key: cord-336071-t7c0drft authors: Chiyomaru, Katsumi; Takemoto, Kazuhiro title: Global COVID-19 transmission rate is influenced by precipitation seasonality and the speed of climate temperature warming date: 2020-04-14 journal: nan DOI: 10.1101/2020.04.10.20060459 sha: doc_id: 336071 cord_uid: t7c0drft The novel coronavirus disease 2019 (COVID-19) became a rapidly spreading worldwide epidemic; thus, it is a global priority to reduce the speed of the epidemic spreading. Several studies predicted that high temperature and humidity could reduce COVID-19 transmission. However, exceptions exist to this observation, further thorough examinations are thus needed for their confirmation. In this study, therefore, we used a global dataset of COVID-19 cases and global climate databases and comprehensively investigated how climate parameters could contribute to the growth rate of COVID-19 cases while statistically controlling for potential confounding effects using spatial analysis. We also confirmed that the growth rate decreased with the temperature; however, the growth rate was affected by precipitation seasonality and warming velocity rather than temperature. In particular, a lower growth rate was observed for a higher precipitation seasonality and lower warming velocity. These effects were independent of population density, human life quality, and travel restrictions. The results indicate that the temperature effect is less important compared to these intrinsic climate characteristics, which might thus be useful for explaining the exceptions. However, the contributions of the climate parameters to the growth rate were moderate; rather, the contribution of travel restrictions in each country was more significant. Although our findings are preliminary owing to data-analysis limitations, they may be helpful when predicting COVID-19 transmission. The novel coronavirus disease 2019 (COVID-19) became a rapidly spreading worldwide 11 epidemic; thus, it is a global priority to reduce the speed of the epidemic spreading. 12 Several studies predicted that high temperature and humidity could reduce COVID-19 13 transmission. However, exceptions exist to this observation, further thorough 14 examinations are thus needed for their confirmation. In this study, therefore, we used a 15 global dataset of COVID-19 cases and global climate databases and comprehensively 16 investigated how climate parameters could contribute to the growth rate of COVID-19 17 cases while statistically controlling for potential confounding effects using spatial 18 analysis. We also confirmed that the growth rate decreased with the temperature; 19 however, the growth rate was affected by precipitation seasonality and warming velocity 20 rather than temperature. In particular, a lower growth rate was observed for a higher 21 precipitation seasonality and lower warming velocity. These effects were independent of 22 population density, human life quality, and travel restrictions. The results indicate that the 23 temperature effect is less important compared to these intrinsic climate characteristics, 24 which might thus be useful for explaining the exceptions. However, the contributions of 25 the climate parameters to the growth rate were moderate; rather, the contribution of travel 26 restrictions in each country was more significant. Although our findings are preliminary 27 owing to data-analysis limitations, they may be helpful when predicting COVID-19 28 transmission. 29 The world-wide spreading of coronavirus disease 2019 (COVID-19) [ when 30 and more cases were confirmed in cumulative counts, as described previously 95 [7]. We confirmed that similar conclusions were obtained at the different cut-off values 96 (using the data within 30 days starting from the date when 50 and more cases were 97 confirmed). 98 We obtained climate parameters from several databases based on the observation area 100 latitudes and longitudes available in the dataset [1] . was also considered to remove spatial autocorrelation in the regression residuals. 166 Specifically, the Moran eigenvector approach was adopted using the function 167 SpatialFiltering in the R package spatialreg (version 1.1.5). As with the OLS regression 168 analysis, full models were constructed, and then the best model was selected based on 169 AICc values. The spatial filter was fixed in the model-selection procedures [33] . 170 The contribution (i.e., non-zero estimate) of each explanatory variable to the growth rate 171 of COVID-19 cases was considered significant when the associated p-value was less than 172 0.05. 173 The data in 300 areas were investigated (Figure 1 ). The OLS regression analysis (Table 1) 175 and spatial analysis ( and Argentina were warm in March; however, they show low precipitation seasonality 214 ( Figure 2) . Thus, the spreads might occur in these areas. Moreover, Europe and the USA 215 might have undergone rapid spreads because they show low precipitation seasonality; on 216 the other hand, the spread might have reached a peak relatively quickly in China because 217 of relatively high precipitation seasonality. 218 The contribution of solar radiation is currently ambiguous. Solar radiation showed a 219 positive association with the growth rate of COVID-19 cases. However, the results were 220 less robust; in particular, the contribution was statistically significant in spatial analysis 221 (Table 2 ), but not in the full and averaged models in the OLS regression (Table 1) . Thus, 222 it remains possible that the contributions partly observed in the analyses are artefacts. 223 Assuming the positive association, the result is inconsistent with the fact that solar (UV) 224 radiation is expected to reduce infection disease (e.g., influenza) transmission [13]. 225 Moreover, a pairwise correlation analysis showed no association between the growth rate 226 and solar radiation (Spearman's rank correlation coefficient r = -0.06, p = 0.31). 227 The contributions of wind speed and precipitation were also limited. This is inconsistent 228 with previous studies [8, 9] ; however, statistical significances were not evaluated in these 229 studies. This discrepancy might be due to differences in the data analyses between this 230 study and previous studies. In particular, previous studies used the measures based on the 231 number of confirmed cases; however, these measures may be affected by the difference 232 of COVID-19 testing between areas. Hence, further examinations may be needed, given 233 the importance of these climate parameters in infectious disease transmission [13, 17] . 234 Non-climate parameters were also associated with the growth rate of According to the estimates of the models of the OLS regression analysis and spatial 236 analysis, the contribution of travel restrictions was most significant than those of the 237 climate parameters; in particular, travel restrictions showed a negative association with 238 the growth rate. This result may be an extension of the result that the reduction of 239 Tables Table 1. Influence of explanatory variables on the growth rate of COVID-19 cases based on the ordinary least squared regression approach. The results of the full model, best model, and averaged model are shown, respectively. The abbreviations of variables are as follows: Tmean (monthly mean temperature), DTR (monthly diurnal temperature range), Tseasonality (temperature seasonality), Pseasonality (precipitation seasonality), UV (monthly solar radiation index), WV (warming velocity), PD (population density), HDI (human development index), and Ban (travel restrictions). R 2 denotes the coefficient of determination for full and best models based on the OLS regression. SE is the standard error. Values in brackets are the associated p-values. Full . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.10.20060459 doi: medRxiv preprint An interactive web-based dashboard to track COVID-19 in 285 real time Clinical features of patients 287 infected with 2019 novel coronavirus in Wuhan Why inequality could spread COVID-19 High temperature and high humidity reduce the 292 transmission of COVID-19 The impact of temperature and 294 absolute humidity on the coronavirus disease 2019 (COVID-19) outbreak -evidence 295 from China Role of temperature and humidity 297 in the modulation of the doubling time of COVID-19 cases Temperature dependence of COVID-19 transmission Roles of meteorological 302 conditions in COVID-19 transmission on a worldwide scale Spread of SARS-CoV-2 Coronavirus likely to be constrained by 305 climate Preliminary evidence that 307 higher temperatures are associated with lower incidence of COVID-19, for cases reported 308 globally up to 29th Temperature significant 311 change COVID-19 Transmission in 429 cities Influenza seasonality: Lifting the fog Environmental Role in Influenza Virus Outbreaks Effects of temperature, humidity, 318 and diurnal temperature range on influenza incidence in a temperate region. Influenza 319 Other Respi Viruses Urbanization and humidity shape the intensity of influenza epidemics in U.S. cities Changing rapid weather variability 324 increases influenza epidemic risk in a warming climate Seasonality and 327 the dynamics of infectious diseases Human impacts and climate change influence nestedness 330 and modularity in food-web and mutualistic networks Decomposing the effects of ocean environments on 333 predator-prey body-size relationships in food webs Network resilience of mutualistic ecosystems and 336 environmental changes: an empirical study 339 Methods to account for spatial autocorrelation in the analysis of species distributional 340 data: a review The 343 effect of human mobility and control measures on the COVID-19 epidemic in China An investigation of 346 transmission control measures during the first 50 days of the COVID-19 epidemic Epidemic curves made 349 easy using the R package incidence Global COVID-19 transmission rate is influenced by 352 precipitation seasonality and the speed of climate temperature warming Very high resolution 355 interpolated climate surfaces for global land areas Macroecological 358 trends in nestedness and modularity of seed-dispersal networks: human impact matters The 361 influence of late Quaternary climate-change velocity on species endemism 364 Population Density, Revision 11 The 366 human footprint and the last of the wild Gridded global datasets for Gross Domestic 369 Product and Human Development Index over 1990-2015. Sci Data Model selection and information theory in 375 geographical ecology Age specificity of cases and attack rate of novel 377 coronavirus disease (COVID-19) The authors would like to thank Editage (www.editage.com) for English language 281 editing. 282 283