You are here:
Ecosystems and spatiotemporal mosquito-borne disease models across a gradient of urbanization (ACES 12/03/18)
Citation:
Myer, M. AND JohnM Johnston. Ecosystems and spatiotemporal mosquito-borne disease models across a gradient of urbanization (ACES 12/03/18). 2018 ACEs Conference, Washington, DC, December 03 - 06, 2018.
Impact/Purpose:
Presented at 2018 ACES.
Description:
Vector-borne diseases are increasing in geographic extent and incidence in the United States and worldwide. The combined impacts of human-mediated ecological change and a changing climate provide new habitats for disease vectors, enable the rapid spread of infectious disease, and change the dynamic processes that govern interactions between vector, reservoir, host, and disease agent. Understanding those processes is an important part of efforts to identify, predict, and prevent vector-borne disease. We worked with three counties and municipalities to identify mosquito-borne disease surveillance datasets and develop models to determine indicators of disease incidence. In Suffolk County, New York and Nassau County, New York we obtained West Nile Virus surveillance data from 2008-2015 and 2001-2015 respectively, and in Brownsville, Texas we obtained Aedes aegypti mosquito trap data from 2017. Suffolk and Nassau Counties, on Long Island, New York, represent a gradient of development from rural to urban and allowed us to examine the differences in West Nile Virus indicators between those development types. In Brownsville, we adapted our modeling methods to a drastically different climate and densely urban area. We used the INLA SPDE (Integrated Nested Laplace Approxmations, Stochastic Partial Differential Equations) Bayesian method to fit spatiotemporal models of West Nile Virus incidence in Suffolk and Nassau counties, and of Aedes aegypti presence in Brownsville. These methods allowed us to identify spatial and temporal patterns and more accurately model our response. In each modeling context, we began with a substantial pool of remote-sensed and publicly available data including land use/land cover, vegetation indices, census data, and location-specific variables of interest. Our results indicated that in the more rural and exurban Suffolk County, on-site septic disposal was a significant predictor of West Nile incidence, while the presence of woody wetlands was associated with a reduction in disease incidence. Septic systems in the area are a major contributor to nitrogen pollution and eutrophication, which may have a connection to reduction in healthy wetland habitat. In suburban and urban Nassau County, we found that areas with both high vegetative index and high intensity development had lowered disease incidence, indicating that there is a ‘sweet-spot’ effect of intermediate development that contributes to West Nile spread in suburban areas. Older housing stock also had an association with increased West Nile incidence, paralleling results from studies in Chicago and Baltimore. Our investigation in Brownsville is ongoing, and preliminary results point to Aedes aegypti populations being associated with older housing stock, vegetation height, and degree of urban development. These studies highlight the differences in factors influencing mosquito-borne disease dynamics across a gradient of urbanization. In rural and exurban areas, disease incidence was affected by wetlands and the presence of septic systems which can adversely affect water quality, while suburban and urban areas provided ample habitat for disease-bearing mosquitoes in areas of medium-intensity development and old houses. We conclude that in urban areas, human development is a primary mediator of West Nile virus incidence, while ecosystem services deriving from healthy wetland and surface water habitats prevail in less-developed and rural areas.
