Abstract
Seasonal influenza epidemics are associated with various meteorological factors. Recently absolute humidity (AH) has garnered attention, and some epidemiological studies show an association between AH and human influenza infection. However, they mainly analyzed weekly surveillance data, and daily data remains largely unexplored despite its potential benefits. In this study, we analyze daily influenza surveillance data using a distributed lag non-linear model to examine the association of AH with the number of influenza cases and the magnitude of the association. Additionally, we investigate how adjustment for seasonality and autocorrelation in the model affect results. All models used in the study showed a significant increase in the number of influenza cases as AH decreased, although the magnitude of the association differed substantially by model. Furthermore, we found that relative risk reached a peak at lag 10-14 with extremely low AH. To verify these findings, further analysis should be conducted using data from other locations.