Characteristics of Ground Ozone Concentration over Beijing from 2004 to 1 2015 : Trends , Transport , and Effects of Reductions 2 3

Characteristics of Ground Ozone Concentration over Beijing from 2004 to 1 2015: Trends, Transport, and Effects of Reductions 2 3 Nianliang Cheng 1,2,3 , Yunting Li 1 ,Dawei Zhang 1,4* ,Tian Chen 5 , Feng Sun 1,6 ,Chen Chen 1,6 ,Fan Meng 2,3 4 Beijing Key Laboratory of Airborne Particulate Matter Monitoring Technology,Beijing Municipal Environmental 5 Monitoring Center,Beijing 100048, China 6 College of Water Sciences,Beijing Normal University,Beijing 100875,China 7 State Key Laboratory of Environmental Criteria and Risk Assessment, Chinese Research Academy of Environmental 8 Sciences, Beijing 100012, China 9 Department of Environmental Science and Engineering,Tsinghua University, Beijing 100084, China 10 Beijing Municipal Environmental Protection Bureau, Beijing 100044, China 11 Department of Environmental Science, Peking University, Beijing 100871, China 12 Corresponding author at:Beijing Municipal Environmental Monitoring Center,Beijing 100048, China.E-mail address: 13 zhangdawei@bjmemc.com.cn (W.D. Zhang). 14 Abstract: Based on the hourly ozone monitoring data during 2004–2015 in urban area and at 15


Introduction
Ground-level ozone, one of the most important secondary air pollutants in the atmosphere, is generated through photochemical reactions between nitrogen oxides(NO x ) and volatile organic compounds (VOCs) (Trainer et al., 2000;Sillman, 1999).High concentrations of ozone near the ground are harmful to human health, ecosystems, and global climate (Fiore et al., 2009).
In recent years, elevated regional ozone concentration and atmospheric oxidation capacity in China have attracted increasing attention (Lin et al., 2008;Zhang et al., 2007).Numerous studies have analyzed the concentration variations of ozone and its photochemical reactions with its precursors based on the measurements over a short period or satellite data (Chan et al., 2003;Wang et al., 2012;Vingarzan, 2004).Most studies in China were mainly concentrated in city cluster regions, such as Pearl River delta (Li et al., 2011;Wei et al., 2012;Zhang et al., 2013), Yangtze River delta (Li et al., 2014;Ding et al., 2013;Ran et al., 2009), and Beijing-Tianjin-Hebei regions(BTH) (Tang et al., 2009;Shao et al., 2009;Lu et al., 2010).These studies focused on the chemical characteristics of ozone, with few discussions on the variation trends and ozone transport and its influencing factors especially by regional reduction measures within a long period because of the lack of observed data in Beijing (An et al., 2006;Chou et al., 2009;Yuan et al., 2009) and other limiting facors.
Different to the continuously decreasing ground ozone concentrations in urban sites in the US (Pollack et al., 2013), recent limited studies performed in China, particularly in BTH area, suggested that ozone concentrations in both regional background and urban areas are increasing (Meng et al., 2009;Wang et al., 2008) due to large NO x emissions.Few long-term studies analyzed the trends of ground-level ozone in Beijing (Lu et al., 2010),let alone analyze the trends of daily maximum 8 h average ozone concentration(O 3 8h) and daily maximum 1 h ozone concentration(O 3 1h) and effects of urbanization and regional emission reduction measures on ozone concentrations.After the implementation of the new standard of "Ambient Air Quality Standard" (MEP, 2013) in 2013, the levels of O 3 1h and O 3 8h have a direct impact to the ranks of the air quality in Beijing.Furthermore, the increasing ozone pollution of Beijing obtained much public concerns from Beijing Municipal Government and the whole society (Ding et al., 2013;Wang et al., 2013) especially in Summer.The executive meeting of the State Council examined and adopted 'The Control Measures of Beijing Air Pollution during 2012-2020" (http://zhengwu.beijing.gov.cn/gzdt/gggs/t1225355.htm).According to the regulation, the non-attainment hours of ozone in Beijing will decrease by 30% than that in Atmos.Chem. Phys. Discuss., doi:10.5194/acp-2016-508, 2016 Manuscript under review for journal Atmos.Chem.Phys.Published: 10 August 2016 c Author(s) 2016.CC-BY 3.0 License.2010 and should be controlled at about 200 hours annually.Therefore, the results of previous studies were far from the current needs.
Air quality security programs were implemented from Aug 20 to Sep 3 in 2015 to guarantee the air quality for the parade on the 70th Victory Memorial Day for the Chinese People's War of Resistance against Japanese Aggression.Chinese government established numerous emission reduction measures, such as reducing coals, industrial adjustment, joint prevention measures, and limitation of vehicles (particularly heavy-duty buses and trucks from outside Beijing, and odd-even license plate policy on roads within urban Beijing).As regional emission reduction measures can not be copied and costs a lot of manpower and material resources,it offers a precious opportunity to study the changes in ozone and its precursors during the period of air quality assurance.This paper aims to investigate the temporal trends of O 3 1h and O 3 8h in different sites in Beijing and verify the importance of ozone transport.Also, we evaluated the changes on ozone concentration after the reduction measures during the Sep 3 military parade in 2015.

Site distribution
Beijing is located at 115.7°-117.4°E, 39.4°-41.6°N.This area is at the northwest edge of the North China Plain and close to the edge of the semi-desert zone.Its terrain exhibits a dustpan shape, and it is surrounded by mountains in three directions.The average altitude of Beijing is 43.5 m, and the general altitude of mountains is in the range of 1 000-1 500 m, which is not conducive to pollutant diffusion.The total area of Beijing is 16410.54km 2 , in which 62% are mountains.Its total forest coverage in the plain region is about 15%, which is lower than that in whole city (38%).Beijing exhibits a temperate continental monsoon climate, where it is hot and rainy in summer and cold and dry in winter.Over the past decade, the annual average rainfall is less than 450 mm, 80% of which is concentrated in June, July, and August (BJEPB 2014;Beijing Statistics Bureau, 2014).
As the capital of China, the air quality monitoring network in Beijing is more advanced than that of the remaining regions of China (BJEPB, 2014).In 2001, an air quality monitoring network that obtains 35 monitoring stations was established by the Beijing Municipal Environmental Monitoring Center (BJMEMC, http://zx.bjmemc.com.cn/,Fig. 1).The 35 monitoring stations cover all districts that contain different environment types defined by regional background, such as suburbs, city, and residential.Twelve monitoring sites (DL, DS, GY, TT, WSXG, AT, NZG, WL, GC, SY, CP, HR) in urban area and one background station

Monitoring instruments
The monitoring instruments of ozone are all the 49C ozone analyzer instruments produced by Thermo Fisher Corporation (USA).The minimum limit of ozone analyzer instrument is 1×10 −9 , and the zero and cross drifts are 0.4%/24 h and ±1% /24 h, respectively.
An ozone calibrator (49IPS) traceable to the Standard Reference Photometer maintained by the WMO World Calibration Center was used to calibrate the ozone analyzers.Ozone monitoring instrument at each station had a zero cross calibration every three days, precision audit every three month, and an accuracy check every six months to ensure the monitoring quality of ozone in Beijing.Thermo Fisher 42C NO-NO 2 -NO x analyzer was used to monitor NO and NO 2 concentrations with a limit of 0.05×10 −9 , zero drift of 0.025×10 −9 /24 h, and span drift of ±1%/24h.Operation procedure strictly followed the "The Specification of Environmental Air Quality Automatic Monitoring Technology" (HJ/T193-2005, was in an evident upward trend with an annual concentration growth rate (AAGR) of 1.79 ppbv• yr −1 (correlation coefficient R=0.82,highly correlated) and an higher increase of 2.84 ppbv• yr −1 during May to September(MSAGR,R=0.87,highly correlated) from 2004 to 2015.Variation of O 3 8h was in an overall upward trend with AAGR of 1.14 ppbv• yr −1 (R=0.88,highlycorrelated) and MSAGR of 1.68 ppbv• yr −1 (R=0.85,highlycorrelated) during May to September, respectively from 2004 to 2015.For the variations of ozone concentration at DL background station, O 3 1h was in an overall upward trend with AAGR of 2.05 ppbv• yr −1 (R=0.81,highly correlated) and MSAGR of 0.14 ppbv• yr −1 (R=0.10,micro relevant), whereas O 3 8h was in a slightly downward trend (AAGR=−0.47ppbv• yr −1 , R=−0.42,weak correlation,real relevant; MSAGR=−0.70ppbv• yr −1 , R=−0.40, real relevant).
(r is the pearson correlation coefficient; X is the ozone concentration;Y presents the year;r between 0 and 0.3 representing the micro relevant, r between 0.3 and 0.5 representing the real relevant, r between 0.5 and 0.8 representing the significant relevant,and r between 0.8 and 1.0 representing highly relevant) For the variations of O  , 2006), or high concentrations of the regional zone and its precursors transport (Parrish et al., 2014).Also,the rapid growth of population and industrialization have driven substantial increases in ozone background concentrations in BTH area (Willem et al., 2015).that in urban sites.For the spatial distribution of ozone,it was lower in central urban area and relatively higher in the northern and western area with good vegetation.Ozone has a lifespan of several days; consequently, high ozone concentrations can be found in regions distant from precursor emission sources (Seinfeld 2004;Kalabokas et al., 2000), and several chemical ozone destruction reactions existing in urban center, such as R2-R3, are absent in background areas (Saitanis 2003;Pablo et al., 2013).In Beijing, NO x concentration in the urban center of the city was typically higher because of the large amount of vehicle population, which consumed and titrated a certain amount of ozone.The ozone peaks at DL station from May to September in different years was obviously 1-h behind than that of urban sites, which was closely related to the regional ozone transport (R4).Most of the ozone was generated during the transport of its precursors from emission districts to surroundings or background sites.In summer, high temperature, strong solar radiation, low humidity, and small southwest wind in Beijing strengthen photochemical pollution; moreover, ozone and its precursors, such as NO x , CO, and VOCs, are transported to the downwind area, hence the reduced ozone peak concentration in downwind area (Carnero et al., 2010;Shan et al., 2010).

Emission reductions on ozone concentrations
Comparing the ozone peaks in urban Beijing during S2 and S3 stages between 2004Beijing during S2 and S3 stages between -2014Beijing during S2 and S3 stages between and 2015(the year of taking emission reduction measures to ensure the regional air quality), the ozone peak in 2015 was 3h earlier and 0.  (Seinfeld ,2006) .
0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 0 0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 0 As we all know,NO x and VOCs emission control can considerably affect the O 3 concentration, but ozone generation is not a simple linear relationship with its precursors (Sillman, 1999).Ozone pollution is mainly concentrated in summer, and biogenic emissions accounted for a majority of the total VOCs.Therefore, the emission reduction of VOCs via anthropogenic measures connot make it higher than that of NO x .To ensure the air quality during the military parade(S2 and S3 stages) in 2015, NO x and VOCs emission control in Beijing and its surrounding areas lasted for almost a month, and VOCs emission control measures was much stricter than NO x (MEP, 2015); thereby, ensuring the reduction of VOCs emission(45%) is higher than that of NO x (30%).While for the temporal distribution of ozone during APEC meeting air quality assurance period, regional VOCs emission(about 30%) was equal to that of NO x ( about 30%)(MEP, 2015) and it was easily affected by the relatively unfavorable diffusion conditions in Autumn in Beijing which lead to the concentrations of NO x and CO two times larger than that of Sep 3 military parade period.So different emission reduction ratio between NO x and VOCs and different weather conditions led to different VOC(ppbvC)/NOx(ppbv)ratios during Sep 3 military parade period and APEC meeting.If the NO x levels are so high that it is not consumed before the end of the day,then ozone is VOCs sensitive, and decreasing NO x would cause increased ozone formation during APEC meeting.This phenomenon of concentrations of most of the air pollutants decreased, whereas concentrations of ozone increased during APEC meeting period which was consistent with the study of Wang (Wang et al,2015b).Whereas,the higher VOCs emission reduction caused the slight decrease in urban area but significant decrease at downwind DL background station during Sep 3 military parade period.Above all,success of air quality protection during the Sep 3 military parade proved that the current governance policy is correct and far-sighted.Moreover, ozone pollution is typically a regional rather than a local issue.Thus, in the future, clean air action plan in Beijing should be implemented on the basis of the lessons from regional air pollution prevention and control mechanism to promote the continuous improvement of regional air quality unswervingly and jointly.Combined with multiple observation stations of ozone in a 4.2 Diurnal variation of ozone peaks obtained at the downwind DL station were about 1 h later than that of the urban area from May to October in different years and concentration of ozone at downwind background station was much higher than that of urban sites.Moreover, difference of ozone peaks between urban sites and DL background station was significantly becoming smaller in recent years, which may be related to regional ozone transport and the expansion urbanization of Beijing.
4.3 Based on the joint efforts of regional air pollution prevention and control,Beijing achieved Sep 3 military blue.Average concentrations of CO, NO 2 , and O 3 in S2 (Aug 20~31,2015) and S3 (Sep 01~03,2015) decreased by 31.48%,43.97%, and 13.21% in urban sites, and by 20.93%, 57.10%, and 23.62% at DL site, respectively compared with those in S1(Aug 01~19,2015) and S4 (Sep 04~30,2015).A reduction of local anthropogenic emissions such as VOCs and NO x could reduce ozone efficiently especially in downwind areas of Beijing and made the ozone peaks decrease significantly and appear 2~3h earlier compared to the scenarios of no emission reductions.Compared to the increasing ozone during APEC period,to decrease the ozone concentration in Beijing, emissions of VOCs should be reduced larger than that of NO x in Beijing and the policy of regional air pollution joint prevention and control should still be promoted unswervingly and jointly in the futher.

Fig. 1
Fig. 1 Distribution and classification of observation sites in Beijing Atmos.Chem.Phys.Discuss., doi:10.5194/acp-2016-508,2016 Manuscript under review for journal Atmos.Chem.Phys.Published: 10 August 2016 c Author(s) 2016.CC-BY 3.0 License.http://kjs.mep.gov.cn/hjbhbz/bzwb/dqhjbh/jcgfffbz/200601/t20060101_71675.htm), and the equipment was regularly calibrated and maintained by technicians.regression and statistical tests, such as Pearson's correlation analysis were implemented to investigate the trends of O 3 1h and O 3 8h in urban area and at DL station in Beijing (Fig. 2), for the interannual variations of ozone concentrations in urban Beijing, O 3 1h Fig. 2 Variation trends of O 3 1h and O 3 8h in urban Beijing (a1, a2) and at DL background station (b1, b2) from 2004 to 2015 (including annual trends and trends from May to Sep) and the linear fitting equations(95% confidence interval, twotailed ,**highly relevant)

Fig. 3 Fig. 4
Fig. 3 Average concentrations of O 3 1h and O 3 8h at urban sites and DL background site from May to Sep during the periodsof 2004of  -2007of  , 2008of  -2012of  , and 2013of  -2015 in Beijing    in Beijing

Fig. 5
Fig. 5 was the diurnal variation of ozone in urban Beijing and at DL station from May to September during 2004-2015 and Table2presented the statistics of ozone peaks at DL station and urban sites from May to September during 2004-2015.In general,ozone concentration at DL station was higher than that of urban sites, and peaks of ozone concentration at DL station from May to September in different years was 1.01-1.56times significantly decreased from 18.20 ppbv to 2.72 ppbv during 2004-2010 and 2011-2015.This change may be related to the expansion of urbanization of Beijing.With city expansion and economic development, the district near DL stationwas urbanized and easily influenced by anthropogenic emissions.Santini et al.(2010) found Beijing urban extent estimated from Landsat data was from 1105km 2 to 4139km 2 between 2000 and 2009.Jacobson et al. (2015)pointed that urbanization decreases the concentrations of many surface chemicals due to their vertical dilution but increases near-surface ozone.

Fig. 5
Fig. 5 Diurnal variations of ozone in urban areas and at DL station from May to September during 2004-2015 in Beijing

Fig. 6
Fig. 6 Concentrations of main air pollutants in urban Beijing from August 01to September 30 in 2015.303 304 305 91 ppbv lower compared to the average ozone peaks during the period of 2004-2014.Whereas comparing the diurnal variations of ozone at DL site during S2 and S3 stages between 2004-2014 and 2015, the ozone peak in 2015 was 10.98 ppbv lower and 2 h earlier compared to that during the period of 2004-2014.The ozone peaks between urban sites and DL site were much closer (only about 0.48ppbv) after the a Atmos.Chem.Phys.Discuss., doi:10.5194/acp-2016-508,2016 Manuscript under review for journal Atmos.Chem.Phys.Published: 10 August 2016 c Author(s) 2016.CC-BY 3.0 License.reduction of anthropogenic emissions in 2015.The earlier ozone peaks indicated the approximate photochemical equilibrium(R5-R7) of O 3 ,NO and NO 2 was moved up in Beijingduring the day due to regional emission reductions.Therefore,a reduction of anthropogenic emissions such as VOCs and NO x in urban areas made the ozone peaks decrease significantly and appear 2~3h earlier compared to the scenarios of no emission reductions which was a very interesting phenomenon and first found in Beijing and it could also reduce ozone concentration efficiently especially at background sites or downwind areas by the weakened regional transport which was coincident with the study of Seinfeld

Fig. 7
Fig. 7 Diurnal variations of ozone in urban Beijing (a) and at DL station (b) during S2 and S3 stages from 2004 to 2015 compared to that of APEC meeting air quality assurance period(Nov 1~Nov 11,2014) in urban Beijing (c) from 2004 to 2015 Atmos.Chem.Phys.Discuss., doi:10.5194/acp-2016-508,2016 Manuscript under review for journal Atmos.Chem.Phys.Published: 10 August 2016 c Author(s) 2016.CC-BY 3.0 License.long period, the numerical models should also be combined to further analyze the ozone formation, so as to develop effective ozone pollution control measures.4 Conclusions 4.1 Annual concentration of daily maximum 1 h ozone (O 3 1h) was all increasing at urban sites(1.79ppbv• yr −1 ) and DL background station(2.05ppbv• yr −1 ) while daily maximum 8 h average ozone concentration(O 3 8h) was increasing in urban area(1.14ppbv• yr −1 ) and slightly decreasing at DL background station(-0.47 ppbv• yr −1 ) from 2004 to 2015 due to different ozone sensitivity regimes and ratios of NO 2 /NO.

Table 1
Statistics of NO 2 /NO in urban areas and at DL site during 2008-2015 in Beijing

Table 2
Peak ozone concentration from May to September during 2004-2015 at DL station compared with the urban city of

Table 3
Average concentrations of main air pollutants at urban sites at four stages in Beijing in 2015 Atmos.Chem.Phys.Discuss., doi:10.5194/acp-2016-508,2016 Manuscript under review for journal Atmos.Chem.Phys.Published: 10 August 2016 c Author(s) 2016.CC-BY 3.0 License.In order to eliminate the influence of meteorological factors, we counted the observed variations of meteorological elements from Aug 20 to Sep 05 at GXT station in Beijing between 2010 and 2015 (We only collected meteorological data for the past five years).From

Table 4 ,
the temperature and wind speed at ground which can affect ozone concentration directly changed slightly during the study periods.Average temperature was fluctuating between 23.6~24.3℃and it was was lowest in 2011 and highest in 2013.Average of wind speeds was fluctuating between 1.4~1.9m• s -1 ,suggesting the atmosphere was generally stable in August in Beijing.Average relative humidity and suface pressure was also possessing the same characteristics.Frequency of the north wind at 850hPa directly affects vertical diffusion of ozone.Average frequency of the north wind at 850hPa was between 12.1% and 41.2% and it changed to 24.3% from Aug 20 to Sep 05,2015 indicating the atmosphere was also relatively stable in the vertical direction It is reported that both sunshine hours and visibility in BTH region have been decreasing in the past decades (Yang Table4.Average meteorological elements from Aug 20 to Sep 05 at GXT station in Beijing between 2010 and 2015.