63 - Optimizing vector surveillance coverage using multivariate environmental similarity surface maps

Abstract: Surveillance is arguably the most important tool in combating the current and emerging public health threats posed by arthropod vectors. Data including where hosts, vectors, and pathogens co-occur allow for the development of cost effective and sustainable control efforts, and provide valuable scientific insights to agencies and researchers. Programs are increasingly benefitted by new trapping and monitoring technology, but in general having more surveillance sites will improve data collection and our ability to detect and address issues. This ideal comes into conflict with limitations on workforce that is common across the range of public health agencies, leading to a need to analyze how surveillance can be carried out most efficiently. One potential approach involves using the sites that best represent the diversity of environmental or sociological conditions contained in a heterogeneous landscape. This analysis is commonly conducted in species distribution modeling and conservation research, and employs a technique called multivariate environmental similarity surface (MESS). MESS maps depict how well a set of points (in this case trapping sites) represents the ranges and combinations of various environmental factors present across a landscape, creating a gradient ranging from highly similar to highly novel. To test this application, existing surveillance data was partitioned by trap and subjected to various MESS analyses using climate, land use, and sociological factors. Results revealed that selecting subsets of trapping sites that maximized overall environmental coverage improved extrapolated predictions from various models compared to randomly selected or spatially distinct sites, and also characterized the asymptotic relationship between the number of sites and overall coverage. Deploying this strategy when resources are limited will improve surveillance overall, and this approach is especially beneficial for detecting novel species and modeling vector populations. Future extensions of this analysis may be used when establishing new sites that maximize environmental coverage.