The challenge of improving visibility in Beijing

The “Blue Sky Project” was proposed in 1998 to investigate by how much emissions should be reduced to increase blue sky frequency in Beijing, which hosted the Summer Olympics in 2008. This paper focuses on the temporal variation of visibility and its dependence on meteorological conditions and suspended particles at Beijing using the hourly observed visibility data at Beijing Capital International Airport (BCIA) from 1999 to 2007. It has been found that about 47.8% (24.2%) of the hours in Beijing are “bad” (“good”) hours with visibility below 10 km (equal or higher than 20 km) between 1999 and 2007. Due to the high Relative Humidity (RH), summer is the season with the lowest mean visibility in a year. Although PM 10 index was reported in a decreasing trend (Chan and Yao, 2008), the increase of RH has resulted in a decreasing trend of visibility over BCIA in the summer from 1999 to 2007. To ensure blue sky (“good” visibility) for Olympics 2008, daily mean PM10 index should have been reduced from 81 to 44. This requires that not only vehicle emissions, but also other emissions should be limited. Observations verify that blue-skyhour rate increased significantly after mean PM 10 index was reduced to 53 during Olympics 2008, however, the visibility of 2009 returned to the mean level from 1999 to 2007 during the period 8–24 August. RH (aerosol) contribute 24% (76%) of the improvement of visibility during August 2008.

Beijing hosted 2008 summer Olympics. Atmospheric visibility, as well as air quality, were important factors determining the success of the 2008 Beijing Olympic Games. Measures have been taken to improve the air quality of Beijing, including moving highpolluting industries out of Beijing, replacing coal fuel with natural gas, and phasing out leaded gasoline in recent years. Pollutants such as sulfates and nitrates have 5 been reduced significantly since 1998 (Chan and Yao, 2008), and the PM 10 (Aerosol particulate matters having diameter <10 µm) concentration has also been reduced after 2003 (Chan and Yao, 2008). The Chinese government has planned to increase Blue Sky frequency by limiting vehicle emission during Beijing 2008 Olympics. In this regard, understanding the relationship between visibility and meteorological conditions and 10 aerosols is essential as a first step toward the development of a program to improve visibility in Beijing.
Visibility at a given location is controlled by the physical and chemical properties of the particulate matter and the ambient relative humidity (RH). Numerous studies carried out all over the world indicate that visibility impairment is largely due to the 15 light scattering of ambient aerosols. Concentrations of these aerosols are governed by meteorological conditions and the emission source strength (Sequeira and Lai, 1998;Sloane, 1983Sloane, , 1984Lee, 1983;Dayan and Levy, 2005;Tsai, 2005;Vautard, 2009). Qin and Yang (2000) illustrate a clear trend of decreased visibility in Beijing during 1980Beijing during -1994 the visibility data at Beijing Local Time (LT) 14:00 in clear days. Decreased 20 trend of visibility during 1973-2007 are also detected on the basis of the daily visibility, which was defined by a minimum of four synoptic observations per day from National Climate Data Center (Chang et. al., 2009). There is, however, no comprehensive visibility study based on hourly data for recent years in general. In this paper, the hourly operational meteorological data, including atmospheric visibility, RH, wind direction and Introduction The daily mean concentrations of PM 10 , a leading pollutant, in the Beijing city have been measured in the form of the ambient air pollution index (API). The relationships between API index of PM 10 and PM 10 concentration is given in Table 1. The daily PM 10 15 data is spacially averaged from all monitoring stations in the whole city and have been obtained from Beijing Environmental Protection Bureau (BJEPB: http://www.bjee.org. cn/api/index.php).
A "good" hour, is defined as the hour with a visibility equal or higher than 20 km in this study. A "bad" hour is defined as the hour with a visibility lower than 10 km. A "very Introduction  Fig. 2 (Fig. 3). The data of precipitation and fog hours are excluded 15 from analysis. Summer is the season with the lowest visibility. Mean visibility in summer is generally below 10 km during the night and morning, and exceeds 10 km only in the afternoon. Note that July is the "worst" month with a mean visibility of 9.4 km. The mean visibility in August, the month summer Olympics 2008 was held, is 9.8 km from 1999 to 2007. For the other seasons, mean visibility generally exceeds 10 km during 20 the day time, and is lower than 10 km during the night time. October is the "best" month with a mean visibility of 12.8 km. The moisture abundance in the summer is responsible for the relatively low visibility (Fig. 3).
The lowest visibility of 9.7 km occurs at 01:00 LT, and the highest visibility, which is 13.7 km, appears at 15:00 LT. The relatively good visibility in the afternoon is mainly 25 because of the high temperature (not shown) from solar heating that reduces the RH (Fig. 3), and on the other hand, the higher PBL (Planetary Boundary Layer) in the afternoon may provide suitable conditions for vertical mixing of pollutants, leading to 6203 Introduction

Tables Figures
Back Close

Full Screen / Esc
Printer-friendly Version Interactive Discussion better visibility. In general, the annual mean visibility is decreasing at −0.13 km per year from 1999 to 2007, however the decreasing trend is not significant at the 95% confidence level. While there is an obvious decreasing trend of visibility in most months, summer is the main season that contributes to the decrease of annual mean visibility (Fig. 2). The 5 decreasing trend in July is significant at the 95% confidence level. The decreasing rate peak is about −0.5 km per year in July, and it could be explained by the obvious increase trend of RH in summer (Fig. 3). The mean visibility trends are negative for every hour and reaches the minimum value of −0.17 km per year in the late afternoon (17:00 LT), when a slight increase of RH was found in Fig. 3, although none of them 10 are significant.

Dependence of visibility on PM 10 and meteorological conditions
Several former studies indicated that the decrease of visibility was related to the increase of PM 10 in urban city Xian and Taiwan (Che et al., 2006;Tsai, 2005). The daily mean visibility of BCIA in relation to daily mean PM 10 index of Beijing from 2005 to 15 2007 is shown in Fig. 4. The day with daily mean RH greater than 90% are excluded from analysis (Chang et al., 2009). Daily average PM 10 index in spring is higher than those in other seasons. The correlation index between visibility and ln(PM 10 index) for four seasons are statistically significant at the 99% confidence level. "Good day" hardly occur when PM 10 index is greater than 150. The visibility may vary from 2 km to 45 km 20 when PM 10 index is below 100, and from which, most days with visibility below 10 km occur in the summer due to the high RH (Fig. 3). Figure 5 demonstrates the occurrence frequency at different visibility and RH at BCIA from 1999 to 2007. Results show that low visibility is highly associated with high RH. Most "very bad" hours occur when the RH is 90∼100%. There are a few cases of 25 "very bad" hours caused by sand storm when RH is about 20∼40%. For most hours with visibility higher than 2 km and lower than 10 km, the RH is around 60∼100%. The RH has a broad range of 10∼100% for visibility from 10 km to 15 km, but with a high 6204 Introduction

Tables Figures
Back Close

Full Screen / Esc
Printer-friendly Version Interactive Discussion frequency around 30∼40%. Results also show that good visibility is highly related to low RH. For "good" hours with visibility equal or higher than 20 km, the RH is mostly around 20∼40%. Results in this study are consistent with previous work that showed reduction in visibility under high RH (Malm and Day, 2001;Sequeira and Lai, 1998).
Although the annual PM 10 concentration has been reported to decrease from 1999 5 to 2005 in Beijing (Chan and Yao, 2008), the increase of moisture over the years in the summer may intensify the hygroscopic growth of the aerosol particles, and therefore strengthen light extinction and reduce visibility. Specific humidity reveals an increasing rate of 0.16 g kg −1 per year in the summer from 1999 to 2007 (not shown). This explains the trend of decreasing visibility from 1999 to 2007. Surface moisture increase is also 10 reported by Dai (2006) in a global range and is found to be associated with global warming.
To understand visibility variation at a location, it is important to examine the relationship between visibility and the wind that can represent the transport of the pollutants. Figure 6 presents the visibility distribution in relation to wind direction and wind speed.

15
It is obvious that the visibility over BCIA is low under the prevailing southerly, southeasterly and easterly wind conditions when industrial aerosols are carried over to Beijing (Song et al., 2006;Chan et al., 2005). Mountains around the north and west of Beijing also play a significant role in blocking these urban aerosols from being advected out of Beijing. Moreover, "bad" hours occur under weak variable wind conditions; this 20 may be related to the dominance of local urban emissions. There are some cases of lower visibility under strong southerly or southeasterly wind (greater than 12 m s −1 ) condition, which correspond to the low level jet under severe convective system in the summer. "Good" visibility are associated with northerly, northwesterly and westerly wind greater than 3 m s −1 , suggesting that regional transportation from the sparsely-ACPD 10,[6199][6200][6201][6202][6203][6204][6205][6206][6207][6208][6209][6210][6211][6212][6213][6214][6215][6216][6217][6218]2010 The challenge of improving visibility in Beijing  Figure 7 summarizes the relationship of daily mean visibility with daily mean RH and PM 10 index in August from 2000 to 2007. Also observations in 2008, represented by the symbols with "+" observation are shown in Fig. 7 Fig. 7 shows that PM 10 index needs to be 72 (44) in order to get a visibility of 10 (20) km during August. Improvements in visibility have been experienced at UK after the reduction of pollution (Doyle and Dorling, 2002). A recent study shows that 24% of PM 10 in Beijing is contributed by vehicle emission in a 15 wet season (Zhang et al., 2007). Other emissions, such as soil dust, coal combustion, secondary biomass burning, and industrial emission also make 30%, 19%, 18%, and 9% contributions to PM 10 , respectively (Zhang et al., 2007). If the visibility would be 10 km under a RH of 79%, 70% of vehicle emissions should be limited for the Olympics 2008. For RH higher than 64% in August, the visibility would not reach the standard of and 01:00 LT is the hour with lowest visibility in a day. Analysis of the meteorological conditions reveals that high (low) RH is in high accordance with low (high) visibility. Topography plays an important role in blocking pollutant dispersion. A "good" hour in Beijing is associated with northerly, northwesterly, and westerly wind greater than 3 m s −1 ; a "bad" hour is associated with weak variable wind and wind from the south, 20 the southeast, and the east. The data presented in this paper indicate that there is a significant decreasing trend of visibility over BCIA in the summer from 1999 to 2007. The decreasing trend is mostly contributed by the increase of RH. If the mean RH 79% is considered, a "good" day would occur if PM 10 index were reduced to 44. Assuming 24% of PM 10 in Beijing is 25 contributed by vehicles, a "good" day will not happen during Olympics 2008 even if all vehicle emission is limited. Apart from vehicle emission limitation, soil dust, coal combustion and secondary emission limitation should also be considered to improve 6207 Introduction  10,[6199][6200][6201][6202][6203][6204][6205][6206][6207][6208][6209][6210][6211][6212][6213][6214][6215][6216][6217][6218]2010 The challenge of improving visibility in Beijing 6216 ACPD 10,[6199][6200][6201][6202][6203][6204][6205][6206][6207][6208][6209][6210][6211][6212][6213][6214][6215][6216][6217][6218]2010 The challenge of improving visibility in Beijing  Daily mean visibility distribution in relation to RH, PM 10 index, wind speed, and wind direction for August from 2000 and 2008 in Beijing. Blue symbol represents visibility lower than 5 km, green symbol represents visibility lower than 10 km and no less than 5 km, cyan symbol represents visibility lower than 20 km and no less than 10 km, and red symbol represents visibility no less than 20 km. The symbols with "+" represent observation in 2008. Triangle represents wind speed lower than 1.5 m s −1 , circle represents wind speed greater than 1.5 m s −1 and lower than 3 m s −1 , and square represents wind speed greater than 3 m s −1 . Hollow symbols represent wind directions of south, southeast, east, and southwest. Solid symbols represent wind directions of northeast, north, northwest and west. The black solid lines denote the visibility contour of 5 km, 10 km and 20 km. 10,[6199][6200][6201][6202][6203][6204][6205][6206][6207][6208][6209][6210][6211][6212][6213][6214][6215][6216][6217][6218]2010 The challenge of improving visibility in Beijing