Numerical study of boundary layer ozone transport and photochemical production in east Asia in the wintertime



[1] The transport and chemical transformation of boundary layer ozone in east Asia in January 1997 was investigated using the Models-3 Community Multiscale Air Quality (CMAQ) modeling system with meteorological fields calculated by the Regional Atmospheric Modeling System (RAMS). The predicted ozone concentrations were compared with ground-level observations at four remote sites in Japan and it was found that the model reproduces many of the important features in the observations. Examination of several high ozone episodes indicates that elevated ozone levels are found in association with continental outflow. Net ozone production is found to occur during the daytime over the lower marine atmosphere, and this is also evident from the observations, demonstrating the critical role of the rapid transport of ozone precursors from the Asian continent. An analysis of the ozone budget indicates that the supply and loss of boundary layer ozone in east Asia in winter are dominated by photochemistry.

1. Introduction

[2] East Asia is a region of the world with large and rapidly increasing anthropogenic emissions of nitrogen oxides (NOx), sulfur dioxide (SO2), and other species. Recent research suggests that tropospheric ozone concentrations have increased in the lower troposphere over east Asia in recent decades and that the rate of increase is larger than over other areas of the northern mid-latitudes [Oltmans et al., 1998; Lee et al., 1998]. In contrast to Europe, boundary layer ozone in the 1990's was still continuing to increase in this area, and this is probably due to the increase in emissions of precursors on a regional scale [Lee et al., 1998].

[3] Boundary layer ozone has several sources, including photochemical production from anthropogenic and natural precursors, the most important being NOx, hydrocarbons and carbon monoxide (CO). The distribution of these precursors, the ozone generation from them, and the governing transport mechanisms are complex and nonlinear. A three dimensional numerical model provides a means of estimating the importance of the key factors which determine ozone concentrations.

[4] The Models-3 Community Multi-scale Air Quality (CMAQ) modeling system has been developed by the U.S. Environmental Protection Agency for urban and regional scale air quality simulation [Byun and Ching, 1999]. We apply this system with meteorological data provided by the Regional Atmospheric Modeling System (RAMS) [Pielke et al., 1992] to elucidate the processes determining the temporal and spatial distribution of ozone in east Asia. This is the first quantitative study in which the CMAQ system has been used to simulate tropospheric ozone in east Asia, and we focus on wintertime pollutant outbreak episodes.

2. Model Description

[5] The model used in this study has two major components: RAMS and CMAQ. The study domain (shown in Plate 1) is 4000 × 4000 km2 on a rotated polar stereographic map projection with 80 × 80 km2 grid resolution. This region has dramatic variations in topography and land type, with mixtures of industrial and urban centers and rural agricultural regions. There are 23 vertical σz layers unequally spaced from the ground to 19.5 km, with about 9 layers concentrated in the lowest 2 km of the atmosphere in order to resolve the planetary boundary layer.

Figure Plate 1..

Monthly mean wind vector (in m/s) in the surface layer (∼100 m) and ozone mixing ratios (in ppbv) in the boundary layer (∼1 km). Also shown are the locations of (1) Oki, (2) Fukue, (3) Shanghai, (4) Okinawa, and (5) Hateruma.

[6] The three-dimensional meteorological fields for RAMS are obtained from the European Center for Medium-range Weather Forecasts (ECMWF) 1° × 1° reanalysis data sets, and are available every six hours. Sea Surface Temperatures (SST) for RAMS are based on weekly mean values.

[7] The CMAQ modeling system is a multi-scale (urban and regional) and multi-pollutant (ozone, aerosol, and acid deposition) air quality model. This system can be configured with various degrees of complexity and choices of optional science modules. In this study it utilizes meteorological data provided by RAMS and the Carbon Bond mechanism version 4 (CBM-IV) with isoprene chemistry, modified to provide the necessary linkages for aerosol and aqueous chemistry processes [Byun and Ching, 1999].

[8] For CMAQ, the anthropogenic emissions of NOx, CO, and nonmethane hydrocarbon carbons (NMHC) were obtained from the Emission Database for Global Atmospheric Research (EDGAR) 1° × 1° annual global inventory [Oliver et al., 1996]. NOx emissions from soils and natural hydrocarbon emissions were obtained from the Global Emissions Inventory Activity (GEIA) 1° × 1° monthly global inventory [Benkovitz et al., 1996] for the month of January. NMHC emissions were apportioned appropriately among the lumped-hydrocarbon categories used in CBM-IV.

[9] Initial and boundary conditions were chosen to reflect the east Asian situation. Recent measurements were used wherever possible [Uno et al., 1997; Talbot et al., 1997]. To specify the initial conditions, concentration values were generally chosen at the lower end of their observed range so as to allow the emissions and chemical reactions to bring them closer to their actual values during the 5-day initialization period [Liu et al., 1996; Carmichael et al., 1998].

3. Model Results and Discussion

[10] The simulation period covers December 27, 1996 to January 31, 1997, but model results are discussed for the last 31 days. In the wintertime a strong monsoon is predominant in east Asia due to strong pressure gradients between the Siberian (Continental) High and Okhotsk (Maritime) Low. The monthly averaged wind flow pattern shown in Plate 1 indicates the typical pollutant transport pathway. A lot of pollutants are transported to the West Pacific region along this path. The typical winter monsoon usually lasts a few days, and after this, a low pressure system develops and moves from the continent of China to the east side of Japan. These changes in the large scale weather pattern are responsible for pollutant transport from the Asian continent to the West Pacific region. Such a transition is usually observed 1–2 times per week, and accounts for the intermittent ozone peaks seen in the observations in Figure 1.

Figure 1.

Comparison between predicted hourly average ozone mixing ratios (solid lines, ppbv) for the lowest model layer (∼50 m above ground) and observed ground-level hourly mean ozone concentrations (dots, ppbv) at (a) Fukue, (b) Okinawa, (c) Hateruma, and (d) Oki.

[11] Plate 2 shows the monthly average net ozone production rate in the boundary layer (∼1 km) for January 1997. From the plate, we can see that net ozone production generally occurs over southern China (with maximum values of 20 ppbv/day) except in the main industrial and urban centers such as Shanghai, Hong Kong, Seoul and Tokyo. The highest emission intensities of NOx are found around these areas, and this may lead to net O3 destruction due to direct chemical removal by nitric oxide. Net ozone production in continental regions north of 35°N is very low due to the low insolation and photochemical activity in the wintertime. Ozone destruction generally exceeds ozone production over remote marine areas because of low NOx concentrations. Note that slight net production over marine regions north of 25°N is produced due to the rapid transport of ozone precursors from the Asian continent, and is seen most clearly in downwind of North China and Japan. Net ozone production has been found to take place in the lower marine troposphere by Crawford et al. [1997] in an analysis of PEM-West B data. The distribution pattern of net ozone production over east Asia is consistent with the MOZART 1 simulation [Mauzerall et al., 2001] for January.

Figure Plate 2..

Average horizontal fluxes of ozone (in mole/m2/s shown as arrows) in the period of January 25–28 in the surface layer (∼100 m) and monthly mean net ozone production (in ppbv/day) in the boundary layer (∼1 km).

[12] Also shown in Plate 2 are the average horizontal fluxes of O3 in the period January 25–28 in the surface layer. This period was characterized by passage of a high-pressure system over southern China. In south China, O3 flowed clockwise out to sea then returned to the southwest and moved southwards into the South China Sea. The transport processes during this period are described below. Analysis of vertical cross-sections shows that the strongest continental outflow in the boundary layer occurs between 25 and 35°N.

[13] Figure 1 shows the time variations of hourly averaged ozone mixing ratios measured at four remote Japanese sites Fukue (Figure 1a), Okinawa (Figure 1b), Hateruma (Figure 1c), and Oki (Figure 1d). Also shown are the results from the model for the lowest model layer, approximately 50 m above the ground. The locations of the observation sites are shown in Plate 1. From Figure 1 we can see that in most cases the model reproduces ozone concentrations reasonably well. For example, elevated ozone concentrations in both observations and predictions are clearly seen at Okinawa and Hateruma around January 13, at Fukue and Oki around January 21, and at all the four observation sites around January 28.

[14] Shown in Figure 2 are the contours of ozone concentration and wind vectors at 1200 JST (Japan Standard Time) and a height of ∼50 m above the ground on January 12, 20, and each day for the period January 25–28. Close examination of Figure 2 shows that these high ozone episodes were associated with continental outflow. For example, the high ozone episode during the period January 25–28 was associated with the occurrence of a cold air outbreak. At 1200 JST on January 25, ozone concentrations above 60 ppbv were located over the central part of China near Wuhan city (30°N, 114°E) in Hubei Province (cf. Figure 2c). The maximum ozone levels reached 75 ppbv. This polluted air mass moved eastward, and the center was located east of the estuary of the Yangtze River (30°N, 120°E), adjacent to Shanghai, 24 hours later, with the 50 ppbv contour line extending to Kyushu Island in Japan (cf. Figure 2d). Because of enhancement from local emissions of ozone precursors, the maximum ozone levels reached 78 ppbv. By January 27, the center of the air mass had moved southeast over the East China Sea, while the area enclosed by the 50 ppbv contour line covered the southern part of Japan and the Okinawa area (cf. Figure 2e). The maximum ozone levels decreased to 69 ppbv as the air mass flowed out over the sea, which reflects the decrease in local emissions of ozone precursors. In the next 24 hours the central region enclosed by the 60 ppbv contour line moved eastward and lay east of Okinawa Island (cf. Figure 2f).

Figure 2.

Snapshots of flow fields and ozone mixing ratios at the lowest model layer (∼50 m above the ground) and at 1200 JST on (a) January 12, (b) January 20, (c) January 25, (d) January 26, (e) January 27, and (f) January 28. The shaded areas represent regions of values greater than 60 ppbv, and contour lines are 35, 40, 45, 50, 55, and 60 ppbv, respectively.

[15] Analysis of the time series of surface O3 mixing ratios at the four remote Japanese sites and the high O3 episode during January 25–28 indicates that the horizontal distributions of trace species are reasonably well simulated. The sources and sinks of O3, CO, NOx, and PAN (peroxyacetyl nitrate) in the boundary layer below 1 km in an area 400 × 400 km2 around Shanghai, Oki and Okinawa are shown in Table 1. These three sites are chosen to represent source regions of ozone precursors, remote areas readily affected by emissions, and remote areas, respectively. In the table, horizontal transport includes the contributions from horizontal transport and diffusion, vertical transport accounts for vertical transport and diffusion, and net transport is the sum of these terms. Negative values indicate mass fluxes out of the sample column. For the O3 budget, we find that O3 was photochemically produced at all the sites. The contribution from in-situ production is larger than that from transport, which implies that supply and loss from the boundary layer are dominated by photochemistry.

Table 1. Sources and Sinks of O3, CO, NOx, and PAN in Boundary Layer for January 1997 (Unit: 108mole)a
 HTVTNetChemistryEmissionDry Depo.
  • a

    HT and VT stand for horizontal and vertical transport; SH, OK, and ON represent Shanghai, Oki, and Okinawa, respectively.

 SH−10.410.90.59.4 −9.5
O3OK5.5−7.2−1.76.6 −4.7
 ON2.7− −4.2
C OOK111.0−57.1−32.8−0.731.6 
 SH−1.2−0.3−1.51.5 0
PANOK−0.1−0.2−0.30.3 0
 ON0.10.10.2−0.2 0

[16] Around Shanghai, intense emissions lead to transport of CO out of the region both horizontally and vertically, and chemical production of CO from oxidation of NMHC dominated direct destruction. Around Oki and Okinawa, CO was horizontally transported in and vertically transported out. More CO was destroyed in Okinawa than in Oki due to the higher insolation and lower NMHC abundance. NOx has a similar transport pattern, but has a shorter chemical lifetime, so half of NOx emitted around Shanghai was converted to other nitrogen species, e.g., PAN, which were then also transported out. Around Okinawa NOx was chemically produced from these longer-lived species, consistent with the greater chemical activity as seen for CO and O3.

4. Summary

[17] The CMAQ modeling system coupled with RAMS was applied to east Asia to investigate the transport and photochemical production of boundary layer ozone during the wintertime. Comparisons between observations and model calculations indicate that the model is able to reproduce many of the observed features. Examination of elevated ozone episodes at four remote observation sites clearly show the influence of continental anthropogenic emissions, and the calculated net ozone production during the daytime over the lower marine troposphere demonstrate the critical role of the rapid transport of ozone precursors from the Asian continent. An ozone budget analysis indicates that in-situ photochemical production plays a more important role than transport processes in determining the boundary layer ozone levels during the wintertime.


[18] We thank Dr. K. Murano of National Institute for Environmental Studies of Japan and Mr. Kamaya of Nagasaki Prefectural Institute for Health and Environment of Japan for providing ozone measurement data at Fukue Island. Dr. Oliver Wild is gratefully acknowledged for his suggestions on improving the manuscript.