Aerosol Chemical Profile of Near-Source Biomass Burning Smoke in Sonla, Vietnam during 7-SEAS Campaigns in 2012 and 2013

This study aimed to investigate aerosol chemical characteristics and to obtain the chemical profile of near-source biomass burning (BB) aerosols at a site (675 m a.s.l.) in Sonla, Northern Vietnam. Particulate matter (PM) with an aerodynamic diameter less than or equal to 2.5 μm (PM2.5) was collected over a 24 h sampling period as part of the Seven South East Asian Studies (7-SEAS) campaign. The studies were conducted when BB was highly active — that is, in the spring of 2012 and 2013. The collected particles were analyzed for carbonaceous fractions and water-soluble components, in addition to the mass concentration. Data obtained were further analyzed to determine the stable species profile by classifying the 5-day air-mass backward trajectories. The average PM2.5 mass concentrations were 51 ± 19 μg m and 57 ± 27 μg m in 2012 and 2013, respectively. Carbonaceous contents dominated BB aerosol, with 59% ± 9% and 58% ± 9% in organic carbon (OC) and 9% ± 3% and 10% ± 3% in elemental carbon (EC) of PM2.5 in 2012 and 2013, respectively. Of the 8 carbonaceous fractions analyzed thermo-optically for PM2.5, OC3 (evolution temperature at 280°C–480°C) was most abundant in OC fractions, and EC1-OP (elemental carbon evolved at 580°C minus the pyrolized OC fractions) was predominant in EC fractions in most occasions. Among the measured water-soluble inorganic ions, NH4 and SO4 widely varied, indicating the influence of different trajectory origins. This finding was confirmed by trajectory classification of aerosol data. The trajectories were also distinguished with respect to char-EC to soot-EC ratio, and water-soluble OC. These characteristics were highest in the trajectory from the BB source area.


INTRODUCTION
Biomass burning (BB) has been established as a significant source of particulate matter (PM) and trace gases in the atmosphere (Andreae and Merlet, 2001;Koe et al., 2001;Streets et al., 2003;Crounse et al., 2009;Chuang et al., 2013a, b;Chuang et al., 2014).Emissions from BB significantly influence both health (IPCC, 2007;Naeher et al., 2007) and global climate (Arola et al., 2007;Pratt et al., 2010).BB generally includes burning of forests, grassland, agricultural waste, as well as open fires and wood combustions (Hays et al., 2002;Gao et al., 2003;Vanderzalm et al., 2003;Fabian et al., 2005;Iinuma et al., 2007;Puxbaum et al., 2007;Mead et al., 2008;Pio et al., 2008;Schmidl et al., 2008a;Engling et al., 2009).Alonso-Blanco et al. (2014) demonstrates that wildfires affect not only the number of emitted particles and the size distribution, causing a clear increase in the number of aerosols in the atmosphere, but they are also responsible for altering the local radiative balance.Chemical profiling of aerosols from BB activities is essential to distinguish various sources of BB aerosols.Resolved chemical species profile of aerosols from BB emissions can be used to evaluate the effects of BB on health and climate.
Several studies have been conducted to characterize BB aerosols (Liu et al., 2000;Yamasoe et al., 2000;Graham et al., 2002;Gao et al., 2003;Salam et al., 2003;Wangkiat et al., 2003;Cao et al., 2005;See et al., 2006;Iinuma et al., 2007;Robinson et al., 2009;Lee et al., 2011;Chuang et al., 2013a, b;Mkoma et al., 2013;Tsai et al., 2013;Chuang et al., 2014).Intensive BB frequently occurs every spring in the Indochina Peninsula, including Myanmar, Thailand, Laos, Cambodia, and Vietnam (Lee et al., 2011;Chuang et al., 2013b;Chang et al., 2015;Wang et al., 2015;Pani et al., 2016a, b;Tsay et al., 2016).Major components of BBinfluenced PM 2.5 (PM with an aerodynamic diameter less than or equal to 2.5 µm) include carbonaceous contents [both organic carbon (OC) and elemental carbon (EC)] (Cao et al., 2005;Lee et al., 2011;Chuang et al., 2013a, b;Mkoma et al., 2013;Chuang et al., 2014).These studies distinguished the emission sources by estimating the fractions of carbonaceous contents.The carbonaceous contents were resolved into eight different fractions depending on the evolution temperature and time interval for each carbon fraction (Chow et al., 1993).The distribution patterns and gradients of such carbon fractions evolving at different temperatures were used to identify the emission sources (Chow et al., 2004;Kim et al., 2004;Cao et al., 2006).Among the five OC fractions, OC1 and OC3 are dominant fractions of OC in PM 2.5 from vegetative burning (Chow et al., 2004).OC1 and OC2 are also considered major carbon fractions, followed by OC3 and pyrolized OC (OP) in the aerosol collected during BB activity (Cao et al., 2005).By contrast, OC2 is the most abundant carbon fraction in motor vehicle exhausts and coal combustion, as reported by Chow et al. (2004) and Cao et al. (2005), respectively.Chuang et al. (2014) have recently differentiated the carbon fractions between the BB and non-biomass-burning (NBB) periods and reported that OC3 and EC1-OP are enriched in OC and EC for BB aerosol, whereas OC2 and EC2 occur more than other fractions in OC and EC for NBB aerosol.
Except for carbonaceous fractions, the ratios of OC to EC indicate various sources of emissions such as gasoline vehicles, diesel vehicles, coal combustion, and BB (Cachier et al., 1989;Watson et al., 2001;Chow et al., 2004;Liu et al., 2006;Chuang et al., 2013b;Chuang et al., 2014).Furthermore, Han et al. (2007) have suggested that the char-EC/soot-EC ratio (where char-EC is EC1-OP and soot-EC is the sum of EC2 and EC3) can more effectively identify the sources of carbonaceous aerosols, compared with the previously used OC-to-EC ratio because the latter can be affected by the formation of secondary organic aerosol (Han et al., 2009).The char-EC is formed at relatively low combustion temperatures such as that during BB activity, whereas soot-EC is formed at high-combustion temperatures, similar to those of coal combustion or from vehicle exhausts (Zhu et al., 2010).Char-EC/soot-EC ratios can be as high as 20 for aerosols from BB activity and less than 1 for those coming from vehicle exhausts (Chow et al., 2004).Cao et al. (2005) found char-EC/soot-EC ratios at 11.6 for BB in Xi'an City, China.However, the char-EC/soot-EC ratio can be as high as 31 for sagebrush burning, as determined in a laboratory experiment (Chen et al., 2007).In contrast to the chamber study, Chuang et al. (2014) reported char-EC/soot-EC ratios to be 3.9 ± 3.5 and occasionally exceeding 10 in the BB dominated air mass trajectory at a site in Mt.Lulin, Taiwan.
Meanwhile, the levels of water-soluble inorganic ions (WSIs), such as potassium ion (K + ), ammonia ion (NH 4 + ), sulfate ion (SO 4 2- ) and nitrate ion (NO 3 -), exhibit a signature pattern for different sources.Major WSIs related to BB aerosols include NH 4 + , K + , NO 3 -, and SO 4 2- (Ryu et al., 2007;Lee et al., 2011;Chuang et al., 2013b).Chuang et al. (2013b) reported abundant cations and anions such as NH 4 + , K + , NO 3 -, and SO 4 2-in PM 2.5 , constituting 2.98%, 2.55%, 1.24%, and 9.96% of the PM 2.5 mass, respectively.However Lee et al. (2011) have reported higher NH 4 + , K + , NO 3 -, and SO 4 2-mass fractions, i.e., 6.9%, 1.6%, 3.6%, and 18.3% of PM 2.5 , respectively, during a BB-influenced period at a site in Mt.Lulin, Taiwan.The abundance of NH 4 + , NO 3 -, and SO 4 2-has been attributed to fossil fuel combustion and anthropogenic activities from other areas by long-range transport (Lee et al., 2011).Ryu et al. (2007) have investigated the chemical properties of BB aerosol in an agricultural waste burning at a rural site in Korea and observed significant NH 4 + , K + , NO 3 -, and SO 4 2-mass fractions corresponding to 6.81%, 2.44%, 9.63%, and 15.90% of PM 2.5 , respectively.Species such as K + is well-established as a BB tracer, and the presence of K + in aerosols from urban areas have been linked to BB (Hai and Oanh, 2013;Hang and Oanh, 2014).Similarly, NH 4 + and SO 4 2-are considered tracers for anthropogenic sources.However, significant presence of such tracers has been observed in aerosols from areas with dominant BB activities.The presence of these aerosols has been attributed to the influence of air masses flow from urban areas (Ryu et al., 2007;Lee et al., 2011).Therefore, trajectory differentiated analysis of the aerosol chemical profile is required to identify the sources.
The abovementioned studies focused on the chemical characterization and identification of tracers of BB aerosols.However, studies on the chemical characteristics of aerosols in Vietnam are rarely reported.In addition, the aerosol source profile from BB activities in Indochina Peninsula is poorly understood.Aerosol source profile is essential in identifying the source influence in the process of source apportionment.This study is aimed at finding signature profiles from the near-source BB aerosol, with reference to the composition of carbonaceous fractions and WSIs in PM 2.5 collected in Sonla, Northern Vietnam in 2012 and 2013.This study demonstrates the influences of different air mass trajectories on the variability of species composition of aerosol.

Sampling Site and Sampling Collection
The sampling site (21°19'55"N, 103°54'18"E, 675 m a.s.l) is located at the Sonla Atmospheric Observatory Station in Sonla, Northern Vietnam (Fig. 1).The Sonla site is approximately 302 km northwest of the capital Hanoi, 140 km north to the border of China, and 110 km east of the border of Lao.The province covers approximately 14,174 km 2 , with a population of about 1,134,000 inhabitants as of 2012 (Vietnam Demographics Profile, 2010 available at http://www.gso.gov.vn/) with a density of about 80 persons per km 2 .The site is located on a high hill in Sonla City and surrounded by grasses and trees.
PM 2.5 was collected using collocated R&P ChemComb Model 3500 Speciation Sampling Cartridges (Thermo Fisher Scientific Co., Inc., Waltham MA, USA) supported by tripod stands, as described by Chuang et al. (2013b).Daily sampling was conducted continuously for 28 days in 2012 and 44 days in 2013.Collection of aerosol mass, WSIs, and carbonaceous contents, as well as the reduction of artifacts was performed following the procedure described by Chuang et al. (2013b).Each sample was collected for 24 h (starting at 8 am local time), and the collected filters were stored in the refrigerator at 4°C and then shipped back to Taiwan for chemical analysis.In addition, 5-day air-mass backward trajectories from the HYSPLIT website (http://ready.arl.noaa.gov/HYSPLIT.php;Draxler and Rolph, 2013)

Chemical Analysis
Teflon filters were conditioned for at least 72 h and then weighed using Mettler MX5 Microbalance (Mettler Toledo Co., Inc., Greifensee, Switzerland) with a sensitivity of ± 1 µg in a stable environment with a temperature of 22°C ± 1°C and a relative humidity of 30%-35%.The filters were used for the analysis of WSIs after weighing.Each Teflon filter was moistened with 200 µL of ethanol and ultrasonically extracted for 90 min in 10 mL of de-ionized distilled water (resistivity > 18 MΩ-cm).The de-ionized distilled water was subsequently passed through a Teflon syringe filter with a pore size of 0.45 µm (Millipore Corp., Billerica, MA, USA).The filtrate was then injected into an ion chromatograph (DX-120 for cations and DX-1000 for anions) to measure WSIs (including Na + , NH 4 + , K + , Mg 2+ , Ca 2+ , Cl -, NO 3 -, and SO 4 2-).Meanwhile, the quartz fiber filters were analyzed for OC and EC fractions.The analysis was conducted using the DRI Model 2001A OC/EC Carbon Analyzer (Atmoslytic Inc., Calabasas, CA, USA), which employs a thermal optical reflectance correction scheme for pyrolized OC (OP) by following the United States Interagency Monitoring of Protected Visual Environments (IMPROVE) protocol for monitoring OC and EC fractions (Chow et al., 1993;Chow et al., 2004;Watson et al., 2005).The IMPROVE method separates carbons in collected PM into 8 fractions (OC1, OC2, OC3, OC4, OP, EC1, EC2, and EC3) based on evolution temperatures and time intervals.The resolved EC fractions were further classified into char-EC (EC1-OP) and soot-EC (EC2 + EC3) (Han et al., 2007;Chuang et al., 2013b) for source identification.Water-soluble organic carbon (WSOC) was measured using the Aurora 1030W Analytical TOC analyzer (Global Spec.Inc., East Greenbush, NY, USA), following the procedures described by Yang et al. (2003).

Distribution of PM 2.5
During sampling, the daily mass concentrations of PM 2.5  2) and 76 ± 24 µg m -3 in 2013 (Table 3), respectively.The highest level of PM 2.5 in the BBIC trajectory indicates the greatest effect of BB activities on regional aerosol in the springtime.Meanwhile, the reduction in BBIC standard deviations for both years, in contrast to that for the entire sampling period (10 vs. 19 µg m -3 in 2012; 24 vs. 27 µg m -3 in 2013), indicates the potential of trajectory grouping in reducing data fluctuation.
Temporal variations in the PM 2.5 aerosol bulk contents of WSIs and carbonaceous fractions during the study periods of 2012 and 2013 are also depicted in Figs.3(a) and 3(b), respectively.Among the bulk contents, OC fractions dominate the distribution, indicating the characteristics of BB activity on the PM 2.5 chemical contents.EC fractions are much less than the OC fractions in the PM 2.5 , providing other distinction of carbonaceous contents for near-source BB aerosol.Likewise, bulk anions are comparatively more than the bulk cations in the PM 2.5 at the Sonla site.

Carbonaceous Contents
As shown in Table 1, OC comprises the dominant bulk content in the PM 2.5 mass with the average percentages of the resolved PM 2.5 at 59.0% ± 9.1% in 2012 and 58.1% ± 9.5% in 2013, respectively.In the present study, EC has been calculated to be 9.3% ± 2.5% and 10.0% ± 2.7% of the resolved PM 2.5 in 2012 and 2013, respectively.Among the 8 carbonaceous fractions analyzed for PM 2.5 , the OC3 > OC4 ≒ OC2 > OC1 ≒ OP (pyrolized OC) order in OC fractions and the EC1-OP > EC2 > EC3 order in EC fractions were observed in most occasions; the same trends were observed in both 2012 and 2013 [Table 1 and Figs. 4(a) and 4(b)].The least amount of OC1 in the OCs may be attributed to rapid evaporation of this fraction in BB smoke after emission into the atmosphere (Chuang et al., 2013b).Further classification of aerosol carbonaceous fractions into BBIC, BBSC, and BBSS [Figs. 5(a) and 5(b)] trajectory types confirms that OC3 and EC1-OP are the most abundant fractions of OC and EC contents in BBIC (trajectories originated from BB source area) for the three trajectory types in both years.
The PM 2.5 EC averages were found to be 3.5 ± 1.8 and 3.3 ± 1.7 µg m -3 in 2012 and 2013, respectively.These values are comparable to the EC value (3.3 ± 0.9 µg m -3 ) measured in Chiangmai, Thailand (Chuang et al., 2013b).The average EC concentration is higher than those reported by others, i.e., in the range of 0.4-0.7 µg m -3 at the Dongsha Island site (Chuang et al., 2013a); 1 ± 0.33 µg m -3 by Mkoma et al. (2013) in Tanzania; and 0.4-1.0 µg m -3 (five-year average) at the Mt.Lulin site (Chuang et al., 2014).Trajectoryclassified observations indicated a higher EC concentration in the BBIC trajectory in both 2012 and 2013 (Tables 2 and  3).Such a high concentration of EC indicates the warming potential of BB aerosol.
In the literature, the ratios of OC to EC have been used to provide information on the emission sources (Chow et al., 2004;Dan et al., 2004;Cao et al., 2005;Han et al., 2007Han et al., , 2009)).Cao et al. (2005) reported OC-to-EC ratio of 5.1 in the winter, which was attributed to coal combustion for residential heating and BB.By contrast, Cao et al. (2006) reported low OC-to-EC ratio of 1, which was attributed to vehicle and traffic-related sources.Moreover, the OC-to-EC ratios were 4.75 in BB aerosol in Phimai, Thailand (Li et al., 2013) and 5.7 in Chiangmai, Thailand (Chuang et al., 2013b).The averages of OC-to-EC ratio for PM 2.5 are 6.8 ± 1.8 and 6.1 ± 1.5 in 2012 and 2013, respectively (Table 1).The wellabove OC-to-EC ratios confirmed the BB contributions to the aerosol collected in the present study.However, the trajectory differentiated values for OC-to-EC ratios for the BBSC, BBIC and BBSS trajectories were 7.6 ± 2.4, 6.3 ± 1.0, and 6.3 ± 1.2 in 2012, respectively (Table 2).In 2013, the OC-to-EC ratios for the BBSC, BBIC, and BBSS trajectories were 5.8   11.2 Note: the value listed under "% resolved" is obtained from a fraction of the resolved PM 2.5 (sum of OC, EC, and WSIs).± 1.5, 6.0 ± 1.1, and 6.7 ± 2.0, respectively (Table 3).The trajectory differentiated values for OC-to-EC ratios in both years are comparable to each other, but without distinction.
Using the ratio of char-EC (EC1-OP) to soot-EC (the sum of EC2 and EC3) in PM 2.5 is demonstrated to effectively distinguish the carbonaceous influence from BB (Han et al., 2007(Han et al., , 2009)).The difference in combustion temperaturethat is, char-EC with low combustion temperature such as BB activities and soot-EC with higher combustion temperature such as coal combustion and internal engine combustionprovides the rationale for the use of char-EC to soot-EC ratio in source apportionment of carbonaceous contents in aerosol (Zhu et al., 2010).The char-EC to soot-EC ratio has been suggested to be as high as 20 for BB and less than 2.0 for coal combustion and vehicle exhausts (Chow et al., 2004;Chen et al., 2007).Similarly, Chuang et al. (2013b) found that the char-EC to soot-EC ratios from BB sources were 9.4.A lower char-EC to soot-EC ratio within the 0.6-1.4range in the pristine sea and anthropogenic air masses has been reported in the study by Chuang et al. (2013a).In the present study, the averages of char-EC to soot-EC ratios were calculated to be 22.4 ± 15.0 and 9.6 ± 10.1 for 2012 and2013, respectively (Table. 1).These two values clearly indicated that the collected aerosols are influenced by BB activities for both years.The higher values for 2012 may be attributed to lower concentration of EC2 fraction (approximately 2 to 3 times less than that of 2013).Higher EC2 is generally observed in aerosols from diesel vehicle exhausts (Cao et al., 2006) and lower EC2 from BB (Cao et al., 2005).Similarly, lower EC2 fractions were also reported in a chamber-based study where wildland fuels were burnt using laboratory combustion facility (Chen et al., 2007).Therefore, the contribution of BB activity to aerosol in Sonla was relatively stronger than that of vehicle exhausts in 2012 than 2013.Further trajectory analysis of the data indicated a clear distinction among three different trajectory types in terms of the char-EC to soot-EC ratio.The char-EC to soot-EC ratios for the BBSC, BBIC, and BBSS trajectories were 17.6 ± 8.5, 35.5 ± 10.4, and 12.0 ± 14.6 in 2012, respectively (Table 2) and 4.3 ± 2, 14.9 ± 12, and 4.1 ± 1.5 in 2013, respectively (Table 3).The highest char-EC to soot-EC ratio was obtained in the BBIC trajectory, thereby confirming the influence of BB activity.The other values suggest the influence of anthropogenic Table 2. Trajectory-wise data for PM 2.5 aerosol mass, resolved water-soluble inorganic ions (WSIs), carbonaceous content, water-soluble organic carbon (WSOC), as well as ratios of different components as mass concentrations (µg m -3 ) and weight percentages (%) of the resolved PM 2.5 collected in Sonla, Northern Vietnam in 2012.22.0 ± 6.6 7.5 ± 3.9 35.0 ± 8.4 Note: the value listed under "% resolved" is obtained from a fraction of the resolved PM 2.5 (sum of OC, EC, and WSIs).

2012
activities with respect to BBSC and BBSS trajectory types.Such results indicate that the air mass from Indochina is highly influenced by BB activities, and the air mass from South China is affected by mixed fossil fuel combustion and BB.The air mass from the BBSS trajectory exerts an anthropogenic influence as it passes through Hanoi (the capital city of Vietnam).Kumagai et al. (2010) showed that secondary formation and emissions from BB contributed to WSOC in the Kanto Plain, Japan.The average quantities of WSOC observed in the present study were determined to be 13.4 ± 6.7 µg m -3 (37.1% of the resolved PM 2.5 and 25.2% of PM 2.5 ) and 12.1 ± 6.6 µg m -3 (33.6% of the resolved PM 2.5 and 19.9% of PM 2.5 ) in 2012 and 2013, respectively (Table 1).Levels of WSOC observed in the present study are comparable with those in studies that found major BB influence-for instance, 2.2-39.6 µg m -3 (Graham et al., 2002); 11-46 µg m -3  (Mayol-Bracero et al., 2002); 4.4-52.6µg m -3 (Decesari et al., 2006); 0.57-18.45µg m -3 (Sullivan et al., 2006), and 0.8-40.6 µg m -3 (Wonaschutz et al., 2011).Chuang et al. (2013b) have reported WSOC as high as 25.2% PM 2.5 in Chiangmai, Thailand.By contrast, Duong et al. (2011) have reported relatively WSOC as low as 6%-11% of PM 2.5 in the Los Angeles Basin and outflow regions during the 2010 CalNex field campaign.Table 1 also shows WSOC in OC to be 62.8% ± 8.2% and 57.8% ± 15.5% in 2012 and 2013, respectively.Such high percentages of WSOC in OC confirm that the OC near the BB source is mainly water-soluble (Mayol-Bracero et al., 2002;Saarikoski et al., 2007;Pio et al., 2008).The WSOC-to-OC ratio has reportedly fallen within the 45%-75% range from BB smoke in Amazon, Brazil (Mayol-Bracero et al., 2002) and 57%-68% from BB smoke aerosol in northern Europe (Saarikoski et al., 2007).Similarly, Chuang et al. (2013b) have demonstrated a WSOC-to-OC ratio of 61.6% ± 6.8% in the BB aerosol at the Chiangmai, Thailand site.Tables 2 and 3 show that WSOC distributions in the three classified trajectory types are comparable.However, the quantities are highest in the BBIC trajectory in both years (65.4% and61.3% for WSOC in OC for 2012 and2013, respectively).
with average concentrations of 6.0 ± 3.0 µg m -3 and 5.5 ± 2.4 µg m -3 , comprising 18.2% ± 7.8% and 17.6% ± 7.7% of the resolved PM 2.5 mass in 2012 and 2013, respectively.In addition to SO 4 2-, other major WSIs observed were NH 4 + followed by NO 3 -and K + .The concentrations of these WSIs were comparable in 2012 and 2013.The concentrations of SO 4 2-and NH 4 + widely varied, as revealed in Figs.6(a) and 6(b).The wider variations in SO 4 2-and NH 4 + fractions of the resolved PM 2.5 indicate that the aerosol collected at the Sonla site have additional contributions from anthropogenic sources from South China and northern Vietnam because of emissions from coal-fired power plants near the border between Vietnam and China (Hien et al., 2004;Oanh et al., 2006;Streets et al., 2009;Cohen et al., 2010).Such differences suggest a mixed source where lower value of SO 4 2-is expected from BB activity and higher value of SO 4 2-is possibly due to anthropogenic inputs.To verify such variations, the data have been further analyzed for air mass trajectory.
Trajectory-wise average concentrations of SO 4 2-in the BBSC, BBIC, and BBSS trajectories for 2012 and 2013 are depicted in Tables 2 and 3, as well as Figs.7(a) and 7(b).The SO 4 2-concentrations in the BBSC type are 8.0 ± 3.4 µg m -3 (21.7% ± 6.9% in the resolved PM 2.5 ) and 7.3 ± 2.9 µg m -3 (25.9% ± 7.5% in the resolved PM 2.5 ) for 2012 and 2013, respectively.The highest SO 4 2-concentration is found in the BBSC type among three trajectory types and is comparable with the results from the literature, indicating the occurrence of regional aerosol loading from anthropogenic sources from South China to northern Vietnam.The SO 4 2average concentration was evaluated to be 8.9 µg m -3 (18.1% in PM 2.5 ) at the Hanoi site, according to in a study on longrange transport pollution from China by Hien et al. (2004).Similarly, SO 4 2-average concentration was 10.4 ± 5.7 µg m -3 (19.6% in PM 2.5 ) in the air mass from China during the dry season (Co et al., 2014).Likewise, Hang and Oanh (2014) have reported 8 ± 4.8 µg m -3 (12.7% in PM 2.5 ) of SO 4 2-in the air mass trajectory originating from the Pacific Ocean, off the coast of South Korea and Japan, and passing over Southern China before arriving at the site in Northeastern Vietnam.
NH 4 + has also exhibited a pattern similar to that of SO 4 2-.The NH 4 + average concentrations in the BBSC type are 2.9 ± 1.0 µg m -3 (8.1% ± 2.0% in the resolved PM 2.5 ) and 2.9 ± 1.0 µg m -3 (10.4% ± 2.6% in the resolved PM 2.5 ) for 2012 and 2013, respectively.Among the three trajectory types, the BBSC trajectory has the highest NH 4 + concentration, and the quantities observed are comparable to those in the literature, indicating the combination of NH 4 + and SO 4 2-in the compound form (r = 0.98 and r = 0.94 for 2012 and 2013, respectively).Hien et al. (2004) reported 1.9 µg m -3 (3.8% in PM 2.5 ) NH 4 + at the Hanoi site, which was attributed to long-range transport of pollution from China.Similarly, NH 4 + average concentration was 4.2 ± 2.5 µg m -3 (7.9% in PM 2.5 ) in the air mass from China during the dry season (Co et al., 2014).Likewise, 2.8 ± 1.5 µg m -3 (4.6% in PM 2.5 ) of NH 4 + was reported in the air mass trajectory that originated from the Pacific Ocean and passed over South China before arriving at the site in Northeastern Vietnam (Hang and Oanh, 2014).By contrast, high concentration of NH 4 + (7.9 ± 4.8 µg m -3 ; 10.1% in PM 2.5 ) in the dry season was found at the Hanoi site, highlighting the long-range transport for such high NH 4 + , according to Hai and Oanh (2013).Moreover, the highest value of K + (a BB tracer) in the BBIC trajectory clearly indicated the dominance of BB at the sampling site (Tables 2 and 3).However, the levels of NO 3 -in terms of mass concentration and percentage of the resolved PM 2.5 in different trajectories showed a slight variation.The slight variation in NO 3 -suggested the lack of gas-to-particle transformation at the Sonla site because of near BB sources, as reported by Lee et al. (2011).Another plausible explanation for the lack of NO 3 -enhancement is the nitric acid phase partitioning to coarse particles such as crustal dust and sea salt (Sorooshian et al., 2013;Prabhakar et al., 2014), which exceeded the upper size limit detected in the present study.

Near-Source and Prescribed-Source Profile
In the atmosphere, the BB aerosol source profile can be observed by collecting aerosol at a site with visible fire spots in the surrounding environment (near-source) and directly monitoring a known BB plume such as a prescribed BB.Table 4 depicts a comparative analysis of the composition of aerosols collected from near-source and prescribed BB in terms of the resolved PM 2.5 in weight percentage (%).The abundance of SO 4 2-mass fraction is separated in categories of more than and less than 10%.Pratt et al. (2011) have characterized two prescribed BB plumes by using instruments loaded in a C-130 aircraft.In a plume, the submicron SO 4 2mass fraction increased from 0.6% (2-4 min of age) to 17% (81-88 min of age).This result indicated that SO 4 2-in the surrounding air was mixed immediately with the transported BB plume to obtain a high SO 4 2-mass fraction.Cayetano et al. (2014) found prefrontal air masses originating from polluted regions of the Asian continent influenced the particulate properties with enhanced SO 4 2levels.Therefore, the SO 4 2-mass fraction in the aerosol can be used as an index to determine the closeness of a sample to the BB source.In addition, SO 4 2-mass fraction in aerosol can also be low for areas influenced predominantly by agricultural activities (Mkoma et al., 2013) or on the periphery of the Amazon basin (Kundu et al., 2010).

Enrichment of NH 4
+ , NO 3 -and SO 4 2-with an increase in distance from the source or particle aging has been discussed by Reid et al. (2005).The data clearly show comparable K + contribution to PM 2.5 in the majority of the studies.Relatively higher contribution of K + (~14% of PM 2.5 ) was reported by Mkoma et al. (2013) for soil resuspension or enriched K + in plant tissues in Tanzania, East Africa.OC had the highest contribution to PM 2.5 in all of these studies, as shown in Table 4.The fine particles in smoke from prescribed fires contain OC up to 91% of the resolved PM 2.5 mass (Lee et al., 2005) and organics in 89% of PM 2.5 mass (Pratt et al., 2011).Such high OC content has been observed in fresh dry smoke fine particulates (Reid et al., 2005).The contribution of OC to PM 2.5 is relatively less in aerosols from the near-source BB than that from the prescribed BB.EC contribution to PM 2.5 exhibits the reverse trend of OC.Both trends of OC and EC agree to the aging effect of prescribed BB aerosol observed by Pratt et al.Table 4. PM 2.5 aerosol mass, water-soluble inorganic ions, carbonaceous content, water-soluble organic carbon (WSOC), as well as the ratio of organic carbon over elemental carbon (OC/EC) as mass concentrations (µg m -3 ) and weight percentages (%) of the resolved PM 2.5 observed in the near-source and prescribed BB.Note: the resolved PM 2.5 is the sum of OC, EC, and water-soluble inorganic ions.
(2011).Hence, the OC-to-EC ratio follows the pattern of OC contribution to PM 2.5 , i.e., higher OC-to-EC ratio values are obtained in the case of prescribed BB, compared to that of the near-source BB.Similarly, the WSOC content is relatively low in the aerosols originating from near-source BB than that in the prescribed BB.

CONCLUSIONS
The chemical species and their distinctive concentration gradients in PM 2.5 at a site in Sonla, Northern Vietnam were thoroughly investigated.The influence of air mass trajectories on aerosol components was analyzed.The enhancement of PM 2.5 mass concentration and tracer concentration gradients revealed the influence of BB activity.OC was the most abundant bulk chemical in the resolved PM 2.5 .The orders OC3 > OC4 ≒ OC2 > OC1 ≒OP in OC fractions and EC1-OP > EC2 > EC3 in EC fractions occurred most frequently among the 8 resolved carbonaceous fractions.This pattern was consistent in different trajectories for both 2012 and 2013.Therefore, such dominance of OC3 and EC1-OP previously reported in field and chamber studies can be considered suitable tracers for near-source BB aerosol.
The OC-to-EC ratios observed in this study were comparable to the findings of studies related to BB aerosols.However, the OC-to-EC ratios varied only slight in terms of trajectories.By using the char-EC-to-soot-EC ratio, the highest value was associated with the BBIC trajectory that originates from the BB source area.Hence, the char-EC-tosoot-EC ratio can more efficiently distinguish carbonaceous aerosol from BB.Meanwhile, the WSOC in OC exceeds 50%, indicating that most OC in BB aerosol is water-soluble.
The trajectory-classified data also revealed that the highest WSOC-to-OC ratio was obtained in the BBIC trajectory type (trajectories originated from BB source area).Hence, the abundance of WSOC in OC can also be used as an index for near-source BB aerosol.
The weight percentages of major WSIs, such as SO 4 2and NH 4 + , exhibited wider ranges, thereby indicating the influence of anthropogenic sources, in addition to BB.This observation is verified by trajectory types.The trajectory with air mass from South China (BBSC) has higher SO 4 2and NH 4 + content than BBIC and BBSS.Aerosol chemical tracers or indices carry the signature of the sources, which only appeared markedly when the air masses were transported from the right sources.The BBIC trajectory has enhanced BB tracers such as OC3, EC1-OP, gradient of carbonaceous contents, char-EC to soot-EC ratio, and WSOC content.
A comparative analysis of the composition of aerosols collected from near-source and prescribed BB indicates that SO 4 2-in the surrounding air is mixed fast with the transported BB plume, resulting in a high SO 4 2-mass fraction.Fine particles in smoke from prescribed fires contained OC at around 90% of the resolved PM 2.5 mass, whereas that from near-source BB was relatively less.By contrast, the fraction of EC in PM 2.5 increased with distance from the BB source site.Therefore, higher OC-to-EC ratios were obtained in the prescribed BB than in the near-source BB.

Fig. 1 .
Fig. 1.Geographic location of the Sonla site and Hanoi, Vietnam in Indochina Peninsula (adapted from Google Earth).

Fig. 3 .
Fig. 3. Temporal variations in PM 2.5 aerosol mass concentrations and bulk of inorganic ions and carbonaceous fractions during the study periods in (a) 2012 and (b) 2013 in Sonla, Northern Vietnam.

Fig. 4 .Fig. 5 .
Fig. 4. Distributions of carbonaceous fractions as weight percentages of the resolved PM 2.5 during the study periods in (a) 2012 and (b) 2013 in Sonla, Northern Vietnam.

Fig. 6 .Fig 7 .
Fig. 6.Distributions of water-soluble inorganic ions as weight percentages of resolved PM 2.5 during the study periods in (a) 2012 and (b) 2013 in Sonla, Northern Vietnam.

Table 1 .
Measurements of PM 2.5 aerosol mass, water-soluble inorganic ions (WSIs), carbonaceous content, water-soluble organic carbon (WSOC), and ratios of different components as mass concentrations (µg m -3 ) and weight percentages (%) of the resolved PM 2.5 collected in Sonla, Northern Vietnam in 2012 and 2013.