Atmospheric Pollution And Human Health In A Chinese Megacity (APHH .

APHH-Beijing Technical Report to improve understanding of the processes by whichpollutants are transformed or removed throughtransport, chemical reactions and photolysis, andthe rates of formation and conversion of particulatematter (PM) via atmospheric reactions;1. BACKGROUND AND OBJECTIVESAPHH-Beijing is the first major UK-China project jointlyfunded by UKRI (NERC, MRC) and NSFC, as part of thethe UK-China Research and Innovation PartnershipFund. The APHH-Beijing programme includes fiveseparate but related projects: to improve understanding on how the detailedproperties of PM evolve and can influence theirphysical properties and behaviour in the atmosphereand elucidate the mechanisms whereby thoseproperties may interact and give feedback on urbanscale and regional meteorology; Sources and Emissions of Air Pollutants in Beijing(AIRPOLL): lead PIs - Roy Harrison, Kebin He. to exploit new satellite observations and regionalmodels to place the in-situ campaigns into awider context; An Integrated Study of Air Pollution Processes inBeijing (AIRPRO): lead PIs - Ally Lewis, Pingqing Fu. Effects of Air Pollution on Cardiopulmonary Diseasein Urban And Peri-Urban Residents in Beijing(AIRLESS): lead PIs - Frank Kelly, Tong Zhu. to determine the personal exposure of Beijinginhabitants to key health-related pollutants andassess the association between air pollution exposureand key cardiopulmonary measures; Air Pollution Impacts on Cardiopulmonary Diseasein Beijing: An integrated study of Exposure Science,Toxicogenomics and Environmental Epidemiology(APIC-ESTEE): lead PIs - Miranda Loh, Zhiwei Sun. to determine the contribution of specific activities,environments and pollution sources to the personalexposure of the Beijing population to air pollutants; Integrated Assessment of the Emission-HealthSocioeconomics Nexus and Air Pollution MitigationSolutions and interventions in Beijing (INHANCE):lead PIs - Dabo Guan, Shu Tao. to enhance our understanding of the health effects insusceptible individuals over time periods when thereare large fluctuations in pollutants compared withnormal controls and to identify health outcomes ofair pollution; andOverall Science Coordinator: Zongbo Shi. to estimate physical and mental impacts of airpollution and examine how Beijing can improve itsair quality more cost-effectively.AIRPOLL is by far the most comprehensiveinvestigation of air pollutant sources and emissions inBeijing to date. AIRPRO focused on understanding thebasic processes controlling gas and aerosol pollution,meteorological dynamics, and the links betweenthem, within Beijing’s atmosphere (after emission).AIRLESS and APIC-ESTEE focused on cardiovascularand respiratory disease and sought to addressuncertainties between personal exposure, toxicologyof air pollutants and acute cardiopulmonary responsein Beijing. INHANCE was designed as an enablerproject to deliver integrated and science-based policydesign, drawing upon all investigations. The objectivesof APHH-Beijing were1:In the past 4 years, APHH-Beijing has successfullydelivered the key objectives of the programme andadditional multidisciplinary science through integrationof individual projects. The integration has led to highlysuccessful field campaigns1 as well as the publication ofover 400 scientific papers including in multidisciplinaryjournals and 47 papers in the APHH-Beijing ACP/AMTspecial Issue, with further publications pending. Thisreport aims to capture some of the research highlightsand successful practices, lessons learned and remainingscience questions to support the design of futureresearch programmes. to determine the emission fluxes of key air pollutantsand to measure the contributions of different sources,economic sectors and regional transport to airpollution in Beijing;1

2. SCIENCE HIGHLIGHTS AND THEIRPOLICY IMPLICATIONS Total black carbon emission from the downscaledMEIC overestimate the actual flux by a factor of59 and 47 during the winter and summer periods,respectively 4. Emissions of black carbon in the citycentre were dominated by vehicle emissions.APHH-Beijing has generated a series of new andpublished scientific results. Here, we highlighted a fewnovel findings, focusing on those that have potentialpolicy implications.2.1 High concentrations of black carbon are observed duringwinter haze events, but flux measurements indicate thatthese originate from outside of central Beijing 4.First Application of Eddy Covariance to AirPollutant Emission Flux Measurements inChina (UKCEH, York, Lancaster, Manchester,GIG, Tsinghua, IAP) Surprisingly, in summer, the city centre's surfacelayer is a sink rather than a source of PM1. This isdue to the higher temperatures in the surface layercompared with the air above, leading to evaporationof (semi-)volatile aerosol components such asammonium nitrate. This a process not resolved bymost atmospheric chemistry and transport models 3.EmissionAPHH-Beijing made the first direct observations ofemission fluxes in Beijing of a range of air pollutantsincluding NOx, NH3, CO, CO2, O3 and individual volatileorganic compounds (VOCs), particulate matter (PM)and its chemical components (black carbon, NH4 , NO3-,SO42-, Cl- and organic matter). This was achieved bythe application of the eddy covariance technique witha range of fast-response analsyers at the Institute ofAtmospheric Physics (IAP) meteorological tower incentral Beijing in winter 2016 and summer 2017. Thismethod provides a direct test of emission inventoriesand new scientific process understandings. Keyoutcomes are:NOx Emission / mg m 2 h 1Toluene Benzene / nmol m 2 s 170706060505040403030202010100 The NOx emission fluxes are similar in summerand winter (3.55 vs. 4.41 mg m 2 h 1), peakingduring the morning and evening rush hour periods,indicating a major traffic source. The measured fluxis substantially lower than that downscaled from theMEIC (Multi-resolution Emission Inventory) based onthe commonly used proxy method (Fig. 1) 2.0612Inventory18000Hour of day612180MeasuredFigure 1. Measured NOx and toluene benzene versusdownscaled MEIC emissions based on a proxy method incentral Beijing2. Diurnal variation from the downscaledMEIC estimates are used for the diurnal cycles foremissions of all sectors (sum of emissions from transport,industry, residential and power). Although the concentrations of aromatic VOCs (suchas toluene and benzene) in central Beijing are veryhigh, their surface-atmosphere flux is very low, andthey must originate largely from outside of centralBeijing. Downscaling from MEIC would result in anorder of magnitude higher emissions in the centre ofBeijing than measured 2.Policy Implications: The observation-based approach developed withinthe APHH-Beijing programme can be appliedelsewhere to validate local emission inventories andprovide direct policy-relevant insight to emissionscontrol choices. Organic matter dominates chemically speciated PM1fluxes at IAP (black carbon, NH4 , NO3-, SO42-, Cl- andorganic aerosol mass). Emission of primary aerosolcomponents in PM1 is dominated by cooking sources,followed by traffic sources and solid fuel burningin winter 3. Emission inventory requires continuous updating andis validated by flux observations. The scope for policy interventions focusing onanthropogenic VOC emission from central Beijing islimited. Emissions controls in regions surroundingthe megacity are necessary to achieve substantialreductions in VOC concentrations within the city.2

at pedestrian level and higher domain-averagedconcentrations over the area. In addition, canyonswith highly even or highly uneven building heights oneach side of the street tend to lower the urban-scale airpollution concentrations at pedestrian level 7. Source apportionment of the measured PM1 fluxidentified cooking (oil) emissions as the largest singlePM1 source, suggesting that a more stringent controlof this local source may be needed within the city. Local emissions overall make a small contributionto concentrations of PM2.5 compared with longrange transport of pollutants which suggests thattargeting further emission reduction measures at thewider region rather than central Beijing itself wouldbe advisable. Receptor modelling shows that road traffic(gasoline diesel) contributed only 12% and 6% ofparticulate organic carbon (OC) in Beijing duringwinter 2016 and summer 2017, respectively 8. An optimized gridded inventory considering resultsfrom the flux observations was developed. Using thisnew inventory and high resolution ADMS modelling,we showed that the contribution of primaryemissions from road traffic to PM2.5 is much smallerthan to NO2 (Fig. 2) 9.2.2. Traffic emissions and their contribution toair pollution in Beijing (Peking, Birmingham,Edinburgh, Lancaster, Reading, IAP, York)APHH-Beijing has applied multiple methods to quantifythe contribution of vehicle emissions to air pollution inBeijing and surrounding regions, including:Policy Implications: Beijing, Tianjin and Hebei should adopt differentiatedpolicies for the control of road vehicle pollution basedon their differing vehicle emissions contributions. High resolution traffic emission inventories showthat in the Beijing-Tianjin-Hebei (BTH) region, lightduty passenger vehicles and heavy-duty trucks arethe largest contributors to road traffic primary PM2.5emissions, whereas heavy duty trucks and busestogether contribute about 50% of road traffic NOxemissions; light-duty passenger vehicles contribute56% of total road traffic hydrocarbons. Emissionscenario analyses showed that eliminating oldvehicles could reduce emissions efficiently but theunit cost for emission reduction increases fromBeijing, to Tianjin and to Hebei 5.Roadside traffic and urban increments70Road IncrementUrban Increment5030200UrbanIncrement4030Rural Background4010 The roadside increment of PM2.5, defined as thedifference between PM2.5 levels at roadside and urbanbackground sites, only accounts for 7.5% of the cityaverage, suggesting a very small contribution ofprimary road traffic to urban PM2.5 mass. But roadtraffic represents a significant source of NO2 (Fig. 2).The urban increment in PM2.5 (difference betweenPM2.5 at urban background and rural sites) is alsovery small, suggesting a limited contribution of urbanemissions to Beijing’s PM2.5 levels 6.RoadIncrement60NO2 (µg m 3)50Rural BackgroundPM2.5 (µg m 3)60702010TrafficUrban BG0RuralTrafficUrban BGRuralSpatial maps simulated using the MEIC Opt emissions inventoryNO2 (µg m 3)PM2.5 (µg m 3)10 597.5-100100-105105-110110-12510 -160Figure 2. (top) Average roadside and urban increments ofPM2.5 and NO2 (de-weathered data) over three years (20162018) for Beijing 6; (bottom) Spatial maps of mean PM2.5 (a)and NO2 (b) concentrations for the winter campaign period(5 November to 10 December 2016), simulated using theMEIC Optimised emissions inventory. Mean measuredconcentrations at monitoring sites (NO2 and PM2.5) and theIAP field site (NO2) are represented by coloured dots 9. Using a newly developed urban-scale traffic pollutiondispersion model with a horizontal resolution of 5minside street canyons, the effects of canyon geometryon the distribution of NOx and CO from traffic emissionswere investigated in central Beijing. It was found thatan increase in building height leads to heavier pollutioninside canyons and lower pollution outside canyons3

70% of summertime conditions, a reduction in VOCsleads to an improvement in ozone air quality 10. Further controls on traffic emissions, including thetransition to an electric fleet, are needed to reduceNO2 concentrations but this will have a limitedbenefit for reduction of the concentration of primaryfine particles. An improved Random Forest algorithm shows thatO3 increased by 14.8( 5.3)% in Beijing due tothe COVID-19 lockdown in spring 2020 althoughthe total gaseous oxidant (Ox NO2 O3) showedcorrespondingly limited change; a slower pace of VOCemission reduction, relative to that for NOx, could riska further increase in O3 pollution 12. Street canyon geometry strongly influences humanexposure to traffic pollutants in the populatedmegacity. Careful planning of street layout andcanyon geometry and consideration of traffic demandand local weather patterns may significantly reduceinhalation of unhealthy air by urban residents.2.3 The first, fully observation-driven assessment of theimpacts of haze on urban photochemistry throughthe effects of aerosol on photolysis, demonstratesthat haze has a major impact on ozone and hydroxylradicals in the urban boundary layer. Reducedhaze has the potential to worsen O3 pollution byenhancing photochemistry 13.Ozone Pollution and Oxidant Chemistry (York,Leeds, Peking, Lancaster, IAP, Birmingham)APHH-Beijing contributes to an improved understandingof ozone pollution in the Beijing region. The dominant primary source of OH radicals wasnitrous acid (HONO), present at levels up to 11 ppb inwinter (compared to 1-2 ppb in London) 14. Central Beijing presents an unusual gas phase chemicalmixture, unlike that observed in many other megacitiesaround the world. It reflects NOx emissions of a largemodern transport fleet alongside VOCs reflective ofhigh levels of industrial activity and solvent-relatedemissions. This atypical urban combination createshigh potential for summertime photochemical ozoneproduction and can lead to unusually elevated ( 150ppb) ground-level urban ozone concentrations, due tolimited availability of surface NO in the city centre toremove O3 by chemical reaction 10. Commercial NO2 analysers used in the air qualitynetwork may overestimate the true abundance of NO2due to retrieval of HONO as NO2 (but we note followthe established international methodologies) 15. OH levels were largely maintained during hazeconditions in wintertime, despite significantreductions in actinic flux, indicating that localproduction of secondary pollutants continues duringhaze events, likely linked to precursor abundance 14.066 The key VOCs in Beijing are indicative of solventusage, and the most abundant species are smalloxygenated VOCs, such as methanol, acetaldehydeand acetone in winter 11.0548440010.53pm [O3] ppb1200.252824020 018 016[VOC] multiplication0362140 Biogenic compounds make significant contribution tothe photochemical ozone creation potential (POCP),photochemical peroxyacyl nitrates (PAN) creationpotential (PPCP) and potential OH reactivity emittedfrom the city but currently contribute only a smallproportion of total reactivity in the atmosphere 11.This is due to the relatively low emission flux at thecity centre.040.1250.125 0.250.5124NO emission multiplication8Figure 3. O3 isopleth diagram as a function of NO and VOCs(the central diamond indicating the average of Beijingunder typical 3 pm conditions during summer APHHBeijing campaign). 1 in the x or y axis means NO emissionor concentration of all measured VOCs during APHH-Beijingsummer campaign; 0.5 means half of the NO emission or VOCconcentration during APHH-Beijing summer campaign 10. Average summertime Beijing ozone formationconditions are shown in Fig. 3, with the centraldiamond indicating typical 3 pm conditions. As canbe seen from the diamond position on the contourplot, to further reduce NO (moving left on the x axis)would lead to further increases in ozone in the city. In4

conditions brought about by synoptic influencesare paramount for the formation of haze episodesin Beijing. During stagnant synoptic conditions, thefeedback is stronger, so there is a cycle of increasingstagnation and increasing concentrations 18. Detailed model analyses demonstrated that theoxidation products of larger VOCs in the Beijingatmosphere produce more organic peroxy radicalsthan was previously thought, implying greater localozone production under high NOx conditions 16. Further reductions in Black Carbon (BC)concentrations at the surface in the centre ofBeijing (due to local emission reductions) thanabove the planetary boundary layer (due to regionalpollution and high level emissions) leads to heatingaloft and limited mixing and thus more stagnantmeteorological conditions during subsequent daysof a pollution event 19.Policy Implications: Beijing ozone control needs to consider acceleratingthe control of VOCs, especially aromatic VOCs fromfuel evaporation and alcohols, ketones and aldehydesfrom domestic and industrial solvent consumption. There may be regional and city sources of solvent usethat would benefit from stricter abatement measures,as the relatively high VOC concentrations are notprimarily a consequence of road transport emissions. Locally emitted BC in Beijing could reach the upperboundary layer through vertical mixing when synopticconditions change. This upper level BC would thenwork to suppress boundary layer development andenhance haze further 19. Biogenic VOC emissions may become increasingly moreimportant to atmospheric chemistry. Governmentsshould discourage or restrict the planting of thosetypes of trees that contribute to air pollution.Scattering aerosolsAbsorbing aerosolsReduced buoyantturbulence2.4 Interaction between Haze and Meteorologyand Implications for Haze Event Prevention(Manchester, Reading)ResucedAerosols scatter andabsorb radiationPM2.5PBL HeightReduced volumefor aerosol mixingIncI n c r e a s e d s econdAPHH-Beijing developed a novel method to determinemixing layer height and applied a Large Eddy Simulationmodel modified for an urban boundary layer thatcan resolve turbulent exchange and treats pollution,radiation and dynamics fully interactively 17.reasesar yIncreased aerosolconcentrationsaeroso A new method was developed to determine mixedlayer height using automatic lidar and ceilometers 17.rmlfo Aerosol particles can scatter and absorb solarradiation to cause net cooling at the surface andwarming above, which alters the thermal profile of theatmosphere, reducing turbulence due to buoyancy.Reduced turbulent mixing suppresses boundary layerdevelopment during the day, minimises the verticalmixing of pollutants and increases surface aerosolconcentrations. This positive feedback loop betweenaerosols, radiation and meteorology can lead tosustained periods of stagnation and has been foundto enhance pollution events (Fig. 4) 18.a tionRHFigure 4. Aerosol – radiation – meteorology feedback loop18.Policy Implications: Reducing black carbon (BC) emissions fromboth inside and outside Beijing could limit thesuppression of the boundary layer and potentiallyreduce the aerosol-meteorology feedback effects onhaze formation. Emission control should not be limited to groundlevel BC, but also needs to consider the upper level(above the boundary layer) BC, likely to be associatedwith elevated sources outside the city. The contribution of aerosol-meteorology feedbacksalone is not enough to suppress the boundary layer tolevels observed in Beijing. The initial meteorological5

It is essential to control urban ammonia emissions. If pollution events are predicted to occur, emissionreduction measures should be taken in advance toreduce the influence of the aerosol cooling effect onthe mixing layer height, which may effectively limitthe occurrence or severity of heavy pollution events.4Normalised flux32.5. Sources of Ammonia and Possible ControlMeasures (IAP, UKCEH)APHH-Beijing has applied multiple approaches toapportion the sources of ammonia, which is a keyprecursor to PM2.5 formation.IsopreneMonoterpeneNH321000:00 High-resolution flux observations found a significantlocal inner-city ammonia emission3 with the diurnalcycles following those of biogenic sources andevaporative emissions rather than fossil fuel markers(Fig. 5) 11. This suggests that urban ammonia emissionsare unlikely to be dominated by traffic sources,contradicting the MEIC emission inventory, but thatas substantial contribution derives from temperaturedriven sources, such as evaporation from wastewater,urban fertiliser use or deposited particles on surfaces 3.06:0012:00Hour of day18:0000:00Figure 5. Diurnal cycle of normalized fluxes of ammonia,isoprene, and monoterpene3,11.2.6 Evaluation of the effectiveness of air pollutionaction plans (Birmingham, Tsing

levels and the impact of air pollution on health. 10. Increases in air pollution are associated with a deleterious mental health effects. Policy suggestion 8. Advocate personal protection measures, such as using face-masks and air purifiers, particularly by people with pre-existing conditions, the elderly and the young, during pollution events.