Analysis and prediction of COVID-19 epidemic characteristics in Brazil based on meteorological synergy

Since the outbreak of COVID-19 (December 2019), this disaster has caused a lot of death cases and economic losses. Climate is one of the most important issues of virus spreading. This research explored the relationship between climate factors and the COVID-19 cases in Brazil using Spearman’s analysis and the Generalized Additive Models. The results indicated that new cases were negatively correlated with air temperature, dew point temperature, wind speed, while positively correlated with air pressure. Also, the response of COVID-19 to meteorological factors can be divided into five types: monotonic rising, monotonic falling, single valley, single peak, and fluctuating. The main response patterns to air temperature and dew point temperature are monotone valley and monotone falling with nonlinear, and the main response modes to air pressure are monotone rising. In the single valley pattern, the threshold scale of the minimum air temperature was 18-23.5℃, and the threshold scale of dew point temperature was 10-21℃. Furthermore, this study constructed a COVID prediction model based on the synergistic effect of multiple meteorological factors, which could effectively forecast the epidemic spread over half-year in Brazil over more than half a year. According to this result, it is possible to make the prevention and forecast of COVID-19 by combining with the threshold range, which has the value of application for reducing the occurrence of diseases and the number of deaths.