Developing a hybrid model for early-warning against mosquito-borne diseases in Germany

Abstract text

Driven by globalisation and climate change, mosquito-borne viruses have emerged in the last decades, with Usutu virus (USUV) transmission first detected in Germany in 2010, and West Nile virus (WNV) in 2018. Mechanistic (e.g., epidemiological R₀-models) and correlative (e.g., environmental niche models, ENMs) modelling approaches are commonly used to evaluate the risk of transmission, but there is a gap in literature that integrates both streams for more informative results. Most studies also employ static data instead of regularly updated data, limiting their ability to make near-real time predictions and guide mosquito control measures. In this study, a "hybrid" model was developed in which estimates of the spatial-temporal abundance of mosquito populations generated from ENMs were used to refine a mechanistic R₀ model based on temperature-dependent and taxa-specific transmission parameters. Real-time mosquito surveillance data collected from traps across Germany provided an accurate estimate of vector abundance, and allowed deriving the vector-to-host ratio parameter fed into the model. Nation-wide climate data updated on an hourly basis also served as model input to yield short-term forecasts presented as risk maps, which can be easily interpreted and used as a tool for risk assessment. Results so far suggest that the role of Culex torrentium in WNV transmission may have been grossly underestimated, as its high vector competence for this virus generated high R₀ values. Integrative models that take advantage of regularly updated climate and mosquito surveillance data could be of great value to better guide decision-making by public health authorities regarding surveillance plans and preventive measures.