Human-environment interactions impact Aedes aegypti abundance in an urban desert setting

Abstract

Background: The Sonoran Desert of southwestern North America is naturally hostile to Aedes aegypti mosquitoes, but their populations have thrived in Arizona cities since initially reported over two decades ago. With growing concern about the emergence and spread of Aedes-borne diseases like Zika, chikungunya, and dengue, there is a pressing need to understand the factors that promote Ae. aegypti survival and abundance. Though well studied in more tropical climates, little is known about these factors in arid regions. This study aimed to investigate the multi-level geographic factors that enable Ae. aegypti abundance in desert cities. Methods & Materials: The Maricopa County Environmental Services Vector Control Division collects mosquitoes weekly from ∼700 CO2-baited Encephalitis Vector Survey (EVS) traps distributed throughout Maricopa County, which includes the Phoenix metropolitan area. Weekly counts were reported from each year 2014 to 2016. We used 1m-resolution satellite images from the National Agricultural Imagery Program from 2013 and 2015 to classify microgeographic differences in land cover. We assessed land surface temperature, rainfall, normalized difference vegetation index (NDVI), land cover/use, and human demographic data from the U.S. Census as potential predictors of the presence and abundance of adult Ae. aegypti in each trap. Results: A total of 83,750 female adult Ae. aegypti mosquitoes were captured between 2014 and 2016. High-density traps consistently tended to be located in the north-central and southeastern parts of the county. The percent of positive traps and the average number of mosquitoes per trap peaked in October each year, following annual rains from July-September. Nearly all human socioeconomic status variables were inversely associated with the presence of Ae. aegypti mosquitoes in crude logistic regression, though the direction and strength of the associations with mosquito counts from negative binomial regression were less consistent. Further model fitting and analyses of geospatial relationships are underway. Conclusion: Desert urban residents may be susceptible to disease outbreaks where Ae. aegypti mosquitoes are abundant. Better knowledge of factors promoting mosquito survival and distribution in these settings will support both long- and short-term control efforts to prevent epidemic and endemic vector-borne disease transmission in arid habitats.