Impact of climate change on SARS-CoV-2 epidemic in China

respectively, and TP lagged 17. In the multifactor analysis, the mean lag days of the six weather factors were used to obtain a total lag of 11 days, and the validation showed that the lag order of the GR was consistent with the simultaneous consideration of the six weather factors. In the single-climate factor analysis, for every one unit increase in each factor, the GR decreased by 23.29±3.00% in 10 days (in H, P-value < 0.001), decreased by 0.56±0.08% in 9 days (in T, P-value < 0.001), increased by 0.67±0.31% in 17 days (in TP, P-value = 0.002 < 0.005), decreased by 0.17±0.03% in 8 days (in UV, P-value < 0.001), decreased by 12.01±0.02% in 11 days (in WS, P-value < 0.001), and increased by 4.81±1.02% in 11 days (in SP, P-value < 0.001). In the multiclimate factor analysis (in 11 days), the GR decreased by 16.61±20.28% for each 1% increase in H; 1.58±1.11% for each 1˚C increase in T; 0.90±0.50% for each 1 mm increase in the TP; 0.30±0.21% for each 1 kJ�m-2�h-1 increase in UV radiation; 16.47±2.93% for each 1 m�s-1 increase in the WS; and 15.02±1.63% increase in the GR for each 1 kPa increase in the SP. The model interpretability was 0.85 (P-value < 0.001) and the coefficients of each factor are shown in S3 Table. Among the six weather factors, H exerted the greatest effect (23.06%), followed by T (20.92%) and UV (18.65%). The effect of TP was negligible at 6.04%, and the correlation between TP and GR was minimal, indicating that the effect of TP was insignificant; the effect of each factor is shown in Fig 9. Discussion Spatial and temporal differences

Regional differences caused by weather. The first case of SARS-CoV-2 was discovered in Wuhan, Hubei Province, on December 8, 2019, and the first case of SARS was discovered in Fig 8. Distribution and ranking of weather sensitivity. Note: To determine weather sensitivity, principal component analysis was performed to calculate the comprehensive scores of six weather factors (UV, H, TP, WS, SP, and T), which were mapped to values of 1 to 100; subsequently, K-means clustering was used to classify them into three types (high, medium, and low). The comprehensive score (before mapping) included two sets, higher or lower than zero, which represent positive and negative directions, respectively. A value closer to zero significances less significant effect. https://doi.org/10.1371/journal.pone.0285179.g008 PLOS ONE Impact of climate change on SARS-CoV-2 epidemic in China PLOS ONE | https://doi.org/10.1371/journal.pone.0285179 July 27, 2023 14 / 21