Regional climatology of particulate carbonaceous substances in the northern area of the east Asian Pacific rim

Authors


Abstract

[1] A year-round observation of atmospheric aerosols and their associated species was conducted from March 2001 to May 2002 on Rishiri Island in the northern area of the east Asian Pacific rim region. Asian outflows brought continental air masses to this area during the period from the autumn to the spring although marine air masses from the high-latitudinal ocean often broke into this area during the midwinter. In contrast, marine air masses were predominant over this area in the summer. Particulate elemental carbon (EC) would be mainly transported with biomass smoke particles, and seasonal variation in its concentrations was well correlated with the air mass alternation, showing higher concentrations during the period from the autumn to the spring with some decreases in the midwinter. The concentrations of particulate organic carbon (OC) showed a similar seasonal trend with those of the particulate EC, but relatively high concentrations were found in the summer due to photochemical secondary productions. The particulate OC that is vaporized by heating at higher temperatures (OCHT) would be mainly carried with the biomass smoke, and the particulate OC that is vaporized at lower temperatures (OCLT) would be mainly caused by secondary production processes. Summer enhancements of the secondary species, such as OC and nss-SO42–, caused decreases in the ratio of EC/PM2.5, which would contribute to the high single scattering albedo of fine aerosols in the summer. Aerosols in the Asian outflows in this area are relatively “black” in the winter, although the aerosol mass loading increases in the spring.

1. Introduction

[2] The atmospheric transport of crustal and anthropogenic materials from the Asian continental region to the remote Pacific Ocean have a significant impact on the global climate forcing and atmospheric chemistry. Many studies have focused on the chemical properties, long-range transport, and depositional flux of Asian dust [Duce et al., 1980; Uematsu et al., 1983] and anthropogenic aerosols [Prospero et al., 1985; Prospero and Savoie, 1989] over the remote Pacific Ocean. Recently, the Asia-Pacific Regional Aerosol Characterization Experiment (ACE-Asia), integrating in situ measurements, remote sensing observations, and model studies, was carried out over the northwestern North Pacific in the spring of 2001 to understand the physical, chemical, and radiative properties of the aerosols in the “Asian outflows” [Huebert et al., 2003]. Several intensive aircraft campaigns, such as the Transport and Chemical Evolution over the Pacific (TRACE-P), the Pacific Exploration of Asian Continental Emission-A (PEACE-A), and the PEACE-B, have also been conducted over the northwestern North Pacific during the period from 2001 to 2002, which characterized the chemical properties of the Asian outflows [Jacob et al., 2003; Parrish et al., 2004].

[3] The east Asian Pacific rim is a region where anthropogenic and mineral aerosols first flow out from the Asian continental region. In order to understand the transport patterns and chemical properties of the Asian outflows, the long-term observations of aerosols and their associated species should be conducted in the east Asian Pacific rim. In past studies, several researchers have reported the seasonal characteristics of the outflows from the measurements of aerosols [e.g., Mukai and Suzuki, 1996; Matsumoto et al., 1998] and gaseous species [e.g., Pochanart et al., 1999; Tanimoto et al., 2000] on remote islands around the Japanese Islands, and found the frequent long-range transport of continental air masses during the winter-spring periods. These studies have provided a better understanding of the chemical properties and transport patterns of the Asian outflows over the east Asian Pacific rim.

[4] In order to comprehensively discuss the geographical characteristics of the chemical and physical properties of the Asian outflow over the northern area of the east Asian Pacific rim, we conducted a year-round observation of the atmospheric aerosols and their associated species from March 2001 to May 2002 on Rishiri Island near the northern tip of Japan. This paper mainly focuses on the carbonaceous substances in accumulation-mode aerosols that are important species for the estimation of the radiative properties of the Asian outflows.

[5] Accumulation-mode aerosols make large contributions to atmospheric radiation fields [Hobbs, 1993]. Especially, particulate carbonaceous substances including organic carbon (OC) and elemental carbon (EC) have been the focus of recent studies due to their importance when discussing the radiative properties of aerosols. Particulate OC effectively scatters solar radiation [Gelencsér, 2004], contributes to the formation of cloud droplets [Novakov and Penner, 1993], and affects the cloud albedo [Takemura et al., 2000]. Soot particles consist of EC that is the principal light-absorbing species in atmospheric aerosols [Seinfeld and Pandis, 1998]. Furthermore, soot particles frequently adsorb hydrophilic materials [Okada et al., 1992] and are incorporated into cloud droplets, which reduces the cloud albedo [Kaufman and Nakajima, 1993]. In spite of their importance to aerosol radiative properties, annual observations of particulate carbon have been rarely reported over the east Asian Pacific rim compared to ionic and mineral species.

[6] In past studies, annual observations of aerosol species over the east Asian Pacific rim have been conducted on remote islands such as the Oki Islands [Mukai et al., 1990; Mukai and Suzuki, 1996], Cheju Island [Carmichael et al., 1996, 1997], the Nansei Islands [Suzuki and Tsunogai, 1988; Kaneyasu and Takada, 2004], and the Ogasawara Islands [Matsumoto et al., 1997, 1998]. These islands are located in the southern area of the east Asian Pacific rim; southwestern area of the Japan Sea, the East China Sea, and the western North Pacific Ocean. Recent studies, however, have suggested that heavy plumes of anthropogenic substances frequently reach the northern area of the east Asian Pacific rim during the spring [Matsumoto et al., 2003b].

[7] In order to estimate the transport budget of continental aerosols from the Asian continental region to the remote North Pacific and depositional flux of airborne materials to the marginal seas, such as the Okhotsk Sea and the Japan Sea, long-term observations of aerosol species should be conducted on a remote site in the northern area of the east Asian Pacific rim that has been rarely focused on in the past aerosol studies compared to the southern area. The pioneering study of the annual aerosol observations in the northern area were conducted on Okushiri Island, about 400 km south of Rishiri in the Japan Sea, by Tsunogai et al. [1985], although they focused only on mineral particles. Recently, Okuda et al. [2006] reported the long-term observation of trace metals in aerosols on Rishiri Island, showing the frequent transport of particulate contaminants from the Asian continental region to Rishiri.

[8] In this paper, we report the seasonal characteristics of carbonaceous substances in accumulation-mode aerosols over the northern area of the east Asian Pacific rim in order to discuss the regional climatology of aerosol chemistry and its implication to the atmospheric radiation field over this area.

2. Observations

2.1. Site Description

[9] Observations were carried out on Rishiri Island near the northern tip of Japan. The location of the observation site is shown in Figure 1. The measurements of particulate carbonaceous substances, ion species, and carbon monoxide (CO) were made at an observatory that was constructed close by the Rishiri National Acid Rain Monitoring (RNARM) station under the Acid Deposition Monitoring Network in East Asia (EANET). The mass concentrations of aerosols with d < 2.5μm (PM2.5) and nitrogen monoxide (NO) concentrations were continuously monitored at the RNARM station. Detailed explanations of the site have been reported by Tanimoto et al. [2002]. There is a small residential area and a small fishing port about 1 km south of the observatory. The influence of local contaminations from them on the measurements was found at rare intervals, which can be assessed by the spike-like enhancement of particulate elemental carbon (EC) with 3 times higher concentrations than the just prior 12-hour-averaged concentrations. Local contaminations were also cross-checked using the CO and NO concentrations. The data affected by them were removed from the following discussion.

Figure 1.

Location of the sampling site (solid square).

2.2. Instruments

2.2.1. Particulate Carbonaceous Substances

[10] Measurements of the particulate carbonaceous substances in the accumulation-mode range were made using an ambient particulate carbon monitor (ACPM) (Rupprecht and Patashnik Co., Inc., Series 5400). Details of the sampling and analytical procedures have been described in our previous papers [Matsumoto et al., 2003c], of which brief descriptions are given in the following.

[11] The air sample was taken from the rooftop of the observatory, introduced into a cyclone with a 50% effective cutoff diameter of 2.5 μm to eliminate the coarse particles, and then delivered to an ACPM at 16.7 L min−1. The ACPM collected the ambient aerosols by an impactor, of which the 50% effective cutoff diameter is 0.14 μm at the flow rate, with intervals of 4 hours, and then automatically measured the concentrations of the carbonaceous substances in the aerosols by thermal analysis. The collection plates of the impactor were heated at 50°C to minimize the adsorption of gaseous organic matter during the sampling. After sampling, the impaction plate was sequentially heated in four steps; 200°C, 250°C, 340°C, and 750°C. In this study, the OC and total carbon (TC) were defined as the carbonaceous substances evolved below 340°C and below 750°C, respectively. The difference between the amounts of the TC and OC was then defined as the amount of EC.

[12] The ACPM measurements include several uncertainties. The ACPM cannot collect the particles with diameters smaller than 0.14 μm. Höller et al. [2002] reported that the fraction of EC with d < 0.14 μm can be estimated to be about 30% of the EC with d < 2.5 μm at a coastal rural site in Japan. Meanwhile, a smaller contribution (<10%) of EC with d < 0.14 μm to that with d < 2.5 μm was found in the remote marine atmosphere [Kaneyasu and Murayama, 2000]. With regard to the EC measurements, on the other hand, the possibility of their overestimation is also pointed out because of the influences of pyrolysis OC, since a thermal technique without any correction for the pyrolysis was used in the analytical mode of the ACPM [Rice, 2004].

[13] Comparison experiments for the EC measurement between the ACPM and the integrated filter-thermal optical method have found good correlations, but significant overestimations of EC from the ACPM. The ratios of EC from the ACPM to that from the thermal optical method ranged from 1.2 to 1.9 probably because of the effect of the pyrolysis OC [Watson and Chow, 2002; Lim et al., 2003; Rice, 2004]. On the other hand, underestimations of the EC measurement by the ACPM compared to the thermal optical method have also been reported with the ratio of 0.65 [Cowen et al., 2001]. The ACPM measurements of the EC can cause both an overestimation and underestimation, of which the direction and magnitude would be dependent on the chemical and physical properties of the aerosols, such as the chemical composition of the OC (pyrolysis OC concentration) and size distribution of the EC.

[14] OC measurements from the ACPM also include several uncertainties. Our preliminary experiment found that the ACPM with an absolute filter in front of the inlet detected significant concentrations of the OC [Matsumoto and Uematsu, 2007], which would be caused by the adsorption of organic gases and lead to an overestimation of the OC [Matsumoto et al., 2003a]. On the other hand, the ACPM readings are also in danger of showing a negative bias due to evaporation of the organic matter collected on the plate heated at 50°C [Cowen et al., 2001]. Comparison experiments of the OC measurements between the ACPM and the integrated filter-thermal optical method have also shown both an overestimation and underestimation with the ratios (the ACPM/the integrated filter-thermal optical method) from 0.2 to 1.5 [Watson and Chow, 2002; Lim et al., 2003; Rice, 2004]. The direction and magnitude of the uncertainties would also be dependent on the chemical and physical properties of aerosols, such as the size distribution of the OC, chemical composition of the OC (volatility of OC), and concentration of the gaseous semivolatile OC.

2.2.2. Particulate Ion Species, PM2.5, CO, and NO

[15] For the chemical analysis of ion species in fine aerosols, ambient aerosols segregated into two size fractions (d < 2.5 μm and d > 2.5 μm) were collected at intervals of about 7 days or less on PTFE fiber filters (Toyo Roshi Kaisha, Ltd., PF040) by a high-volume dichotomous virtual impactor air sampler equipped with an automatic filter change system (Kimoto Electric Co., Ltd., Model ACS21) at the flow rate of about 180 L min−1. The filter samples were ultrasonically extracted with ultrapure water (specific resistivity >18 MΩcm), and then analyzed for major ion species by ion chromatography (Dionex, DX-120), of which the analytical conditions have been described in our previous paper [Matsumoto et al., 2003c].

[16] The PM2.5 measurements were made using a TEOM ambient particulate monitor (Rupprecht and Patashnik Co., Inc., Series 1400). Details of the precision and accuracy of the TEOM were reported by Jaques et al. [2004]. The concentrations of CO and NO that were used for the detection of the local contaminations in our measurements were also continuously monitored using a modified NDIR photometer instrument, of which the details were described by Tanimoto et al. [2002], and a commercial chemiluminescence NO analyzer, respectively. The one-hour averaged data of the PM2.5 and NO were supplied by the Acid Deposition and Oxidant Research Center [2004].

3. Results and Discussion

3.1. Seasonality of Air Mass Alternation

[17] In order to clarify the origins of the air masses arriving at Rishiri, the 7-day backward trajectories were calculated with modeled vertical velocities using the hybrid single-particle Lagrangian integrated trajectory (HY-SPLIT 4) model (NOAA Air Resources Laboratory, http://www.arl.noaa.gov/ready/hysplit4.html). The trajectories were calculated once daily at 0000 GMT. The starting altitudes were set at the top of the boundary layer over the site, which was estimated from vertical profiles of the potential temperature obtained from the Wakkanai Observatory of the Japan Meteorological Agency located about 50 km east of Rishiri Island. The trajectories were classified into seven types according to their directions as shown in Figure 2.

Figure 2.

Classification of air trajectories into seven categories. “OS,” “BS,” “PO,” “RU,” “NC,” “SC,” and “JP” denote “Okhotsk Sea,” “Bering Sea,” “Pacific Ocean,” “Russia,” “North China,” “South China,” and “Japan,” respectively.

[18] Figure 3 shows the appearance frequency of each air mass from March 2001 to May 2002. The continental air masses flowed out to the site with a high frequency during the period from the autumn to the spring. In particular, the air masses from the region of RU and NC were frequently found at the site during this period. In contrast to the continental air masses, the marine air masses, especially from the OS region, often covered the site in the summer.

Figure 3.

Monthly appearance frequency of the air trajectories based on the categories shown in Figure 2. UK denotes air masses that were not classified into any category. Shaded columns (RU, NC, SC, and JP) are continental regions, and the other three columns (OS, BS, and PO) are marine regions.

[19] These air mass alternations are caused by synoptic-scale weather systems. Continental anticyclones are developed over East Siberia and, therefore, air masses flow out from the continental region to the North Pacific during the period from autumn to spring. In the transitional seasons of spring and autumn, especially, traveling anticyclones frequently move eastward and effectively transport air masses from the continental region to the east Asian Pacific rim. Alternatively, a marine anticyclone is developed over the Pacific Ocean in the summer, which brings clean marine air masses to the site. As examples, the monthly mean sea level pressure distributions over the western North Pacific region in the spring (March 2002) and the summer (August 2001) are shown in Figures 4a and 4b, respectively.

Figure 4.

Monthly mean sea level pressure distributions around Japan (a) in the spring (March 2002), (b) in the summer (August 2001), and (c) in the midwinter (December 2001).

[20] It is interesting to note that, during the midwinter from December to January, the air masses from the OS region were often found at the site as shown in Figure 3. The monthly mean sea level pressure distribution in December 2001 is also shown in Figure 4c. During this period, large low-pressure systems often stagnated and developed east of Japan, and this weather system would cause transport of the air masses from the OS region to the site. This weather system is important for aerosol chemistry during the winter period as discussed below.

3.2. Seasonal Characteristics of the Particulate EC

3.2.1. Seasonal Variation of the Particulate EC Concentration

[21] There are few reports on the annual measurement of particulate EC at remote sites over the east Asian Pacific rim, in spite of its importance for the atmospheric radiation field. Recently, Kaneyasu and Takada [2004] reported the annual measurements of particulate EC at the Nansei Islands (Amami and Miyako Islands) over the eastern edge of the East China Sea, showing a winter-spring enhancement of its concentrations.

[22] The concentrations of the particulate EC at Rishiri are summarized in Table 1, and their temporal variations are plotted in Figure 5a. Higher concentrations were found during the period from the autumn to the spring and lower in the summer, meaning that the continental outflows bring particulate EC to this area. In the midwinter from December to January, the particulate EC concentrations decreased, which corresponded to the frequent intrusion of the marine air masses from the OS region.

Figure 5.

Temporal variation in the concentrations of the particulate EC. (a) Open circles and solid squares indicate daily and monthly means, respectively. (b) Daily means for the samples in continental air masses. (c) Daily means for the samples in marine air masses.

Table 1. Concentrations of Particulate EC and OCa and the Ratios of OC/EC in All, Marine, and Continental Air Masses
 ECOCOC/EC
AverageSDAverageSDAverageSD
  • a

    Unit is μg m−3.

In All Air Masses
Spring0.270.210.690.453.242.24
Summer0.120.110.610.315.942.05
Autumn0.210.230.700.634.221.89
Winter0.260.190.450.261.920.40
Total0.230.200.620.433.702.13
 
In Marine Air Masses
Spring0.100.070.420.265.033.85
Summer0.070.040.510.257.171.74
Autumn0.060.030.300.135.881.99
Winter0.100.040.210.062.300.51
Total0.080.050.410.245.503.00
 
In Continental Air Masses
Spring0.330.210.760.462.671.18
Summer0.200.140.770.344.481.52
Autumn0.270.250.810.673.471.26
Winter0.310.190.530.261.790.27
Total0.300.210.710.462.831.38

[23] Scavenging by precipitation would be a significant removal process for the aerosols, and rainfall amounts at Rishiri showed seasonality with increasing in the summer. The mean concentrations of the particulate EC for the rainy (≥1 mm/day) and no-rain days, however, were 0.26 and 0.29 μg m−3, respectively, in the transitional seasons, and 0.11 and 0.12 μg m−3, respectively, in the summer. Statistically significant differences between the EC concentrations for the rainy and no-rain days were not found with a significance level of 5% (P = 0.05) in both seasons, which suggests that the rainfall events cannot be regarded as a major factor to explain the seasonal variation in the EC concentrations. Air mass alternations would be the principal factor for the regional climatology of the particulate EC.

[24] The temporal variations in the particulate EC concentrations in each air trajectory are shown in Figures 5b and 5c. Higher concentrations of the particulate EC were found in the continental air masses with the mean concentration of 0.30 μg m−3, whereas its concentrations decreased to lower values with the mean of 0.08 μg m−3 in the marine air masses.

[25] It is interesting to note that the seasonal variation in the particulate EC concentrations were also found in the continental air masses; relatively lower concentrations were frequently detected in the summer. A similar seasonal trend has also been found at Rishiri for the atmospheric radon concentration in the continental air masses (K. Yoshioka, private communication, 2005). The mixing layer depth and air mass transport altitude did not show a clear seasonality. It is possible that marine air masses frequently penetrate into the continental region and have an influence on the continental air masses in the summer, which can cause a summer decrease in the particulate EC concentrations in the continental air masses. The influence of the summer monsoon on the atmospheric chemistry of Eurasian continental air masses should be investigated in a future study. In contrast to the case of the continental air masses, the concentrations of the particulate EC in the marine air masses show a small seasonal variation with low concentrations due to a small influence of the sources.

3.2.2. Potential Source of the Particulate EC

[26] EC in the aerosols is mainly derived from combustion processes including both fossil fuel and biomass combustions and, therefore, have been regarded as a signal for anthropogenic aerosols and biomass smoke particles. Figure 6 shows the relationship between the concentrations of the EC and nss-K+ in the fine aerosols. Nss-K+ is included in biomass smoke particles with high concentrations [e.g., Novakov and Corrigan, 1996; Lee et al., 2005] and has been regarded as a tracer for biomass combustions. Although nss-K+ is also included in mineral aerosols [Minakawa and Uematsu, 2001], no significant correlation between the nss-Ca2+and nss-K+ was found in fine aerosols from our observation, which implies that a large portion of the nss-K+ in the fine aerosols was derived from the biomass combustion. A good correlation between the concentrations of the EC and nss-K+ suggests that the particulate EC is transported with the biomass smoke particles from forest fires and/or biomass fuel combustion.

Figure 6.

Relationship between EC and nss-K+ concentrations in the fine aerosols.

[27] The relationship between the EC and nss-SO42– in fine aerosols was more scattered than the case for nss-K+ as shown in Figure 7, although past researchers found good correlations between both species [e.g., Kaneyasu and Murayama, 2000]. Biomass combustion processes are not significant sources for particulate nss-SO42– [Novakov and Corrigan, 1996]. The fact that the concentrations of EC were better correlated with nss-K+ than with nss-SO42– suggests that the EC is more strongly affected by the biomass smoke transport than by the fossil fuel combustion sources over this area. A multiple linear regression between the concentrations of the EC (dependent valuable), nss-K+ and nss-SO42– (independent valuables) also suggests a greater contribution of the variation of the nss-K+ concentrations to that of the EC concentrations. The standardized partial regression coefficients for independent valuables of the nss-K+ and nss-SO42– concentrations are 0.65 and 0.30, respectively.

Figure 7.

Relationship between EC and nss-SO42– concentrations in the fine aerosols. Open and shaded circles indicate the samples from May to August and those from September to April, respectively. Solid line indicates linear regression for all samples, and dashed line indicates that for the samples from May to August.

[28] Photochemical oxidation of SO2 to nss-SO42– would have a seasonal variability and, moreover, SO2 is also derived from the oxidation of oceanic dimethylsufide (DMS), which can also be reasons for the poorer correlation between the EC and nss-SO42–. Indeed, the samples from May to August had high ratios of nss-SO42– to EC as shown in Figure 7. This is consistent with the seasonal trend in the ratios of EC/PM2.5 as mentioned below.

[29] The slopes of the regression lines for the samples from all seasons and for those from May to August were 4.0 and 7.3, respectively. Kaneyasu and Murayama [2000] reported that the slope of the regression line between the EC and nss-SO42– in fine aerosols over the central North Pacific Ocean was approximately 9.1. Compared with nss-SO42–, the EC was relatively enriched in the aerosols obtained by our observations, which suggests a significant influence of the biomass smoke transport or a small influence of the photo-oxidation processes on the aerosols over the northern area of the east Asian Pacific rim. Previous researchers have suggested that frequent occurrences of boreal forest fires have significant impacts on the atmospheric chemistry over the northern area of the east Asian Pacific rim [Tanimoto et al., 2000].

3.2.3. Optical Implication of the Particulate EC

[30] Although the outbreaks of both the particulate EC and PM2.5 are accelerated during the period from autumn to spring in this area, the proportions of EC to PM2.5 would be an important factor when discussing the radiative properties of the aerosols in the Asian outflows. Figure 8a shows the ratios of the EC concentration to PM2.5 during our observations, showing higher values in the winter, and a decrease after the middle spring, which is somewhat different from the trend in the absolute concentrations of the particulate EC and PM2.5.

Figure 8.

Temporal variation in the ratios of the particulate EC concentration to PM2.5. (a) Open circles and solid squares indicate daily and monthly means, respectively. (b) Daily means for the samples in continental air masses. (c) Daily means for the samples in marine air masses.

[31] An increase in the ratios of the EC to aerosol mass contributes to a decrease in the single scattering albedo of the aerosols [Panchenko et al., 2001; Takeuchi et al., 2004]. Our result shows that aerosols in the Asian outflows over the northern area of the east Asian Pacific rim are relatively “black” in winter, although the aerosol mass loading is increasing in the transitional seasons, especially in spring. No statistically significant difference between the ratios of the EC to PM2.5 for the rainy days and for the no-rain days was found for P = 0.05. This suggests that the particulate EC would be strongly associated with hygroscopic species in the fine aerosols over this area, which has been also suggested by past studies [e.g., Okada et al., 1992].

[32] Temporal variations in the ratios of the EC to PM2.5 in each air trajectory are also shown in Figures 8b and 8c. A seasonal variation in the ratios of the EC to PM2.5 was also found in the continental air masses with lower ratios in the late spring and summer. In the late spring and summer, aerosol production through secondary formation processes would be accelerated and secondary particulate species would significantly contribute to the PM2.5, which can cause decreases in the ratios of the EC to PM2.5. Similar seasonal trends were also found in the marine air masses, but the ratios were lower than that in the continental air masses, which would also be attributable to the large contribution of secondary particulate species from precursor gases to PM2.5 in the marine air masses. Similar to the EC concentrations, the influences of marine air masses with lower ratios of the EC to PM2.5 on summer continental air masses may also contribute to the seasonal trend of the ratios in the continental air masses.

3.3. Seasonal Characteristics of the Particulate OC

3.3.1. Seasonal Variation in the Particulate OC Concentration

[33] Particulate OC is an important component to control the hygroscopic properties of aerosols and contributes to the scattering of incoming solar radiation through a direct process by the aerosol itself and an indirect process via cloud formation. Similar to the particulate EC, there is a small data set of annual measurements of particulate OC at a remote site over the east Asian Pacific rim. Concentrations of the particulate OC at Rishiri are also summarized in Table 1, and their temporal variation is shown in Figure 9a. Although the concentrations of the particulate OC showed a similar seasonal trend to the particulate EC with higher values in the autumn and the spring, its seasonal variation was not as clear when compared to that of the particulate EC.

Figure 9.

Temporal variation in the concentrations of the particulate OC. (a) Open circles and solid squares indicate daily and monthly means, respectively. (b) Daily means for the samples in continental air masses. (c) Daily means for the samples in marine air masses.

[34] Temporal variations in the concentrations of the particulate OC in each air trajectory are shown in Figures 9b and 9c. Higher concentrations of the particulate OC were found in the continental air masses. Summer enhancement of the OC was clearly found in the air masses from oceanic regions, which suggests that the emission of the precursor gases from the oceanic biosphere and subsequent secondary productions would be important sources for the particulate OC in the marine air masses. In the continental outflows, on the other hand, a seasonal variation in the concentrations of the particulate OC was not pronounced. In the summer, the concentrations of the particulate substances in the continental air masses may decrease as well as the particulate EC, whereas the production of the particulate OC via secondary processes would be accelerated during the air mass transport. In addition, complex sources of particulate OC and its precursor gases, such as fossil fuel and biomass combustions, terrestrial vegetation, and the oceanic biosphere, would also make the seasonality of the particulate OC unclear.

[35] Temporal variation in the ratios of the OC/EC is plotted in Figure 10 and summarized in Table 1, showing that the OC was enriched in the aerosols compared to the EC in the summer. This is mainly caused by the intrusion of the marine air masses with enhanced particulate OC in the summer. Continental air masses also have high OC/EC ratios in the summer due to secondary productions.

Figure 10.

Temporal variation in the ratios of OC/EC. Open circles and solid squares indicate daily and monthly means, respectively.

3.3.2. Thermal Properties of the Particulate OC

[36] Although it is necessary for understanding the atmospheric roles and sources of particulate OC to identify the chemical compounds of the OC, there are many compounds in the OC fraction of the ambient aerosols. In past studies, some researchers have divided organics into several groups with similar chemical properties such as water solubility [e.g., Kleefeld et al., 2002] and/or polarity [e.g., Daisey et al., 1984] for the discussion of atmospheric behaviors and sources of the particulate OC. We have divided the organics on the basis of their volatility into two groups, i.e., the OC that is vaporized by heating at lower temperatures below 200°C (OCLT) and vaporized at higher temperatures between 200 and 340°C (OCHT), in order to discuss atmospheric behaviors of the particulate OC. It is expected that the OCHT mainly includes nonvolatile organic species, such as humic-like substances, whereas the OCLT mainly consists of semivolatile species, such as low-molecular-weight carboxylic acids. The volatility classification of aerosol species has been reported by several workers, showing that secondary aerosol production processes increase the concentrations of the species vaporized in the lower temperature ranges [e.g., O'Dowd et al., 2000].

[37] The concentrations of the particulate OCHT showed a similar seasonal variation to those of the particulate EC with higher values from the autumn to the spring. The relationships between the concentrations of the particulate EC and OCHT are plotted in Figure 11, showing good correlations. These results means that the carbonaceous compounds found in the OCHT fraction have source regions and transport mechanisms similar to those of the EC. Biomass smoke transport may introduce the particulate OCHT into the atmosphere together with the EC. For instance, past studies of macromolecular organics have suggested that biomass burning is a possible source of humic-like substances in the aerosols [e.g., Mukai and Ambe, 1986].

Figure 11.

Relationships between the concentrations of the particulate EC and OCHT for the daily means in the summer (open circles), in the winter (solid circles), and in the transitional season of spring and autumn (shaded circles). Dashed, solid, and dotted lines indicate linear regressions for the samples in the summer, in the winter, and in the transitional seasons, respectively.

[38] Some portions of the OCHT, however, are expected to be derived from secondary production processes. Figure 11 suggests that the relationships between the EC and OCHT have somewhat different trends among the four seasons. In the winter, the slope of the OCHT/EC was gentle and the correlation between both species was the highest. In the summer, on the other hand, relatively scattered plots were found in the relationship between both species, probably due to secondary production of organics in the OCHT fraction.

[39] Figure 12 shows the relationship between the concentrations of the particulate EC and OCLT. Although a statistical significant correlation can be found, the correlation was less than the OCHT case. Large portions of the carbonaceous compounds allocated into the OCLT fraction have different sources and/or production processes from the EC and OCHT. Secondary production processes may be important sources of the particulate OCLT. Indeed, as shown in Figure 12, the relationship between the OCLT and EC during the period from May to August shows that the OCLT was enriched in this season. Organic compounds derived from secondary productions tend to be semivolatile. For example, the low-molecular-weight dicarboxylic acids that have been detected in the ambient particulate OC [e.g., Kawamura and Usukura, 1993] are semivolatile compounds in equilibrium with the gaseous phase and mainly derived from photo-oxidation processes [Limbeck et al., 2005].

Figure 12.

Relationships between the concentrations of the particulate EC and OCLT for the daily means. Open and shaded circles indicate the samples from May to August and those from September to April, respectively. Solid line indicates linear regression for all samples, and dashed line indicates that for the sample from May to August.

4. Conclusions

[40] In order to understand the seasonal characteristics of the chemical and physical properties of the Asian outflows, we have conducted a year-round observation of atmospheric aerosols and their associated species from March 2001 to May 2002 on Rishiri Island over the northern area of the east Asian Pacific rim. The major findings of this study are summarized below.

[41] The Asian outflows were accelerated during the period from the autumn to the spring and brought particulate substances to the area. Alternatively, marine air masses covered the area and secondary productions became significant in the processes controlling the concentrations of the particulate substances in the summer. In the midwinter, the low-pressure systems east of Japan brought clean marine air masses from the high-latitudinal ocean region to the area, which resulted in low concentrations of the particulate substances. This type of atmospheric blocking in the midwinter has not been found in the observations from the southern area of the east Asian Pacific rim [e.g., Matsumoto et al., 1998; Kaneyasu and Takada, 2004].

[42] The particulate EC concentration showed higher values in the Asian outflows. The particulate EC is mainly transported with biomass smoke particles. The seasonal variability was also found in the continental air masses; relatively lower concentrations were frequently detected in the summer.

[43] The concentrations of the particulate OC showed an unclear seasonal trend. In the transitional seasons, Asian outflow brought high concentrations of the particulate OC to this area. In the summer, secondary production processes contributed to the enhancement of the secondary OC. The particulate OCHT may be mainly carried with the biomass smoke and showed a similar seasonal variation to the particulate EC. On the other hand, the particulate OCLT showed higher concentrations in the summer, which may be caused by secondary production processes and contribute to the secondary OC.

[44] The ratio of the EC to PM2.5 showed higher values during the winter, and decreased after the middle spring. During the period from the late spring to the summer, secondary species, such as OC and nss-SO42– from the photo-oxidation processes, increased in the fine aerosols, which can increase their single scattering albedo. Aerosols in the Asian outflows in this area are relatively “black” in winter, although the aerosol mass loading increases in spring.

Acknowledgments

[45] The authors express our gratitude to M. Sawaki, K. Ohta, Y. Miyata, M. Sasakawa, T. Suzuki, and H. Akimoto for their generous support during our observations. The authors also gratefully acknowledge the NOAA Air Resources Laboratory (ARL) for use of the HYSPLIT transport and dispersion model. This study was financially supported by the Core Research for Evolutional Science and Technology (CREST) of the Japan Science and Technology Corporation (JST).

Ancillary