Modeling the Spatial Distribution of the Vectors of Dengue Fever in Iranusing the Maximum Entropy Model and Genetic Algorithm Publisher



Haghi S¹ ; Karimi M¹ ; Hanafibojd AA² Show Affiliations
Source: Iranian Journal of Remote Sensing and GIS Published:2024

Abstract

Introduction: Dengue fever is a contagious viral disease transmitted by two species of mosquitoes, Aedes aegypti and Aedes albopictus, which is spreading rapidly worldwide. The increase in global temperatures, climate change, rainfall patterns, and urbanization have significantly influenced the distribution of these species, creating new areas for their presence. Iran is considered vulnerable to these species, making it essential to determine their potential distribution ranges to implement effective population control programs. Habitat suitability models, which are algorithmic models, predict suitable spatial distributions for species establishment. The main objective of this study is to model the spatial distribution of dengue fever vectors in Iran using available global and Asian-level data due to the lack of sufficient local data on vectors. The study innovatively utilizes heterogeneity layers as an auxiliary factor for analyzing presence points and reducing spatial autocorrelation. It also compares two species distribution models based on presence data to select the optimal modeling approach. Materials and Methods: The models used in this study include the Maximum Entropy (MaxEnt) method and a genetic algorithm called GARP. These models can detect non-linear and influential relationships between species and environmental variables to develop prediction models. Necessary information layers include species presence points and independent environmental variable layers. 2,780 presence points were collected from various databases for both species (1,926 for Aedes aegypti and 854 for Aedes albopictus). To reduce spatial autocorrelation among dengue fever vector presence data, a heterogeneity layer of topography was created using the SDM toolbox in ArcGIS, removing points with similar elevation conditions from the modeling process. Principal Component Analysis (PCA) in ArcGIS was used to assess the correlation between biological variables, with variables showing correlations above 0.75 being excluded from the analysis. Population density, climate, vegetation cover density, elevation, and soil organic carbon variables were considered and incorporated into the model. Finally, habitat suitability was modeled globally with a spatial resolution of 5 kilometers for both species using the MaxEnt method and GARP. Results and Discussion: The area under the curve (AUC) values were calculated at 0.942 and 0.948 for Ae. aegypti and Ae. albopictus, respectively. The research results showed that the northern and southern provinces of Iran have higher habitat suitability for both species. However, Ae. aegypti has a higher distribution probability in the southern parts towards the east along the Omani seashore. In the implementation of the MaxEnt method for Ae. albopictus, the provinces in the west of Iran were also determined as favorable, although this was not correctly modeled on a smaller scale. In February 2019, a small number of eggs and adults of Ae. aegypti were discovered in Bandar-e-Lengeh city, which was exactly predicted by this research. Conclusion: The results of this study can be used in planning to manage the population of these vector insects and control the disease while monitoring the populations during epidemic seasons. © 2020, Shahid Beheshti University. All rights reserved.