Observation and Research on the Typical Atmosphere of Beijing Olympic Games by Optical Remote Sensing

1. Introduction

The 29^th Olympic and Paralympics Games were held from August 8^th to September 17^th, 2008 in China’s capital – Beijing. Air quality is a critical concern to China and the world. To ensure good air quality during the Olympic Games, Beijing Municipal Government and the Environmental Protection Agency of China launched a program called “Air Quality Guarantee Plan for the 29^th Olympics in Beijing (AQGP)”. The period starting from July 1st, 2008 marked an intensified endeavour to improve Beijing's air quality by shutting down heavy polluters. 3.3 million motor vehicles were removed from city streets on alternate days, depending on whether the license plate ended in an odd or even number during the period of July 20^th to Sep. 20^th, 2008 (the alternative day-driving scheme). At the same time, the adjacent cities and provinces also did their best to close major polluters that may have influence on the air quality of Beijing. The exceptional strict pollution control measures have hugely reduced the source emissions, making a green Olympic Games achievable. Beijing's air quality has witnessed significant improvements due to enhanced pollution control efforts and new environment criteria.

Beijing's air quality was intensively monitored by environment protection authorities and research institutions before, during, and after the Olympic Games. The rich scientific data collected during the period have become solid evidences for judging the impact of exposure to air pollution on human health. To evaluate the effectiveness of the AQGP program, Beijing Municipal Environmental Protection Agency led a plan named “Assessment of AQGP and Air quality Monitoring, Forecasting and Early Warning”.

There are numerous publications (Cermak & Knutti, 2009; Huang et al., 2010; Y. Li et al., 2010; X. G. Liu et al., 2009; Mijling et al., 2009; Shao et al., 2009; Shen et al., 2011; Shou-bin et al., 2009; Simonich, 2009; Streets et al., 2007; X. Y. Tang et al., 2009; W. T. Wang et al., 2009; Westerdahl et al., 2009; Yao et al., 2009) to now on evaluation of these pollution measures. Hou. et. al. (Hou et al., 2010) measured the levels of Beijing residents' exposure to PM₁₀ during three different time periods. Their study showed that during the Olympic Games, population-weighted PM₁₀ exposure came down by 46% and 19%, respectively, compared with the pre-OG and the post-OG periods, indicating that in addition to favourable weather conditions, enhanced traffic and emission control policies and measures have produced a noticeable effect on PM₁₀ reduction.

For persistent organic pollutants (POPs), Zhang et. al. (L. F. Zhang et al., 2010) systematic studied hexachlorobenzene in the ambient air before and after the Olympic Games. Hexachlorobenzene concentration was found decreased sharply in the winter of 2008 comparing with that of 2007 due to the implementation of a series of ‘‘Green Olympic’’ policies. Gas-particle partitioning shows that the increase of hexachlorobenzene levels in winter time was mainly contributed by the high total suspended particulate from combustion processes such as coal-burning and traffic emission.

Xin et. al. (Xin et al., 2010) used Beijing-Tianjin-Hebei Atmospheric Environment Monitoring Network to monitor and provide warnings of the atmospheric quality in Beijing and its surrounding area during the Beijing 2008 Olympic Games. During the Olympic Games, the mean concentration of SO₂, PM_2.5, NO₂, O₃_8h max, and O_x were 12.5±4, 56±28, 23±4, 114±29, 95±17 μg/m³ in the region, respectively, and fell by 51.0%, 43.7%, 13%, 20.2%, and 18.9%, respectively, compared to the prophase mean concentration before the Olympic Games. After the Olympic Games, SO₂, PM_2.5 and NO_x increased significantly as the temporary atmospheric pollution control measures was terminated. Wang et. al’s study (S. X. Wang et al., 2010) showed that the mobile source NO_X and non-methane volatile organic compounds (NMVOC) reduced by 46% and 57%, respectively, during the Olympic Games period due to strict pollution policies.

Zhou et. al. investigated the effects of pollution measures on urban motor vehicle emissions in Beijing. (Zhou et al., 2010) Comparison between the emission intensity before and during the 2008 Olympics shows a reduction of 44.5% and 49.0% in daily CO and NOx emission from motor vehicles. Their results suggest that reasonable traffic system improvement strategies along with vehicle technology improvements can contribute to controlling total motor vehicle emissions in Beijing after the Olympic Games. Wang et. al. (Wang & Xie, 2009) investigated the traffic-related air pollution in the urban streets. The average reduction of PM₁₀, CO, NO₂ and O₃ are 28%, 19.3%, 12.3%, and -25.2%, respectively. Another study (X. Wang et al., 2009) evaluated on-road emission factors and black carbon profiles. Diesel trucks are believed to be a major source of summertime BC in Beijing. A mobile laboratory(M. Wang et al., 2009) also was employed to evaluate changes in on-road air pollutants.

Liu et. al. (J. F. Liu et al., 2009) measured the hourly concentrations of BTEX (Benzene, Toluene, Ethylbenzene, m,p-Xylene and o-Xylene) in the urban area of Beijing. They found that during the games, the mean daytime concentrations of benzene, toluene, ethylbenzene, m,p-xylene and o-xylene were 2.37, 3.97, 1.92, 3.51 and 1.90 μg/m³, respectively, and were 52.8%, 63.9%, 56.4%, 56.8%, and 46.9% lower than those after the games. Secondary pollutants and regional impact also were studied. (T. Wang et al., 2010)

He et. al. (S. Z. He et al., 2010) measured atmospheric H₂O₂ and organic peroxides in urban Beijing (at the Peking University campus), from 12^th July to 30^th September, before and during the Beijing Olympic Games full-scale control. They concluded that the heterogeneous removal of H₂O₂ is faster than that of methyl hydroperoxide (MHP), as indicated by the strong negative correlation between the H₂O₂‐to‐MHP ratio and the aerosol surface area.

Wang. et. al. (B. Wang et al., 2010) assessed variation of ambient non-methane hydrocarbons by gas chromatography equipped with a quadrupole mass spectrometer and a flame ionization detector (GC/MSD/FID) as parts of the field Campaign for the Beijing Olympic Games Air Quality program (CareBeijing). Their findings demonstrate the effectiveness of the air quality control measures enacted for the 2008 Olympics and indicate that controlling vehicular emissions could be the most important measure to improve the air quality of Beijing.

The abovementioned studies used all kinds of methods to evaluate the air quality during Beijing Olympic Games, but nearly none used optical remote sensing methods. Here we will introduce our investigations on the air quality of Beijing Olympic Games using all kinds of optical remote sensing techniques. Due to the limited room in this chapter, we will only introduce part of our results here.

The emission flux of important pollution sources was measured with mobile differential optical absorption spectroscopy (DOAS) and light detection and ranging (LIDAR). The sources are Capital Iron and Steel (CIS), Yanshan Petrochemical (YP), and Capital Airport (CA), respectively. It was found that CIS contributed most to emissions of NO₂, SO₂, and particulate matter, but the overall emission flux from June to September was decreasing, representing the effectiveness of air pollution control policy. It was also found that the emission flux came from YP declined as well, while the NO₂ emission of CA only had a slight drop.

The MAX-DOS system located in the region of pollution sources monitored the tropospheric column density of SO₂ and NO₂, which reflected the whole layer concentration of SO₂ and NO₂ around pollution sources. The results showed that the column density of SO₂ and NO₂ of CIS, YP and SO₂ decreased significantly, where the column density of NO₂ of CIS and YP decreased by 30-40%, whereas that of CA reduced less.

LIDAR measurements proved that emissions of particulate matter can be reduced to some extent by the alternative day-driving scheme and plant production restriction. Weather is the major factor for cleaning of particulate matters, implying that the emission control policy can only reduce the accumulation of pollutant in some extent, but the decisive factor of pollutant elimination is still weather.

The observation to 4 main pollutants (SO₂, NO₂, CO, PM₁₀) in all stations showed their concentration reached the China National Level II Standard and the guideline values of WHO. The concentration of NH₃ was found to be higher in the daytime while lower during the night-time. Varieties of optical remote sensing techniques were used to monitor the air quality of 2008 Beijing Olympic Games, and were proved to be critical in air quality monitoring system.

The tropospheric nitrogen dioxide over the Beijing Olympic Games venues was measured by ground-based Multi Axis Differential Optical Absorption Spectroscopy (MAX-DOAS), and was compared with the result of Ozone Monitoring Instrument (OMI) (Duncan et al., 2010). It showed that the results of MAX-DOAS were higher than that of OMI, and can up to 2.4 times the result of OMI; both instruments had good correlation (r=0.64) when it is clear. There was more difference between the results of both instruments for the existing of cloud when it was rainy or cloudy, with a correlation coefficient of 0.19, but the results of MAX-DOAS had better consistency with LP-DOAS, with a correlation coefficient of 0.92.

2. Experimental methods - optical remote sensing

An integrated spatiotemporal monitoring system (Fig. 1) for regional complex air pollution was constructed which comprises a variety of optical remote sensing techniques, such as differential optical absorption spectroscopy (DOAS)(Winer & Biermann, 1994), light detection and ranging (LIDAR)(Vierling et al., 2008), Fourier transform infrared spectroscopy (FTIR)(Movasaghi et al., 2008), tunable diode laser absorption spectroscopy (TDLAS) (Lackner, 2007), tapered element oscillating microbalance (TEOM) (Patashnick & Rupprecht, 1991). It was worth pointing out that all the instruments mentioned here were developed in our institute (Chen et al., 2009; He et al., 2009; Y. B. He et al., 2010; Jin et al., 2010; S. W. Li et al., 2009; S. W. Li et al., 2008; Lu et al., 2010a, 2010b; Luo et al., 2011; Si et al., 2010; Y. Y. Tang et al., 2010; J. Xu et al., 2010; Z. Y. Xu et al., 2010; S. A. Zhang et al., 2009; Zhou et al., 2010). We established 11 monitoring stations including 3 super monitoring stations and 2 mobile monitoring vehicles. From June 10^th to September 20^th of 2008, the air quality of Beijing area was monitored continuously and in real-time at these stations. Enormous data on all kinds of pollutants were acquired to assess the consequences of air pollution source control policies implied during the Olympic Games.

3. Experimental results and discussions

The integrated spatiotemporal monitoring system consists of 11 monitoring stations equipped with a variety of aforementioned optical remote sensing instruments. There are 3 super monitoring stations and 2 mobile monitoring vehicles. Air quality of Beijing area was monitored continuously and in real-time in these stations from June 10^th to September 20^th,

Enormous data on all kinds of pollutants were acquired to assess the consequences of air pollution source control policies implied during the Olympic Games. Here we will choose some typical data to demonstrate applications of optical remote sensing techniques.

Figure 1.

Integrated spatiotemporal monitoring system for regional complex air pollution

3.1. Aerosol extinction & backscattering coefficients

We’ve done continuously measurements to air quality over the Olympic Games main sports center – “Bird’s Nest” using double wavelength Raman LIDAR with three channels. The Raman LIDAR was installed at the Institute of Remote Sensing Applications (IRSA), Chinese Academy of Sciences, which is located about 500 meters north of the “Bird’s Nest”. The data acquired from Raman LIDAR are used to analyze the optical characteristics of representative weather during the period of Olympic Games.

Fig. 2 shows the aerosol extinction coefficient spatial and temporal distributions during Aug. 7^th – Aug. 25^th, 2010. Obviously, one can find that the aerosol extinction coefficients on Aug. 8^th and Aug. 9^th are considerably large, suggesting pretty heavy air pollution, and thereafter the atmosphere became clear due to several rainfalls.

Figure 2.

Aerosol extinction coefficient vertical distributions during Aug. 7^th ~ 25^th, 2008

The aerosol extinction coefficient at some wavelength can give the distribution of pollutants, but it cannot provide the particle size information. Fig. 3 shows the spatial and temporal distribution of backscatter signal and extinction coefficient at 355 nm & 532 nm from Aug. 7^th to 9^th, 2010. As can be seen from the figure, the backscatter signals at 355 nm and 532 nm illustrate the spatial distribution of small aerosol particles clearly, and the heights of most of these small particles are lower than 600 meters. The most abundant distribution is at ~400 meters, and this is verified by the ratio of aerosol extinction coefficients at two different wavelengths. The height distribution and diurnal change of boundary layer are clearly demonstrated in the spatial -temporal distribution (Fig. 3).

Figure 3.

Ratio vertical distribution of backscatter signal (upper pannel) and extinction coefficient (lower pannel) of 355/532 nm from Aug. 7^th to 9^th, 2008

It is worth noting that Yang (Yang et al., 2010) also used a Dual-wavelength Polarized LIDAR to analyzes the variation of aerosol extinction coefficients during the Beijing Olympic Games. Their results showed that

1. during the Beijing Olympic Games, the aerosol extinction coefficient decreased to about 42.3% in the surface layer (below 250 m) compared with that in 2007 under almost the same meteorological conditions, indicating the effectiveness of local air pollution control measures in Beijing areas;

2. the analysis of the aerosol extinction coefficient obtained before (Jul. 20^th–Aug. 7^th, 2008) and during the Beijing Olympic Games (Aug. 8^th–Aug. 24^th, 2008) shows a maximum decrease of daily aerosol extinction coefficient in the layer of 0.5−1.5 km.

To elucidate such a phenomenon, the transport of PM₁₀ from surrounding areas to Beijing was simulated by NAQPMS (the Nested Air Quality Prediction Modeling System) for the period from Jul. 20^th to Aug. 24^th, 2008. Analysis of PM₁₀ transport from neighboring areas indicated a decrease of 36.6% during the Olympic Games.

Aerosol optical depth (AOD) is a manifestation of air pollution to some extent, and angstrom wavelength exponent (AE) is an indicator of the relative size of aerosol particles. Fig. 4 illustrates the distribution of aerosol optical depth and angstrom exponent between Aug. 7^th and Aug. 25^th, 2010. One can clearly see that the AODs in the first 3 days are large, and the AOD of Aug. 8^th is much larger. The data of Aug. 10^th, Aug. 11^th, Aug 14^th, and Aug. 21^st are missed due to rains in these days. According to the variations of AOD, we found that the AOD rose significantly after rainfalls, hinting a procedure of pollutants accumulation. The average Angstrom exponents calculated from Fig. 4 is 2.64. Fig. 4 shows that the heavy pollution on Aug 8^th-9^th is caused by large aerosol particles. The AEs are very small after the rainfalls on Aug 14^th and Aug. 21^st, meaning the average aerosol particle size is small, and this agrees well with other atmospheric fine particle size measurements.

Figure 4.

Distribution of aerosol optical depth and Angstrom exponent during 20:00~23:59 during Aug. 7^th ~ 25^th, 2008

From the measurements of Raman LIDAR, it was concluded that:

1. the distribution of pollutants on Aug 7-9^th exhibits period variation with time, and the heavy pollution on Aug. 8-9^th is attributed to large aerosol particles;

2. The echo signals of LIDAR at different wavelengths clearly depicts the spatial distribution of aerosol particle size. The altitude of most small particles is about 400 meters. The AOD become much larger after rains, suggesting an accumulation process of pollutants;

3. the average AE is 2.64 during the period of Olympic Games, and its diurnal changes are in very good agreement with the data of ground-based atmospheric fine particle measurement.

3.2. Measurement of SO₂, NO₂, O₃ and PM₁₀

5 of the 11 monitor stations are used to measure the conventional pollutants in Beijing and surrounding areas. These stations are located in Yanshan Petrochemical (YP), Yugang Town (YT), Institute of Remote Sensing Applications, Chinese Academy of Sciences (IRSA), Capital Airport (CA), and Yongledian Town (YLD), respectively. The concentration of gaseous pollutants, i.e., SO₂, NO₂, and O₃, was measured using DOAS. We tested the performance of the DOAS system and calibrateed the lamp spectra every day. The comparison of these data with other instruments in IRSA and YP was used to ensure the accuracy and reliability of field measurements. TEOM was used to get the PM₁₀ data. The measurements started from Jun. 20^th of 2008, and ended in September, 2008. We will introduce the time series, average diurnal change, regional transport, and comparison of pollutant concentration in YP, YT, IRSA, and CA.

3.2.1. Time series of pollutant concentration

Fig. 5 is the hourly average time series of pollutants. The PM₁₀ concentration of IRSA station is much higher than that of YT because IRSA is very close to the Bird’s Nest, and there were more human activities near this station, while YT is in suburb area. The maximum concentration of PM₁₀ in IRSA station occurred at 13:00, July 4^th, and is 470.6 μg/m³. The maximum concentration of PM₁₀ in YT station occurred at 22:00, July 27^th, and is 354.0 μg/m³. The sources of PM₁₀ are very complicated, e.g., coal and dust pollution, industrial emissions, vehicle emissions, and pollutants transport. The weak southerly wind from July 4^th to Aug. 10^th hindered the diffusion of pollutants, and made the concentrations of PM₁₀ and O₃ measured in each station are pretty high.

Figure 5.

Time series of observed SO2, NO2, O3 and PM10

O₃ came from the photochemical reactions during the daytime, and exhibited apparent diurnal variation trend, i.e., the peak value occurred in the afternoon, as the valley happened in the night. The maximum hourly average O₃ concentration occurred at 13:00, Jun. 22^nd in IRSA station, and the value is 324 μg/m³. The concentrations of SO₂ and NO₂ were under the requirement of National Atmosphere Environment Quality Level II, which are 500 μg/m³ and 120 μg/m³. The main source of SO₂ came from combustion of sulfur compounds. The peak values of SO₂ concentration were more frequently found in YP and YT stations. NO₂ was found to have a contrary trend with O₃ in concentration change, i.e., the concentration was high during the night, but became lower in the day due to low photolysis concentration. There are lots of rainy days before July due to Mongolian cold- and low-swirl. In addition, the emission reduction policy started to effect since July. These two factors caused the concentration of NO₂ generally was high as a result of lacking of photolysis. The concentrations of primary pollutants, SO₂ and NO₂, have significant falling with the implementation of emission reduction policy, particularly during the Olympic Games period.

3.2.2. Diurnal variation of pollutants in different monitoring sites

Fig. 6 presents the average diurnal variations of SO₂, NO₂, O₃, and PM₁₀ at different monitoring sites in different stages. The average diurnal variation curves of urban site (Fig. 6a, IRSA) and suburban site (Fig. 6b, YT) show that the concentration of PM₁₀ only changed slightly, and the PM₁₀ concentration of IRSA site is higher than that of YT site. The PM₁₀ concentration measured in IRSA site was found decreasing continuously from Jul. 20^th (The alternative day-driving scheme starting date) to September. The sharpest decreasing occurred between 01:00 ~ 12:00, and the PM₁₀ concentration average diurnal variations at other stages were nearly equal. The PM₁₀ concentration measured at YT site had a significant decline in September than before, and the main difference happened between 02:00 ~ 10:00. The average difference between urban and suburban changed from 50% of Jul. 1^st – Jul. 19^th to 25% of Sept. 6^th – Sept. 17^th.

The measured NO₂ concentration in all the monitor sites exhibited lower value from Jul. 20^th to Aug. 24^th. The diurnal variation has an obvious double-peak. The first peak appeared at 05:00 ~ 09:00, suggesting that it may be related with local traffic emission and human activity. The second peak came out at 19:00 ~ 22:00, and this is originated from concentration accumulation due to atmospheric inversion effects of near-ground atmosphere and lack of photolysis in the night.

The comparison of NO₂ diurnal variations in different monitoring sites showed that the concentration of NO₂ in Capital Airport (CA) changed very little. This is reasonable when considering the large number of air flights every day, and the effect of alternative day-driving scheme is relative small. The measured NO₂ concentration in July is much lower than that of June, and there is no clear peak value in the morning. The NO₂ concentration of all sites during Sept. 6^th –Sept. 16^th increased a little bit compare with previous measurements, especially at around 19:00 in Capital Airport and 21:00 in IRSA.

The NO₂ concentrations were found to decrease most in IRSA and CA sits from the average diurnal variation of SO₂. The intermittent SO₂ peaks appeared in YT site (16:00) before applying alternative day-driving scheme and YP site (10:00) during Jul. 1^st – Jul. 19^th. The NO₂ average diurnal variations generally showed single peak, and only YP and YT sites showed double-peak. The NO₂ concentration is higher during the night-time in urban sites, while high NO₂ concentration was found in other sites, e.g., YT and YP sites were found to have peak value between 07:00 ~ 09:00.

Figure 6.

Average diurnal patterns of SO₂, NO₂, O₃ and PM₁₀ in 4 monitoring stations

In contrast with NO₂ and SO₂, the measured O₃ concentration increased somewhat at the beginning of applying alternative day-driving scheme. This can be explained by that O₃’s originate in Beijing is basically controlled by VOCs. The decreasing of NOx resulted in local O₃’s consumption declination too, thus caused the increasing of O₃ concentration. The statistic data after Jul. 20^th demonstrated that the O₃ concentration decreased continuously. O₃ concentration exhibits typical diurnal variation pattern, that is, the peak value appeared in the daytime, while the concentration became lower due to the reaction with NO. Measurements in urban sites showed that the O₃ concentrations in Jul. 20^th - Aug. 24^th are comparable to that of June, but slightly lower. The peak O₃ concentration appeared earlier from 16:00 of Jul. 1^st - Jul. 19^th to 13:00 of Jul. 20^th - Aug. 24^th, while the peak appeared time from Sept. 6th – Sept. 17^th is almost the same as that the beginning of alternative day-driving scheme and the difference between peak and valley values became smaller and smaller. O₃ concentrations measured in YP and YT sites gave apparent double-peak features, where the first peak appeared around 13:00, and the second peak appeared around 16:00. The second peak probably is related with O₃’s regional transport. Although YT site is far from urban area, O₃ concentration was found to increase at the beginning stage of alternative day-driving scheme, but it became much lower afterwards, and the second peak became not so obvious. We found the same trend for O₃ in Capital Airport, whose peak value appeared at around 14:00. The maximum concentration of O₃ during Jul. 1^st - Jul. 19^th increased by nearly 40 μg/m³ compare with that of June, and become a little lower afterwards.

3.2.3. Observation of pollutants regional transport

We will take the measured data of Aug. 8^th in four sites as an example to investigate the effect of the regional transport in southwest direction. SO₂ and O₃ were chosen as tracer pollutants. Fig. 7 shows the air mass backward trajectory at 200 m, 500 m, and 1000 m heights in 24 hours using HYSPLIT model, where the star labeled the location of IRSA. One can find from Fig. 7 that the main wind direction is southwest, and the air mass transportation is from southwest (YP) to northeast (IRSA).

Figure 7.

Air trajectories for different altitudes of Aug. 8th, 2008

Fig. 8 presents the diurnal variations of SO₂ and O₃ on Aug. 8^th, 2008. The YP monitor site got SO₂ peak value at 12:00 first, and then the peak values were found in YT, IRSA, and CA sites at 13:00, 14:00, and 15:00, respectively. The time gap between the peak values is about 1 hour. SO₂ was found to move from southwest (YP) to northwest (IRSA, CA). Same as SO₂, the O₃ peak value was found in YP site (12:00), and then YT, IRSA, and CA sites. The delay between peaks is about 1 hour as well. The O₃ concentration of YP site was found to be the lowest, and keep at 125-128 μg/m³ level during 12:00-16:00, and lowered a little bit at night-time, but the second clear peak was found after 21:00. YT and IRSA sites then found O₃ concentration peak values during night-time at 22:00 and 23:00, respectively. These peak values may be related with pollutants regional transport. Such findings are consistent with the SO₂ vertical column density measurements of MAX-DOAS.

Figure 8.

Diurnal variation of SO₂ (a) and O₃ (b) measured at 4 monitoring stations on Aug. 8^th, 2008

3.2.4. Comparison of pollutants concentration in different monitoring sites

Here we present the diurnal averages of PM₁₀, SO₂, and NO₂ concentration; the maximum hourly average concentration of O₃ measured during Jun. 20^th – Sept. 22^nd in every monitoring site. Table 1 gives the concentrations of conventional pollutants, PM₁₀, SO₂, NO₂ and O₃, measured at different stages between June and September, 2008, where diurnal averages value were used for PM₁₀, SO₂, and NO₂, while maximum hourly average values were used for O₃.

PM10: Fig. 9 presents the comparison of PM₁₀ daily average concentration and statistical diurnal average concentration measured in IRSA and YT sites. Since IRSA site is close to the Bird’s Nest and there are more human activities around the site, the concentration measured here was about 40% higher than that of YT site. The PM₁₀ variation patterns in two sites are similar, and high concentration was found in two periods (Jul. 24^th – Jul. 28^th, Aug. 4^th –Aug. 7^th). The respirable particulate matters accumulate due to the static stable weather and resulted in the PM₁₀ concentration in IRSA exceeded the National Atmospheric Environmental Level II standard (150 μg/m³). We can see from Table 1 that the PM₁₀ concentration of IRSA which located in urban area was larger than that of YT, and the reductions were 12.1% and 9.9%, respectively. The measurements at urban sites showed that of PM₁₀ concentration had a big growth in September.

SO₂: Fig. 10 presents the variation of NO₂ daily average concentration and the contrast of statistical average concentration. No concentration over standard was found during the whole campaign. Since there were lots of precipitations between Jun. 28^th and Jul. 4^th, the concentrations of pollutants like NO₂ were relatively low due to wet sedimentation. The NO₂ concentration of CA was higher than other sites, and the highest concentration

Figure 9.

Diurnal mean variation (a) and statistical average daily mean of PM₁₀ concentration (b) during the campaign

Site	Pollutants; (μg/m^3)	Measurement Periods
June	7.1-7.19	7.20-8.24	8.8-8.24	9.6-9.17	7.1-7.19/June; Decease	7.20-8.24/7.1-7.19 Decrease	9.6-9.17/8.8-8.24; Decrease
YT	NO2	51.7	32.0	15.6	13.9	28.6	38.1%	51.3%	-105.6%
SO2	22.7	18.2	15.6	11.4	18.5	19.8%	14.3%	-61.7%
O3	176.1	200.4	177.8	155.9	166.1	-13.8%	11.3%	-4.03%
PM10		90.0	81.3	66.3	75.9		9.6%	-14.5%
IRSA	NO2	54.5	33.2	21.9	18.7	32.8	39.1%	34.0%	-75.8%
SO2	19.4	12.3	10.2	7.7	15.1	36.6%	17.1%	-97.0%
O3	185.0	208.3	179.7	136.8	144.7	-12.6%	13.7%	-2.5%
PM10		139.0	122.2	80.8	94.4		12.1%	-16.9%
YP	NO2	37.8	26.6	10.3	10.1	24.1	29.6%	61.3%	-142.5%
SO2	19.1	18.4	14.4	12.6	18.8	3.7%	21.7%	-50.0%
O3	145.4	157.4	135.7	111.6	125.0	-8.3%	13.8%	-8.9%
CA	NO2	57.8	32.5	27.5	28.4	20.5	43.8%	15.4%	-35.8%
SO2	37.2	25.1	16.8	10.7	18.1	32.5%	33.1%	-68.3%
O3	106.0	157.6	153.7	124.1	129.7	-48.7%	2.5%	-3.3%

Table 1.

Concentrations of routine pollutants at 4 monitoring stations from June to September, 2008

measured on Jun. 27^th was 56.8 μg/m³. The average concentrations of each site was found to decrease more than that before implementation of alternative day-driving scheme, especially for IRSA site where the reduction up to 35%. Due to the continuous rain showers between Aug. 8^th and Aug. 15^th, the NO₂ concentrations in several sites were found decreased considerably because of NO₂ purge. CA’s SO₂ concentration decreased significantly during the period of Jul. 20^th – Aug. 24^th compare with the beginning of the alternative day-driving scheme. After the Olympic Games, the SO₂ concentrations measured in all sites were higher than that of August. Specifically, the concentration of SO₂ was found to increase by 97% (Table 2). One can find the CA statistical diurnal average concentration is the highest in all the sites, followed by YT, YP and IRSA. This can be explained by there are more flights in CA area.

Figure 10.

Daily mean variation and statistical average daily mean of SO₂ during the whole campaign

NO₂: Fig. 11 gives the NO₂ diurnal average concentration and the comparison of statistical average concentration. Same as SO₂, there was no concentration over standard was found. There were lots of rains between Jun. 25^th and Jul. 1^st. NO₂ accumulated due to lack of photolysis in this period, and resulted in high concentration during Jun. 27^th and Jun. 29^th. As the weather changed and photochemical activity was enhanced after Jul. 1^st, the NO₂ was decomposed through photolysis and NO₂ concentrations at all sites decreased. The static weather pollutants accumulated on Jul. 29^th and Aug. 1^st-Aug. 8^th resulted in high NO₂ concentration. With the coming of Olympic Games opening ceremony, the NO₂ concentration in CA and IRSA increased gradually, but decreased abruptly on Aug. 8^th. Same as that of SO₂, NO₂ concentration in BA site was higher than other sites. With the implementation of emission reduction policy, NO₂ concentration in YP and YT decreased noticeably during Jul. 20^th to Aug. 24^th. The concentration in all sites increased to some extent in September, especially in YP and YT sites, and the growths are 142% and 105%, respectively. Judged from statistical diurnal average concentration, CA and IRSA had a higher NO₂ concentration due to human activity, then followed by YT and YP.

Figure 11.

Daily mean variation and statistical average daily mean of NO₂ during the whole campaign

O₃: Fig. 12 gives the O₃ diurnal hourly concentration and the comparison of statistical hourly maximum average concentration. As a result of rainy weather, the low photochemical activities led to low O₃ concentration. The concentration exceeding standard happened mainly during the static stable weather period, and they are Jul. 11^th – Jul. 25^th, and Aug. 2^nd – Aug. 6^th. In particular, the O₃ concentration in several sites increased at a rate of about 40% from Aug. 1^st to Aug. 6^th with the maximum happened on Aug. 3^rd, and then decreased at a rate of 7%/day. From Aug. 8^th to the end of Olympic Games, O₃ concentration measured in 4 sites meet the National Air Quality Standard Level II overall and were kept in a relative low level except on some days. From the statistical hourly average maximum concentration, YT site had the highest O₃ concentration, followed by IRSA and CA, and YP had the lowest O₃ concentration.

Figure 12.

Maximum hourly ozone concentrations on a daily basis and statistical average of the maximum hourly ozone during the whole campaign

The exceeding standard hourage of each site is shown in Fig. 20. One can find that O₃ concentrations go beyond the standard on Jun. 22^nd, Jul. 3^rd - Jul. 4^th, Jul. 7^th, and Jul. 16^th. It is worth to mention that the period over standard in YP, YT and IRSA on Jun. 22nd was found to last up to 9 hours. With the implementation of pollution reduction policy, O₃ concentration in YP decreased significantly. The O₃ concentration in all sites meet the requirement of National Air Quality Standard Level II except O₃ concentration of YT site exceeded the standard occasionally.

SO₂ and NO₂ concentrations were found to meet the National Air Quality Level II standard during the whole campaign. The SO₂ and NO₂ concentration declined in the first stage of alternative day-driving scheme (Jul. 1^st – Jul. 19^th). The reduction of NO₂ was far larger than that of SO₂, and reduced up to 30% in all the sites, especially in CA site. All the pollutants concentration continue to decrease in the second stage (Jul. 20^th – Aug. 24^th), and NO₂ concentration in YP site which is in suburban decreased most significantly and up to 60%, then followed by YT site. The SO₂ concentration in CA site decreased most drastically and up to 33%.

Figure 13.

Hourage of O₃ exceeding standard

The O₃ concentration was found to go up then go down at the beginning of the alternative day-driving scheme. This can be explained by that Beijing is located in the VOCs control area generated by O₃. The reduction of NOx due to alternative day-driving scheme will cause O₃ concentration to increase. Compare with June, O₃ in CA site increase most significant and up to 48.7%, and that of YT and IRSA increase by about 13%. The regional transport existed due to the influence of the southwest wind. The YT and IRSA sites located in the downwind of YP site can find peak O₃ in turn.

Since PM₁₀ is more closely affected by human activities, the PM₁₀ concentrations in monitoring sites in urban area were higher. The PM₁₀ concentrations of IRSA which is in urban area were found to have more reduction than that of YT site during the Olympic Games period compare to the first stage, and decreased by 12.1% and 9.6%, respectively.

When comparing the Olympic Games period with the first stage of emission reduction policy, SO₂, NO₂, O₃, and PM₁₀ concentrations in all monitoring sites decreased by 14%-33%, 15%-6%, 2.5%-14%, 10%-12%, respectively. But the concentrations of these pollutants increased during the Paralympic Games period.

3.3. Comparison of NO₂ measured by MAX-DOAS and OMI

Fig. 14 presents the comparison of the tropospheric NO₂ colmun density of MAX-DOAS and OMI from Jun. 10^th to Sept. 14^th. The MAX-DOAS setup was installed close to the Bird‘s Nest. One can easily find the results of ground-basd MAX-DOAS are higher than that of OMI, and can up to 2.4 times on some days (Fig. 14a). This can be explained by spatial resolution difference between MAX-DOAS and OMI. The OMI’s spatial resolution is 13×24 km² whereas MAX-DOAS has much higher resolution. This means the results of OMI is the average of a large area including the urban and suburban area. We know that the NO₂ of suburban area is much lower than that of urban area. This was validated by the average NO₂ column density measured in 3 sites (Fig. 14b). Fig. 14b showed the column density ratio of MAX-DOAS to OMI decrease to 1.5 - 1.8 (Table 2). One can image that with more sites, the MAX-DOAS average will be close to OMI measurements.

	OMI; (10^15molec./cm^2)	MAX-DOAS; (10^15molec./cm^2)	Average; (10^15molec./cm^2)
June	8.54	12.46	13.13
August	7.81	15.91	14.36
September	9.05	19.91	15.36

Table 2.

Comparison of monthly average between OMI and MAX-DOAS

Figure 14.

Comparison between the results of OMI and MAX-DOAS,(a) is the result of comparison between OMI and MAX-DOAS near the Olympic venues,(b) is comparison between OMI and the average of three sites, IRSA, CIS and YT, which are in the same pixel of OMI

It is worth to mention that Yu (Yu et al., 2010) also assessed the effect of the air quality ensuring measures during the game period using the tropospheric NO₂ column density retrieved from Ozone Monitoring Instrument (OMI) onboard AURA satellite. About 40% reduction in tropospheric NO₂ column density over the Beijing area is obtained from the assessment during July to August, 2008, a key period of air quality ensuring measures for the Beijing 2008 Olympic Games.

4. Conclusion

China hosted XXIX Olympic and Paralympics Games in its capital, Beijing, from August to September of 2008. To improve the air quality during this period, extreme stringent pollutions control policies were applied to Beijing and its neighbouring provinces. A variety of optical remote sensing techniques were employed on the measurements of the air quality of Beijing at different stages of Olympic Games. The aerosol extinction coefficients and aerosol optical depth between August 7^th and August 25^th were obtained using Raman LIDAR. Differential Optical Absorption Spectroscopy (DOAS) was utilized to evaluate the density of SO₂, NO₂, O₃. The density of SO₂, NO₂, O₃ from July 20^th to August 24^th were reduced by 14%-33%, 15%-61%, and 2.5%-14%, than that of July 1^st to July 19^th, respectively. The troposphere NO₂ column density acquired using multi-axis differential DOAS (MAX-DOAS) was found to be much higher than that from OMI with a maximum factor of 2.4. PM₁₀ was found to decline by 10% - 12% during the Olympic Game period using TEOM. The results were compared with data from other groups using similar methods or different methods like satellite observations or air quality modelling system, and demonstrated the efficiency of pollution control measures to improve the air quality during the XXIX Olympic Games in 2008.

Using the optical remote techniques aforementioned along with other techniques like satellite observation, an integrated spatiotemporal monitoring system for regional complex air pollution was constructed. The results demonstrated the efficiency of pollution control measures in surrounding areas in improving air quality in Beijing. Optical remote sensing techniques were proved to be critical in air quality monitoring system.

Acknowledgments

The experimental and theoretical work done at AIOFM were supported by grants from National Science Foundation of China (40905010, 40805015), National High Technology Research and Development Program (2009AA063006), and Special Fund of Environmental Protection and Public Service (201109007).