The Importance of Human Population Characteristics in Modeling Mosquito Vectors: A Comparative Analysis of Model Components

Document Type


Publication Date

November 2016


The current Zika virus epidemic in the Western hemisphere is representative of the confluence of global climate change and infectious disease expansion, and vector modeling represents a pertinent and timely method to analyze the environment associated with Zika-carrying mosquitoes. Among many mosquito species distribution models, there are varying opinions on which variables are most predictive and, consequently, should be included in modeling efforts. While climate variables (e.g.,mean temperature, mean precipitation) are routinely included, some argue that human population dynamics, in the form of population density and socioeconomic status, should also be included. This project aimed to test the importance of including human population characteristics by modelling the Zika virus vector Aedes aegypti in the Southeastern United States with climate variables, population density, and poverty characteristics. A. aegypti occurrences, global climate data, and population characteristics were obtained from publicly available sources and sampled at a resolution of 2.5 arc-minutes. Data pre and post-processing was completed in ArcMap 10.3 and models were created in Maxent v.3.3.3k. Four models were developed for this project: a climate-only model, a climate and population density model, a climate and poverty model, and a combined model with climate, population density, and poverty. Models were evaluated by comparing test and training area under the curve metrics, omission and commission errors, and variable jackknifing results. The climate-only model performed poorly compared to models with human population characteristics. The combined model was the best fit, though the model with climate and population density had a lower commission rate (21.0% and 20.6%, respectively). Jackknife results for the full model showed that population density was the most significant contributor to the model. This research indicates that more consideration should be given to human population characteristics when modelling mosquito habitats.


Atlanta, GA

Copyright Statement

Abstract is also available through the American Society of Tropical Medicine.

This document is currently not available here.