We present results from a spatial model of raccoon rabies in which a latitudinal gradient in the seasonal reproduction of the host population resulted in an unexpected trend in the spatial synchronization of epidemics. Models of disease dynamics often assume that host demographic processes are constant across space. However, as a disease spreads it may encounter local host populations that are governed by different demographic processes that can in turn change epidemiological dynamics. Using our model, we consider a north-south gradient in the length of the seasonal parturition period of raccoons, the primary wildlife host for rabies virus in the eastern United States. A short birth pulse, associated with the relatively short northern growing season, characterizes northern populations. A longer birth pulse associated with the longer growing season characterizes southern populations. Our biological intuition was that the short northern birth pulse would tend to entrain local host dynamics and consequently entrain disease dynamics resulting in spatial synchronization of epidemics. In the absence of strong seasonal forcing we expected the local host dynamics in southern populations to proceed along locally determined trajectories resulting in spatially asynchronous disease dynamics. 

Results/Conclusions

Contrary to intuition and previous results from childhood diseases, the short northern birth pulse resulted in spatially asynchronous epidemics of raccoon rabies. On the other hand, weak seasonal forcing resulted in epidemics that were spatially synchronized. Results from our model for a continuum of birth pulses between the short northern and long southern birth pulses showed an increasing gradient in the spatial synchronization of rabies epidemics. The latitudinal gradient in the spatial synchronization of epidemics has potentially important consequences for the allocation of limited public health resources to the surveillance of wildlife diseases. In the case of rabies, surveillance in a northern population will require sampling at most or all locations in order to detect the local epidemics. Since epidemics in southern populations are spatially synchronous the epidemiological dynamics for an entire region can be assessed through surveillance at a few locations.