To better understand the organic aerosol (OA) sources and formation processes in Northeast Asia, we studied atmospheric aerosol (total suspended particulate matter (TSP)) samples collected from Sapporo, northern Japan over 1 year period for the measurement of radiocarbon (14C) in aerosol total carbon (TC) and water-soluble organic carbon (WSOC). We also measured various organic tracers and carbonaceous components: organic carbon (OC), WSOC, and elemental carbon (EC). We found that the percent modern carbon (pMC) in TC and WSOC increased during spring/summer with maximum (85% and 117%, respectively) in May. The temporal variations of pMC in WSOC and TC showed a good agreement with changes in the contributions of biogenic OA tracers to OC and WSOC. We also found that emissions of pollen in spring and fungal spores from soil in summer/autumn as well as secondary OA formation from biogenic VOCs (BVOCs) in summer/autumn are responsible for the enhanced pMC. Although emissions from biomass burning are significant in winter, enhanced fossil fuel combustion lowers the pMC in WSOC and TC. Throughout the year, modern carbon is more enriched in WSOC fraction than in TC, indicating that WSOC is more associated with biological activity. This study warrants a need to reconcile the atmospheric models, considering seasonal differences in BVOC emissions, particularly in temperate regions, in estimating secondary OA budget and its climatic impacts.
 Carbonaceous aerosols impact the Earth's climate system directly by scattering and absorbing solar radiation and indirectly by acting as cloud condensation nuclei (CCN) [Novakov and Penner, 1993; Ramanathan et al., 2001]. They also cause adverse effects on human health [Nel, 2005]. In particular, water-soluble organic carbon (WSOC) alters the hygroscopic properties of aerosol particles and increases their CCN activity [Asa-Awuku et al., 2011]. Enhanced CCN number concentrations in the atmosphere should result in the increase of indirect radiative forcing by aerosol, causing a more reflective cloud [Twomey, 1977] and less precipitation [Albrecht, 1989]. High aerosol loadings are commonly observed in the Asian atmosphere due to enhanced industrial activities. In many regions of East Asia, aerosol optical depth (AOD) has been reported to be high (>0.3) with significant contribution from anthropogenic aerosols [Carmichael et al., 2009]. A substantial increase of longwave radiative forcing on the Earth surface due to the increased cloudiness and cloud liquid water by human activities has also been reported in Asian regions [Huang et al., 2006]. Asian aerosols and their precursors are further transported over the Pacific Ocean, causing significant changes in the atmospheric chemical compositions in the outflow regions of the Northern Hemisphere [Uno et al., 2011].
 Organic aerosols (OA) that represent a large fraction (20–90%) of the submicron aerosols [Kanakidou et al., 2005] are comprised of primary OA (POA) and secondary OA (SOA); the former is emitted directly from fossil fuel combustion, biomass burning, soil dust and biota whereas the latter is formed by photo-oxidation of gaseous precursors emitted from various sources including plants, biomass burning, and fossil fuel combustion [Robinson et al., 2007]. Global models predict that biomass burning emissions and SOA formation from biogenic volatile organic compounds (BVOCs) are the two major sources of OA [Kanakidou et al., 2005]. Regional studies suggest that anthropogenic sources could contribute 50% of the OA at northern midlatitudes and even more in densely populated areas of North America, Western Europe, and East Asia [de Gouw and Jimenez, 2009]. Aerosol measurements of radiocarbon (14C), a unique tracer to distinguish fossil and modern carbon, demonstrated that the percent modern carbon (pMC) in carbonaceous aerosols is large [Gustafsson et al., 2009; Schichtel et al., 2008; Szidat et al., 2004]. However, 14C data alone cannot discriminate the different sources of modern carbon: biomass burning, biological emissions, and SOA from BVOCs.
 Recently, a combined study of 14C with elemental and molecular tracers has become a prominent technique to identify and apportion the sources of carbonaceous aerosols [Gelencsér et al., 2007; Gilardoni et al., 2011; Yttri et al., 2011]. However, OA sources are still poorly understood due to a lack of year-round observations of pMC and/or limited studies on organic molecular tracers [Gilardoni et al., 2011; Yttri et al., 2011] that provide more specific source information [Fu et al., 2010; Simoneit, 2002]. For example, Gilardoni et al.  reported a negligible contribution (5–110 ng m−3) of primary biogenic organic carbon (OC) compared to biogenic secondary OC (up to 7500 ng m−3 in spring) in a European rural site. Yttri et al.  also reported a greater contribution of biogenic SOA to TC in PM1 and PM10 at the rural (68% and 56%, respectively) and urban (38% and 36%) sites in Norway. They stressed a dominance of modern OC regardless of season, location, and particle size. On the other hand, Fu et al.  reported high concentrations of sugars (6.8–1760 ng m−3) with the highest value in April, the spring bloom season at Jeju Island, which attributed for pollen (biogenic POA) emissions in spring. Further, biogenic POA and SOA largely contribute to the budget of WSOC [Sullivan et al., 2006] that may possibly be used as a proxy for SOA in fine particles, when the contribution of biomass burning is minor [Wonaschütz et al., 2011]. However, both direct [Kirillova et al., 2010; Weber et al., 2007] and indirect [Szidat et al., 2004] measurements of 14C in WSOC are very limited and there is no combined study of 14C in WSOC and SOA tracers, although about 70–90% of WSOC is identified as modern [Kirillova et al., 2010; Szidat et al., 2004; Weber et al., 2007].
 Here we report a combined study of 14C in WSOC fraction and total carbon (TC) together with various organic tracers in aerosol samples from Sapporo, northern Japan. TC, OC, WSOC, and elemental carbon (EC) were also measured. We discuss seasonal changes in the pMC of WSOC and TC in terms of the contribution of various biological sources to modern carbon in aerosols and the meteorology and thus growing season of vegetation. We also examine the variability of aerosol mass, carbon contents and pMC in TC and WSOC of fine particles collected from Sapporo (PM3.0, n = 3) and Xi'an, China (PM2.5, n = 1) to compare with the concurrent TSP samples.
2 Materials and Methods
2.1 Site Description
 Sapporo is located in the western part of the Hokkaido Island, northern Japan (43.07°N, 141.36°E), as shown in Figure 1. The population of the city was 1.9 million in 2010 (http://www.stat.go.jp/english/). The major industries in Sapporo include retail and tourism as well as manufacturing of foods, beverages, paper and pulp and machinery (http://en.wikipedia.org/wiki/Sapporo), from which the input of organics into the atmosphere is expected to be small. The humid continental climate prevails in Sapporo with a wide range of temperature between the summer and winter and the ground surfaces are covered with snow from late December to early April. The regional climate is strongly affected by the East Asian monsoon, which forms as a result of the thermal difference between the Asian continent and the Pacific Ocean, that controls the air mass source regions and compositions of natural and anthropogenic organic aerosols transported to the western North Pacific regions [Kaneyasu et al., 2000; Koike et al., 2006; Zhang et al., 2011], passing over Hokkaido. Thus, Sapporo, a network site of Aerosol Characterization Experiments in Asian Pacific region (ACE-Asia) [Huebert et al., 2003], is an ideal location for collecting air masses delivered from Siberia, China, and surrounding oceans [Aggarwal and Kawamura, 2008; Yamamoto et al., 2011].
2.2 Aerosol Sampling
 Sampling of total suspended particulate matter (TSP) (n = 21) was performed from 2 September 2009 to 5 October 2010 for approximately two consecutive weeks each using a pre-combusted quartz fiber filter (450°C, 4 h) and high-volume air sampler (~65 m3 h−1) on the rooftop (approximately 20 m above the ground) of the building of the Institute of Low Temperature Science, Hokkaido University, Sapporo. The university campus is surrounded mainly by residential area and the sampling point is about 500 m away from the main streets and the sampler was set up to avoid entraining the building emissions. PM3.0 samples were also collected in summer (n = 2) and autumn (n = 1) 2010 using high volume cascade impactor (Tisch Environmental, USA) at the same site. In addition, one TSP and one PM2.5 sample were collected in autumn (30 October to 3 November) 2009 from Xi'an, China (34.41°N, 109.04°E). Each sample was placed in a pre-combusted glass jar with a Teflon-lined screw cap to avoid potential contamination and stored in dark cold room at −20°C for 2 ~ 3 months prior to analysis.
 It should be noted that aerosols collected on quartz fiber filters might have positive (adsorption of gas phase organics onto the filter) and negative (volatilization of semi-volatile organics from the particles) sampling artifacts [Malm et al., 2011; Turpin et al., 2000]. In the present study, the evaporative loss of some semi-volatile organics from the aerosol particles should be more significant than the adsorbed organic vapors on quartz filter for longer time of sampling period and thus the reported concentrations may be underestimated. However, we consider that the evaporative loss of organics should be minimal because the concentration ratios of lower molecular weight (LMW) (≤4 rings) polycyclic aromatic hydrocarbons (PAHs) to higher MW (HMW) (≥5 rings) PAHs were found to be comparable for different seasons (0.41 ± 0.07, 0.63 ± 0.09, 0.56 ± 0.14, and 0.43 ± 0.07 in autumn, winter, spring, and summer, respectively), despite that ambient temperatures in Sapporo varied significantly (average temperature for each sample period ranged from −2.21 to 24.5°C) and that evaporative loss of LMW PAHs could be more likely than HMW PAHs [Coutant et al., 1988]. The higher LMW/HMW PAH ratios in winter may be caused by enhanced emission of LMW-PAHs due to increased consumption of kerosene (see section 3.2).
2.3 Aerosol Analyses
 The TSP mass in each filter was measured gravimetrically using analytical balance (Mettler-Toledo; AB204) by mass difference of the filters before and after sampling, which were conditioned in a desiccator for approximately 48 h.
2.3.1 Measurements of Carbonaceous Components
 Organic carbon (OC) and elemental carbon (EC) were measured using OC/EC analyzer (Sunset Laboratory Inc., USA) following Interagency Monitoring Protected Visual Environments (IMPROVE) thermal/optical evolution protocol and assuming carbonate carbon to be negligible. A filter disc (1.4 cm in diameter) was placed in a quartz sample boat inside the thermal desorption chamber and then stepwise heating was applied and the evolved CO2 was measured by a non-dispersive infrared (NDIR) detector system [Pavuluri et al., 2011]. The transmittance of laser light (660 nm) through the filter punch was used for setting up OC/EC split point and thereby OC correction. The analytical errors in duplicate analyses were within 0.7% for OC and 4.3% for EC. The sum of OC and EC was considered as TC.
 To determine water-soluble organic carbon (WSOC), an aliquot of filter disc (1.4 cm in diameter) was extracted with 15 ml organic free Milli Q water (18.3 MΩ) under ultrasonication for 20 min. The extracts were then passed through syringe filter (Millex-GV, 0.45 µm, Millipore) and then WSOC was measured using TOC analyzer (Shimadzu TOC-VCSH). The analytical error in duplicate analyses was within 9%.
2.3.2 Determination of Radiocarbon in TC and WSOC
 Radiocarbon (14C) in aerosol TC and WSOC fractions was determined by accelerator mass spectrometer (AMS) using one and three discs (1.6 cm in diameter) of filter sample, respectively. WSOC was extracted with organic free Milli Q water (18.3 MΩ) under ultrasonication for 20 min and filtrated with syringe filter (Millex-GV, 0.45 µm, Millipore). The extracts were adjusted to pH 8 ~ 9 with 0.05 M KOH and concentrated to dryness using a rotary evaporator under vacuum. WSOC was re-dissolved in about 1 mL methanol and transferred into glass vial. Before graphitization, WSOC (and filter disc for TC) were transferred into quartz tube (25 cm × 9 mm outer diameter) and dried under N2 blow and combusted to CO2 in the presence of CuO (850°C, 5 h) and Ag wires. CO2 produced from TC and WSOC was then converted to a graphite target using a microscale technique for 14C analysis [Tanaka et al., 2000; Uchida et al., 2004]. 14C contents were measured at AMS facility (NIES-TERRA) of the National Institute for Environmental Studies, and expressed as pMC [Stuvier and Polach, 1977] after δ13C correction. pMC values were calculated by normalizing to the standard material NIST SRM 4990c (HOx II) with a known pMC value (134.07). Our precision for 3–5 measurements of 1 mg of IAEA C6 standard is 0.3–0.5% [Tanaka et al., 2000]. Analytical errors in duplicate analysis of pMC in TC and WSOC in two sets of samples were 0.4–2% and 0.3–7%, respectively.
 It should be noted that pMC of atmospheric CO2 has increased, reaching a maximum (185% of pre-1950s level) in 1964 due to significant input of 14C from nuclear bomb test in the 1950s and early 1960s [Levin and Hesshaimer, 2000; Schichtel et al., 2008]. Since then, the pMC has steadily declined due to the banned nuclear test and exchange with the ocean and biosphere as well as the addition of fossil carbon, reaching a value of 107% in 2003 in midlatitudes of the Northern Hemisphere [Levin and Kromer, 2004]. Interestingly, wood burning of old trees (e.g., 127% for 50 years old) can increase pMC of atmospheric aerosols [Lemire et al., 2002]. Therefore, if biogenic and/or biomass burning emissions are only the source of aerosol carbon, the pMC of aerosol carbon can be greater than 100%.
2.3.3 Measurements of Organic Tracers
 Organic tracer compounds were extracted with dichloromethane/methanol (2:1; v/v) under ultrasonication for three times for 10 min each. The extracts were combined and then passed through quartz wool packed in a Pasteur pipette and then concentrated using a rotary evaporator under vacuum and dried under N2 blow down, followed by derivatization with 50 μL of N,O-bis-(trimethylsilyl)triflouroacetamide (BSTFA) with 1% trimethylsilyl chloride and 10 μL of pyridine at 70°C for 3 h. After the reaction, derivatives were diluted by addition of 140 μL of n-hexane containing 1.43 ng μL−1 of the internal standard (C13n-alkane) and then analyzed using a capillary gas chromatograph (Hewlett-Packard 6890) coupled to a mass spectrometer (Hewlett-Packard 5973) (GC/MS) [Simoneit et al., 2004; Fu et al., 2010]. The mass spectrometer was operated on the electron ionization (EI) mode at 70 eV and scanned in the m/z range 50–650. Data were processed with the Chemstation software. Individual compounds (TMS derivatives) were identified by comparison of mass spectra with those of authentic standards or literature data. Recoveries for the authentic standards and surrogates that were spiked onto pre-combusted quartz filters were better than 80% and the analytical errors in duplicate analyses were less than 10%.
2.4 Estimation of Ambient Secondary Organic Carbon
 Based on the measured organic tracer data set, the contributions of toluene-, isoprene-, α-pinene- and β-caryopyllene-derived secondary organic carbon (SOC) were estimated using a tracer-based method proposed by Kleindienst et al. . The SOC was calculated using the measured concentrations of tracer compounds in aerosol samples and the laboratory-derived tracer mass fraction (fSOC) factors of 0.0079 ± 0.0026 for toluene, 0.155 ± 0.039 for isoprene, 0.231 ± 0.111 for α-pinene and 0.023 ± 0.005 for β-caryophyllene [Kleindienst et al., 2007] by the following equation:
 where Σi[traceri] is the sum of the concentrations of the selected suite of tracers for a same precursor in µg m−3.
 It should be noted that Kleindienst et al.  derived mass fractions of secondary organic aerosol (SOA) tracers by using ketopinic acid as the surrogate for all the SOA tracers, whereas in this study we quantified SOA tracers using authentic standards together with some surrogates that were further quantified by GC/FID. 2,3-Dihydroxy-4-oxopentanoic and β-caryophyllic acids were used as toluene- and β-caryophyllene-SOA tracers, respectively, in this study, which are exactly the same as those in Kleindienst et al. . As isoprene-SOA tracers, six compounds including 2-methylglyceric acid, three C5-alkene triols (cis-2-methyl-1,3,4-trihydroxy-1-butene, 3-methyl-2,3,4-trihydroxy-1-butene and trans-2-methyl-1,3,4-trihydroxy-1-butene), and two diastereoisomeric 2-methyltetrols (2-methylthreitol and 2-methylerythritol) were used, whereas as α-pinene-SOA tracers four compounds such as 3-hydroxyglzutaric, pinonic, pinic, and 3-methyl-1,2,3-butanetricarboxylic acids were used in this study, although Kleindienst et al.  reported three and nine tracers for isoprene and α-pinene, respectively. However, we consider that the uncertainties due to the differences between the two quantification methods as well as the tracers used for SOA estimation should be within the uncertainties raised by the tracer-based approach itself (22–54%) due to the standard deviation of fSOC mentioned above.
2.5 Backward Air Mass Trajectories
 Ten-day backward air mass trajectories arriving in Sapporo at 500 m above the ground level were computed for every 48 h using HYSPLIT model [Draxler and Rolph, 2012] during each sample period. The three-dimensional plots of the trajectories are depicted in Figure 2. The air masses mainly originated from Siberia passing over Northeast Asia in autumn, winter, and spring, whereas in summer they mostly originated from the East China Sea and/or western North Pacific Ocean passing over coastal and/or Japanese Main Island. The air parcels mostly traveled at lower than 2000 m above the ground level. Hence, Sapporo aerosols should have been influenced by long-range transport of air masses that are enriched with marine and terrestrial biogenic emissions in summer and biogenic/anthropogenic emissions in rest of the year.
3 Results and Discussion
3.1 TSP and Carbonaceous Components and Their Temporal Variations
 Atmospheric loading of TSP in Sapporo ranged from 13.5 µg m−3 to 73.8 µg m−3 (average 30.0 ± 12.7 µg m−3), in which TC accounted for 7.60–23.7% (19.2 ± 4.03%). OC (range 1.62–8.33 µg m−3; average 3.77 ± 1.59 µg m−3) was found to be significantly higher than EC (0.57–2.84 µg m−3; 1.67 ± 0.63 µg m−3). WSOC ranged from 0.63–2.39 µg m−3 (12.3 ± 0.48 µg m−3) that accounted for 15.6–57.4% (35.0 ± 12.3%) of OC. TSP masses stayed low during winter with minimum (13.5 µg m−3) in February and then suddenly peaked (73.8 µg m−3) in early April (Figure 3a). They gradually decreased until mid summer and stayed low in rest of the year. In contrast, TC, OC, and WSOC showed a gradual decrease from mid autumn to late winter and then increased until early summer, except for few cases, whereas EC did not show any clear trend but is somewhat similar to that of TSP in winter and summer (Figure 3a).
 The higher TSP loading during early spring might be due to significant contribution of Asian dusts that are often transported to the western North Pacific by strong westerly winds [Huebert et al., 2003]. In fact, the average local winds suddenly rose from 3.2 to 4.4 m s−1 in March and to 4.6 m s−1 in April. A similar wind pattern is also likely at regional scale. The backward air mass trajectories arriving in Sapporo during early spring originate from Siberia and pass over Mongolia and Northeast China (Figure 2c). Further, the temporal variations of crustal elements, i.e., Ca, Al, and Fe (C. M. Pavuluri et al., Year-round observations of water-soluble ionic species and trace metals in Sapporo aerosols: Implication for significant contributions from terrestrial biological sources in Northeast Asia, submitted to Atmospheric Chemistry and Physics, 2013), are consistent with those of TSP from winter to mid summer, indicating a significant long-range transport of Asian dust from arid regions in Mongolia and China. In fact, massive dust storms occurred on 20 March 2010 in Mongolia and 31 March 2010 in Northeast China (http://earthobservatory.nasa.gov) and the dust plume covered over downwind regions including Hokkaido [Liu et al., 2011, see Figures 4 and 6 therein for MODIS AOD coverage and corresponding model output for 21 March 2010], which confirms that the contribution of Asian dust plume to Sapporo aerosols is significant in early spring. However, the inconsistent peaks of TC, OC, and WSOC with TSP (Figure 3a) suggest that the influence of polluted plume is insignificant on OC and WSOC in early spring, although we could not preclude such influence on EC. Hence, we presume that OC and TC are mainly influenced by biospheric sources; biomass burning, biological emissions and SOA formation from BVOCs on local to regional scales.
3.2 Temporal Variations of Percent Modern Carbon (pMC) in TC and WSOC and Organic Tracers
 We found a gradual decrease in pMC of both TC (pMCTC) and WSOC (pMCWSOC) from autumn to winter with minima in December (TC, 26.9%) or January (WSOC, 57.6%), followed by a progressive increase towards the late spring maximum (Figure 3b) with an average of 57.1 ± 13.5 and 81.8 ± 14.1, respectively (n = 21). The minimum pMCTC (Figure 3b) is consistent with the increased concentrations of EC (Figure 3a), hopanes and toluene-derived SOA (Figure 4a) as well as the increase in the contributions of hopanes and toluene-derived SOA (tracers of fossil fuel combustion) to OC (Figure 3c). In addition, pMCTC showed a negative correlation with the fractions of EC in TC (Figure 5a) and sum of hopanes and toluene-derived SOA in OC (Figure 5b). This pattern indicates that fossil fuel combustion on local to regional scale significantly contributes to the lower pMC in winter aerosols. This is consistent with the fact that consumption of kerosene in winter increases 20 times over that in summer in Japan [Moriwaki and Kanda, 2004]. However, the correlation coefficient (r2 = 0.27) for pMCTC and fraction of fossil fuel combustion-derived EC (ECff) in TC, which is estimated using median mass ratios of levoglucosan (Lev) to TC (15.0) and OC to TC ratio (0.78) in biomass burning aerosols (PM10) [Yttri et al., 2011], is smaller than that (r2 = 0.34) obtained for pMCTC and fraction of EC in TC (Figure 5a). The weaker correlation may be caused by large uncertainties of ECff estimation because the mass ratios of Lev/TC and OC/TC significantly depend on type of biomass, burning conditions and particle size [Yttri et al., 2011, and references therein].
 On the other hand, the temporal variations of pMCTC and pMCWSOC showed maxima in May (85% and 117%, respectively, see Figure 3b), being consistent with the increased concentrations of sums of lipid class compounds (plant wax n-alkanes (C23-C37) and fatty alcohols (C18-C32); Figure 4a) and biogenic organic tracers such as sucrose for pollens, mannitol for fungal spores, and isoprene-, α-pinene- and β-caryophyllene-derived SOA tracers (Figure 4b). The temporal variations of pMCTC and pMCWSOC are also consistent with the increase in the contributions of lipid biomarkers to OC (Figure 3c) and of biogenic organic and SOA tracers to WSOC (Figure 3d), respectively. Further, pMCWSOC showed a positive correlation with summed contributions of biogenic organic tracers and levoglucosan, a biomass-burning product [Simoneit, 2002] (Figure 5c). The temporal variations of WSOC and OC concentrations (Figure 3a) are similar to those of pMCTC and pMCWSOC (Figure 3b). These similarities suggest that the increased pMC in spring and summer are involved with the contributions of modern carbon from biomass burning, primary bioaerosol and/or SOA formation from BVOCs.
 Averaged molecular distributions of lipid biomarkers such as n-alkanes (C20-C38), fatty acids (C12-C32), and fatty alcohols (C18-C32) are characterized by an odd-, even- and even-carbon-numbered predominance with a maximum at C29, C16 and C26, and C26, respectively. Their carbon preference index (CPI) ranged from 1.97 to 8.11 (average 4.50) for n-alkanes, 3.06–8.92 (average 5.21) for fatty acids and 5.18–23.8 (average 10.6) for fatty alcohols. Such molecular distributions and CPI values suggest that they were most likely derived from terrestrial higher plant waxes [Fu et al., 2008, and references therein]. It is also noteworthy that the CPI index for n-alkanes is comparable to those reported in the aerosols from mountain area in Japan (>5) [Kawamura et al., 1995] and China (1.12–8.03, average 4.42) [Fu et al., 2008] and higher than those of urban aerosols from Chinese cities (1.16 ± 0.12) [Wang et al., 2006] and Tokyo (1.1–2.8, average 1.5) [Kawamura et al., 1995] as well as remote marine aerosols from the tropical North Pacific (1.6–3.2) [Gagosian et al., 1982], but similar to those from the western North Pacific (1.8–14.6, average 4.5) [Kawamura et al., 2003], where the Asian outflow is significant. These results and comparisons confirm that the organics were mainly derived from terrestrial biological sources rather than the emissions from marine biota. However, we do not preclude a minor contribution of marine sources at least in summer, when the air masses were mostly originated from oceanic region (Figure 2d).
 In Hokkaido, the growing season starts in May and extends into October when daily average ambient temperatures are ≥10°C [Toma et al., 2011]. The average temperature in Sapporo was ≤10°C from 11 November 2009 to 1 May 2010; hence local BVOC emission is insignificant until the end of April. In fact, atmospheric CO2 flux measurements (from forest ecosystem to air) in November to May in Sapporo showed a positive value (approximately 0.1 mg CO2 m−2 s−1) [Miyazaki et al., 2012], indicating an overall release of CO2 and negligible photosynthesis by vegetation. However, pMCTC and pMCWSOC started to increase from mid and late winter toward spring, respectively (Figure 3b). This earlier increase in pMC suggests that modern carbon in winter is not derived from local vegetation, but is involved with the long-range atmospheric transport from Eurasia (Figure 2). However, local contribution of biomass burning to OC and WSOC cannot be excluded in early winter because wood burning is often utilized in Sapporo for domestic heating.
 Contributions of levoglucosan to WSOC increased from autumn to winter with a maximum in December. However, the combined contributions of levoglucosan and biogenic SOA tracers to WSOC are at most 5% in winter (Figure 3d). This result suggests that major portion of water-soluble organic compounds in the aerosols is not detectable by the GC/MS technique used. They may include water-soluble amino acids/proteins that originate from bacteria and other bioaerosols as well as polymers derived from α-dicarbonyls such as glyoxal of biogenic origin [Yu et al., 2011]. Low molecular weight dicarboxylic acids should also be abundant [Kawamura and Sakaguchi, 1999], contributing significantly to WSOC.
3.3 Implication of Seasonal Changes in pMC for Aerosol Sources in Northeast Asia
 During May when pMC maximizes (Figure 3b), we detected an increased level of sucrose (Figure 4b), whose contribution to WSOC is highest in early May (Figure 3d). Sucrose can be derived from pollen emitted from the local vegetation near the sampling site [Fu et al., 2012]. We also found simultaneous increase in the contributions of α-pinene- and β-caryophyllene-derived carbon to WSOC (Figure 3d). In fact, the air masses during 17–31 May (QFF3544) originate from the southern tip of Siberia/Russian Far East spending most of the time over northern Hokkaido (Figure 2c), where thick forest vegetation emits BVOCs. These results indicate that pollen emissions as well as oxidation of mono- and sesqui-terpenes are the main cause for the increased pMCWSOC during 17–31 May (Figure 3b). Higher pMC has been reported in coarse particles collected in Tokyo during spring [Shibata et al., 2004], although pollens or pollen tracers were not determined. In April to May, pollens are abundantly emitted in Sapporo and surrounding areas, suggesting that the peak of pMC is mainly caused by local and regional vegetation. Because sucrose and biogenic SOA contribute less than 9% of WSOC in spring, there are other classes of water-soluble organics including small dicarboxylic acids, which are produced by photochemical oxidation of various organic precursors [Kawamura and Sakaguchi, 1999]. The other classes of water-soluble organics should also contribute to maximize the pMC of aerosols. In contrast, the mass fraction of levoglucosan in WSOC stays low during spring (Figure 3d).
 In summer, levels of pMC in WSOC are as high as 96.8% (Figure 3b) being consistent with enhanced contributions of α-pinene- and β-caryophyllene-SOA tracers to WSOC in summer (Figure 3d). Interestingly, β-caryophyllene SOA is a dominant contributor to WSOC in June to July followed by α-pinene SOA whereas isoprene SOA dominates in late August to early September (Figure 3d) when the ambient temperature is highest (24.5 ± 1.1°C and 23.0 ± 2.3°C, respectively). It is important to note that isoprene emission is highly associated with ambient temperature because the isoprene provides thermal protection for leaves to lower the leaf surface temperature when ambient temperature significantly increases [Sharkey and Singsaas, 1995]. These results demonstrate that, although the sources of WSOC are variable, emissions of BVOCs followed by photo-oxidation largely control the pMCWSOC in Northeast Asia. Similarly, higher pMCTC (ca. 60%) were obtained in summer (Figure 3b), although the values are lower than pMCWSOC by 20%. Interestingly, the mass fractions of water-insoluble organic compound classes (plant wax n-alkanes, fatty acids, and fatty alcohols) in OC are rather small in summer to autumn (Figure 3c). At present, we cannot specify the cause of higher pMCTC in August at the level of compound class. However, we suspect that increased emissions of BVOCs and bioaerosols should be responsible during the growing season [Guenther, 1997]. Again, biomass burning is not a major source of enhanced pMC in summer because the mass fraction of levoglucosan in WSOC is very small (Figure 3d), although its fraction gradually increased toward autumn.
 In autumn, pMC levels of both TC and WSOC stayed still relatively high until the end of the growing season (early November) in Hokkaido and then gradually decreased toward early winter (Figure 3b). Surprisingly, the averaged pMC in autumn is comparable to that in summer (Table 1). BVOC emissions from local vegetation in Sapporo should significantly decline in October when the defoliation starts. Ambient temperatures in Sapporo decrease to <10°C in late autumn and thus local BVOC emissions decrease [Fuentes et al., 1998] and the production of SOA become insignificant after November on local scale. Hence, the relatively high levels of pMC in autumn cannot be explained by the local emission of BVOCs. Instead, a long-range atmospheric transport of air masses from Eurasia is likely responsible to the high pMC of aerosols. The air masses may contain abundant BVOCs emitted from terrestrial higher plants and biomass burning products emitted from forest fires in Siberia (Figure 2a). The leaf area index reaches a maximum by May in boreal aspen forest [Fuentes et al., 1998], June in Siberia [Kobayashi et al., 2010], and July in northern Japan [Wang et al., 2004] and stays high for the following several months. Hence, significant emissions of BVOCs from boreal evergreen conifer may be important source of high pMC. In fact, α-pinene SOA significantly contributes to WSOC until mid November (Figure 3d). Increased contribution of mannitol to WSOC in autumn (Figure 3d) also demonstrates that fungal activities emit modern carbon as spores from the fallen leaves to the air during defoliation season.
Table 1. Percent Modern Carbon (pMC) in Total Carbon (TC) and Water-Soluble Organic Carbon (WSOC) of Atmospheric Aerosols from Sapporo, Northern Japan During September 2009 and October 2010a
Component (Aerosol Size)
Autumn, September–November; Winter, December–February; spring, March–May; summer, June–August; Annual: September 2009 to August 2010.
55.2 ± 9.0
36.9 ± 9.4
68.7 ± 14.0
60.8 ± 6.2
56.6 ± 14.1
55.8 ± 2.4
81.7 ± 10.2
59.8 ± 1.9
88.5 ± 17.5
88.3 ± 5.4
81.5 ± 14.8
85.2 ± 0.9
 Interestingly, the ratio of pMCWSOC to pMCTC is higher (1.6) in winter than in spring (1.3) and other seasons (1.5). This result may indicate that biomass burning produces WSOC [Sullivan et al., 2006], causing an increase in pMCWSOC/pMCTC ratio in winter. This explanation is also consistent with the fact that combustion products of fossil fuels are less water-soluble due to the abundant presence of hydrocarbon-like materials [Miyazaki et al., 2006]. It is important to note that modern carbon accounts for more than a quarter of TC and half of WSOC even in winter (Table 1). This finding suggests that, in addition to fossil fuel combustion, the winter aerosols are significantly affected by biogenic POA and SOA. In fact, average contributions (%) of biomarkers (i.e., fatty acids and alcohols) to OC in winter (1.21 ± 0.33 and 0.30 ± 0.06) are comparable to those in spring (0.99 ± 0.28 and 0.48 ± 0.22). The concentrations of α-pinene-derived compounds (Figure 4b) and their contributions to WSOC (Figure 3d) are significant even in winter although the contributions are approximately five times lower than those in summer/autumn (Figures 4d). The air masses that arrived in Sapporo during winter mostly originating from Siberia and passing over Northeast Asia (Figure 2b) should have been enriched with biogenic POA and SOA.
 Long-range atmospheric transport of terrestrial biomarkers from Siberia to northern Japan by the Asian winter monsoon has been demonstrated by stable carbon and hydrogen isotopic compositions of fatty acids in fresh snow in Sapporo [Yamamoto et al., 2011]. Meanwhile, emission of monoterpenes depends on temperature, light intensity, and leaf oil production with a clear seasonal maximum in spring or summer. However, their emissions are still significant in Japan during winter, although the emission rate is declined by a factor of 4–10 compared to that in spring [Matsunaga et al., 2011] or summer [Yatagai et al., 1995]. These previous studies further support our findings that the contributions of biogenic POA and SOA are also important in winter.
3.4 Comparison of pMC in TC and WSOC between Fine Particles and Concurrent TSP
 The mass fractions of fine particles (three PM3.0 samples from Sapporo and one PM2.5 sample from Xi'an, China) in concurrent TSP were found to be 40–64% (Table 2). Contributions of WSOC, OC, EC, and TC in the fine particles to those in TSP were accounted for 54–100%, 47–62%, 49–82%, and 49–67%, respectively (Table 2). These results suggest that most of OA, especially WSOC exist in fine particles. Further, we found that pMC of both TC and WSOC in fine particles is in general comparable to those of TSP samples (Table 2). However, the PM3.0 samples that were simultaneously collected in 17 July to 2 August and 21 September to 5 October in Sapporo showed that pMCTC and pMCWSOC in PM3.0 are lower than those of TSP. These results may suggest that modern carbon is more enriched in coarse particles that are probably derived from bioaerosols such as pollens and spores. In contrast, Xi'an samples did not show a significant difference in pMC between PM2.5 and TSP possibly due to less important contributions of pollens and spores that are mostly present in large particles.
Table 2. Concentrations of TSP and PM3.0 (PM2.5), Carbonaceous Components of TSP and PM3.0 (PM2.5) and pMC in TC and WSOC as Well as Their Ratios in Sapporo, Japan and Xi'an, Chinaa
 Throughout the campaign in Sapporo, we found that pMCWSOC values are always higher than pMCTC by 2.1–40.4% (average 24.6 ± 8.6%). Similar results (27.4 ± 3.5%) were obtained for fine particles from Sapporo (PM3.0) and Xi'an (PM2.5 and TSP) (Table 2). Systematically higher pMC in WSOC than in TC indicates that WSOC fraction is more enriched with modern carbon. It is consistent with the knowledge, that is, BVOC-derived SOA and biomass burning products are highly water-soluble [Sullivan et al., 2006; Wonaschütz et al., 2011]. Hence, WSOC derived from these sources is enriched with modern carbon.
 Finally, it is of interest to note that the pMC values of Xi'an aerosols (PM2.5 and TSP) (Table 2) are significantly lower than those of Sapporo aerosols (Table 1). This result is reasonable because Xi'an is an industrial city with heavy air pollution. However, pMCWSOC in both PM2.5 and TSP from Xi'an are approximately 70% (Table 2), suggesting that a significant contribution of SOA and/or biomass burning products to OA is still important even in the industrial city of China.
 Based on the seasonal distributions of organic tracers, we discovered a changing contribution of different biogenic sources to modern carbon in aerosols following the meteorology and thus changes in growing season of vegetations. The temporal variations of modern and fossil carbon in aerosols have been characterized by enhanced primary emissions of pollens in spring, increased SOA production from α-pinene in summer, enhanced SOA production from isoprene and primary emissions of fungal spores in autumn, and enhanced fossil fuel combustion and biomass burning products in winter. The radiocarbon measurement of WSOC fraction combined with organic tracer analyses also demonstrated that water-soluble OA, which can act as CCN and thus affect the Earth climate, is more enriched with modern carbon that is derived from photo-oxidation of BVOCs and biomass burning than water-insoluble OA. Based on modern carbon and organic tracer analyses, this study supports previous understanding, that is, biological and photochemical activities largely control the organic aerosol composition and water-solubility and thus influence the hygroscopic properties of atmospheric particles in Northeast Asia. The present study further requires taking the seasonal BVOC emission and subsequent OA formation into the consideration of modeling SOA budget and its impacts on climate system on regional to global scales.
 This study was in part supported by the Environment Research and Technology Development Fund (B-0903) of the Ministry of the Environment, Japan. We thank G. Wang for the help of collecting TSP and PM2.5 samples from Xi'an, China. Phil Meyers is acknowledged for the English correction for the early version of the manuscript.