ozone production in summer in the megacities of tianjin and

Atmos Chem Phys 12 7531ndash7542 2012wwwatmos-chem-physnet1275312012doi105194acp-12-7531-2012copy Author(s) 2012 CC Attribution 30 License

AtmosphericChemistry

and Physics

Ozone production in summer in the megacities of Tianjin andShanghai China a comparative study

L Ran1 C S Zhao1 W Y Xu1 M Han2 X Q Lu2 S Q Han3 W L Lin 4 X B Xu4 W Gao5 Q Yu5F H Geng5 N Ma1 Z Z Deng1 and J Chen1

1Department of Atmospheric and Oceanic Sciences School of Physics Peking University Beijing China2Tianjin Academy of Environmental Sciences Tianjin China3Tianjin Institute of Meteorological Science Tianjin China4Chinese Academy of Meteorological Sciences China Meteorological Administration Beijing China5Shanghai Meteorological Administration Shanghai China now at Key Laboratory of Middle Atmosphere and Global Environment Observation Institute of Atmospheric PhysicsChinese Academy of Sciences Beijing China now at Meteorological Observation Center China Meteorological Administration Beijing China

Correspondence toC S Zhao (zcspkueducn)

Received 27 December 2011 ndash Published in Atmos Chem Phys Discuss 10 April 2012Revised 18 July 2012 ndash Accepted 4 August 2012 ndash Published 17 August 2012

Abstract Rapid economic growth has given rise to a signif-icant increase in ozone precursor emissions in many regionsof China especially in the densely populated North ChinaPlain (NCP) and Yangtze River Delta (YRD) Improved un-derstanding of ozone formation in response to different pre-cursor emissions is imperative to address the highly nonlin-ear ozone problem and to provide a solid scientific basis forefficient ozone abatement in these regions A comparativestudy on ozone photochemical production in summer hasthus been carried out in the megacities of Tianjin (NCP) andShanghai (YRD) Two intensive field campaigns were carriedout respectively at an urban and a suburban site of Tianjin inaddition to routine monitoring of trace gases in Shanghaiproviding data sets of surface ozone and its precursors in-cluding nitrogen oxides (NOx) and various non-methane hy-drocarbons (NMHCs) Ozone pollution in summer was foundto be more severe in the Tianjin region than in the Shang-hai region based on either the frequency or the duration ofhigh ozone events Such differences might be attributed to thelarge amount of highly reactive NMHCs in Tianjin Industryrelated species like light alkenes were of particular impor-tance in both urban and suburban Tianjin while in Shang-hai aromatics dominated In general the ozone problem inShanghai is on an urban scale Stringent control policies onlocal emissions would help reduce the occurrence of high

ozone concentrations By contrast ozone pollution in Tianjinis probably a regional problem Combined efforts to reduceozone precursor emissions on a regional scale must be under-taken to bring the ozone problem under control

1 Introduction

China has recently experienced an acceleration of economicgrowth and urbanization Elevated anthropogenic emissionsof primary gaseous pollutants have accordingly been notedin the last few decades with a continuous increase to beforeseeable in the near future (Streets and Waldhoff 2000Klimont et al 2001 Van der A et al 2006 2008) It is there-fore not surprising that air quality degradation has been ex-tensively reported (eg Chan and Yao 2008 and referencestherein) The North China Plain (NCP) and the Yangtze RiverDelta (YRD) are two of the most populated and industrial-ized regions in China and currently suffer from air pollu-tion in many aspects (Shao et al 2006) The ground-levelozone problem is among the most stubborn environmentalissues given the ubiquitous existence of the precursors NOxand volatile organic compounds (VOCs) that would photo-chemically generate ozone in the presence of sunlight ashas been well documented in a set of research articles (eg

Published by Copernicus Publications on behalf of the European Geosciences Union

7532 L Ran et al Ozone production in summer in Tianjin and Shanghai

Haagen-Smit 1952 Haagen-Smit et al 1953 Seinfeld andPandis 1998) Thus there is an urgent need to tackle theozone problem in these regions where precursor emissionssignificantly rise in order to avoid possible threats to the pub-lic and ecosystems that might be potentially raised by highconcentrations of ozone and its precursors (Chameides et al1994 Jacobson 2002)

Previous studies related to ozone issues in the NCP andYRD regions were carried out separately In the NCP re-gion most observations and analysis were performed at re-gional background sites or in the megacity of Beijing (egGao et al 2005 Lin et al 2008 2009 2011 Tang et al2010 Lu et al 2010 Xu et al 2011a) In preparation forthe 2008 Olympics the relocation of old industrial facilitiesaway from Beijing and other stringent emission restrictionswere implemented to mitigate local air pollution It turnedout that the control of secondary pollutants such as ozone wasless successful than most primary gases as a result of regionalpollution A considerable contribution was demonstrated tobe made by regional sources like those in Tianjin (Streets etal 2007 Wang et al 2010) However current research re-garding the ozone problem in Tianjin is rather limited (Hanet al 2011 Ran et al 2011 Xu et al 2011b) In the YRDregion the expansion of population and economics centersaround the megacity of Shanghai where ozone issues havebeen more elaborately addressed but mainly in the thrivingdowntown area (eg Cheung and Wang 2001 Zhao et al2004 Geng et al 2008 Tang et al 2008 Ran et al 2009)In an effort to understand ozone behaviour in responses toenhanced precursor emissions and to design feasible and ef-fective emission control strategies more work needs to bedone in these fast developing regions

In this paper observations regarding ozone and its precur-sors from an intensive field campaign in summer 2010 in ur-ban Tianjin and from routine monitoring in summer 2009 inboth urban and suburban Shanghai were comparatively ana-lyzed and discussed with the dataset from the 2009 Haze inChina (HaChi) summer campaign in suburban Tianjin (Ranet al 2011) This would hopefully cast more light on un-derstanding and handling current ozone problem in the twomegacities

2 Experiments and methodology

21 Sites

Observations of gaseous pollutants were conducted respec-tively at an urban and a suburban site in the megacities ofTianjin and Shanghai Sitting on the eastern edge of the pol-luted NCP Tianjin is adjacent to the Bohai Bay as illustratedin Fig 1a The Tieta site (3906prime N 11710prime E) where the2010 summer campaign took place is located in a thicklysettled urban district of Tianjin The largest petrochemicalcomplex in China has recently been built up in the Tianjin

Binhai New Area which is situated about 40 km southeastof urban Tianjin and extends nearly 150 km along the coastalline The suburban site Wuqing (3923prime N 11701prime E) is thelocation where the 2009 HaChi summer campaign was car-ried out (site description refers to Ran et al 2011) The siteXujiahui (3112prime N 12126prime E) is a routine monitoring sta-tion in the crowded metropolitan area of Shanghai (Fig 1b)which is overburdened with dense population high trafficloading and numerous tower buildings Near the coast andalmost 50 km south of the urban center is a petrochemical in-dustrial area wherein the site Jinshan (3045prime N 12121prime E)is situated The residential population there is much less thanin the urban center and the land cover is generally well veg-etated

To support related discussions the observational data ofsurface ozone during July and August 2009 at another threeNCP sites are also used The urban site in the megacity ofBeijing is located in the courtyard of China Meteorologi-cal Administration (CMA 3957prime N 11619prime E) which isbetween the busy 2nd and 3rd ring roads in the northwestarea of urban Beijing (Lin et al 2011) The site CMA isabout 110 km northwest of Tianjin (Fig 1a) Gucheng (GCH3908prime N 11540prime E) is a polluted rural site about 110 kmsouthwest of Beijing and 130 km west of Tianjin (Lin et al2009) Shangdianzi (SDZ 4039prime N 11707prime E) is a WMOregional atmosphere background station on the north rimof the NCP region showing footprints of polluted regionalbackground and clean natural background according to winddirection (Lin et al 2008)

22 Measurements and data processing

Continuous monitoring of ozone and NOx took place in JulyndashAugust 2009 at both sites of Shanghai and in suburban Tian-jin while observations were made during JulyndashAugust 2010in urban Tianjin The commercial Thermo EnvironmentalInstrument (TEI Inc USA) Model 49C a UV Photomet-ric Analyzer was used to measure ozone concentrations atboth sites of Tianjin CMA GCH and SDZ Surface ozonein Shanghai was observed using the UV Absorption OzoneAnalyzer (EC9810BECOTECH) based on the same opera-tion principle As for NOx measurements TEI 42CTL NO-NO2-NOx Analyzer at Wuqing simultaneously determinedthe ambient concentrations of NO and NOx utilizing chemi-luminescence techniques and a heated molybdenum NO2 toNO converter This instrument was able to accurately mea-sure the amount of nitrogen oxides to a trace level of lessthan 50 pptv TEI 42i used at Jinshan and the ECOTECHOxides of Nitrogen Analyzer (EC9841B) used in urban ar-eas of Tianjin and Shanghai were operated on the basis ofsimilar techniques as TEI 42CTL Besides carbon monox-ide (CO) was measured by TEI 48C CO Analyzer during thesame periods of ozone measurements at both sites of TianjinMultipoint calibrations and daily maintenance were carefullyperformed at all sites following USEPA recommendations on

Atmos Chem Phys 12 7531ndash7542 2012 wwwatmos-chem-physnet1275312012

L Ran et al Ozone production in summer in Tianjin and Shanghai 7533

5

55

5

10

10

10

10

15

15

15

1515

15

20

20

20

20

GCH

SDZ

CMA

WuqingTieta

( )a

113

E 114

E 115

E 116

E 117

E 118

E 119

E 120

E37

N

38

N

39

N

40

N

41

N

5

55

5

10

10

10

10

10

15

15

15

Xujiahui

Jinshan

( )b

116

E 117

E 118

E 119

E 120

E 121

E 122

E29

N

30

N

31

N

32

N

33

N

NO

21

5sdot

-2)

0

5

10

15

20

25

Bohai Bay

Fig 1 Average NO2 distributions during March 2009ndash2010 obtained from OMI (Ozone Measuring Instrument) Level 3 data in a resolutionof 025 times 025 (a) in the NCP(b) in the YRD The white lines are contour lines of NO2 tropospheric column amounts The black linesdisplay the borders of provinces and coastlines The grey lines show the borders of counties

quality assurance and quality control (USEPA 2008) Mix-ing ratios (ppbv) of ozone NOx and CO were acquired at aresolution of one minute Hourly averages were calculatedfor each hour with at least 75 valid data

Sampling protocols of non-methane hydrocarbons(NMHCs) differed among the four sites At Tieta a 32 lcanister was used to continuously collect ambient airsamples over each 2-h period (0730ndash0930 1030ndash12301330ndash1530 1630ndash1830 and 1930ndash2130 LT) from 22to 28 August 2010 At Wuqing samplings were conductedduring five periods of time (0730ndash0930 1100ndash13001400ndash1600 1700ndash1900 and 2100ndash2300 LT) between6 and 13 August 2009 An 8 l Teflon bag was used tocollect instant air samples at a 30-min interval Each 8 l airsample thus represents the average condition for the 2-hsampling period At both sites of Shanghai ambient air wascontinuously sampled into a 6 l canister over each 3-h periodfrom midnight Thus eight samples were obtained on eachday of 29 July to 6 August at Jinshan and 25 to 31 Augustat Xujiahui respectively Nonetheless analytic methods fordetermining specific NMHC species were generally the samefor all air samples A total of 52 hydrocarbons including28 alkanes 8 alkenes and 16 aromatics were identified andquantified (in mixing ratios ppbv) using a coupled GasChromatographyMass Spectrometry (Agilent Tech Ran etal 2009) in the laboratory according to USEPA methodsTO-14A and TO-15 (USEPA 1999a b) The carbon-atombased concentration (in ppbC) was another straightforwardexpression for the measurements To scale the relativereactivity of various hydrocarbons the propene equivalentconcentration (Propy-Equiv in ppbC) was calculated foreach species As defined in Chameides et al (1992) thePropy-Equiv concentration was derived by weighting thecarbon-atom based concentration of a single species withthe corresponding OH reaction rate constant and then beingnormalized to that of propene The reaction rate coefficientsof individual species with OH radical were mostly adoptedfrom Atkinson (1990) and also from Middleton et al (1990)This widely used scaling approach neglected the impacts of

subsequent chemical mechanisms on ozone production afterinitial OH oxidation yet provided a simple way for directintercomparisons

Meteorological parameters were recorded in a temporalresolution of one minute by the Automatic Weather Stationinstalled at each site Limited transport and wet depositionof gas pollutants were expected on days with wind speedless than 4 m sminus1 and without precipitation in the daytimeSince winds were usually stronger in the suburbs than in ur-ban areas it was believed to be reasonable that data withwind speed less than 6 m sminus1 in the two suburbs should alsobe counted in Based on such rules 19 and 14 days wereselected at Tieta and Wuqing respectively At Xujiahui andJinshan the respective numbers were 10 and 11 Photochem-ical process was supposed to predominantly contribute to ob-served ozone concentrations on these selected days Selectedozone and NOx data are used in Sect 33 to analyze their di-urnal cycles under photochemistry-dominant circumstances

23 Calculation of ozone production rates

Ozone production is an essential term for determining theextent to which in-situ ozone could be ascribed to pho-tochemistry According to our current knowledge on thechemical mechanism it is hydroperoxy radical (HO2) pro-duced from the oxidation of CO and organic peroxy radi-cals (RO2) produced from the oxidation of NMHCs that ox-idize NO into NO2 and subsequently lead to net productionof ozone by NO2 photodissociation (Seinfeld 1989 Atkin-son 1990 2000 Jenkin and Clemitshaw 2000) The instan-taneous production rate of ozone usually denoted asP (O3)could thereby be written as

P(O3) = [NO](k[HO2] +

sumi

ki[RiO2]) (1)

wherek is the rate constant for the reaction of NO and HO2andki is the rate constant for each reaction of NO andRiO2(Trainer et al 2000) Several methods have been developedto estimateP (O3) based on measurements and applied to the

wwwatmos-chem-physnet1275312012 Atmos Chem Phys 12 7531ndash7542 2012


0 50 100 150 2000

005

01

015

02

025

Ozone (ppbv)

Pro

babi

lity

Dis

trib

utio

n

(a) Tieta Urban TianjinWuqing Suburban TianjinXujiahui Urban ShanghaiJinshan Suburban Shanghai

0 50 100 150 2000

005

01

015

02

025

03

NOx (ppbv)

Pro

babi

lity

Dis

trib

utio

n

(b) Tieta Urban TianjinWuqing Suburban TianjinXujiahui Urban ShanghaiJinshan Suburban Shanghai

0 50 100 150 200

0

005

01

015

Ozone (ppbv)

Pro

babi

lity

Dis

trib

utio

n

(c) Tieta Urban TianjinWuqing Suburban TianjinCMA Urban BeijingGCH Rural NCPSDZ Background NCP

0 1 2 3 4 50

002

004

006

CO (ppmv)

Pro

babi

lity

Dis

trib

utio

n

(d) Tieta Urban TianjinWuqing Suburban Tianjin

Fig 2Probability distributions of(a) ozone and(b) NOx (both 5 ppbv per bin) in summer at the four sites(c) ozone at five sites in summer2009 in the NCP (5 ppbv per bin)(d) CO at both sites in Tianjin (005 ppmv per bin)

analysis of ozone photochemical process (Frost et al 1998and references therein Kleinman 2000)

In this paperP (O3) is calculated using the concentra-tions of model simulated peroxy radicals and measured NOA photochemical box model NCARMM (NCAR MasterMechanism Version 23) with a detailed description of thechemical mechanism is used to estimate the short-lived rad-ical species (Madronich and Calvert 1990 Herring et al1997 Madronich and Flocke 1999) Model simulations areinitialized by available measurements at 0200 pm localtime at each site Each run lasts for half an hour for radi-cals to reach a steady state before losing initial informationTwo types of chemical inputs are arranged One is to averagethe midday NMHC data over the sampling period and adjustinitial NOx concentration for each run (1 ppbv and from 2 to50 ppbv with an interval of 2 ppbv) The other is to perform asimulation for each NMHC sampling day by inputting mid-day NMHC data and 30-min averages of measured NOx Inaddition the set of input parameters other than location infor-mation are completely the same Regardless of the reversiblepathway for peroxy radicals to react with NO in forming per-nitric acid and organic nitrates that would quickly be decom-posed back to their reactants at a prescribed temperature of303 K in the model runs peroxy radicals mainly react withNO to yield NO2 in an amount given by the stoichiometriccoefficients At the end of each run simulated HO2 and RO2are used forP (O3) calculation according to Eq (1)

3 Results and discussion

31 Probability distributions of ozone and NOx

Ozone probability distributions (5 ppbv per bin) in all loca-tions are illustrated in Fig 2a The occurrence of 1-h aver-age ozone concentrations exceeding 80 ppbv is usually asso-ciated with high ozone episodes that pose threats to humanhealth A relatively higher probability of hourly ozone av-

erages over 80 ppbv in Tianjin indicates more severe ozonepollution and a higher ozone exposure risk in summer in thisregion than in Shanghai

A close similarity could be found in ozone probabilitydistributions at both sites of Tianjin implying a good re-gional mixing over the polluted NCP However we shouldbe cautious to conclude so quickly given the disagreement ofthe observational periods Surface ozone measured at CMAGCH and SDZ in the NCP were then further examined Theprobability distributions of surface ozone at the five NCPsites generally resemble each other (Fig 2c) especially atGCH CMA Wuqing and Tieta which lie within a clus-ter of urban centers in the NCP The frequency distribu-tions of ozone daily maximum and ozone episode (1-h av-eragegt 80 ppbv) duration are generally similar among thosesites also suggesting ozone pollution in the NCP is probablya regional problem This is consistent with the existence of ahigh-ozone belt peaking between Beijing and Tianjin demon-strated by Lin et al (2009) on the basis of the regional dis-tribution of ozone tropospheric column amounts It is foundthat CO probability distributions at both sites of Tianjin arebasically the same (Fig 2d) also in support of well regionalmixing of gas pollutants in the NCP Nevertheless an ex-tended network of monitoring sites would be needed to pro-vide more solid evidence for regional characteristics

In contrast ozone probability distribution in urban Shang-hai is completely different from that in the suburb (Fig 2a)Ozone probability distribution peaks at about 20 ppbv at Jin-shan while hourly averages are largely (sim 60 ) in the rangeof 0ndash20 ppbv at Xujiahui where the frequency of ozone con-centrations between 0 and 5 ppbv is even as high as 22 Strong NO emission from busy traffic should be most respon-sible for such a high frequency of very low ozone level inthe downtown in comparison with the suburban site Jinshanwhere fewer automobiles are there

NOx as a short-lived species is often characterizedas more representative of local impacts than long-range



Table 1Average mixing ratios and fractions of the three NMHC categories at the four sites

Tieta Wuqing Xujiahui JinshanPPPPPPPSites

ConcUrban Tianjin Suburban Tianjin Urban Shanghai Suburban Shanghai

ppbC ppbC ppbC ppbC

Alkanes 1132 366 2318 529 544 448 243 244Alkenes 245 79 373 85 47 39 41 41Aromatics 1719 555 1696 386 623 513 712 715

Total 3096 4387 1214 996

transport or regional mixing Figure 1 displays the averagedistribution of tropospheric NO2 columns from Ozone Mea-suring Instrument (OMI) in the NCP and YRD regions OMIis a key instrument on NASArsquos Earth Observing Systemrsquos(EOS) Aura satellite with a spatial resolution of 13 times 24 km2

at nadir OMI NO2 Level 3 data used in this paper are griddedproduct with a horizontal resolution of 025 times 025 in lati-tude and longitude Generally NO2 (NOx) pollution is moresevere in the NCP than in the YRD Besides average dis-tributions of OMI NO2 exhibit differences in NO2 concen-trations within the megacity of Shanghai in contrast to thehigh-NO2 belt over the entire Tianjin-Beijing region Thismight also imply that ozone issues in Tianjin are probably ona regional scale while on an urban scale in Shanghai

In-situ NOx observations provide us more details on asmaller scale that satellite data may not be able to distinguishDifferent features of NOx probability distributions could berecognized between urban and suburban areas (Fig 2b) Amuch greater portion of the hourly data falls into bins withlow concentrations at suburban sites than urban sites Jin-shan is the cleanest location when referring to NOx sincetraffic burden is rather light in this area Additionally south-easterly winds dominate in the summer and bring in clean airmasses from the East China Sea Influenced by limited sourceemissions as well as the dilution observed NOx concentra-tions are often below 10 ppbv and hardly exceed 40 ppbvThe average value is about 11 ppbv close to the concentra-tion corresponding to the highest probability At Wuqing av-erage NOx concentration is about 20 ppbv below which al-most half of the data are The probability distribution reachesits maximum around 10ndash15 ppbv with a sharp decline fromthe peak value toward either direction It is noteworthy thatthe frequency for 0ndash5 ppbv unlike at Jinshan is almost zeroat Wuqing Polluted events with NOx concentrations in ex-cess of 50 ppbv could also be occasionally encountered inthis area As an industrial area near the transport hub insidethe polluted NCP Wuqing is apparently more polluted thanJinshan In the urban centers of both megacities frequenciesof hourly NOx averages distribute similarly below 100 ppbvThe curves display a maximum between 20ndash25 ppbv andshow the main part in the range of 10ndash40 ppbv

32 Characterization of NMHCs

Average carbon-atom based concentrations of total NMHCsand each category in all locations are listed in Table 1 Ahigher level of total NMHCs is found in the Tianjin region(300ndash400 ppbC) about 2 to 3 times larger than the over-all concentration in the Shanghai region In both urban cen-ters automobile emissions undoubtedly make a substantialcontribution to the NMHC composition The traffic indica-tor methyl tert-butyl ether (Chang et al 2003) is averagelyabout about 14 ppbC at Tieta and 6 ppbC at Wuqing whereas15 and 05 ppbC at Xujiahui and Jinshan respectively (notshown in this paper) The NMHC group in urban atmosphereof the two megacities is similarly composed of about 50 aromatics followed by alkanes and alkenes In the two in-dustrialized suburbs aside from the marked difference in to-tal concentrations the compositions also differ remarkablyOver 50 of the measured NMHCs are made up of alkanesat Wuqing with n-hexane alone to contribute 11 Evapo-ration of solvents used in electronics and pharmaceutical in-dustries probably provides abundant sources for heavy alka-nes (C ge 5) as well as certain aromatic species like toluene(10 ) and xylenes (7 ) Due to such large evaporation ofheavy alkanes from industrial processes in this area it is un-surprising yet noteworthy that total NMHC concentration atWuqing is substantially larger than that at the urban site Ti-eta The predominant component in Jinshan is the aromaticgroup (amounting to 72 ) that could be largely releasedfrom the process of petroleum refining Toluene accounts forroughly 50 of the mixture As a matter of fact carbon-atom based concentrations of toluene are very close (around50 ppbC) at Jinshan Tieta and Wuqing under the influencesof petrochemical emissions The ambient level of toluene atXujiahui is only a half of that in other locations

NMHC reactivity is more relevant to ozone formation incomparison with measured mixing ratio Based on Propy-Equiv concentrations the highest NMHC reactivity is foundat Wuqing with a value of about 200 ppbC (Fig 3) In urbanTianjin average Propy-Equiv concentration of total NMHCsis about 130 ppbC NMHC reactivities at both sites of Shang-hai are less than one quarter of the NMHC reactivity atWuqing and close to each other The ratio of Propy-Equiv



0 50 100 150 200 250 30001

02

03

04

05

06

07

08

PropyminusEquiv Conc (ppbC)

Pro

pyminus

Equ

iv C

onc

pp

bC C

onc

Tieta Wuqing Xujiahui Jinshan Std

Fig 3NMHC reactivity at each site of Tianjin and Shanghai Mark-ers represent average Propy-Equiv concentrations during the sam-pling period (x-axis) as well as the ratio of Propy-Equiv concen-tration to carbon-atom based concentration (y-axis) Standard devi-ations are given in black lines

concentration to carbon-atom based concentration could betreated as a simplified indicator for reactivity related compo-sition Such ratios (Fig 3) reveal that the NMHC group atWuqing consists of more reactive species Reactivity compo-sitions at Jinshan and Tieta are similar leaving the share ofreactive species in Xujiahui to be the lowest

The variability of a dataset also has useful implications fordata interpretation Figure 3 gives standard deviations of bothPropy-Equiv concentrations and the ratios of Propy-Equivconcentration to carbon-atom based concentration Duringthe sampling period a stable weather condition has beenfound in urban Tianjin Variations in temperature and pres-sure were very limited and wind speeds barely exceeded3 m sminus1 in favour of local pollution and accumulation Undersuch conditions a distinct diurnal cycle of reactivity relatedcomposition repeated every day with a low valley in the af-ternoon Large variability in Propy-Equiv concentrations isattributable to this diurnal variation At Wuqing the sam-pling period could be subdivided into two periods either de-pending upon NMHC composition in terms of Propy-Equivfractions or surface winds (also see Ran et al 2011) Ele-vated total NMHC reactivity together with percentages of re-active components was found in the second period when airparcels passed by the sampling site from the well vegetatedwestern area As a consequence a high degree of variabil-ity is found both in Propy-Equiv concentrations and the ra-tios of Propy-Equiv to carbon-atom based concentration Thesmallest variability in reactivity related composition has beenfound in Xujiahui although wind direction varied and dailymaximum wind speed ranged from 4 m sminus1 to 8 m sminus1 duringthe sampling period implying dominant emission sources to

be local and stable The largest variability in NMHC reac-tivity is found at Jinshan This could be mainly explainedby the distinct diurnal variation of isoprene which rises tothe highest level of about 4 ppbC around midday and falls tonearly zero at night due to the variability of emission rate inresponse to temperature and sunlight (Guenther et al 19911993) Since easterly winds generally dominate during thesampling period hydrocarbons of anthropogenic origin dis-play a very limited variation in the composition

The relative ozone forming potential of individual speciesis assessed in terms of OH-reactivity weighted equivalentconcentration (Table 2 also see the average column inFig 6) The top 10 species account for at least about 70 of the total reactivity in all locations At urban sites topspecies that are most relevant to ozone formation exhibitsomewhat similar ranking with mp-xylene and toluene to bethe most important Consistent with results of previous inves-tigation in Xujiahui (Ran et al 2009) total NMHC reactivityis mainly contributed by aromatic species which are thoughtto be largely emitted by automobiles and solvent evaporationin urban centres (Lu et al 2003 Song et al 2007 Cai et al2010) However the relative importance of isoprene differssignificantly from negligible (ranked below 50th) in previ-ous study to considerable (within top 5) in this study Theseasonal variation of vegetative cover as well as biogenicemission intensity should be responsible for such a differ-ence between the annual isoprene average in previous studyand summertime level in this study Unlike in urban areas itis both aromatic and alkene group that mainly constitute top10 species at Wuqing Isoprene is found to be most influentialon the total reactivity This was ascribed to strong isopreneemission in the well vegetated suburb and a much higherOH reactivity of isoprene compared to most anthropogenicspecies Trans-2-butene and cis-2-butene might be associatedwith production of petroleum and manufacture of syntheticrubber Their fractions amount to about 19 of the total re-activity Another particular aspect of top species at Wuqingis the importance of some industry related heavy alkaneslike n-hexane which is a widely used cleaning solvent inmany industrial processes and also contained in gasoline andpetroleum products At Jinshan toluene is particularly im-portant to total OH reactivity followed by mp-xylene andisoprene The components of top species are almost exclu-sively aromatics and light alkenes which should be emit-ted by the large petrochemical complex given that Jinshanis a less populated residential area with relatively low trafficloading compared to the urban area

Last but not least characterizations of NMHC data in-volved in the discussion are subject to impacts of differ-ent sampling protocols at the four sites This might lead touncertainties in the analysis results to some extent Long-term measurements with the same sampling duration and fre-quency would be critical to gain more detailed informationon NMHC characteristics and also be helpful in inferringemission sources from measured NMHC composition



000 300 600 900 1200 1500 1800 2100 24000

20

40

60

80

100

120

Ozo

ne (

ppbv

)

(a)

Tieta Urban TianjinWuqing Suburban TianjinXujiahui Urban ShanghaiJinshan Suburban Shanghai

0

10

20

30

40

50

60

70

80

NO

x (p

pbv)

NOx (Tieta)NOx (Wuqing)NOx (Xujiahui)NOx (Jinshan)

000 300 600 900 1200 1500 1800 2100 2400

TN

MH

Cs

10 (

ppbC

)

(b)

NMHCs (Tieta)NMHCs (Wuqing)NMHCs (Xujiahui)NMHCs (Jinshan)

Fig 4Average diurnal cycles of(a) ozone and(b) its precursors on selected days (the selection criteria have been described in Sect 22)

Table 2Top 10 species sorted by contributions to total OH reactivity at the four sites

Tieta Urban Tianjin Wuqing Suburban Tianjin Xujiahui Urban Shanghai Jinshan Suburban ShanghaiSpecies Species Species Species

mp-Xylene 200 Isoprene 160 mp-xylene 244 Toluene 315Toluene 96 trans-2-Butene 106 Toluene 111 mp-xylene 165trans-2-Butene 91 cis-2-Butene 81 Ethylbenzene 77 Isoprene 112124-Trimethylbenzene 74 mp-Xylene 75 Isoprene 50 13-Butadiene 54cis-2-Butene 71 3-Methylhexane 60 o-Xylene 43 Ethylbenzene 37Isoprene 49 n-Hexane 49 Propene 37 Propene 30Propene 48 Toluene 48 n-Pentane 36 o-Xylene 30Ethylbenzene 46 135-Trimethylbenzene 41 124-Trimethylbenzene 36 124-Trimethylbenzene 23o-Xylene 35 124-Trimethylbenzene 41 Isopentane 29 trans-2-Butene 18n-Pentane 31 123-Trimethylbenzene 34 trans-2-Butene 29 n-Hexane 18Top 10 species 741 Top 10 species 695 Top 10 species 692 Top 10 species 802

33 Diurnal cycles of ozone and its precursors

As mentioned in Sect 22 selected ozone and NOx dataare used to discuss their average diurnal variations underphotochemistry-dominant conditions Average diurnal cycleof NMHCs during the sampling period is also examined foreach site

In general ozone concentrations undergo a diurnal patternof reaching a maximum in the afternoon and a minimum inthe early morning A slight decline could be observed frommidnight to the dawn in the polluted NCP and urban Shang-hai (Fig 4a) as a response to NOx accumulation within thenocturnal planetary boundary layer Nighttime NOx concen-trations remain such a high level (averagely ranging fromabout 20 ppbv to more than 30 ppbv among the three loca-tions as shown in Fig 4b) that efficient depletion of ozonehas been caused before sunrise Occasionally surface ozonecould be completely consumed up by strong local NO emis-sions On the contrary nighttime ozone stays stably at about20 ppbv at Jinshan along with an average NOx of 10 ppbvthat reveals low NO emissions and NO titration of ozone atnight Ozone concentrations rise rapidly once photochemical

processes start up The large increase in surface ozone con-centrations could also be attributable to vertical mixing thatbrings in ozone-rich air from the residual layer aloft Theturning point where ozone accumulation rates change fromnegative to positive values exhibits approximately 1-h delay(sim 0630 am) in urban Shanghai in comparison with otherlocations Since fresh NO emissions are markedly strength-ened during morning rush hours in urban areas ozone ac-cumulation is minimized or even completely inhibited due toNO consumption especially when NMHC concentration andreactivity are not high enough to accelerate ozone produc-tion as in the case of Xujiahui Ozone concentrations increasewith an observed accumulation rate ranging from 10ndash20 ppbvper hour preceding noon arrive at its daily maximum in theafternoon and go down thereafter The highest average dailymaximum of nearly 110 ppbv is found at Wuqing followedby a value of about 95 ppbv at Tieta and Jinshan The lowestaverage peak concentration is slightly over 80 ppbv at Xuji-ahui Averagely the duration of hourly ozone concentrationsthat violate the threshold of 80 ppbv is the longest in Tian-jin (sim 6 h) followed by Jinshan (sim 4 h) and Xujiahui (only1 h) On about 25 of the observational days the occurrence



1 2 4 6 8 10 20 30 400

5

10

15

20

25

30

35

40

45

50

NOx (ppbv)

P(O

3) (p

pbv

h)

TietaWuqingXujiahuiJinshan

Fig 5 Simulated instantaneous production rates of ozone(ppbv hminus1) as a function of NOx (ppbv) Model results using thetwo chemical inputs are illustrated in the same colour for each siteLines represent model results from prescribed NOx concentrationsand NMHC measurements averaged over the whole sampling pe-riod Markers represent model results from both measured NOx andNMHCs on each day of the sampling period if available

of daytime ozone exceedances (gt 80 ppbv) lasting for morethan 6 h was observed in Tianjin region while none was en-countered in Shanghai There were only 3 of the observa-tional days on which daily ozone exceedances lasted for 4 hat Xujiahui and 19 at Jinshan This reveals a more persis-tent ozone exposure risk in the Tianjin region than in Shang-hai Besides it is notable that concurrent existence of highozone and NOx happened sometimes at night in Tianjin asan evidence for regional mixing of aged air mass rich in NOxand ozone in this polluted region

Average ozone daily peak increases from the urban centerto the suburb in both regions (Fig 4a) Calculated morningNMHCsNOx ratios are exclusively larger than 10 at Wuqingand mainly between 8 and 10 in the urban center (Fig 4b)At theses high NMHCsNOx ratios the generation of netozone is limited by the competition between the formationof organic nitrates by peroxy radicals reacting with NONO2and the pathway to form NO2 by peroxy radicals reactingwith NO (Finlayson-Pitts and Pitts Jr 1986 NRC 1991)Any reduction in available NOx would limit the formation ofozone However organic nitrates often quickly undergo ther-mal decomposition in hot summer Higher ozone productionrates could still be expected under higher NMHC reactivityThus reactive NMHCs abundantly released from a numberof nearby industrial facilities in Wuqing possibly lead to thehigher ozone peak in there than in the urban center In theShanghai region a different story is told Ozone formationhas previously been found to be typically NOx-inhibited inXujiahui (Geng et al 2008 Ran et al 2009) as also in-

dicated by measured morning NMHCsNOx ratios closelyaround 4 in this study NOx is mainly emitted by motor ve-hicles and greatly inhibits net ozone production During thedaytime strong local traffic emission of NO helps maintainthe highest NOx concentration and the lowest ozone peakamong all locations The pronounced double peaks in NOxdiurnal variation are associated with heavy traffic during rushhours causing ozone daily peak shifted to an earlier timeAt Jinshan the level of NMHCs is comparable to that in ur-ban Shanghai while NOx concentrations are much lower andno diurnal variation is found Calculated morning ratios ofNMHCsNOx are variable from one chemical regime to an-other

34 Ozone production rates

Model simulated ozone production rates for the four sitesare shown in Fig 5 as a function of NOx concentrationsThe four curves that are calculated using the average NMHCcomposition around midday at the four sites differ greatly Ateach given NOx concentration the highestP (O3) is found atWuqing followed by Tieta Xujiahui and Jinshan in goodagreement with the relative magnitude of their prescribedNMHC reactivity in model runs The discrepancies ofP (O3)

between the four locations increase with increasing NOx con-centration Above several ppbv of NOx which is easily en-countered in these areas ozone production rates are foundto be strongly dependent on NMHC reactivity This mightexplain in part the observed differences of photochemicalozone from one location to anotherP (O3) undergoes simi-lar variation in response to NOx change at each site that is toincrease almost linearly at first and then go down after reach-ing the peak value The concentration of NOx to which max-imumP (O3) corresponds however shifts from about 6 ppbvat Jinshan to 20 ppbv at Wuqing As for specific simulationcases that rely on measurements made on a particular day aconsiderable scatter has been found for each site partly dueto a difference in average and specific NMHC compositionand partly due to the use of measured NO instead of sim-ulated one In spite of the dispersion case simulations foreach site generally fall into the same chemical regimes ex-cept for that in Jinshan According to the model results ac-tive ozone photochemical production found in Wuqing wouldburst into even more drastic enhancement when bumping intoair parcels rich in NOx In the two urban centers ozone pro-duction is more or less inhibited by high NOx level Ozoneproduction seems to be highly variable in Jinshan given thelarge variability of the NMHC reactivity shown in Fig 3This implies detailed analysis for each specific case favor-ably coupled with meteorology is needed in this area

Diurnal fractions of different NMHC groups are examinedin Fig 6 The composition of NMHC reactivity is relativelystable throughout the day in urban centers while more vari-able in the suburbs as a result of isoprene emission in the day-time Around midday isoprene is of particular importance



Table 3A summary for comparisons of ozone NOx and NMHCs attributes in the four locations

Items to make comparisons Tianjin Shanghai

Tieta Urban Wuqing Suburban Xujiahui Urban Jinshan Suburban

Ozone Pollution Episodes Frequency High Markedly high Low Moderate

(gt 80 ppbv) Duration Long (gt 6 h) Long (gt 6 h) Negligibly short Relatively long

Diurnal Startup Dawn Dawn 1 h delay Dawn

Variation Peak Time Around 1500 Around 1500 Shift to 1 h earlier Around 1500

NOx Average Polluted Moderate Polluted Clean

Pollution Episodes gt 50 ppbv Readily Occasionally Often Barely

Diurnal Variation Double Peaks Evitable Evitable Pronounced Not found

NMHCs Concentration High High Low Low

Reactivity Moderate High Low Low

Composition Key Groups AromaticsLight alkenes

AromaticsLight alkenesIsoprene

Aromatics AromaticsIsoprene

20

40

60

80

100

Tieta

20

40

60

80

100

Wuqing

20

40

60

80

100

N

MH

C R

eact

ivity

Bas

ed F

ract

ions

Xujiahui

20

40

60

80

100

000 300 600 900 1200 1500 1800 2100 2400 Average

Jinshan

Light Alkanes Heavy Alkanes Alkenes Isoprene Toluene Xylenes Other Aromatics

Fig 6Average diurnal fractions of different NMHC groups at eachsite

to fostering high ozone concentrations in the suburbs andless contributing in urban areas High ozone production ratesfound at Wuqing should be ascribed to biogenic isoprene andindustry related alkenes As suggested in Ran et al (2011)it would be more feasible to reduce anthropogenic emissionsrather than background biogenic emissions for the purposeof ozone abatement in particular to make restriction criteri-ons for industrial emissions In both urban areas aromaticsthat largely come from motor vehicles are considerably re-sponsible for ozone formation Since ozone production ratesdepend on NOx and NMHCs which are both largely fromautomobile emissions in cities a second thought should begiven before implementing control strategies on automobileemissions that may lead to the migration of chemical regimeAlthough aromatics are key species related to ozone formingpotentials in Jinshan on a daily average it is isoprene that

contributes most around midday when photochemical pro-cess is most active Thereby we suggest that even effectiveindustrial emission reductions would be of little use to miti-gate ozone pollution in this area

4 Conclusions

A comparative study is carried out in this paper using a setof observational data at an urban and a suburban site in themegacities of Tianjin (NCP) and Shanghai (YRD) as an ef-fort to understand and handle current ozone problem in thesefast developing megacities where a significant increase inozone precursor emissions has been noted

A distinction between the characteristics of ozone andits precursors has been found between Tianjin and Shang-hai (Table 3) In the summer high ozone concentrations(gt 80 ppbv) of long duration (gt 6 h) were frequently encoun-tered in both urban and suburban Tianjin while the occur-rence of high ozone concentrations lasted for a shorter period(usuallylt 4 h) and had a much lower frequency in Shanghaiindicating more severe ozone pollution in the Tianjin regionSuch differences in ozone behavior are thought to be largelyattributable to higher NMHC concentration and reactivity inTianjin than in Shanghai Model simulations based on mea-surements also reveal similar dependence of ozone produc-tion rates upon NMHC reactivity

On average total NMHC concentration as well as reactiv-ity in Tianjin is approximately 2 to 3 times larger than thatin Shanghai NMHC compositions in the two urban centersare to some extent similarly determined by emissions fromautomobiles and influenced by regional industrial sourcesIndustry related species like light alkenes and heavy alkanesare important components of total NMHC mixture in subur-ban Tianjin while aromatics dominate in suburban Shanghai



Although anthropogenic sources play a major role in con-tributing to total NMHC concentrations isoprene makes con-siderable contribution to total reactivity in the suburbs

The probability distributions of NOx in the two urban cen-ters fairly resemble each other Strong local traffic emissionsof NO in the daytime maintain a high level of NOx and exertan inhibition on ozone production in urban Shanghai Dis-tinct double-peak diurnal variations of NOx have been ob-served on transport limited days as a result of heavy trafficduring rush hours In urban Tianjin where NOx and NMHCemissions are significant observational analysis shows thatozone formation is sensitive to both precursors At WuqingNOx concentrations are often below 25 ppbv although pol-luted episodes with concentration exceeding 50 ppbv couldalso occasionally occur Ozone production is completely re-stricted by available NOx With limited source emissions inJinshan NOx concentrations remain close to about 10 ppbvwithout displaying a diurnal variation The sensitivity ofozone photochemistry is found to be quite variable in thisarea

Finally it is found that ozone problem extends over alarger area in the Tianjin region than in the Shanghai re-gion based on both in-situ ozone measurements and OMINO2 observations We thus conclude that the ozone prob-lem in Shanghai is on an urban scale Stringent restrictionson local emissions would help reduce the occurrence of highozone concentrations By contrast ozone pollution in Tianjinis probably a regional problem Combined efforts to reduceozone precursor emissions on a regional scale are urgentlyneeded to bring the ozone problem under control Futureinvestigations involving more extensive observations wouldhelp to further our understanding of the ozone problem inthese two megacities as well as extend the application of cur-rent conclusions

AcknowledgementsThis research was funded by 973 Program(2011CB403402) the National Natural Science Foundation ofChina (NSFC) under Grant No 40875001 and the Basic ResearchFund of Chinese Academy of Meteorological Sciences (2008Z011)The National Special Science and Technology Program for Non-profit Industry Ministry of Environmental Protection of China(200909022) and Tianjin Fundamental Research Program (10JCY-BJC05800) also supported this work The OMI NO2 daily griddedproduct was produced from the OMI NO2 Level-2 StandardProduct by the Atmospheric Chemistry and Dynamics branchCode 6133 NASA Goddard Space Flight Center Greenbelt MDThe Principal Investigator is James F Gleason

Edited by T Petaja

References

Atkinson R Gas-phase tropospheric chemistry of or-ganic compounds a review Atmos Environ 24A 1ndash41doi101016jatmosenv200710068 1990

Atkinson R Atmospheric chemistry of VOCs and NOx AtmosEnviron 34 2063ndash2101doi101016S1352-2310(99)00460-42000

Cai C J Geng F H Tie X X Yu Q and An JL Characteristics and source apportionment of VOCs mea-sured in Shanghai China Atmos Environ 44 5005ndash5014doi101016jatmosenv201007059 2010

Chameides W L Fehsenfeld F Rodgers M O CardelinoC Martinez J Parrish D Lonneman W Lawson DR Rasmussen R A Zimmerman P Greenberg J Mid-dleton P and Wang T Ozone precursor relationships inthe ambient atmosphere J Geophys Res 97 6037ndash6055doi10102991JD03014 1992

Chameides W L Kasibhatla P S Yienger J and Levy IIH Growth of Continental-Scale Metro-Agro-Plexes regionalozone pollution and world food production Science 264 74ndash77doi101126science264515574 1994

Chan C K and Yao X H Air pollution in mega cities in ChinaAtmos Environ 42 1ndash42doi101016jatmosenv2007090032008

Chang C C Lo S J Lo J G and Wang J L Analysis ofmethyl tert-butyl ether in the atmosphere and implications as anexclusive indicator of automobile exhaust Atmos Environ 374747ndash4755doi101016jatmosenv200308017 2003

Cheung V T F and Wang T Observational study of ozone pollu-tion at a rural site in the Yangtze Delta of China Atmos Environ35 4947ndash4958doi101016S1352-2310(01)00351-X 2001

Finlayson-Pitts B J and Pitts Jr J N Atmospheric chemistryFundamentals and experimental techniques John Wiley andSons New York 1986

Frost G J Trainer M Allwine G Buhr M P Calvert JG Cantrell C A Fehsenfeld F C Goldan P D HerweheJ Huebler G Kuster W C Martin R McMillen R TMontzka S A Norton R B Parrish D D Ridley B AShetter R E Walega J G Watkins B A Westberg H Hand Williams E J Photochemical ozone production in the ruralsoutheastern United States during the 1990 Rural Oxidants in theSouthern Environment (ROSE) program J Geophys Res 10322491ndash22508doi10102998JD00881 1998

Gao J Wang T Ding A J and Liu C B Observational study ofozone and carbon monoxide at the summit of mount Tai (1534masl) in central-eastern China Atmos Environ 39 4779ndash4791doi101016jatmosenv200504030 2005

Geng F H Tie X X Xu J M Zhou G Q Peng L Gao WTang X and Zhao C S Characterizations of ozone NOx andVOCs measured in Shanghai China Atmos Environ 42 6873ndash6883doi101016jatmosenv200805045 2008

Guenther A B Monson R K and Fall R Isoprene and monoter-pene emission rate variability observations with eucalyptus andemission rate algorithm development J Geophys Res 9610799ndash10808doi10102991JD00960 1991

Guenther A B Zimmerman P R Harley P C Monson R Kand Fall R Isoprene and monoterpene emission rate variabilitymodel evaluations and sensitivity analyses J Geophys Res 9812609ndash12617doi10102993JD00527 1993



Haagen-Smit A J Chemistry and physiology of LosAngeles Smog Ind Eng Chem 44 1342ndash1346doi101021ie50510a045 1952

Haagen-Smit A J Bradley C E and Fox M M Ozone forma-tion in photochemical oxidation of organic substances Ind EngChem 45 2086ndash2089doi101021ie50525a044 1953

Han S Q Bian H Feng Y C Liu A X Li X J Zeng Fand Zhang X L Analysis of the Relationship between O3 NOand NO2 in Tianjin China Aerosol and Air Quality Research11 128ndash139doi104209aaqr2010070055 2011

Herring J A Jaffe D A Beine H J Madronich S and BlakeD R High-latitude springtime photochemistry Part II sensitiv-ity studies of ozone production J Atmos Chem 27 155ndash178doi101023A1005822017405 1997

Jacobson M Z Atmospheric pollution history science and regu-lation Cambridge University Press 2002

Jenkin M E and Clemitshaw K C Ozone and other secondaryphotochemical pollutants chemical processes governing theirformation in the planetary boundary layer Atmos Environ 342499ndash2527doi101016S1352-2310(99)00478-1 2000

Kleinman L I Ozone process insights from field experiments-part II Observation-based analysis for ozone production AtmosEnviron 34 2023ndash2033doi101016S1352-2310(99)00457-42000

Klimont Z Cofala J Schopp W Amann M Streets D GIchikawa Y and Fujita S Projections of SO2 NOx NH3 andVOC emissions in East Asia up to 2030 Water Air Soil Pollut130 193ndash198doi101023A1013886429786 2001

Lin W Xu X Zhang X and Tang J Contributions of pol-lutants from North China Plain to surface ozone at the Shang-dianzi GAW Station Atmos Chem Phys 8 5889ndash5898doi105194acp-8-5889-2008 2008

Lin W Xu X Ge B Z and Zhang X C Characteristics ofgaseous pollutants at Gucheng a rural site southwest of BeijingJ Geophys Res 114 D00G14doi1010292008JD0103392009

Lin W Xu X Ge B and Liu X Gaseous pollutants in Bei-jing urban area during the heating period 2007ndash2008 variabilitysources meteorological and chemical impacts Atmos ChemPhys 11 8157ndash8170doi105194acp-11-8157-2011 2011

Lu K D Zhang Y H Su H Brauers T Chou C C Hofzuma-haus A Liu S C Kita K Kondo Y Shao M Wahner AWang J L Wang X S and Zhu T Oxidant (O3+NO2) pro-duction processes and formation regimes in Beijing J GeophysRes 115 D07303doi1010292009JD012714 2010

Lu S H Bai Y H Zhang G S and Ma J Study on the charac-teristics of VOCs source profiles of vehicle exhaust and gasolineemission Acta Sci Natur Uni Pek 39 507ndash511 2003

Madronich S and Calvert J G Permutation reactions of organicperoxy radicals in the troposphere J Geophys Res 95 5697ndash5715doi101029JD095iD05p05697 1990

Madronich S and Flocke S The role of solar radiation in atmo-spheric chemistry in The Handbook of Environmental Chem-istry edited by Boule P Springer-Verlag Berlin Heidelberg 1ndash26 1999

Middleton P Stockwell W R and Carter W P L Aggrega-tion analysis of volatile organic compound emissions for regionalmodeling Atmos Environ 24A 1107ndash1133doi1010160960-1686(90)90077-Z 1990

NRC Rethinking the ozone problem in urban and regional air pol-lution Committee on Tropospheric Ozone National ResearchCouncil National Academy Press 1991

Ran L Zhao C S Geng F H Tie X X Tang X Peng LZhou G Q Yu Q Xu J M and Guenther A Ozone photo-chemical production in urban Shanghai China Analysis basedon ground level observations J Geophys Res 114 D15301doi1010292008JD010752 2009

Ran L Zhao C S Xu W Y Lu X Q Han M Lin W LYan P Xu X B Deng Z Z Ma N Liu P F Yu J LiangW D and Chen L L VOC reactivity and its effect on ozoneproduction during the HaChi summer campaign Atmos ChemPhys 11 4657ndash4667doi105194acp-11-4657-2011 2011

Seinfeld J H Urban air pollution state of the science Science243 745ndash752doi101126science2434892745 1989

Seinfeld J H and Pandis S N Atmospheric chemistry andphysics from air pollution to climate change Wiley Interscience1998

Shao M Tang X Y Zhang Y H and Li W JCity clusters in China air and surface water pollutionFront Ecol Environ 4 353ndash361doi1018901540-9295(2006)004[0353CCICAA]20CO2 2006

Song Y Shao M Liu Y Lu S H Kuster W Goldan P andXie S D Source apportionment of ambient volatile organiccompounds in Beijing Environ Sci Technol 41 4348ndash4353doi101021es0625982 2007

Streets D G and Waldhoff S T Present and future emissions ofair pollutants in China SO2 NOx and CO Atmos Environ 34363ndash374doi101016S1352-2310(99)00167-3 2000

Streets D G Fu J S Jang C J Hao J M He K BTang X Y Zhang Y H Wang Z F Li Z P Zhang QWang L T Wang B Y and Yu C Air quality during the2008 Beijing Olympic Games Atmos Environ 41 480ndash492doi101016jatmosenv200608046 2007

Tang W Y Zhao C S Geng F H Peng L Zhou G Q GaoW Xu J M and Tie X X Study of ozone ldquoweekend ef-fectrdquo in Shanghai Sci China Ser D-Earth Sci 51 1354ndash1360doi101007s11430-008-0088-2 2008

Tang X Wang Z F Zhu J Gbaguidi A E Wu Q Z Li jand Zhu T Sensitivity of ozone to precursor emissions in urbanBeijing with a Monte Carlo scheme Atmos Environ 44 3833ndash3842doi101016jatmosenv201006026 2010

Trainer M Parrish D D Goldan P D Roberts J and Fehsen-feld F C Review of observation-based analysis of the regionalfactors influencing ozone concentrations Atmos Environ 342045ndash2061doi101016S1352-2310(99)00459-8 2000

USEPA Compendium of Methods for the Determination of ToxicOrganic Compounds in Ambient Air Second Edition Com-pendium Method TO-14A Determination of Volatile OrganicCompounds (VOCs) in Ambient Air Using Specially PreparedCanisters With Subsequent Analysis by Gas ChromatographyUS Environmental Protection Agency 1999a

USEPA Compendium of Methods for the Determination of ToxicOrganic Compounds in Ambient Air Second Edition Com-pendium Method TO-15 Determination of Volatile OrganicCompounds (VOCs) in Air Collected in Specially-Prepared Can-isters and Analyzed by Gas ChromatographyMass Spectrometry(GCMS) US Environmental Protection Agency 1999b



USEPA Quality Assurance Handbook for Air Pollution Measure-ment Systems Volume II Ambient Air Quality Monitoring Pro-gram EPA-454B-08-003 2008

Wang T Nie W Gao J Xue L K Gao X M Wang X F QiuJ Poon C N Meinardi S Blake D Wang S L Ding A JChai F H Zhang Q Z and Wang W X Air quality duringthe 2008 Beijing Olympics secondary pollutants and regionalimpact Atmos Chem Phys 10 7603ndash7615doi105194acp-10-7603-2010 2010

Van der A R J Peters D H M U Eskes H Boersma K F VanRoozendael M De Smedt I and Kelder H M Detection ofthe trend and seasonal variation in tropospheric NO2 over ChinaJ Geophys Res 111 D12317doi1010292005JD0065942006

Van der A R J Eskes H J Boersma K F Van Noije T P CVan Roozendael M De Smedt I Peters D H M U and Mei-jer E W Trends seasonal variability and dominant NOx sourcederived from a ten year record of NO2 measured from space JGeophys Res 113 D04302doi1010292007JD009021 2008

Xu J Ma J Z Zhang X L Xu X B Xu X F Lin WL Wang Y Meng W and Ma Z Q Measurements ofozone and its precursors in Beijing during summertime im-pact of urban plumes on ozone pollution in downwind rural ar-eas Atmos Chem Phys 11 12241ndash12252doi105194acp-11-12241-2011 2011a

Xu W Y Zhao C S Ran L Deng Z Z Liu P F Ma NLin W L Xu X B Yan P He X Yu J Liang W D andChen L L Characteristics of pollutants and their correlation tometeorological conditions at a suburban site in the North ChinaPlain Atmos Chem Phys 11 4353ndash4369doi105194acp-11-4353-2011 2011b

Zhao C S Peng L Sun A D Qin Y Liu H L Li W Land Zhou X J Numerical modeling of tropospheric ozone overYangtze Delta region Acta Sci Circum 24 525ndash533 2004



Haagen-Smit 1952 Haagen-Smit et al 1953 Seinfeld andPandis 1998) Thus there is an urgent need to tackle theozone problem in these regions where precursor emissionssignificantly rise in order to avoid possible threats to the pub-lic and ecosystems that might be potentially raised by highconcentrations of ozone and its precursors (Chameides et al1994 Jacobson 2002)

Previous studies related to ozone issues in the NCP andYRD regions were carried out separately In the NCP re-gion most observations and analysis were performed at re-gional background sites or in the megacity of Beijing (egGao et al 2005 Lin et al 2008 2009 2011 Tang et al2010 Lu et al 2010 Xu et al 2011a) In preparation forthe 2008 Olympics the relocation of old industrial facilitiesaway from Beijing and other stringent emission restrictionswere implemented to mitigate local air pollution It turnedout that the control of secondary pollutants such as ozone wasless successful than most primary gases as a result of regionalpollution A considerable contribution was demonstrated tobe made by regional sources like those in Tianjin (Streets etal 2007 Wang et al 2010) However current research re-garding the ozone problem in Tianjin is rather limited (Hanet al 2011 Ran et al 2011 Xu et al 2011b) In the YRDregion the expansion of population and economics centersaround the megacity of Shanghai where ozone issues havebeen more elaborately addressed but mainly in the thrivingdowntown area (eg Cheung and Wang 2001 Zhao et al2004 Geng et al 2008 Tang et al 2008 Ran et al 2009)In an effort to understand ozone behaviour in responses toenhanced precursor emissions and to design feasible and ef-fective emission control strategies more work needs to bedone in these fast developing regions

In this paper observations regarding ozone and its precur-sors from an intensive field campaign in summer 2010 in ur-ban Tianjin and from routine monitoring in summer 2009 inboth urban and suburban Shanghai were comparatively ana-lyzed and discussed with the dataset from the 2009 Haze inChina (HaChi) summer campaign in suburban Tianjin (Ranet al 2011) This would hopefully cast more light on un-derstanding and handling current ozone problem in the twomegacities

2 Experiments and methodology

21 Sites

Observations of gaseous pollutants were conducted respec-tively at an urban and a suburban site in the megacities ofTianjin and Shanghai Sitting on the eastern edge of the pol-luted NCP Tianjin is adjacent to the Bohai Bay as illustratedin Fig 1a The Tieta site (3906prime N 11710prime E) where the2010 summer campaign took place is located in a thicklysettled urban district of Tianjin The largest petrochemicalcomplex in China has recently been built up in the Tianjin

Binhai New Area which is situated about 40 km southeastof urban Tianjin and extends nearly 150 km along the coastalline The suburban site Wuqing (3923prime N 11701prime E) is thelocation where the 2009 HaChi summer campaign was car-ried out (site description refers to Ran et al 2011) The siteXujiahui (3112prime N 12126prime E) is a routine monitoring sta-tion in the crowded metropolitan area of Shanghai (Fig 1b)which is overburdened with dense population high trafficloading and numerous tower buildings Near the coast andalmost 50 km south of the urban center is a petrochemical in-dustrial area wherein the site Jinshan (3045prime N 12121prime E)is situated The residential population there is much less thanin the urban center and the land cover is generally well veg-etated

To support related discussions the observational data ofsurface ozone during July and August 2009 at another threeNCP sites are also used The urban site in the megacity ofBeijing is located in the courtyard of China Meteorologi-cal Administration (CMA 3957prime N 11619prime E) which isbetween the busy 2nd and 3rd ring roads in the northwestarea of urban Beijing (Lin et al 2011) The site CMA isabout 110 km northwest of Tianjin (Fig 1a) Gucheng (GCH3908prime N 11540prime E) is a polluted rural site about 110 kmsouthwest of Beijing and 130 km west of Tianjin (Lin et al2009) Shangdianzi (SDZ 4039prime N 11707prime E) is a WMOregional atmosphere background station on the north rimof the NCP region showing footprints of polluted regionalbackground and clean natural background according to winddirection (Lin et al 2008)

22 Measurements and data processing

Continuous monitoring of ozone and NOx took place in JulyndashAugust 2009 at both sites of Shanghai and in suburban Tian-jin while observations were made during JulyndashAugust 2010in urban Tianjin The commercial Thermo EnvironmentalInstrument (TEI Inc USA) Model 49C a UV Photomet-ric Analyzer was used to measure ozone concentrations atboth sites of Tianjin CMA GCH and SDZ Surface ozonein Shanghai was observed using the UV Absorption OzoneAnalyzer (EC9810BECOTECH) based on the same opera-tion principle As for NOx measurements TEI 42CTL NO-NO2-NOx Analyzer at Wuqing simultaneously determinedthe ambient concentrations of NO and NOx utilizing chemi-luminescence techniques and a heated molybdenum NO2 toNO converter This instrument was able to accurately mea-sure the amount of nitrogen oxides to a trace level of lessthan 50 pptv TEI 42i used at Jinshan and the ECOTECHOxides of Nitrogen Analyzer (EC9841B) used in urban ar-eas of Tianjin and Shanghai were operated on the basis ofsimilar techniques as TEI 42CTL Besides carbon monox-ide (CO) was measured by TEI 48C CO Analyzer during thesame periods of ozone measurements at both sites of TianjinMultipoint calibrations and daily maintenance were carefullyperformed at all sites following USEPA recommendations on



5

55

5

10

10

10

10

15

15

15

1515

15

20

20

20

20

GCH

SDZ

CMA

WuqingTieta

( )a

113

E 114

E 115

E 116

E 117

E 118

E 119

E 120

E37

N

38

N

39

N

40

N

41

N

5

55

5

10

10

10

10

10

15

15

15

Xujiahui

Jinshan

( )b

116

E 117

E 118

E 119

E 120

E 121

E 122

E29

N

30

N

31

N

32

N

33

N

NO

21

5sdot

-2)

0

5

10

15

20

25

Bohai Bay








P(O3) = [NO](k[HO2] +

sumi

ki[RiO2]) (1)




0 50 100 150 2000

005

01

015

02

025

Ozone (ppbv)

Pro

babi

lity

Dis

trib

utio

n


0 50 100 150 2000

005

01

015

02

025

03

NOx (ppbv)

Pro

babi

lity

Dis

trib

utio

n


0 50 100 150 200

0

005

01

015

Ozone (ppbv)

Pro

babi

lity

Dis

trib

utio

n


0 1 2 3 4 50

002

004

006

CO (ppmv)

Pro

babi

lity

Dis

trib

utio

n

















ppbC ppbC ppbC ppbC


Total 3096 4387 1214 996









0 50 100 150 200 250 30001

02

03

04

05

06

07

08


Pro

pyminus

Equ

iv C

onc

pp

bC C

onc










000 300 600 900 1200 1500 1800 2100 24000

20

40

60

80

100

120

Ozo

ne (

ppbv

)

(a)


0

10

20

30

40

50

60

70

80

NO

x (p

pbv)


000 300 600 900 1200 1500 1800 2100 2400

TN

MH

Cs

10 (

ppbC

)

(b)












1 2 4 6 8 10 20 30 400

5

10

15

20

25

30

35

40

45

50

NOx (ppbv)

P(O

3) (p

pbv

h)



























20

40

60

80

100

Tieta

20

40

60

80

100

Wuqing

20

40

60

80

100

N

MH

C R

eact

ivity

Bas

ed F

ract

ions

Xujiahui

20

40

60

80

100

000 300 600 900 1200 1500 1800 2100 2400 Average

Jinshan





4 Conclusions










Edited by T Petaja

References


























































5

55

5

10

10

10

10

15

15

15

1515

15

20

20

20

20

GCH

SDZ

CMA

WuqingTieta

( )a

113

E 114

E 115

E 116

E 117

E 118

E 119

E 120

E37

N

38

N

39

N

40

N

41

N

5

55

5

10

10

10

10

10

15

15

15

Xujiahui

Jinshan

( )b

116

E 117

E 118

E 119

E 120

E 121

E 122

E29

N

30

N

31

N

32

N

33

N

NO

21

5sdot

-2)

0

5

10

15

20

25

Bohai Bay








P(O3) = [NO](k[HO2] +

sumi

ki[RiO2]) (1)




0 50 100 150 2000

005

01

015

02

025

Ozone (ppbv)

Pro

babi

lity

Dis

trib

utio

n


0 50 100 150 2000

005

01

015

02

025

03

NOx (ppbv)

Pro

babi

lity

Dis

trib

utio

n


0 50 100 150 200

0

005

01

015

Ozone (ppbv)

Pro

babi

lity

Dis

trib

utio

n


0 1 2 3 4 50

002

004

006

CO (ppmv)

Pro

babi

lity

Dis

trib

utio

n

















ppbC ppbC ppbC ppbC


Total 3096 4387 1214 996









0 50 100 150 200 250 30001

02

03

04

05

06

07

08


Pro

pyminus

Equ

iv C

onc

pp

bC C

onc










000 300 600 900 1200 1500 1800 2100 24000

20

40

60

80

100

120

Ozo

ne (

ppbv

)

(a)


0

10

20

30

40

50

60

70

80

NO

x (p

pbv)


000 300 600 900 1200 1500 1800 2100 2400

TN

MH

Cs

10 (

ppbC

)

(b)












1 2 4 6 8 10 20 30 400

5

10

15

20

25

30

35

40

45

50

NOx (ppbv)

P(O

3) (p

pbv

h)



























20

40

60

80

100

Tieta

20

40

60

80

100

Wuqing

20

40

60

80

100

N

MH

C R

eact

ivity

Bas

ed F

ract

ions

Xujiahui

20

40

60

80

100

000 300 600 900 1200 1500 1800 2100 2400 Average

Jinshan





4 Conclusions










Edited by T Petaja

References


























































0 50 100 150 2000

005

01

015

02

025

Ozone (ppbv)

Pro

babi

lity

Dis

trib

utio

n


0 50 100 150 2000

005

01

015

02

025

03

NOx (ppbv)

Pro

babi

lity

Dis

trib

utio

n


0 50 100 150 200

0

005

01

015

Ozone (ppbv)

Pro

babi

lity

Dis

trib

utio

n


0 1 2 3 4 50

002

004

006

CO (ppmv)

Pro

babi

lity

Dis

trib

utio

n

















ppbC ppbC ppbC ppbC


Total 3096 4387 1214 996









0 50 100 150 200 250 30001

02

03

04

05

06

07

08


Pro

pyminus

Equ

iv C

onc

pp

bC C

onc










000 300 600 900 1200 1500 1800 2100 24000

20

40

60

80

100

120

Ozo

ne (

ppbv

)

(a)


0

10

20

30

40

50

60

70

80

NO

x (p

pbv)


000 300 600 900 1200 1500 1800 2100 2400

TN

MH

Cs

10 (

ppbC

)

(b)












1 2 4 6 8 10 20 30 400

5

10

15

20

25

30

35

40

45

50

NOx (ppbv)

P(O

3) (p

pbv

h)



























20

40

60

80

100

Tieta

20

40

60

80

100

Wuqing

20

40

60

80

100

N

MH

C R

eact

ivity

Bas

ed F

ract

ions

Xujiahui

20

40

60

80

100

000 300 600 900 1200 1500 1800 2100 2400 Average

Jinshan





4 Conclusions










Edited by T Petaja

References





























































ppbC ppbC ppbC ppbC


Total 3096 4387 1214 996









0 50 100 150 200 250 30001

02

03

04

05

06

07

08


Pro

pyminus

Equ

iv C

onc

pp

bC C

onc










000 300 600 900 1200 1500 1800 2100 24000

20

40

60

80

100

120

Ozo

ne (

ppbv

)

(a)


0

10

20

30

40

50

60

70

80

NO

x (p

pbv)


000 300 600 900 1200 1500 1800 2100 2400

TN

MH

Cs

10 (

ppbC

)

(b)












1 2 4 6 8 10 20 30 400

5

10

15

20

25

30

35

40

45

50

NOx (ppbv)

P(O

3) (p

pbv

h)



























20

40

60

80

100

Tieta

20

40

60

80

100

Wuqing

20

40

60

80

100

N

MH

C R

eact

ivity

Bas

ed F

ract

ions

Xujiahui

20

40

60

80

100

000 300 600 900 1200 1500 1800 2100 2400 Average

Jinshan





4 Conclusions










Edited by T Petaja

References


























































0 50 100 150 200 250 30001

02

03

04

05

06

07

08


Pro

pyminus

Equ

iv C

onc

pp

bC C

onc










000 300 600 900 1200 1500 1800 2100 24000

20

40

60

80

100

120

Ozo

ne (

ppbv

)

(a)


0

10

20

30

40

50

60

70

80

NO

x (p

pbv)


000 300 600 900 1200 1500 1800 2100 2400

TN

MH

Cs

10 (

ppbC

)

(b)












1 2 4 6 8 10 20 30 400

5

10

15

20

25

30

35

40

45

50

NOx (ppbv)

P(O

3) (p

pbv

h)



























20

40

60

80

100

Tieta

20

40

60

80

100

Wuqing

20

40

60

80

100

N

MH

C R

eact

ivity

Bas

ed F

ract

ions

Xujiahui

20

40

60

80

100

000 300 600 900 1200 1500 1800 2100 2400 Average

Jinshan





4 Conclusions










Edited by T Petaja

References


























































000 300 600 900 1200 1500 1800 2100 24000

20

40

60

80

100

120

Ozo

ne (

ppbv

)

(a)


0

10

20

30

40

50

60

70

80

NO

x (p

pbv)


000 300 600 900 1200 1500 1800 2100 2400

TN

MH

Cs

10 (

ppbC

)

(b)












1 2 4 6 8 10 20 30 400

5

10

15

20

25

30

35

40

45

50

NOx (ppbv)

P(O

3) (p

pbv

h)



























20

40

60

80

100

Tieta

20

40

60

80

100

Wuqing

20

40

60

80

100

N

MH

C R

eact

ivity

Bas

ed F

ract

ions

Xujiahui

20

40

60

80

100

000 300 600 900 1200 1500 1800 2100 2400 Average

Jinshan





4 Conclusions










Edited by T Petaja

References


























































1 2 4 6 8 10 20 30 400

5

10

15

20

25

30

35

40

45

50

NOx (ppbv)

P(O

3) (p

pbv

h)



























20

40

60

80

100

Tieta

20

40

60

80

100

Wuqing

20

40

60

80

100

N

MH

C R

eact

ivity

Bas

ed F

ract

ions

Xujiahui

20

40

60

80

100

000 300 600 900 1200 1500 1800 2100 2400 Average

Jinshan





4 Conclusions










Edited by T Petaja

References









































































20

40

60

80

100

Tieta

20

40

60

80

100

Wuqing

20

40

60

80

100

N

MH

C R

eact

ivity

Bas

ed F

ract

ions

Xujiahui

20

40

60

80

100

000 300 600 900 1200 1500 1800 2100 2400 Average

Jinshan





4 Conclusions










Edited by T Petaja

References






























































Edited by T Petaja

References

























































ozone production in summer in the megacities of tianjin and

Documents