Roles of meteorological conditions in COVID-19 transmission on a worldwide scale

Biqing Chen; Hao Liang; Xiaomin Yuan; Yingying Hu; Miao Xu; Yating Zhao; Binfen Zhang; Fang Tian; Xuejun Zhu

To elucidate the contribution of each meteorological factor to the case counts, we first performed single-factor non-linear regression modeling for each meteorological variable in the Wuhan dataset as well as in the discovery dataset. Temperature and wind speed were fitted into quadratic models; relative humidity was fitted into a cubic model; visibility was fitted into two models, an inverse model when modeling in the discovery dataset and a quadratic model when modeling in the Wuhan dataset because distribution of visibility in the two datasets was different. We used these fitted models to calculate a predicted value for case counts for each studied site, and then compared this predicted value with the real observed case counts by calculating a Pearson's correlation coefficient between them. Model fitting results showed that using the Wuhan dataset for single-factor modeling produced better model fitness. There was 0.40, 0.24, and 0.35 correlation between the observed data for Wuhan and values predicted by average air temperature, relative humidity, and visibility, separately, while wind speed alone could not explain much of the variance in confirmed case counts ( Figure 4 ). According to the equation, SARS-CoV2 transmission reaches a peak when the air temperature is 8.07 ℃, or when the wind speed is 16.1 mile/hr, or when the visibility is 2.99 statute miles to nearest tenth, or when the relative humidity is 64.6%.