulation of predators). Here, we evaluate a ‘cascade fueling’ hypothesis for disease spread by predators. In this hypothesis, predators fuel disease, despite depressing host density, because the resource release (via trophic cascade) increases parasite production per host. Specifically, we investigate how midge predators, Chaoborus punctipennis, spread epidemics of a virulent fungal parasite, Metschnikowia bicuspidata in a dominant zooplankton grazer Daphnia dentifera. First, we detect signatures of cascade fueling in multi-year field survey of Midwestern freshwater lakes. Then, we evaluate predictions of a simple mathematical model with a mesocosm experiment. In the experiment, we track host densities, resources, and infection prevalence under three levels of midge predation. Together, our combination of field data, mathematical model, and mesocosm experiment introduces a general mechanism for predator-mediated disease spread through trophic cascades.
Our field data, model predictions, and experimental results confirmed the importance of predator cascade-fueled epidemics. In the field data, lakes with intense midge predation had high algal resources, high per-host parasite yields and experience large epidemics. In the dynamic model, predators triggered cascades by reducing host density. Subsequent increases in host’s resources stimulated parasite production from infected hosts. Net effect of predation spreads infection despite the simultaneous culling of hosts, if parasite production responds sensitively to elevated resource density. The mesocosm experiment confirms all three model predictions necessary to support cascade-fueled mechanism. First, midge predation depressed total host density. Second, treatments with midges had elevated resources. Third, at the end of experiment, both infection prevalence and density of infected hosts were higher in treatments with midges. However, infection reached peaks later, indicating a density-resource-parasite production lag inherent in cascade fueling. Since cascade-fueled feedbacks are realized with a delay, predation eventually ends up increasing the size of epidemics despite immediate reduction in host density and, often, disease suppression in the initial stages of epidemics. Overall, our study provides a general mechanism by which predators increase parasite burden in wildlife populations through strong indirect effects on resources.