Impact and Correlation of Air Quality and Climate Variables with COVID-19 Morbidity and Mortality in Dhaka, Bangladesh

The COVID-19 pandemic unexpectedly stopped the steady life and enhanced environmental quality. To apprehend the transmission of COVID-19 and the improvement of air quality, we have studied air quality indicators (PM2.5, PM10, AQI, and NO2), CO2 emission, and climate variables (temperature, relative humidity, rainfall, and wind velocity) in the extremely polluted and densely populated Southeast Asian megacity Dhaka, Bangladesh from March to June 2020. The Kendall and Spearman correlations were chosen to test the connotation of air quality and climate variables with COVID-19 morbidity and mortality. The average concentrations of PM2.5, PM10, and CO2 were 65.0 with STD of 37.9 and 87.1 with STD of 52.8 microgram m-3, and 427 with STD of 11.8 ppm, respectively. The average PM2.5 and PM10 drastically reduced up to 62% during COVID-19 lockdown in Dhaka comparing with March 2020 (before lockdown). Comparing with the same period in 2019, PM2.5 reduced up to 33.5%. The average NO2 concentration was 35.0 micromol m-2 during the lockdown period in April, whereas 175.0 micromol m-2 during March (before lockdown). A significant correlation was observed between COVID-19 cases and air quality indicators. A strong correlation was obtained between climate variables and the total number of COVID-19 morbidity and mortality representing a favorable condition for spreading the virus. Our study will be very expedient for policymakers to establish a mechanism for air pollution mitigation based on scientific substantiation, and also will be an essential reference for the advance research to improve urban air quality and the transmission of the SARS-CoV-2 virus in the tropical nations.

1 1 1 2 1 3 *Corresponding author: +8801817061160; asalam@gmail.com; asalam@du.ac.bd 1 4 1 5 1 6 1 7 1 8 1 9 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 . CC-BY 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 (which was not certified by peer review) preprint The copyright holder for this this version posted September 13, 2020. . https://doi.org/10.1101/2020 1 4 0 This study includes two stages of analysis: 1) to assess the effect of lockdown on air quality 1 4 1 indicators (PM 2.5 and PM 10 ), CO 2 emission, and 2) to study the correlation of climate variables 1 4 2 (average temperature, relative humidity, rainfall, and wind velocity) and air quality parameters with to examine the correlation between variables and air quality parameters. Air quality data of Dhaka Dhaka. Data set for COVID-19 was taken from the COVID-19 archive from the Directorate General humidity, wind speed, and rainfall) was taken from weatherforyou.com and darksky.net. Overview of the Air Quality and Climate variables 1 5 2 . CC-BY 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 certified by peer review) preprint The copyright holder for this this version posted September 13, 2020. . https://doi.org/10.1101/2020 8 The summary of PM 2.5 , PM 10 , CO 2 , temperature, relative humidity, wind velocity, and rainfall 1 5 3 between March 8 (1st day of confirmed COVID-9 cases in Bangladesh) and June 16, 2020, in 1 5 4 Dhaka megacity has been given in Table 1. The total average PM 2.5 and PM 10 were 65.0 ± 37.9 and 1 5 5 87.1 ± 52.8 µg m -3 , which were about 2.6 and 1.7 times higher than WHO guideline values for 24 h 1 5 6 average, respectively. The average PM 2.5 was 50.8±19.6 µg m -3 during the lockdown and 94.9 ± 1 5 7 30.6 µg m -3 before lockdown in Dhaka megacity. The lowest average relative humidity was 39.1 %, 1 5 8 the highest was 99.8 %, and the average rainfall was 0.23 ± 0.31 inch hour -1 . The average CO 2 1 5 9 emission was 427 ± 11.8 ppm, which is slightly lower (456 ± 63.4 ppm) than the value in January 1 6 0 2020.

6 1
As control measures during COVID-19 pandemic, precautions, such as lockdown were enforced by 1 6 2 restricting vehicle movements, closing government offices, schools, colleges, universities, shopping 1 6 3 malls, and "stay in home order" on March 26, 2020, have been taken by the Government of Bangladesh. Due to the restrictions, the concentrations of PM 2.5 and PM 10 have reduced drastically 1 6 5 up to 62% during April. The times series of daily average concentrations of PM 2.5 , PM 10 , and CO 2 1 6 6 have given in Figs. 2(a), 2(b) and 2(c) mentioning the events, e.g., lockdown begins, Garments industry open, Eid festival, lockdown ends and offices, etc. The daily average global CO 2 were also 1 6 8 drastically reduced up to 17% with a maximum of 26% and a minimum of 4% in the early April 1 6 9 compared to the previous year in 2019 due to the restricted activates during COVID-19 pandemic 1 7 0 (Le Quéré et al., 2020). In general, the values of the air quality parameters were getting lower with 1 7 1 the progress of the lockdown (Fig. 2). When the lockdown was curbed on May 30, 2020, we can average concentrations of PM 10 , PM 2.5 , and CO 2 emission were getting lower from January through 1 7 4 . CC-BY 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 certified by peer review) preprint The copyright holder for this this version posted September 13, 2020. . https://doi.org/10. 1101/2020 June 2020. January and February were almost similar but much higher than April, May, and June. The reduction was not only for COVID-19 precautions activities, but also rainfall from April to 1 7 6 June ( Fig. 2(d)), especially for PM 2.5 , PM 10 , and CO 2 ), but NO 2 reduction in April 2020 is only for 1 7 7 traffic emission (Fig. 5). However, we have identified and quantified the contribution of reduction previous year 2019 ( Fig. 3 and Fig. 4). As a case study, we have differentiated between "before 1 8 0 lockdown" and "during lockdown" PM 2.5 emission in Fig. 3. Several recent papers reported that a decline in PM 2.5 is associated with lockdown and affected  The satellite images of NO 2 distribution over the Dhaka city have been given in Fig. 4 at different 1 9 2 periods before and during the lockdown, and also during limited scale lockdown in 2020; and also, before lockdown over Dhaka. The comparison between April 13-27 in 2020 and April 15-29 in 1 9 6 . CC-BY 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 certified by peer review) preprint The copyright holder for this this version posted September 13, 2020. . https://doi.org/10.1101/2020.09.12.20193086 doi: medRxiv preprint 1 0 2019 was also indicating a very drastic reduction of NO 2 concentration over Dhaka. During April 1 9 7 2020, the vehicles were very limited in the street of Dhaka due to the complete shutdown of offices, 1 9 8 schools, colleges, universities, constructions, industries, and also "stay in-home" order. These are 1 9 9 the major causes of the drastic reduction of NO 2 over Dhaka city because the foremost source of 2 0 0 NO 2 emissions is from traffic vehicles. The NO 2 reduction was 40% (on average) in many cities of  Temporal variation of total confirmed COVID-19 cases, new cases, and daily deaths with PM 10 , 2 0 7 PM 2.5 , and AQI in Dhaka has given in Fig. 5. The relationship between aerosol particle number and between COVID-19 morbidity and mortality cases and air quality index (PM 10 , PM 2.5 , and AQI) 2 1 2 was not linear (Fig. 5). The reason may be the particle sizes -PM 2.5 and PM 10 are much larger 2 1 3 compared to the size of the SARS-CoV-2 virus. However, long term exposure to a high 2 1 4 concentration of particulate air pollution might be attributed to bronchial asthma, chronic 2 1 5 obstructive pulmonary diseases (COPD), chronic heart diseases, which might be one of the causes 2 1 6 of the relatively high COVID-19 death rate in Dhaka. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 13, 2020. The variation of total COVID-19 confirmed cases, daily new cases, and total deaths with climate 2 2 0 variables (temperate and relative humidity) in Dhaka from March 8, 2020 to June 16, 2020 has 2 2 1 given in Fig. 6. The total number of COVID-19 confirmed cases and total deaths were increasing 2 2 2 sharply in Dhaka. The temperature and relative humidity curve were also going up. The number of total and deaths cases with temperature and relative humidity in Dhaka city (Fig. 6). Climate variables (e.g., temperature, relative humidity, wind velocity and rainfall) have a significant impact with climate variables aggravated the pulmonary diseases and injured the respiratory pathways, and   The Kendall and Spearman correlation tests were performed between the empirical data of PM 2.5 , 2 3 5 PM 10 , CO 2 , and AQI and climate variables (temperature, relative humidity, wind velocity, and 2 3 6 rainfall) with COVID-19 morbidity and mortality. In both Kendall and Spearman correlation tests, 2 3 7 we found significant negative correlation with air quality indicators (PM 2.5 , PM 10 , CO 2 , and AQI) 2 3 8 with total confirmed COVID-19 cases, daily new cases, and mortality in Dhaka, Bangladesh ( Table   2  3  9 2). The magnitude of the correlation coefficients of PM 2.5 , PM 10 , CO 2 , and AQI were lower in 2 4 0 . CC-BY 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 certified by peer review) preprint

Kendall and Spearman Correlation Tests
The copyright holder for this this version posted September 13, 2020. . https://doi.org/10.1101/2020.09.12.20193086 doi: medRxiv preprint 1 2 Kendall tests than the Spearman correlation. The air quality was good during lockdown period but CO, and NO 2 have significant negative correlation with total COVID-19 morbidity and mortality. Italy).

5 5
The climate variables (average temperature, relative humidity, wind velocity, and rainfall) have  Wuhan, China showed a strong association between diseases spread and weather conditions, with a 2 6 2 . CC-BY 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 certified by peer review) preprint
The copyright holder for this this version posted September 13, 2020. . https://doi.org/10.1101/2020.09.12.20193086 doi: medRxiv preprint 1 3 prediction of "warm weather will play an important role in suppressing the virus." However, 2 6 3 temperature and relative humidity may play a significant role in COVID-19 transmission and also 2 6 4 have a substantial role in the mortality rate (Auler et al., 2020;Poole, 2020;Sajadi et al., 2020;2 6 5 Chen et al., 2020a;Ma et al., 2020;Wang et al., 2020b). Other meteorological indicators such as 2 6 6 wind velocity and rainfall also showed a significant positive correlation with COVID-19 cases in 2 6 7 Dhaka, Bangladesh.  Therefore, our study will supplement new knowledge for decision-makers that air quality and 2 7 5 climate variables are two important constraints that need to be addressed for controlling the 2 7 6 transmission of the SARS-CoV-2 virus in the developing countries. Air quality parameters (PM 2.5 , PM 10 , AQI, NO 2 , and CO 2 ) and climate variables (average 2 8 0 temperature, relative humidity, rainfall, and wind velocity) were studied before and during the correlation tests were performed to see the correlation between COVID-19 morbidity and mortality 2 8 4 . CC-BY 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 certified by peer review) preprint The copyright holder for this this version posted September 13, 2020. . https://doi.org/10.1101/2020.09.12.20193086 doi: medRxiv preprint 1 4 cases with climate variables and air quality indicators. The air quality has drastically reduced in 2 8 5 Dhaka due to COVID-19 lockdown in April 2020 (on average, about 2.5 times lower PM 2.5 and 6 2 8 6 times lower NO 2 observed than a normal period). PM 2.5 , PM 10 , CO 2 , air quality index (AQI) has no 2 8 7 positive but a significant correlation (within 1.0%) with the number of COVID-19 confirmed cases 2 8 8 and mortality. Average temperature, minimum temperature, and wind velocity have a strong 2 8 9 positive correlation with COVID-19 confirmed morbidity and mortality cases in Dhaka. These 2 9 0 findings will be helpful for policymakers to understand the mitigation mechanism of the air quality 2 9 1 and also will be useful for future research to control the SARS-CoV-2 virus as well as air quality. Authors acknowledge the data support to Air visual (www.iqair.com), European space agency 2 9 5 (NO 2 satellite images), DGHS, Bangladesh (COVID-19 related data at www.dghs.gov.bd) and for 2 9 6 climate variables data (www.darksky.net and www.weatherforyou.com). This research did not receive any specific grant from any funding agencies in the public,  . CC-BY 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 certified by peer review) preprint The copyright holder for this this version posted September 13, 2020. . https://doi.org/10.1101/2020  . CC-BY 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 certified by peer review) preprint The copyright holder for this this version posted September 13, 2020.  . CC-BY 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 certified by peer review) preprint The copyright holder for this this version posted September 13, 2020.  CC-BY 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 certified by peer review) preprint The copyright holder for this this version posted September 13, 2020. . https://doi.org/10. 1101/2020