Can highly virulent disease persist without evolving lower levels of virulence? We use a suit of models to address this question in several systems. We work with plague in prairie dogs, rabies in bats and avian influenza in waterfowl. In an NSF funded project on plague, we are studying the relative importance of different modes of transmission within prairie dog towns (Webb et al. PNAS 2006), plague persistence at the landscape level, and the evolution of resistance to plague in prairie dog hosts. In bat rabies, we are studying how hibernation impacts rabies persistence and factors that predict disease dynamics. In a USDA funded project on avian influenza, we are studying how migratory dynamics impacts the prevalence of the disease. We use traditional dynamical systems models, metapopulation models, frequentist and Bayesian statistical approaches and computer simulation in addressing these questions.

The response of biodiversity to a changing environment very likely constrains the way ecosystem function responds to environmental change.  The main goal of this NSF funded project is to incorporate the dynamic response of functional diversity into our understanding of how diversity impacts the stability and resilience responses of ecosystems to environmental change.  Among candidate theoretical approaches for achieving this goal, trait-based modeling approaches are particularly well suited, given their inherent structure for simultaneously describing the relationship between functional diversity and ecosystem function and the response of functional diversity to temporal and spatial variability in the environment.  The most well-developed of these trait-based approaches (Norberg et al. PNAS 2001) aggregates inter- and intra-specific variability in functional traits into a biomass distribution of the functional trait.  Using a framework similar to quantitative genetics, environmental drivers produce changes in the trait distribution which in turn translate into changes in ecosystem function.  A weakness of current versions of trait-based approaches is that they deal with only a single trait and environmental driver, while real ecosystems are characterized by organisms with multiple, correlated traits responding to multiple, correlated environmental drivers.  Together with my collaborators (J. Norberg, Stockholm Univ., V. Savage, Harvard), we are extending Norberg et al.’s analytic trait-based framework to multiple, correlated traits and environmental drivers (Savage et al. 2007).  Because of the promise of trait-based approaches in addressing long-term ecosystem responses, we are also testing the predictive ability of trait-based approaches using long-term biodiversity studies conducted at multiple LTERs.  We are currently supported by grants from NSF and the Swedish National Research Council.