Environmental Process and Convergence Belt of Atmospheric NO2Pollution in North China

VOL.25 NO. 6ACTA METEOROLOGICA SINICA2011Environmental Process and Convergence Belt of Atmospheric NO2Pollution in North China*WEI Peng' (尉鹏), REN Zhenhai2t(任阵海), sU Fuqing?(苏福庆), CHENG Shuiyuan'(程水源),ZHANG Peng3(张鹏), and GAO Qingxian2(高庆先)1 College of Environmental & Energy Engineering, Beijing University of Technology, Beijing 1000222 Chinese Research Acaderny of Environmental Sciences, Beijing 1000123 National Satelite Meteorological Center, China Meteorological Administration, Beijing 100081(Received October 16, 2010; in final form October 8, 2011)ABSTRACTBoth surface environmental monitoring and satellite remote sensing show that North China is one ofthe regions that are heavily polluted by NOz. Using the NO2 monitoring data from 18 major cities inthe region, the tropospheric NO2 column density data from the Ozone Monitoring Instrument (OMI) onthe Aura satellite, and the observations from the China Meteorological Administration network, this paperanalyzes a regional NO2 pollution event in February 2007 over North China, examines the convergence of thepollutant, and identifies its correlation with the atmospheric background conditions. The results show thatdaily mean NO2 concentrations derived from surface observations are consistent with the mean values of theOMI measurements, with their correlation coefficient reaching 0.81. The correlations of NO2 concentrationwith general weather patterns and sequential changes of temperature structure from 925 hPa down to thesurface indicate that the weather fronts, high pressure and low pressure systems in the atmosphere playrole in changing the temporal and spatial evolutions of NO2 through removing, accumulating or convergingof the pollutant, respectively. It is also found that the eastern Taihang Mountains is most heavily pollutedby NO2 in North China. Based on a model that correlates NOz column density with surface wind vector,the relation of the NO2 concentrations in six major cities in North China to the surrounding wind feldis analyzed. The results show that the maximum wind field is associated with the highest frequency opollution events, and under certain large scale atmospheric conditions together with the topographic effectsmall- and meso-scale wind fields often act to transport and converge pollutants, and become a major factorin forming the heaviest NO2 pollution event in North China. Analysis of the causes for the severe NOpollution event in this study may shed light on understanding, forecasting, and mitigating occurrences ofheavy NO2 pollution.Key words: atmospheric pollution, environmental process, convergence belt, satellite remote sensingCitation: Wei Peng, Ren Zhenhai, Su Fuqing, et al., 2011: Environmental process and convergence belt ofatmospheric NO2 pollution in North China. Acta Meteor. Sinica, 25(6), 797-811, doi: 10.1007/s13351-011-0610-x.1. IntroductionPrevious studies show that during 1996- 2006, adecreasing trend in NO2 concentration appeared glob-Nitrogen dioxide (NO2) is a major atmospherically, especially along the eastern coast of the Unitedpollutant and a main precursor of acid rain and photo- States (Kim et al, 2006) and Europe (Richter et al,chemical smog, which has serious effects on atmo-2005). In contrast, as a result of the rapid economicspheric environment quality. High NO2 concentration development in recent years, China now becomes amay lead to a higher mortality of lung cancer, 80 it isregion with the highest NO2 concentration and thedetrimental to human health. Also as a precursor offastestNO2 increase rate; especially, the easternozone, higher NO2 concentration often leads to peak-China has witnessed a rapid NO2 increase sincevalued ozone concentration.2000 (Sachin et al., 2009). Therefore, studies on the●Supported by the Fundamental Rescarch Fund of the Chinese Research Academy of Environmental Sciences (2010KYYW08).↑Corresponding author: weipeng - 1981 @hotmail.com.CThe Chinese Meteorological Society and Springer- Verlag Berlin Heidelberg 2011中国煤化工MYHCNMHG798ACTA METEOROLOGICA SINICAVOL.25 .characteristics and mechanisms of severe NO2 pollu-also analyzed the pollution sources. They suggestedtion events are of particular interest and urgency to that human activities may affect the spatial distribu-the environmental science community.tion of NO2 concentration (Zhang x. Y., et al, 2007).The concentration changes of pollutants in a reGOME and SCIAMACHY data, andgional air pllution process are subject to impacts of with an atmospheric chemistry transport model, Yueatmospheric conditions, as the removal, accumulation,t al. (2009) investigated the seasonal NO2 variationsand convergence of air pollutants are closely related and drew simnilar conclusions. Wang et al. (2009) con-to the changes in weather patterns (Ren et al, 2005;firmed the impacts of human activities on NO2 columnDing, 2005). A convergence belt usually representsa density in the troposphere in several geographic zoneszone with the highest concentration of a given pollu-of China.tant within a certain region, and its genesis and re-In the next five years (2011 -2015), the China en-moval with time and its spatial variation determine vironmental protection authority will make greater ef-the occurrence probability of severe pollution eventsforts to control the NO2 pollution. But so far, neitherin China (Su et al., 2004a).comprehensive investigations on the spatial distribu-In recent years, with advances in space-basedtion and temporal evolution of pollutants in Chinaremote sensing technologies, the sensorsnor studies on the seasonal NO2 variations have beenMACHY (Scanning Imaging Absorption Spectromecarried out at home and abroad. Using high resolu-ter for Atmosplheric CHartographY) and OMI (Ozonetion and large coverage satellite remote sensing ob-Monitoring Instrument)， among others, have beenservations, this paper focuses on the pending issuesused to obtain the distributions of large-scale NO2 col-about air pollution control in China and investigatesumn density in a continuously updated manner suchenvironmental processes and convergence belts of reas in the GOME (Global Ozone Monitoring Experi-gional NO2 pollution in China based on the previouslyment) and GOME-2. Relative to traditional methods,proposed concepts of atmospheric environmental pro-satellite-based remote sensing provides higher resolu-cesses (Ren et al., 2005) and pollution convergencetion, wider coverage and simultaneous measurements,belts (Su et al., 2004b). The study aims to provideand makes reliable observations available for the at- useful scientific reference for the control of air pollu-mospheric environmental studies. Therefore, scientiststion and planning of pollution reduction in China.worldwide have carried out a range of investigations onUsing the ground NO2 measurements in ChinaNO2 pollution based on stelite data. Some comparedand the remote sensing products of global troposphericNO2 column density observations from space with sur-NO2 column density from the US National Aeronau-face NO2 concentration measurements (Petritoli et al，tics and Space Administration (NASA), and select-2004; Ordonez et al., 2006; Lamsal et al, 2008). Us-ing NO2 concentration evolution sequence in Febru-ing GOME and SCIAMACHY data, Richter et alary 2007 in North China, this study ventures to in-(2005) analyzed the global distribution of monthlyvestigate the distribution characteristics and causes ofmean values of NO2 column density in the troposphereatmospheric environment processes and convergencein 1996 2004. They suggested a significant NO2 in belts of NO2 pollution, and ilustrates the seasonalcrease in eastern China. Using GOME data in a threevariation and infuential factors of NO2 in China. Thedimensional model, Van der et al. (2006) also pre study combines ground NO2 concentration measuresented a global picture of the total NO2 distribution ments and sea level pressure field data with a correla-in the troposphere. Focusing on NO2 pollution events tion vector model.in China, many similar studies were carried out byChinese scientists. Using SCIAMACHY data, Zhang2. DataX. Y., et al. (2007) showed both temporal and spatialNO2 distributions over China in 1997 -2006, and theyThe remote sensirg-dota " inthis_ naper are .中国煤化工fYHCNMH G .NO.6WEI Peng, REN Zhenhai, su Fuqing, et al.801vironmental process of atmospheric pollution, namely, tration in all these cities varied from 0.02 mg m-3a vlley-peak sequence showing concentration varia- to 0.10 mg m-3, mostly below the atmnospheric en-tions. The pollutants may be renoved by a frontal vironmental criteria II level (0.08 mg m-3). Afterzone, get accumulated by a stable high pressure syn- small fuctuations and accumulation in a few days,optic weather system, or get trapped and converged in on 5 February, the heaviest pollution occurred in Bei-a certain region by a low pressure system, exhibiting jing, Tianjin, Tangshan, Zhengzhou, and Taiyuan.distinctive evolution features associated with weather The daily average concentration in Beijing was 0.144and climate.mg m-3, exceeding the Criteria II level (0.12 mgBased on stllite observations and analysis of m-3). The peak values in Tangshan, Zhengzhou, andweather patternrs in February 2007, surface NO2 con- Pingdingshan were 0.075, 0.072, and 0.073 mg m-3, recentration measurements will be examined in this sec-spectively. On 6 February, all the other cities reachedtion. Its regional distribution and temporal variation their respective peak values. The daily average NO2characteristics during the severe pollution-prone win- concentration in Tianjin and Xuzhou both registeredter season over North China will be discussed.0.106 mg m-3. On 7 February, NO2 concentration4.1 Surface NO2 measurements in northernin all the cities began to fuctuate and decline. On9 February, a moderate fuctuation of NO2 concen-Chinatration took place once again till the day when NO2Figure 4 shows variations of surface NO2 con-concentration was falling down to the valley, endingcentration in 18 cities of northern China in Februarywith all cities below the Criteria II level. This first2007, based on the surface NOz measurements from process in February lasted for 10 days, during whichthe urban atmospheric environmental monitoring net-NO2 concentration in all the cities was in the range ofwork. It is dscernible that during 1-12 February, NO2 67%- 92% relative to the Criteria II level.concentration in all the cities generally underwent aDuring the first process in 4- 5 February, thevalley- peak-valley process, or a regional atmospheric single-day NO2 concentration over Tianjin rose by 0.04environmental process. On 1 February, NO2 concen- mg m~ -3, an increase of 66%. The NO2 concentration0.160.14 .一- Baoding- o- Beijing).12 t--Dalian-✧Datong .? 0.10十Handan中◆ JinanJining0.08Jiaozuo, Kaifeng0.0- Luoyang. Pingdingshan0.04. Qinhuangdao- Qingdao0.02 .- Rizhao- Shenyang- Shijiazhuang0.002/112/162/21Observation dateFig.4. Variations of NO2 concentration in 18 cities in northern (中国煤化工YHCNMH G .802ACTA METEOROLOGICA SINICAVOL.250.07over Beijing on 4 February was 0.076 mg m^ -3 and went3.▲OMI observed◆Surface observed0.06up to 0.144 mg m 3 next day (5 February), an increase2.0.05置of 94%, went down to 0.065 mg m" -3 on 6 Februaryby a decrease of 55%. Therefore, this environmental曾1.5-.04昌process can be characterized as: in 1-2 days, NO2 con-0.03centration in all the cities rose rapidly and reached thee 1.00.02peaks; subsequently, each city experienced 4 environ-0.50.01会mental processes in 11-14, 14-18, 18 23, and 23 -27Jo.002/12/16 2/21 2/26February, respectively.Observation date4.2 Satellite obseruationsFigure 5 gives average values of NO2 column den-Fig. 5. The OMI observed average NO2 column densitysity in northern China derived from OMI data inand the surface observed average daily NOz concentrationFebruary 2007 and a variation sequence of averagein 18 cities of northern China in February 2007.daily NO2 concentration measurements in 18 cities ofy-=0.010x+0.026northern China from the urban atmospheric environ-r'=0.667mental monitoring network. Because the OMI obser-品曾g 0.vation is performed around 1330 LT each day, the timeto calculate the daily surface NO2 concentration mea-surements starts from 1200 LT of the previous day. 0.0 0.51.0 1.52.0 2.5 3.0to 1200 of the current day. Therefore, the OMI curveOMI- -based average NO2 column densitydifers from the surface observed curve of NO2 concen-(10"* mole cm了tration by one day sometimes. As shown inFig. 5, on1, 10, 14, and 24 February, the average NO2 concentra Fig. 6. Correlation between OMl-based average NO2 col-tion from the monitoring network in China (32° 43°N，umn density and ground measured average daily NO2 con-110- 122.5E) reached the lowest level. The aver- centration over the 18 cities of northern China in Februaryage concentrations were 0.045, 0.035, 0.029, and 0.031 2007.mg m~3, respectively. The tropospheric NO2 column4.3 Association between environmental pro-densities were 0.97x1016, 0.64x1016, 0.60x 1016, andcesses of NO2 pollstion and weather pat-0.12x 1016 molec cm- -2, respectively. From the mea-terns in North Chinasurements made in the 18 cities in northern China,the regional average NO2 concentration peaked on 6, .13, 17, 21, and 27 February, repectively The aver- tin is subjet to the impacts of large scale weatherage concentratioas were 0.059, 0.049, 0.037, and 0.036 patterns (Chen et al, 2008; Su et al, 2004a; Kangmg m" -3. The corresponding NO2 column densitieset al, 2009a, b). When the concentration rises andwere 2.50x1016, 2.12x1016, 1.63x1016 , and 1.46x1016 reaches a peak, the corresponding weather patternmolec cm-2, rspectively. It is quite clear that the av- may be a stable continental high pressusre system thaterage tropospheric NO2 column densities from OMI has lasted for several days. When the pllutants dropare in general consistent with the trend of average and reach 8 vlley, this may orrespond to a front.daily NO2 concentration variations in these cities fromFrom Fig. 7, it is seen that the lowest values werethe ground measurements. Figure 6 shows a good con-found in the atmospberic environmental process on 10sisteney between the two with a crrelation cofcient and 13 February 2007. The corresponding SLP feldsof 0.81. Therefore, OMI observations can also reflectand NO2 column density distributions on 10 and 13the temporal variation trend in the proces of a severe February are given in Figs. 7a and 7b. As shown inNO2 pollution event.Fig.7b, when中国煤化Irth ChinaYHCNMHGNO.6WeI Peng, ren Zhenhai, sU Fuqing, et al.807homogenous spatial distribution of pollution sources characteristics in North China, this paper uses &or NO2 column density while the curve denotesthe real distribution, with the Gini cofficients be- distribution of NO2 concentrations with surface winding 0.36 and 0.67， respectively， indicating thatfields (Xu et al., 2005). Figure 12 shows the sur-he spatial distribution of pollutants and pollutionface vector winds from the Meteorological Informa-sources are not consistent, and both the transport tion Comprehensive Analysis and Process System (MI-and convergence of pollutants in the atmosphere CAPS). Figure 13 shows the vector-correlation fieldare important for the spatial distribution of the derived from OMI-based tropospheric NO2 columnpollutants.density data and the surface winds, in which vector5.2 Distribution of NO2 concentration and thelength represents the quantity of the correlation coef-ficient, and vector direction represents the pollutant8urface wind fieldconvergence channels. The equations are as followsTo demonstrate the regional NO2 convergence (Xu et al., 2004, 2005):R=一OP.(Xo, Y%)Oun(X,Y) - NOP.(Xo, Yo)~ Aun(X_______(1)(OP(Xo, Y))2 - N((Xo, Yo)~)2 2 (Qun(Xo, Y))2)- N(um(X,Y)N)2)△Pn(Xo, Y)Qun(X,Y) - NAPn(Xo, Yo)N Oun(X;, Y})NR,=i(2)1 E(Op.n(Xo, Y))2 - N((Xo, Y%)~)2 ((Ovn(Xo, Y))2)- N(x(.,)M)SPn(Xo, Yo)! Pn(Xo, Yo)一P'(Xo, Yo)NOun(Xo, Y)un(Xo, Yo)一u(Xo, Yo)NOvn(Xo,Yo),vn(Xo, Yo) - %(Xo,Y)N )where (Xo,Yo) is the correlation coefficient between u, 1 components of the wind fled.the evolution of a single point pollution process andFrom Eqs. (1) and (2), the following expressionthe u, U components of a wind field around the obser- can be obtained:vation point. Pn(Xo, Y) is the average concentrationvalues at (Xo, Yo), and un(Xo, Yo) and vn(Xo, Y) arer= Rni + Rurj.(3)1._0.9+ (a0.(t虽o.0.7& 0.40.30.2直0.20.0.O.0.6.0).6Cumulative percentage of BeraCumulative percentage of acraFig.11. Lorenz curves of NOz pollution sources (a) and OMI observations of column density area percentage (b).Pollution source: 2006; NOz column denaity: February 2007.中国煤化工MYHCNMH G .810ACTA METEOROLOGICA SINICAVOL.25that the pollutants emitted from Jinzhong basin con-tions of small- and meso scale NO2 concentrations canverge in the Taiyuan area under the impacts of the be captured by using the OMI-based observation data.basin convergence, leading to a higher NO2 concen-North China is the most heavily NO2 polluted re-tration in Taiyuan than in its surrounding areas (note: gion in terms of severity and extent across the country.Taiyuan is located in northern Jinzhong basin). FigChina features evident seasonal NO2 patterns, i.e., theure 14d shows that there exists a convergence zone heaviest pollution in winter (January) and the lightestbetween the southeasterly wind from the North China in summer (August). Seasonal diference of air pres-Plain and the local wind of the Taihang Mountains sure fields and that of precipitation distributions areunder the influence of a large scale southeasterly wind the main factors for seasonal NO2 variations.field (thus forming southwesterly winds from Shiji-Based on analysis of synoptic weather patterns,azhuang to Bejing along the Taihang Mountains), so taking into account sellite observations and the con-NO2 from the North China Plain and that from thecept of atmospheric environmental pollution process,Xinzhou basin meet in Shijiazbuang, the most heav- it is found that the evolution of large-scale synopticily NO2 polluted area in North China (note: Shijipatterns in the pressure field and vertical tempera-azhuang is located to the east of the Taihang Moun- ture structure of the boundary layer are the main fac-tains). Figure 14e describes the correlation of NO2 col-tors for NO2 concentration variations in North China.umn density with wind fields in Zhengzhou. It shows Their roles are diferent in an atmospheric environ-no evident pollutants around Zhengzhou. Figure 14f mental proces: fronts remove pollutants while highshows weak vector convergence zones existing aroundpressure systems accumulate them, but low pressureJinan and the north part of the Yimeng Mountains. In systems tend to make them converge.summary, NO2 concentration in the convergence zoneSatellite data show that heavy NO2 pollution inunder small- and meso scale wind fields induced by North China often concentrates over the eastern Tai-terrain and near surface fAow fields of the large -scalehang Mountains, southern Yanshan Mountains, Fenhesynoptic systerm is evidently higher than that in its River valley, and northern Yimeng Mountains. Usingsurrounding areas. Apparently, the vector correlationthe Lorentz curves and through the Gini officientfield model can be used to capture the regional distri- calculations, it is found that no overlapped correla-bution of NO2 concentrations.tions exist between distributions of heavy pollutionareas and those of pollution sources. Using the vec-6. Conclusionstor correlation field model, the authors calculated theNO2 column density and the surrounding wind fieldsUsing OMI-based NO2 data, MICAPS data and for six cities in North China, showed the convergencesurface NO2 concentration measurements from 18ranges and channels of pollutants, and concluded thatcities in northern China, the authors investigated andthe convergence of small- and meso-scale wind fieldsanalyzed the environmental process of regional NO2 generated by terrains and surface wind fields of large-pollution, pollutant convergence zones, and correla-scale synoptic systems is the major factor for shapingtions of the atmospheric background with regionalup a heavy NO2 pollution belt.NO2 pollution over this region in February 2007. The .The combined use of synoptic process analysisconclusions are 88 follows.and the vector correlation model is a useful approachSynchronous evolutions were found in temporal to infer a heavy pollution region, which can be usedand spatial sequences of satellite-based NO2 column for environment prediction of heavy pollution events.density and ground -based NO2 measurements, withREFERENCEStheir correlation coficient reaching 0.81. OMI-basedobservations not only reveal heavy pollution eventsChen, z. H., s. Y. Cheng, J. B. Li, et al., 2008: Reover North China but also have the advantage of widerlationship between atmospberic pollution processescoverage and higher resolution, so regional distribu-and synoptic中国煤化工ermn China.IYHCNMHGNO.6WEI Peng, REN Zhenhai, SU Fuqing, et al.811Atmos. Environ., 24(42), 6078- 6087.inter-comparison. Environ. Poll, 157(6), 1873-Ding Yihui, 2005: Advanced Synoptic Meteorology (Sec-1878.ond Edition). China Meteorological Press, 83 ppSu Fuqing, Yang Mingzhen, Zhong Jihong, et al, 20048:(in Chinese)The effects of synoptic type on regional atmosphericKang Na, Gao Qingxian, Wangcontamination in North China. Res. Environ. Sci,2009b: Analysis of regional pollution process and17(1), 16-20. (in Chinese)application of system cluster method. Res. Envi-Ren Zhenhai, Gao Qingxian, et al, 2004b: Con-ron. Sci, 22(10), 1120-1127. (in Chinese)vergence system of air contaminationXin Jinyuan, Lin Yongyao, et al, 2009a: Meteoro-layer above Bejjing and North China: Transporta~logical condition analysis of a typical fine particletion convergence in boundary layer. Res. Environ.pollution event in the piedmont region of Beijing,Sci, 17(1), 22 25. (in Chinese)Res. Environ. Sci., 22(9), 1014-1020. (in Chinese)Wang Yueqi, Jiang Hong, Zhang Xiuying, et al, 2009:Kim, s. W., A. Heckel, S. A. McKeen, et al, 2006:Temporal-spatial distribution of tropospheric NO2Satellite observed US power plant NOx emission re-in China using OMI satellite remote sensing data.ductions and their impact on air quality. Geophys.Res. Environ. Sci, 22(8), 932 -937. (in Chinese)Res. Lett, 33(L22), 812- 817.Xu Xiangde, Zhou Li, Zhou Xiji, et al, 2004: Infu-Lamsal, L. N, R. V. Martin, A. van Donkelaar, et al, .ence domain of the surrounding pollution sources in2008: Ground-level nitrogen dioxide concentrationsheavy urban air pollution process. Sci. China (Ser.inferred ftom the satellite-borne ozone monitoringD), 34(10), 958- -966.instrument. J. Geophys. Res, 113(D16), 308-323.Zhou Xiuji, Han Xue, et al, 2005: Spatial structureOrd6nez, C., A. Richter, M. Steinbacher, et al, 2006:and scale of air pollution sources influencing an ur-Comparison of 7 years of satellite -borne and ground-ban cluster. Sci. China (Ser. D), 35(1), 1-19.based tropospheric NO2 measurements around MiYue Jie, Lin Yunping, Deng Zhaoze, et al, 2009: Seasonallan, Italy. J. Geophys. Res., 111(D05), 310 -322.variations of tropospheric NO2 over megalopolis inPetritoli, A,， P. Bonasoni, G. Giovanelli, et al.,eastern China using satellite remote-sensing dataand chemistry-transport model. Acta Scientiarumsaelite-borne measurements of tropospheric nitro-Naturalium Universitatis Pekinensis, 45(3), 431-gen dioxide in the Po basin. J. Geophys. Res,438. (in Chinese)109(D15), 307- -324.Zhang, Q, D. G. Streets, K. He, et al, 2007a: MajorVan der A., D. H. M. U. Peters, H. Eskes, et al, 2006: .components of China's anthropogenic primary par-Detection of the trend and seasonal variation inticulate emissions. Environ. Res. Lett, 2(2007),tropospheric NO2 over China. J. Geophys. Res,045027- 045034.111(D12317), doi: 10.1029P2005D006594.Ren Zhenhai, Su Fuqing, Gao Qingxian, et al., 2005:China, 1995- 2004: The view from the ground andAnalysis of the serious atmospheric pollution eventthe view from space. J. Geophys. Res, 112(D22),caused by emissions in boundary layer. J. Atmos.1-18.Sci, 29(1), 57 -63. (in Chinese)一, D. G. Streets, G. R. Cermichael, et al, 2009: AsianRichter, A., J. P. Burrows, H. Ni, et al, 2005: Increase inemissions in 2006 for the NASA INTEX-B mission.tropospheric nitrogen dioxide over China observedAtmos. Chem. Phys, 14(9), 5131-5153.from Space. Nature, 7055(437), 129-132.Zhang, X. Y., Zhang P., Zhang Y., et al., 2007: The trend,Sachin, D. G, R. J. Van der A., G. Beig, et al,, 2009: .seasonal cycle, and sources of tropospheric NO2 overSatellite derived trends in NO2 over the major g!obalChina during 1997 -2006 based on satellite measurehotspot regions during the past decade and theirment. Science China (Ser. D), 37(10), 1409- 1416.中国煤化工MHCNMH G