Early warning system of the seasonal west nile virus infection risk in humans in northern greece, 2020-2024.

This study introduces a novel forecasting tool for West Nile Virus (WNV) risk at the municipal level, based on a previously developed climate-dependent epidemiological model coupled with a data-driven model and field data in an ensemble framework. The modelling system was evaluated at the municipal level in the regions of Central Macedonia (RCM) and Thessaly (RTH) in Greece for the period 2020-2024 and classifies the municipalities according to the risk level of occurrence of a WNV human case (WNVhc) within each forecast season. The modelling system produces seasonal forecasts updated monthly throughout the entire mosquito breeding season (April-September). From the first forecast of the year, WNVhc were correctly predicted at 71% of the infected municipalities in RCM and 60% of the infected municipalities in RTH, percentages that increased by at least 20% after July when surveillance data were incorporated. The observed number of infected humans was within the predicted range, highlighting the part of the region with an expected outbreak. The approach shows that the coupling of a climate-dependent epidemiological model at local level with a data-driven model, incorporating entomological data to predict the risk of WNVhc, can be a useful tool in planning control strategies.