Examining the status of improved air quality due to COVID-19 lockdown and an associated reduction in anthropogenic emissions

Clean air is a fundamental necessity for human health and well-being. The COVID-19 lockdown worldwide resulted in controls on anthropogenic emission that have a significant synergistic effect on air quality ecosystem services (ESs). This study utilised both satellite and surface monitored measurements to estimate air pollution for 20 cities across the world. Sentinel-5 Precursor TROPOspheric Monitoring Instrument (TROPOMI) data were used for evaluating tropospheric air quality status during the lockdown period. Surface measurement data were retrieved from the Environmental Protection Agency (EPA, USA) for a more explicit assessment of air quality ESs. Google Earth Engine TROPOMI application was utilised for a time series assessment of air pollution during the lockdown (1 Feb to 11 May 2020) compared with the lockdown equivalent periods (1 Feb to 11 May 2019). The economic valuation for air pollution reduction services was measured using two approaches: (1) median externality value coefficient approach; and (2) public health burden approach. Human mobility data from Apple (for city-scale) and Google (for country scale) was used for examining the connection between human interferences on air quality ESs. Using satellite data, the spatial and temporal concentration of four major pollutants such as nitrogen dioxide (NO2), sulfur dioxide (SO2), carbon monoxide (CO) and the aerosol index (AI) were measured. For NO2, the highest reduction was found in Paris (46%), followed by Detroit (40%), Milan (37%), Turin (37%), Frankfurt (36%), Philadelphia (34%), London (34%), and Madrid (34%), respectively. At the same time, a comparably lower reduction of NO2 is observed in Los Angeles (11%), Sao Paulo (17%), Antwerp (24%), Tehran (25%), and Rotterdam (27%), during the lockdown period. Using the adjusted value coefficients, the economic value of the air quality ESs was calculated for different pollutants. Using the public health burden valuation method, the highest economic benefits due to the reduced anthropogenic emission (for NO2) was estimated in US$ for New York (501M $), followed by London (375M $), Chicago (137M $), Paris (124M $), Madrid (90M $), Philadelphia (89M $), Milan (78M $), Cologne (67M $), Los Angeles (67M $), Frankfurt (52M $), Turin (45M $), Detroit (43M $), Barcelona (41M $), Sao Paulo (40M $), Tehran (37M $), Denver (30M $), Antwerp (16M $), Utrecht (14 million $), Brussels (9 million $), Rotterdam (9 million $), respectively. In this study, the public health burden and median externality valuation approaches were adopted for the economic valuation and subsequent interpretation. This one dimension and linear valuation may not be able to track the overall economic impact of air pollution on human welfare. Therefore, research that broadens the scope of valuation in environmental capitals needs to be initiated for exploring the importance of proper monetary valuation in natural capital accounting.


142
It is now well-established by many data-driven experiments that the accelerated rate of is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020 5 (lockdown equivalent period) and 2020 (lockdown period). The concentration of four key air 174 pollutants, nitrogen dioxide (NO2), sulfur dioxide (SO2), carbon monoxide (CO), and aerosol 175 index (AI) concentration, was computed for both 2019 and 2020 using Sentinel 5P data. For 176 six cities, i.e., Chicago, Denver, Detroit, Los Angeles, New York, and Philadelphia, the ground 177 monitored air pollution data was collected for a more explicit assessment of air quality ESs. 178 However, the ground monitored data was not adequate for the spatial evaluation for most of 179 the cities considered in this study. Therefore, the in-situ data was only used for time series 180 assessment of air pollutions, and the satellite measured pollution estimates were utilised for the 181 spatially explicit appraisal and economic valuation. Human mobility data, including driving 182 and transit for the selected cities, were collected from Apple (for city-scale) and Google (for 183 country scale) mobility reports. In addition to this, the gridded human settlement data and 184 population density data (pixel format) were collected from the Socio-Economic Data 185 Application Center, National Aeronautics and Space Application data center (SEDAC,186 NASA). For evaluating the total air pollution reduction of these 20 cities in a more accurate 187 way, the Geographical Information System (GIS) enabled city boundary (shapefile format) was 188 extracted from the OpenStreetMap (OSM) application. Two consecutive steps were followed 189 to get the boundary of these cities. First, the OSM relation identifier number (OSM id) was 190 generated for all the 20 cities using Nominatim, a search engine for OpenStreetMap data. Then,191 the OSM relation id of each city was ingested in the OSM polygon creation application 192 interface, which generates the geometry (both actual and simplified) of the relation id in poly, 193 GeoJSON, WKT or image formats. The formatted image geometry of the cities was then 194 imported in ArcGIS Pro software, and the city boundary was extracted using an automatic 195 digitisation function. CartograpHY (SCIAMACHY) programme (Veefkind et al., 2012). TROPOMI measures the 210 concentration of key tropospheric constituents at 7 × 7 km 2 spatial unit. This default spatial 211 scale was downscaled into 1km × 1km scale for city-scale analysis and subsequent 212 interpretation. In this study, the spatial and temporal variability of four key air pollutants was 213 extracted and mapped from the TROPOMI measurements using the Google Earth Engine cloud 214 platform. For this purpose, an interactive application called TROPOMI Explorer App, 215 developed by Google developers teams (Google, 2020;Braaten, 2020), was utilised to facilitate 216 quick and easy S5P data exploration and to examine the changes in air pollution in both cross-217 . CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020 6 sectional and longitudinal way. Spatial visualisation and time series charts for the selected air 218 pollutants were also prepared with the help of this TROPOMI Explorer application. The other 219 accessories of this application, such as NO2 time series inspector, NO2 temporal comparison, 220 NO2 time-series animation, were also utilised for different computational purposes. 221 2.3.2 Ground pollution data 222 Ground monitored air quality data was available only for a few cities considered in the 223 study, including Chicago, Denver,Detroit,Los Angeles,New York,and Philadelphia. Thus,224 these cities were selected for the ground data-driven analysis. Ground monitored data for these 225 cities were collected from the U.S. Environmental Protection Agency (US EPA). This data is 226 available for a daily scale and for six key pollutants, such as CO, NO2, O3, PM2.5, PM10, and 227 SO2, respectively. The in-situ air pollution concentration at a daily scale was considered only 228 for the time series assessment of pollution concentration. Additionally, the said in-situ data had 229 not been used for any validation and calibration of satellite pollution estimates. The time series 230 (2000-2020) air quality index (AQI) of these selected cities were also generated using the 231 multilayer time plot function. The overall AQI values were sub-divided into six groups, i.e., 232 good, moderate, unhealthy for sensitive population groups, unhealthy, very unhealthy, and 233 hazardous, respectively. In addition to this, the single year AQI data was also extracted for the 234 selected cities from the EPA. The number of unhealthy days for each pollutant was measured 235 using the EPA AQI plot function. The combination of two different pollutants, such as CO and 236 NO2, PM10 and PM2.5, was permuted to assess the yearly AQI status of the cities. As several 237 studies reported the increment of O3 due to the reduction of GHG emissions, this study also 238 evaluated the O3 exceedances for the current year compared to the average O3 concentration 239 of the last 5 and 20 years. This particular task was implemented using the EPA Ozone 240 exceedances plot function (EPA, 2020).  244 The accelerating increases of air pollution in cities is a major concern across the world 245 (Chan and Yao, 2008;Kim Oanh et al., 2006;Mayer, 1999;Guttikunda et al., 2014;Abhijith 246 et al., 2017;Rai et al., 2017;Pilla and Broderick, 2015). Various policies have been 247 implemented for managing the city-based air pollution that mainly originated from 248 anthropogenic activities from specific sources and sectors (Kumar et al., 2015;Kumar et al., 249 2016;Baró et al., 2014;Feng and Liao, 2016;Zhang et al., 2016 (Guerriero et al., 2016;Castro et al., 2017;Jeanjean et al., 2017;7 the economic values of the NO2, SO2, CO, aerosol reduction to gauge the economic benefits of 261 these functions. Since this study has considered the air pollution reduction at the city scale, the 262 public health burden and mean externality valuation approaches were utilised for estimating 263 economic damage due to air pollution and to calculate the economic values of air quality 264 services ( Baro et al., 2014;Matthews and Lave, 2000). Unit social damage price due to air 265 pollution was estimated for 2020 using the U.S consumer price index (CPI) inflation calculator 266 (U.S Bureau of Labor Statistics, 2020). Additionally, using the most updated price conversion 267 factors, the mean externality values for the key pollutants was estimated as: CO = 956 $ t -1 , 268 NOx = 5149 $ t -1 , SO2 = 3678 $ t -1 , PM10 = 7907 $ t -1 .

269
The public health burden valuation approach has also been utilised for economic 270 valuation of air quality ESs (Kumar et al., 2020a, Etchie et al. 2018Hu et al., 2015;Sharma et 271 al., 2020;Sahu and Kota, 2017;COMEAP, 2009 Gridded population data from SEDAC, NASA, was utilised for this task. Pollution and gridded 283 population data for the same time period were used for estimations of PWAC.

284
Following, the health burden (HB), which refers premature deaths attributable to short-285 term exposure to air pollutants was estimated for the study period (1 February to 11 May 2020).

286
The reduction in health burden (ΔHB) was also measured by calculating the difference between 287 the previous and later HB estimates. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. analysis (OECD, 2014;WHO, 2015), ecosystem service studies (Zhang et al., 2018(Zhang et al., , 2020.

315
The economic benefits due to avoided premature mortality were estimated as follows: is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.  York, Philadelphia, etc. (Fig. 1). Moreover, among the 20 cities, the highest NO2 reduction is 370 recorded in Tehran, and the lowest reduction is found in Los Angeles and Sao Paulo (based on 371 1 st Feb to 11 th May pollution data). The SO2 emission is evaluated and presented in Fig. S1.

372
An incremental trend of SO2 emission is observed during the study period. For most cities, SO2 373 concentration was increased during the study period. However, for exceptions, a slight decrease 374 in SO2 emission is observed in Rotterdam, Frankfurt, London, and Detroit cities (Fig. S1). The 375 spatial distribution of CO is also evaluated using GEE cloud application and Sentinel 5P data is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020 York, Philadelphia, Milan, Madrid, etc. (Fig. 2). At the same time, CO emission was increased 379 in Cologne, Denver (Fig. S2). The spatial distribution of aerosol concentration is also 380 calculated and presented in Fig. S3. Aerosol concentration is also found to be decreased during 381 the COVID lockdown with restricted human activities. is also evaluated and presented in Fig. 2  Los Angeles, while a comparably low CO concentration is documented for Sao Paulo, Denver,405 Madrid, Barcelona, and Brussels (Fig. 2). Except for a few cities, the concentration of NO2, 406 CO, and aerosol has been reduced substantially ( Fig. 3  and Rotterdam (26.72%), respectively ( Fig. 3 and Table. 2). For CO, the maximum reduction 412 was recorded for New York (4.24%), followed by Detroit (4.09%), Sao Paulo (3.88%), Rotterdam (0.01%) ( Fig. 3 and Table. 2). The temporal variability of NO2, SO2, CO, and 416 aerosol concentration is shown in Fig. 4, Fig. 5, Fig. S4, Fig. S5, Fig.S6, Fig. S7, Fig. S8. Both all the 20 cities considered in this study (Fig. S4, Fig. S5). However, for SO2, an incremental 420 trend was observed for most of the cities (Fig. S6).

421
. CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020 Using the ground monitored data, the daily air quality index (AQI), and a cumulative 422 number of good AQI days for the six American cities was computed and presented in Fig. 6. 423 The ground monitored data for these six cities have been considered only for time series 424 assessment and subsequent interpretation. In all cases, it has been found that AQI is reduced 425 significantly due to lockdown led reduction in human mobility and traffic emission. In the left 426 panel, the grey color indicates the five years average AQI and light blue shade demonstrating 427 the average AQI range in the last 20 years. Based on the AQI ranges, four AQI classes were 428 characterised, such as good, moderate, unhealthy for sensitive groups, and unhealthy (Fig. 6). 429 A comparably higher cumulative number of good AQI days is recorded during the lockdown 430 period for all five cities, except Chicago (Fig. 6). Using the EPA AQI interactive plot function 431 application, the daily AQI of the US cities were analysed and presented in Fig. 7, Fig. S9, Fig.   432 S10. The daily NO2 and SO2 AQI suggest that all the cities are benefitted by having good 433 quality air due to anthropogenic pollution switch-off and restricted human mobility that 434 collectively improved the air quality ecosystem services in these cities. The multi-year daily 435 time series plot (Fig. 8, Fig. 9, Fig. 10 (Jan to May 2020) period. Fig. 8 shows that in all cities, the NO2 AQI status is mostly good 441 during the lockdown period compared to the long-term average AQI in these cities. Fig. 9 442 shows the PM2.5 AQI status, which also found improving during the lockdown period. The  multi-year time series plot was prepared after combining all the pollutants that suggest that the 446 air quality is improved substantially, which is supported by the lower AQI recorded during the 447 lockdown period compared to the long-term AQI recorded in these cities. Among the six cities, 448 the hazardous to very unhealthy air quality is common in Los Angeles, compared to the other 449 five US cities considered in this study.  451 Using both Google and Apple human mobility information, the effect of lockdown and 452 its striking impact on human outdoor activities is measured and presented in Fig. 11, Fig. S12, 453 Fig. S13, Fig. S14, Fig. S15. The driving and transit mobility was calculated using the Apple 454 mobility data. Mobility on January 13 was taken as a baseline, and further changes in human 455 mobility during the lockdown period was calculated from the baseline mobility. The driving 456 counts reduced most significantly in Paris, followed by Madrid, London, Antwerp, and 457 Brussels (Fig. 11). Whereas, such changes were comparably lower in Chicago, Cologne, 458 Denver, Los Angeles, New York (Fig. 11). Transit counts also reduced significantly in Paris, 459 followed by Utrecht, Sao Paulo, New York, Milan, Chicago, Antwerp, and Brussels (Fig. 11). 460 Using the Google human mobility records, the changes in different mobility such as retail and 461 recreation, grocery and pharmacy stores, transit, parks and outdoor, workplace visitor, and time 462 spent at home were measured. Transport related mobilities were reduced most significantly in 463 the Latin American countries, followed by a few Middle East and Southeast Asian countries, 464 and American countries (Fig. S13). Parks and outdoor activities were found to be reduced is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

Human mobility and its paramount effect in lowering the pollution levels
The copyright holder for this this version posted August 24, 2020. .

12
outdoor activities are seen to be increased in a few European countries as well (Fig. S13). The 467 highest reduction in retail and recreation is found in India, Turkey, UK, and few Latin 468 American countries due to lockdown and associated restrictive measures. (Fig. S14). 469 Considering grocery and pharmacy-related mobilities, the highest reduction is being observed 470 in the Latin American countries and a few European countries. Whereas grocery related 471 mobility was found to be increased in the USA, few African and European countries (Fig. S14). 472 Workplace related mobility is reduced significantly in Peru, Bolivia, India, Spain, Turkey,   is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020.08.20.20177949 doi: medRxiv preprint 13 levels, the economic cost attributed to air pollution led health burdens was reduced significantly 510 (  transferring ideas from place to place would be easy, which eventually establishes more trust 532 and transparency in applying the scientific findings to solve real-life problems. Evaluating the 533 reliability of remote sensing data is always a matter of concern. Since this study has evaluated 534 the air pollution in cities, which itself is very sensitive in nature, proper and careful evaluation 535 is required to verify the accuracy of satellite estimates to draw a data-driven conclusion that   546 In this study, the spatial and temporal distribution and changes in different air pollution is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

Anthropogenic emission and ecosystem services
The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020.08.20.20177949 doi: medRxiv preprint 14 NO2 reduction was observed for Netherlands (70%), Japan (64%), Macao (60%), Lebanon 553 (55%), Italy (54%), India (54%), Monaco (54%), North Korea (51%), Hungary (50%), and 554 Kuwait (50%), respectively. While an incremental trend of NO2 emission was found in the 555 Island countries, i.e., Kiribati (213%), Howland Island (136%), Jarvis Island (129%), Nauru 556 (93%), Pacific Islands (Palau) (81%) along with other countries such as Indonesia (74%), Nepal 557 (57%), Mozambique (56%), Norfolk Island (55%), and Jan Mayen (52%), where COVID1-9 558 lockdown has not implemented or followed strictly (Fig. S11 574 The connection between human mobility and air pollution levels in selected cities were 575 also examined in this research. Both Apple and Google mobility data were used for this 576 purpose. Results derived from both the report suggest that due to the mandatory lockdown and 577 resulted in limited outdoor human activities, mobility has been reduced significantly across the 578 world. This drastic reduction of human mobility could contribute to the reduced level of air 579 pollution observed in the last few months. For most of the cities considered in this study, human 580 mobility has been reduced up to 80% from the baseline mobility. The highest reduction in 581 mobility was found in the European cities. To prevent infection, the authorities in these cities 582 implemented preventive measures, which included partial lockdown in different sectors, 583 including restricted outdoor social activities. This mandatory imposition of lockdown has 584 resulted in a reduced level of traffic volume in cities (Fig. 11, Table. S6). The mobility analysis 585 thus suggests that by introducing sustainable transport plans and policies, air pollution in the 586 urban regions can be minimised to a certain extent. The periodic and temporary lockdown can 587 also be adopted in the highly polluted cities if no other alternatives are feasible at the place. A 588 similar strategy has already been adopted by New Delhi Government by introducing 589 "odd/even" transport scheme where private vehicles with odd digit (1, 3, 5, 7, 9) registration 590 numbers will be allowed on roads on odd dates and vehicles with even digit (0, 2, 4, 6, 8) is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

643
The present research has made an effort to investigate the human impact on the natural 644 environment by taking COVID-19 lockdown and its resultant reduction of air pollution. Both  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. .

J.
Hyg   is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.    is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.   is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.   is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020.08.20.20177949 doi: medRxiv preprint Fig.1 Spatial distribution and changes in NO2 concentration (µmol/m2) derived from Sentinel 5P TROPOMI data. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint Chicago Denver New York Philadelphia Fig. 8 Multi-year daily time series plot shows the variation of air quality status (NO 2 ) from 2000 to 2020. Due to lock down and associated reduction of air pollution, air quality status is improved in all the selected cities in USA.  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.  is the author/funder, who has granted medRxiv a license to disp h was not certified by peer review) The copyri this version posted August 24, 2020. . https://doi.org/10. 1101/2020.08.20.20177949 doi: int Fig. S1 Spatial distribution of SO 2 (μmol m -2 ) in the selected cities in 2019 and 2020 (from Feb to May). Spatial maps in third panel shows the spatial difference in SO 2 concentration between 2019 and 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020.  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020.08.20.20177949 doi: medRxiv preprint

Parks and outdoor spaces
Fig. S13 Spatial variability of public transport and parks/outdoor mobility during the lockdown period.
. CC-BY 4.0 International license It is made available under a perpetuity.

Time spent at home
Fig. S15 Spatial variability of workplace and residential mobility during the lockdown period.
. CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the prep (which was not certified by peer review) preprint The copyright holder fo this version posted August 24, 2020.  Table. 1 Summary statistics of mean NO 2 , SO 2 , CO, and Aerosol concentration during 2019 and 2020 (Feb to May). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020.08.20.20177949 doi: medRxiv preprint Table. S6 Changes in human mobility (%) from the baseline (mobility on 13 th January) during the lockdown period (1 st February to 11 th May 2020).

City
Jan (   is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted August 24, 2020. . https://doi.org/10.1101/2020