key: cord-0823685-q3abp9o7
authors: Wu, Yu; Jing, Wenzhan; Liu, Jue; Ma, Qiuyue; Yuan, Jie; Wang, Yaping; Du, Min; Liu, Min
title: Effects of temperature and humidity on the daily new cases and new deaths of COVID-19 in 166 countries
date: 2020-04-28
journal: Sci Total Environ
DOI: 10.1016/j.scitotenv.2020.139051
sha: c3c0a8ba2dc4e9f7ca6e4152f3266a1616e1a63f
doc_id: 823685
cord_uid: q3abp9o7

Abstract The coronavirus disease 2019 (COVID-19) pandemic is the defining global health crisis of our time and the greatest challenge facing the world. Meteorological parameters are reportedly crucial factors affecting respiratory infectious disease epidemics; however, the effect of meteorological parameters on COVID-19 remains controversial. This study investigated the effects of temperature and relative humidity on daily new cases and daily new deaths of COVID-19, which has useful implications for policymakers and the public. Daily data on meteorological conditions, new cases and new deaths of COVID-19 were collected for 166 countries (excluding China) as of March 27, 2020. Log-linear generalized additive model was used to analyze the effects of temperature and relative humidity on daily new cases and daily new deaths of COVID-19, with potential confounders controlled for, including wind speed, median age of the national population, Global Health Security Index, Human Development Index and population density. Our findings revealed that temperature and relative humidity were both negatively related to daily new cases and deaths. A 1 °C increase in temperature was associated with a 3.08% (95% CI: 1.53%, 4.63%) reduction in daily new cases and a 1.19% (95% CI: 0.44%, 1.95%) reduction in daily new deaths, whereas a 1% increase in relative humidity was associated with a 0.85% (95% CI: 0.51%, 1.19%) reduction in daily new cases and a 0.51% (95% CI: 0.34%, 0.67%) reduction in daily new deaths. The results remained robust when different lag structures and the sensitivity analysis were used. These findings provide preliminary evidence that the COVID-19 pandemic may be partially suppressed with temperature and humidity increases. However, active measures must be taken to control the source of infection, block transmission and prevent further spread of COVID-19.

J o u r n a l P r e -p r o o f July 2003 (Tan et al., 2005; Ma et al., 2020) . A study in China that was based on case studies in Hong Kong, Guangzhou, Beijing, and Taiyuan indicated that the outbreaks of SARS were significantly associated with variations in temperature (Tan et al., 2005) .

Emerging laboratory and epidemiological data suggest that environmental conditions may affect the current COVID-19 pandemic (Jon et al., 2020) . A published laboratory study by Chin et al.(2020) reported that SARS-CoV-2 was highly stable at 4 °C but sensitive to heat. The virus survival time was shortened to 5 mins as the incubation temperature increased to 70 °C.

Epidemiological studies have explored the relationship between COVID-19 and meteorological parameters; however, findings are controversial (Ma et al., 2020; Xie and Zhu, 2020; Yao et al., 2020) . A study by Xie and Zhu (2020) reported that a 1 °C rise was associated with a 4.861% increase in the daily confirmed cases of COVID-19, when mean temperature (lag 0-14) was below 3 °C. A study by Ma et al.(2020) reported a positive association between daily deaths of COVID-19 and diurnal temperature range, and a negative association for relative humidity.

However, a study by Yao et al.(2020) reported that COVID-19 transmission did not exhibit an association with temperature in Chinese cities.

COVID-19 continues to spread globally, and a second wave of COVID-19 appears probable (Leung et al., 2020) 

Relative humidity is the ratio of the actual water vapor pressure to the saturation water vapor pressure at the prevailing temperature, calculated using the following equation:

where E (hPa) donates the vapor pressure of air at temperature t(°C). E w (hPa) donates the saturated vapor pressure of the pure horizontal liquid surface at dry bulb temperature t (°C). The dew point is the temperature at which air must be cooled to become saturated with water vapor.

J o u r n a l P r e -p r o o f where E 0 denotes saturation vapor pressure at a reference temperature T 0 (273.15 K) which equals 6.11 mb. A is a constant of 17.43 and B is a constant of 240.73. t (°C) is the actual temperature or dew point .

Five variables were included in the model as potential confounders: wind speed population density (https://worldpopulationreview.com/countries/countries-by-density/). Wind is a crucial factor in the transmission of respiratory infectious diseases and it may modulate the dynamics of various vectors and pathogens (Ellwanger and Chies, 2018). Median age of the national population is an indicator of population aging; the incidence of severe cases is higher in countries with higher levels of population aging (Verity et al., 2020) . GHSI is the first comprehensive assessment of global health security capabilities to be employed in 195 countries;

the 2019 GHSI report scored (out of 100) the country-level capacity for "early detection and reporting for epidemics of potential concern." HDI is a summary measure of average achievement In the present study, 166 countries with confirmed cases as of March 27, 2020 were included.

Descriptive analyses were performed for all the data. A log-linear GAM was used to analyze the associations between meteorological factors (temperature and relative humidity) and daily new cases and daily new deaths of COVID-19 (Samet et al., 2000) . First, the basic models were constructed, including meteorological factors (temperature and relative humidity). Second, the variables were controlled to adjust for regional variation, including wind speed, median age of the population, GHSI, and country. Third, the day-of-week and the penalized smoothing spline functions were incorporated to control the time trend and cycle. The core GAM equation was as follows:

where t is the day of the observation. Considering that the number of daily new cases or daily new deaths in some countries is 0, Y t is the number of daily new cases or daily new deaths on day t plus one. α is the intercept; β is the regression coefficient; X t is the weather variables on day t;

Wind is the wind speed; Country is a categorical variable for country; Median.Age is the median age of each countries' population; GHSI is the Global Health Security Index; DOW is a categorical variable indicating the date of the week; s() refers to the smoother, which is based on the penalized smoothing spline; Time is the date of the observation; df is the degree of freedom.

were then considered using single lag days (lag 0, 1, 2, 3). The cumulative effects of average exposure over multiple days were then assessed using additional analyses (lag 01, 02, 03) to control for the possible misalignment of a single lag day exposure.

All analyses were performed using R software (version 3.6.0) with the "mgcv" package (version 1.8-28). The results were expressed as percentage changes and 95% confidence intervals (CIs) in daily new cases and daily new deaths of COVID-19 associated with a 1 unit increase in weather variables. All tests were two-sided, and a value P < 0.05 was considered statistically significant.

Two other variables were included in the sensitivity analysis: HDI and population density. Because of the inconsistent outbreak time of COVID-19 in different countries, the transmission mechanism and transmission rate of COVID-19 may differ. Therefore, countries that first reported case relatively early and countries with more cases were selected to fit the core model to determine the stability of results.

J o u r n a l P r e -p r o o f countries. We selected countries that reported their first case over 10 days before data collection and countries with over 100 cumulative cases to fit the core model. Among countries with over 10 days since the first reported case, a 1 °C increase in temperature was associated with a 3.05% and 1.22% reduction in daily new cases and daily new deaths , respectively, and a 1% increase in relative humidity was associated with a 0.87% and 0.51% reduction in daily new cases and daily new deaths , respectively. In countries with over 100 cumulative cases, a 1 °C increase in temperature was associated with a 2.82% and 1.25% reduction in daily new cases and daily new deaths , respectively; a 1% increase in relative humidity was associated with a 0.86% and 0.53% reduction in daily new cases and daily new deaths , respectively. The results of the comparisons of these datasets with the total data were not significant. These results demonstrate a robust effect of temperature and relative humidity on daily new cases and new deaths of COVID-19 (Table 2 ).

The COVID-19 pandemic is the defining global health crisis of our time and the greatest challenge facing the world (United Nations Development Programme, 2020). Our findings revealed that temperature and humidity were inversely correlated with daily new cases and deaths of COVID-19. For every 1 °C increase in temperature, daily new cases of COVID-19 reduced by 3.08% (95% CI: 1.53%, 4.63%) and daily new deaths reduced by 1.19% (95% CI: 0.44%, 1.95%);

for every 1% increase in humidity, daily new cases of COVID-19 reduced by 0.85% (95% CI: 0.51%, 1.19%), and daily new deaths reduced by 0.51% (95% CI: 0.34%, 0.67%), according to analyses of 166 countries. After adjusting for potential factors and lag days, the negative relationships remained robust. Furthermore, we hypothesized that the effects of temperature and Journal Pre-proof J o u r n a l P r e -p r o o f humidity on daily cases and deaths of COVID-19 may not be evident in countries where the community transmission of COVID-19 had not occurred because the proportion of imported cases was high. Therefore, countries where the time of onset was fewer than 10 days prior, or countries with fewer than 100 cumulative confirmed cases were excluded in the sensitivity analysis. The model-fitting results remained stable.

Few studies have investigated the association of temperature and humidity with COVID-19

incidence and death rates. A preprint study, which investigated the total confirmed cases in 429 cities globally from January 20 to February 4, 2020 indicated that the cumulative number of confirmed cases reduced by 0.86 for every 1 °C increase in the minimum temperature of higher-temperature cities . Another preprint study by Bannister-Tyrrell et al.(2020) reported that as of February 29, 2020, average temperature increases of 1 °C were negatively correlated with the predicted number of cases worldwide (excluding Hubei Province).

Furthermore, a study reported that relative humidity was inversely related to daily deaths of COVID-19 (r = −0.32), with the largest reduction in lag 3 [−11.41% (95% CI: −19.68%,−2.29%)]

(Ma et al., 2020). These results accord with our findings.

Other studies have reported conflicting results. A study in 122 cities in China demonstrated that when the mean temperature (lag 0-14) was below 3 °C, the daily confirmed cases of COVID-19 increased by 4.861% (95% CI: 3.209, 6.513%) for every 1 °C rise in temperature (Xie and Zhu, 2020) . However, no correlations were observed when the mean temperature was above Relative humidity(%) -0.86% -1.25% -0.47% <0.01 -0.53% -0.73% -0.33% <0.01 Table 2 : Estimated changes with 95% confidence intervals in daily new cases and daily new deaths percentage change (%) associated with each 1 unit increase in temperature and relative humidity among countries with over 10 days since the first reported case and countries with over 100 cumulative cases

Correction, Derivation and Application of Saturated Water Vapor Pressure Empirical Formula. Meteorological,Hydrological and Marine Instruments(04)

Stability of SARS-CoV-2 in different environmental conditions. The Lancet Microbe

Epidemiology and clinical presentations of the four human coronaviruses 229E, HKU1, NL63, and OC43 detected over 3 years using a novel multiplex real-time PCR method

Severe acute respiratory syndrome-related coronavirus: The species and its viruses -a statement of the Coronavirus Study Group

Clinical Characteristics of Coronavirus Disease 2019 in China

Do weather conditions influence the transmission of the coronavirus (SARS-CoV-2)? Retrieved from

Low ambient humidity impairs barrier function and innate resistance against influenza infection

National Health Commission & State Administration of Traditional Chinese Medicine. Diagnosis and Treatment Protocol for Novel Coronavirus Pneumonia

Association between viral seasonality and meteorological factors

The National Morbidity, Mortality, and Air Pollution Study. Part I: Methods and methodologic issues

Absolute humidity and the seasonal onset of influenza in the continental United States

High Contagiousness and Rapid Spread of Severe Acute Respiratory Syndrome Coronavirus 2. Emerging Infectious Disease journal

Potential Factors Influencing Repeated SARS Outbreaks in China

An initial investigation of the association between the SARS outbreak and weather: with the view of the environmental temperature and its variation

Review of aerosol transmission of influenza A virus

Coronavirus Resource Center

Aerosol and Surface Stability of SARS-CoV-2 as Compared with SARS-CoV-1

Estimates of the severity of coronavirus disease 2019: a model-based analysis. The Lancet Infectious Diseases

J o u r n a l P r e -p r o o f e-media-briefing-on-covid

World Health Organization, 2020c. Director-General's opening remarks at the media briefing on COVID-19 -11

Facemask shortage and the novel coronavirus disease (COVID-19) outbreak: Reflections on public health measures

Association between ambient temperature and COVID-19 infection in 122 cities from China

No Association of COVID-19 transmission with temperature or UV radiation in Chinese cities

We are very grateful to Wei Li for guiding the use of the software.

Bannister-Tyrrell, M., Meyer, A., Faverjon, C., Cameron, A., 2020. Preliminary evidence that higher temperatures are associated with lower incidence of COVID-19, for cases reported globally up

The authors declare that they have no conflict of interest.