Re-emergence of Aedes aegypti (Linnaeus) in Egypt under climate changes

Aedes aegypti, the primary vector of several medically significant arboviruses, including dengue fever, yellow fever, chikungunya, and Zika viruses, was successfully eradicated from Egypt in 1963. However, reports of its re-emergence and associated dengue outbreaks in southern Egyptian governorates since 2011 have raised concerns. This study aimed to model the current and future distribution of Ae. aegypti in Egypt. Locally collected occurrence data were combined with bioclimatic, anthropogenic, and biological environmental variables to identify key factors driving the distribution of Ae. aegypti. The modeling of maximum entropy (MaxEnt) showed good performance (AUC mean = 0.975; TSS mean = 0.789) and identified the density of the human population, the annual precipitation and the normalized difference vegetation index (NDVI) as key determinants of the habitat suitability of Ae. aegypti. The present-day predictions highlight the Nile Valley, Nile Delta, Fayoum Basin, Red Sea coast, and South Sinai as suitable habitats. The model projects a potential range expansion for Ae. aegypti under future climate change scenarios, particularly in the Nile Delta region. This expansion is expected to increase the suitable area for Ae. aegypti by 61-68% by 2050 and 64-69% by 2070, depending on the climate scenarios. These findings can support decision-making regarding vector control and disease prevention strategies to protect both local populations and international travelers; in particular, Egypt is one of the world’s most important tourist destinations. The study considers bioclimatic, anthropogenic, and biological environmental variables to understand Aedes aegypti habitat. Bioclimatic variables: Nineteen bioclimatic variables were sourced from the WorldClim database. To avoid the influence of highly correlated variables, the researchers conducted a principal component analysis and selected three key variables (annual precipitation, mean temperature of the warmest quarter, and temperature seasonality). Anthropogenic variable: Human population density data was obtained from the WorldPop dataset through the Google Earth Engine platform, using the median range from 2000 to 2020. Biological environmental variable: Normalized difference vegetation index (NDVI) data, representing vegetation cover, was acquired from the Moderate Resolution Imaging Spectroradiometer (MODIS) MOD13A2 product on the Google Earth Engine platform, calculating the median range from 2000 to 2015. Future Projections: To predict the impact of future changes, the researchers used five general circulation models (GCMs) from the WorldClim database and projected human population density data from Figshare for the years 2050 and 2070. They used two emission scenarios: a low emission scenario (ssp126; SSP1-RCP2.6) and a high emission scenario (ssp585; SSP5-RCP8.5). For future NDVI projections, the median NDVI range from 2000 to 2015 served as a proxy for the present, while the median NDVI for 2020 acted as a proxy for the future. Habitat suitability model: The researchers used MaxEnt software (version 3.4.4) to predict the current and future distribution of Aedes aegypti. MaxEnt is a modeling technique that uses presence-only data and is effective for projecting species distribution shifts under climate change. The model used 75% of the occurrence records as training data and 25% as test data. Tenfold bootstrapped replicates were used with a random seed to enhance model performance. The model's performance was evaluated using the area under the receiver operating characteristic curve (AUC) and the true skill statistic (TSS). To summarize the model results for current conditions, mean values across all runs were used. For future conditions, the mean across all means of all GCMs for each SSP scenario was calculated. The uncertainty index of the model predictions was estimated using the range (maximum-minimum) of predictions from ten replicate MaxEnt runs for current conditions and the range across all future model combinations within each SSP for future conditions. # Re-emergence of Aedes aegypti (Linnaeus) in Egypt under climate changes ## General Information Created: December 2024 Authors: * Mustafa M. Soliman ([msoliman@cu.edu.eg](mailto:msoliman@cu.edu.eg)) | ORCID: 0000-0002-0905-4473 * Abdallah M. Samy ([samy@sci.asu.edu.eg](mailto:samy@sci.asu.edu.eg)) | ORCID: 0000-0003-3978-1134 * Magdi S. El-Hawagry ([elhawagry@cu.edu.eg](mailto:elhawagry@cu.edu.eg)) | ORCID: 0000-0001-9162-5265 ## Dataset Description This dataset supports research modeling the current and future distribution of Aedes aegypti mosquitoes in Egypt using MaxEnt. The study combines locally collected data with bioclimatic, anthropogenic, and biological variables to identify key environmental factors influencing mosquito distribution. ### Data Collection * 48 Ae. aegypti occurrence records spanning 1925-2023 * Collected from nine Egyptian governorates: * Alexandria * Assiut * Aswan * Beni Suef * Cairo * Fayoum * Minya * Qena * Red Sea ### Data Sources 1. Bioclimatic Variables: * Nineteen variables from WorldClim database 2. Anthropogenic Data: * Human population density from WorldPop * Future population projections from Figshare (2050 and 2070) 3. Biological Data: * Normalized difference vegetation index (NDVI) from MODIS MOD13A2 product (Google Earth Engine platform) 4. Climate Projections: * Five general circulation models (GCMs) from WorldClim * Projection years: 2050 and 2070 * Emission scenarios: * Low emission (ssp126; SSP1-RCP2.6) * High emission (ssp585; SSP5-RCP8.5) ## File Structure ├── Aedes_rarefied_points.csv # Occurrence records data ├── Layers_current/ # Current bioclimatic layers │ └── *.asc # ASCII format climate data ├── Future_layers/ # Projected climate scenarios │ └── *.asc # ASCII format projected data └── Bias_file_rarefy_200/ # Bias files for MaxEnt modeling ## File Formats * Occurrence data: CSV format * Environmental layers: ASCII (.asc) format * All spatial data are clipped to the study area ## Usage Notes * The occurrence records were compiled from previous literature * All environmental layers are spatially aligned and processed for use in MaxEnt modeling * Bias files are included to account for sampling bias in the modeling process