Predicting short-life-cycle species suitability via dynamic species distribution models: Applications in dengue and...

Species Distribution Models (SDMs) are used to estimate environmental suitability and the potential distribution of organisms based on occurrence records and climatic variables. However, traditional approaches rely on long-term climatic averages, producing static predictions with no ability to capture seasonal variations that are particularly relevant for short life-cycle species, such as disease vectors. In this study, we developed Dynamic Species Distribution Models (DSDMs) for Ae. aegypti and An. darlingi by integrating monthly occurrence records and corresponding climatic variables from 2016 to 2022, using the MaxEnt algorithm. As an application in spatial epidemiology, we estimated the Incidence Rate Ratio (IRR) of dengue and malaria in Brazil, for regions and states, defining exposed populations as those located in areas with high climatic suitability, predicted by the DSDMs. The models showed satisfactory performance (Ae. aegypti: AUC=0.8, TSS=0.6; An. darlingi: AUC=0.9, TSS=0.8) and identified vapor pressure (vap) as the most influential climatic predictor for both vectors. For Ae. aegypti, climatic suitability was distributed across all regions, with strong seasonality and higher values concentrated between spring and early summer. IRR maps indicated higher incidence of dengue in the Southeast and South and significant associations between suitability, case counts, and IRR (p < 0.01). For An. darlingi, climatic suitability was concentrated in the Northern region, particularly within the Amazon Basin, reflecting the vector's dependence on humid environments. Our results demonstrate that incorporating temporal dynamics into SDMs enables the identification of seasonal windows of increased vulnerability to transmission, supporting more targeted surveillance and control strategies under increasing climatic variability.