A Humidity-Driven Prediction System for Influenza Outbreaks

Abstract:

Recent studies have highlighted the role of absolute (or specific) humidity conditions as a leading explanation for the seasonal behavior of influenza outbreaks in temperate regions. If the timing and intensity of seasonal influenza outbreaks can be forecast, this would be of great value for public health response efforts. We have developed and implemented a SIRS (Susceptible-Infectious-Recovered-Susceptible) type numerical prediction system that is driven by specific humidity to predict influenza outbreaks. For the humidity, we have explored using both satellite data from the AIRS (Atmospheric Infrared Sounder) instrument as well as ERA-Interim re-analysis data. We discuss the development, testing, sensitivities and limitations of the prediction system and show results for influenza outbreaks in the United States during the years 2010-2014 (modeled in retrospect). Comparisons are made with other existing prediction systems and available data for influenza outbreaks from Google Flu Trends and the CDC (Center for Disease Control), and the incorporation of these datasets into the forecasting system is discussed.