Seasonal variations of the transport of black ... - Wiley Online Library

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 116, D21307, doi:10.1029/2011JD015830, 2011

Seasonal variations of the transport of black carbon and carbon monoxide from the Asian continent to the western Pacific in the boundary layer R. L. Verma,1 Y. Kondo,1 N. Oshima,2 H. Matsui,1 K. Kita,3 L. K. Sahu,4 S. Kato,5 Y. Kajii,5 A. Takami,6 and T. Miyakawa7 Received 19 February 2011; revised 17 August 2011; accepted 21 August 2011; published 10 November 2011.

[1] Continuous in situ measurements of the mass concentration of black carbon (BC) aerosols and mixing ratio of carbon monoxide (CO) were made at Cape Hedo on Okinawa Island, Japan, a remote site located in the East China Sea, from March 2008 to May 2009. For the first time, we show temporal variations of BC and CO at Hedo in Asian outflows throughout the year. Annual average concentrations of BC and CO were 0.29 mg m−3 and 150 ppbv, respectively. The origins of the observed air masses were determined by using 5‐day back trajectories, suggesting that about 51% of the air masses arriving at Hedo were from the Chinese region during spring and winter, while about 78% of air masses were of maritime origin during summer. Because of the more frequent transport of Chinese air to Hedo in spring and winter, the average and background concentrations of BC and CO in these seasons were higher by about a factor of 2 than those in summer and fall. Air masses from north China made the largest contributions to elevating the BC levels at Hedo because of the high BC emission rate and frequency of transport. The observed DBC/DCO ratio systematically decreased with the decrease in model‐calculated transport efficiency (TEcal BC). On the basis of this result, we derive region‐specific DBC/DCO ratios by selecting data with TEcal BC > 80%. The annually averaged DBC/DCO ratios for air originated from north and south China were 7.0 ± 3.3 and 7.5 ± 4.6 ng m−3 ppbv−1, respectively, about half the annual BC/CO emission ratio derived from the emission inventory of Zhang et al. (2009). We evaluate the CO emission inventory of Zhang et al. (2009) for China by comparing observed (ground‐based and aircraft) and model‐calculated CO values. The comparison indicates that the CO emissions from China were underestimated by about a factor of 2. Citation: Verma, R. L., Y. Kondo, N. Oshima, H. Matsui, K. Kita, L. K. Sahu, S. Kato, Y. Kajii, A. Takami, and T. Miyakawa (2011), Seasonal variations of the transport of black carbon and carbon monoxide from the Asian continent to the western Pacific in the boundary layer, J. Geophys. Res., 116, D21307, doi:10.1029/2011JD015830.

1. Introduction [2] Incomplete combustion of fossil fuels and biomass burning are major sources of black carbon (BC) aerosols. BC particles absorb incoming solar radiation, acting as a direct radiative forcing agent [Intergovernmental Panel on Climate Change (IPCC), 2007; Ramanathan and 1 Department of Earth and Planetary Science, Graduate School of Science, University of Tokyo, Tokyo, Japan. 2 Meteorological Research Institute, Tsukuba, Ibaraki, Japan. 3 Faculty of Science, Ibaraki University, Mito, Japan. 4 Physical Research Laboratory, Ahmedabad, India. 5 Division of Applied Chemistry, Faculty of Urban Environmental Sciences, Tokyo Metropolitan University, Tokyo, Japan. 6 National Institute for Environmental Studies, Tsukuba, Japan. 7 Research Center for Advanced Science and Technology, University of Tokyo, Tokyo, Japan.

Copyright 2011 by the American Geophysical Union. 0148‐0227/11/2011JD015830

Carmichael, 2008, and references therein]. BC, as an indirect radiative forcing agent, interacts with cloud microphysical processes by changing the size distributions and lifetimes of cloud droplets [Kristjansson, 2002; Haywood and Boucher, 2000; Chuang et al., 2002; Conant et al., 2002]. In addition, BC has a deleterious effect on human health [Lighty et al., 2000; Jansen et al., 2005; Alessandrini et al., 2006; Suglia et al., 2007]. [3] In an estimate by Bond et al. [2004], the amount of BC emitted from East Asia is about 30% of the total global anthropogenic BC emissions. BC emissions in this region from anthropogenic sources are increasing because of rapid industrial and economic growth [Streets et al., 2003; Ohara et al., 2007; Zhang et al., 2009]. An estimate by Ohara et al. [2007] suggests that overall BC emissions in Asia may have increased by 5% in 2010 from the level in 2000. It has been suggested that large BC emissions from Asia can have substantial impacts on climate in Asia [Menon et al., 2002;

D21307

1 of 22

D21307

VERMA ET AL.: TRANSPORT OF BC AND CO FROM EAST ASIA

Ramanathan et al., 2001; Ramanathan and Carmichael, 2008; Naidu et al., 2009] in addition to their effects on the global climate [IPCC, 2007]. [4] A large fraction of carbon monoxide (CO) is produced from incomplete combustion of fossil fuels and biomass, similarly to BC. This, together with a typical lifetime of a few weeks at midlatitudes [Weinstock, 1969; Koch et al., 2009 and references therein], makes CO as a useful tracer for quantitative investigations of transport processes of BC, especially when the emission rate of CO is well known. [5] Streets et al. [2003] estimated the CO emission rate in East Asia for the year 2000. However, this inventory was assessed to be underestimated by about a factor of 2 [Carmichael et al., 2003; Allen et al., 2004; Tan et al., 2004; Suntharalingam et al., 2004; Palmer et al., 2003]. Zhang et al. [2009] updated the rate of CO emissions from anthropogenic sources in East Asia for the year 2006. However, this inventory still needs to be validated. [6] Accurate long‐term measurements of BC and CO are necessary in assessing the transport of BC and CO and the effects of BC on wide regions downstream of the large Asian sources. An evaluation of transport patterns and characterization of East Asian air masses for aerosols and trace gasses have been made using aircraft [Verma et al., 2009; Carmichael et al., 2003; Hatakeyama et al., 2001, 2004] and surface measurements [Sahu et al., 2009; Koga et al., 2008; Takami et al., 2007; Li et al., 2007; Matsumoto et al., 2003]. Other studies have discussed the transport patterns of CO and ozone for this region [Narita et al., 1999; Kajii et al., 1998; Pochanart et al., 1999; Kato et al., 2004]. However, detailed studies of the transport of BC and CO, including their seasonal variation, are limited mainly because of the lack of long‐term reliable BC and CO measurements in Asian outflows. In order to overcome this difficulty, we made accurate measurements of BC and CO using the Continuous Soot Monitoring System (COSMOS) [Miyazaki et al., 2008; Kondo et al., 2009, 2011a] and Non‐ Dispersive Infrared Absorption (NDIR)‐based instrument, respectively, at Cape Hedo (hereafter referred to as Hedo) on Okinawa Island, Japan, for 15 months (March 2008 to May 2009). [7] In this study, we focus on four issues based on the analysis of BC and CO at Hedo: (1) elucidation of the seasonal variations of the concentrations of BC and CO in relation to the variations in transport pathways from the Asian continent and wet deposition of BC during transport, (2) derivation of region‐specific DBC/DCO ratios for air least impacted by wet deposition to estimate BC‐CO emission ratios, (3) estimation of the seasonal variation of the transport efficiency of BC, and (4) the evaluation of the CO emission inventory of Zhang et al. [2009]. An evaluation of the BC emissions of Zhang et al. [2009] was made by Kondo et al. [2011b] using the same data set in combination with regional‐scale chemical transport models (CTMs).

2. Measurements 2.1. Hedo 2.1.1. Observation Site [8] Continuous in situ measurements of the mass concentration of BC and mixing ratio of CO were made at Hedo

D21307

(26.87°N, 128.26°E) on Okinawa Island, Japan, in the East China Sea from March 2008 to May 2009. The observation periods are classified as spring (March–May), summer (June–August), fall (September–November), and winter (December–February). We obtained observational data from two springs, and one each from summer, fall, and winter. As shown in Figure 1, Hedo is located on the mainland of Okinawa Island, approximately 60 m above mean sea level. Hedo is an ideal observational site for studies of long‐range transport of BC and CO for the East Asian region. As suggested by Suthawaree et al. [2008, and references therein] and Takami et al. [2005, 2007], there are no large local sources around the observation site. 2.1.2. Experimental Setup [9] BC mass concentrations in the fine mode (PM2.5, i.e., particles with aerodynamic diameters smaller than 2.5 mm) were measured using a filter‐based absorption photometer, COSMOS [Miyazaki et al., 2008; Kondo et al., 2009, 2011a]. The instrument monitors changes in transmittance across an automatically advancing quartz fiber filter tape at 565 nm wavelength (l). The changes in the transmittance are converted to BC mass concentrations using the mass absorption cross section (MAC) determined by comparison with a single‐particle soot photometer (SP2) based on the laser‐ induced incandescence technique (LII) and the thermal‐ optical transmittance (TOT) techniques [Kondo et al., 2009, 2011a]. The stability of MAC is achieved by removing volatile aerosol components before BC particles are collected on filters with the use of an inlet heated at 400°C. The BC mass concentrations are given in the units of mass per unit volume of air (mg m−3) at standard temperature and pressure (STP; 273.15 K and 1013 hPa). The accuracy of the BC mass concentrations measured by COSMOS has been estimated to be about 10% by comparison with those measured by SP2 and TOT instruments [Kondo et al., 2009, 2011a]. BC was measured with an integration time of one minute. The estimated lower limit of detection (LOD) was about 0.047 mg m−3. We used one‐hour average BC data for the present analysis. [10] The interference of coarse particles (e.g., dust) on the BC measurements was negligibly small on average in previous measurements in Asia [Kondo et al., 2009]. We confirmed this point by correlating the artifact component of BC (BCART) with coarse particles (DPM = PM10 − PM2.5). Mass concentrations of the particulate matter with size cutoffs at 2.5 mm (PM2.5) and 10 mm (PM10) were also measured at the Hedo observation site using a Tapered Element Oscillating MicroBalance (TEOM, RP1400) with an accuracy of 7% and Beta x‐ray Absorption (Thermo‐ FH62‐C14) with an accuracy of 10%, respectively, for the same period of the measurements of the BC and CO. For this analysis, we have used the PM2.5 and PM10 observed during winter and spring, because the lidar backscattering coefficients at 532 nm and 1064 nm and the depolarization ratio at 532 nm at Hedo suggested that the dust storms originating over the continent were expected to influence BC measurements at Hedo during these seasons (Yumimoto et al. [2008]; Sugimoto et al. [2011]; http://www-lidar.nies. go.jp/SKYNET/index.html). [11] The observed BC concentrations were highly (r2 = 0.78) correlated with PM2.5 for DPM < 50 mg m−3 (not shown). Because of the high correlation, we assumed that

2 of 22

D21307


D21307

Figure 1. Location of the Cape Hedo observatory. The scale represents the annual anthropogenic emissions of BC (Gg yr−1 grid−1) for the year 2006 with a spatial resolution of 0.5° × 0.5° by Zhang et al. [2009]. The regions north China (NC), south China (SC), Korea (KR), and Japan (JP) defined in this study are shown with solid white lines. BC can be approximately expressed as a function of PM2.5, namely, [BC*] = (0.0295 × [PM2.5]) – 0.098. BCART was estimated by subtracting [BC*] from the observed BC concentrations ([BCobs]). Namely, [BCART] = [BCobs] – [BC*]. [12] Figure 2 shows the correlation between BCART and coarse particles (DPM). In Figure 2, we also added the average (±1 standard deviation (SD)) and median values of BCART calculated within each 5 bins of DPM. As can be seen in Figure 2, no correlation (r2 = 0.00) was observed between BCART and DPM. The average and median values also remained close to zero for the entire range of DPM. Thus, the interference of coarse particles on the BC measured by the COSMOS at Hedo was assessed to be negligible. [13] CO was measured continuously using a commercial instrument TECO (Thermo Environmental Instruments, Inc.) Model 48C, based on the NDIR technique [Suthawaree et al., 2008]. The instrument was calibrated at least once per year using a 1.80 ppmv CO standard gas (Nippon Sanso Inc.). Changes in the calibration factor between each calibration did not exceed 5%. Zero air generated by a TECO Model 96 was sampled every hour for 20 min to correct for the drift of the background signal of the instrument. The measurement uncertainty was 8 ppbv or 6%, whichever was greater. 2.2. Aircraft [14] The Aerosol Radiative Forcing in East Asia (A‐ FORCE) aircraft measurement campaign was conducted over East Asia during March–April 2009, covering a region from 26°N to 38°N. The objectives of this campaign were to investigate transport and removal processes of aerosols, their physical and chemical properties, and cloud microphysical properties in Asian outflows. Details of the A‐

FORCE campaign are given elsewhere [Kondo et al., 2011b; N. Oshima et al., Wet removal of black carbon in Asian outflow: Aerosol Radiative Forcing in East Asia (A‐FORCE) aircraft campaign, submitted to Journal of Geophysical Research, 2011]. [15] During the A‐FORCE campaign, we measured the mixing ratios of CO using a Vacuum Ultraviolet (VUV) Resonance Fluorescence instrument (AL5002, Aero‐Laser GmbH) with a time resolution of 1 s [Gerbig et al., 1999]. Calibration of the instrument was made regularly and the accuracy of the instrument was estimated to be 2%, with

Figure 2. Correlation of nonanthropogenic component of BC (BCART) with coarse particles (DPM) (dust). See text for detailed descriptions of BCART and DPM.

3 of 22

D21307

D21307


Table 1a. Emissions of BC and CO (Gg/yr) in East Asian Regionsa Industrial

Power

Residential

Transportation

Total

Region

BC

CO

BC

CO

BC

CO

BC

CO

BC

CO

North China (NC) South China (SC) Japan (JP) Korea (KR)

339 (36) 227 (27) 8.8 (15) 6.3 (14)

40598 (48) 32931 (40) 1181 (20) 1243 (41)

21 (2) 14 (1.7) 0.14 (0.3) 0.42 (0.9)

1270 (1.5) 1105 (1.4) 14.6 (0.3) 4.5 (0.2)

477 (51) 504 (59) 7.4 (13) 22 (47)

25838 (30) 29525 (36) 463 (7.9) 1037 (34)

106 (11) 106 (12) 41 (72) 18 (38)

17653 (21) 18578 (23) 4207 (72) 738 (24)

943 (100) 851 (100) 57 (100) 47 (100)

85360 (100) 82139 (100) 5866 (100) 3025 (100)

a From Zhang et al. [2009]. The emission quantities of BC and CO given in this table are confined to the domain regions selected in Figure 1. The value given in parentheses is the percentage contribution.

a precision of about 0.5% for 10 s averaged data. We use 1 min averaged CO data in section 7.

3. BC and CO Emissions in East Asia [16] Estimates of BC and CO emissions were made for the East Asian region for the year 2000 with a spatial resolution of 1° × 1° to support the atmospheric modeling and analysis of observations taken during the NASA Transport and Chemical Evolution over the Pacific (TRACE‐P) and Asia Pacific Regional Aerosol Characterization (ACE‐Asia) missions [Streets et al., 2003]. The inventory was updated by Zhang et al. [2009] for the year 2006 with a spatial resolution of 0.5° × 0.5° to support the NASA Intercontinental Chemical Transport Experiment‐Phase B (INTEX‐ B). The Zhang et al. [2009] inventory mainly emphasized the estimation of BC and CO from anthropogenic sources, namely, the industry, power, residential, and transportation sectors. Figure 1 shows the spatial distribution of BC (Gg yr−1 grid−1) for the East Asian region (10°N–53°N to 90°E– 150°E) [Zhang et al., 2009]. The spatial distribution of CO was similar to that of BC (not shown). The East Asian region (Figure 1) considered for this study was divided into north China (NC: 33°N–40°N, 100°E–123°E and 40°N– 50°N, 100°E–130°E), south China (SC: 20°N–33°N, 100°E– 123°E), Korea (KR: 33.5°N–40°N, 123°E–129.5°E), and Japan (JP: 30°N–46°N, 130°E–147°E), because these regions are the major sources of BC. Division of the Chinese region into NC and SC is rather arbitrary. As can be seen in Figure 1, most BC emissions are located in highly populated urban and industrial regions of China [Zhang et al., 2009; Streets et al., 2003; Cao et al., 2006]. The transport of BC to the observation site varied depending on the pattern of transport, reflecting the distribution of BC sources. [17] Table 1a summarizes the sector‐wise BC and CO emissions from each region within East Asia. The quantities of the emissions of BC and CO given in Table 1a are confined only to the domain regions shown in Figure 1. The total BC emissions from East Asia, comprising China (NC + SC), Japan, and Korea, are about 1.90 Tg (1 Tg = 109 g). About 95% (1.79 Tg) is contributed by the Chinese region. In China (both NC and SC), a large fraction of BC is emitted from the residential (∼55%) and industrial (∼32%) sectors, while a smaller fraction is from transportation (∼12%) and power (∼2%). In Japan, the transportation sector (∼72%) is a major source of BC, and in Korea, the residential (∼47%) and transportation (∼38%) sectors are the most important sources. However, these inventories are subject to an uncertainty of 208% for BC for China [Zhang et al., 2009]. [18] Kondo et al. [2011b] estimated BC emissions for the whole of China, and the estimated value was about 1.92

(±40%) (Tg/yr), a value close to that reported by Zhang et al. [2009] for China. According to Li et al. [2007], BC emissions from China may vary from 1.3 Tg to 2.6 Tg. The impact of biomass burning (BB) at the Hedo observation site may not be significant, since the ratio of the BC emissions from BB to the total (anthropogenic + BB) estimated using the inventory of Streets et al. [2003] is only 0.13 in East Asia [Kondo et al., 2011b]. Also, BB is mainly confined to 10°N–25°N and 90°E–110°E [Allen et al., 2004], from which the Hedo observation site received hardly any air masses. Air mass transport is discussed in section 4.3. [19] As discussed above, Streets et al. [2003] estimated the CO emission rate in China for the year 2000 to be 116 Tg/yr, including the contributions from the combustion of fossil fuels, biofuels, and biomass (TRACE‐P inventory). This inventory was updated by Zhang et al. [2009] for INTEX‐B for the year 2006. Table 1b summarizes the changes in the estimates of BC and CO emissions in China from the year 2000 to 2006. The BC emissions from China estimated by Zhang et al. [2009] for the year 2006 agreed well with those estimated from the BC observations at Hedo [Kondo et al., 2011b]. [20] Evaluations of the TRACE‐P CO inventory [Streets et al., 2003] were made by Carmichael et al. [2003], Allen et al. [2004], Tan et al. [2004], Suntharalingam et al. [2004], and Palmer et al. [2003] using 3‐D CTMs. These studies, based on the comparison of model‐calculated CO with that observed in the Asian continental outflows, suggested that the total CO emissions from China for TRACE‐P [Streets et al., 2003] were underestimated by about a factor of 2, possibly because of the underestimation of CO emissions from combustion of biofuels and domestic and industrial coal [Carmichael et al., 2003; Tan et al., 2004] and old power plants [Suntharalingam et al., 2004]. On the basis of these suggestions, Streets et al. [2006] reestimated the CO emission rate in China to be 158 Tg/yr, about 36% higher than their previous estimate, by incorporating updated statistical fuel data and improving the emission factors for the domestic and industrial sources. Table 1b. Emission Rates of BC and CO (Tg/yr) in China Reference

Year

BC

CO

Streets et al. [2003] Ohara et al. [2007] Ohara et al. [2007] Streets et al. [2006] Zhang et al. [2009] Ohara et al. [2007] Zhang et al. [2009]

2000 2000 2001 2001 2001 2003 2006

1.05 1.09 1.10 NAa 1.60 1.14 1.81

116 137 141 158 142 158 167

4 of 22

a

NA, not available.

D21307


D21307

Figure 3. Spatial distribution of precipitation (1° × 1°) in the East Asian region for spring, summer, fall 2008, and winter 2009. Seasonally averaged precipitation data from the Global Precipitation Climatology Project (GPCP) were used. [21] Ohara et al. [2007] estimated the CO emission rate in China from the year 1995 to 2003 from anthropogenic sources, excluding the contribution from biomass burning, as summarized in Table 1b. The CO emission rate increased by about 15% between the year 2001 and 2003, according to their estimate. [22] Zhang et al. [2009] also estimated the increase in the rate of CO emissions from the anthropogenic sources in China to be about 18% from the year 2001 to 2006 (Table 1b). However, this estimate still needs to be validated. [23] We made a detailed comparison of the CO observed at Hedo (ground‐based) and aircraft (A‐FORCE campaign) with that calculated using the three‐dimensional (3‐D) Community Multiscale Air Quality (CMAQ) model using the CO inventory of Zhang et al. [2009]. The calculation of CO using CMAQ is discussed in section 4.5, and the evaluation of the CO emission inventory of Zhang et al. [2009] using the ground‐based and aircraft measurements is discussed in section 7.

4. Meteorology of Data Analysis 4.1. Precipitation [24] Figure 3 shows the spatial distribution of the seasonally averaged precipitation level for the East Asian region, expressed in millimeters per day per grid box (mm d−1 grid−1). Data from the Global Precipitation Climatology Project (GPCP) were used to estimate daily precipitation

levels with a horizontal resolution of 1° × 1° [Huffman et al., 2001; Adler et al., 2003; Matsui et al., 2011]. As can be seen in Figure 3, the level of precipitation was higher over the continent, especially over south China, during spring, summer, and fall. The precipitation level was lowest during winter. Thus, it is anticipated that the transport of BC from the Asian continent to Hedo may generally have been influenced by the seasonal variations in precipitation through wet deposition. We investigate this point in section 6. 4.2. Pressure and Winds [25] The flow of air masses is governed by the wind field, created by pressure gradients. Figure 4 shows maps of the seasonal variation of average sea level pressure (hPa) with wind vectors (u, v) plotted using data from the National Centers for Environmental Prediction (NCEP) Final (FNL) operational global analysis data with a resolution of 1° × 1°. As can be seen in Figure 4, the regions of high pressure were located over northwestern China and Mongolia, centered at about 50°N, 90°E, during winter, spring, and fall, with low pressure over the western Pacific Ocean. On the other hand, the pressure pattern in summer was the opposite of that in winter and spring. The pressure maps suggest the dominance of air mass flow from the Asian continent to the western Pacific during winter, spring, and fall, whereas in summer, the wind flow pattern is from the western Pacific to the continent.

5 of 22

D21307


D21307

Figure 4. Seasonal mean sea level pressure (hPa) over East Asia and mean horizontal winds (u, v). National Centers for Environmental Prediction (NCEP) Final (FNL) operational global analysis data with a resolution of 1° × 1° were used for the pressure and winds. 4.3. Backward Trajectories [26] To study the transport pattern influencing the measurement site, we analyzed 5 day back trajectories at hourly intervals, each starting from 26.87°N, 128.26°E (Hedo location) at 950 hPa pressure height (about 500 m), based on the method described by Tomikawa and Sato [2005] using the NCEP FNL data. Figure 5 shows the trajectories and relative humidities (RHs) from the NCEP FNL data along the trajectories. One trajectory per day is shown in Figure 5, each starting at 1200 LT or 300 UTC. The general transport pathways shown by the back trajectories are consistent with those predicted from the mean general wind fields (shown in Figure 4). Namely, the majority of air masses were transported from the Asian continent (Mongolia and north China)

during winter and spring. However, air masses from the western Pacific Ocean prevailed for most of the period in summer. [27] Fall marked the transition of the two seasons, with extreme climatic conditions in summer and winter. At the beginning of fall (September and early October), the climatic conditions of summer prevailed, making the air masses flow from the Pacific region to the Asian continent. Later, as fall progressed, the origins of the air masses shifted to continental. Reflecting this, the origins of air masses sampled at Hedo in fall typically alternated between maritime and continental. Irrespective of the season, the air masses arriving from the Chinese region were transported from higher altitudes (not shown) and were dry (Figure 5). Air masses arriving from the western Pacific were transported at higher

6 of 22

D21307


D21307

Figure 5. Five day back trajectories, each starting from the location of Hedo (26.87°N, 128.26°E) at 950 hPa pressure height (about 500 m in altitude) and 1200 LT or 300 UTC. Relative humidity (RH) along the trajectories is also shown. These trajectories were calculated using the trajectory model of the National Institute of Polar Research (NIPR, Japan), developed by Tomikawa and Sato [2005]. pressure levels (Figure 5) and were more humid (Figure 5). The pressure scale is not shown in Figure 5. 4.4. Air Mass Classification [28] On the basis of the back trajectory analysis, the origins of air masses were classified as north China (NC), south China (SC), Korea (KR), Japan (JP), Marine (MA), and free troposphere (FT). The air masses were defined as NC, KR, SC, or JP, if their back trajectories originated in the planetary boundary layer (PBL) of NC, KR, SC, or JP. In this study,

we have defined the PBL as being below the 750 hPa atmospheric pressure level ( 24 h) in a specified region were included for the statistics.

4.5. Model‐Calculated CO and Transport Efficiency of BC (TEcal BC) [32] We used hourly results of CO mixing ratios and BC mass concentrations calculated using the CMAQ model [Byun and Ching, 1999; Binkowski and Roselle, 2003], covering the entire period of the measurements at Hedo. We also calculated CO using the CMAQ model for the A‐ FORCE aircraft measurements along the flight tracks. CMAQ was driven by the Weather Research and Forecasting (WRF) model [Skamarock et al., 2005], with a horizontal resolution of 81 km on a Lambert conformal map projection consisting of 117 × 69 grid cells with the center at 30°N and 110°E, covering the entire Asian region [Kondo et al., 2011b]. There were 21 vertical levels from the ground surface up to 100 hPa on terrain‐following coordinates. The SAPRC99 mechanism [Carter, 2000] was used for gas‐ phase chemistry, containing 80 chemical species and 214 chemical reactions. The initial and boundary conditions of CO mixing ratios for the model domain were set to 170, 150, and 125 ppbv in the lower, middle, and upper troposphere, respectively. The initial and boundary conditions for BC mass concentrations were set to zero. The BC and CO

[29] As can be seen in Figure 5, air masses arriving from North China often passed over the Korean region before reaching Hedo, making it difficult to clearly separate the air masses between NC and KR. We have separated NC and KR by estimating the degree of mixing using the RT for their respective regions. The ratio of the RT spent in NC to the total RT was used for this purpose. Rf ðNCÞ ¼ RTðNCÞ=½RTðNCÞ þ RTðKRÞ

ð1Þ

Based on this parameter, we defined KR for Rf (NC) < 0.2, Mixed (NC + KR) for 0.2 < Rf (NC) < 0.8, and NC for Rf (NC) > 0.8. We follow this classification for NC, KR, and mixed (MX) throughout the paper. The definitions of the other categories of air masses (SC, JP, MA, and FT) remain unchanged. To ensure robustness in the analysis, i.e., that it represents the dominant contributions from the categories, we used only data with RT > 24 h in the specified regions for the statistical analyses throughout this study. [30] Figure 6 shows the fractions (%) of air masses of different origins during the measurement period in five seasons. In spring 2008 and 2009, about 34% and 22% of air masses originated from the Chinese region (NC + SC) (Figure 6). If MX air masses are included, 45% and 48% of air masses were substantially influenced by emissions in China in these years.

Figure 7. Correlation between monthly averaged precipitation levels observed by the Global Precipitation Climatology Project (GPCP) and those calculated by the Weather Research Forecasting (WRF) model within 20°N–50°N and 100°E–145°E domain.

8 of 22

D21307


Table 2. Monthly Averaged Precipitation Levels Observed by the Global Precipitation Climatology Project (GPCP) and Estimated by the Weather Research Forecasting Model (WRF) Months

GPCP (mm d−1)

WRF (mm d−1)

March 2008 April 2008 May 008 June 2008 July 2008 August 2008 September 2008 October 2008 November 2008 December 2008 January 2009 February 2009 March 2009 April 2009 May 2009

2.01 2.33 4.76 5.36 4.74 4.49 3.48 2.38 1.93 1.22 1.28 1.25 2.03 2.84 3.40

2.66 2.70 4.15 4.68 4.11 3.83 2.50 1.54 2.04 1.63 2.02 2.21 2.82 3.06 2.85

D21307

the median BC vertical profile obtained by aircraft over the East China Sea in spring 2009 (Aerosol Radiative Forcing in East Asia (A‐FORCE) aircraft campaign) [Kondo et al., 2011b]. [33] The transport efficiency of BC (TEcal BC) was estimated for the East Asian BC emissions using the hourly results of BC mass concentrations calculated by the CMAQ model. Following the method of Kondo et al. [2011b], the TEcal BC value was derived as follows: TEcal BC ð%Þ ¼ ½BCðCMAQ ModifiedÞ=½BCðCMAQ NoWetDepÞ 100

emission inventories of Zhang et al. [2009] with a grid resolution of 0.5° × 0.5° in latitude and longitude were used. Kondo et al. [2011b] made detailed comparisons of BC calculated using CMAQ and the Eulerian, Multiscale Tropospheric Aerosol Chemistry and dynamics Simulator (EMTACS) models [Kajino and Kondo, 2011] with observations. The CMAQ and EMTACS simulations reproduced well the temporal variations of surface BC mass concentrations observed at Hedo during the entire measurement period and

ð2Þ

Here [BC](CMAQ‐Modified) represents the BC mass concentration calculated by the CMAQ baseline simulation using the modified scheme of wet deposition of BC [Kondo et al., 2011b]. [BC](CMAQ‐NoWetDep) represents BC mass concentration calculated by another CMAQ simulation without including the effect of wet deposition [Kondo et al., 2011b]. We used this parameter for the analysis of the effect of wet deposition of BC. [34] We show a comparison of the WRF precipitation simulation with observations (GPCP, see section 4.1) for each month, because the validity of the WRF precipitation is important for calculation of the transport efficiency of BC. Figure 7 shows correlation of the monthly averaged precipitation levels observed by the GPCP with those estimated by the WRF simulation within 20°N–50°N and 100°E– 145°E domain; the numerical data are given in Table 2.

Figure 8. Time series of the hourly BC and CO concentrations during spring and summer of 2008 and winter 2009. Air mass categories are shown with gray horizontal lines: (1) north China (NC), including mixed (MX), (2) south China (SC), and (3) Japan (JP). Maritime (MA) and free tropospheric (FT) air masses are not indicated. 9 of 22

D21307


Figure 9. Time series of the monthly values of BC, CO, DBC/DCO ratios, and the model‐calculated transport efficiency of BC (TEcal BC). Monthly mean DBC/DCO ratios were derived using three methods: (1) the average and (2) median of the (DBC/DCO)i data points, shown with filled circles and pluses, respectively, and the (3) ratio of the sums of (DBC)i and (DCO)i (open circles). The lower and upper extents of the vertical bars represent the 10th and 90th percentiles of the values, respectively. The lower and upper limits of the boxes are the 25th and 75th percentiles, respectively. The precipitation levels observed by the GPCP and calculated by the WRF model are well correlated (r2 = 0.81). The WRF simulation generally reproduced well the monthly averaged precipitation over East Asia during the measurement period with at most 76% of the discrepancy in February 2009.

D21307

Air mass categories, in particular, from the source regions of NC (including MX), SC, and JP, are also shown with horizontal bars in this figure. In general, BC and CO are well correlated throughout the measurement period. [36] Figure 9 shows the monthly mean and median values of BC and CO (Figure 9, first and second panels). The statistical values (mean, median, and background) of BC and CO for each season are summarized in Table 3. The average concentrations of BC and CO observed during spring 2008 and winter 2009 were 0.41 mg m−3 and 0.38 mg m−3, and 193 ppbv and 179 ppbv, and they were 0.12 mg m−3 and 0.18 mg m−3, and 86 ppbv and 115 ppbv in summer and fall, respectively (Table 3). The average concentrations of BC and CO in spring and winter were higher by about a factor of 2 than those in summer and fall, reflecting stronger influence of the continental outflows in winter and spring. In summer and early fall, Hedo was generally covered by relatively clean marine air (Figures 4 and 5). These features are also clearly seen in Figure 9 (first and second panels). [37] Short‐term reduction in the pollution levels because of the implementation of emission control measures, for example, occurred during the 2008 Beijing Olympics (8– 24 August). Mijling et al. [2009] and Yu et al. [2010] reported, on the basis of satellite measurements, that there was a reduction in the tropospheric NO2 column by about 20%– 60% in Beijing and surrounding cities during the Olympic Games. However, we did not observe such impacts on the BC measurements at Hedo, because the transport of the air masses was from the Pacific to the continent during the 2008 Olympics (see Figure 5). [38] The episodic fluctuations of BC and CO concentrations were seen prominently during winter and spring (Figure 8). These fluctuations were associated with transport of high‐BC and CO air masses from the Asian continent, as identified by trajectory analysis (Figure 5). The transport was often associated with the passage of cold fronts and strong winter monsoons [Chien and Kuo, 2006; Liu et al., 2008; Suthawaree et al., 2008; Takami et al., 2005, 2007; Hatakeyama et al., 2001, 2004]. During summer, the fluctuations of BC concentrations were much smaller because of the prevailing inflow of marine air. 5.2. Characterization of the Air Masses [39] Figure 10 shows seasonally averaged values of BC and CO in each category for each season. Table 4 shows more detailed statistics, including the ranges of the variabilities. The BC and CO concentrations were highest in Chinese air masses (NC and SC) for all seasons except summer. Similarly, the concentrations of BC and CO in MX air masses were generally higher than those of KR but lower than those of NC. This indicates the major influence of the NC air masses over KR air with some mixing of NC and KR air. The average concentrations of BC and CO were generally lower in MA and FT air masses than those in other air masses in winter and spring. This suggests that FT air was not strongly influenced by emissions in the Asian continent.

5. Variations of BC and CO

6. DBC/DCO

5.1. Seasonal Variations [35] Figure 8 shows time series of BC and CO observed at Hedo during spring and summer 2008 and winter 2009.

6.1. Seasonal Variations of DBC/DCO and TEcal BC [40] Figure 11 shows time series of the background levels of BC (BCBG) and CO (COBG) for each month. The BCBG

10 of 22

D21307

D21307


Table 3. Average ± Standard Deviation (SD), Median (25th, 75th Percentiles), and Background Concentrations (5th Percentile) of BC and CO Observed at Hedo During Five Seasons BC (mg m−3) Season

Average ± SD

Spring 2008 Summer 2008 Fall 2008 Winter 2009 Spring 2009 All data set

0.414 0.118 0.183 0.379 0.363 0.291

± ± ± ± ± ±

0.377 0.115 0.209 0.500 0.305 0.349

Median 0.307 0.081 0.117 0.228 0.272 0.168

(0.129, (0.044, (0.055, (0.116, (0.129, (0.080,

CO (ppbv) Background

Average ± SD

Median

Background

0.039 0.020 0.025 0.058 0.053 0.029

193 ± 66 86 ± 30 115 ± 47 179 ± 83 175 ± 55 150 ± 72

179 (153, 230) 77 (66, 94) 109 (81, 141) 155 (116, 197) 169 (139, 198) 141 (97, 183)

96 56 52 104 109 61

0.549) 0.151) 0.236) 0.432) 0.558) 0.373)

and COBG values were defined as the 5th percentiles of the monthly observed data. Table 3 summarizes the seasonal levels of BCBG and COBG. The average of BCBG and COBG values for spring 2008 and 2009 and winter 2009 was about 0.05 mg m−3 and 103 ppbv. These were about 0.023 mg m−3 and 54 ppbv from summer and fall. The amplitudes of the seasonal variations of BCBG and COBG values were about a factor of 2. [41] Although both BC and CO are emitted from incomplete combustion, BC has a shorter residence time (5– 11 days) [Koch et al., 2009, and references therein] than CO, since it is removed from the atmosphere through wet deposition. CO is removed from the atmosphere through the reaction with the hydroxyl radical (OH) with a lifetime of about 2 months at northern midlatitudes [e.g., Weinstock, 1969]. Therefore DBC (BC – BCBG) to DCO (CO – COBG) ratios should be proportional to the transport efficiency of BC, if the DBC/DCO ratios in the source regions are stable. [42] Figure 12 shows the temporal variations of the DBC/ DCO ratios and TEcal BC during spring 2008 and winter 2009. The DBC/DCO ratio was derived for each data point. As can be seen in Figure 12, many of the peaks in the DBC/D CO ratios coincided with those of TEcal BC during winter and spring, when a large amount of BC was transported from the source regions of East Asia (Figures 1 and 5). However, no clear correspondence between DBC/DCO and TEcal BC were seen for summer (not shown). [43] Figure 9 (third and fourth panels) shows the variations of the monthly mean DBC/DCO ratio and TEcal BC. The monthly mean values of the DBC/DCO ratios were derived by using three methods: (1) the average and (2) median of the (DBC/DCO)i data points, shown as filled circles and pluses, respectively, and (3) the ratio of the sums of (DBC)i and (DCO)i (open circles). ½DBC=DCOall ¼

X

ðDBCÞi = i

X i

ðDCOÞi :

ð3Þ

The DBC/DCO ratios calculated using the three methods agreed well (Figure 9), supporting the statistical reliability of the different methods. [44] The average DBC/DCO ratios showed large seasonal variations. They were about 3.0–7.0 ng m−3 ppbv−1 in winter and spring. They reached minimum values of about 1.5–2.0 ng m−3 ppbv−1 in late spring and summer. The general pattern of seasonal variation of the DBC/DCO ratios was similar to that of TEcal BC. The large seasonal variation of the average DBC/DCO ratios was driven by the seasonal variations of the dominance of largely different air masses associated with alternating winter and summer

Figure 10. Variations of BC and CO concentrations among various categories in a season, as well as for the intercomparison among seasons.

11 of 22

D21307

D21307


Table 4. Seasonal Statistics, Comprising the Average ± SD, Median (25th, 75th Percentiles), and Background Level (5th percentile) for cal obs obs BC and CO, and [DBC/DCO]all ± d [DBC/DCO]all and Transport Efficiencies of BC (TEcal BC ± dTEBC and TEBC ± dTEBC ) for the Air a Masses Arriving at Hedo From Different Regions Spending at Least 24 h of Residence Time in Each Source Region BC (mg m−3) Number Region of Data

Average ± SD ± ± ± ± ± ± ±

0.423 0.426 0.240 0.165 0.235 0.284 0.230

Mediane

SC NC MX KR JP MA FT

222 232 161 31 292 313 168

0.545 0.855 0.476 0.345 0.244 0.230 0.259

0.383 0.790 0.441 0.312 0.155 0.148 0.154

(0.199, (0.520, (0.320, (0.258, (0.095, (0.056, (0.059,


149 0 0 35 177 1316 15

0.237 ± 0.175 NA NA 0.398 ± 0.276 0.154 ± 0.131 0.078 ± 0.060 0.101 ± 0.081

0.227 (0.108, NA NA 0.319 (0.255, 0.108 (0.060, 0.059 (0.034, 0.091 (0.039,


41 125 194 78 442 563 90

0.437 0.441 0.326 0.236 0.126 0.069 0.235

± ± ± ± ± ± ±

0.592 0.379 0.239 0.120 0.084 0.062 0.144

0.210 0.308 0.285 0.258 0.109 0.049 0.224

(0.132, (0.180, (0.139, (0.141, (0.063, (0.030, (0.093,


95 297 184 26 84 139 312

1.129 0.649 0.393 0.219 0.261 0.343 0.176

± ± ± ± ± ± ±

0.998 0.777 0.333 0.187 0.195 0.328 0.122

1.000 0.365 0.354 0.116 0.176 0.238 0.139

(0.407, (0.209, (0.152, (0.075, (0.135, (0.118, (0.085,


49 159 242 53 148 96 182

0.789 0.656 0.492 0.358 0.252 0.172 0.161

± ± ± ± ± ± ±

0.698 0.372 0.338 0.292 0.202 0.146 0.099

0.600 0.658 0.473 0.252 0.178 0.094 0.143

(0.361, (0.422, (0.234, (0.115, (0.099, (0.071, (0.113,

CO (ppbv) Average Background ± SD

Mediane

[DBC/DCO]allb obs d c Background (ng m−3 ppbv−1) TEcal BC| (%) TEBC | (%)

Spring 2008 (Total Number of Data Points = 1419) 0.836) 0.126 226 ± 86 224 (172, 274) 86 1.098) 0.322 265 ± 65 248 (226, 287) 186 0.550) 0.167 217 ± 53 215 (172, 247) 157 0.383) 0.127 201 ± 31 197 (180, 213) 162 0.298) 0.045 169 ± 24 168 (154, 179) 143 0.307) 0.024 149 ± 60 148 (94, 179) 70 0.412) 0.035 163 ± 52 156 (127, 193) 102 Summer 2008 (Total Number of Data Points = 1692) 0.302) 0.031 112 ± 27 113 (90, 125) 72 NA NA NA NA NA NA NA NA 0.399) 0.036 146 ± 50 136 (123, 165) 60 0.248) 0.034 110 ± 34 99 (84, 136) 66 0.106) 0.016 74 ± 22 70 (62, 80) 55 0.113) 0.028 127 ± 6 126 (123, 130) 120 Fall 2008 (Total Number of Data Points = 1533) 0.362) 0.085 159 ± 57 135 (121, 161) 108 0.600) 0.071 169 ± 39 159 (144, 192) 121 0.412) 0.038 172 ± 53 158 (134, 194) 112 0.315) 0.026 142 ± 29 138 (130, 152) 98 0.171) 0.030 107 ± 22 101 (92, 117) 82 0.088) 0.014 73 ± 25 65 (53, 87) 46 0.336) 0.033 129 ± 23 131 (114, 139) 90 Winter 2009 (Total Number of Data Points = 1137) 1.430) 0.149 307 ± 133 282 (209, 370) 149 0.712) 0.097 234 ± 105 191 (156, 313) 129 0.453) 0.062 205 ± 80 185 (148, 236) 126 0.329) 0.052 225 ± 71 201 (181, 271) 150 0.361) 0.058 146 ± 62 124 (108, 146) 92 0.374) 0.061 169 ± 62 160 (115, 196) 100 0.235) 0.045 132 ± 35 119 (108, 150) 98 Spring 2009 (Total Number of Data Points = 929) 1.007) 0.058 252 ± 114 210 (183, 266) 162 0.883) 0.093 241 ± 81 218 (187, 270) 163 0.659) 0.080 192 ± 46 187 (166, 206) 133 0.592) 0.063 193 ± 38 191 (172, 209) 152 0.335) 0.047 180 ± 34 174 (160, 191) 148 0.236) 0.035 142 ± 31 138 (122, 150) 104 0.173) 0.057 140 ± 19 137 (127, 146) 115

4.28 ± 3.66 5.26 ± 3.11 4.23 ± 4.75 3.26 ± 3.41 2.81 ± 3.59 NA NA

62 ± 25 83 ± 20 76 ± 17 57 ± 18 59 ± 22 NA NA

57 ± 52 73 ± 50 72 ± 100 NA NA NA NA

3.78 ± 2.19 NA NA 4.14 ± 3.00 2.45 ± 3.76 NA NA

60 ± 16 NA NA 47 ± 15 56 ± 13 NA NA

NA NA NA NA NA NA NA

5.12 ± 4.06 4.91 ± 3.11 2.99 ± 1.70 2.93 ± 1.74 2.97 ± 1.50 NA NA

54 ± 21 79 ± 23 52 ± 23 42 ± 15 46 ± 16 NA NA

NA 77 ± 61 NA NA NA NA NA

5.40 ± 2.62 4.71 ± 2.83 3.49 ± 2.93 1.41 ± 2.02 5.06 ± 1.35 NA NA

61 ± 23 72 ± 20 64 ± 21 44 ± 24 60 ± 19 NA NA

73 ± 45 66 ± 49 61 ± 65 NA NA NA NA

5.43 ± 3.23 4.50 ± 2.93 5.45 ± 3.25 4.10 ± 2.66 3.05 ± 2.49 NA NA

49 ± 20 73 ± 20 74 ± 23 63 ± 21 64 ± 19 NA NA

NA 66 ± 48 76 ± 72 NA NA NA NA

a

MX, mixed; NA, not available. [DBC/DCO]all = ∑DBC/∑DCO (equation (3)). d [DBC/DCO]all is the standard deviation (SD). cal c TEcal BC is the model‐calculated transport efficiency of BC, derived from equation (2). dTEBC is the SD. d obs TEobs BC is the observed transport efficiency of BC, derived from equation (4). dTEBC is derived from equation (5). For example, the values of 57 ± 52 for SC in Spring 2008 are given as 57 = 4.28/7.56 × 100 and 52 = 57 × [ (3.66/4.28)2 + (2.65/7.56)2) ]1/2. e Values in parentheses are the 25th and 75th percentiles. b

monsoons. The summer minimum in DBC/DCO ratios was consistent with that in TEcal BC, which was generally caused by wet deposition of BC. 6.2. DBC/DCO–TEcal BC Correlation [45] We investigated the direct correlation between DBC/ DCO and TEcal BC for NC + SC + MX air for winter and spring, as shown in Figure 13. On average, DBC/DCO and TEcal BC were positively correlated; however, the scatter was also substantial, leading to a moderate r2 value of 0.41. The positive correlation between DBC/DCO and TEcal BC indicates that one of the important causes of the variability in the BC mass concentrations was wet removal of BC during transport from China, as partly seen by their correlated temporal variations (Figure 12).

[46] In addition to wet deposition, DBC/DCO at Hedo may change depending on the BC‐CO emission ratios in source regions. In order to investigate this point, we extracted the DBC/DCO data least influenced by wet deposition by selecting the data with TEcal BC > 80%, as has been done by Kondo et al. [2011b]. [47] Figure 14 shows the DBC‐DCO correlations for NC and SC air masses for spring 2008 and winter 2009, after the selection of the data (TEcal BC > 80%). DBC and DCO were highly correlated (r2 = 0.69–0.88). Considering the high correlations, we derived reference ratios, [DBC/DCO]ref, least influenced by wet deposition (TEcal BC > 80%). The [DBC/DCO]ref values were calculated by the ∑i (DBC)i)/∑i (DCO)i ratios, instead of the slopes of the correlations, for consistency in the methodology of the statistical analysis

12 of 22

D21307


D21307

Figure 11. Time series of the monthly background values of BC (BCBG) and CO (COBG). Background levels were given as the 5th percentile of the monthly data set, and then interpolated for each data point. (Table 4). We also calculate the SD of the hourly (DBC/D CO)i data points before and after the data selection (TEcal BC > 80%) for each season, namely, d [DBC/DCO]all and d [DBC/DCO]ref. The [DBC/DCO]all and [DBC/DCO]ref ratios and their SD (i.e., d [DBC/DCO]all and d [DBC/D CO]ref) for each season are summarized in Tables 4 and 5a, respectively. [48] Further, we have also calculated [DBC/DCO]ref with TEcal BC > 90%, summarized in Table 5b. The ratio of [DBC/ cal DCO]ref with TEcal BC > 80% to [DBC/DCO]ref with TEBC > 90% were varied over 0.75–0.94, 0.93–0.99, 0.76–0.97, 0.89–0.92, and 0.79–0.94 for SC, NC, MX, KR, and JP, respectively (Tables 5a and 5b). However, if the number of data points is sufficient, the value of the [DBC/DCO]ref ratios changes by less than 5%. For example, in the case of NC the [DBC/DCO]ref ratios for all data set is 7.00 ng m−3

−3 ppbv−1 at TEcal ppbv−1 at BC > 80%, while it is 7.37 ng m cal cal TEBC > 90%. We used [DBC/DCO]ref with TEBC > 80% for further interpretation. [49] In deriving the [DBC/DCO]ref values, we used TEcal BC as a meteorological parameter representing the effect of wet deposition. In addition to TEcal BC, we used the accumulated precipitation along each trajectory (APT) using the GPCP data (daily data with a resolution of 1° × 1°), discussed in section 4.1. The APT was useful in representing the wet deposition of BC during synoptic‐scale transport from the PBL to the FT [Matsui et al., 2011]. However, the correlation of DBC/DCO with APT was much poorer than that with TEcal BC, despite good agreement between the monthly averaged precipitation levels predicted by the WRF simulation and the GPCP data. The poor correlation for APT may reflect complex processes in controlling the wet deposition

Figure 12. Time series of the variations of the DBC/DCO ratio (excluding DCO < 20 ppbv) and model‐ calculated transport efficiency (TEcal BC) during spring 2008 and winter 2009. Black lines represent 10 point running averages of the DBC/DCO ratios. 13 of 22

D21307


D21307

[50] The [DBC/DCO]all ratios for NC air in winter and spring were 4.50–5.26 ng m−3 ppbv−1 (Table 4), and the [DBC/DCO]ref ratios for these air masses were 6.79– 7.24 ng m−3 ppbv−1 (Table 5a). Figure 15 and Tables 4 and 5a indicate that seasonally averaged DBC/DCO ratios for NC and SC were rather stable. 6.3. Observed Transport Efficiency of BC (TEobs BC ) [51] Based on the discussion in sections 6.1 and 6.2, we derived the transport efficiency of BC from the observation using following equation. TEobs BC ð%Þ ¼ ½DBC=DCOall =½DBC=DCOref 100:

ð4Þ

cal The SD of the hourly (TEcal BC)i data points (dTEBC) was obs calculated for each season. The variability of TEBC (dTEobs BC ) was calculated using the following equation:

h

2 ½DBC=DCOall =½DBC=DCOall 2 i1=2 þ ½DBC=DCOref =½DBC=DCOref : ð5Þ

obs TEobs BC ð%Þ ¼ TEBC ð%Þ

Figure 13. DBC/DCO ratio versus TEcal BC for NC, MX, and SC air with DCO > 20 ppbv for winter and spring. The statistical DBC/DCO ratios for each bin of TEcal BC are also shown. The average DBC/DCO ± 1SD ratios (filled circles and vertical bars), median ratios (open circles), and the ratio of the sum of DBC (∑i (DBC)i and DCO (∑i (DCO)i) (open triangles) are shown. of BC in the PBL and difficulties in representing the wet deposition with a transport time scale of 2–3 days using the daily GPCP precipitation data. Wet deposition of BC is generally efficient for transport from the PBL to the FT because of the higher degree of supersaturation of water vapor. Wet deposition of BC by mixing processes in the PBL may not be efficient, although clouds often form near the top of the boundary layer [Stull, 1988]. The CMAQ calculations (i.e., TEcal BC) using hourly results of the WRF simulation may represent better the degree of influence of wet deposition on BC in the PBL, including precipitation in the PBL and entrainment of FT air influenced by wet deposition into the PBL.

cal The seasonally averaged values of TEcal BC with dTEBC and obs obs TEBC with dTEBC for each season are compared in Figure 16 and Table 4. The TEobs BC values for NC air in winter and spring were about 70%, and the TEcal BC values were comparable but slightly higher. [52] It should be noted that for the selection procedure, we used TEcal BC as a meteorological parameter representing a measure of the wet deposition. The use of TEcal BC itself does cal not directly influence the absolute values of TEobs BC . TEBC obs influences the TEBC values through the choice of the threshold TEcal BC value. However, the effect is small, as discussed in section 6.3. This uncertainty is unavoidable for any parameters representing wet deposition or any other criteria in selecting the air masses least impacted by wet deposition. [53] The TEobs BC values for NC air were similar for winter and spring (about 70%). This is the result of little difference in the [DBC/DCO]all ratio between winter and spring, because the [DBC/DCO]ref ratios were similar (Table 4). Precipitation increased from winter to spring, especially over the Yellow Sea and East China Sea, over which BC

Figure 14. Scatterplots of DBC‐DCO for NC and SC air during (left) spring 2008 and (right) winter 2009 for the air masses with RT > 24 h and TEcal BC > 80%. The slopes of the DBC/DCO correlations (in ng m−3 ppbv−1) are also given. 14 of 22

D21307

D21307


Table 5a. Seasonal Statistics of [DBC/DCO]ref ± d [DBC/DCO]ref (ng m−3 ppbv−1) for the Air Masses Arriving at Hedo From Different Regions for RT > 24 h and TEcal BC > 80% SC Season Spring 2008 Summer 2008 Fall 2008 Winter 2009 Spring 2009 All data set

NC

KR

MX

JP

[DBC/DCO]ref Data Point [DBC/DCO]ref Data Point [DBC/DCO]ref Data Point [DBC/DCO]ref Data Point [DBC/DCO]ref Data Point 7.56 ± 2.65 NA NA 7.37 ± 2.76 NA 7.47 ± 4.55

72 ‐ ‐ 31 ‐ 127

7.24 ± 2.56 NA 6.42 ± 3.06 7.09 ± 3.09 6.79 ± 2.17 7.00 ± 3.25

142 ‐ 71 101 66 376

NA NA NA NA 5.82 ± 1.52 5.73 ± 1.95

was transported, as shown in Figure 3. However, the change in the increase in precipitation did not decrease the TEobs BC value. This is also reflected in the poor correlation of the DBC/DCO ratio with APT (not shown). [54] The DBC/DCO ratios, and therefore TEobs BC , were correlated with TEcal BC, clearly indicating the effect of wet deposition of BC. The CMAQ calculations for TEcal BC for individual air masses show no significant seasonal difference in the average TEcal BC (Figure 16), although the monthly average TEcal BC has shown some seasonal variation (Figure 9, fourth panel). The absolute values of TEcal BC were somewhat higher than the TEobs BC values (Figure 16). The similarity in cal the seasonal variation of TEobs BC and TEBC also supports the representation of wet deposition of BC qualitatively. However, it should be noted that wet deposition by CMAQ may have been underestimated to some extent considering that obs the TEcal BC was larger than TEBC by about 10% on average. 6.4. Region‐Specific [DBC/DCO]ref [55] Region‐specific [DBC/DCO]ref values can be used to evaluate the regional emission inventory of BC and CO, as has been suggested in previous studies conducted in urban regions of East Asia [Kondo et al., 2006; Han et al., 2009; Verma et al., 2010]. An accurate estimation of emissions is essential for the modeling studies, because the uncertainties in model simulations are associated with those present in the input emission inventories. The [DBC/DCO]ref values were derived for NC, SC, MX, JP, and KR combining the entire measurement period (five seasons) after the selection of air masses with RT > 24 h and TEcal BC > 80%. The ratios are given in Table 5a. The [DBC/DCO]ref observed in the air masses of NC and SC, the major emission regions of East Asia, were 7.00 ng m−3 ppbv−1 and 7.47 ng m−3 ppbv−1, respectively. The air masses that arrived at Hedo from NC and SC could have been well mixed, influenced by emissions from all sectors summarized in Table 1a, which shows

‐ ‐ ‐ ‐ 17 27

5.87 ± 4.76 NA NA 5.73 ± 3.71 7.13 ± 5.22 6.23 ± 4.24

66 ‐ ‐ 42 116 260

6.92 ± 2.69 NA NA NA 4.69 ± 2.27 5.49 ± 3.24

68 ‐ ‐ ‐ 38 154

that the majority of emissions were from the industrial and residential sectors. [56] Table 6 illustrates a comparison of the observed [DBC/DCO]ref with other observations conducted in East Asia. Han et al. [2009] observed the slopes of DBC/DCO at an urban measurement site of Beijing in north China (NC) during four seasons (2005–2006). The BC measurements were made by using the thermal optical transmittance technique (TOT). The BC concentrations measured by TOT and COSMOS were shown to agree to within about 10% in Asia [Kondo et al., 2009, 2011a]. Their observed values of the DBC/DCO slopes were 3.4 ng m−3 ppbv−1, 4.8 ng m−3 ppbv−1, 5.8 ng m−3 ppbv−1, and 3.5 ng m−3 ppbv−1 during spring, summer, fall, and winter, respectively. These values are interpreted to represent the [DBC/DCO]ref ratios in the Beijing area, because the observation site was located much closer to large BC sources than Hedo. The [DBC/DCO]ref observed at Hedo is significantly higher, by about a factor of 2, than those of Han et al. [2009]. Hedo received well‐ mixed air masses carrying the impacts of various sources. Hence, the difference between [DBC/DCO]ref values observed at Hedo and those of Han et al. [2009] in Beijing suggests that the ratios in the Beijing area did not represent values in north China, in general. [57] Li et al. [2007] measured the slope of the BC‐CO correlations at a rural site about 70 km southeast downwind of the Beijing metropolitan area during spring 2005. The value of 9.2 ng m−3 ppbv−1 of the BC/CO slope of Li et al. [2007] is about 25% higher than that observed at Hedo (Table 5a). However, the BC measured by Li et al. [2007] may have interferences from aerosols other than BC, because their inlet was not heated. [58] Sahu et al. [2009] derived DBC/DCO slopes at Jeju Island in South Korea, located in the East China Sea, during spring 2005. The BC measurements were made by a Particle Soot Absorption Photometer (PSAP) with a heated inlet. BC measured with a PSAP and COSMOS agreed to within

Table 5b. Seasonal Statistics of [DBC/DCO]ref ± d [DBC/DCO]ref (ng m−3 ppbv−1) for the Air Masses Arriving at Hedo From Different a Regions for RT > 24 h and TEcal BC > 90% SC Season Spring 2008 Summer 2008 Fall 2008 Winter 2009 Spring 2009 All data set

NC

KR

MX

JP

[DBC/DCO]ref Data Point [DBC/DCO]ref Data Point [DBC/DCO]ref Data Point [DBC/DCO]ref Data Point [DBC/DCO]ref Data Point 8.01 ± 2.52 NA NA 9.83 ± 1.29 NA 8.64 ± 5.18

52 ‐ ‐ 15 ‐ 78

7.80 ± 2.28 NA 6.57 ± 3.08 7.45 ± 3.12 6.89 ± 2.41 7.37 ± 3.30

119 ‐ 64 72 43 294

NA NA NA NA 6.34 ± 0.71 6.44 ± 2.06

‐ ‐ ‐ ‐ 8 12

7.43 ± 5.22 NA NA 7.56 ± 3.32 7.38 ± 4.18 7.08 ± 3.11

cal [DBC/DCO]ref = ∑DBC/∑DCO (equation (3)), after the data selection with TEcal BC > 80% or TEBC > 90%.

a

15 of 22

39 ‐ ‐ 23 96 180

8.71 ± 1.29 NA NA NA 4.97 ± 2.57 6.07 ± 3.25

26 ‐ ‐ ‐ 18 56

D21307


D21307

based) and during the A‐FORCE aircraft measurement campaign held in spring 2009.

Figure 15. Seasonal trends of the [DBC/DCO]ref ratios in NC (open circles) and SC (open triangles) air masses using the data with RT > 24 h and TEcal BC > 80%. The [DBC/D CO]all ratios calculated using the entire data set for the categories of NC (closed circles) and SC (closed triangles) are also shown. BC/CO emission ratios calculated from the inventory of Zhang et al. [2009] (red triangles) are also shown for comparison. Vertical bars are d [DBC/DCO]all and d [DBC/DCO]ref given in Tables 4 and 5a, respectively. about 10% [Kondo et al., 2009]. Their DBC/DCO slopes ranged from 8.1 to 9.8 ng m−3 ppbv−1 in the air masses transported from China and Korea, respectively, although the data were not selected by TEcal BC. The [DBC/DCO]ref ratios of this study for China (both NC and SC) are somewhat lower with these values. Year‐to‐year variability and spatial variability in the [DBC/DCO]ref may explain the difference. However, it should also be noted that the [DBC/ DCO]ref value at Jeju was based on a much smaller number of observations than that used for the present study. [59] The [DBC/DCO]ref ratios for KR and JP air masses were 5.73 ng m−3 ppbv−1 and 5.49 ng m−3 ppbv−1, respectively. These values are similar to the values of 4.2– 6.2 ng m−3 ppbv−1 measured in Gwangju (KR) [Park et al., 2005] and 5.7 ng m−3 ppbv−1 in Tokyo [Kondo et al., 2006]. [60] Note that the value of the DBC/DCO ratio measured in urban regions of East Asia is about 5.0 (3.4–6.2) ng m−3 ppbv−1 on average [Han et al., 2009; Park et al., 2005; Kondo et al., 2006; Verma et al., 2010], indicating the influence of specific sources of BC, for example, traffic and industry. On the other hand, the DBC/DCO ratio measured at remote or rural observation sites in East Asia is about 7.0 (4.7–9.8) ng m−3 ppbv−1 on average, based on this and previous studies [Sahu et al., 2009; Li et al., 2007; Matsumoto et al., 2003], suggesting the combined influence from various sources of BC emissions.

7.1. Hedo [62] Figure 17 shows the time series of the monthly average ± standard deviation (dDCO) and median values of DCO calculated using the CMAQ model and those observed at Hedo. For the calculation of DCO from the CMAQ model‐calculated CO (COCMAQ), we follow the same method that was applied for the calculation of DBC and DCO from the observed data, as discussed in section 6.1. The observed DCO (DCOObs) has shown a strong seasonal variation. On the other hand, DCOCMAQ did not show strong seasonal variations. The difference between DCOObs and DCOCMAQ was large during spring and winter, when Hedo was under the influence of transport of CO from the source regions of East Asia (Figure 5), whereas the difference was less during summer (Figure 17), when the transport pattern was from the western Pacific to the continent (Figure 5). [63] The CO emission inventory of Zhang et al. [2009] used in this study is an annual average that did not include the seasonal variations in the CO emissions. We calculated the DCOCMAQ/DCOObs ratios in the air mass categories of source regions. The correlations between DCOCMAQ and DCOObs in NC and SC were moderate (not shown). The values of the DCOCMAQ/DCOObs slopes were 0.38 (r2 = 0.28) in NC and 0.30 (r2 = 0.30) in SC during spring 2008, while these were 0.48 (r2 = 0.52) in NC and 0.30 (r2 = 0.32) in SC during winter 2009. [64] Because correlation coefficients for DCOCMAQ‐D COObs were moderate (r2 ranged from 0.28 to 0.52), we calculated the DCOCMAQ/DCOObs ratio using the same method as for the calculation of the DBC/DCO ratios (equation (3)). Figure 18 shows the seasonal variations of the DCOCMAQ/DCOObs ratios for the air mass categories NC and SC. The DCOCMAQ/DCOObs ratio calculated for NC and SC using the entire data set is also shown in Figure 18. Table 7 summarizes the DCOCMAQ/DCOObs ratios for NC,

7. Evaluation of CO Emissions [61] An evaluation of the BC emissions estimated by Zhang et al. [2009] for China was made by Kondo et al. [2011b]. In this section, we evaluate the CO emission inventory of Zhang et al. [2009] by comparing the concentrations of CO calculated by using the CMAQ model with the concentrations of CO observed at Hedo (ground‐

Figure 16. Seasonal trends of the observed transport efficiency (TEobs BC ) in NC (closed circles) and SC (closed triangles) and model‐calculated transport efficiency (TEcal BC) in NC (open circles) and SC (open triangles). Vertical bars obs are dTEcal BC (±1SD, Table 4) and dTEBC (derived from equation (5), Table 4).

16 of 22

D21307

Not required no Not required Not required yes Not required yes Sunset EC‐OC (TOT) PSAP (L. Abs.) Sunset EC‐OC (TOT) Sunset EC‐OC (TOT) PSAP (L. Abs) ACPM (TOT) PSAP (L. Abs.) urban rural urban urban remote remote Aircraft measurement over Nagoya city

a COSMOS, continuous soot monitoring system; PSAP, Particle Soot Absorption Photometer; ACPM, ambient carbon particulate monitor; L. Abs., Light absorption; TOT, thermal‐optical transmittance; NA, not available; NC, north Chinese air masses; SC, south Chinese air masses; JP, Japanese air masses; KR, Korean air masses.

Han et al. [2009] Li et al. [2007] Andreae et al. [2008] Park et al. [2005] Sahu et al. [2009] Matsumoto et al. [2003] Kondo et al. [2006]

This study

NC: 7.09, SC: 7.37 JP: NA, KR: NA 3.5 NA NA NA NA ‐ NA NC: 6.42, SC: NA JP: NA, KR: NA 5.8 NA 7.9 NA NA ‐ NA NC: 7.24, 6.79, SC: 7.56 JP: 6.92, 4.69, KR: 5.82 3.4 9.22 NA 4.2 – 6.2 8.1 – 9.8 7.88 6.3 yes COSMOS (L. Abs.) remote

Hedo, Okinawa Island, Japan (2008–2009) Beijing, China (2005–2006) Xianghe (Beijing), China (2005) Guangzhou, China (2004) Gwanju, South Korea (2001) Gosan, Jeju Island, South Korea (2005) Rishiri Island, Japan (2001) Nagoya, Japan (2003)

Fall Spring Heated Inlet Instrument and Method of BC Analysis Type of Location Location and Year of Measurements

a Table 6. Comparison of Observed BC/CO Ratios ([DBC/DCO]ref, (ng m−3 ppbv−1) With TEcal BC > 80%) With Other Observations in East Asia

Winter

Reference


D21307

SC, MX, KR, and JP. The DCOCMAQ/DCOObs ratios ranged between 0.41 and 0.53 in NC during spring and winter. NC was the dominant source of air masses arriving at the Hedo during these seasons (Figure 5). The overall DCOCMAQ/D COObs ratios for NC and SC were 0.49 and 0.39 ppbv/ppbv, respectively (Table 7). 7.2. Aircraft Measurements [65] Figure 19 shows vertical profiles of the average ± d (SD) concentrations of COObs and COCMAQ at 26°N–33°N during A‐FORCE. Average and background concentrations of COObs and COCMAQ for each 1 km altitude bin are shown in Figure 19. Background concentrations of CO were defined as the 5th percentile for each 0.5 km altitude bin and were interpolated for each data point. The concentrations of COObs and COCMAQ near the surface were 241 and 165 ppbv, respectively, with a difference of about 76 ppbv. The difference between the average values of COObs and COCMAQ was larger near the surface and generally decreased with increasing altitude. [66] Figure 20 (left) shows the vertical profiles of the average ± d concentrations of DCOObs and DCOCMAQ at 26°N–33°N during A‐FORCE. The DCOObs and DCOCMAQ values were calculated by subtracting their background concentrations, given in Figure 19. The DCO observed during spring 2009 at Hedo (ground‐based) is also shown in Figure 20 (left). The concentration of DCOObs observed during A‐FORCE near the surface agreed with that of Hedo (ground‐based) (Figure 20, left). The average concentrations of DCOObs and DCOCMAQ increased up to an altitude of 2–3 km. The pattern of the vertical profile of DCOCMAQ was similar to those observed (Figure 20, left). The similarity in the profiles of DCOObs and DCOCMAQ suggests that the CMAQ model represented well the transport of CO from the PBL to the free troposphere on average. [67] Figure 20 (right) shows the vertical profile of the DCOCMAQ/DCOObs ratios for 26°N–33°N. The DCOCMAQ/ DCOObs ratios were calculated as the ∑DCOCMAQ/ ∑DCOObs ratio for each 1 km altitude bin. The DCOCMAQ/ DCOObs ratio observed at Hedo (ground‐based) for NC

Figure 17. Time series of the monthly average ±d (1 SD) and median values of the observed DCO (DCOObs) and those calculated using the CMAQ model (DCO CMAQ ). The DCOObs and DCOCMAQ values were calculated by subtracting their background concentrations from the observed and model‐calculated CO values. Background CO concentrations were defined as the 5th percentile of the monthly data, interpolated for each data point. 17 of 22

D21307

0.08 0.33 0.58 0.64 0.20 0.69 ± ± ± ± ± ± 0.17 0.19 0.30 0.54 0.41 0.28 161 ‐ 194 184 242 781 0.45 ± 0.24 NA 0.40 ± 0.46 0.46 ± 0.43 0.40 ± 0.22 0.41 ± 0.52 31 ‐ 78 26 53 223 0.30 ± 0.28 NA 0.21 ± 0.28 0.48 ± 0.28 0.48 ± 0.13 0.32 ± 0.34 232 ‐ 125 297 159 813 a

DCOCMAQ/DCOObs = ∑DCOCMAQ/∑DCOObs.

0.41 ± 0.32 NA 0.72 ± 0.51 0.53 ± 0.43 0.43 ± 0.18 0.49 ± 0.54 222 149 41 95 49 556 0.59 0.57 0.66 0.25 0.25 0.55 ± ± ± ± ± ± 0.34 0.51 0.78 0.37 0.25 0.39 Spring 2008 Summer 2008 Fall 2008 Winter 2009 Spring 2009 All data set

DCOCMAQ/DCOObs

JP MX KR NC

Data Points

DCOCMAQ/DCOObs

Data Points

DCOCMAQ/DCOObs

Data Points

DCOCMAQ/DCOObs

Data Points

18 of 22

SC

[69] The [DBC/DCO]ref ratios derived for each region should represent the BC/CO emission ratios. The region‐ specific [DBC/DCO]ref ratios of this study are useful in evaluating existing inventories, because, for example, the recent inventory of Zhang et al. [2009] still contains an uncertainty of 208% for BC and 70% for CO. Here, the [DBC/DCO]ref ratios derived for NC and SC air in different seasons are compared with the emission ratios calculated from the recent inventory of Zhang et al. [2009]. [70] Table 8 shows the BC/CO emission ratios from different sectors in China (NC and SC), Japan, and Korea. Comparing the [DBC/DCO]ref (Table 5a) with those of BC/ CO emission ratios (Table 8), the emissions ratios predicted by the emission inventory for the NC and SC regions are factors of 1.8 (=12.7/7.00) and 1.6 (=11.9/7.47) larger than [DBC/DCO]ref measured at Hedo, respectively. The BC/CO emission ratios for the power and residential sectors are much higher than the [DBC/DCO]ref ratios. Table 1a indicates that the contributions of BC and CO emissions from the power sector are only about 2%. The residential sector is estimated to make large contributions to the emissions of these species. From these considerations, it may be possible

DCOCMAQ/DCOObs

8. Comparison of [DBC/DCO]ref With Emission Ratios

Season

using the entire data set (Table 7) is also shown in Figure 20 (right). The vertical straight line in Figure 20 (right) shows the average DCOCMAQ/DCOObs ratio in the PBL (below 2 km). The A‐FORCE DCOCMAQ/DCOObs ratio near the surface agreed well with that for Hedo (ground‐based). It should be noted that the DCOCMAQ/DCOObs ratio was stable at about 0.5 up to 3–4 km, suggesting an effective uplifting of CO over the Asian continent up to this altitude. [68] The results of the comparison of DCOObs and DCOCMAQ using ground‐based and A‐FORCE aircraft measurements clearly demonstrate that the emission rate of CO from China calculated by Zhang et al. [2009] is underestimated by about a factor of 2.

Table 7. Ratio of CMAQ Model Calculated CO to That Observed at Hedo (DCOCMAQ/DCOObs ± d [DCOCMAQ/DCOObs)a

Figure 18. Seasonal variations of the COCMAQ/DCOObs ratios in NC (close circle) and SC (open circle) air masses using data with RT > 24 h. The DCOCMAQ/DCOObs ratio, calculated from the entire data set for the categories NC and SC, are also shown. Vertical bars are d [DCOCMAQ/D COObs], given in Table 7.

292 35 442 84 148 1143

VERMA ET AL.: TRANSPORT OF BC AND CO FROM EAST ASIA Data Points

D21307

D21307

D21307


Table 8. BC/CO (ng m−3 ppbv−1) Emission Ratios Estimated From Inventorya Region North China (NC) South China (SC) Japan (JP) Korea (KR)

Industrial Power Residential Transportation Total 9.6 7.9 8.6 5.8

18.8 14.8 12.0 105.0

21.2 19.5 18.3 23.9

6.9 6.5 11.1 27.7

12.7 11.7 11.3 17.5

a

From Zhang et al. [2009].

Figure 19. Vertical profile of the average ±d (1SD) of CO (COCMAQ and COObs) for the period of the Aerosol Radiative Forcing in East Asia (A‐FORCE) aircraft campaign (26°N– 33°N) over East Asia during spring 2009. Black closed circles and black open circles represent the average values of the observed CO (COObs) and their background concentrations (BG), respectively. BG was defined as the 5th percentile within each 0.5 km altitude bin. Horizontal bars represent standard deviations (1d). Red closed circles and red open circles represent the average and BG of COCMAQ, respectively.

that the BC/CO emission ratios in the residential sector are largely overestimated, since BC/CO emission ratios from industry and transportation agree well with [DBC/DCO]ref, within the 10%–20% variations (Tables 5a and 8). Kondo et al. [2011b] made a detailed comparison of the BC observed at Hedo and that predicted by CMAQ using the BC emission rates of Zhang et al. [2009]. It has been estimated that the total BC emission flux over China is very close (1.06 ± 0.03) to the estimate of Zhang et al. [2009], with an uncertainty of about 40%. This suggests that CO emissions from the residential sector in China are largely underestimated. It was also demonstrated that the emission rate of CO in China derived by Zhang et al. [2009] was underestimated by about a factor of 2. However, identification of the sectors causing the underestimation of the CO emissions is not entirely conclusive, because there can be substantial uncertainties in the estimates of the BC and CO emissions from the industry and transportation sectors in the Zhang et al. [2009] inventory. [71] According to Zhang et al. [2009], BC/CO emission ratios undergo seasonal variation, reaching a maximum in winter, mainly because of heating in this season. Figure 21

Figure 20. (left) Vertical profiles of DCO (DCOCMAQ and DCOObs) during A‐FORCE over East Asia (26°–33°N) during spring 2009. Black closed circles and triangles represent the average DCOObs and DCO CMAQ , respectively. Horizontal bars represent standard deviations (1 sigma). DCO Obs and DCOCMAQ were calculated by subtracting their background concentrations from COObs and COCMAQ, shown in Figure 19. The red closed circle is the average ± d of the DCO observed at Hedo (ground‐ based) during spring 2009. (right) Vertical profile of DCOCMAQ/DCOObs ratios. The DCOCMAQ/DCOObs ratios were calculated as the ∑DCOCMAQ/∑DCOObs ratios for each 1 km altitude bin. The red closed circle is the ground‐based DCOCMAQ/DCOObs ratio for spring 2009. The vertical blue solid line is the average DCOCMAQ/DCOObs ratio for the planetary boundary layer (below 2 km). 19 of 22

D21307


Figure 21. Ratio of the observed [DBC/DCO]ref in spring to that in winter for NC and SC. The emission ratios for the Chinese region (NC + SC) calculated from the inventory of Zhang et al. [2009] (red triangles) are also shown for comparison. Vertical bars are the uncertainties calculated from the square‐root of the sum of the squares of (d [DBC/ DCO]ref/ [DBC/DCO]ref) for spring and winter (Table 5a). shows a comparison of the observed seasonal variations of [DBC/DCO]ref ratios with those predicted by the inventory. The present data show almost no change between winter and spring, although the inventory predicts about a 20% change. This result suggests the uncertainty in the seasonal variation of the BC/CO emission ratios predicted by the inventory. [72] Relative variabilities in [DBC/DCO]ref (i.e., d [DBC/ DCO]ref/ [DBC/DCO]ref) measured at Hedo are calculated to be 46% (=3.25/7.00) and 61% (=4.55/7.47) for NC and SC air, respectively (Table 5a). Here we estimate the uncertainty of [DBC/DCO]ref represented by the square root of the sum of the squares of the uncertainty in the measurements of 10% for BC and 6% for CO (i.e., (102 + 62)1/2). The calculated uncertainty for [DBC/DCO]ref is about 12%. It is much smaller than 46%–61% of the relative variability in [DBC/DCO]ref and 219% (= (2082 + 702)1/2) of uncertainty in BC/CO emission ratios calculated from the square root of the sum of the squares of the uncertainty of 208% for BC and 70% for CO given in the inventories of Zhang et al. [2009]. The region‐specific [DBC/DCO]ref ratios estimated in the present study are, thus, reliable and are useful for studies of the transport of BC and validation of emissions inventories.

9. Conclusion [73] BC mass concentrations were measured continuously by COSMOS with an accuracy of about 10%, together with CO at Hedo, a remote site located in the East China Sea, from March 2008 to May 2009. For the first time, we have made a statistical analysis of the temporal variations of BC in Asian outflows throughout the year. Annual average concentrations of BC and CO were 0.29 mg m−3 and 150 ppbv, respectively. Interference from coarse particles (mostly dust particles) on the BC measurements was assessed to be negligible. [74] The origins of the observed air masses were determined by using 5 day back trajectories together with an analysis of the seasonally averaged wind and precipitation

D21307

fields. The seasonal variation of the transport of air masses over the East China Sea was controlled by alternating winter and summer monsoons. In spring and winter, the dominant source region air masses reaching Hedo was from China. During summer, most of the air masses were transported from the western Pacific. Because of the more frequent transport of Chinese air to Hedo in spring and winter, the average and background concentrations of BC and CO in these seasons were higher by about a factor of 2 than those of summer and fall. [75] Irrespective of season, higher concentrations of BC and CO were observed in the air masses arriving from NC and SC. For these air masses, the DBC/DCO ratios were generally correlated with the model‐calculated transport efficiency of BC (TEcal BC), suggesting the effect of wet removal in controlling BC transport. The validity of TEcal BC has been partially supported by the agreement between the monthly average precipitation in East Asia predicted by the WRF simulation and the GPCP data. The average TEcal BC for winter and spring was highest (72–81%) for NC air among all air mass categories, reflecting the lowest precipitation rates. This, together with the highest BC emission rate and frequency of transport, caused NC air to make the largest contributions to elevating BC levels at Hedo. [76] The DBC/DCO ratios in air masses influenced by Chinese emissions showed significant correlations with TEcal BC, clearly indicating that wet deposition is also an important controlling factor of the BC concentration at Hedo. We derived the [DBC/DCO]ref ratios least impacted by wet deposition by selecting data using TEcal BC (>80%) as a meteorological parameter. The transport efficiencies of BC, TEobs BC (equation (4)), estimated using the reference ratios were about 70% for NC and SC air. The TEobs BC values were similar for winter and spring, despite the increase in precipitation over the outflow region in spring. [77] Evaluation of the CO emission inventory of Zhang et al. [2009] was made using ground‐based and aircraft measurements. The DCOCMAQ/DCOObs ratios for ground‐ based and aircraft measurements were 0.49 (for NC at Hedo) and 0.52 in the PBL (aircraft), respectively. [78] We compared the [DBC/DCO]ref ratios with the BC/ CO emission ratios derived from the emission inventories of Zhang et al. [2009]. The derived BC/CO emission ratios were 1.8 and 1.6 times larger than the [DBC/DCO]ref ratios for NC and SC air, respectively. Since the overall BC emissions estimate by Kondo et al. [2011b] agreed well with those of the Zhang et al. [2009] inventory for China, it is therefore likely that the emission rate of CO was underestimated by about a factor of 2 by Zhang et al. [2009]. An indication of the underestimation of CO was also shown by the comparison of the CMAQ model‐calculated CO with that observed with ground‐based and aircraft measurements. [79] Acknowledgments. The work was supported by the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) and the global environment research fund of the Japanese Ministry of the Environment (A‐0803 and A‐1101). The CO measurements were supported by the Okinawa Prefecture Institute for Health and Environment and the PM10 by the Ministry of Environment (Japan) and EANET. The trajectory calculation program used in this paper was developed by Y. Tomikawa of the National Institute of Polar Research and K. Sato of the University of Tokyo, Japan. We also thank K. Kawana for her assistance with the field measurements and K. Ram for providing comments on this paper.

20 of 22

D21307


References Adler, R. F., et al. (2003), The Version 2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979‐Present), J. Hydrometeorol., 4, 1147–1167, doi:10.1175/1525-7541(2003)0042.0.CO;2. Alessandrini, F., H. Schulz, S. Takenaka, B. Lentner, E. Karg, H. Behrendt, and T. Jakob (2006), Effects of ultrafine carbon particle inhalation on allergic inflammation of the lung, J. Allergy Clin. Immunol., 117(4), 824–830, doi:10.1016/j.jaci.2005.11.046. Allen, D., K. Pickering, and M. Fox‐Rabinovitz (2004), Evaluation of pollutant outflow and CO sources during TRACE‐P using model‐calculated, aircraft‐based, and Measurements of Pollution in the Troposphere (MOPITT)‐derived CO concentrations, J. Geophys. Res., 109, D15S03, doi:10.1029/2003JD004250. Andreae, M. O., O. Schmid, H. Yang, D. Chanda, J. Z. Yu, L. M. Zeng, and Y. H. Zhang (2008), Optical properties and chemical composition of the atmospheric aerosol in urban Guangzhou, China, Atmos. Environ., 42, 6335–6350, doi:10.1016/j.atmosenv.2008.01.030. Binkowski, F. S., and S. J. Roselle (2003), Models‐3 Community Multiscale Air Quality (CMAQ) model aerosol component: 1. Model description, J. Geophys. Res., 108(D6), 4183, doi:10.1029/2001JD001409. Bond, T. C., D. G. Streets, K. F. Yarber, S. M. Nelson, J.‐H. Woo, and Z. Klimont (2004), A technology‐based global inventory of black and organic carbon emissions from combustion, J. Geophys. Res., 109, D14203, doi:10.1029/2003JD003697. Byun, D. W., and J. K. S. Ching (1999), Science algorithms of the EPA Models‐3 Community Multiscale Air Quality (CMAQ) modeling system, Rep. EPA/600/R‐99/030, U.S. Environ. Prot. Agency, Washington, D. C. Cao, G., X. Zhang, and F. Zheng (2006), Inventory of black carbon and organic carbon emissions from China, Atmos. Environ., 40, 6516–6527, doi:10.1016/j.atmosenv.2006.05.070. Carmichael, G. R., et al. (2003), Evaluating regional emission estimates using the TRACE‐P observations, J. Geophys. Res., 108(D21), 8810, doi:10.1029/2002JD003116. Carter, W. (2000), Documentation of the SAPRC‐99 chemical mechanism for VOC reactivity assessment, final report to California Air Resources Board, Rep. 92–329, Univ. of Calif., Riverside, 8 May. Chien, F. C., and H. Y. Kuo (2006), Topographic effects on a wintertime cold Front in Taiwan, Mon. Weather Rev., 134(11), 3297–3316, doi:10.1175/MWR3255.1. Chuang, C. C., J. E. Penner, J. M. Prospero, K. E. Grant, G. H. Rau, and K. Kawamoto (2002), Cloud susceptibility and the first aerosol indirect forcing: Sensitivity to black carbon and aerosol concentrations, J. Geophys. Res., 107(D21), 4564, doi:10.1029/2000JD000215. Conant, W. C., A. Nenes, and J. H. Seinfeld (2002), Black carbon radiative heating effects on cloud microphysics and implications for aerosol indirect forcing: 1. Extended Köhler theory, J. Geophys. Res., 107(D21), 4604, doi:10.1029/2002JD002094. Gerbig, C., S. Smitgen, D. Kley, A. Volz‐Thomas, H. Dewey, and D. Haaks (1999), An improved fast response vacuum UV resonance fluorescence CO instrument, J. Geophys. Res., 104, 1699–1704, doi:10.1029/ 1998JD100031. Han, S., et al. (2009), Temporal variations of elemental carbon in Beijing, J. Geophys. Res., 114, D23202, doi:10.1029/2009JD012027. Hatakeyama, S., K. Murano, F. Sakamaki, H. Mukai, H. Bandow, and Y. Komazaki (2001), Transport of atmospheric pollutants from East Asia, Water Air Soil Pollut., 130, 373–378, doi:10.1023/A:1013877000169. Hatakeyama, S., A. Takami, F. Sakamaki, H. Mukai, N. Sugimoto, A. Shimizu, and H. Bandow (2004), Aerial measurement of air pollutants and aerosols during 20– 22 March 2001 over the East China Sea, J. Geophys. Res., 109, D13304, doi:10.1029/2003JD004271. Haywood, J., and O. Boucher (2000), Estimates of the direct and indirect radiative forcing due to tropospheric aerosols: A review, Rev. Geophys., 38, 513–543, doi:10.1029/1999RG000078. Huffman, G. J., R. F. Adler, M. M. Morrissey, D. T. Bolvin, S. Curtis, R. Joyce, B. McGavock, and J. Susskind (2001), Global precipitation at one‐degree daily resolution from multisatellite observations, J. Hydrometeorol., 2, 36–50, doi:10.1175/1525-7541(2001)0022.0.CO;2. Intergovernmental Panel on Climate Change (IPCC) (2007), IPCC Fourth Assessment Report: Climate Change 2007, Cambridge, Univ. Press, Cambridge, U. K. Jansen, K. L., T. V. Larson, J. Q. Koenig, T. F. Mar, C. Fields, J. Stewart, and M. Lippmann (2005), Associations between health effects and particulate matter and black carbon in subjects with respiratory disease, Environ. Health Perspect., 113, 1741–1746, doi:10.1289/ehp.8153. Kajii, Y., K. Someno, H. Tanimoto, J. Hirokawa, and H. Akimoto (1998), Evidence for the seasonal variation of photochemical activity of tropo-

D21307

spheric ozone: Continuous observation of ozone and CO at Happo, Japan, Geophys. Res. Lett., 25, 3505–3508, doi:10.1029/98GL02602. Kajino, M., and Y. Kondo (2011), EMTACS: Development and regional‐ scale simulation of a size, chemical, mixing type, and soot shape resolved atmospheric particle model, J. Geophys. Res., 116, D02303, doi:10.1029/ 2010JD015030. Kato, S., Y. Kajii, R. Itokazu, J. Hirokawa, S. Koda, and Y. Kinjo (2004), Transport of atmospheric carbon monoxide, ozone, and hydrocarbons from Chinese coast to Okinawa island in the Western Pacific during winter, Atmos. Environ., 38, 2975–2981, doi:10.1016/j.atmosenv. 2004.02.049. Koch, D., et al. (2009), Evaluation of black carbon estimations in global aerosol models, Atmos. Chem. Phys., 9, 9001–9026, doi:10.5194/acp-99001-2009. Koga, S., T. Maeda, and N. Kaneyasu (2008), Source contributions to black carbon mass fractions in aerosol particles over the northwestern Pacific, Atmos. Environ., 42, 800–814, doi:10.1016/j.atmosenv.2007.09.052. Kondo, Y., et al. (2006), Temporal variations of elemental carbon in Tokyo, J. Geophys. Res., 111, D12205, doi:10.1029/2005JD006257. Kondo, Y., et al. (2009), Stabilization of the mass absorption cross section of black Carbon for filter‐based absorption photometry by the use of a heated inlet, Aerosol Sci. Technol., 43(8), 741–756, doi:10.1080/ 02786820902889879. Kondo, Y., L. Sahu, N. Moteki, F. Khan, N. Takegawa, X. Liu, M. Koike, and T. Miyakawa (2011a), Consistency and traceability of black carbon measurements made by laser‐induced incandescence, thermal optical transmittance, and filter‐based photo‐absorption techniques, Aerosol Sci. Technol., 45(2), 295–312, doi:10.1080/02786826.2010.533215. Kondo, Y., N. Oshima, M. Kajino, R. Mikami, N. Moteki, N. Takegawa, R. L. Verma, Y. Kajii, S. Kato, and A. Takami (2011b), Emissions of black carbon in East Asia estimated from observations at a remote site in the East China Sea, J. Geophys. Res., 116, D16201, doi:10.1029/ 2011JD015637. Kristjansson, J. E. (2002), Studies of the aerosol indirect effect from sulfate and black carbon aerosols, J. Geophys. Res., 107(D15), 4246, doi:10.1029/2001JD000887. Li, C., L. T. Marufu, R. R. Dickerson, Z. Li, T. Wen, Y. Wang, P. Wang, H. Chen, and J. W. Stehr (2007), In situ measurements of trace gases and aerosol optical properties at a rural site in northern China during East Asian Study of Tropospheric Aerosols: An International Regional Experi m e n t 2 0 0 5 , J . G e o p h y s . R e s. , 1 1 2 , D 2 2 S 0 4 , d o i : 1 0 . 1 0 2 9/ 2006JD007592. Lighty, J. S., J. M. Veranth, and A. F. Sarofim (2000), Combustion aerosols: Factors governing their size and composition and implications to human health, J. Air Waste Manage. Assoc., 50, 1565–1618. Liu, M. C., M. T. Yeh, S. Paul, Y. C. Lee, D. J. Jacob, M. Fu, J. H. Woo, G. R. Carmichael, and D. G. Streets (2008), Effect of anthropogenic emissions in East Asia on regional ozone levels during spring cold continental outbreaks near Taiwan: A case study, Environ. Model. Softw., 23, 579–591, doi:10.1016/j.envsoft.2007.08.007. Matsui, H., et al. (2011), Seasonal variation of the transport of black carbon aerosol from the Asian continent to the Arctic during the ARCTAS aircraft campaign, J. Geophys. Res., 116, D05202, doi:10.1029/ 2010JD015067. Matsumoto, K., Y. Uyama, T. Hayano, H. Tanimoto, I. Uno, and M. Uematsu (2003), Chemical properties and outflow patterns of anthropogenic and dust particles on Rishiri Island during the Asian Pacific Regional Aerosol Characterization Experiment (ACE‐Asia), J. Geophys. Res., 108(D23), 8666, doi:10.1029/2003JD003426. Menon, S., J. Hansen, L. Nazarenko, and Y. Luo (2002), Climate effects of black carbon aerosols in China and India, Science, 297, 2250–2253, doi:10.1126/science.1075159. Mijling, B., R. J. van der A, K. F. Boersma, M. Van Roozendael, I. De Smedt, and H. M. Kelder (2009), Reductions of NO2 detected from space during the 2008 Beijing Olympic Games, Geophys. Res. Lett., 36, L13801, doi:10.1029/2009GL038943. Miyazaki, Y., Y. Kondo, L. K. Sahu, J. Imaru, N. Fukushima, and M. Kanod (2008), Performance of a newly designed continuous soot monitoring system (COSMOS), J. Environ. Monit., 10, 1195–1201, doi:10.1039/ b806957c. Naidu, C. V., K. Durgalakshmi, K. Muni Krishna, S. Ramalingeswara Rao, G. C. Satyanarayana, P. Lakshminarayana, and L. Malleswara Rao (2009), Is summer monsoon rainfall decreasing over India in the global warming era?, J. Geophys. Res., 114, D24108, doi:10.1029/2008JD011288. Narita, D., et al. (1999), Seasonal variation of carbon monoxide at remote sites in Japan, Chemosphere Global Change Sci., 1, 137–144, doi:10.1016/S1465-9972(99)00023-9. Ohara, T., H. Akimoto, J. Kurokawa, N. Horii, K. Yamaji, X. Yan, and T. Hayasaka (2007), An Asian emission inventory of anthropogenic

21 of 22

D21307


emission sources for the period 1980–2020, Atmos. Chem. Phys., 7, 4419–4444, doi:10.5194/acp-7-4419-2007. Palmer, P. I., D. J. Jacob, D. B. A. Jones, C. L. Heald, R. M. Yantosca, J. A. Logan, G. W. Sachse, and D. G. Streets (2003), Inverting for emissions of carbon monoxide from Asia using aircraft observations over the western Pacific, J. Geophys. Res., 108(D21), 8828, doi:10.1029/ 2003JD003397. Park, S. S., M. S. Bae, J. J. Schauer, S. S. Ryu, Y. J. Kim, S. Y. Cho, and S. J. Kim (2005), Evaluation of the TMO and TOT methods for OC and EC measurement and their characteristics in PM2.5 at an urban site of Korea during ACE‐Asia, Atmos. Environ., 39, 5101–5112, doi:10.1016/j.atmosenv.2005.05.016. Pochanart, P., J. Hirokawa, Y. Kajii, and H. Akimoto (1999), Influence of regional‐scale anthropogenic activity in northeast Asia on seasonal variations of surface ozone and carbon monoxide observed at Oki, Japan, J. Geophys. Res., 104(D3), 3621–3631, doi:10.1029/1998JD100071. Ramanathan, V., and G. Carmichael (2008), Global and regional climate changes due to black carbon, Nat. Geosci., 1, 221–227, doi:10.1038/ ngeo156. Ramanathan, V., et al. (2001), The Indian Ocean Experiment: An integrated analysis of the climate forcing and effects of the great Indo‐Asian haze, J. Geophys. Res., 106, 28,371–28,398, doi:10.1029/2001JD900133. Sahu, L. K., Y. Kondo, Y. Miyazaki, M. Kuwata, M. Koike, N. Takegawa, H. Tanimoto, H. Matsueda, S. C. Yoon, and Y. J. Kim (2009), Anthropogenic aerosols observed in Asian continental outflow at Jeju Island, Korea, in spring 2005, J. Geophys. Res., 114, D03301, doi:10.1029/ 2008JD010306. Skamarock, W. C., J. B. Klemp, J. Dudhia, D. O. Grill, D. M. Barker, W. Wang, and J. G. Powers (2005), A description of the advanced research WRF version 2, NCAR Tech. Note NCAR/TN‐468+STR, 100 pp., Natl. Cent. for Atmos. Res., Boulder, Colo. Streets, D. G., et al. (2003), An inventory of gaseous and primary aerosol emissions in Asia in the year 2000, J. Geophys. Res., 108(D21), 8809, doi:10.1029/2002JD003093. Streets, D. G., Q. Zhang, L. Wang, K. He, J. Hao, Y. Wu, Y. Tang, and G. R. Carmichael (2006), Revisiting China’s CO emissions after the Transport and Chemical Evolution over the Pacific (TRACE‐P) mission: Synthesis of inventories, atmospheric modeling, and observations, J. Geophys. Res., 111, D14306, doi:10.1029/2006JD007118. Stull, R. B. (1988), An Introduction to Boundary Layer Meteorology, 684 pp., Springer, New York. Sugimoto, N., Y. Hara, A. Shimizu, K. Yumimoto, I. Uno, and M. Nishikawa (2011), Comparison of surface observations and a regional dust transport model assimilated with lidar network data in Asian dust event of March 29 to April 2, 2007, SOLA, 7A, 13–16. Suglia, S. F., A. Gryparis, R. O. Wright, J. Schwartz, and R. J. Wright (2007), Association of black carbon with cognition among children in a prospective birth cohort study, Am. J. Epidemiol., 167(3), 280–286, doi:10.1093/aje/kwm308. Suntharalingam, P., D. J. Jacob, P. I. Palmer, J. A. Logan, R. M. Yantosca, Y. Xiao, M. J. Evans, D. G. Streets, S. L. Vay, and G. W. Sachse (2004), Improved quantification of Chinese carbon fluxes using CO2/CO correlations in Asian outflow, J. Geophys. Res., 109, D18S18, doi:10.1029/ 2003JD004362. Suthawaree, J., S. Kato, A. Takami, H. Kadena, M. Toguchi, K. Yogi, S. Hatakeyama, and Y. Kajii (2008), Observation of ozone and carbon

D21307

monoxide at Cape Hedo, Japan: Seasonal variation and influence of long‐range transport, Atmos. Environ., 42, 2971–2981, doi:10.1016/j. atmosenv.2007.12.053. Takami, A., T. Miyoshi, A. Shimono, and S. Hatakeyama (2005), Chemical composition of fine aerosol measured by AMS at Fukue Island, Japan during APEX period, Atmos. Environ., 39, 4913–4924, doi:10.1016/j. atmosenv.2005.04.038. Takami, A., T. Miyoshi, A. Shimono, N. Kaneyasu, S. Kato, Y. Kajii, and S. Hatakeyama (2007), Transport of anthropogenic aerosols from Asia and subsequent chemical transformation, J. Geophys. Res., 112, D22S31, doi:10.1029/2006JD008120. Tan, Q., W. L. Chameides, D. Streets, T. Wang, J. Xu, M. Bergin, and J. Woo (2004), An evaluation of TRACE‐P emission inventories from China using a regional model and chemical measurements, J. Geophys. Res., 109, D22305, doi:10.1029/2004JD005071. Tomikawa, Y., and K. Sato (2005), Design of the NIPR trajectory model, Polar Meteorol. Glaciol., 19, 120–137. Verma, R. L., et al. (2010), Temporal variations of black carbon in Guangzhou, China, in summer 2006, Atmos. Chem. Phys., 10, 6471–6485, doi:10.5194/acp-10-6471-2010. Verma, S., J. Worden, S. Payra, L. Jourdain, and C. Shim (2009), Characterizing the long‐range transport of black carbon aerosols during Transport and Chemical Evolution over the Pacific (TRACE‐P) experiment, Environ. Monit. Assess., 154, 85–92, doi:10.1007/s10661-008-0379-2. Weinstock, B. (1969), Carbon monoxide: Residence time in the atmosphere, Science, 166, 224–225, doi:10.1126/science.166.3902.224. Yu, H., P. Wang, X. M. Zong, X. Li, and D. Lu (2010), Change of NO2 column density over Beijing from satellite measurement during the Beijing 2008 Olympic Games, Chin. Sci. Bull., 55(3), 308‐313, doi:10.1007/s11434-009-0375-0. Yumimoto, H., I. Uno, N. Sugimoto, A. Shimizu, Z. Liu, and D. M. Winker (2008), Adjoint inversion modeling of Asian dust emission using lidar observations, Atmos. Chem. Phys., 8, 2869–2884, doi:10.5194/acp-82869-2008. Zhang, Q., et al. (2009), Asian emissions in 2006 for the NASA INTEX‐B mission, Atmos. Chem. Phys., 9, 5131–5153, doi:10.5194/acp-9-51312009. Y. Kajii and S. Kato, Division of Applied Chemistry, Faculty of Urban Environmental Sciences, Tokyo Metropolitan University, 1‐1 Minami‐ ohsawa, Hachioji, Tokyo 192‐0397, Japan. K. Kita, Faculty of Science, Ibaraki University, 2‐1‐1 Bunkyo, Mito, Ibaraki 310‐8512, Japan. Y. Kondo, H. Matsui, and R. L. Verma, Department of Earth and Planetary Science, Graduate School of Science, University of Tokyo, Hongo 7‐3‐1, Bunkyo‐ku, Tokyo 113‐0033, Japan. ([email protected]‐ tokyo.ac.jp) T. Miyakawa, Research Center for Advanced Science and Technology, University of Tokyo, 4‐6‐1 Komaba, Meguro‐ku, Tokyo 153‐8904, Japan. N. Oshima, Meteorological Research Institute, 1‐1 Nagamine, Tsukuba, Ibaraki 305‐0052, Japan. L. K. Sahu, Physical Research Laboratory, Ahmedabad 380 009, India. A. Takami, National Institute for Environmental Studies, 16‐2 Onogawa, Tsukuba, Ibaraki 305‐8506, Japan.

22 of 22