• Akimoto, H., and H. Narita, 1994: Distribution of SO2, NOx and CO2 emissions from fuel combustion and industrial activities in Asia with 1° × 1° resolution. Atmos. Environ.,28A, 213–225.

  • Altshuller, A. P., 1989: Sources and level of background ozone and its precursors and impact at ground level. Atmospheric Ozone Research and Its Policy Implications, Elsevier Science, 127–151.

  • Angle, R. P., and H. S. Sandhu, 1988: Urban and rural ozone concentrations in Alberta, Canada. Atmos. Environ.,23, 215–221.

  • Bell, G. J., P. Peterson, and P. C. Chin, 1970: Meteorological aspects of atmospheric pollution in Hong Kong. Royal Observatory Tech. Note 29, Hong Kong, 35 pp.

  • Bower, J. S., and Coauthors, 1989: Surface ozone concentrations in the U.K., in 1987–1988. Atmos. Environ.,23, 2003–2016.

  • ——, G. F. J. Broughton, M. T. Dando, K. J. Stevenson, J. E. Lampert, B. P. Sweeney, and J. Wilken, 1994: Assessing recent surface ozone concentrations in the U.K. Atmos. Environ.,28, 115–128.

  • Chan, L. Y., H. Y. Liu, K. S. Lam, and T. Wang, 1996: Observation of total column ozone and vertical ozone distribution at subtropical Hong Kong. Proc. XVIII Quadrennial Ozone Symp., L’Aquila, Italy, in press.

  • Chin, P. C., 1986: Climate and weather. A Geography of Hong Kong, T. N. Chiu and C. L. So, Eds., Oxford University Press, 69–85.

  • Chung, Y. C., 1977: Ground-level ozone and regional transport of air pollutants. J. Appl. Meteor.,16, 1127–1136.

  • EPD, 1994: Air quality in Hong Kong: Results from the Air Quality Monitoring Network. Hong Kong Government Printer, 62 pp.

  • Gladstone, K. P., H. Niki, P. B. Shepson, J. W. Bottenheim, H. I. Schiff, and H. S. Sandhu, 1991: Photochemical oxidant concentrations in two Canadian prairie cities: Model evaluation. Atmos. Environ.,25B, 243–254.

  • Heywood, G. S. P., 1953: Surface pressure-patterns and weather around the year in Hong Kong. Royal Observatory Tech. Memo. 6, Hong Kong, 9 pp.

  • Lee, B. Y., and E. Koo, 1987: Atmospheric stability in the eastern Victoria Harbor, Hong Kong. Royal Observatory Tech. Note 39, Hong Kong, 36 pp.

  • Lin, X., M. Trainer, and S. C. Liu, 1988: On the nonlinearity of the tropospheric ozone production. J. Geophys. Res.,93, 15 879–15 888.

  • Liu, C. M., S. C. Liu, and S. H. Shen, 1990: A study of Taipei ozone problem. Atmos. Environ.,24A, 1461–1472.

  • ——, C. J. Lin, H. W. Chang, and S. M. Chou, 1994: Ozone monitoring in Taiwan. Atmospheric Ozone as a Climate Gas, W. C. Wang and I. S. A. Isaksen, Eds., NATO Asi Series, Vol. 32, Springer-Verlag, 289–304.

  • Logan, J. A., 1985: Tropospheric ozone: Seasonal behavior, trend and anthropogenic influences. J. Geophys. Res.,90 (10), 463–510.

  • McFarland, M., D. Kely, J. W. Drummond, A. L. Schmeltekopt, and R. H. Winkler, 1979: Nitric oxide measurements in the equatorial Pacific region. Geophys. Res. Lett.,6, 605–608.

  • McKendry, I. G., 1993: Ground-level ozone in Montreal, Canada. Atmos. Environ.,27B, 93–103.

  • Ogawa, T., and N. Komala, 1989: Diurnal and seasonal variations of the tropospheric ozone in tropical Asia. Ozone in the Atmosphere, Deepak, 437–440.

  • ——, and A. Miyata, 1989: Seasonal behavior of the tropospheric ozone in Japan. Atmospheric Ozone, Proceedings of the Quadrennial Ozone Symposium, C. S. Zerefos and A. Ghazi, Eds., D. Reidel, 754–758.

  • Piotrowicz, S. R., D. A. Boran, and C. K. Fischer, 1986: Ozone in the boundary layer of the equatorial Pacific Ocean. J. Geophys. Res.,91, 13 113–13 119.

  • Sunwoo, Y., and G. R. Carmichael, 1994: Characteristics of background surface ozone in Japan. Atmos. Environ.,28, 25–37.

  • Tsui, W. S., and Y. C. Lee, 1989: Survey of ambient ozone levels in Hong Kong. Pollution in the Metropolitan and Urban Environment Polmet 89, J. Boxall, Ed., Elsevier Science, 430–434.

  • Tsuruta, H., K. Shinya, T. Mizoguchi, and T. Ogawa, 1989: Seasonal behavior of the troposphere ozone in rural Japan. Ozone in the Atmosphere, Deepak, 433–436.

  • Ueda, H., and G. R. Carmichael, 1995: Formation of secondary pollutants during long-range transport and its contribution to air quality in east Asia. TAO,6, 487–500.

  • UKPORG, 1990: Ozone in the United Kingdom. AEA Harwell Laboratory, United Kingdom Photochemical Oxidants Review Group Final Rep. 169 pp. [Available from AEA Harwell Laboratory, Didcot OX11 ORA United Kingdom.].

  • van Aalst, R. M., 1989: Ozone and oxidants in the planetary boundary layer. Atmospheric Ozone Research and Its Policy Implications, Elsevier Science, 573–587.

  • Yeung, K. K., W. L. Chang, B. Wan, F. Kimura, and T. Yoshikawa, 1990: Simulation of boundary layer flow in Hong Kong. Atmos. Environ.,25A, 2161–2172.

  • View in gallery

    Map showing the air-quality monitoring stations in Hong Kong and some nearby locations.

  • View in gallery

    Monthly variations of averaged ozone and carbon monoxide concentrations in Hong Kong.

  • View in gallery

    Typical synoptic charts representing the six types of pressure patterns in Hong Kong.

  • View in gallery

    Typical back air trajectory for (a) summer monsoon (30 May 1996) and (b) winter monsoon (13 Dec 1996).

  • View in gallery

    Windroses for Waland Island.

  • View in gallery

    Diurnal variation of averaged ozone concentrations in four typical months.

  • View in gallery

    Diurnal variation of averaged O3, NO2, and NO concentrations in July and October.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 242 242 37
PDF Downloads 112 112 32

Surface Ozone Pattern in Hong Kong

View More View Less
  • 1 Department of Civil and Structural Engineering, The Hong Kong Polytechnic University, Kowloon, Hong Kong, China
© Get Permissions
Full access

Abstract

Surface ozone (O3) and its precursors in rural and urban areas of Hong Kong are analyzed through the seasonal, temporal, and spatial variation patterns. The seasonal O3 shows a unique pattern with a major peak in autumn and a trough in summer. The spring and winter seasons are the transition periods with a relatively small peak in spring. The seasonal alternation of the prevailing oceanic and continental air masses, plus the climate system associated with the Asian monsoon system, are the governing factors for the temporal O3 pattern in Hong Kong. The O3 imported by these air masses is found to be the dominating factor for the fluctuation of ambient O3 in Hong Kong. The aged air masses associated with the continental outflow from China carry with them anthropogenic air pollutants emitted from the blooming industrial and urban neighborhoods north of Hong Kong in Guangdong Province, China. Under favorable meteorological conditions for photochemical O3 formation in southeast China, the O3 level reaches a maximum in autumn. The absence of a local urban or a summer O3 peak suggests that the local O3 formation is not the dominant source of O3 in summer. The absence of an elevated ground-level O3 peak in the spring season is an indication that the stratospheric intrusion process of O3 is not a significant source of surface O3 in Hong Kong. The authors’ analysis also shows that the emission of O3 precursors from motor vehicles and the complex topography within the territories has a local effect on the spatial O3 distribution and diurnal O3 pattern in Hong Kong.

Corresponding author address: Dr. L. Y. Chan, Department of Civil and Structural Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong, China.

9598225@polyucc.polyu.edu.hk

Abstract

Surface ozone (O3) and its precursors in rural and urban areas of Hong Kong are analyzed through the seasonal, temporal, and spatial variation patterns. The seasonal O3 shows a unique pattern with a major peak in autumn and a trough in summer. The spring and winter seasons are the transition periods with a relatively small peak in spring. The seasonal alternation of the prevailing oceanic and continental air masses, plus the climate system associated with the Asian monsoon system, are the governing factors for the temporal O3 pattern in Hong Kong. The O3 imported by these air masses is found to be the dominating factor for the fluctuation of ambient O3 in Hong Kong. The aged air masses associated with the continental outflow from China carry with them anthropogenic air pollutants emitted from the blooming industrial and urban neighborhoods north of Hong Kong in Guangdong Province, China. Under favorable meteorological conditions for photochemical O3 formation in southeast China, the O3 level reaches a maximum in autumn. The absence of a local urban or a summer O3 peak suggests that the local O3 formation is not the dominant source of O3 in summer. The absence of an elevated ground-level O3 peak in the spring season is an indication that the stratospheric intrusion process of O3 is not a significant source of surface O3 in Hong Kong. The authors’ analysis also shows that the emission of O3 precursors from motor vehicles and the complex topography within the territories has a local effect on the spatial O3 distribution and diurnal O3 pattern in Hong Kong.

Corresponding author address: Dr. L. Y. Chan, Department of Civil and Structural Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong, China.

9598225@polyucc.polyu.edu.hk

Introduction

Ozone (O3) is a secondary pollutant formed in the atmosphere through a series of photochemical reactions from natural and anthropogenic precursors [NOx, CO, and volatile organic compounds (VOCs)]. In the free troposphere and the planetary boundary layer in a given locality, direct emission of O3 is of minor importance. The main formation (sources) and removal (sinks) processes contributing to its budget are vertical and horizontal transportation, in situ photochemical production and destruction, and wet and dry deposition (Altshuller 1989; van Aalst 1989). The variation of the precursor levels and the meteorological conditions of the atmosphere influence the extent of these chemical and physical processes. The superimposition of these production and destruction processes in time and spatial scale leads the O3 in the atmosphere to show a strong spatial and temporal variation pattern.

In midlatitudes, at remote sites far from anthropogenic sources, O3 shows a maximum level in spring and a minimum level in autumn or winter. This variation pattern closely follows that in the lower stratosphere. The indication is that there is substantial contribution of the vertical stratospheric intrusion of O3 to ground-level O3. For areas with large anthropogenic sources, such as power plant, industrial, and traffic emissions, O3 shows a maximum in summer due to the elevated photochemical formation under the abundance of its precursors and favorable meteorological conditions.

In the Northern Hemisphere, surface O3 measurements are frequently concentrated in the developed countries and have a good spatial coverage. However, in the rapid blooming Asian region, routine measurements are unevenly distributed. In Japan, tropospheric O3 levels are high with maximum values occurring in spring (Ogawa and Miyata 1989; Tsuruta et al. 1989; Sunwoo et al. 1994; Udea and Carmichael 1995). The stratospheric O3 intrusion into the troposphere, the long-range transport of anthropogenic pollutants, and the tropospheric O3 formation caused by them are believed to be responsible for this high value. At low latitudes in the tropical region, tropospheric O3 is found to be low compared with that in the mid- or high latitudes as a result of the inactivity of the stratospheric and tropospheric transport (Piotrowicz 1986) and the low photochemical formation due to the lack of precursor sources in the large oceanic areas (McFarland et al. 1979).

Situated on the coast between the Asian continent and the South China Sea, Hong Kong is under the strong influence of the summer and winter Asian monsoons. Air masses from these two totally different origins experience very different long-range transport of O3 to this area. In this paper, the surface O3 pattern is analyzed through the study of the temporal and spatial variations of the rural and urban O3 levels, utilizing information on the production and transport of O3 and its precursors as well as the variation of the climatological and meteorological conditions.

Geography and climate of Hong Kong

Hong Kong is a well-developed city with a population of over 6 million people. It is situated at 22°N, 114°E, lying just within the Tropic of Cancer on the boundary of the continent of Asia and the South China Sea and the North Pacific Ocean (Fig. 1). It is at the tip of the blooming Pearl River delta of Guangdong Province in southern China and close to some large industrial and urban centers. The nearest Chinese city, Shenzhen, special economic zone, with 4 million people is roughly 35 km north from the urban center of Hong Kong, while the central city of Guangdong Province, Guangzhou, is 120 km northwest of Hong Kong. The blooming urban center, Zhuhai, and the Portuguese colony, Macau, are located about 70 km to the west. All these cities lie within the fast-growing industrial zone of the Pearl River delta and its vicinity.

The climate of Hong Kong is governed by the monsoons. The territory lies in the subtropics, and it experiences a variation of weather unusual for tropical countries. It is dominated by the Asian continental high pressure built up during autumn and matured in winter, followed by its collapse in spring and transits to low pressure in summer. This seasonal cycle brings a high proportion of dry easterly or northeasterly winds in autumn and winter, transitional wind directions in spring, and moist southerly or southwesterly winds in summer. The change in these prevailing wind directions always marks the alternation of the two totally different air masses reaching Hong Kong.

The annual mean air temperature varies within a range of some 13°C, from a monthly average of 15.8°C in January to 28.8°C in July. A summary of the monthly means of the surface meteorological parameters observed at the head office of Royal Observatory Hong Kong (ROHK) and Waland Island (Fig. 1) is given in Table 1.

Pollutant emission in Hong Kong

In metropolitan Hong Kong, the majority of the population lives in the urban area and a few new towns, which account for about 15% of the total land area. The residential, commercial, and industrial activities and the major trunk roads are close together. The main built-up areas and urban centers are situated at both sides of the Victoria Harbor and are characterized by a high population density and high-rise buildings. There is a complex road network in these areas with high traffic flow. The main industrial areas with medium and small industries are located at the northwest end of Victoria Harbor and northeast of the territory in Tai Po and Shatin regions next to Tolo Harbor.

Within the territory, there are dominant sources of air pollutants and O3 precursors related to motor vehicle, combustion, industrial and commercial emissions, and energy consumption. Table 2 summarizes the emission inventory for these three types of precursors (NOx, CO, and VOCs) in 1994 in Hong Kong. There were a total of 148 064, 80 075, and 16 556 tonnes of NOx, CO, and VOC, respectively, emitted in 1994 (EPD 1996, personal communication). Among all the sources, power plants and vehicle emissions accounted for the majority of these pollutants. There were about 1500 km of roads and 490 000 registered vehicles, and a total of 470 and 870 million liters of petrol and diesel, respectively, sold in 1994. The vehicular emissions accounted for 88% of VOC, 79% of CO, and 32% of NOx. There were altogether three power stations generating some 8342 MW of electricity in 1991. These power plant emissions accounted for some 62% or 91 850 tonnes of NOx in 1994. Figure 1 shows the site location of these plants. The one in Castle Peak is the largest coal-fired power plant in the world with an annual capacity of 4350 MW, while the one on Tsing Yi Island provides some 1520 MW. These power plants are located in the west of the territory and for much of the time are located downwind of the prevailing easterly winds. The huge amount of emission from these power plant sources together with those from incinerators and fuel combustion processes in urban and industrial areas makes Hong Kong a significant source of O3 precursors for the local areas and cities in the vicinity.

On the other hand, due to its geographical location, Hong Kong may also act as a receptor of the huge anthropogenic emission from this Asian region (Akimoto and Narita 1994) by long- and medium-range transport under the influence of the monsoon system.

Monitoring sites and available data

Mean hourly ground level O3, NOx, and CO data from the Environmental Protection Department (EPD) and The Hong Kong Polytechnic University (HKPU) are used for analysis. In Hong Kong, ambient air pollution level has been measured and monitored by EPD with a fixed site air-quality monitoring network since 1983. Until 1994, there were altogether five sites measuring ambient O3, covering different periods of time and different area types. A brief description of the site characteristics, pollutants measured, and the period of O3 measurement is given in Table 3. The locations of the sites are shown in Fig. 1. In general, the monitoring sites are located on the rooftops of buildings, and due attention has been given to avoid the effect of the local pollution sources and wind sheltering by topography or neighboring buildings.

In these stations, all the O3 measurements are taken by Dasibi 1003-RS UV photometric ozone monitors. The NOx are measured by the Monitor Labs’ 8840 dual-channel chemiluminescent analyzers for the continuous detection of NO and NO2. The O3 analyzers are regularly calibrated by a TECO Model 49 PS O3 calibrater. Flow calibration is performed at constant intervals with a Gillian flow calibrator. The data obtained are recorded by a Microdata M1600L datalogger initially and later by Enhanced Air-Quality-Monitoring Telemtry System through telemetry to the sophisticated and reliable error-checking system, DATAPAK, for central processing. The hourly, daily, monthly, and annual average results are finally derived from 5-min data with over 67% of data being valid (EPD 1994).

Background O3 and CO levels have been monitored by HKPU background air-quality monitoring stations since 1994. The station is located at the top of a cliff at Cape D’Aguila, Hok Tsui (HT), in the most southern tip of Hong Kong Island (Fig. 1). The site is in a relatively remote and rural area away from local vehicular and industrial emissions. The O3 level is measured by the UV photometric analyzer (Thermal Environmental Instrument Inc., Model 49). The instrument is subjected to a daily check against artifact by zero air and the response of the instrument is checked daily using the internal O3 source. The detection limit is 2 ppb. The CO level is measured by a gas filter correlation, nondispersive infrared absorption instrument (Thermal Environmental Instrument Inc., Model 48) with a modification for a low-concentration measurement. The analyzer is subjected to routine calibration by a standard addition method. The valid data of O3 in 1994 from this measurement are over 76%.

In addition, the meteorological data from ROHK, in particular, the wind speed measurement from Waland Island next to the HT station, and the meteorological parameters from the head office of ROHK in urban areas are used for analysis in this paper. The EPD network data from 1990 to 1994 and the HKPU background monitoring station data in 1994 are used for analysis.

Result and discussion

Seasonal variation

In Hong Kong the spring months are March, April, and May; summer months are June, July, and August; autumn months are September, October, and November;and winter months are December, January, and February. Figure 2 shows the monthly averaged time series of surface O3 for the typical remote, rural, urban, industrial, and urban new town stations: HT, JB, CW, KC, and TP, respectively. Monthly CO variation for HT is also shown in Fig. 2. The period of the data covered is as shown in Table 3. All the stations show similar seasonal variations with O3 troughs in summer and the O3 peaks in autumn. The winter and spring seasons are usually transition periods with a relatively small peak in spring for the rural and remote sites. The transition is characterized by a sharp change from its summer trough to its autumn peak.

Note that this O3 pattern in Hong Kong is quite different when compared with surface O3 observations at midlatitudes, Great Britain and Canada, in the Northern Hemisphere, for example. Normally these O3 observations show a seasonal cycle with spring maximum and autumn minimum similar to those in the free troposphere or lower stratosphere (Angle and Sandhu 1988; McKendry 1992; Bower et al. 1989 and 1994). This usual 1993 spring surface O3 peak and autumn minimum is not apparent in Hong Kong. The surface O3 in Hong Kong shows a different pattern from that in the free troposphere. The vertical O3 profile measurement by Chan and his coworkers (Chan 1996) has shown that the commonly observed springtime O3 maximum is also apparent but to a much lesser extent in the free troposphere in Hong Kong. Short-term cross-country and vertical-spiral aircraft measurements suggest that there are occasionally higher O3 concentrations in the downwind edge of the urban plume and in the spring season (Tsui and Lee 1989). This suggests that the direct vertical stratospheric or tropospheric exchange process is not the dominant source of surface O3 in Hong Kong. When compared to the measurement in East Asia, the O3 pattern seems to share the same O3 trough in summer with the tropical and midlatitude regions (Ogawa and Komala 1989; Ogawa and Miyata 1989; Tsuratagw 1989 and Sunwoo et al. 1994). When compared with the results from Taiwan, it has a similar peak in the autumn season (Liu et al. 1994). The absence of the usual urban and summer O3 peak suggests that the summer local photochemical O3 production is not significant in Hong Kong. It is interesting to note that the seasonal O3 pattern is different from those of the tropical and the midlatitude regions in Asia. This unique pattern is probably due to the prevailing air masses and climatic characteristics associated with the monsoon systems experienced in Hong Kong.

As mentioned in section 3, the climate of Hong Kong is governed by the Asian Monsoon System, which is a climatological phenomenon manifested by a marked change of airmass flow and wind direction between summer and winter due to the change of the thermal structure of the Asian continent. The system involves different air masses and produces noticeable effects on the weather and climate of the whole Asia–Pacific region. Such seasonal variation of the monsoon system always induces a distinctly marked variation of the pressure distribution at sea level in this region. We have looked at the daily weather map of Hong Kong and its vicinity for 1994. Following Heywood (Heywood 1953) and Chin (Chin 1986), we classify the surface pressure pattern into six main catalogs: northerly (N), northeasterly (NE), easterly or southeasterly (E), south or southwesterly (S), trough (T), and cyclonic (C). The typical synoptic situations associated with these pressure patterns are shown in Fig. 3. The detailed description of the pressure patterns, such as the centers of anticyclone and cyclone, the period of occurrence and predominance, the associated weather conditions, and the prevailing surface wind in Hong Kong, is shown in Table 4. Table 5 shows the results for 1994 as the monthly distribution frequency of these six types of pressure patterns.

In summer, the heating of the Asiatic continent results in a very low pressure developed over central Asia (central Asia low). The pressure system is dominated by a strong cyclone located over the northwestern Indian–Pakistani subcontinent and a weak anticyclone over the Aleutian Islands (Aleutian high) of the northern Pacific Ocean. Added to these, there is a strong equatorial high pressure system over Australia and New Guinea (equatorial high) in this season. The interaction of these pressure systems results in the surface pressure distribution pattern types T, S, and C, respectively. Type T is dominated by a low pressure trough with an axis extending approximately east–west over or to the south of Hong Kong. Type S results in a quasi-stationary low pressure area over China, and type C is the circulation of a traveling cyclone. Such a synoptic situation led to an inflow of tropical–Pacific and equatorial air masses from a southeastern and southwestern direction, respectively, toward the center of the Asian continent through Hong Kong. Figure 4a shows typical airmass trajectories associated with this summer monsoon. The paths of the trajectories indicate that the air masses originated from the equatorial tropical region traveling through the South China Sea to Hong Kong. These tropical maritime air masses are warm and humid in nature. When reaching Hong Kong, they are either easterly, southeasterly, south, or southwesterly winds, depending on the actual type of the airmass flow. Figure 5 shows the windroses of the four seasons recorded in Waland Island, which can be considered representing the prevailing wind of Hong Kong due to its location away from a geographical barrier (Fig. 1). The windrose of the summer season shows that the dominant surface wind is generally from the south and east directions. This fact, together with the high air temperature (27.8°–28.8°C), dewpoint temperature (24.4°–24.9°C), and relative humidity (80%–82%) shown in Table 1, supports our discussion above.

On the other hand, the prevailing air mass and the atmosphere condition associated with the summer monsoon is under upward convective motion (Bell et al. 1970, 1977). Thus, the air mass is also believed to have originated from the lower troposphere or boundary layer close to the surface of the ocean from the tropical region. This tropically originated marine air is well known to be low in O3 (Piotrowicz et al. 1986; Ogawa and Miyata 1985; Logan 1985) since there is a lack of sources of precursors, NOx, CO, methane, and nonmethane hydrocarbon, in the ocean (McFarland et al. 1979). In fact, the O3 concentration measured (16–22 ppb) in the remote HT station is comparable in magnitude to the background O3 of the tropical air masses (Logan 1985; Altshuller 1989). Accompanying such large-scale motion, the meteorological conditions in summer are unstable. Table 6 shows the monthly frequency distribution of Pasquill stability classes at the International Airport of Hong Kong (Fig. 1) determined by Smith’s nomogram using insolation, wind speed, and cloud cover as parameters adopted from Lee and Koo (Lee and Koo 1987). Note that the occurrence of unstable classes A, B, and C from June to August is much higher than the other seasons. Added to that, summer is the rainy season. Therefore, the local photochemical O3 formation and accumulation is limited despite high solar radiation and high air temperature in this season.

In early autumn months, September or October, the situation starts to reverse. The summer monsoon weakens and the northeast winter monsoon gradually becomes prevalent. It is because the temperature over central Asia starts to decrease due to the thermal cooling of the underlying continent. An anticyclone high pressure gradually developed over this area, while the intense cyclone over the northwestern Indian–Pakistani subcontinent quickly disappears. At the same time, the Aleutian high and the equatorial high are gradually replaced by the Aleutian low and equatorial low as the season progresses from summer toward autumn. These pressure systems result in a shift of the surface pressure to be more dominated by the anticyclone over China. Pressure distribution types N, NE, and E are characterized, respectively by anticyclones over China, which extends over the East China Sea and southern Japan; one centered east of 130°E; and another anticyclone centered north of 20°N, which often has a ridge extending to Taiwan. The monthly frequency distribution of these surface pressure pattern types in Hong Kong for 1994 is described in Table 5. The geographical location of these pressure systems results in a moderate pressure gradient between central Asia and northwestern Pacific Ocean and between central Asia and the Australian and New Guinea regions. This pressure gradient brings in a cool and dry air mass to Hong Kong in winter, and it may originate from the north as far as Manchuria or eastern Siberia.

A typical air trajectory in the autumn period shows that the air mass travels a long path over continental Asia. It curves across the warm seas to the south of Japan and then flows through a large area of continental mainland. When reaching southern China, it follows the coastline blowing over the East China Sea and the Taiwan Straits to the rapidly blooming, industrialized southeastern part of China and then to Hong Kong. The windrose (Fig. 5b) indicates that the prevailing air mass is generally from the north, northeast, or east direction. The air is drier and cooler than that in summer (Table 1). It is actually continental in nature and the low-level cold monsoon moves beneath the upper warm westerly (Bell et al. 1970, 1977). Due to the large-scale downward motion driven by the moving anticyclone, the air mass is believed to be from the free troposphere or lower troposphere. The air is aged in nature and very often heavily laden with pollutants from these areas (Bell et al. 1970, 1977). It is also rich in O3 precursors. From Fig. 2, it can be seen that CO in the remote HT station gradually increases in this period and reaches its maximum level in winter. As CO has a long lifetime in the atmosphere, its higher level in this season indicates the influence of the anthropogenic nature of the air masses.

On the regional basis, there is fine weather in the central and southeast parts of China in the autumn season. As shown in Table 1, the climate in this region and Hong Kong has minimal cloud cover (53%–63%), fairly long periods of bright sunshine (182–195 h), strong mean daily global solar radiation (13.4–16.5 MJ m−2), fairly high temperature (21.4°–27.6°C), and with a dominant weak stationary high pressure system. From Table 6, the relative high frequency of the stable classes E and F + G (13.3%–15.3% and 19.3%–24.9%, respectively) indicates that the atmosphere is quite stable in the autumn season. This atmospheric condition around southeastern China is favorable for the photochemical production of the background O3 in the free troposphere. Added to that, the anthropogenic precursors emitted from the highly industrial and urban areas in the southern parts of China, especially the Pearl River delta region, enhanced the opportunity to form O3 in the aged air masses. Therefore, the high O3 level measured in Hong Kong is greatly influenced by long-range transport related to the industrial and urban O3 precursor emission from mainland China. Also, under such fine weather and stable atmosphere, the local O3 formation will be higher than that in the summer months. Hence, the surface O3 peak in autumn is a result of the increase in background O3 due to the long-range transport aged air masses from China due to the continental outflow as well as the local photochemical formation under favorable meteorological conditions.

The transition of the autumn O3 peak to summer trough is more gradual and is related to the gradual change of the regional climate. In winter the northern monsoon is more frequent and in spring there is the shifting from the northeast to the southwest monsoon.

As the season progresses to deep autumn and winter, the anticyclone over central Asia, the Aleutian high, and equatorial low further intensify and become more mature. Such changes give rise to a more frequent occurrence of the types N and NE pressure patterns in this region (Table 5). The strengthening of this northern monsoon in late autumn results in the lessening of the curving of the air mass across the East China Sea; a typical air trajectory is shown in Fig. 4b. From late November to late December, the weather is still influenced by warming of the cold northeast winds as they blow over the warm water of the East China Sea. This warming creates a fairly favorable atmospheric condition for the O3 formation and this explains why there is a fairly high O3 concentration. However, as winter further progresses to January and February, the anticyclone becomes fully established and cold fronts very often move southward across China. The air mass becomes more directly from the north in the polar region. It is continental in nature with the lowest relative humidity (68%–78%). As it reaches Hong Kong, the temperature falls rapidly and sometimes reaches only 10°C. The change of the atmospheric conditions in the region continues and the wind becomes more from a north and north-northeasterly direction (Fig. 5). It also comes with very low sunshine hours (98–182 h) and minimum mean daily solar radiation (10.7–12.0 MJ m−2). Therefore, the in situ photochemical production of the background O3 in the free troposphere as well as in the urban plume decreases. The O3 concentration drops eventually to a lower level.

During the early spring period, the air pressure over central Asia begins to fall. The continental outflow thus decreases in strength, while the maritime air from over the Pacific Ocean and the South China Sea gradually develop and start to inflow through Hong Kong to the low pressure cell over central Asia. By late spring and early summer, the central Asian low, Aleutian high, and the equatorial high further increase in strength. The airflow becomes more from the south and west directions and is more and more of marine origin. The surface pressure pattern in Table 5 shows that there is a gradual decrease of types N and NE with an increase in types E, SE, and S patterns from early spring (February) to early summer (May). The windrose (Fig. 5b) demonstrates that the prevailing air mass flows more regularly from the southeasterly direction. This inflowing air mass is of tropical and oceanic origin, which dilutes the O3 and its precursors in the continental air mass. In late spring, unstable weather begins to dominate. The intrusion of the warm maritime air mass from the southeasterly direction always leads to the formation of the advection fog and mist in the boundary layer in this season. It is characterized by thick cloud cover, short sunshine hours (96–154 h), and the formation of troughs of low pressure. These factors lead to the lower photochemical production in the boundary layer and thus the surface O3 concentration continues to drop through the summer season.

Spatial ozone distribution pattern in Hong Kong

From Fig. 2, we see that the O3 level within the territories can be classified into five different groups related to land use. The ranging orders of O3 concentration are remote areas, rural areas, urban areas with commercial usage, industrial areas, and new towns. In the remote and rural HT station, the O3 concentration is the highest. Although part of the August data is missing, the annual average of this station is close to 29.5 ppb with a monthly maximum of 44 ppb in October and November and an annual monthly minimum of 16 ppb in July. The seasonal average values are 42, 29, 28, and 19 ppb from autumn to summer, respectively. The rural JB station shows a similar seasonal O3 pattern as in HT. The minimum monthly average is 10.8 ppb in July, the maximum monthly average is 30.5 ppb in October, and the annual average is 18.6 ppb.

In urban and industrial stations, the monthly minimum O3 levels range from 1.5 to 6.5 ppb in summer or winter and the monthly maximum O3 levels range from 7.4 to 20.9 ppb in autumn, which are lower than that in rural and remote stations. The urban commercial CW station shows a monthly maximum of 19.9 ppb in October and a minimum of 6.5 ppb in June, while the industrial KC station has a maximum of 20.9 ppb and a minimum of 5.4 ppb in October and July, respectively. In the new town TP station the maximum and minimum are 7.4 and 1.5 ppb in October and December, respectively. The annual averaged concentrations in the urban commercial CW station and urban industrial KC station are close—10.8 and 10.1 ppb, respectively. In the new town TP station the annual average is lowest, reaching only 4.9 ppb.

Due to the geographical location, the remote HT station and the rural JB station are situated at the upwind during the summer monsoon and downwind during the winter monsoon in relation to seasonally transported air masses. Thus, the relatively closely observed O3 pattern around the year between these remote and rural stations suggests that the background O3 is the dominant source in Hong Kong. In fact, the O3 concentration in these stations with the monthly average from 10.8 to 44 ppb are close to the background O3 level (20–50 ppb) estimated for this region (Logan 1985; Altshuller 1989). Moreover, the lower O3 concentration in the urban areas indicates that the urban areas of Hong Kong act as a net sink for the background O3 and that the local photochemical formation due to local precursor emissions is not significant. In fact, this O3 sink is a common feature observed in many countries in the Northern Hemisphere, such as in Great Britain and Canada. In these two countries, the urban stations in central London (Bower et al. 1989; UKPORG 1990) and Alberta (Angle and Sandhu 1988) show lower O3 concentrations than their counterparts in the rural areas. This can be explained by the fact that the fresh precursor emissions from traffic and other sources cause direct chemical scavenging of O3.

In an urban environment, the photochemical O3 formation, destruction, and accumulation are dominantly determined by the photochemical reactions between O3, and its precursors, NOx and VOC. It has been observed (Chung 1979) and shown by model evaluation (Gladstone et al. 1991) that the formation and accumulation of O3 tend to occur over the timescale of several hours and that the direct reaction between O3 and NO tends to occur in less than an hour, leading to the so-called NOx scavenging or titration effect. Such a titration effect lowers the ambient O3 in the urban areas.

In Hong Kong, the metropolitan area is well ventilated by the strong prevailing wind and the channeling effect (Yeung et al. 1990), created along both sides of Victoria Harbor due to its coastal location and complex topography. These local meteorological and atmospheric conditions provide a favorable atmospheric environment to disperse the primary O3 precursors and inhibit the in situ formation and accumulation of O3. Indeed, Bell et al. (1970, 1977) has shown that even under light wind conditions, pollutants generated from local sources will be dispersed within 2–3 h. Thus, the titration effect of the fresh O3 precursors, especially NO, emitted from the metropolitan area of Hong Kong leads to the lower O3 levels in the urban stations in our study.

Diurnal variation

The diurnal O3 variation in April, July, October, and January for the four typical seasons for the HT, JB, CW, KC, and TP areas are shown in Fig. 6. Also, the diurnal variation of O3, NO, and NO2 at maximum (October) and minimum (July) O3 concentrations for these stations are shown in Fig. 7. The diurnal curves shown in these two figures are such that they represent the hourly averaged data available during the period, as shown in Table 3. From Figs. 6 and 7, we see that there are distinct diurnal O3 variation patterns according to land use and the location of the monitoring stations.

In the remote HT station the diurnal pattern shows a small dip in the morning around 0600–0800 LST and an afternoon peak around 1500–1700 LST in January, April, and July. However, in October, it only shows a single peak around 1600 LST in the afternoon. In the rural JB station, the O3 shows a relatively less diurnal change than that in HT. It is especially true for April and October. In general, the diurnal variation in this station follows the pattern in HT but with less fluctuation. In the remote HT station, the diurnal O3 reaches a minimum of 13 ppb in July and a minimum of 38 ppb in October. In this remote station, the maximum diurnal peak (20 ppb) in July is less than the maximum peak (58 ppb) in October by 38 ppb and it reflects the insignificant photochemical formation as mentioned in section 5a. On the contrary, the high magnitude in October with a maximum of 58 ppb reflects the substantial O3 formation in the prevailing aged air mass under favorable meteorological conditions in southeastern China. In the rural JB station the substantial diurnal change in the summer month, July is probably due to its downwind location in this season in the prevailing southwesterly wind. The precursors’ emission from the upwind urban areas contributes to substantial photochemical O3 formation in the midday. Thus, O3 reaches its maximum at noontime.

For all seasons, the urban, industrial and new town stations exhibit consistent bimodal variation patterns. The diurnal cycle is characterized by a comparable midafternoon (1200–1400 LST) peak, corresponding to maximum solar radiation, and an early morning (0200–0500 LST) peak. It is also characterized by two O3 troughs in the morning (0700–0900 LST) and evening (1700–1900 LST), corresponding to the rush hour with maximum traffic flow. A similar bimodal diurnal or double-peaked pattern with a morning peak has actually been observed for all the monitoring stations in Taiwan (Liu 1990). From Fig. 6, it is observed that for all the stations, the afternoon peaks increase as well as broaden from urban to rural and remote stations. Also, the early morning peak is more obvious for the urban and industrial stations.

The observed diurnal cycle of O3 in the atmosphere can be explained in terms of the change of physical conditions of the boundary layer and the chemistry of O3 and its precursors.

For the rural and urban stations, Fig. 7 shows that the O3 variation throughout the day in these stations is in an opposite pattern to that of NO. The occurrence of the O3 peaks and troughs always coincide with the NO minimum and maximum, respectively. The O3 troughs are most likely due to the titration effect of the fresh vehicular emission since it is rich in O3 precursors, especially NO, which acts as a direct scavenging agent to the incident background or ambient O3 in the early stages of oxidation (Chung 1979; Angle and Sandhu 1989). In these stations, this phenomenon is especially apparent with fresh NO, NO2, and VOC emissions from traffic and industries. In between these O3 troughs, the O3 peak at noontime can be at least partly explained by the increase of the in situ photochemical formation due to the increase of the temperature and solar radiation that reach their maximum at midday (Lin et al. 1988).

In addition, there are diurnal changes in the atmospheric boundary layer depth. In morning, the heating of the earth surface by solar radiation causes the air to become unstable and to rise. This results in an increase in the mixing depth of the boundary layer and brings in mixing with air mass aloft in the troposphere by vertical mixing. The depth of this mixing layer reaches its maximum at noon when the solar radiation becomes most abundant. Therefore, the increase in the mixing depth of the boundary layer and mixing with O3-rich air mass aloft in the troposphere causes an increase of surface O3 from the morning to its peak at noon. The higher mixing depth also acts as a factor for diluting the fresh precursors and lowers the scavenging of the O3. This effect, together with the photochemical formation, explains why the O3 reaches its maximum in midday. At night, when most of the industries stop their operation and traffic flow subsides to reach its minimum at dawn, the O3 eventually regains its normal background level. When the traffic starts to pick up again, the O3 gradually decreases and thus shows an early morning peak at dawn. Liu and his coworkers (Liu et al. 1990) had shown that strong local circulation is the primary cause of such early morning O3 peak. In a coastal location such as Hong Kong, temperature inversion is rare (Bell et al. 1970, 1977), and the strong prevailing wind inhabits the formation of a strong local circulation such as the one that happens in Taiwan. Therefore, the cause of the observed early morning O3 peak in the urban areas of Hong Kong is different from that in Taiwan.

Conclusions

In this study, the surface O3 pattern of Hong Kong is analyzed through interaction with its precursors and the influence of climatic and meteorological conditions. Using available data from the Environmental Protection Department and the Hong Kong Polytechnic University network, the seasonal, temporal, and spatial variation patterns are obtained.

The surface O3 pattern in Hong Kong is quite different when compared with surface O3 observations at midlatitude countries in the Northern Hemisphere, such as Great Britain and Canada. Instead of a seasonal cycle as in these countries with a spring maximum and an autumn minimum, Hong Kong experiences a summer minimum like other countries in East Asia and a distinct autumn maximum. There is only a relatively small peak observed in spring for the remote and rural stations.

The long-range transport of oceanic and continental air masses associated with the Asian monsoon in summer and in winter carrying with it distinctly different O3 and precursor contents is found to be the governing factor for the seasonal O3 variation. The aged air mass associated with the continental outflows from continental China in autumn and winter is rich in anthropogenic air pollutants transported from the blooming industrial and urban neighborhoods north of Hong Kong. Under favorable weather conditions over southeastern China in autumn, the O3 level reaches a maximum. The transition of the autumn peak to summer trough is more gradual with the eventual retreat of the winter monsoon and the development of the summer monsoon. The marine air masses reaching Hong Kong in summer are known to be low in O3 since there is a lack of source of precursors in the ocean. A summer O3 minimum is observed. The absence of a local urban or summer O3 peak suggests that the local photochemical O3 formation is not the dominant source in Hong Kong. The absence of an elevated ground level O3 peak in the spring season also indicates that the stratosphere intrusion process of O3 is not a dominant source of surface O3 in Hong Kong.

It is noted that there is spatial variation due to the complex topography and the type of land use and diurnal variation due to traffic emissions and meteorological condition. Yet, on the whole, the surface O3 pattern in Hong Kong is dominated by the Asian monsoon system. Hong Kong’s geographical location, which receives both continental outflow from China in autumn and winter as well as marine air masses from the Pacific Ocean during late spring and summer, has the greatest influence on the surface ozone pattern in Hong Kong.

Acknowledgments

The authors would like to thank the Environmental Protection Department and Royal Observatory of Hong Kong for providing the necessary data. This study is supported by a Hong Kong Polytechnic University research grant and a grant from the Research Grant Council of Hong Kong.

REFERENCES

  • Akimoto, H., and H. Narita, 1994: Distribution of SO2, NOx and CO2 emissions from fuel combustion and industrial activities in Asia with 1° × 1° resolution. Atmos. Environ.,28A, 213–225.

  • Altshuller, A. P., 1989: Sources and level of background ozone and its precursors and impact at ground level. Atmospheric Ozone Research and Its Policy Implications, Elsevier Science, 127–151.

  • Angle, R. P., and H. S. Sandhu, 1988: Urban and rural ozone concentrations in Alberta, Canada. Atmos. Environ.,23, 215–221.

  • Bell, G. J., P. Peterson, and P. C. Chin, 1970: Meteorological aspects of atmospheric pollution in Hong Kong. Royal Observatory Tech. Note 29, Hong Kong, 35 pp.

  • Bower, J. S., and Coauthors, 1989: Surface ozone concentrations in the U.K., in 1987–1988. Atmos. Environ.,23, 2003–2016.

  • ——, G. F. J. Broughton, M. T. Dando, K. J. Stevenson, J. E. Lampert, B. P. Sweeney, and J. Wilken, 1994: Assessing recent surface ozone concentrations in the U.K. Atmos. Environ.,28, 115–128.

  • Chan, L. Y., H. Y. Liu, K. S. Lam, and T. Wang, 1996: Observation of total column ozone and vertical ozone distribution at subtropical Hong Kong. Proc. XVIII Quadrennial Ozone Symp., L’Aquila, Italy, in press.

  • Chin, P. C., 1986: Climate and weather. A Geography of Hong Kong, T. N. Chiu and C. L. So, Eds., Oxford University Press, 69–85.

  • Chung, Y. C., 1977: Ground-level ozone and regional transport of air pollutants. J. Appl. Meteor.,16, 1127–1136.

  • EPD, 1994: Air quality in Hong Kong: Results from the Air Quality Monitoring Network. Hong Kong Government Printer, 62 pp.

  • Gladstone, K. P., H. Niki, P. B. Shepson, J. W. Bottenheim, H. I. Schiff, and H. S. Sandhu, 1991: Photochemical oxidant concentrations in two Canadian prairie cities: Model evaluation. Atmos. Environ.,25B, 243–254.

  • Heywood, G. S. P., 1953: Surface pressure-patterns and weather around the year in Hong Kong. Royal Observatory Tech. Memo. 6, Hong Kong, 9 pp.

  • Lee, B. Y., and E. Koo, 1987: Atmospheric stability in the eastern Victoria Harbor, Hong Kong. Royal Observatory Tech. Note 39, Hong Kong, 36 pp.

  • Lin, X., M. Trainer, and S. C. Liu, 1988: On the nonlinearity of the tropospheric ozone production. J. Geophys. Res.,93, 15 879–15 888.

  • Liu, C. M., S. C. Liu, and S. H. Shen, 1990: A study of Taipei ozone problem. Atmos. Environ.,24A, 1461–1472.

  • ——, C. J. Lin, H. W. Chang, and S. M. Chou, 1994: Ozone monitoring in Taiwan. Atmospheric Ozone as a Climate Gas, W. C. Wang and I. S. A. Isaksen, Eds., NATO Asi Series, Vol. 32, Springer-Verlag, 289–304.

  • Logan, J. A., 1985: Tropospheric ozone: Seasonal behavior, trend and anthropogenic influences. J. Geophys. Res.,90 (10), 463–510.

  • McFarland, M., D. Kely, J. W. Drummond, A. L. Schmeltekopt, and R. H. Winkler, 1979: Nitric oxide measurements in the equatorial Pacific region. Geophys. Res. Lett.,6, 605–608.

  • McKendry, I. G., 1993: Ground-level ozone in Montreal, Canada. Atmos. Environ.,27B, 93–103.

  • Ogawa, T., and N. Komala, 1989: Diurnal and seasonal variations of the tropospheric ozone in tropical Asia. Ozone in the Atmosphere, Deepak, 437–440.

  • ——, and A. Miyata, 1989: Seasonal behavior of the tropospheric ozone in Japan. Atmospheric Ozone, Proceedings of the Quadrennial Ozone Symposium, C. S. Zerefos and A. Ghazi, Eds., D. Reidel, 754–758.

  • Piotrowicz, S. R., D. A. Boran, and C. K. Fischer, 1986: Ozone in the boundary layer of the equatorial Pacific Ocean. J. Geophys. Res.,91, 13 113–13 119.

  • Sunwoo, Y., and G. R. Carmichael, 1994: Characteristics of background surface ozone in Japan. Atmos. Environ.,28, 25–37.

  • Tsui, W. S., and Y. C. Lee, 1989: Survey of ambient ozone levels in Hong Kong. Pollution in the Metropolitan and Urban Environment Polmet 89, J. Boxall, Ed., Elsevier Science, 430–434.

  • Tsuruta, H., K. Shinya, T. Mizoguchi, and T. Ogawa, 1989: Seasonal behavior of the troposphere ozone in rural Japan. Ozone in the Atmosphere, Deepak, 433–436.

  • Ueda, H., and G. R. Carmichael, 1995: Formation of secondary pollutants during long-range transport and its contribution to air quality in east Asia. TAO,6, 487–500.

  • UKPORG, 1990: Ozone in the United Kingdom. AEA Harwell Laboratory, United Kingdom Photochemical Oxidants Review Group Final Rep. 169 pp. [Available from AEA Harwell Laboratory, Didcot OX11 ORA United Kingdom.].

  • van Aalst, R. M., 1989: Ozone and oxidants in the planetary boundary layer. Atmospheric Ozone Research and Its Policy Implications, Elsevier Science, 573–587.

  • Yeung, K. K., W. L. Chang, B. Wan, F. Kimura, and T. Yoshikawa, 1990: Simulation of boundary layer flow in Hong Kong. Atmos. Environ.,25A, 2161–2172.

Fig. 1.
Fig. 1.

Map showing the air-quality monitoring stations in Hong Kong and some nearby locations.

Citation: Journal of Applied Meteorology 37, 10; 10.1175/1520-0450(1998)037<1153:SOPIHK>2.0.CO;2

Fig. 2.
Fig. 2.

Monthly variations of averaged ozone and carbon monoxide concentrations in Hong Kong.

Citation: Journal of Applied Meteorology 37, 10; 10.1175/1520-0450(1998)037<1153:SOPIHK>2.0.CO;2

Fig. 3.
Fig. 3.

Typical synoptic charts representing the six types of pressure patterns in Hong Kong.

Citation: Journal of Applied Meteorology 37, 10; 10.1175/1520-0450(1998)037<1153:SOPIHK>2.0.CO;2

Fig. 4.
Fig. 4.

Typical back air trajectory for (a) summer monsoon (30 May 1996) and (b) winter monsoon (13 Dec 1996).

Citation: Journal of Applied Meteorology 37, 10; 10.1175/1520-0450(1998)037<1153:SOPIHK>2.0.CO;2

Fig. 5.
Fig. 5.

Windroses for Waland Island.

Citation: Journal of Applied Meteorology 37, 10; 10.1175/1520-0450(1998)037<1153:SOPIHK>2.0.CO;2

Fig. 6.
Fig. 6.

Diurnal variation of averaged ozone concentrations in four typical months.

Citation: Journal of Applied Meteorology 37, 10; 10.1175/1520-0450(1998)037<1153:SOPIHK>2.0.CO;2

Fig. 7.
Fig. 7.

Diurnal variation of averaged O3, NO2, and NO concentrations in July and October.

Citation: Journal of Applied Meteorology 37, 10; 10.1175/1520-0450(1998)037<1153:SOPIHK>2.0.CO;2

Table 1.

Monthly normal of meteorological elements from 1961–90 observed at ROHK.

Table 1.
Table 2.

Annual air pollution emission inventory in Hong Kong in 1994 (tonnes).

Table 2.
Table 3.

Ozone monitoring stations in Hong Kong.

Table 3.
Table 4.

Surface pressure patterns and associated weather in Hong Kong and its vicinity.

Table 4.
Table 5.

Monthly frequency distribution of surface pressure pattern types in Hong Kong (1994).

Table 5.
Table 6.

Monthly percentage frequency distribution of atmospheric stability at the International Airport of Hong Kong (1979–84).

Table 6.
Save