Received 6 December 2012; revised 25 July 2013; accepted 1 August 2013; published 9 September 2013.

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH: ATMOSPHERES, VOL. 118, , doi: /jgrd.50702, 2013 Seasonal variations of Asian black carbon outflow to the Pacific: Contribution from anthropogenic sources in China and biomass burning sources in Siberia and Southeast Asia H. Matsui, 1 M. Koike, 1 Y. Kondo, 1 N. Oshima, 2 N. Moteki, 1 Y. Kanaya, 3 A. Takami, 4 and M. Irwin 1 Received 6 December 2012; revised 25 July 2013; accepted 1 August 2013; published 9 September [1] The Community Multiscale Air Quality model with a source and process tagged method (CMAQ/PASCAL) was used to understand source regions and types (anthropogenic (AN) and biomass burning (BB)) of Asian black carbon (BC) outflow to the Pacific during The model simulations generally reproduced absolute concentrations and temporal (seasonal, monthly, and day-to-day) variations of BC mass concentrations, observed by both surface and aircraft measurements in outflow regions in East Asia. These model simulations show that both the total eastward flux and transport efficiency (fractions transported from sources) of BC are highest during spring (26 kg s 1 and 33% at 150 E) and lowest during summer (8 kg s 1 and 20% at 150 E). These seasonal variations of Asian BC outflow are generally controlled by transport patterns (monsoons, frontal passages, and convection) and emissions from the following three sources: (1) AN emissions from China (China AN), (2) BB emissions from Southeast Asia and South China (SEA BB) during February April, and (3) BB emissions from Siberia and Kazakhstan (Siberia BB) during April July. In our simulations, China AN dominates the total eastward BC flux on a 3 year average (61%, 17%, and 6% from China AN, Siberia BB, and SEA BB, respectively, at 150 E). In contrast, SEA and Siberia BB account for 30 50% of the total eastward BC flux (150 E and 175 E) during spring and summer, and they increase the seasonal variability of the Asian BC outflow flux. BC from Siberia BB is also found to be transported to the Pacificmoreefficiently than BC from other sources. Although the magnitudes of BB emissions are highly uncertain, our results suggest that the control of Siberia BB will be important in terms of the transboundary transport of BC to the Pacific, North America, and the Arctic. Citation: Matsui, H., M. Koike, Y. Kondo, N. Oshima, N. Moteki, Y. Kanaya, A. Takami, and M. Irwin (2013), Seasonal variations of Asian black carbon outflow to the Pacific: Contribution from anthropogenic sources in China and biomass burning sources in Siberia and Southeast Asia, J. Geophys. Res. Atmos., 118, , doi: /jgrd Introduction [2] Black carbon (BC) particles are considered to be a major contributor to global warming because they efficiently absorb solar radiation, leading to a heating of the atmosphere (aerosol direct effect) [e.g., Menon et al., 2002; Ramanathan and Carmichael, 2008]. Absorption by BC particles can also modify the vertical stability of the atmosphere, convective activity, cloud formation (semidirect effect), and potentially 1 Department of Earth and Planetary Science, Graduate School of Science, University of Tokyo, Tokyo, Japan. 2 Meteorological Research Institute, Ibaraki, Japan. 3 Research Institute for Global Change, Japan Agency for Marine-Earth Science and Technology, Kanagawa, Japan. 4 National Institute for Environmental Studies, Ibaraki, Japan. Corresponding author: H. Matsui, Department of Earth and Planetary Science, Graduate School of Science, University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo , Japan. (matsui@eps.s.u-tokyo.ac.jp) American Geophysical Union. All Rights Reserved X/13/ /jgrd the circulation and hydrological cycle of the atmosphere [Hansen et al., 1997; Jacobson, 2002; Koch and Del Genio, 2010]. Once BC particles are transported and deposited onto snow/ice surfaces over the polar regions, they lower surface reflectance and promote faster melting of snow/ice sheets (known as the snow-albedo feedback) [Clarke and Noone, 1985; Flanner et al., 2007, 2009; Hansen and Nazarenko, 2004]. BC particles freshly emitted to the atmosphere are generally hydrophobic but gradually become hydrophilic through atmospheric aging processes (condensation, coagulation, and/or photochemical oxidation processes) [e.g., Oshima et al., 2009; Zaveri et al., 2010; Matsui et al., 2013]. Hydrophilic BC particles can act as cloud condensation nuclei and modify cloud microphysical properties, such as cloud brightness, longevity, and precipitation patterns (aerosol indirect effects) [e.g., Koch et al., 2011]. [3] BC is mainly emitted from anthropogenic (AN) combustion and biomass burning (BB) (AN is defined as sources other than biomass burning in this study). East and Southeast Asia (hereafter referred to simply as Asia) are one of the largest sources of anthropogenic BC in the world 9948

2 [Streets et al., 2003; Bond et al., 2004; Dentener et al., 2006; Zhang et al., 2009]. A number of previous studies were conducted to understand the transport of polluted air emitted from Asia to the Pacific, North America, and the Arctic [e.g., Ramanathan et al., 2001; Bey et al., 2001; Huebert et al., 2003; Jacob et al., 2003; Koike et al., 2003; Liu et al., 2003, 2008; Jaffe et al., 2003; Carmichael et al., 2003; Tang et al., 2003, 2004; Takemura et al., 2003; Kondo et al., 2004; Liang et al., 2004; Oshima et al., 2004; Nakajima et al., 2007; Adhikary et al., 2010; Matsui et al., 2011a, 2011b, 2011c; Oshima et al., 2012]. Several studies focused on the spatial distributions and transport processes of BC over Asia and the Pacific [e.g.,ramanathan et al., 2001; Uno et al., 2003; Clarke et al., 2004; Park et al., 2005; Hadley et al., 2007; Oshima et al., 2012]. For example, Uno et al. [2003] conducted BC simulations during the ACE-Asia campaign (spring 2001) and estimated the BB contribution from Southeast Asia and China to BC mass concentrations over the outflow regions. Hadley et al. [2007] simulated BC transport across the Pacific into North America during April 2004 and estimated a very large contribution of BC from Asia. These studies are important, as they show the impact of BC particles from Asia on the radiation budget and air quality over the Pacific. However, since most of them focused on spring periods, few studies have been conducted for long-term BC simulations over Asia and the Pacific. Therefore, it remains an important issue to understand seasonal variations of BC concentrations, their source regions and types (AN or BB), and their transport flux over the Pacific. [4] Another potentially important source of BC over the Pacific is BB emissions from Siberia and Kazakhstan (hereafter referred to as Siberia BB) in late spring and summer. Previous studies have shown the impact of Siberia BB on air quality over the outflow regions [e.g., Jeong et al., 2008; Tanimoto et al., 2009]. Some studies, focused on BC, showed that Siberia BB largely contributes to BC concentrations in Japan [Kaneyasu et al., 2007] and over the Arctic [Warneke et al., 2009, 2010; Matsui et al., 2011a, 2011b]. However, there are few modeling studies focused on the transpacific transport of BC from Siberia BB sources. The relative importance of Siberia BB, Asia AN, and Asia BB sources to BC outflow over the Pacific is still not understood well. [5] In this study, we simulate the BC mass outflow from the Asian continent to the Pacific during , using the Community Multiscale Air Quality (CMAQ) model with the Process, Age, and Source Region Chasing Algorithm (PASCAL) [Matsui and Koike, 2012] (section 2). Both surface and aircraft measurements (section 3) are used to validate BC mass concentrations in the CMAQ/ PASCAL model simulations (sections 4 and 5). Then, we examine seasonal variations of spatial distributions, source contributions (regions and types), and transport efficiencies and patterns of BC outflow over the Pacific (section 6). We mainly focus on the three largest BC sources over the Asian continent: AN emissions from China (China AN), BB emissions from Southeast Asia and South China (SEA BB), and BB emissions from Siberia and Kazakhstan (Siberia BB). 2. Regional Three-Dimensional Model 2.1. WRF-CMAQ/PASCAL Model [6] The meteorological field was simulated using the Weather Research and Forecasting (WRF) model version [Skamarock et al., 2008]. The chemical field was simulated by the CMAQ model version 4.7 with PASCAL [Matsui and Koike, 2012] using the meteorological results simulated by WRF. In the CMAQ model, aerosol dynamics were simulated with the fifth-generation CMAQ aerosol module. The CMAQ model represents the aerosol size distribution by three lognormal distributions (Aitken, accumulation, and coarse modes). The aerosol mass concentrations of sulfate, nitrate, ammonium, BC, primary and secondary organic aerosols, nonreactive dust, sea salt, and aerosol-phase water are predicted by calculating emission rates, gas-phase and in-cloud chemistry, nucleation, coagulation, gas-aerosol equilibrium, and dry and wet deposition. In this study, the modified wet deposition scheme was used following the studies of Kondo et al. [2011a] and N. Oshima (submitted manuscript, 2013). More detailed descriptions of the CMAQ model have been given elsewhere [Byun and Ching, 1999; Binkowski and Roselle, 2003]. Other schemes and options of the WRF andcmaqmodelsusedinthisstudyaresimilartothose of Matsui et al. [2009]. [7] In the CMAQ model, BC particles in the accumulation mode are assumed to be completely activated within clouds and removed by precipitation (hydrophobic BC is not considered), while the aging process of BC plays a key role in determining the lifetime of BC in the atmosphere. Recent measurements and modeling studies have shown that photochemical activity leads to the rapid conversion of fresh, nonspherical BC to aged, coated, spherical BC particles [Moffet and Prather, 2009; Matsui et al., 2013]. Matsui et al. [2013] simulated BC aging processes using a BC mixing state resolved (MS-resolved) WRF-chem over East Asia during the Aerosol Radiative Forcing in East Asia (A-FORCE) periods and showed that most BC particles have already experienced sufficient aging processes over the Asian continent. Therefore, the assumption used in CMAQ may be reasonable as a first approximation. Matsui et al. [2013] showed that the simple mixing state treatment (simulation without hydrophobic BC) led to the underestimation of BC mass concentrations by 15 20% in the planetary boundary layer and up to 35% in the upper troposphere over the Asian continent and western Pacific. The underestimation of the BC flux at 150 E (discussed in section 6) was also estimated to be 33% (10 N 60 N and surface 100 hpa in March and April 2009), with little latitudinal dependence (not shown). This result suggests that the main conclusions obtained in this study (e.g., source contribution, the latitudinal dependence of transport efficiency) will not change due to the treatment of the BC mixing state, though there are some uncertainties in the absolute concentrations and fluxes of BC. [8] PASCAL, which was developed and implemented into the CMAQ model in our previous study, is a new source and process apportionment method, tracing source regions, ages (transport time from emissions), and physical and chemical processes three-dimensionally for primary and secondary aerosols and their precursor gases by using new variables 9949

3 Figure 1. Simulation domain used in this study (shaded area). Gray lines show the definition of source regions in the CMAQ/PASCAL simulations. Colored lines show the source groups defined in this study: China (red, 13 source regions), Japan (blue, one source region), Korea (orange, one source region), Siberia (green, seven source regions), and Southeast Asia (black, three source regions). (tags) [Matsui and Koike, 2012]. PASCAL is not only free from the nonlinearity of secondary aerosol formation processes but is also computationally efficient. In this study, PASCAL is used to trace source regions and types (AN and BB) and transport efficiency (fractions transported from sources, TE BC ) of BC mass concentrations three-dimensionally. TE BC values are defined as the ratio of BC concentrations in the base simulation (with removal processes) to those assuming no removal processes (the sum of BC concentrations in the base simulation and DDEP and WDEP tags: DDEP and WDEP are the loss amounts by dry and wet deposition processes, respectively, traced in PASCAL). The definition is conceptually consistent with our previous studies [Oshima et al., 2012; Matsui et al., 2011a, 2011b]. PASCAL is described in further detail by Matsui and Koike [2012] Simulation Setups [9] Figure 1 shows the model domain used in this study (denoted by the shaded area). The horizontal grid spacing is 81 km ( grids), and there are 20 vertical levels from the surface to 100 hpa. The simulation period was 1 December 2007 to 31 December 2010 (totaling 3 years and 1 month). The first month was used for model spin-up. Figure 1 also shows the definition of source regions in the CMAQ/PASCAL simulations used in this study (gray lines): 25 specific regions, one for the rest of the region, and one for the initial and boundary conditions. Two tags were assigned to each source region to trace both anthropogenic (AN) and biomass burning (BB) contributions. In this study, we mainly focus on five source groups, defined by colored lines in Figure 1: China (13 source regions), Japan (one source region), Korea (one source region), Siberia (seven source regions), and Southeast Asia (three source regions). In addition, we defined three subregions in China: North (20 N 28 N), Central (28 N 36 N), and South China (36 N 50 N). Meteorological parameters (temperature, wind field, and water vapor mixing ratio) were nudged by the National Centers for Environmental Prediction (NCEP) Final (FNL) Operational Global Analysis data in the WRF simulation. [10] The horizontal grid spacing of 81 km may be considered coarse. However, since BC transport over East Asia and over the Pacific is generally controlled by synopticscale (~1000 km scale) meteorological variations (e.g., passages of high- and low-pressure systems), the simulation with the horizontal resolution of 81 km can capture these transport processes reasonably well, as shown in section 4. Our previous study [Verma et al., 2011] also showed the performance of precipitation amounts simulated by WRF- CMAQ model through the comparison with Global Precipitation Climatology Project observation data. Verma et al. [2011] showed that the monthly averaged precipitation amounts and their monthly variations over East Asia (using a WRF simulation with similar meteorological data (NCEP FNL) and resolution, i.e., 81 km) were generally reproduced well, suggesting the validity of precipitation amounts in our simulation Emissions [11] The emission inventories used in this study are as follows: anthropogenic emissions in East Asia for the year 2006 [Zhang et al., 2009], daily BB emissions estimated from Moderate Resolution Imaging Spectroradiometer (MODIS) satellite measurements (described in the next paragraph), volcanic emissions in East Asia for the year 2000 [Streets et al., 2003], monthly biogenic emissions from the Global Emissions Inventories Activity [Guenther et al., 1995], and online sea salt emissions [Gong, 2003; Zhang et al., 2005]. The latter three emission sources (volcanic, biogenic, and sea salt) are negligible for BC mass. As the injection height of emissions (e.g., sources from BB and power plants) itself is an uncertain parameter, all emissions used in this study were introduced to the lowest layer in CMAQ/PASCAL Table 1a. BC Emissions From Anthropogenic (AN) and Biomass Burning (BB) Sources Used in This Study Area (60 E 180 E) Source Type Reference Year Units 10 N 30 N 30 N 50 N 50 N 70 N BB This study 2006 Gg a BB This study 2008 Gg a BB This study 2009 Gg a BB This study 2010 Gg a AN Zhang et al. [2009] 2006 Gg a BB / (BB + AN) a 2006 % a Ratios of BB emissions estimated in this study to sum of AN [Zhang et al., 2009] and BB emissions. 9950

4 Table 1b. CO Emissions From Anthropogenic (AN) and Biomass Burning (BB) Sources Used in This Study Area (60 E 180 E) Source Type Reference Year Units 10 N 30 N 30 N 50 N 50 N 70 N BB This study 2006 Tg a BB This study 2008 Tg a BB This study 2009 Tg a BB This study 2010 Tg a AN Zhang et al. [2009] 2006 Tg a BB / (BB + AN) a 2006 % a Ratios of BB emissions estimated in this study to sum of AN [Zhang et al., 2009] and BB emissions. simulations. Sensitivity simulations suggest that the uncertainty in the injection height of BB emissions is less than 20% for BC concentrations at the surface and less than 10% for the BC flux at 150 E, as shown in Appendix A. AN emissions at higher latitudes (> 50 N) are not considered in this study, as BC emissions from them (~ 40 Gg a 1 within 50 N 70 N and 60 E 180 E based on Bond et al. [2004]) are sufficiently small compared with BB sources at higher latitudes and AN sources at lower latitudes (10 N 50 N) (Table 1a). [12] Daily BB emissions were calculated by the following equation [e.g., Seiler and Crutzen, 1980; Chang and Song, 2010]. M i ¼ A B CF EF i (1) where M i is the mass of emissions for species i, A is the area burned (m 2 ), B is the available fuel load (kg m 2 ), CF is the combustion factor (fraction of fuel that is combusted), and EF i is the emission factor for species i (mass of species per dry biomass burned). The values of A were estimated from MODIS Level 3 8-Day Daily Composite Fire Products (MOD14A1 and MYD14A1), which have a spatial resolution of 1km. We mainly used MOD14A1 (terra); MYD14A1 (aqua) was used only for the periods when MOD14A1 was not available. We assumed that the burned area is 180 ha for each pixel assigned to high-confidence fire, following the assumption of Stohl et al. [2007]. This assumption accounts for both the area burned by the detected fire itself and undetected fires in its vicinity on the same day [Stohl et al., 2007]. The values of B, CF, and EF i used in this study are based on previous studies [e.g., Reid et al., 2009] and are summarized in Table 2. Four vegetation types (forest, shrubland/wetland, savannas/ grasslands, and croplands) were estimated from MODIS Yearly Land Cover Type Product for the year Note that the estimation of BB emissions is highly dependent on the assumption shown above and has large uncertainties, as described further below. [13] Tables 1a and 1b show the estimated BB emissions for BC (Gg a 1 ) and CO (Tg a 1 ) within three latitudinal regions (longitude within 60 E 180 E). The BB emissions are also compared with previous studies (Table 1c). BB emissions estimated in this study are generally within the range of previous studies for both BC and CO over North (30 N 70 N) and South Asia (10 N 30 N) and globally, as shown in Table 1c. For example, our CO estimation is consistent with Stroppiana et al. [2010] within 50 S 71 N. BC estimation is Table 1c. Comparison of BC and CO Emissions From Biomass Burning Sources With Previous Studies Emission Species Unit Area This Study a Reference Ratio Reference North Asia CO Tg a 1 60 E 180 E, 30 N 50 N (2000) Streets et al. [2003] CO Tg m 1 60 E E, 33 N 60 N 25 (April 2008) 29 (April 2008) 0.86 FLAMBE, Fisher et al. [2010] CO Tg a 1 45 E 180 E, 26 N 71 N (2003) ~ 1 Stroppiana et al. [2010] BC Gg a 1 60 E 180 E, 50 N 70 N (1996) Bond et al. [2004] BC Gg a 1 30 E E, 33 N 60 N (2008) FLAMBE, Wang et al. [2011] BC Gg a 1 60 E 180 E, 30 N 50 N (2000) Streets et al. [2003] BC Gg a 1 60 E 180 E, 30 N 50 N (1996) Bond et al. [2004] South Asia CO Tg a 1 60 E 180 E, 10 N 30 N (2000) Streets et al. [2003] CO Tg m 1 50 E E, 0 N 33 N 11 (April 2008) 51 (April 2008) 0.22 FLAMBE, Fisher et al. [2010] CO Tg a 1 63 E 180 E, 26 N 50 S (2003) ~ 1 Stroppiana et al. [2010] BC Gg a 1 60 E 180 E, 10 N 30 N (2000) Streets et al. [2003] BC Gg a 1 60 E 180 E, 10 N 30 N (1996) Bond et al. [2004] BC Gg a 1 60 E E, 0 N 33 N (2008) FLAMBE, Wang et al. [2011] Global CO Tg m 1 Global 44 (April 2008) 104 (April 2008) 0.42 FLAMBE, Fisher et al. [2010] CO Tg a 1 Global (2003) < 1 Stroppiana et al. [2010] CO Tg a 1 Global ( ) ~ 1 GFED3, van der Werf et al. [2010] BC Gg a 1 Global (2008) FLAMBE, Wang et al. [2011] BC Gg a 1 Global ( ) ~ 1 GFED3, van der Werf et al. [2010] a Emissions in this study are for the years of

5 Table 2. Parameters for Biomass Burning Emissions Used in This Study Parameters Species Unit Forest Shrubland/Wetland Savannas/Grasslands Croplands Fuel load amount (B) a kg m Combustion factor (CF) a Emission factor (EF) b CO g kg EF BC g kg EF OC g kg EF NO x gkg EF NH 3 gkg EF SO 2 gkg a Values are derived from Reid et al. [2009]. b Values are derived from Chang and Song [2010] and references therein. also generally consistent with Streets et al. [2003] and Bond et al. [2004] within 10 N 50 N, although the studies focused on different time periods. Therefore, we consider that our BB emissions over the Asian continent are within the range of previous emission inventories. [14] The amounts of BC and CO emissions differ considerably between studies, suggesting large uncertainties in the estimations of BB emissions: the range of CO emissions is Tg a 1 ( Tg a 1 in this study), and that of BC emissions is 2 11 Tg a 1 (2 2.5 Tg a 1 in this study) on global scale (Table 1c). As a result, we cannot necessarily validate our BB emissions quantitatively beyond these uncertainties. The reduction of the uncertainties in BB emissions is one of the most important remaining issues facing the community, though it is beyond the scope of this paper. [15] Figures 2a 2c show day-to-day variations of BB emissions in three latitudinal bands. Figures 2d and 2e show regional distributions of BB emissions in 2008 and Most BB occurred during February April at latitudes 10 N 30 N (over South China and Southeast Asia) and during March May at latitudes 30 N 50 N (over Central and North China). As a result, daily BC emissions from BB sources are sometimes greater than those from AN sources (assuming no temporal variations) in spring, while yearly BC emissions from BB sources are not dominant (< 13% of total), as shown in Table 1a. Most BB occurred during April July at latitudes 50 N 70 N, and the amounts of BB emissions are the largest in 2008 and the smallest in 2010 (Figures 2a, 2d, and 2e and Table 1a). The interannual variability of fire activity may be linked with the Arctic Oscillation, El Nino-Southern Oscillation, temperature anomaly, and human activity [e.g., van der Werf et al., 2006, 2010]. 3. Measurements [16] We used both surface and aircraft BC measurements to validate our CMAQ/PASCAL model simulations. The locations of the long-term surface measurements during the simulation periods are shown in Figure 3: Fukue (32.75 N, E), Hedo (26.87 N, E), and Happo (36.68 N, E) stations in Japan. At these stations, BC mass concentrations in the fine mode (PM 2.5, i.e., particles with aerodynamic diameter smaller than 2.5 μm) were measured by a filter-based absorption photometer, the continuous soot monitoring system (COSMOS) [Miyazaki et al., 2008; Kondo et al., 2011b; Kanaya et al., 2013]: March 2009 to March 2010 at Fukue, February 2008 to December 2010 at Hedo, and January 2008 to September 2010 at Happo during the simulation periods. The accuracy of the BC mass concentrations measured by COSMOS is estimated to be about 10% [Kondo et al., 2011b]. Further details of these measurements are presented by Kondo et al. [2011a] and Verma et al. [2011] for the Hedo station and Liu et al. [2013] for the Happo station. [17] We also used aircraft measurements during the Aerosol Radiative Forcing in East Asia (A-FORCE) campaign [Oshima et al., 2012; Moteki et al., 2012; Koike et al., 2012]. This campaign was conducted over the Yellow Sea and the East China Sea in March and April 2009 (Figure 3). A number of aerosol vertical profiles were obtained from near the surface to 9 km in altitude. BC mass concentrations were measured using a single particle soot photometer (SP2) for the dry diameter ranges nm [Moteki and Kondo, 2007, 2010; Kondo et al., 2011a, 2011b]. Oshima et al. [2012] consider the majority of ambient BC mass (> 90%) comprises particles of diameters within the size range of the SP2 used during A-FORCE. We have excluded all data sampled from within cloud, basedonthedefinition by Oshima et al. [2012]. Oshima et al. [2012] defined the observed TE BC values, and their definition is essentially the same as for our model simulations (section 2.1). The A-FORCE campaign is explained in further detail by Oshima et al. [2012]. 4. Results and Validation of Model Simulations 4.1. Comparison With Surface Measurements [18] Figure 4 shows temporal variations of BC mass concentrations at the Fukue, Hedo, and Happo stations for both measurements and model simulations. Simulated BC concentrations were chosen from a horizontal and vertical grid closest to each site or flight track (for surface stations, see this section; for aircraft measurements, see section 4.2). Observed BC mass concentrations have distinct seasonal variations at the Fukue and Hedo sites: the highest concentrations are during spring (March May) and winter (December February), moderate concentrations are during fall (September November), and the lowest concentrations are observed during summer (June August). BC mass concentrations at Happo are also the highest in spring, while seasonal variability is smaller because Happo is the most remote site at the highest altitude (~ 1840 m above sea level). [19] These seasonal variations are generally controlled by the East Asian winter/summer monsoon [e.g., Liu et al., 2002, 2003; Liang et al., 2004; Bey et al., 2001]. In winter, 9952

6 (c) (d) (e) Figure 2. (a c) Day-to-day variations of BC emissions from anthropogenic (red line) and biomass burning sources (green, blue, and red lines) used in this study. (d, e) Regional distributions of BC emissions from biomass burning sources in 2008 and 2010 ( degree grids). the Siberian High over Mongolia and Siberia and the Aleutian Low over the Pacific Ocean lead to strong northwesterly wind in the lower troposphere over the northern part of China, the Yellow Sea, and the East China Sea (> 25 N). As a result, pollution from East Asian countries is transported to the western Pacific efficiently. In summer, the Pacific High is located southeast of Japan, leading to a southerly flow over the East China Sea and the south of Japan (< 40 N in the lower troposphere). The outflow from China is considerably restricted in lower altitudes during this season, and moist clean air is transported from the southern marine area. Spring and fall seasons correspond to the transition period between the winter and summer monsoons. During spring, pollution is transported most frequently to the western Pacific (e.g., the South China Sea) associated with the frequent passages of cold fronts. [20] Model simulations tend to underestimate BC mass concentrations observed at Hedo in spring However, they reproduced both absolute concentrations and seasonal variations of observed BC mass concentrations for other periods very well: normalized mean biases (NMB) and Pearson s correlation coefficients (R) for monthly mean concentrations were 0.7% and 0.91 at Fukue, 22% and 9953

7 Figure 3. Locations of surface measurements (circles) and flight tracks of the A-FORCE aircraft campaign (blue lines) used in this study at Hedo, and 4% and 0.65 at Happo, respectively, during the simulation periods (Table 3). The variability of BC mass concentrations (vertical bars in Figure 4) was also reproduced well by the model simulations. [21] PASCAL can provide information on the source regions and types and transport efficiency (TE BC )ofbc mass concentrations (section 2). Figure 5 shows the results at Hedo, for example. The statistics at the three measurement sites are summarized in Tables 4a and 4b. These results suggest that BC observed at these three sites is transported mostly from AN sources on a 3 year average (90 95%), while the BB contribution becomes larger in spring (10 20%) (Figure 5 and Table 4a). For AN sources, the contribution from China is dominant in spring, fall, and winter (60 80% of total BC), but the AN contribution from Japan becomes larger in summer (25 35% of total BC). These are consistent with the wind fields controlled by the East Asian monsoon, shown above. The contribution from Korea is slightly higher at Fukue (14% of total BC on 3 year average) than at other sites, due to the close proximity of Fukue to Korea. For BB sources, the contribution from China is the largest in early spring (10% of total BC), and Siberia BB contributes 20 30% (of total BC) during the late spring in 2008 and [22] TE BC values have clear seasonal variations: 50 75% in spring, fall, and winter and 25 40% in summer at all three sites during the simulation periods (Table 4b). The lower TE BC values in summer are most likely due to moist air masses influenced by larger amounts of precipitation (resulting in efficient wet removal) during their transport from the south. In contrast, relatively dry air masses influenced by smaller amounts of precipitation are transported from the continent during other seasons. TE BC values are the highest at Fukue (64%) and lower at Hedo and Happo (45 50%). Since Hedo is the most southern site influenced by moist air, and Happo is a mountain site, the air masses transported to both of these sites have more chance to be influenced by wet removal processes during transport. (c) Figure 4. Monthly mean BC mass concentrations at Fukue, Hedo, and (c) Happo sites for both measurements (blue) and model simulations (red). The vertical bars show the standard deviations of BC mass concentrations. 9954

8 Table 3. Statistics for BC Mass Concentrations at the Surface Measurement Sites Station Period Mean Concentration (μgm 3 ) Observation Calculation a NMB (%) Fukue All Spring Summer Fall Winter Hedo All Spring Summer Fall Winter Happo All Spring Summer Fall Winter a Values are calculated for the periods when measurements were available Comparison With the A-FORCE Aircraft Campaign [23] Figure 6 shows the vertical profile of observed BC mass concentrations (in STP) and TE BC values during the R A-FORCE aircraft campaign. Both BC mass concentrations and TE BC values have maxima within the planetary boundary layer (PBL) (> 800 hpa) and decrease with altitude. Large emission sources near the surface and the wet removal of BC during uplift transport (from the PBL to the free troposphere) are responsible for such vertical profiles [Oshima et al., 2012]. Model simulations reproduce this profile reasonably well for both BC mass concentrations and TE BC values, although BC mass concentrations are overestimated by 30% (277 ng m 3 in measurements and 371 ng m 3 in model simulations, average of all data during A-FORCE) and TE BC values are underestimated by 10% (0.78 in measurements and 0.70 in model simulations on average). However, these values are generally within their variability (25th 75th percentiles shown as horizontal bars in Figure 6) at each altitude level. A more detailed description of TE BC vertical profiles is given by Kondo et al. (in preparation). [24] Figure 7a shows a vertical profile observed during flight 8 (March 30) conducted over the Yellow Sea. There are two maxima of observed BC mass concentrations in the PBL and middle troposphere. Model simulations successfully reproduced these two observed peaks, although the absolute concentrations were underestimated within the PBL and overestimated in the middle troposphere. Figure 7b shows the contribution of source regions (c) (d) Figure 5. Source regions and types and transport efficiency of BC mass concentrations at Hedo site estimated by CMAQ/PASCAL simulations: BC mass concentrations from anthropogenic (red) and biomass burning (green) sources, contribution of anthropogenic sources and their source regions to the total BC mass concentrations, (c) contribution of biomass burning sources and their source regions to total BC mass concentrations, and (d) transport efficiency of BC (fractions transported from sources). 9955

9 Table 4a. Source Regions and Types of BC Mass Concentrations at the Fukue, Hedo, and Happo Sites Contribution (%) Station Source Region Source Type All Period Spring Summer Fall Winter Fukue AN Southeast Asia AN China AN Korea AN Japan BB Southeast Asia BB China BB Siberia AN Total BB Total Hedo AN Southeast Asia AN China AN Korea AN Japan BB Southeast Asia BB China BB Siberia AN Total BB Total Happo AN Southeast Asia AN China AN Korea AN Japan BB Southeast Asia BB China BB Siberia AN Total BB Total in the model simulations during this vertical profile. BC in the PBL is mostly from AN sources in Korea and North China, but BC in the middle troposphere is a mixture of AN emissions from Central and North China. Therefore, we conclude that the source regions of BC are distinct between altitudes. 5. Case Study: BC Sources and Transport Mechanisms [25] Since the model simulations generally reproduced observed BC mass concentrations and their temporal variations, we hereby examine the spatial distribution of BC mass concentrations transported from the Asian continent to the Pacific. As a case study, we discuss two BC outflow events seen over Japan and the western Pacific inmay 2008, with model validation for short-term (day-to-day) BC variations. [26] Figure 8a shows a time series plot of BC mass concentrations at Happo in May Model simulations (red line in Figure 8a) reproduced observed BC mass concentrations and their temporal variations well. There are two high BC concentration events with high TE BC values (> 80%) during this month (Figures 8a and 8b): the first is on 6 10 May when high BC was mostly transported from China AN sources, and the other is on May when BC was transported from Siberia BB sources (mixed with BC from China AN sources). We note that the latter event cannot be reproduced successfully without including the contribution from Siberia BB emissions in the model simulations. [27] Figure 9 shows BC mass concentrations and relative humidity (RH) at a sigma level of 0.91 (~ 1 km) before and during two high-bc events. During the first event (an AN source) there is a low-pressure system around 130 E and 50 N, and a cold front is seen from the center of the low pressure southward (5 May, Figure 9a) and southeastward (7 May, Figure 9c). BC is emitted from the AN sources in the central and northern parts of China located over dry regions behind the cold front on 5 May (Figures 9a and 9b) and is transported efficiently from the Asian continent to Japan and the western Pacific by 7 May (Figure 9d). Dry intrusion such as this is an important mechanism in the efficient transportation of pollution to outflow regions [e.g., Fuelberg et al., 2010; Liang et al., 2004]. [28] During the second event (comprising both AN and BB sources), high concentrations of BC were seen over China (20 N 40 N) and around 45 N 50 Non16May, 5 days before the second event was observed at Happo (Figures 9e and 9f). BC at around 45 N 50 N was transported from BB sources in Siberia (emitted at latitudes 50 N 60 N) and had high concentrations (greater than 3 μg m 3 ). BC over China was mostly emitted from AN sources in China. A low-pressure system was formed on 17 May (the center is around 120 E and 50 N), and BC air masses from Siberia BB were gradually transported to the south by this low-pressure system. These BC air masses then merged with BC from other AN sources in China by 19 May (2 days before the second event at Happo) (not shown). The high BC concentrations are located over dry regions behind the low-pressure system and are transported from the Table 4b. Transport Efficiency of BC Mass Concentrations (%) at the Fukue, Hedo, and Happo Sites Station All Period Spring Summer Fall Winter Fukue Hedo Happo

10 Figure 6. Vertical profiles of BC mass concentrations and transport efficiency of BC during the A-FORCE aircraft campaign (data from all the flights shown in Figure 2). Medians (squares) and 25th 75th percentiles of 1 min data (horizontal bars) are shown for both measurements (blue) and model simulations (red). continent to Japan and the western Pacific quite efficiently (TE BC > 80%), with the eastward movement of the lowpressure system (Figures 9g and 9h). [29] Similar high-bc events influenced by BB sources in Siberia were simulated three times in April May 2008 and four times in April May 2009 at Happo (Figures 10a and 10c). In some events, the model severely underestimated observed BC concentrations when Siberia BB emissions were not considered, suggesting their importance during the high-bc events. Siberia BB contributions were also high at the Fukue and Hedo sites during most of the high- BC events influenced by BB sources (Figures 10b and 10d). Therefore, the transport of high BC concentrations from BB sources in Siberia (> 50 N) to the western Pacific was found to be a common phenomenon within the midlatitudes (25 N 50 N). 6. Seasonal Variations of Asian BC Outflow 6.1. BC Mass Concentrations and Eastward Flux Spring [30] Figure 11a shows the regional distribution of mean BC mass concentrations at a sigma level of (3 4 km) in spring (3 years 3 months). This altitude corresponds to that of the maximum eastward BC flux in spring, as shown later. High BC mass concentrations are seen from the Asian continent to the western Pacific at latitudes 20 N 60 N. This regional distribution arises from a combination of Figure 7. Vertical profile of BC mass concentrations obtained over the Yellow Sea during the A-FORCE campaign (flight 8 conducted on 30 March): observed (blue) and simulated (red) BC mass concentrations and simulated contribution from individual source regions (black line shows total BC mass concentrations). 9957

11 Figure 8. Time series plots of observed (blue) and simulated (red and green) BC mass concentrations and transport efficiency of BC at Happo in May China AN (Figure 11b), SEA BB (Figure 11c), and Siberia BB sources (Figure 11d). The individual BC peaks from SEA BB, China AN, and Siberia BB appear over lower latitudes (20 N 30 N), midlatitudes (30 N 50 N), and higher latitudes (50 N 60 N), respectively. [31] Figure 12a shows the mean eastward BC mass flux at 150 E during spring. The BC flux is a product of BC mass concentration and westerly wind velocity, which increases with altitude in this latitude range. The flux is generally large at altitudes around 3 5 km and latitudes 25 N 55 N. The individual peaks of BC flux at latitudes around 30 N, 40 N, and 50 N correspond to the sources from China AN (Figure 12b), SEA BB (Figure 12c), and Siberia BB (Figure 12d), respectively, in accordance with Figure 10: (1) At the peak around 30 N, BB sources from South China and Southeast Asia account for 28% and 8% of the total BC flux, respectively (their sum is comparable to the contribution from China AN (44%)); (2) At the peak around 40 N, AN sources from North, Central, and South China account for 39%, 26%, and 2.0% of the total BC flux, respectively; and (3) At the peak around 50 N, BB sources from Siberia and North China account for 39% and 13% of the total BC flux, respectively. We note that BC from SEA BB and Siberia BB have maxima in March and May, respectively, reflecting temporal variations in BB activities in these regions (section 2.3). The peak altitudes of the BC flux are also different among the individual emission sources. The peaks from the China AN and the Siberia BB contributions (40 N and 50 N, respectively) are seen at lower sigma levels ( corresponding to 2 4 km) because BC mass concentrations from China AN and Siberia BB have maxima in the lower troposphere (sigma level of ) (Figure 12e). This is likely due to relatively weak upward transport associated with midlatitude cyclones (N. Oshima, submitted manuscript, 2013). The peak BC flux from SEA BB is seen around a sigma level of 0.5 (~ 5 km), because BC mass concentrations from SEA BB were the highest in the middle troposphere (sigma level of ). This reflects stronger upward transport at lower latitudes due to convection (N. Oshima, submitted manuscript, 2013), although a greater fraction of BC is removed by wet deposition processes (Figure 12f). This feature is generally consistent with CO eastward flux from BB sources in Southeast Asia reported in previous studies [e.g., Bey et al., 2001; Liu et al., 2002]. These three BC flux peaks can only be identified near the Asian continent at longitudes between 125 E and 150 E, as they merged into one broad peak by 175 E. [32] Figure 12f shows the latitudinal and altitude dependence of TE BC values at 150 E during spring. TE BC values are the highest at latitudes 40 N 55 N in the lower troposphere (50 60%) and decrease southward (20 30% at latitudes 20 N 30 N) and with altitude. This feature is consistent with the latitudinal dependence of TE BC values during the A-FORCE campaign (N. Oshima, submitted manuscript, 2013). The latitudinal dependence of TE BC values is likely due to precipitation amounts (drier air masses and smaller amounts of precipitation at higher latitudes but larger amounts of precipitation at lower latitudes by moister air masses and convection processes) and source regions of emissions (BC is emitted from more eastern part of the continent at higher latitudes resulting in a shorter transport distance from sources to 150 E). This result implies that BC from Siberia BB (higher latitudes) is more efficiently transported to the Pacific than that from China and Southeast Asia (lower and midlatitudes). The latitudinal (source regions and types) dependence of BC transport is discussed further in section Summer, Fall, and Winter [33] Figures 13a 13c show mean eastward BC mass flux at 150 E in summer, fall, and winter. The latitudinal and altitude dependence of BC flux is largely different between seasons. During summer, a peak of BC flux is seen around 50 N in the lower troposphere. These air masses were mostly transported from Siberia BB (91% of total flux at this peak) likely due to cyclones. We note that this peak is not continuous but is formed by occasional high BC flux events. The BC flux at this peak is not very large, because westerly flows in summer are weaker than other seasons (section 4.1). During summer, the maximum BC mass concentrations at 150 E reach 0.24 μgm 3 (sigma level around 0.9), which is greater than the peak BC concentrations in spring (0.18 μgm 3 ). In contrast, BC mass 9958

12 % µg m -3 (c) (d) % µg m -3 (e) (f) % µg m -3 (g) (h) % µg m -3 Figure 9. Snapshots of (a, c, e, and g) relative humidity and (b, d, f, and h) BC mass concentrations in model simulations at a sigma level of 0.91 (~ 1 km) at 00 UT on 5 May (Figures 9a and 9b), 7 May (Figures 9c and 9d), 16 May (Figures 9e and 9f), and 22 May (Figures 9g and 9h). Arrows show horizontal winds. 9959

13 (c) (d) Figure 10. Time series plots of observed (blue) and simulated (red and green) BC mass concentrations at (a and c) Happo and (b and d) Hedo during April May 2008 (Figures 10a and 10b) and 2009 (Figures 10c and 10d). concentrations from AN sources in North and Central China are predicted at around 40 N in the middle and upper troposphere but are considerably smaller in magnitude. BC ascended over Central China continuously during summer but was not transported to the Pacific efficiently due to weak westerly flow. This result is generally consistent with previous studies, which showed that the pollution (such as CO) over Central China could be lifted to the upper troposphere by convection and be transported by Westerlies in the midlatitudes during summer. However, a large portion of this outflow is transported southwestward, and only a limited fraction is transported to the Pacific [Liang et al., 2004; Liu et al., 2002]. In addition, BC is mostly removed from the atmosphere by wet deposition during convection with heavy precipitation. In fact, TE BC values around 40 N and a sigma level of 0.4 were very low ( ) at 150 E (not shown). [34] During fall, a peak is seen at latitudes 40 N 50 N in the lower troposphere. At the peak, AN sources from North and Central China account for 64% and 15% of the total flux, respectively. Eastward and upward BC transport is generally due to midlatitude cyclones. [35] During winter, two peaks are seen around 30 N and 40 N in the middle troposphere. The peak around 30 N is mainly from AN sources in Central China (39% of the total flux, around the Sichuan Basin), South China (21%), and Southeast Asia (16%, mostly from India). The peak around 30 N is seen at higher altitude than that at 40 N due to orographic uplift transport around the Sichuan Basin (100 E 105 E). The uplift transport from the Sichuan Basin was also predicted frequently during other seasons. However, the corresponding BC flux reached a maximum in winter due to the strongest westerly flow. The importance of orographic uplift for the transport of BC during the A-FORCE period is also shown in N. Oshima (submitted manuscript, 2013). in detail. At the peak around 40 N, the contribution from North and Central China to the total BC flux was 54% and 29%, respectively. This transport is mainly due to northwesterly flow induced by the winter monsoon and midlatitude cyclones. While the BC flux in winter is larger than other seasons due to stronger westerly flow, the maximum BC mass concentrations are not so high (0.07 μg m 3 around a sigma level of 0.9) at 150 E. The contribution from BB sources is negligible in fall and early winter (between September and January, Figure 2). [36] The latitudinal and altitude pattern of TE BC in summer, fall, and winter is generally similar to that in spring (not shown), while absolute TE BC values are different between seasons: maximum in spring and winter and minimum in summer (section 6.2) Importance of BB Sources Within Asian BC Outflow [37] Figure 14 shows the relationship between BC emissions over the Asian continent and eastward BC flux at 150 E, for three sources (China AN, SEA BB, and Siberia 9960

14 µg m -3 µg m -3 (c) (d) µg m -3 µg m -3 Figure 11. Regional distribution of mean BC mass concentrations at a sigma level of (3 4 km) in spring (3 years 3 months): total BC mass concentrations, BC from anthropogenic sources in China, (c) BC from biomass burning sources in South China and Southeast Asia, and (d) BC from biomass burning sources in Siberia. Arrows show horizontal winds. BB). The BC flux (net flux) was calculated within latitudes 10 N 70 N from the surface to a sigma level of 0.2 (8 9 km). The statistical values for individual seasons and sources are shown in Table 5. During spring, the mean eastward flux was 5.5, 13.3, and 2.5 kg s 1 for Siberia BB, China AN, and SEA BB, respectively (Figure 14a). Within the total eastward BC flux, the contribution from BB sources was estimated to be 37% (Table 5), suggesting the importance of BB sources in spring. The ratios of eastward flux at 150 E to emissions over the continent (defined as transport fraction, TF) were highest in spring and were estimated to be 32%, 23%, and 11% for Siberia BB, China AN, and SEA BB, respectively (Figure 14a). TF is used here (instead of TE BC ) because it is possible to define TF values for each source region, whereas TE BC values cannot be defined for each source in PASCAL. We note that the definitions of TF and TE BC are similar but not identical. TF is the ratio of transported BC to the amounts of emissions and is dependent on both transport pathway (westerly or not) and removal processes during transport. TE BC is the ratio of BC to maximum transport amounts (equivalent to the amounts in assuming no deposition processes) and is only dependent on removal processes. TF values for all sources were lower than TE BC values, but they had a similar spatial and seasonal tendency in our study area. The TF values increase with latitude (Siberia BB > China AN > SEA BB), consistent with the latitudinal dependence of TE values shown in section The latitude dependence of TF values is mainly due to that of precipitation amounts (smaller amounts of precipitation at higher latitudes) and source regions of emissions (emitted from more eastern parts of the continent at higher latitudes), as shown in section [38] During summer, the mean eastward flux was 4.7 and 3.4kgs 1 for Siberia BB and China AN, respectively (Figure 14b). The contribution from SEA BB is negligible. The TF values in summer were the lowest and were estimated to be 26% for Siberia BB and 6% for China AN. Since BC from Siberia BB is transported more efficiently (compared with China AN), the eastward flux from Siberia BB is greater than that from China AN, while the total amount of emissions are a factor of 3 smaller than China AN. These results are consistent with the implication in section that BC from Siberia BB is transported to the Pacific moreefficiently. [39] The mean eastward flux from China AN was 7.0 kg s 1 infalland11.1kgs 1 in winter (Figures 14c and 14d). The TF values were 12% in fall and 19% in winter. The contribution from BB sources is small, except for February (SEA BB starts in February). Since BC from SEA BB is not transported to the Pacific efficiently, its contribution becomes much smaller during transport (at 150 E) than emission sources, as shown in Figure 14d. [40] Figure 15 shows the monthly mean eastward BC mass flux at 125 E, 150 E, and 175 E (3 year average). It is evident from Figure 15 that in addition to the AN emissions (specifically, China AN), the flux from BB emissions is greatest in spring (March April), leading to an enhancement in the BC flux in this season. On the contrary, the fluxes from 9961

15 MATSUI ET AL.: TAGGED SIMULATION OF ASIAN BC OUTFLOW µg m-2 s-1 µg m-2 s-1 (d) (c) µg m-2 s-1 µg m-2 s-1 (f) (e) µg m-3 Figure 12. Mean eastward BC flux, mass concentrations, and transport efficiency at 150 E during spring (3 years 3 months): total eastward BC flux, BC flux from anthropogenic sources in China, (c) BC flux from biomass burning sources in South China and Southeast Asia, (d) BC flux from biomass burning sources in Siberia, (e) total BC mass concentrations, and (f) transport efficiency of BC. 9962

16 (c) µg m -2 s -1 µg m -2 s -1 µg m -2 s -1 Figure 13. Mean eastward BC flux at 150 E in summer, fall, and (c) winter. both AN and BB emissions reach a minima in August and September. Therefore, BB emissions increase the seasonal variability of eastward BC flux over the western Pacific. As described above, these results are due to a combination of the seasonal variations in BB emissions, transport pathway (eastward or not), westerly wind velocity, and TE BC values. TE BC values were the highest in spring and winter and the lowest in summer at three longitudes: mean TE BC values of 45 60% (125 E), 30 35% (150 E), and 20 30% (175 E) during spring and winter and 30% (125 E), 20% (150 E), (c) (d) Unit: kg s -1 Figure 14. Relationship between BC emissions (kg s 1 ) over the continent and eastward BC mass flux (kg s 1 ) at 150 E for three sources from biomass burning sources in Siberia (Siberia BB), anthropogenic sources in China (China AN), and biomass burning sources in South China and Southeast Asia (SEA BB) for (a d) each season. The values below the arrows correspond to the transport fraction (TF); the ratio of eastward BC mass flux to emissions for individual sources. 9963

17 Table 5. Statistics of Eastward BC Mass Flux and Its Contribution From Individual Source Regions and Types at 150 E Item Unit All Period Spring Summer Fall Winter Eastward flux kg s Transport efficiency % Southeast Asia AN % South China AN % Central China AN % North China AN % Korea AN % Japan AN % Southeast Asia BB % South China BB % Central China BB % North China BB % Siberia BB % Total AN % Total BB % and 15% (175 E) during summer (Figure 15). These seasonal variations are consistent with the results at surface measurement sites (section 4.1) and TF values (shown above). [41] In summary, the contributions to the total eastward BC flux were estimated to be 61%, 17%, and 6% for China AN, Siberia BB, and SEA BB, respectively, over the western Pacific (at 150 E) on 3 year average (Table 5 and Figure 15). AN sources are the largest sources of BC over the western Pacific. However, BB sources in South China, Southeast Asia, Siberia, and Kazakhstan contribute 30 50% of total eastward BC flux in spring and summer and increase its seasonal variability. BC from Siberia and Kazakhstan sources is transported more efficiently than that from other sources and is found to contribute more than 20% of total eastward flux during April August. As mentioned in section 2.3, the estimation of BB emissions currently has large uncertainties. However, our results suggest that the control of BB emissions in Siberia and Kazakhstan will be important in terms of transboundary transport of BC to the Pacific, North America, and the Arctic. Therefore, there is an urgent need to reduce the uncertainties in the estimation of BB emission amounts. Finally, since BB emissions in Siberia and Kazakhstan may continue to increase [Stocks et al., 1998; Westerling et al., 2006; Soja et al., 2007], their impact on the Pacific would be more important in the future. 7. Summary [42] In this study, we examined source regions and types (anthropogenic (AN) and biomass burning (BB)) of BC outflow from the Asian continent to the Pacific during using a tagged version of the CMAQ model (CMAQ/ PASCAL). Model simulations reproduced absolute BC mass concentrations and temporal (seasonal, monthly, and day-today) variations observed at three surface measurement sites (Hedo, Fukue, and Happo) very well. Model simulations also reproduced the vertical profiles of BC mass concentrations and their transport efficiency (TE BC, fraction transported from sources) during the A-FORCE aircraft campaign, which was conducted over the Yellow Sea and the East China Sea in March and April [43] A case study (May 2008) showed that BC from BB sources in North Asia (> 50 N, Siberia BB) could be transported to Japan and the western Pacific quite efficiently (TE BC > 80%). This is because high BC concentrations from Siberia BB are located over dry regions behind a cold front, and they were transported southeastward with the movement of the low-pressure system. High-BC events influenced by Siberia BB were simulated three times in April May 2008 and four times in April May 2009 over Japan. The transport of high BC concentrations from Siberia BB (> 50 N) to the (c) Figure 15. Monthly mean eastward BC flux at 125 E, 150 E, and (c) 175 E. Colored bars show the contribution from individual source regions and types (left axis). The black lines show the transport efficiency of BC (right axis). 9964