Air Quality and Its Relationship with COVID-19 Mortality in Hotspot Places of India: A Post-lockdown Analysis

The COVID-19 pandemic spread over the world like the wind with more than 400,000 documented cases as of March 24 th , 2020. In this regard, strict lockdown measures were imposed in India on the same date to stop the virus spread. Thereafter, various lockdown impacts were observed and one of the immediate effects was a reduction in air pollution levels across the world. In this study, we have selected 14 major hotspot places where the COVID-19 cases were >1000 (as of 1 st June 2020) that represents more than 70% mortalities of India. We assessed the impact of lockdown on different air quality indicators including ground (PM 2.5 , PM 10 , NO 2 , SO 2 , O 3 , and AQI) and tropospheric nitric oxide (NO 2 ) concentrations through ground monitoring stations and Sentinel-5 satellite data respectively. We have found highest reduction in NO 2 (-48.68%), PM 2.5 (-34.84%) and PM 10 (-33.89%) air pollutant (unit in µg/m 3 ) levels post-lockdown. Moreover, tropospheric NO 2 (mol/m 2 ) concentrations were also improved over Delhi, Mumbai, Kolkata, Thane, and Ahmedabad metro cities. Interestingly, air pollutant indicators have been correlated with different periods (as of 1 st and 15 th June 2020) of COVID-19 mortalities data to assess the bounding between these variables. Accordingly, we have found a strong positive correlation of mortalities data with ground PM 10 (R 2 =0.145; r =0.38) and AQI (R 2 =0.17; r =0.412) indicators and this relationship has been improved signicantly on second time point. The correlation nding suggests that the long-term bad air quality may aggravate the clinical symptoms of the disease.


Introduction
The COVID-19 outbreak started from Wuhan city, capital of Hubei province (Raibhandari et al., 2020) when the rst case of coronavirus was reported in December 2019 . The Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) gradually spread throughout the world and become a global health issue (Chen et al., 2020a;Gilbert et al., 2020) and later World Health Organization (WHO) has declared it a pandemic Cucinatta & Venelli, 2020). In India, the rst SARS-CoV-2 case was reported in the Kerala state in Jan 2020 (Gautam & Hens, 2020) and gradually spread in Maharashtra, Gujarat, Delhi, and rest of the states. Later, a countrywide lockdown was declared by the Prime Minister of India on the midnight of March 24th, 2020 that has been extended till April and further extended till June 2020 due to the fear of spreading the virus. By this countrywide lockdown, transportation and industrial activities were almost stopped and as a bene cial outcome, air pollution levels were drastically reduced in various cities of India (Sharma et al., 2020) reported by the Central Pollution Control Board (CPCB). The different pollution indexes were adopted throughout the world that includes various air pollutants and in India, Air Quality Index (AQI) often, also is termed as Air Pollution Index (API) is adopted to express the magnitude of air pollution (Shenfeld, 1970;Ott and Thorn, 1976;Murena, 2004). In the initial stage, some signi cant pollutants were not included which harms the respiratory system (Radojevic and Hassan, 1999;Qian et al., 2004). In this context, Indian National Air Quality Standards (INAQS) have included various air pollutants (PM2.5, PM10, O3, SO2, NO2, NH3, and Page 3/18 CO) to calculate the AQI through more than 200 monitoring stations across India (http://www.cpcb.nic.in).
India is one of the most polluted countries and 21 Indian cities are in the list of world's most 30 polluted cities (https://www.iqair.com/us/world-most-polluted-cities) that takes almost 12.4 million lives every year (Balakrishnan et al., 2018). In such conditions, the COVID-19 lockdown bene cially improved the air quality throughout the globe, and a decline in different air pollutants have been reported worldwide and in India employing ground monitoring stations and remote sensing datasets (Chauhan and Singh, 2020; Nakada and Urban, 2020). The drastic deduction (more than 50%) in different air pollutants were highlighted in megacity Delhi, India (Mahato et al., 2020).
Interestingly researches have established a relationship between air pollution and mortalities related to SARS-CoV-2 (Zhu et al., 2020;Conticini et al., 2020). In this regard ground and tropospheric pollutants (PM and NO2) levels were correlated (Qin et al., 2020). Therefore, people living in polluted regions generally inhale toxic pollutants from the past couple of years are more prone to it that makes the immune system weaker (Viehmann et al., 2015;Schraufnagel et al., 2019). In this study, we have targeted only those 14 hotspots regions of India where the COVID-19 cases and mortalities (represents more than 70%) are reported high. Accordingly, this research aims to assess the reduction of AQI and different air pollutants (PM 2.5 , PM 10 , NO 2 , O 3 , and SO 2 ) at the ground monitoring station and extract the changes in Tropospheric NO2 concentration employing remote sensing data. Moreover, the study also tried to establish the relationship between air pollutants and COVID-19 mortalities.

Material And Methods
To assure the post-lockdown changes in air pollution, we employed In-situ air quality monitoring data for more than 200 stations of India. Later, the Inverse Distance Weighted (IDW) interpolation technique was used to get the raster values for maximum area. These AQI were assessed on 25 th February 2020 (a month before lockdown), 25 th March 2020 ( rst day of lockdown) and 25 th April 2020 (a month after lockdown) different dates to monitor the variations. Moreover the detailed investigation were performed using ground and tropospheric pollutant indicators for Ahmedabad, Aurangabad, Bhopal, Chennai, Delhi, Hyderabad, Indore, Jaipur, Jodhpur, Kolkata, Mumbai, Pune, Surat, Thane cities (Fig. 1). The criterion behind selecting these major places is that we have considered only those places where the number of cases was more than 1000 as of 1st June 2020 ( Table 1). The statistical analysis was performed using Air Quality Index (AQI), and other pollutants i.e, PM 2.5 , PM 10 , NO 2 , SO 2 and O 3 (units in µg/m 3 ) monthly average data of mentioned places were procured for pre-lockdown (February 25th, 2020 and March 24th, 2020) and post-lockdown (March 25th, 2020 and April 24th, 2020) periods from Central Pollution Control Board (CPCB) portal (https://cpcb.nic.in/). At the same time, Remote sensing data was analyzed using Google Earth Engine that was collected from Copernicus Sentinel-5 Precursor Tropospheric Monitoring Instrument (S5p/TROPOMI) to know the spatial variations in tropospheric NO2 (mol/m 2 ) concentrations during the same average monthly time periods for the same study areas. Moreover, the linear regression and correlation analysis were performed between different air pollutants with COVID-19 mortalities (as of 1st June 2020) obtained from different portals handling by the Ministry of Health & Family Welfare (https://www.mohfw.gov.in & https://www.mygov.in/covid-19). A similar investigation was performed after 2 weeks (as of 15th June 2020) of updated COVID-19 mortalities (Table 1) to assess the variations in the relationship between these variables.  In India, three weeks of lockdown was declared at midnight on the 24 th of March 2020. Thereafter, the country has observed a remarkable reduction of about -36.10% in AQI on 25 th March 2020 compared to a month before (25 th Feb 2020) lockdown as per our statistical calculation using more than 200 air quality monitoring station data. The overall AQI of India has been reduced drastically as per the investigations.
The maximum and minimum AQI values are visible and highlighting the improvements over the period of time (pre, during and post lockdown dates) in India (Fig. 2a-c). The average AQI before a month of lockdown was 128 that dropped down at 89 (on a very rst day of lockdown) and further decreased after a month of lockdown that was 72. This continuous reduction in AQI clearly indicates that the atmospheric pollution of India has been improved a lot under the COVID-19 lockdown. March 2020 (Pre-lockdown) (Fig. 3a-f). The PM 2.5 was >100 µg/m 3 at Jodhpur, Surat, and Thane cities, whereas Delhi, Jodhpur, Mumbai and Ahmadabad were the leading cities in PM 10 (most of them had >110 µg/m 3 ) concentrations before lockdown period (Fig 3a, b). These regions have more tra c and industrial burden which was the main responsible reason for the high level of PM concentrations. A similar pattern has found in NO2 levels where these pollutants were recorded higher at Indore, Thane, Jodhpur, and Ahmedabad stations (Fig. 3c) whereas SO 2 concentration was more at Aurangabad, Pune and Thane cities in a pre-lockdown phase that have been reduced considerably (Fig 3d). Interestingly, in 6 out of 14 cities, and increased has been observed in the O 3 level after lockdown (Fig. 3e). Aurangabad, Chennai, Delhi, Kolkata are few cities where the O 3 level has increased partially. The signi cant improvement has been noticed in AQI levels of all cities among which Ahmedabad, Delhi, Jodhpur, and Kolkata cities were leading in poor air quality (Fig. 3f). The changes (positive and negative percentage) in different air pollutants were calculated ( Fig. 4 and Table 2) and accordingly it has been found that the average AQI is dropped almost -31.59% in studied cities. The highest average reduction has been found in NO 2 (-48.68) compare to other pollutants. The same trends of reductions were also observed in SO 2 (-37.76%), PM 2.5 (-34.84%), PM 10 (-33.89%) and O 3 (-9.06%) air pollutant indicators. Results reveal that the NO 2 levels have reduced more in Thane, Mumbai, and Kolkata cities after imposed lockdown with 77%, 74%, and 68% respectively. These cities are known for heavy tra c loads with less density of roads that increase the burden of NO 2 levels. A similar trend of reduction is observed in both Particulate Matter (PM 2.5 and PM 10 ) with little variations. The remarkable improvement in PM2.5 concentrations was observed at Pune, Thane, and Ahmedabad that was reduced -63%, -56%, and -43% respectively, however, it was high before the lockdown period at Jodhpur city. The prominent source of PM 2.5 is organic aerosols and motor vehicle tra c that are totally anthropogenic induced activities that stopped due to lockdown. The PM 10 is highly controlled by construction sites, burning activities, industrial sources, and dusts factors that make Delhi, Jodhpur and Mumbai cities more prone in PM 10    Moreover, tropospheric NO 2 (mol/m 2 ) pollutant concentrations were also mapped to observe the temporal variation through remote sensing data. Accordingly, we found a massive improvement in Delhi, Mumbai, Thane, Ahmedabad, Chennai and Hyderabad cities as the highest NO2 concentrations (red color) scale showed 0.0001 mol/m 2 that is totally invisible post-lockdown compared to pre-lockdown (Fig. 5).

Relationship of Air Pollutants and COVID-19 Mortalities
The considered 14 places have covered almost >70% of COVID-19 mortalities and its growth rate is high in these hotspot regions compared to the rest of Indian places. The Mumbai and Delhi are main COVID-19 hotspot and polluted places in India and across the world. People are inhaling these toxic pollutants and dying from past decades, therefore, to nd out the relationship between COVID-19 mortality and atmospheric pollution is an important task. In this regard, our linear regression results have shown satisfactory positive relationships with PM 2.5 , PM 10 , and AQI pollutant indicators (Fig. 6a, b and f). The analysis showed promising association between COVID-19 deaths and PM 10 (R 2 =0.145; r =0.38, p=0.039), AQI (R 2 =0.17; r =0.412, p=0.21) and PM 2.5 (R 2 =0.107; r =0.-32, p=0.081) air pollutants. The weaker/negative correlation between these variables have found with SO 2 and O 3 air pollutants (Fig. 6d,   e). The NO 2 pollutant at the ground insigni cant relationship (Fig. 6c), whereas high concentrations of tropospheric NO 2 (mol/m 2 ) over Mumbai, Delhi, Thane and Ahmedabad places (Fig. 5) indicates it a contributing factor, as the COVID-19 deaths are more in these regions compared to rest of investigated places. However, we could make this relationship stronger when less vulnerable COVID-19 places would be correlated with good air quality regions. We have taken the COVID-19 mortalities data again after 2 weeks (as of 15th June 2020) to analyze the depth relationship between these two variables and our results again corroborate with signi cant improvements. As per the updated COVID-19 mortalities data, this relationship and correlation with PM 10 (R 2 =0.207; r=0.455, p=0.036) and AQI (R 2 =0.18; r =0.425, p=0.044) have made stronger with signi cant improvements than before (Fig. 7a, b).

Conclusion
This study investigates the impact of lockdown on air quality that improved signi cantly and the detailed study was conducted on 14 major COVID-19 susceptible places of India. Our results reveal that the higher reductions were observed in NO 2 , PM 10 and PM 2.5 (µg/m 3 ) pollutants with a rate of -48.68%, -34.84% and concentrations were also decreased especially over the Delhi, Mumbai, Thane, Pune, Kolkata and Ahmedabad places where the concentration was high before imposing lockdown. We tried to establish the relationship between COVID-19 mortalities (on two different time datasets) with different air pollutants and observed satisfactory positive correlations with PM 10 and AQI indicators. Interestingly, we found improvement in a relationship when correlated again with updated COVID-19 mortalities data. A similar attempt has been made in different regions of the world that strongly support our results (Wu et al., 2020;Chen et al., 2020). However, the SARS-CoV-2 is a communicable disease but at the same time people living under poor air quality with weakened immune systems are vulnerable or certain diseases also face higher risks and are experiencing the coronavirus pandemic with particular anxieties.

Competing interests
The authors declare no competing interests.  Figure 1 Study area: The major COVID-19 vulnerable places marked with red color along with respective codes and air quality monitoring stations.

Figure 2
Air quality index levels in India: (a) a month before, (b) the next day and a month after (c) country lockdown.
Page 14/18  The average tropospheric NO2 concentration variations in the study area during a one-month pre-and post-lockdown period.