Examining the relationship between air masses and Aedes aegypti presence in US counties using explainable artificial intelligence

Abstract Aedes aegypti mosquitoes are primary vectors for viral diseases including Zika, Dengue, Chikungunya, and Yellow Fever. Understanding the climatic factors influencing their proliferation is crucial for targeted vector control strategies. This study investigates the relationship between air mass (AM) categories and the annual and seasonal presence of Aedes aegypti in Manatee County, Florida, using data spanning from 1979 to 2024 sourced from the Global Biodiversity Information Facility. We employed Spearman correlations and explainable artificial intelligence (XAI), specifically a gradient boosting classifier integrated with Shapley additive explanations (SHAP), to investigate these relationships. Our findings highlight a pronounced seasonal cycle, with mosquito presence peaking in summer months (July-August) and minimizing during winter (December-March). At the annual scale, significant correlations were identified between mosquito presence and specific air mass conditions: humid warm (HW) AMs positively correlated (ρ = 0.39), suggesting enhanced presence under humid and warm environments, whereas cold (C) AMs negatively correlated (ρ = −0.34), indicative of reduced presence under colder conditions. Seasonal analyses using SHAP and Spearman correlations revealed that during summer, dry cold (DC) and dry (D) AM categories significantly decrease mosquito populations, collectively accounting for 22% of the model’s explanatory power. Conversely, warm AM emerged as a critical positive predictor in summer, contributing to 20% of the model’s predictive capability. Further, DC air masses in summer negatively impacted autumn mosquito presence (9.2% SHAP contribution) possibly by suppressing conditions favorable for proliferation in the peak breeding summer months. Comparative analysis with other counties exhibiting distinct climatic profiles provides broader applicability. These insights guide mosquito-borne disease risk management, guiding strategic interventions tailored to regional climatic conditions.