Spatiotemporal trends of selected air quality parameters during force lockdown and its relationship to COVID-19 positive cases in Bangladesh.

Worldwide improved air quality in different cities is reported influenced by lockdown came in force due to COVID-19 pandemic; however, as expected, such changes might have been different in different places. And what is still not very clear whether air quality pollutants have some link to account COVID-19 positive cases and death tolls. This study aims to evaluate the spatiotemporal variability of air pollutants and their relationship to COVID-19 positive cases in major cities in Bangladesh. The relevant data of air pollutants and COVID-19 positive cases are collected, analyzed, discussed for lockdown period of 26 March to 26 April 2020 in comparison to data for same period averaging over 2013 to 2019 for eight major cities in Bangladesh. To characterize air pollutants affected by lockdown, trend and rate of changes were carried out using Mann-Kandle and Sen's slope methods, while spatial variability across the cities was done using ArcGIS and statistics within ArcGIS. The substantial reduction of mean concentrations in the range of 30-65%, 20-80%, 30 - 80%, 65 - 90% and 75 - 90% across the cities is found during lockdown compared to typical mean in previous years for the PM2.5, PM10, SO2, CO, and NO2 concentrations in air. Among the cities studied, it is seen that relatively lesser reduction in Dhaka, Gazipur and Narayanganj and moderate reduction in Chittagong, Rajshahi, Khulna and Barisal, while significantly bigger reduction in Sylhet influenced by the city attributes and climatic variabilities. Among all the pollutants studied, the increasing trends of NO2 and CO in Dhaka, Gazipur and Narayanganj are unexpected even in lockdown pointing the effectiveness of lockdown management. Correlation among the air pollutants and confirmed COVID-19 cases across the cities depict foggy relationship, while PCA integrated over the cities revealed association with gaseous pollutants pointing stronger effects of NO2. This relationship illustrates air pollution health effects may increase vulnerability to COVID-19 cases.

Introduction

Air is an important component of the environment and hence, maintain quality air for all the living organisms is deemed necessary, unless otherwise it can substantially affect environmental health and livelihood. It has been seen that air pollution reached to an unacceptable level illustrating significantly higher air quality index (AQI) / air pollution index (API) than the threshold value set by relevant agencies mostly following the guidelines prescribed by the World Health Organization (WHO) and United States Environmental Protection Agency (USEPA). Among many cities in the globe, cities located in South Asia in particular are experiencing air pollution hazard that is ranked sixth of the top ten killers indicating its potential threats to human health and lives along with environmental health, while often ignored. This crucial component of the environment is getting severely polluted by receiving pollutants input from the unplanned urbanization for socio-economic development compromising environmental health, and anthropogenic input from motor vehicles on roads, industrial activities. In a nutshell all these activities causes great damage to the human health, agricultural production, mortality and morbidity in major cities of developing world like Bangladesh (H. A. Haque et al., 2017; Saha, 2011). The air in urban areas of Bangladesh are found with elevated concentrations of air quality pollutants, especially in dry period (November to April) influenced by meteorological parameters, such as of scarce rainfall, northwesterly wind and relatively comfortable temperature with low humidity (CASE-DoE, 2019). A few studies reported that brick kilns within the close proximity, burning dumped wastes (household and industrial) in open air, vehicular emissions, sand fields along the banks of the rivers, construction activities, coal and biomass burning as fuel etc. are primarily responsible for air pollution in most of the cases (Hoque et al., 2014; Rana et al., 2016). In this period, the air pollutants such as, volatile organic compound (VOC), carbon dioxide (CO2), carbon monoxide (CO), oxygen (O2), sulfur dioxide (SO2), nitrogen oxide (NOx), hydrogen sulfide (H2S), suspended particulate matter (SPM), particulate matter (PM10 and PM 2.5) have been found to be increasing in trends in major cities of Bangladesh (Kitada and Azad, 1998; Mohiuddin, 2018; Zahangeer Alam et al., 2018). To address air pollution control, the Government of Bangladesh (GoB) had issued a set of limit values for the criteria air pollutants, e.g., PM, Pb, SO2, CO, NOx and O3 in 2005. The notable measures were adopted meanwhile for improved the air quality, the major of which were as lead free gasoline in 1999, introduction of cleaner fuel CNG to transport sector in early 90's, phase-out of 2-stroke 3-wheeled baby taxis in early 20's, increased the chimney height for brick kilns, and adoption of new brick burning and control law – 2013 (revised) which prohibits burning fossil fuels in brick manufacturing sector (CASE-DoE, 2019). While these interventions see improved air environment during the time of execution, the sustain benefits hardly go long due to the massive influx of populations in urban centers, increased requirement and use of excessive fuel, lack of action against inappropriate actions until the forced lockdown came as blessing disguise for environment healing due to present COVID-19 pandemic worldwide.

The first case of COVID-19 pandemic was identified in Wuhan, Hubei Province of China at the end of 2019 and then it spread over most countries in the world till to date of writing (IEDCR, 2020). In Bangladesh, first three infections of COVID-19 cases were registered on March 8, 2020 (IEDCR, 2020) and thereafter, country wide complete lockdown came in force on 26 March 2020 as per declaration by the GoB on 16 March 2020. It was then found to be continued in several phases until 26 April 2020 (Shawon, 2020) when government allowed the factories and other economic sectors including garments, export-import zones to be reopened in limited scale maintaining health regulations set by the GoB guided by the WHO. Nevertheless that limited opening causes increased in human activities and business as usual like it was before the shutdown (M. Haque and Ahamad, 2020) and hence, open up the door for infections. While the COVID-19 lockdown reduced the human activities up to 90% due to contagion, the environmental pollution was found to reduce about 30% in Spain, USA, Italy and Wuhan (Muhammad et al., 2020). A few studies further in this domain also reported improved air quality influenced by COVID-19 lockdown from major cities of the world. As for example, the concentration of PM2.5, PM10, SO2, CO, and NO2 in the major three cities of central China have reduced by 30.1%, 40.5%, 33.4%, 27.9% and 61.4%,respectively, during COVID-19 lockdown and subsequent AQI reduction of 32.2%, 27.7% and 14.9% as compared with the concentrations during 2017–19 (Xu et al., 2020). Another study by Sharma et al. (2020) reported maximum reduction of PM2.5 in most of the regions in India. In this align, the concentrations of PM2.5, PM10, CO, and NO2 are seen declined by 41%, 52% and 28%, respectively in six megacities of India (Jain and Sharma, 2020). A study in Almaty, Kazakhstan by Kerimray et al., (2020) informed 21%, 35% and 49% reduction of PM2.5, NO2 and CO concentrations, respectively, during lockdown period. Furthermore, a similar declined trend in the concentrations of PM2.5, PM10, SO2, CO, and NO2 in major cities of Bazil, Chaina, Spain and Morocco, and an increasing trend for Ozone (O3) in all cases were described by Dantas et al. (2020), Otmani et al. (2020), Tobías et al. (2020), Xu et al. (2020). As appeared, this forced measure gives the impression as blessing disguise for environmental healing: however, changes in air pollutants and its relation of COVID-19 cases and deaths might have different trends in different places influenced by different attributes.

As air pollutants' health impacts are strongly linked to respiratory diseases like asthma, reduction or difficulties in oxygen supply to blood, sore throat, breathing difficulties etc., and all of which are declared as common symptoms of COVID-19 positive cases by WHO, it is therefore expected that cities with significantly higher air pollution may see the higher number of cases and death tolls. A very limited studies exists that evaluated the link between air quality and corona virus infections, such as, Travaglio et al. (2020) testified a link of corona cases and lethality with NO and NO2 from their study in England and, another similar plots describing the link of NO2 levels and COVID-19 deaths at a regional level in Italy, Spain, France and Germany by Ogen (2020) at Martin Luther University Halle-Wittenberg in Germany. Analyzing the relationship between air quality pollutants and COVID-19 cases and deaths, it appears that the links are yet murky and deemed necessary to explore at different places with different climatic and urban settings to clarify what is still not clear from the available facts and figures. In particular, people living with pollution over the decades like living in urban slums in Dhaka and Chittagong to answer the question whether people living in a densely populated settings is more susceptible to get affected and dying from COVID-19 cases or not. The positive cases in Bangladesh across the major cities to date of 26 April 2020 are 3615 since its first report on 8 March 2020. Among the total cases 95% belongs to the central belt of country, e.g., Dhaka, Gazipur and Narayanganj. This study therefore seeks to characterize the benefits of lockdown on air pollutants at different regions of Bangladesh and their variabilities along with investigation of the relationship exists between COVID-19 registered cases and air pollutants in spatiotemporal scales.

Materials and methods

Bangladesh is a promising developing country located in 20°34’N to 26°38’N latitude and 88°01′E to 92°41′E longitude (see Fig. 1 ) and is housed of about 163 million people in 148,460 km2 area including waterbodies, hills, forest. This is one of the most densely populated countries in the world with an average of 1252 people living in a square km area, while the situation is even worst considering urban population in major cities. The country is surrounded by the Bay of Bangle in the south, the Assam Hills in the east and lofty Himalayas to the north. Bangladesh is in the tropical monsoon region where typically warm temperature and high humidity are very common. Although there are six seasons in Bangladesh, the major four seasons are identified as pre-monsoon (March to May), monsoon (June to September), post-Monsoon (October to November) and winter (December to February).

Fig. 1

Location Map of CAMS stations in major cities of Bangladesh.

Among the months in a year, the dry period (November to April) generally experienced severe air pollution than other time of a year when relatively low humidity and little or no rain persist. The typical wind speed range 0.5 to 1 ms−1 (primarily northwesterly to east direction) and the temperature is found to vary between 15 °C to 25 °C. In this study the hourly concentration of PM10, PM2.5, NO2, SO2 and CO were collected from the available 11 continuous air monitoring stations (CAMS) located in 8 major cities in Bangladesh (see Fig. 1) during COVID-19 lockdown (26 March to 26 April 2020). A summary of few attributes of the CAMS and studied cities covering different climatic parameters and characteristics of urban settings is presented in Table 1 .

Table 1

Climatic attributes of studied cities in Bangladesh (Sources: BBS, 2015 and BMD, 2021).

Latitude	Longitude	Name of the stations	Name of the City	City Code	Populationa(thousands)	Rainfallb (mm)	Average Temperatureb (°C)	R. Humidity b (%)
					Minimum-Maximum (Average)[Coefficient of Variations, CV]
23.76	90.38	CAMS-1 (S-Bhaban)	Dhaka	DHK	13,798	0–318 (101)[0.884]	22.2–30.7(22.3)[0.058]	48–81(66.5)[0.101]
23.76	90.39	CAMS-2 (BARC)
23.78	90.36	CAMS-3 (D-Salam)
23.99	90.42	CAMS- 4 (Gazipur)	Gazipur	GAZ	4046
23.63	90.51	CAMS-5 (Narayangonj)	Narayangonj	NAR	3490
22.36	91.80	CAMS-6 (TV Station, Chittagong)	Chittagong	CTG	8990	0–418(96)[1.108]	23.9–29.6(22.1)[0.052]	62–84(75.5)[0.057]
22.32	91.81	CAMS-7 (Agrabad-Chittagong)
22.48	89.53	CAMS-8 (Sylhet)	Sylhet	SYL	4408	0–1004(261)[0.857]	22.0–28.1(19.9)[0.053]	46–84(72)[0.101]
24.36	88.61	CAMS-9 (Khulna)	Khulna	KHU	2650	0–347(64)[1.155]	23.7–32.3(22.4)[0.065]	57–86(74.5)[0.073]
24.89	91.80	CAMS-10 (Rajshahi)	Rajshahi	RAJ	3000	0–240(40)[1.102]	22.8–31.7(20.4)[0.085]	44–77(64)[0.116]
22.71	90.36	CAMS-11 (Barisal)	Barisal	BAR	2776	0–412(78)[1.113]	23.5–30.4(22.2)[0.058]	57–96(78)[0.067]

population in 2020.

Monthly average of March and April for last three decades.

As seen in Table 1, the populations in Dhaka and Chittagong are just few folds higher than the other cities that gives the impression of how these cities are struggling to manage healthy living housing for too many people in urban settings. The reason behind the higher population in Dhaka and Chittagong is due to the job opportunities in public and private sectors as these two cities houses a significant number of industries and commercial hubs, private offices, school and colleges in comparison to other cities in Bangladesh. Considering the climatic parameters, temperature doesn't show much variability while the monthly average rainfall amount during March and April is found to be varied from as low as 40 mm in Rajshahi to the highest of 261 mm in Sylhet. The relative humidity is found in the range of 64 to 78 within the studied cities. The fact of variation among the sites are confirmed and is even better explained taking range and coefficient of variation in consideration. As Dhaka and Chittagong cities primarily are with commercial and business hubs, the air is seemed to be polluted by emissions from the industries, road traffic sectors, development and constructions activities, household uses of gas as energy, open air burning of waste using fossil fuels etc. Masum and Pal (2020). In a nutshell, the geographical locations, urban settings and regional climate makes the differences in studied cities.

As seen in Fig. 1, among the major cities, Dhaka, Gazipur and Narayanganj are centrally located, Chittagong, Khulna and Barisal are port cities situated in the southern belt, Sylhet in the north eastern belt and Rajshahi is in western part of the country. Bangladesh is divided into eight major hydrologic regions align with air quality monitoring stations. It is seen that Dhaka, the capital of Bangladesh housed the highest population followed by Chittagong, Sylhet, Gazipur, Narayanganj, Rajshahi, Barisal and Khulna. The central belt of the country is seen to be over crowded housing approximately 50% of total population, while the rest is divided over other region (Table 1).

The data measuring tools for selected pollutants studied here are found as Met One Model BAM-1020 beta attenuation monitors (BAMs) for PM, ultraviolet fluorescence for SO2, non-dispersive infrared radiation (NDIR) for CO and chemiluminescence for NO2 managed the air quality monitoring stations operated by the Department of Environment (DoE). The measuring stations are well connected with central data server where data are stored in every hour from different stations and are then processed and published following quality control and quality assurance checked. Upon collection, data were checked for missing data, homogeneity and integrity. Thereafter the air quality index (AQI) of each pollutant was calculated using Eq. (1) as per US Environmental Protection Agencies (USEPA) guidelines (EPA, 2006). Maximum of the AQI estimated from the pollutants has been considered as AQI.Where, Ip=The index value for pollutant p, Cp=The concentration of pollutant p, BPHi=The Breakpoint ≥ Cp, BPLo=The Breakpoint ≤ Cp, I Hi=The AQI value corresponding BPHi and ILo=The AQI value corresponding BPLo.

The influenced of forced lockdown on changed air quality pollutants in Bangladesh was reviewed considering a few perspectives: firstly, spatiotemporal variations of hourly concentration of PM10, PM2.5, NO2, SO2 and CO were evaluated to identify the benefits of COVID-19 lockdown at different sites. The trend analysis was performed using Mann-Kendall (M-K) analysis with their significant level. The magnitude of change in concentration of air pollutants during COVID-19 were determined using Sen's Slope method. In Mann Kendall analysis, the number of sequential values in studied data series is denoted by n. If n is 9 or less, the absolute value of S is compared directly to the theoretical distribution of S derived by Mann and Kendall (Gilbert, 1987). The Mann-Kendall test statistic S is calculated by using following equation.

Where,xj and xk are the sequential data values. When S bears a positive value, it indicates an upward or increasing trend and if the value is negative, it indicates downward trend or decreasing trend. If n is at least 10 or more than 10, the test follows a normal distribution and hence normal approximation test is used with expectation (E) and variance of S as VAR(S) using the formulation shown below:

Here, q is the number of tied groups and tp is the number of data points in the pth tied group in the dataset. The standardized test statistic (Z) is calculated as follows:

Where, the value of Z is the Mann-Kendall test statistic which follows a standard normal distribution with mean being 0 and variance being 1. In this study, the confidence intervals of 90%, 95%, 99%, 99.99%(p < 0.10, p < 0.05, p < 0.01 and p < 0.001, respectively) were taken to classify the significance of positive and negative trends. Furthermore, the Sen's slope estimator, is known as the slope of the linear trend, has been estimated using the Theil–Sen estimator (Sen, 1968). It is a nonparametric method used to determine the true slope of an existing trend where the trend can be assumed to be linear (as change per time). The slope (Q) estimates of N pairs of data are first computed by as follows:

Where xj and xk are data values at times j and k (j > k) respectively. The median of these N values of Q is the Sen's estimator of the slope.

A spatial distribution of concentrations for both types of pollutants (particulate and gaseous) were done using ArcGIS version 10.4 statistics. The trends of changes in mean concentrations of the air pollutants presented in map were classified as extreme, high, moderate and mild considering respective confidence interval of 99.99% (***), 99% (**), 95% (*) and 90 (+) (p < 0.001, p < 0.01, p < 0.05, p < 0.10), respectively. Secondly, the descriptive statistics for the gaseous and particulate pollutants in air were discussed and compared for both lockdown period (March to April 2020) and typical period of same span for 2019 for identifying immediate impacts. As the AQI is a good communicator of air pollution status of an area, the numbers of days that exceeds BNAAQS limit of AQI were also counted and presented with descriptive statistics. In addition, the mean concentration of pollutants during COVID-19 lockdown period (March to April 2020) was also compared with that of found for the same period derived from the large data over 2012 to 2019 to evaluate further.

The correlation coefficients between the air pollutants and confirmed COVID-19 cases was performed using Spearman's rank correlation coefficient. Spearman's rank correlation coefficient (rs) is a method to determine the correlation between variables that are not normally distributed using a monotonic function. The coefficient can be determined using the following equation.where n indicates the number of alternatives, and di represents the difference between the ranks of two parameters. In this study, confidence intervals of 95% and 99% (p < 0.05, p < 0.01) were taken for the analysis. During correlation analysis, along with the current date COVID-19 cases with same day climatic variables, the climatic parameters that are lagged by 3, 5 and 7 days, were also considered based on the assumption that the people got affected on date may not necessarily exposed to same day climatic parameters as symptom of COVID-19 positive cases initiation takes couple of days to two weeks in many cases. Finally, to evaluate the relationship among air pollutants and the number of COVID-19 registered and published cases during the studied period, Principal Component Analysis (PCA) using statistical software SPSS (Ver. 20) was carried out. In PCA, the factor loadings for each of the member of air pollutants have been taken into consideration for clustering, and grouping of air pollutants are decided based on the variations revealed by the principle components (Mahapatra et al., 2012).

Results and discussions

Descriptive statistics of air pollutants

The descriptive statistics of air pollutants studied during COVID-19 lockdown (March to April 2020) and during the same period in 2019 along with their respective permissible limits set by GoB are presented in Table 2, Table 3 . Table 2 illustrate the gaseous pollutants CO, NO2 and SO2 concentrations at 8 different sites in Bangladesh. As seen in Table 2, in general gaseous pollutants concentrations are substantially lower than their respective standard concentrations set by GoB for both time periods. However, distinct reduction of the mean concentrations of CO and NO2 with various magnitude at different sites is evident during COVID-19 lockdown compared to their concentrations during typical period in 2019. The significant reduction of CO concentrations during COVID-19 lockdown are in Sylhet (14 times lower) and Gazipur (13 times lower), whereas at other cities reduction are with 6 to 8 times lower. In this line, the significant reduction of NO2 concentrations is in Sylhet (28 times lower) and Rajshahi (25 times lower), whereas, moderate reduction of 18 times is seen in Khulna and Barisal, while other cities showed 7 to 12 times low during lockdown period. Nonetheless, considering the maximum concentration, the concentrations of CO and NO2 are seen to exceed the standard limits (see Table 2) in Dhaka and Gazipur. The reduction of gaseous pollutants emission during lockdown in Bangladesh is primarily linked to the significant reduction in vehicle and industrial exhausts emission when complete shutdown of offices, business, industries, markets etc. was in place. While increase CO and NO2 levels are primarily from the heavy vehicle, lorries, hospital vehicles that were given permission to operate for medical emergency and transport of consumable goods for daily needs during complete shutdown. Based in particulate pollutants, it is seen that there are even greater number of exceedance cases than the limit concentrations as per Bangladesh standards (see Table 3). It is surprising that even in lockdown period, the maximum PM2.5 concentrations as presented in Table 3, are seen to exceed the limit value in Dhaka, Gazipur and Narayanganj illustrating unhealthy air environment with potential threat to human and aquatic lives being exposed to. However, taking the mean concentrations of PM2.5 and PM10 in attention, Sylhet showed significantly lower values of 22 and 35 μgm−3, respectively, while the highest PM2.5 and PM10 are found in Narayanganj (38 μgm−3) and Gazipur (58 μgm−3) during lockdown period in 2020 compared to previous period in 2019 where Sylhet exhibit the lowest PM2.5 (83 μgm−3) and Narayanganj showed the highest PM10 (158 μgm−3). The specific reason of higher concentrations of selected air pollutants even in lockdown period at aforesaid sites are linked to heavy vehicles transporting goods, transports related to medical emergencies, burning solid waste in open field primarily. In addition, exhaust emission from diesel fuel used for heavy trucks and lorries might have also seen the incremental emission in air. The PM concentrations reduction across the city during lockdown period in comparison to same period in 2019 merely range 3 to 6 times only describing the presence of finer dust in air suspension influenced by the climatic parameters. For example, Sylhet showed lower particulate matters in air due to amount of rain received during lockdown compared to other cities.

Table 2

Descriptive statistics of gaseous pollutants in air during COVID-19 lockdown (March – April 2020) and for same period in 2019

	Para-meters	DHK	DHK	GAZ	GAZ	NAR	NAR	CTG	CTG	SYL	SYL	KHU	KHU	RAJ	RAJ	BAR	BAR
CO(mg m−3)	Min	1.11	0.07	0.93	0.07	0.29	0.07	0.53	0.05	0.57	0.05	0.05	0.04	0.5	0.06	0.78	0.05
	Max	43.55	0.57	66.86	0.46	19.5	0.57	16.9	0.35	23	0.32	21.1	0.59	18.3	0.38	2.7	0.29
	Mean	2.98	0.32	3.93	0.31	2.1	0.31	1.72	0.24	3.11	0.22	2.41	0.29	1.62	0.28	1.46	0.23
	Std. dev.	5.65	0.09	10.6	0.08	3.59	0.09	2.48	0.06	3.94	0.05	4.61	0.1	3.17	0.07	0.39	0.05
	CV	1.89	0.28	2.69	0.27	1.71	0.28	1.44	0.25	1.27	0.23	1.91	0.33	1.95	0.26	0.27	0.2
	std. value⁎	40
NO₂(μg m−3)	Min	3.29	4	2.68	2.63	8.64	4.31	4.83	1.51	5.3	0.64	17.9	1.83	0.96	1.85	1.64	0.83
	Max	276	58.6	120	26.8	204	62.4	152	10.9	247	4.7	238	27.2	147	10.3	175	6.6
	Mean	129.8	12.23	57.3	8.64	99	12.2	45.2	3.72	50	1.79	91.8	4.7	84.5	3.32	33.4	1.87
	SD	67.47	10.24	17.2	5.92	45.7	10.6	32	2.25	42.5	1.15	57	4.57	40.6	1.9	33.6	1.13
	CV	0.98	1.59	1.07	1.28	0.87	1.64	1.34	1.13	1.6	1.21	1.17	1.83	0.91	1.08	1.89	1.15
	std. value⁎	100
SO₂(μg m−3)	Min	6.24	6.29	1.60	5.64	12	5.87	0	3.1	2.06	0.83	3.91	1.77	0	1.1	1.17	0.71
	Max	127.4	31.3	85.9	23.8	87.2	31	101.8	16.7	22.5	20.9	91.35	58.5	54.3	10.44	100.8	12.19
	Mean	32.62	12.7	21	13.8	39.7	11.7	17.7	8.6	9.53	7.9	24.17	12.19	15.95	5.45	21	5.38
	Std. dev.	26.36	4.96	20.5	4.72	20.4	4.93	18	3.52	5.43	5.51	24.4	11.38	14.17	2.8	23.65	3.39
	CV	2.11	1.02	2.56	0.88	1.36	1.1	2.6	1.07	1.49	1.83	2.62	2.41	2.32	1.33	2.92	1.64
	std. value⁎	365

DHK – Dhaka; GAZ – Gazipur; NAR – Narayanganj; CTG – Chittagong; SYL – Sylhet; KHU – Khulna; RAJ – Rajshahi; BAR - Barisal.

The values of pollutants under city codes (e.g. DHK) are in normal condition previous year, while theses under city codes underlines (e.g. DHK) are lockdown period.

Bangladesh National Ambient Air Quality Standard (BNAAQS); Values shown in bold exceed the BNAAQS limit values.

Table 3

Descriptive statistics of particulate pollutants in air during COVID-19 lockdown (March – April 2020) and for same period in 2019⁎

	Para-meters	DHK	DHK	GAZ	GAZ	NAR	NAR	CTG	CTG	SYL	SYL	KHU	KHU	RAJ	RAJ	BAR	BAR
PM2.5(μg m−3)	Min	45.0	5.0	38.9	4.2	35.2	8.3	30.3	6.6	27.5	8.0	25.3	17.3	44.9	12.4	33.1	7.9
	Max	212.7	117.3	220.9	116.5	268.2	120.6	200.0	62.6	151.7	39.9	217.0	61.7	175.1	56.4	213.2	37.8
	Mean	128.4	34.7	136.0	34.0	157.9	38.1	102.3	25.5	83.1	22.1	104.4	34.4	104.6	33.5	113.4	24.2
	Std. dev.	41.8	23.3	43.8	23.3	62.3	23.3	36.2	12.3	27.7	9.1	54.4	11.9	32.4	11.5	43.4	8.6
	CV	0.33	0.67	0.32	0.69	0.39	0.61	0.35	0.48	0.33	0.41	0.52	0.35	0.31	0.34	0.38	0.35
	Std. value	65
PM₁₀(μg m−3)	Min	97	29.5	104	32.7	117	25.2	82.4	14.3	58	11.6	0	29.3	105	20.4	67.2	11.5
	Max	371	75.9	319	81.6	428	70.4	292	103	227	59.5	233	93.9	330	86.3	285	63
	Mean	214	54.6	223	58	296	51.1	180	51.6	146	35.2	131	54.3	211	50.7	168	39
	SD	63.3	13.7	59.1	14.4	78.2	12.7	47.6	18	40.4	14	62.2	18.1	58.9	18.1	51.8	14.2
	CV	0.3	0.25	0.27	0.25	0.26	0.25	0.26	0.35	0.28	0.4	0.47	0.33	0.28	0.36	0.31	0.36
	Std. value	150
AQI	Min	97	0.32	98.9	0.31	77	0.31	82.4	0.24	58	0.22	29.2	0.29	105	0.28	41.3	0.23
	Max	353	181	334	180	385	183	441	144	223	98.9	354	143	296	132	303	94.9
	Mean	219	83.3	216	82.5	247	89	202	67.4	150	60.1	170	85.7	194	83.7	173	64.9
	Std. dev.	71.6	37	73.4	36.7	86.7	36.7	84.4	28.8	45.6	23.1	94.4	28.4	56.2	28	77.8	22.1
	CV	0.33	0.44	0.34	0.44	0.35	0.41	0.42	0.43	0.3	0.38	0.56	0.33	0.29	0.33	0.45	0.34
	std. value	100
No. of days AQI ≥ 100	43	7	49	7	52	9	31	3	19	0	43	13	40	10	23	0	No. ofdays AQI ≥ 100

DHK – Dhaka; GAZ – Gazipur; NAR – Narayanganj; CTG – Chittagong; SYL – Sylhet; KHU – Khulna; RAJ – Rajshahi; BAR - Barisal.

The values of pollutants under city codes (e.g. DHK) are in normal condition previous year, while theses under city codes underlines (e.g. DHK) are lockdown period.

Bangladesh National Ambient Air Quality Standard (BNAAQS); Values shown in bold exceed the BNAAQS limit values.

The air quality index (AQI) biased heavily by PM2.5 showed very similar trend in lockdown period highlighting that except Barisal and Sylhet, the maximum values exceeding Bangladesh standard AQI of 100 are found in Rajshahi (132), Khulna and Chittagong (143) while at Dhaka, Gazipur and Narayanganj around 180. The similar findings are also reported studies elsewhere by Kerimray et al. (2020), Sharma et al. (2020) and Tobías et al. (2020). The urban settings around Dhaka and Chittagong see the dust pollution in dry period influenced by the construction activities, dust derived from the congested road traffic sectors along with industrial emissions. As most of the garment industries, other industries, number of vehicles are located in Dhaka, Gazipur, Narayanganj and Chittagong, the finer fraction of particulate emissions are higher than other cities in Bangladesh. For example, 13 cement factories and 320 brick kilns are in Narayanganj and 410 brick kilns around Gazipur compared to 200 kilns in Dhaka and 100 in Chittagong and much lower in other cities. Nonetheless, the northeasterly wind in dry period may also contribute particulate matters blowing out dust from part of India and Nepal to Bangladesh may further add to the existing amount. Among the cities, distinct variations are therefore expected and usual as discussed. Based on the air pollutants at different cities, it is seen that Sylhet appears to be clean city while city around central part of Bangladesh e.g. Narayanganj, Dhaka and Gazipur appear severely polluted, while Rajshahi, Khulna and Chittagong seem to be moderately polluted. The standing of cities in relation to air pollution in this study slightly differ from previous studies by CASE-DoE (2019), Begum et al. (2013) and Hoque et al. (2014) though Sylhet still ranked first influenced by significantly higher rainfall around the year compared to other part of Bangladesh.

Temporal trends of air quality pollutants during COVID-19 lockdown

Results of trend analysis for gaseous air pollutants during COVID-19 in 8 major cities of Bangladesh is shown in Table 4 and Fig. 2 . Table 4 presents the rate of change while the graphical representations using map as shown in Fig. 2 illustrates level of significance of those changes at 8 different cities under investigation. As seen in Table 4 and Fig. 2, both positive and negative trends are observed during lockdown period. Except Sylhet and Rajshahi, all other cities showed positive rate of changes of NO2 concentrations with high level of significance (see Fig. 2). In the central portion of country, the cities i.e., Dhaka, Gazipur, Narayangonj show extreme increase of NO2 at a rate 0.135 ppb day−1, 0.124 ppb day−1 and 0.118 ppb day−1. The similar trend is found for CO with flat increment and decrement. Considering SO2 concentrations, except Gazipur that exhibits statistically slight increase rate of 0.021 ppb day−1, Sylhet and Barisal showed substantially higher reduction rate of 0.057 ppb day−1 and 0.086 ppb day−1, respectively, whereas Chittagong and Khulna revealed moderate reduction at rate of 0.057 ppb day−1 and 0.126 ppb day−1, respectively.

Table 4

Daily changes in mean concentration of gaseous air pollutants during COVID-19 lockdown period.

Pollutants	Dhaka	Gazipur	Narayangonj	Chittagong	Sylhet	Khulna	Rajshahi	Barisal
CO	0.003	0.003	0.002	−0.003	−0.003	0.004	−0.005	−0.002
SO2	−0.023	0.021	−0.046	−0.057	−0.181	−0.126	−0.055	−0.086
NO2	0.136	0.124	0.118	0.018	−0.005	0.011	−0.034	0.061

Fig. 2

Spatial variability of gaseous air pollutants trends across major cities of Bangladesh during lockdown period (March–April, 2020). In the key- Extreme trend (99.99% confidence level), High trend (99% confidence level), Moderate trend (95% confidence level), Mild trend (90% confidence level or less).

While the decreasing trends are expected, the increasing trend is unexpected and surprising in shutdown period. The increasing trend probably link to emission from emergency service vehicles allowed in shutdown period, e.g., ambulances, utilities services vans, lorries, trucks for medical emergencies and transportations of consumable goods for daily needs influenced by city wide attributes, such as population, offices, industries, hospitals etc. and climatic variabilities.

Furthermore, the concentrations for gaseous air pollutants during COVID-19 lockdown (March–April) in 8 major cities of Bangladesh were compared with the mean concentrations derived for data of March–April in last 7 years (2013–19) and with March–April of 2019 to check the progressive and immediate changes. The results as obtained is presented in Table 5 . As seen Table 5, among the gaseous air pollutants, the maximum reduction in concentration occurs in case of NO2 ranges 75% to 93% across the cities considering 2013 to 2019, while that with 2019, the range is as 80% to 96%. The results yield similar findings of decreasing trends for CO and SO2 (except in Dhaka and Gazipur showing positive compared to 2019 data) across the sites, as discussed earlier illustrating lowest reductions are in the central belt of Bangladesh. A graphical mapping in relation to this can be seen in supplementary file as Fig. S1. Although the magnitude of changes is marginally higher when compared to recent year data (2019), however, no distinct changes in gaseous pollutants are seen for the period of 2013 to 2020, except forced lockdown brought the values down, as expected.

Table 5

Percentage (%) of changes in concentrations of gaseous air pollutants in different cities of Bangladesh.

Name of the City	March–April (2013–19) Vs COVID-19 Lockdown			March–April (2019) Vs COVID-19 Lockdown
	CO	SO₂	NO₂	CO	SO₂	NO₂
Dhaka	−80	−33	−84	−81	38	−87
Gazipur	−80	−9	−75	−91	10	−90
Narayanganj	−46	−51	−75	−76	−65	−80
Chittagong	−81	−35	−93	−79	−79	−90
Sylhet	−88	−29	−92	−68	−2	−96
Khulna	−72	−51	−90	−59	137	−89
Rajshahi	−67	−67	−90	−72	−60	−94
Barisal	−72	−67	−86	−81	−75	−95

Results of trend analysis for particulate matters in air during COVID-19 lockdown measures in 8 major cities of Bangladesh is shown in Table 6 and Fig. 3 . Table 6 presents the rate of change while the graphical representations of changes using map is shown in Fig. 3 with the level of significance of those changes found at 8 different cities under investigation. In addition, Table 6 also present the changes in AQI during several phases along with particulate pollutants. Except PM2.5 in Khulna, all cities exhibit statistically significant reduction in particulate pollutants and AQI considering the mean concentrations during COVID-19 lockdown, as seen in Table 6 and Fig. 3. Furthermore, from Table 6, the reduction changes of PM2.5 concentrations are substantially higher (see Fig. 3) in the order of Barisal, Chittagong, Rajshahi and Sylhet at a rate of 1.573, 0.813, 0.801 and 0.711 μgm−3 day−1, respectively, while the reduction rate are much lower around the central belt e.g., in Dhaka, Gazipur and Narayanganj. Based on PM10 changes the standings are found slightly different in order at a rate of −2.995, −1.317, −1.095 and − 0.887 μgm−3 day−1, respectively, for Barisal, Chittagong, Sylhet and Rajshahi, respectively. The AQI standing follows the PM10 trend across the sites. It is noted that the results follow the similar trends of gaseous pollutants as found in Table 5 confirming city-specific urban settings and climatic variabilities.

Table 6

Daily changes in mean concentration of particulate matters and AQI for the major cities in Bangladesh during COVID-19 lockdown periods (March–April, 2020).

Pollutants	Dhaka	Gazipur	Narayanganj	Chittagong	Sylhet	Khulna	Rajshahi	Barisal
PM2.5	−0.335	−0.334	−0.338	−0.813	−0.711	0.062	−0.801	−1.573
PM10	−0.636	−0.455	−0.642	−1.317	−1.095	−0.001	−0.887	−2.995
AQI	−0.618	−0.589	−0.578	−1.810	−1.686	0.123	−1.634	−3.423

Fig. 3

Spatial variability of particulate air pollutants trends across major cities of Bangladesh during lockdown period (March–April, 2020). In the key- Extreme trend (99.99% confidence level), High trend (99% confidence level), Moderate trend (95% confidence level), Mild trend (90% confidence level or less).

Furthermore, the concentrations for particulate air pollutants along with AQI during COVID-19 lockdown (March–April) in 8 major cities of Bangladesh were compared with the mean concentrations derived for data of March–April in last 7 years (2013–19) and with March–April of 2019 to check the progressive and immediate changes. The results as obtained is presented in Table 7 .

Table 7

Percentage (%) of changes in concentrations of particulate matters in air and AQI in Bangladesh.

Name of the City	March–April (2013–19) Vs COVID-19 Lockdown			March–April (2019) Vs COVID-19 Lockdown
	PM₂.₅	PM₁₀	AQI	PM₂.₅	PM₁₀	AQI
Dhaka	−44	−53	−36	−60	−67	−53
Gazipur	−45	−54	−28	−62	−65	−39
Narayanganj	−38	−70	−23	−62	−80	−46
Chittagong	−60	−57	−47	−62	−66	−66
Sylhet	−55	−63	−35	−64	−72	−49
Khulna	−13	−19	−16	−31	−41	−54
Rajshahi	−43	−56	−36	−66	−64	−48
Barisal	−46	−50	−21	−46	−57	−35

As seen Table 7, both PM2.5 and PM10 show decreasing trends during COVID-19 lockdown period (March to April 2020) and in comparison, with the previous year data of March to April 2019 yields 31 to 66% reduction across the sites, while that for the same period of 2013 to 2019 are found 13% to 60%. A graphical representation across the site can be found in Fig. S2. Although few differences in changes across the sites are observed as expected and discussed earlier influenced by the city-specific attributes, the distinct changes are hardly seen as reported in the study carried out by Travaglio et al. (2020) taking a greater number of cities in the UK in consideration. The results by analyzing the data and information for this study it is clear that the immediate benefits of air pollutants reduction gained since the government adopted few measures such as introduction to cleaner fuel CNG for motor vehicles, enlarged height of brick kilns and industrial chimneys, phase out of 2 stroke 3 wheeler etc. are suppressed by the unplanned urbanizations, increased number of vehicles, industrializations and lack of control and monitoring schemes by DoE, and hence, at later stages enhanced air pollution around central belts, such as in Dhaka, Gazipur Narayanganj and in commercial and port city Chittagong are evident.

Correlation between air pollutants and confirm COVID-19 cases

A limited study to date in the UK by Travaglio et al., 2020 and in few cities in Europe by Ogen, 2020 analyzes the relationship between COVID-19 cases with air pollutants, and found poor to weak correlation. However, no such relationship has been found for south Asian countries so far. Moreover, relationship with the COVID-19 cases and time lag of climatic variables by 3, 5 and 7 days on date of positive cases is evaluated considering the fact that people registered today as COVID positive may got affected a week earlier.

Table 4 displayed the correlation coefficients among air quality pollutants with COVID-19 positive cases registered across the sites in Bangladesh to explore whether air pollutants may have direct or indirect relationship to COVID-19 cases and fatality in Bangladesh. As seen in Table 8 , both positive and negative coefficients with poor to moderate strength (0.360 to 0.627 considering statistically significant relations only) are found to exists across the 8 different sites in Bangladesh. While the number of statically significant correlation coefficients are very limited with COVID-19 cases at different sites, the results are in consistent with previous studies to date elsewhere. The greater number of statistically significant correlations are in Dhaka, Sylhet and Barisal, while Gazipur, Narayanganj, Khulna, Rajshahi exhibit couple of significant correlation coefficients, Chittagong revealed none. Based on the standing of the cities the central belt cities e.g., Dhaka, Gazipur and Narayanganj are found with polluted air environment even in lockdown period compared to other cities around the country and in accordance with higher number of positive cases; however, except Dhaka other two cities did not yield positive correlations with COVID-19 positive cases which is unexpected and left an idea that COVID-19 transmission with air may only be harmful for densely living style that is more dominant in Dhaka city. However, positive correlations of SO2, PM2.5, PM10 and NO2 in Dhaka, PM, SO2, CO and NO2 in Sylhet and similarly in Barisal support the idea of correlations with COVID-19 positive cases and lethality as reported by Travaglio et al. (2020) and Ogen (2020). While, it is well understood that usually people got affected experienced symptoms within a week or so and thereafter got tested and registered in the data base, therefore 3-, 5- and 7-days lag air pollutants with on date COVID-19 positives cases are evaluated and it has been seen that the 5 days lag air pollutants revealed greater number of statistical relationships. This further indicates that the relationship between air pollutants and COVID-19 cases are not straightforward rather complex and need more data to draw concrete decision.

Table 8

Correlation coefficients (r) among the air pollutants and confirmed COVID-19 patients in Bangladesh.

Parameters	Dhaka	Gazipur	Narayanganj	Chittagong	Sylhet	Khulna	Rajshahi	Barisal
PM2.5 (On day)	−0.320	−0.459⁎	−0.305	0.039	−0.346	−0.093	−0.216	−0.129
PM2.5 (3 days lag)	−0.059	0.239	−0.382	0.061	−0.307	0.210	−0.124	0.077
PM2.5 (5 days lag)	−0.308	−0.164	0.218	0.288	0.598⁎⁎	0.379	0.122	0.486⁎
PM2.5 (7 days lag)	0.512⁎	0.115	0.020	0.145	−0.141	0.121	0.344	0.477⁎
PM10 (On day)	0.448⁎	−0.130	0.360⁎	0.139	−0.366	−0.147	−0.184	−0.076
PM10 (3 days lag)	0.568⁎⁎	0.025	0.032	0.240	−0.311	0.237	−0.179	0.092
PM10 (5 days lag)	0.637⁎⁎	−0.041	0.415	0.367	0.622⁎⁎	0.259	0.193	0.583⁎⁎
PM10 (7 days lag)	0.297	0.133	0.198	0.099	−0.075	0.010	0.388	0.392
CO (On day)	−0.024	0.059	−0.025	−0.097	−0.234	0.061	−0.267	−0.377
CO (3 days lag)	0.069	−0.092	−0.090	−0.085	−0.385	0.410⁎	−0.267	0.083
CO (5 days lag)	−0.148	0.072	0.511⁎	0.373	0.577⁎⁎	0.387	0.328	0.189
CO (7 days lag)	−0.238	0.210	−0.119	0.059	−0.176	0.158	0.291	0.236
SO2 (On day)	−0.327	0.114	−0.285	0.056	−0.339	−0.203	−0.155	−0.289
SO2 (3 days lag)	−0.237	0.113	−0.188	0.040	−0.337	0.365	−0.191	−0.051
SO2 (5 days lag)	0.428⁎	0.222	0.555⁎	0.352	0.559⁎	0.385	0.299	0.370
SO2 (7 days lag)	−0.270	0.057	−0.143	0.156	−0.247	−0.449	0.504⁎	0.477⁎
NO2 (On day)	0.371	0.166	0.387⁎	0.046	0.088	0.179	−0.088	0.271
NO2 (3 days lag)	0.415⁎	0.342	0.086	−0.056	0.145	−0.113	0.549⁎⁎	−0.066
NO2 (5 days lag)	0.119	−0.230	0.373	0.306	0.513⁎	0.194	0.372	0.257
NO2 (7 days lag)	0.046	−0.126	−0.363	0.102	−0.099	0.165	0.018	−0.011

The values kept in bold are the values that are found statistically significant correlations with different confidence levels (as mentioned in the table foot note).

*significant at 95% confidence level (p < 0.05); ** significant at 99% confidence level (p < 0.01).

The limited data of COVID-19 positive cases during the lockdown period studied and its murky relation with air quality reported earlier, it is therefore decided to conduct the principle component analysis (PCA) to identify factor loadings of similar groups taking data from 8 different cities of Bangladesh following similar investigation approach by Travaglio et al. (2020) where authors performed correlation analysis gathering data from 120 locations across the UK. The results as obtained from PCA seen two variance groups, as shown in Fig. 4 .

Fig. 4

PCA plots of air pollutants, AQI and COVID-19 positive cases during COVID-19 lockdown period.

Two major clusters of variances (PC1 and PC2) have been found for 7 variables based on their factor loading. In the cluster analysis, particulate matter PM2.5, PM10 and AQI belong to PC1, whereas the gaseous pollutants SO2, NO2, CO and COVID-19 confirmed positive cases belong to PC2. It has been seen that COVID-19 confirmed cases is moderately correlated with gaseous air pollutants. Among the gaseous air pollutants, NO2 shows relatively higher association than CO and SO2. The finding is found in consistent with previous two studies by Travaglio et al., 2020 and Ogen, 2020. Furthermore, PCA has also been performed taking 3-, 5- and 7-days lag air pollutants data with on date COVID-19 cases (see Fig. S3) along with eigen values and PCA loading factors (see Table S1) to get more insight of the relationship or association. It has been seen that the PCA displayed 3D plot with five components, while PC1 and PC4 describing the most positive cases and association between air pollutants (on date and 3-, 5- and 7-days lag). As in Bangladesh the gaseous pollutants are seen substantially low in concentrations across the country even in typical dry period compared to particulate matter, it may help to keep the relatively lesser number of COVID-19 positive cases and death in Bangladesh as well as in countries located in South East Asia to date. While in comparison to the affected population and death tolls elsewhere in the USA, UK and in Europe where air pollution is defined primarily by gaseous pollutants than particulate matters (Tobías et al., 2020; Kerimray et al., 2020; Travaglio et al., 2020; Ogen, 2020). However, to draw a confirmed correlation, inclusion of meteorological parameters, socio-economic factors, virus genome structures etc. need also to be examined. Nevertheless, this relationship pointed out the health affecting air pollutants may remarkably have increased vulnerability to the COVID-19 positive cases and death tolls.

Conclusions

The study evaluates and characterize air pollutants influenced by COVID-19 lockdown for major cities in Bangladesh, and thereafter, examine what is not yet clear whether air quality has some form of relationship with COVID-19 positive cases. The outcomes of the study found that air quality index represents air pollution status in Bangladesh is biased by particulate matters than gaseous pollutants. The investigation revealed that COVID-19 lockdown had substantially reduced air pollutants and thus had positive impact on the air quality in general across the major cities in Bangladesh. Nevertheless, in comparison to air quality data for the same period over 2013 to 2019 as pre-lockdown period, the lockdown exhibit much reduction of CO, SO2, NO2, PM2.5 and PM10 concentrations accounting 65 to 90%, 30 to 80%, 75 to 95%, 30 to 60% and 20 to 80%, respectively, across the eight major cities. The results confirmed substantial spatial variation across the cities, where cities located in the central belt, such as Dhaka, Gazipur and Narayanganj are in the lower range of reduction than other part of the country. Based on temporal trends during COVID-19 lockdown, both positive and negative trends of the air pollutants are observed: while PM2.5, PM10 and SO2 portray negative trends in general for all the cities, the CO and NO2 show unexpectedly increasing trends especially in Dhaka, Gazipur and Narayanganj. Correlation coefficients among the air pollutants and COVID-19 cases across the city depict foggy relationship. However, PCA suggests interesting facts of showing greater association with gaseous pollutants notably with NO2 with COVID-19 positive cases, while no such with particulate pollutants are seen integrated over the major cities in Bangladesh.

Align with studies elsewhere, it is quite clear more work is needed because of significant differences exists in weather pattern, socio-economic environment settings, law and order, virus genome structures and its adaptability with different settings across the countries in the world. Nonetheless, this is clear that health conditions affected by the air pollutants are linked to increase vulnerability to COVID-19.

Declaration of Competing Interest

No potential conflict of interest was reported by the authors. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.