Forecasting invasive mosquito abundance in the Basque Country, Spain using machine learning techniques

Area of Science:

• Environmental Science
• Epidemiology
• Vector Ecology

Background:

• Mosquito-borne diseases pose increasing global health risks due to climate change and species expansion into new territories.
• Invasive mosquitoes in the Basque Country, Spain, heighten the risk of local transmission of diseases like dengue, Zika, and chikungunya.
• Public health systems face challenges managing invasive mosquito populations and potential disease outbreaks in non-endemic areas.

Purpose of the Study:

• To forecast the abundance of invasive Aedes mosquitoes in the Basque Country using machine learning.
• To identify key weather variables influencing mosquito populations.
• To evaluate the predictive accuracy of different forecasting models.

Main Methods:

• Utilized machine learning models (Random Forest, SARIMAX) to predict Aedes mosquito abundance based on egg counts.
• Analyzed the relationship between weather variables (temperature, precipitation, humidity) and mosquito egg counts, including lagged variables.
• Assessed model performance using Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE).

Main Results:

• Temperature, precipitation, and humidity significantly impact mosquito egg abundance.
• The Random Forest model achieved the highest accuracy in forecasting mosquito abundance.
• Including lagged climate and egg count data improved prediction accuracy.

Conclusions:

• Climate-driven forecasting tools are crucial for predicting mosquito abundance in expanding regions.
• Continuous entomological surveillance is essential for refining mosquito spread forecasts.
• These tools support the development and evaluation of effective vector control strategies.