Possible role of meteorological variables in COVID-19 spread : A case study from a 1 subtropical monsoon country , Bangladesh 2

Possible role of meteorological variables in COVID-19 spread: A case study from a 1 subtropical monsoon country, Bangladesh 2 Neaz A. Hasan a, , Md. Sifat Siddik b 3 a Department of Aquaculture, Bangladesh Agricultural University, Mymensingh, Bangladesh 4 b Department of Irrigation and Water Management, Bangladesh Agricultural University, 5 Mymensingh, Bangladesh 6 7 *Corresponding author 8 Neaz A. Hasan 9 Department of Aquaculture, Bangladesh Agricultural University, Mymensingh, Bangladesh. 10 E-mail address: neaz41119@bau.edu.bd 11 Phone: +88-01785583848 12 13 Declaration of competing financial interests 14 The authors declare that they have no known competing financial interests or personal 15 relationships that could have appeared to influence the work reported in this paper. The funders 16 had no role in the design of the study; in the collection, analyses, or interpretation of data; in the 17 writing of the manuscript, or in the decision to publish the results. 18 19 20 21 22 23 Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 28 June 2020 doi:10.20944/preprints202006.0347.v1


Introduction 47
(Neher et al., 2020). Despite the dominant presence of the virus in temperate countries far from 93 the equator (Sajadi et al., 2020), the virus also resides in tropical countries such as Bangladesh. 94 Therefore, the present study aims to assess the correlation between COVID-19 transmission and 95 meteorological variables in a representative tropical country (Bangladesh) in an attempt to 96 address the following research questions: (i) do high temperatures and humidity at the onset of 97 summer curb COVID transmission and (ii) what are the effects of high wind speed from the 98 outset and periodically during summer? 99

Study area 101
Among the first three cases of COVID-19 in Bangladesh, two were from Narayanganj, which is 102 now considered the second largest epicenter (EC) in the country after the capital city, Dhaka. 103 Following these two districts, Gazipur is the closest peripheral city that has been declared a 104 vulnerable district of Bangladesh as it has the densest population and the largest industrial area 105 after Dhaka. These three areas have the most cases of COVID-19 since the initial outbreak. 106 Therefore, these three districts were selected as the representative EC area of Bangladesh to 107 examine the possible effects of climatic factors on the spread of COVID-19. For easy 108 comparison of these EC areas, some non-epicenters (NECs) were selected, namely Narsingdi, 109 Mymensingh, and Tangail (Fig. 1). A number of additional districts that are less vulnerable to 110 COVID-19 are also discussed; these were selected for their adjacency with the EC districts and 111 easy spatial presentation. The study was performed using two types of secondary data sources. Daily and district-wise data 119 for both ECs and NECs were collected from the daily COVID-19 national report published by 120 the IEDCR on behalf of the Director General of Health Services, Ministry of Health and Family 121 Welfare, GoB. Data relating to the main meteorological variables, including air temperature (at 2 122 m height), humidity, and wind speed, within the timeframe of March 8, 2020 to May 17, 2020, 123 were obtained from https://www.timeanddate.com at three-hour intervals. The values of the 124 measured meteorological variables were averaged daily before spatial analysis was carried out. The highest humidity in both ECs and NECs was observed on April 28, 2020. The daily cases 154 following this date were outside the pre-symptomatic period for both areas. Besides the highest 155 humidity date recorded, some peak humidity values were measured on scattered dates; however, 156 the highest daily cases in ECs or NECs were not observed in the following pre-symptomatic 157 period and/or dates. The average daily humidity values followed similar trends in both the ECs 158 and NECs; however, new cases rose or declined in an irregular pattern, which did not appear to 159 be related to trends in humidity (Fig. 3). The wind speed in the ECs and NECs followed different trends. On some particular dates (e.g., 165 April 18, 2020), when peak wind speed (6.61 km/h) was recorded in EC regions, NEC regions 166 experienced one of the lowest wind speed values (4.25 km/h). Following this date, ECs 167 experienced their daily highest number of recorded new cases on the pre-symptomatic date (May 168 1, 2020) than the previous date counts. In contrast, NECs experienced a drop in new cases on the 169 same pre-symptomatic date, compared to their previous date records (Fig. 4)

Spatial analogy of meteorological variables between ECs and NECs 177
The spatial analogy of all meteorological variables for both ECs and NECs was observed at 14 178 days intervals on six different dates over the entire study period. Between these dates, the highest 179 temperature (31.6°C) was recorded on the last study date (May 17, 2020) in all selected areas of 180 ECs, along with one area (Narsingdi) from NECs. From the date-wise temperature record, the

. Wind speed (km/h) and wind direction variation between the ECs and NECs on different 213
dates of the pre-symptomatic period following the date of detection of the first case. 214

Spearman rank and Kendall correlation test 218
Data regarding the COVID-19 cases and meteorological variables were included in the Spearman 219 rank and Kendall correlation test to obtain their specific correlation coefficient values [Spearman 220 correlation coefficient (SCC) and Kendall correlation coefficient (KCC)]. Daily new cases must 221 show a strong correlation with the total number of cases, as new cases are rapidly pushing the 222 tally of total cases each day. Therefore, the correlation between these two similar variables was 223 not included, rather these two variables were examined for their correlation with meteorological 224 variables. Among the meteorological variables, only average wind speed showed moderate 225 * Correlation is significant at the 0.05 level (two-tailed) 245

Discussion 246
The transmission factors of several members of the coronavirus family, including SARS-CoV-2 247 (Yang and Wang, 2020), remain unknown due to their highly variable characteristics (e.g., 248 spread via both water and air, enveloped in layers of fatty molecules or non-enveloped) 249 Our results regarding meteorological variables might be contradictory due to the different data 313 inclusion patterns adopted by different researchers. Some researchers considered new cases vis-314 à-vis meteorological variables, while others included the cumulative incidence rate in their 315 analysis. The asymmetry of population size and population density in different geographical 316 locations is also responsible for variable results regarding the same issue from country to country 317 (Jahangiri et al., 2020). Moreover, even though different countries follow the same strategies of 318 lockdown, testing-tracing-isolating of infected individuals, and social distancing, the degree of 319 enforcement varies. Non-reporting of facts during the initial cases of COVID-19 in various 320 countries has also skewed findings. and events. This warrants the necessity for social distancing. Proper social engineering needs to be applied to modify our behavior in lockdown to minimize the excessive transmission rate of 338 SARS-CoV-2. In addition, hand washing with soaps and sanitizers, avoiding touching the face, 339 using masks, practicing respiratory hygiene at a standard level, etc., can help to effectively 340 flatten the pandemic curve. 341 342 Acknowledgements 343 As part of the current epidemic situation, this study was carried out using the COVID-19 country 344 report of daily data published by the Institute of Epidemiology, Disease Control and Research 345 (IEDCR), Bangladesh. The authors are thankful to the IEDCR for providing free access to 346 download the data. Furthermore, the authors are grateful to timeanddate.com for the daily 347 weather data.