Skip to main content
PMC home page
Elsevier - PMC COVID-19 Collection logoLink to Elsevier - PMC COVID-19 Collection
. 2021 May 4;198:111236. doi: 10.1016/j.envres.2021.111236

Impact of COVID-19 lockdown on NO2 and PM2.5 exposure inequalities in London, UK

Vasilis Kazakos a, Jonathon Taylor b, Zhiwen Luo a,∗
PMCID: PMC9750168  PMID: 33957139

Abstract

Amid the COVID-19 pandemic, a nationwide lockdown was imposed in the United Kingdom (UK) on March 23, 2020. These sudden control measures led to radical changes in human activities in the Greater London Area (GLA). During this lockdown, transportation use was significantly reduced and non-key workers were required to work from home. This study aims to understand how population exposure to PM2.5 and NO2 changed spatially and temporally across London, in different microenvironments, following the lockdown period relative to the previous three-year average in the same calendar period. Our research shows that population exposure to NO2 declined significantly (52.3% ± 6.1%), while population exposure to PM2.5 showed a smaller relative reduction (15.7% ± 4.1%). Changes in population activity had the strongest relative influence on exposure levels during morning rush hours, when prior to the lockdown a large percentage of people would normally commute or be at the workplace. In particular, a very high exposure decrease was observed for both pollutants (approximately 66% for NO2 and 19% for PM2.5) at 08:00am, consistent with the radical changes in population commuting. The infiltration of outdoor air pollution into housing modifies the degree of exposure change both temporally and spatially. Moreover, this study shows that the impacts on air pollution exposure vary across groups with different socioeconomic status (SES), with a disproportionate positive effect on the areas of the city home to more economically deprived communities.

Keywords: COVID-19, Lockdown measures, Population activity, Population-weighted exposure change, Concentration change, Socioeconomic inequalities

1. Introduction

Ambient air pollution levels are strongly associated with human activities, such as transportation, and can have significant population health impacts; in the UK, for example, air pollution is thought to contribute 28,000 to 36,000 excess deaths a year (PHE, 2019). On March 23, 2020, the UK government imposed a nationwide lockdown due to increasing transmission of coronavirus, which subsequently led to radical changes in human activities including the transportation and time-activity behaviours of the population. The COVID-19 lockdown offers a unique natural experiment to evaluate and quantify the impact of rapid changes in people's activity patterns and emissions on air pollution and subsequent population exposure.

Numerous studies have already evaluated the impact of COVID-19 lockdowns on outdoor air quality worldwide (Muhammad et al., 2020). The vast majority of these studies show that radical shutdown measures in big cities led to lower and less variable outdoor concentrations of urban air pollutants (Mahato et al., 2020, Mbandi, 2020, Sharma et al., 2020, Tanzer-Gruener et al., 2020, Zhao et al., 2020). However, most of these studies focus solely on the reduction of outdoor concentrations, and a relative few studies have assessed the impact of lockdowns on population exposure to urban air pollution (Williams, 2020; Zhu et al., 2020). This is important, as exposure to outdoor air pollution also occurs in non-outdoor microenvironments (MEs) due to the infiltration of polluted air; for example, housing is thought to significantly modify population exposures (Taylor et al., 2014).

Exposure is also dependent on the time-activity profiles of the population. In cities under lockdown, much of the population radically changed their daily activities, including working from home instead of their usual workplace and by avoiding all unnecessary travel. For example, the lockdown led to a greater than 70% decrease in public and private transportation in London, likely reducing exposure to outdoor generated air pollution (Williams, 2020). Therefore, to assess spatial and temporal changes in exposure during the lockdown, key factors such as changes in population activity patterns and concentrations in different microenvironments, where people spent their daily time (for example at home, workplaces, in transit, and outdoors) need to be considered.

In addition, there may be important differences in exposure between population groups. Socioeconomic inequalities in concentration and exposure to outdoor pollution are well established (Tonne et al., 2018; Shiels et al., 2017; Stringhini et al., 2017; Rivas et al., 2017; Rotko et al, 2000, 2001) and there is emerging evidence of similar disparities indoors (Ferguson et al., 2020). Several studies have shown a strong connection between communities of either lower or higher socioeconomic position and increased concentrations and exposure to urban air pollution (Hajat et al., 2015). In US, several studies have shown that deprived areas experience higher levels of outdoor pollution exposures (Su et al., 2012; Gray et al., 2013; Hajat et al., 2015). In London and other big cities, it has been suggested socioeconomic inequalities in outdoor levels of traffic-related air pollution are driven by differences in road traffic volume, which affects the amount of emissions (Brook and King, 2017, Padilla et al., 2014, Tonne et al., 2018) Therefore, changes in road traffic following the lockdown provide a unique natural experiment opportunity to investigate exposure disparities across socioeconomic groups, with potential changes in outdoor generated air pollution resulting in differences in exposure changes across different such groups.

In this study, we seek to 1) understand how the COVID-19 lockdown changed population-level outdoor air pollution exposures, and 2) evaluate whether changes in exposures varied across socioeconomic groups and explore the role of traffic-related pollution on exposure inequalities.

To achieve this, we aim to quantify and illustrate the spatio-temporal change in population exposure to outdoor-generated air pollution in London during the lockdown period relative to previous years for the same period. By accounting for the spatial and temporal variability of outdoor air pollution, dwelling Indoor-Outdoor (I/O) ratios (the proportion of outdoor air pollution that infiltrates indoors), and changes in diurnal population activity patterns, we assess the impact of the lockdown on the population exposure levels. Moreover, we also evaluate socioeconomic differences in exposure reduction. Understanding the spatial and temporal distribution of air pollution across different MEs, and subsequent exposure inequalities, is important to develop policies to reduce inequalities and improve sustainable development.

2. Material and methods

2.1. Study period and air quality data

Lockdown measures were applied to the Greater London Area (GLA), UK, on March 23rd, 2020. To examine the impact on short-term air quality, a one-month period (23 March to 22 April) in 2020 was compared against the same calendar period, averaged from 2017 to 2019. The hourly monitoring data for two major traffic-related air pollutants (NO2 and PM2.5) were obtained from the London Air Quality Network (LAQN) (King’s College London,). For NO2, 98 monitoring sites were included, whereas for PM2.5 only 21 monitoring sites were available for the study period. Average hourly concentrations for each hour of the study period were calculated for each monitoring site. The Voronoi Neighbor Averaging (VNA) tool in QGIS was used to spatially interpolate hourly data between monitoring sites, estimating hourly outdoor concentrations pre and post-lockdown at Lower-Super Output Area (LSOA) level (a census unit with an average of 1500 residents).

2.2. Microenvironments and infiltration of outdoor pollutants

Four different MEs have been considered in this work:

  • (i)

    The home;

  • (ii)

    Work, assuming that all individuals work inside buildings;

  • (iii)

    Transport, including public or private transportation (i.e., bus, private car/taxi and train) to travel; and

  • (iv)

    Outdoors, including people who are walking or cycling.

We only consider exposure to outdoor-generated pollution. Estimates of indoor pollution from indoor sources are highly uncertain and have not been considered in this study due to a lack of data. The infiltration of outdoor NO2 and PM2.5 into the home ME was considered using previously derived hourly I/O ratios across GLA for the same calendar period (Taylor et al., 2014). This data includes the hourly average I/O ratios of 1.5 million London dwelling (covering approximately 46% of London dwellings), and accounts for seasonal wind pressures and summertime window opening; here we use hourly average dwelling I/O ratios for April to represent the lockdown period, averaged by LSOA (Fig. S1). For both pollutants, central London shows the lowest I/O ratios, likely due to the newer building stock and the large number of flats in multi-dwelling buildings, where the available surface for infiltration is considerably smaller. The average I/O ratio in the GLA ranges from 0.40 to 0.63 for PM2.5 and 0.15–0.40 for NO2. The I/O ratio is likely to significantly modify population exposure to outdoor air pollution due to the extended amount of time that people spent at home during the lockdown period.

The spatially and temporally resolved I/O ratios provided by Taylor et al. (2014) have been derived only for domestic buildings and are not representative of commercial areas and workplaces. Thus, for the workplace, we have selected representative values according to the available literature. For PM2.5, we selected an average value of 0.60 (Singh et al., 2020; Soares et al., 2014; Hänninen et al., 2011) and for NO2, we chose to use an average value of 0.68 (Hu and Zhao, 2020; Kornartit et al., 2010).

For outdoor air pollution exposure in the transportation ME, we calculated the in-vehicle concentration using a mass balance equation (Smith et al., 2016). The same input values as Smith et al. (2016) were used except for the outdoor concentrations which were updated. As in Smith et al. (2016), the surface area of each commuter was derived as per Song et al. (2009).

2.3. London population data and activity

The spatial distribution of the London population was derived from 2011 census data from the Office of National Statistics (ONS, 2011), representing 95% of households, and was assumed to be the same in both the baseline and study periods. The spatial distribution of the population was considered during daytime (defined as the period from 7:00am to 19:00) population (Fig. S2a in SI) and night-time (defined as the period from 20:00 to 06:00am) population (Fig. S2b in SI). The Census usual resident population was used for the night-time period and the workday population for the daytime period. As expected, under normal circumstances the daytime population density is much higher in Inner London due to much of the population commuting into the city centre, whereas the night-time distribution is much more uniform across the GLA.

2.3.1. Pre-COVID

For the pre-COVID-19 period, we analyzed the amount of people at home, at work, in transportation and outdoors using Census and London Travel Demand Survey (LTDS, 2011) data. We used the Census workplace population (the number of people in each LSOA that were in their workplace during a usual weekday) to calculate the percentage of people normally at work. From the LTDS, the total number of trips per hour of a weekday, and the number of average trips per person were used to calculate the number of people that use public transportation each hour. As there was no data on the movements of populations in each LSOA, the temporal variation of the percentage of people in each ME was estimated using the LTDS data and the daytime and night-time population distributions. LTDS also provides data on the number of people commuting at each hour, defined as travelling between the home and workplace. Thus, at each hour the respective number of commuters was subtracted from the workplace population.

The diurnal variation of the population activity in the four MEs is presented in Fig. 1 A. During the morning and afternoon rush hours, the percentage of people in the transportation ME peaks. During daytime, more than 30% of people are either at work or in transportation, while at the night after 22:00 more than 90% of the population are at home. In this study, children were included in the home population.

Fig. 1.

Fig. 1

Open in a new tab

Diurnal variation of the percentage of people in each ME: a) during the pre-COVID-19 period and b) during the COVID-19 period (first lockdown).

2.3.2. COVID lockdown

For the COVID-19 period, changes in population daily movements between MEs were obtained from App Maps (Apple,) and Google statistics (Google statistics,). Google statistics used the median value of each day of the week in January 2020 (i.e., 5-week period from 3 January until 5 February) as baseline, while App Maps used January 13, 2020. Both datasets show significant changes in population travel and working behavior after March 23rd, with transportation reduced by more than 70%, and more than 75% of the working population remaining at home. The remaining population at work during the lockdown period likely consists largely of key workers, who continued going to their workplace. The data also shows that less than 1% of the total population are outside at most hours of the day. This data was used alongside spatial distribution of the usual resident (night-time) population in order to estimate the variation of the percentage of the population in each ME during the COVID-19 period (Fig. 1B).

2.4. Population-weighted exposure

The population-weighted mean exposure (PE) is estimated from the concentration level in each ME and the amount of people that spent time in those MEs. The PE was calculated as:

PE=∑i=1nCi,t,j×Pi,t,jPT (1)

where PE is the population weighted mean exposure for a population, n is the number of the populated geographical units (here LSOAs); C and P are the mean concentration of the pollutant and the number of people, respectively, for LSOA i, microenvironment j and hour t of the day; and P T is the respective total population.

2.5. Socioeconomic analysis

To compare concentration and exposure across socioeconomic status (SES), we used LSOA – level deprivation data from the 2019 Index of Multiple Deprivation (IMD). The IMD is an overall relative measure of deprivation constructed by combining seven domains of social and economic deprivation (i.e.,’ Income Deprivation’, ‘Employment Deprivation’, ‘Education, Skills and Training Deprivation’, ‘Health Deprivation’, ‘Crime’, ‘Barriers to Housing and Services' and ‘Living Environment Deprivation’). The IMD was linked to population exposure in each LSOA based on the usual resident population distribution.

We then examined the statistical relationship between the IMD and the average change of concentration and exposure to PM2.5 and NO2 at LSOA-level using Spearman's correlation. Our goal was to show the strength of association between the time-averaged air pollution reductions and SES. Spearman's correlation was chosen for the statistical analysis because it is considered as a suitable technique to correlate ordinal variables, such as the ranked IMD data, and has been previously used to correlate UK IMD data with different environmental exposures (Tonne et al., 2018).

3. Results

3.1. Spatial and temporal change in air pollution concentration and exposure

3.1.1. Spatial distribution of concentrations and exposure reduction

Hourly average outdoor concentrations changed significantly following the COVID-19 lockdown. Before the lockdown, the three-year London average (2017–2019) NO2 and PM2.5 concentrations from 23 March to 23 April were 45.1 μg/m3 and 18.2 μg/m3, respectively. After implementation of lockdown measures, the average outdoor concentrations of NO2 and PM2.5 during the same period were 26.7 μg/m3 and 15.7 μg/m3 (Table 1 ), respectively, representing a decrease of 40.9% ± 6% for NO2 and 13.9% ± 4% for PM2.5.

Table 1.

Total exposure and concentrations before (2017–19) and during the lockdown period (2020).

MEs 2017–19
 2020
NO2 (μg/m3) PM2.5 (μg/m3) NO2 (μg/m3) PM2.5 (μg/m3)
Outdoor 45.1 18.2 26.7 15.7
Transportation 37.4 15.1 22.1 13.1
Work 28.9 10.9 17.1 9.4
Home 11.9 9.9 7 8.6
Total Exposure concentration 16.2 10.3 7.7 8.7
Open in a new tab

As changes in outdoor concentrations of NO2 and PM2.5 due to COVID-19 shutdown have been presented and analyzed by several studies, we focus here on changes in population-weighted exposure across different environments. We estimate that transportation was the most highly polluted ME during the lockdown with an average exposure of 22.1 μg/m3 for NO2 and 13.1 μg/m3 for PM2.5, while the average workplace concentration was 17.1 μg/m3 for NO2 and 9.4 μg/m3 for PM2.5. The home ME had the lowest NO2 and PM2.5 concentrations with 7 μg/m3 and 8.6 μg/m3, respectively.

Population and time-weighted exposure is impacted by population activity patterns, I/O ratios and outdoor concentration. The indoor levels of outdoor air pollution are directly affected by the I/O ratios of dwellings, thus modifying exposures to outdoor air pollution. Here, we found that the average population-weighted mean exposure decreased following lockdown from 16.2 μg/m3 to 7.7 μg/m3 (a 52.3% reduction) for NO2, and from 10.3 μg/m3 to 8.7 μg/m3 (a 15.7% reduction) for PM2.5. The fact that a much higher percentage of people were spending their daytime inside their homes (an increase from 50% to 90%), has led to a greater reduction in exposure during the lockdown due to the protective role of housing on outdoor air pollution exposures (Smith et al., 2016).

Fig. 2, Fig. 3a show the concentration and exposure change across London. For NO2, the greatest exposure reductions (55–71%) were observed in Inner London (Fig. 2a). PM2.5 showed the greatest reductions (28%–32%) in East and West areas of Inner London (Fig. 3a). Relatively few areas in West London showed only minor reductions in exposure (<2%). The spatial variation of exposure reduction is also in-part due to changes in the distribution of the population across London and the I/O ratios of the dwellings where they spend their time. The large decrease in exposure in Central London was due to various factors, particularly the more uniform distribution of the population during the lockdown, when the population was not concentrated in central London during working hours (Fig. S2). Additionally, the lower average I/O ratios of dwellings (Fig. S1) and the greater reduction in outdoor concentrations (Fig. 2, Fig. 3) also contributed to reduced exposure. In contrast, some areas in western London, which showed higher I/O ratios (particularly PM2.5) and low reductions in outdoor pollution show comparatively low decreases in overall exposure levels.

Fig. 2.

Fig. 2

Open in a new tab

Maps a) and b) illustrate the spatial distribution of average NO2 concentration and exposure reduction (%) during the lockdown period across London. Maps c) and d) illustrate the spatial distribution of average NO2 concentration and exposure reduction (%) at 08:00am.

Fig. 3.

Fig. 3

Open in a new tab

Maps a) and b) illustrate the spatial distribution of average PM2.5 concentration and exposure reduction (%) during the lockdown period across London. Maps c) and d) illustrate the spatial distribution of average PM2.5 concentration and exposure reduction (%) at 08:00am.

3.1.2. Temporal change in air pollution concentration and exposure

Fig. 4 describes the average hourly reduction in concentration and population exposure to NO2 and PM2.5 during the lockdown. As expected, there is little difference between the concentration (Fig. 4a) and exposure (Fig. 4b) reduction during most hours of the day for both pollutants. However, during morning during rush hours the percent reduction fluctuates differently for both pollutants, which reveals the strong impact of the change in population activity on exposure. Both pollutants show the greatest exposure decrease during morning and evening peak rush hours. During those two time periods, the lowest percentage of people are inside the home relative to other hours of the day (Fig. 1) pre-COVID-19, and thus we expect to observe the most significant changes after lockdown measures at these times. In particular, there was the greatest reduction in population exposure for NO2 (66.1% ± 5.1%) and PM2.5 (19.2% ± 3.9%) at 08:00am.

Fig. 4.

Fig. 4

Open in a new tab

Average diurnal a) concentration reduction (%) and b) exposure reduction (%) during the lockdown period (yellow represents NO2 and blue PM2.5). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

The spatial distribution of the concentration and exposure reduction at the time of the greatest hourly decrease (i.e., 08:00am) is illustrated on Fig. 2, Fig. 3b. NO2 exposures show the highest percent reduction (>65%) in Inner and Northwest London, while PM2.5 exposure is reduced more in the Northeast, South, and parts of Inner London. Because NO2 is strongly related to traffic, the most traffic congested areas of London, such as central London, show the highest exposure change. PM2.5 shows a slightly different and more uniform distribution of exposure reduction, due to factors discussed in section 3.1.1.

3.2. Socioeconomic status

Air pollution concentration and exposure data are summarized to illustrate the differences between IMD classifications. Fig. 5, Fig. 6 present PM2.5 and NO2 concentrations and exposure differences between the two examined periods across each deprivation decile. For PM2.5, the concentration and exposure differences in the most deprived LSOAs (deciles 1, 2, and 3) demonstrate the lowest variability, while the LSOAs with moderate deprivation (i.e., deciles 4,5,6) show the largest variability. For NO2, LSOAs in IMD decile 2 show the highest average and the greatest variability for both concentration and exposure difference, while the least deprived LSOAs (i.e., decile 10) show the lowest variability and slightly lower average difference (8.6 μg/m3) compared to the most deprived (8.9 μg/m3). The magnitude of the variability in each IMD decile is likely influenced by the corresponding spatial variation of I/O ratios and outdoor concentrations among the LSOAs of each decile. The smaller variability across the deprivation deciles observed for PM2.5 reductions relative to NO2 may be explained by the less variable particle concentrations across London (Williams, 2020). Moreover, the reductions in concentration also indicate that highly deprived populations in London are disproportionately impacted by air pollution from traffic sources. For both pollutants, the results demonstrate a negative relationship between deprivation deciles and the average exposure and concentration difference during the study period (Table 2 ). Therefore, disadvantaged areas were associated with higher reduction of concentration and exposure to PM2.5 and NO2. Only a very weak association was found for NO2 with correlations of −0.11 and −0.05 for concentration and exposure, whereas the PM2.5 concentration and exposure difference were more strongly correlated with IMD. All correlations are statistically significant (p-value <0.05). This study provides evidence of weak associations, but in the direction of the predictions of several previous studies that suggest a great concentrations or exposure in the most deprived areas (Brook and King, 2017, Padilla et al., 2014, Tonne et al., 2018)

Fig. 5.

Fig. 5

Open in a new tab

a) Variation of PM2.5 concentration difference and the total population of all LSOAs in each decile, b) Variation of PM2.5 exposure difference and the total population of all LSOAs in each decile.

Fig. 6.

Fig. 6

Open in a new tab

a) Variation of NO2 concentration difference and the number of people in each decile; b) Variation of NO2 exposure and the number of people in each decile.

Table 2.

Spearman's correlation coefficient between deprivation index (IMD) and air pollution concentration (exposure) difference.

Concentration Exposure
IMD NO2 PM2.5 NO2 PM2.5
−0.11* −0.25* −0.05** −0.26*
Open in a new tab

*p-value <0.001, **p-value<0.05.

4. Discussion

Lockdown measures in different parts of the world due to the COVID-19 outbreak have provided an opportunity to evaluate the human impact on the urban environment. In this work, we evaluate the relationship between population exposure and time-activity patterns, including the time spent indoors. We found a high average percent reduction in NO2 exposure (52.3% ± 6.1%) with the greatest decrease in Inner London, while PM2.5 exposure showed a considerably lower average percent reduction (15.7% ± 4.1%). The very high reductions in exposure to both pollutants during the morning rush hours show the strong influence of changes in population commuting. By linking population SES and exposure change, we demonstrate variation in air pollution exposure reduction following lockdown across IMD deciles, and provide evidence supporting the conclusion that deprived communities in London are disproportionately affected by road transport pollution.

Numerous prior research studies have investigated and evaluated the influence of coronavirus on air quality globally, and several approaches can be broadly identified. According to recent literature, reductions in NO2 and PM2.5 concentrations during the lockdown ranged from 10% to greater than 50% worldwide (Fonseca, 2020, Tanzer-Gruener et al., 2020, Williams, 2020, Wu et al., 2021, Zhao et al., 2020) with the highest emission reductions observed during morning rush hours. Here, we estimate an average reduction of approximately 50% and 16% for NO2 and PM2.5, respectively. The radical changes in population activity and the significant change in the spatial distribution of the population are likely to have significantly contributed to this reduction in emissions. As with other studies, we estimated the greatest exposure reductions during morning rush hours and during the evening peak hours, particularly at 08:00am when there was the greatest reduction in population exposure for NO2 (66.1% ± 5.1%) and PM2.5 (19.2% ± 3.9%). The steep decrease in air pollution exposure levels during rush hours reflects the importance of the temporal variation of population activity and spatio-temporal variation of the domestic I/O ratios. Conversely, during night hours and early morning hours, the reduction in exposure was much lower. As the number of night workers is much lower than the number of day or evening workers and over 90% of the population was at home during night or early morning, only minor changes were observed to the population activity patterns at these times.

Many large cities around the world demonstrated lower outdoor concentrations of air pollution during the quarantine measures, improving air quality (Arregocés et al., 2021; Kumar et al., 2020). However, it is worth noting that some studies show higher PM2.5 concentrations in several locations (Daniella Rodriguez-Urrego and Leonardo Rodriguez-Urrego, 2020) relative to the pre-covid period, and the effect of the lockdown on some pollutants might be still questionable. A direct comparison between studies is frustrated by the different periods and sites considered, and the methodologies used to quantify the changes. In the UK, a selection of studies have investigated the impact of the shutdown on the concentration of urban pollutants (Williams, 2020; Fonseca et al., 2020). However, there is little research on how changes in population exposure are distributed across urban areas, accounting for the spatial and temporal variability of the exposures in different MEs. Our novel approach includes hourly average I/O ratios of more than 1.5 million dwellings - averaged by LSOA - and estimates an average population exposure reduction of 66% and 19% for NO2 and PM2.5. For NO2, the highest reduction was observed in Central, Northwest and Southeast London and for PM2.5 in the West and East of Inner London. For both concentration and exposure, NO2 show notably higher reductions than PM2.5 post lockdown. This is likely due to a significant decrease in traffic-rated emissions in London, meaning pollutants that are strongly related to traffic emissions, such as NO2, are more significantly affected. On the other hand, for outdoor PM2.5, the contribution of local transport emissions is smaller than for NO2 (Reis et al., 2018) and particulate pollution may be influenced by other factors (for example, local meteorology, transboundary transport, resuspension and the use of fireplaces).

Health studies have suggested that lower SES populations are more likely to suffer premature mortality from air pollution exposure than higher SES populations (Krewski et al., 2009). Multiple studies have been conducted in large cities and metropolitan areas around the world associating the SES with the air pollution concentration and exposure. Most of them demonstrate high associations between the most deprived areas and high outdoor (Sarmadi et al., 2020; Cakmak et al., 2016; Pinault et al., 2016; Padilla et al., 2014; Gray et al., 2013) and indoor concentrations (Ferguson et al., 2020). Here, we provide new information about the impact of lockdown measures on people across different IMD groups. Results indicate negative associations between the reductions of concentration and exposure during the lockdown period and the area-level deprivation status, where PM2.5 is more strongly correlated than NO2. Several studies conducted in large urban areas have presented similar outcomes (Padilla et al., 2014). In London, (Brook and King, 2017) predicted that reductions in exposure to NO2 would be higher for areas that fall within IMD decile 1 (most deprived) after the implementation of air pollution reduction measures. Furthermore, Tonne et al. (2018) analyzed the relationship between SES and outdoor air pollution, finding an exposure different of 0–1.9 μg/m3 between the highest and lowest household income groups, and greater reductions in air pollution in the least advantaged areas after the activation of the Congestion Charging Zone in London.

The main strengths of our study are the large dataset, including population information at LSOA-level, travel behavior from a representative sample of the London population and the large spatio-temporal variability of the I/O ratios for dwellings. The indoor environment is protective of exposure to outdoor air pollutants and that is usually reflected in much lower exposures when Home MEs have been taken into account. Amid the pandemic lockdown measures, when more than 90% of the population had to stay at their home during the daytime, the incorporation of the spatial and temporal distribution of domestic I/O ratios when estimating the population-weighted exposure significantly modifies the magnitude and distribution of the exposure change.

This study contains several limitations. The limitations is the quality of the derived air pollution data and the absence of meteorological effects. Because our study is based on recent measurements, most of the available concentrations for 2020 have not yet been fully ratified by the LAQN. However, in order to reduce the uncertainty and improve the quality of our data, we did not include any negative or unusually extreme hourly values to our analysis. A few monitoring sites did not provide 100% of the data for the whole study period and some hourly readings were missing (or not included). No sites provided less than 70% of the data (Lang et al., 2019; King's College London, 2015). Temporal and spatial variability of air pollution concentrations are subject to changes in emissions and meteorology, which may impact the exposure levels (Bujin et al., 2020). NO2 levels can be directly linked to the reduction of transport emissions due to its strong relation to traffic (He et al., 2020a, 2020b). However, transboundary transport of PM and precursors from mainland European sources and the associated meteorology play an important role in PM concentrations in London. Thus, post-COVID-19 concentrations might be different than pre-COVID-19 due to reasons that are not directly related to lockdown. The wind conditions during 2020 have been exceptional in many ways across the UK (Carslaw, 2020). Moreover, the lockdown period also coincides with the period of the year where there is an increased frequency of PM2.5 episodes in Europe (Air Quality Expert Group, 2020). Therefore, the lack of accounting for weather conditions in our assessment is likely to have affected our results and some reductions may have been over-estimated. However, our approach of averaging the same calendar period of the previous years might have the benefit of reducing meteorological variability. Another limitation is that exposure to other urban air pollutants was not considered, mostly due to data in availability. In this study we focused on the two most important major air pollutants for London's air quality (https://www.london.gov.uk/). Many air pollutants have common sources, and air pollution reduction strategies that take advantage of these common sources may achieve economies of scale that control strategies that target one pollutant at a time cannot. Moreover, pollutants can also be connected by similar precursors or chemical reactions once in the atmosphere. Thus, control strategies that target one pollutant may affect others, perhaps in unintended ways. A much denser network of monitoring stations was available for the NO2 compared to PM2.5. As the concentration of air pollution can change across small distances, the denser network can lead to higher prediction accuracy. In this work, roadside and urban background sites were included, with roadside sites mostly located within Inner London. The denser NO2 monitoring network and the smaller distances between the sites were able to provide adequate coverage of background sites for non-traffic locations. However, the interpolation of roadside measurements, especially for the less dense PM2.5 network, may have led to an overestimation of the impacts of reduced traffic by interpolating to non-traffic areas. The surrounding urban environment can significantly influence pollutant transport and concentration, and to account for this, high-skilled urban modelling accounting for complex urban morphology is required. However, this kind of advanced modelling was not feasible for this study, but could be incorporated to future studies. Moreover, schools and commercial buildings were assumed to have same values as home microenvironment and children were included in the home population. Finally, the average workplace I/O ratio used in this study was assumed from several European cities (Soares et al., 2014; Hänninen et al., 2004, 2011). Data on I/O ratios in commercial buildings, and for different type of workplaces are scarce. Therefore, it was assumed that the values demonstrated in Europe were also representative for London.

This work utilized Google Statistics and App Maps to determine differences in travel patterns. Both App Maps and Google statistics are based on data sent from users' devices and users that opt-in to location history for their account, respectively. Consequently, those data sources contain limitations in terms of their representativeness of the overall population. Apple Maps has no demographic information about its users, making it difficult determine data representativeness. In the calculations, Google statistics includes only data from users that use their Google account and have opted-in to Location History. Those data also have to meet Google's privacy threshold. Consequently, this location data may not represent the exact behavior of the wider population. As described in methodology section, the IMD is based on seven main domains. The ‘Living Environment’ domain contributes approximately 9% to the production of the overall index and measures the quality of the local environment and the indicators fall into two sub-domains. The ‘indoors’ and the ‘outdoors’, which consists of two elements: air quality and road accidents. However, the already included ‘air quality’ element is not likely to have affected our calculations, because here we examined the associations between the IMD and the reduction of concentration and exposure. Other studies have already used IMD to investigate SES inequalities in air pollution (Brook and King, 2017, Sheridan et al., 2019, Tonne et al., 2018)

Some segments of the working population – so-called essential or key workers - had to continue to travel to work in their original workplace during the lockdown period. When estimating the population-weighted exposure, we assumed that all SES groups are equally likely to stay at home during lockdown, however many essential workers are likely to be low SES individuals. Their total exposure to air pollution may still decrease due to the reduction in outdoor concentration, however the change in their exposure to air pollution would be different from other working groups because their daily activity during the lockdown would be the same as the pre-COVID-19 period. Due to the unavailability of data, essential workers could not be linked with the IMD analysis to investigate how this may impact exposure differences between IMD groups. By using the workplace population for the work ME, and by applying the mean percent reduction for the population that continued going to workplaces during the shutdown, we assume that the percentage of population in work ME during post-COVID-19 period (28%) represents essential workers. This percentage is consistent with the ONS estimate that essential workers are approximately 29.5% of London's workforce (ONS, 2020). While further work is required to understand uncertainties in travel and work patterns of low-SES essential workers, these results allow us to conclude that the lockdown provided significant exposure reductions to low-income communities in London.

5. Conclusions

The implementation of stay-at-home measures due to the global outbreak of COVID-19 has offered a unique opportunity to assess the effect of the rapid changes in population activity patterns on air pollution concentration and population exposure. This study quantified and analyzed spatial and temporal changes in population-weighted mean exposure to air pollution of outdoor origin between the COVID-19 lockdown period and previous 3-year average during the same calendar period. Subsequently, we evaluated socioeconomic variation across the distribution of exposure change. We demonstrate that changes in diurnal population activity and outdoor concentrations have reduced exposure to air pollution, predominately during the morning rush hours. The average exposure to NO2 showed a greater than 50% reduction, which was consistent with the remarkable decrease in traffic levels, a major source of NO2. For PM2.5, the 16% decrease in average exposure could not be linked directly to the reduction in urban traffic, because other factors, such as meteorological conditions, may have affected the magnitude of the change in the outdoor concentrations. While there were not large inequalities in how the exposure change was distributed among people with different SES, our results provided useful evidence about the strength of association between the concentration and exposure reduction, and the impact on the most and the least deprived areas.

By quantifying exposure reduction, and accounting for the significance of the time spent indoors and the spatio-temporal variability of average dwelling I/O ratios, this study offers insight into the effectiveness of extreme traffic-control measures on reducing the outdoor pollution and the exposure. Although these measures are extreme and highly unlikely to be adopted under normal conditions, this natural experiment offers the opportunity to assess the influence of some key elements (e.g., population activity, important indoor MEs) on population exposure, using largely real-world data. The estimated exposure reductions may provide best-case estimates of the degree to which more realistic control strategies for stationary and mobile urban sources, such technological (e.g., new-source certifications, retrofits of existing vehicles, etc.) or non-technological (e.g., management of transportation, etc.) may reduce exposures. The analysis of the SES inequalities across the distribution of the exposure reduction also demonstrates the importance of developing strategies that can reduce existing exposure inequalities.

Credit author assessment

Vasilis Kazakos: Data curation, Data analysis, Writing- Original draft preparation. Zhiwen Luo: Conceptualization, Methodology, Supervision. Jonathon Taylor: Data acquisition, Methodology. All: Writing- Reviewing and Editing.

Declaration of competing interest

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.

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.envres.2021.111236.

Appendix A. Supplementary data

The following is/are the supplementary data to this article:

Multimedia component 1
mmc1.docx (1.3MB, docx)

References

  1. Air Quality Expert Group . ” DEFRA; 2020. Report: Estimation of Changes in Air Pollution Emissions, Concentrations and Exposure during the COVID-19 Outbreak in the UK.https://uk-air.defra.gov.uk/library/reports.php?report_id=1005 [Google Scholar]
  2. Apple Mobility trend”. https://www.apple.com/covid19/mobility/
  3. Arregocés H.A., Rojano R., Restrepo G. Impact of lockdown on particulate matter concentrations in Colombia during the COVID-19 pandemic. Sci. Total Environ. 2021;764:142874. doi: 10.1016/j.scitotenv.2020.142874. 142874. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brook R., King K. ” Oxford Centre for Innovation; 2017. Updated Analysis of Air Pollution Exposure in London. [Google Scholar]
  5. Bujin B., Joshua S.A., Dylan B M., Allen R., Kelley C.W., Julian D.M. 2020. PM2.5 and Ozone Air Pollution Levels Have Not Dropped Consistently across the US Following Societal Covid Response. [DOI] [Google Scholar]
  6. Cakmak S., Hebbern C., Cakmak J.D., Vanos J. The modifying effect of socioeconomic status on the relationship between traffic, air pollution and respiratory health in elementary schoolchildren. J. Environ. Manag. 2016;177:1–8. doi: 10.1016/j.jenvman.2016.03.051. [DOI] [PubMed] [Google Scholar]
  7. 2020. https://airqualitynews.com/2020/03/31/ricardo-an-analysis-of-covid-19-lockdown-on-uk-local-air-pollution/
  8. Ferguson L., Taylor J., Davies M., Shrubsole C., Symonds P., Dimitroulopoulou S. Exposure to indoor air pollution across socio-economic groups in high-income countries: a scoping review of the literature and a modelling methodology. Environ. Int. 2020;143 doi: 10.1016/j.envint.2020.105748. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Fonseca E. ” Environmental Defence Fund Europe; 2020. New Breathe London Data: Covid-19 Confinement Measures Reduce London Air Pollution. [Google Scholar]
  10. Google statistics Mobility changes. https://www.gstatic.com/covid19/mobility/2020-04-17_GB_Mobility_Report_en.pdf
  11. Gray S.C., Edwards S.E., Miranda M.L. Race, socioeconomic status, and air pollution exposure in North Carolina. Environ. Res. 2013;126:152–158. doi: 10.1016/j.envres.2013.06.005. [DOI] [PubMed] [Google Scholar]
  12. Hajat A., Hsia C., O'Neill M.S. Socioeconomic disparities and air pollution exposure: a global review. Curr. Environ. Health Rep. 2015;2:440–450. doi: 10.1007/s40572-015-0069-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Hänninen O.O., Lebret E., Ilacqua V., Katsouyanni K., Künzli N., Srám R.J., Jantunen M. Infiltration of ambient PM2.5 and levels of indoor generated non-ETS PM2.5 in residences of four European cities. Atmos. Environ. 2004;38:6411–6423. doi: 10.1016/j.atmosenv.2004.07.01. [DOI] [Google Scholar]
  14. Hänninen O., Hoek G., Mallone S., Chellini E., Katsouyanni K., Gariazzo C., Cattani G., Marconi A., Molnár P., Bellander T., Jantunen M. Seasonal patterns of outdoor PM infiltration into indoor environments: review and meta-analysis of available studies from different climatological zones in Europe. Air Qual. Atmos. Health. 2011;4:221–233. https://doi:10.1007/s11869-010-0076-5 [Google Scholar]
  15. He L.Q., Zhang S.J., Hu J.N., Li Z.H., Zheng X., Cao Y.H., Xu G.Y., Yan M., Wu Y. On-road emission measurements of reactive nitrogen compounds from heavy-duty diesel trucks in China. Environ. Pollut. 2020;262 doi: 10.1016/j.envpol.2020.114280. [DOI] [PubMed] [Google Scholar]
  16. He M.Z., Kinney P.L., Li T.T., Chen C., Sun Q.H., Ban J., Wang J.N., Liu S.L., Goldsmith J., Kioumourtzoglou M.A. Short- and intermediate-term exposure to NO2 and mortality: a multi-county analysis in China. Environ. Pollut. 2020;261 doi: 10.1016/j.envpol.2020.114165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hu Y., Zhao B. Relationship between indoor and outdoor NO2: a review. Build. Environ. 2020;180:106909. [Google Scholar]
  18. King’s College London London air quality network. LAQN web site. https://www.londonair.org.uk/LondonAir/Default.aspx
  19. King’s College London 2015. http://content.tfl.gov.uk/roadside-air-quality-trends-in-london-identifying-outliers-part-1.pdf
  20. Kornartit C., Sokhi R.S., Burton M.A., Ravindra K. Activity pattern and personal exposure to nitrogen dioxide in indoor and outdoor microenvironments. Environ. Int. 2010;36:36–45. doi: 10.1016/j.envint.2009.09.004. [DOI] [PubMed] [Google Scholar]
  21. Kumar P., Hama S., Omidvarborna H., Sharma A., Sahani J., Abhijith K.V., Debele S.E., Zavala-Reyes J.C., Barwise Y., Tiwari A. Temporary reduction in fine particulate matter due to ‘anthropogenic emissions switch-off’ during COVID-19 lockdown in Indian cities. Sustain. Cities and Soc. 2020;62:102382. doi: 10.1016/j.scs.2020.102382. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Krewski D., Jerrett M., Burnett R.T., Ma R., Hughes E., Shi Y., Turner M.C., Calle E.E., Thun M.J., Beckerman B., DeLuca P., Finkelstein N., Ito K., Moore D.K., Newbold K.B., Ramsay T., Ross Z., Shin H., Tempalski B. Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air pollution and mortality. Res. Rep. Health Eff. Inst. 2009;140:5–114. [PubMed] [Google Scholar]
  23. Lang P.E., Carslaw D.C., Moller S.J. A trend analysis approach for air quality network data. Atmos. Environ. X. 2019;2:100030. doi: 10.1016/j.aeaoa.2019.100030. [DOI] [Google Scholar]
  24. 2011. https://tfl.gov.uk/
  25. Mahato S., Pal S., Ghosh K.G. Effect of lockdown amid COVID-19 pandemic on air quality of the megacity Delhi, India. Sci. Total Environ. 2020;730 doi: 10.1016/j.scitotenv.2020.139086. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Mbandi A.M. Air pollution in africa in the time of COVID-19: the air we breathe indoors and outdoors. Clean Air J. 2020;30 doi: 10.17159/caj/2020/30/1.8227. [DOI] [Google Scholar]
  27. Muhammad S., Long X., Salman M. COVID-19 pandemic and environmental pollution: a blessing in disguise? Sci. Total Environ. 2020;728:138820. doi: 10.1016/j.scitotenv.2020.138820. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. ONS . ” Statistical Bulletin, Office for National Statistics; UK: 2011. Census - Population and Households Estimates for England and Wales.http://www.ons.uk/ons/dcp171778_270487.pdf 2012. [Google Scholar]
  29. ONS . ” Office for National Statistics; UK: 2020. Coronavirus and Key Workers in the UK.https://www.ons.gov.uk/employmentandlabourmarket/peopleinwork/earningsandworkinghours/articles/coronavirusandkeyworkersintheuk/2020-05-15 [Google Scholar]
  30. Padilla C.M., Kihal-Talantikite W., Vieira V.M., Rossello P., Le Nir G., Zmirou-Navier D., Deguen S. Air quality and social deprivation in four French metropolitan areas-A localized spatio-temporal environmental inequality analysis. Environ. Res. 2014;134:315–324. doi: 10.1016/j.envres.2014.07.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Pinault L., Crouse D., Jerrett M., Brauer M., Tjepkema M. Spatial associations between socioeconomic groups and NO2 air pollution exposure within three large Canadian cities. Environ. Res. 2016;147:373–382. doi: 10.1016/j.envres.2016.02.033. [DOI] [PubMed] [Google Scholar]
  32. Public Health England . 2019. Air Pollution Evidence Review.https://www.gov.uk/government/news/public-health-england-publishes-air-pollution-evidence-review [Google Scholar]
  33. Reis S., Liska T., Vieno M., Carvell E.J., Beck R., Clemens T., Dragosits U., Tomlinson S.J., Leaver D., Heal M.R. The influence of residential and workday population mobility on exposure to air pollution in the UK. Environ. Int. 2018;121:803–813. doi: 10.1016/j.envint.2018.10.005. [DOI] [PubMed] [Google Scholar]
  34. Rivas B., Kumar P., Hagen-Zanker A. Exposure to air pollutants during commuting in London: are there inequalities among different socio-economic groups? Environ. Int. 2017;101:143–157. doi: 10.1016/j.envint2017.01.019. [DOI] [PubMed] [Google Scholar]
  35. Rodríguez-Urrego D., Rodríguez-Urrego L. Air quality during the COVID-19: PM(2.5) analysis in the 50 most polluted capital cities in the world. Environ. Pollut. 2020;266:115042. doi: 10.1016/j.envpol.2020.115042. 115042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Rotko T., Koistinen K., Hanninen O., Jantunen M. Sociodemographic descriptors of personal exposure to fine particles (PM2.5) in EXPOLIS Helsinki. J. Expo. Anal. Environ. Epidemiol. 2000;10:385–393. doi: 10.1038/sj.jea.7500104. [DOI] [PubMed] [Google Scholar]
  37. Rotko T., Kousa A., Alm S., Jantunen M. Exposures to nitrogen dioxide in EXPOLIS-Helsinki: microenvironment, behavioral and sociodemographic factors. J. Expo. Anal. Environ. Epidemiol. 2001;11:216–223. doi: 10.1038/sj.jea.7500162. [DOI] [PubMed] [Google Scholar]
  38. Sarmadi M., Moghanddam V.K., Dickerson A.S., Martelletti L. Air Quality, Atmosphere & Health. 2020. Association of COVID-19 distribution with air quality, sociodemographic factors, and comorbidities: an ecological study of US states. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Sharma S., Zhang M.Y., Anshika Gao J.S., Zhang H.L., Kota S.H. Effect of restricted emissions during COVID-19 on air quality in India. Sci. Total Environ. 2020;728 doi: 10.1016/j.scitotenv.2020.138878. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Sheridan C.E., Roscoe C.J., Gulliver J., de Preux L., Fecht D. Inequalities in exposure to nitrogen dioxide in parks and playgrounds in greater London. Int. J. Environ. Res. Publ. Health. 2019;16:3194. doi: 10.3390/ijerph16173194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Shiels M.S., Chernyavskiy P., Anderson W., Best A.F., Haozous E.A., Hartge P., Rosenberg P.S., Thomas D., Freedman N.D., De Gonzalez B. Trends in premature mortality in the USA by sex, race, and ethnicity from 1999 to 2014: an analysis of death certificate data. Lancet. 2017;389 doi: 10.1016/S0140-6736(17)30187-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Singh V., Sokhi R.S., Kukkonen J. An approach to predict population exposure to ambient air PM2.5 concentrations and its dependence on population activity for the megacity London. Environ. Pollut. 2020;257 doi: 10.1016/j.envpol.2019.113623. [DOI] [PubMed] [Google Scholar]
  43. Smith J.D., Mitsakou C., Kitwiroon N., Barratt B.M., Walton H.A., Taylor J.G., Anderson H.R., Kelly F.J., Beevers S.D. London hybrid exposure model: improving human exposure estimates to NO2 and PM2.5 in an urban setting. Environ. Sci. Technol. 2016;50:11760–11768. doi: 10.1021/acs.est.6b01817. [DOI] [PubMed] [Google Scholar]
  44. Soares J., Kousa A., Kukkonen J., Matilainen L., Kangas L., Kauhaniemi M., Riikonen K., Jalkanen J.P., Rasila T., Hänninen O., Koskentalo T., Aarnio M., Hendriks C., Karppinen A. Refinement of a model for evaluating the population exposure in an urban area. Geosci. Model Dev. (GMD) 2014;7:1855–1872. https://gmd.copernicus.org/articles/7/1855/2014/ [Google Scholar]
  45. Song W.W., Ashmore M.R., Terry A.C. The influence of passenger activities on exposure to particles inside buses. Atmos. Environ. 2009;43:6271–6278. doi: 10.1016/j.atmosenv.2009.05.004. [DOI] [Google Scholar]
  46. Stringhini S., Carmeli C., Jokela M. Socioeconomic status and the 25 x 25 risk factors as determinants of premature mortality: a multicohort study and meta-analysis of 1.7 million men and women (vol 389, pg 1229, 2017) Lancet. 2017;389:1194. doi: 10.1016/S0140-6736(16)32380-7. 1194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Su J.G., Jerrett M., Morello-Frosch R., Jesdale B.M., Kyle A.D. Inequalities in cumulative environmental burdens among three urbanized counties in California. Environ. Int. 2012;40:79–87. doi: 10.1016/j.envint.2011.11.003. [DOI] [PubMed] [Google Scholar]
  48. Tanzer-Gruener R., Li J.Y., Eilenberg S.R., Robinson A.L., Presto A.A. Impacts of modifiable factors on ambient air pollution: a case study of COVID-19 shutdowns. Environ. Sci. Technol. Lett. 2020;7:554–559. doi: 10.1021/acs.estlett.0c00365. [DOI] [PubMed] [Google Scholar]
  49. Taylor J., Shrubsole C., Davies M., Biddulph P., Das P., Hamilton I., Vardoulakis S., Mavrogianni A., Jones B., Oikonomou E. The modifying effect of the building envelope on population exposure to PM2.5 from outdoor sources. Indoor Air. 2014;24:639–651. doi: 10.1111/ina.12116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Tonne C., Mila C., Fecht D., Alvarez M., Gulliver J., Smith J., Beevers S., Anderson H.R., Kelly F. Socioeconomic and ethnic inequalities in exposure to air and noise pollution in London. Environ. Int. 2018;115:170–179. doi: 10.1016/j.envint.2018.03.023. [DOI] [PubMed] [Google Scholar]
  51. Williams M. King’s College London; 2020. On Behalf of the Environmental Research Group, “The Effect of Covid-19 Lockdown Measures on Air Quality in London in 2020. [Google Scholar]
  52. Wu C.-L., Wang H.-W., Cai W.-J., He H.-D., Ni A.-N., Peng Z.-R. Building and Environment; 2021. Impact of the COVID-19 Lockdown on Roadside Traffic-Related Air Pollution in Shanghai, China.https://www.sciencedirect.com/science/article/pii/S0360132321001293 107718. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Zhao Y.B., Zhang K., Xu X.T., Shen H.Z., Zhu X., Zhang Y.X., Hu Y.T., Shen G.F. Substantial changes in nitrogen dioxide and ozone after excluding meteorological impacts during the COVID-19 outbreak in mainland China. Environ. Sci. Technol. Lett. 2020;7:402–408. doi: 10.1021/acs.estlett.0c00304. [DOI] [PubMed] [Google Scholar]
  54. Zhu Y.J., Xie J.G., Huang F.M., Cao L.Q. Association between short-term exposure to air pollution and COVID-19 infection: evidence from China. Sci. Total Environ. 2020;727 doi: 10.1016/j.scitotenv.2020.138704. http://www.sciencedirect.com/science/article/pii/S004896972032221X [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Multimedia component 1
mmc1.docx (1.3MB, docx)

Articles from Environmental Research are provided here courtesy of Elsevier

RESOURCES