Emergence of vector-borne diseases in urban environments is a leading public health problem throughout the world. Disease dynamics in the built environment depend on complex interactions between local weather, host abundance, proximity of major disease reservoirs and human intervention within a highly modified and structured space. We speculate that this structure mediates the spread of emerging pathogens. Particularly, we hypothesize that the heterogeneity of urbanized landscapes results in percolation-like patterns of spread rather than traveling waves. West Nile virus is a vector-borne disease of particular concern in the western hemisphere and a model for other vector-borne pathogens (e.g., the four virus serotypes that cause dengue fever and chikungunya virus). The epicenter of West Nile virus emergence in North America is New York City. We tested our hypothesis with data collected by the New York City Department of Health on the spatial distribution and chronosequence of infections of mosquitoes, birds, and humans; mosquito control; and environmental variables.  
Results/Conclusions

Demonstration of concept was performed with a series of analytical and simulation models, tailored to reflect the putative mechanisms dominating West Nile virus transmission in New York City. These models revealed dynamical signatures that are diagnostic of percolation-like spread. Statistical analyses detected these signatures in observations of West Nile infection in New York City in 8 out of 9 years. The spatial distribution of West Nile virus positive dead birds indicated that transmission was disproportionately associated with particular land cover types. Percolation models suggest that there exists a critical fraction of the environment that must be exceeded by transmission promoting land cover types for large-scale propagation of the pathogen to occur. Geographical analysis suggests that New York City is located at this transition.