key: cord-0995511-5exf8tgc
authors: Jephcote, Calvin; Hansell, Anna L.; Adams, Kathryn; Gulliver, John
title: Changes in air quality during COVID-19 ‘lockdown’ in the United Kingdom()
date: 2020-11-20
journal: Environ Pollut
DOI: 10.1016/j.envpol.2020.116011
sha: f4efdbb937ccb119949ece7be121d09ae2c654ce
doc_id: 995511
cord_uid: 5exf8tgc

The UK implemented a lockdown in Spring 2020 to curtail the person-to-person transmission of the SARS-CoV-2 virus. Measures restricted movements to one outing per day for exercise and shopping and otherwise most people were restricted to their dwelling except for key workers (e.g. medical, supermarkets, and transport). In this study, we quantified changes to air quality across the United Kingdom from 30(th) March to 3(rd) May 2020 (weeks 14-18), the period of most stringent travel restrictions. Daily pollutant measurements of NO(2), O(3) and PM(2.5) from the national network of monitoring sites during this period were compared with measurements over the same period during 2017-19. Comparisons were also made with predicted concentrations for the 2020 period from business-as-usual (BAU) modelling, where the contributions of normal anthropogenic activities were estimated under the observed meteorological conditions. During the lockdown study period there was a 69% reduction in traffic overall (74% reduction in light and 35% in heavy vehicles). Measurements from 129 monitoring stations, identified mean reductions in NO(2) of 38.3% (-8.8 μg/m(3)) and PM(2.5) of 16.5% (-2.2 μg/m(3)). Improvements in NO(2) and PM(2.5) were largest at urban traffic sites and more modest at background locations where a large proportion of the population live. In contrast, O(3) concentrations on average increased by 7.6% (+4.8 μg/m(3)) with the largest increases at roadside sites due to reductions in local emissions of NO. A lack of VOC monitoring limited our capacity to interpret changes in O(3) at urban background locations. BAU models predicted comparable NO(2) reductions and O(3) gains, although PM(2.5) episodes would have been more prominent without lockdown. Results demonstrate the relatively modest contribution of traffic to air quality, suggesting that sustained improvements in air quality require actions across various sectors, including working with international and European initiatives on long-range transport air pollutants, especially PM(2.5) and O(3).

The first human cases of a severe acute respiratory syndrome coronavirus (SARS-CoV-2), 27 referred to as COVID-19, were detected in Wuhan City, China, in December 2019. The virus 28 spread rapidly in early 2020, initially from China to Italy and Europe where lockdowns were 29 introduced restricting travel and industrial activity to control spread. These resulted in 30 marked improvements in air quality with dramatic extent of this apparent in international 31 satellite imagery (Menut et al., 2020 , NASA 2020 Shi and Brasseur, 2020) . The UK confirmed its earliest cases on the 31st January 2020, with rapid increase in cases to 34 >5,000 by 20th March 2020 resulting in a national lockdown starting Tuesday 23rd March 35 (UK Government 2020). Restrictions introduced included; 1) people to stay at home, except 36 for limited purposes; 2) closing of all non-essential businesses (including retail and 37 hospitality sectors) and public venues (including community centres, libraries, and places of 38 worship); 3) stopping all gatherings of more than two people in public (Johnson 2020). The 39 media widely reported falls in air pollution of 40%-60% in early April (Khoo, 2020; Telegraph 40 Reporters, 2020). At the same time the UK Department for Environment Food & Rural 41 Affairs (Defra) was warning of air pollution episodes with the Daily Air Quality Index 42 reaching the highest level on its 10-point measurement scale. Hourly measurements of gaseous and particulate pollutants recorded by AURN sites, were 66 accessed in R-Statistical Software via the openair v2.7 application programming interface 67 (API). Measurements of NO 2 , NO X , O 3 , and PM 2.5 were extracted from all monitoring sites 68 with a 95% or above hourly capture rate across weeks 14 to 18 of 2020. This covered a 5- 69 week period of the most stringent restrictions on population movements during an 70 outbreak of COVID-19 in the UK, where employers in non-essential sectors facilitated transition in travel behaviour. The study end date was 5-days before a public holiday, which 74 was anecdotally linked with widespread travel breaches. Daily average concentrations were 75 only calculated for days with a minimum hourly capture rate of 75%. 76 77 For comparison, air quality measurements at the selected sites were obtained across weeks 78 14 to 18 of 2017, 2018, and 2019. AURN sites were only included in the final analysis if they 79 also reported a 95% or above hourly capture rate over this 3-year period. As before, daily 80 average concentrations were calculated for days with a minimum hourly capture rate of 81 75%. Table S1 provides a breakdown of the AURN sites included in this analysis, by type and 82 the availability of pollutant measurements. 107 Hourly measurements of wind-speed (m/s) and wind-direction (°) were directly recorded at 108 the AURN measurement sites. Hourly measurements of other meteorological parameters 109 recorded at UK Met Office stations, were accessed in R-Statistical Software via the worldmet 110 v0.9 API. This measurement data is provided by the National Oceanic and Atmospheric 111 Administration's (NOAA) global integrated surface database of hourly climatological 112 observations, subject to ratification and quality assurance processes (Lott 2004 We obtained hourly air temperature (°C), atmospheric pressure (mb), cloud-ceiling height 115 (m), dew point (°C), and relative humidity (%) for each AURN site using measurements from 116 the nearest Met Office station with a 90% or above data capture on each parameter (66% 117 threshold used for cloud cover). Missing values were imputed using an enhanced polynomial 118 interpolation method, which minimises overshooting and spurious oscillations in time-series 119 data (Stineman, 1980) . On average 1.2%, 1.6%, 1.3%, 1.3% and 17.6% of the hourly 120 measurements for air temperature, atmospheric pressure, dew point, relative humidity, and In terms of meteorological parameters, much of April 2020 was fine and settled, with a high-165 pressure system seeing light easterly winds and benign conditions. The provisional UK mean 166 temperature of 9.1 °C was 1.7 °C above the 1981-2010 long-term average, making it the fifth which fell on the last four days of the month (Met Office 2020). Weekly-average 170 meteorological profiles at urban background sites, recorded consistently low wind-speeds (< 171 4 m/s) in southern and central England, rising to 5-6 m/s in Scotland and Wales (see Table   172 S6). Easterly winds ( Table   198 S8). Table S9 ). However, days coinciding with the three 234 NO 2 episodes, equalled or exceeded the daily-average BTEX concentrations of previous 235 years. VOCs are known to promote photochemical reaction mechanisms, and in part 236 perhaps explain why several days of elevated NO 2 concentrations were observed. concentrations were recorded at suburban sites (see Table 1 Table S6 ). O 3 concentrations are highly influenced by 260 meteorological events, which appear to explain up to 69% of the variation in hourly 261 concentrations. The models report relative humidity as the main meteorological driver of 262 O 3 , which is a function of temperature and pressure. 263 concentrations to their modelled estimates, under a business-as-usual (BAU) scenario. Daily 305 PM 2.5 concentrations were consistently lower than their BAU estimates, but this reduction 306 was 2.6 times greater than the comparisons using reference data. This highlights the 307 importance of accounting for local and regional meteorological conditions, alongside any 308 direct comparison of measurements from previous years. Based on BAU modelling, day-to-day comparisons across all site types reported reductions 382 in NO 2 of -36.8% (95% CI: -45.5 to -19.8) and gains in O 3 of +9.3% (95% CI: +0.1 to +21.6), 383 which are comparable to the differences in the measurement data (see Tables 1 and 2 BVOC emission and NO X scavenging potential in the urban atmosphere (Owen et al., 2003) . 407 408 409 A major strength of our study is that it provides a national comparison of changes in air 410 quality during lockdown with respect to: (a) measurements from the three previous years, 411 and (b) modelled concentrations from a BAU scenario given the weather patterns during 412 lockdown. Limitations apply to BAU, a model with its own uncertainty. As a minimum, 413 interpreted BAU models were able to account for 75% of the variation in the estimates of 414 hourly pollutant concentrations under a series of specified temporal and meteorological 415 parameters. The uncertainty in these hourly estimates is shrunk when converting to a daily 416 mean, using a bootstrapped approach. The 95% confidence intervals for NO 2 , NO x , O 3 and 417 PM 2.5 at each site, on average differ from the mean-daily estimate by 15.7%, 24.3%, 9.7% 418 and 8.4%. Any comparisons between the daily measurement data and BAU estimates, take 419 these 95% confidence intervals into account. We suggest that our results are interpreted as A further strength of this study is that it used monitoring station measurements with 428 information on site types. Some other studies have used satellite data to document changes in air quality on a regional/global scale (Muhammad et al., 2020; Venter et al., 430 2020) but due to their relatively coarse spatial resolution they are unable to distinguish 431 changes by site type (e.g. roadside versus background) which we were able to for the UK. 484 In an effort to comply with air quality limit values, and to protect human health, a number The London CCZ yielded a 15% decrease in vehicle km two years after its implementation, 496 resulting in a 12% reduction in NOx emissions; however emissions were observed to 497 increase on the inner ring road by 1.5%, as some traffic is redistributed across the network 498 (Beevers & Carslaw, 2005) . It was estimated that street-level NO 2 concentrations in the 499 Danish cities of Aalborg, Aarhus, Copenhagen, and Odense reduced by 4-11%, through the 500 enforcement of enforcing Euro-III standards on heavy duty vehicles (Jensen et al., 2011) . 501 J o u r n a l P r e -p r o o f LEZ's targeting < EURO-IV trucks across the Netherlands, on average reported a 9.8% 502 reduction in traffic intensities resulting in a 6.2% reduction in NO 2 concentrations, two-years 503 after their implementation (Boogaard et al., 2012) . NO 2 and PM 2.5 concentrations were 504 respectively found to reduce in the Stockholm LEZ by 4% and 26% over the period 1995-505 2001 (Rapaport, 2002) . LEZs were implemented in the German cities of Berlin, Cologne and 506 Hanover in 2008, in response to revised EU directives on ambient air quality and fine 507 particles. While the LEZs has had no measurable impact on traffic flows, the turnover of the 508 national fleet towards cleaner vehicles has speeded up considerably. In terms of local 509 environmental benefits, areas within the Berlin 'Passive' LEZ have been estimated to 510 experience a 4.5% reduction in PM 10 (Lutz and Rauterberg-Wulff, 2010 The improvements for air quality seen in LEZs are relatively modest compared to those 513 reported from our study, implying that much larger interventions are needed that are 514 scalable nationally and internationally to achieve the magnitude of overall improvement in 515 air quality seen during lockdown. The challenge for reducing air pollution, especially for 516 PM 2.5 and O 3 , is international and major interventions are likely to result in modest 517 improvement in air quality. 518 519 Major improvements to air quality may come from switching to a 'greener' vehicle fleet. 520 Road traffic emission reductions during lockdown may indicate what could be achieved for 521 air quality by having a high proportion (e.g. >50%) of the vehicle fleet switching to electric. 522 Benefits overall to air quality are however likely to be lower than seen during lockdown as 523 BAU traffic volumes would result in normal levels of congestion resulting in higher tailpipe 524 emissions for non-electric vehicles associated with lower traffic speeds and poorer 525 ventilation around roads (AQEG 2019; Grigoratos and Martini 2014) . However, the switch to 526 an electric fleet is not without its own problems. Although for electric vehicles there are no 527 tail-pipe emissions, substantial particulate contributions still arise from the tyres, brakes and 528 general wear and tear of any vehicle (Soret et al., 2014; Timmers and Achten, 2016) . The 529 percentage of fleet electrification needed to acquire a significant level of improvement is 530 between 26-40%, highlighting the difficulties in developing sustainable interventions for 531 improving air quality (Soret et al., 2014) .There are also issues relating to how this energy is 532 sourced, especially when fossil fuels still account for 51% of electricity produced in the UK 533 (BEIS 2017), for example. 

Ozone in the United Kingdom. Department for Environment Food & Rural Affairs

Non-Exhaust Emissions from Road Traffic. Department for Environment Food & Rural Affairs

Estimation of changes in air pollution emissions, concentrations, and exposure during the COVID-19 outbreak in the UK. Department for Environment Food & Rural Affairs

COVID-19 lockdown effects on air quality by NO2 in the cities of Barcelona and Madrid (Spain)

The impact of congestion charging on vehicle emissions in London

Impact of low emission zones and local traffic policies on ambient air pollution concentrations

Cabinet Office Briefing Room-A (COBRA 2020). Slides, datasets and transcripts from press conferences at 10 Downing Street in response to coronavirus

Analysis of air pollution data at a mixed source location using boosted regression trees

A short-term intervention study-Impact of airport closure due to the eruption of Eyjafjallajökull on near-field air quality

Effect of VOC emissions from vegetation on air quality in Berlin during a heatwave

The impact of COVID-19 partial lockdown on the air quality of the city of Rio de Janeiro

Changes in outdoor air pollution due to COVID-19 lockdowns differ by pollutant: Evidence from Scotland

/50/EC of the European Parliament and of the Council of 21 May 2008 on ambient air quality and cleaner air for Europe

Industrial production down in March and

Meteorological drivers and mortality associated with O3 and PM2. 5 air pollution episodes in the UK in 2006

The Interplay between Ozone and Urban Vegetation-BVOC Emissions, Ozone Deposition, and Tree Ecophysiology

Impact of weather types on UK ambient particulate matter concentrations

Non-exhaust traffic related emissions -Brake and tyre wear PM. JRC science and policy reports

Use of the Boosted Regression Tree Optimization Technique to Analyse Air Pollution data

Prime Minister's Office, 10 Downing Street

What are the Impacts on Air Quality of Low Emission Zones in Denmark

Coronavirus lockdown sees air pollution plummet across UK

The quality control of the Integrated Surface Hourly Database, paper presented at the 14th Conference on Applied Climatology

Berlin's low emission zone-top or flop

Impact of lockdown measures to combat Covid-19 on air quality over western Europe

COVID-19 pandemic and environmental pollution: a blessing in disguise?

Airborne Nitrogen Dioxide Plummets Over China

ONS, 2020. Coronavirus and the impact on output in the UK economy

Biogenic Volatile Organic Compound (VOC) Emission Estimates from an Urban Tree Canopy

The impact of synoptic weather on UK surface ozone and implications for premature mortality

The Stockholm environmental zone, a method to curb air pollution from bus and truck traffic

Effect of restricted emissions during COVID-19 on air quality in India

The Response in Air Quality to the Reduction of Chinese Economic Activities during the COVID-19 Outbreak

The potential impacts of electric vehicles on air quality in the urban areas of Barcelona and Madrid (Spain)

The predicted number of air pollution related deaths in the UK during the August 2003 heatwave

A consistently well-behaved method of interpolation. Creative Computing. Telegraph Reporters., 2020. The Telegraph

Non-exhaust PM emissions from electric vehicles

Coronavirus (COVID-19) in the UK

COVID-19 lockdowns cause global air pollution declines with implications for public health risk

Retrieved from Friends of the Earth Europe

Time-series plots of measured daily-average PM 2.5 concentrations for weeks 14 to 18 of 2020, in relation to the range of measurements for corresponding

This study is part-funded by the National Institute for Health Research (NIHR) Health

Public Health England, the Health and Safety Executive and the University of Leicester. The views expressed are those of the author(s) and not necessarily those of the NIHR, Public Health England, the Health and Safety Executive or the Department of Health and SocialCare.J o u r n a l P r e -p r o o f

• NO 2 and PM 2.5 concentrations fell respectively by 38.3% (8.8 µg/m 3 ) and 16.5% (2.2 µg/m 3 ).• O 3 increased by 7.6% overall as the atmospheric chemistry changed.• NO 2 and PM 2.5 improvements were largest at urban traffic sites.• Weather conditions contributed to particulate episodes seen despite traffic reductions.

☒ 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.☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: J o u r n a l P r e -p r o o f