Using machine learning to understand microgeographic determinants of the Zika vector, Aedes aegypti

PLOS ONE RESEARCH ARTICLE Using machine learning to understand microgeographic determinants of the Zika vector, Aedes aegypti Jagger Alexander ID1*, André Barretto Bruno Wilke ID2, Alejandro Mantero1, Chalmers Vasquez3, William Petrie3, Naresh Kumar1, John C. Beier1 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Alexander J, Wilke ABB, Mantero A, Vasquez C, Petrie W, Kumar N, et al. (2022) Using machine learning to understand microgeographic determinants of the Zika vector, Aedes aegypti. PLoS ONE 17(12): e0265472. https://doi.org/ 10.1371/journal.pone.0265472 Editor: Bing Xue, Institute for Advanced Sustainability Studies, GERMANY Received: March 1, 2022 Accepted: December 13, 2022 Published: December 30, 2022 Copyright: This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication. Data Availability Statement: The data underlying the results presented in the study are available from Miami-Dade County Mosquito Control (https://www.miamidade.gov/global/solidwaste/ mosquito/home.page). Funding: This research was supported by CDC (https://www.cdc.gov/) grant 1U01CK000510–05: Southeastern Regional Center of Excellence in Vector-Borne Diseases: The Gateway Program. CDC had no role in the design of the study and 1 University of Miami Department of Public Health, Miami, FL, United States of America, 2 Laboratory for Computational Epidemiology and Public Health, Department of Epidemiology and Biostatistics, Indiana University School of Public Health, Bloomington, IN, United States of America, 3 Miami-Dade County Mosquito Control Division, Miami, FL, United States of America * Abstract There are limited data on why the 2016 Zika outbreak in Miami-Dade County, Florida was confined to certain neighborhoods. In this research, Aedes aegypti, the primary vector of Zika virus, are studied to examine neighborhood-level differences in their population dynamics and underlying processes. Weekly mosquito data were acquired from the Miami-Dade County Mosquito Control Division from 2016 to 2020 from 172 traps deployed around Miami-Dade County. Using random forest, a machine learning method, predictive models of spatiotemporal dynamics of Ae. aegypti in response to meteorological conditions and neighborhood-specific socio-demographic and physical characteristics, such as land-use and land-cover type and income level, were created. The study area was divided into two groups: areas affected by local transmission of Zika during the 2016 outbreak and unaffected areas. Ae. aegypti populations in areas affected by Zika were more strongly influenced by 14- and 21-day lagged weather conditions. In the unaffected areas, mosquito populations were more strongly influenced by land-use and day-of-collection weather conditions. There are neighborhood-scale differences in Ae. aegypti population dynamics. These differences in turn influence vector-borne disease diffusion in a region. These results have implications for vector control experts to lead neighborhood-specific vector control strategies and for epidemiologists to guide vector-borne disease risk preparations, especially for containing the spread of vector-borne disease in response to ongoing climate change. Introduction The Zika virus was the culprit for a recent public health emergency, with over 80,000 reported cases in 2016 alone [1]. The disease associated with Zika virus in humans generally presents mild symptoms lasting for up to one week [2]. However, the virus is of particular risk to pregnant women, whose children have a greater risk of birth defects [3]. Though case numbers PLOS ONE | https://doi.org/10.1371/journal.pone.0265472 December 30, 2022 1 / 16 PLOS ONE collection, analysis, and interpretation of data and in writing the manuscript. Competing interests: The authors have declared that no competing interests exist. Using machine learning to understand microgeographic determinants of the Zika vector, Aedes aegypti have declined since 2016, there are still questions remaining about the rapid global outbreak [4]. In June of 2016, Miami-Dade was the site of the first confirmed Zika virus disease cases in the United States [5, 6]. Unique variations were observed across Miami-Dade County in the spread of the Zika virus. Zika is a vector-borne disease, primarily spread during the bloodmeal of female Aedes mosquitoes. In the United States, Aedes aegypti is the most important mosquito species for the transmission of the Zika virus [7]. Local transmission was generally confined to certain regions across the county and was stopped after mosquito control through the implementation of aerosol insecticides [5, 6]. According to our best knowledge, there are currently no studies examining why the 2016 Zika virus outbreak in South Florida was concentrated in select areas. It is of particular significance to understand the spread of vector-borne disease in Miami-Dade County, as vector mosquito populations in Miami-Dade County may serve as an entry point for vector-borne diseases into the continental United States [8]. Variations in vector mosquito abundance and vector-borne disease spread have been modeled for decades, especially with relation to meteorologic or climatic variables [9–15]. Studies across different continents and habitat types show significant correlations between vector mosquito abundance and temperature, humidity, and precipitation. Though some mechanisms for these variations remain unresolved, some proposed mechanisms include different activity levels and reproductive rates in different temperatures and humidities, different availability of aquatic larval habitats with precipitation, and changing interspecies interactions with weather conditions [8]. Vector species abundance also varies with microgeography, or on the scale of an individual organism’s travel distance, as different habitats offer different ecology, resource availability, and larval habitat potential [16, 17]. Microgeography is especially important in urban environments as a species’ travelling range, which for Ae. aegypti is already short (on the order of 100 meters), is shortened by the presence of manmade barriers [18]. However, urban environments have unique properties that can often support vector species [19–21]. In Miami-Dade County, ongoing research has shown that features of the urban environment such as tire shops [22], cemeteries [23], urban farms [24], and ornamental bromeliads [25] all provide aquatic habitats for mosquito development [26]. A significant reason why these habitats are able to support Ae. aegypti is due to their relationship with weather and climate; water is a determining factor in the physical environment and is necessary for them to breed, while temperature inherently affects mosquito (...truncated)