Aerosol and Air Quality Research, 13: 957 - PDF document

Aerosol and Air Quality Research, 13: 95
Aerosol and Air Quality Research, 13: 957–976, 2013 Copyright © Taiwan Association for Aerosol Research ISSN: 1680-8584 print / 2071-1409 online doi: 10.4209/aaqr.2012.07.0184 Monitoring Supersite in Korea , Suk-Jo LeeDepartment of Environmental Engineering, Chonnam National University, 300 Yongbong-dong, Gwangju 500-757, Korea ABSTRACT Hourly measurements of PM, organic and elemental carbon (OC and EC), inorganic ionic species, and elemental constituents were made between February 1 and March 31, 2011, at a South Area Supersite at Gwangju, Korea. Over the two-month study period, daily PM mass concentration exceeded the 24-hr average Korean NAAQS of 50.0 g/m on 20 days, of which two pollution episodes (episodes I and II) are investigated. Episode I (February 01–08) is associated with Particulate matter (PM) smaller than 2.5 m in aerodynamic diameter (PM) is of great concern due to the transportability in the human body. PM in the atmosphere has been linked to the formation of haze and changes in the atmosphere’s radiative balance (Charlson et al., 1987), as well as adverse effects on human health, crops, and materials (Chameides et al., 1999). Haze pollution has been studied intensively on its impact on air quality, visibility, climate, and public health (Chameides et al., 1999; Kim et al., 2001; Park et al.Aerosol and Air Quality ResearchAlthough long integrated filter-based measurements in the above-mentioned studies have mainly been made to determine aerosol chemical constituents during haze events, their temporal resolution is typically insufficient to observe short-term variations. The daily averaging times used in long integrated filter-based measurements tend to smooth out much of the variability, thus limiting the understanding of the factors influencing of haze formation dynamics and the short-term exposures that result in adverse health impacts. For these reasons, time-resolved measurements of PM2.5 and its chemical components were conducted to better understand sources and the dynamics of haze formation because chemical composition of atmospheric aerosol particles is highly variable in space and time, and is dependent on meteorological parameters (Weber et al., 2003; Park et al.2005b, 2005c, 2006a; Solomon and Sioutas, 2008). It has been indicated that during haze events, excess PM2.5 is often largely due to elevations in the concentration of one or two of the major chemical species. Such elevations are often induced by near-by sources and/or regional sources. These studies indicated that short-term transients in carbonaceous species, nitrate, CO, and NO levels were often observed in the morning hours or sometimes in evening hours. In contrast, sulfate concentrations peak during summertime afternoons when ozone levels are elevated, or can be strongly associated with the regional haze events. As a part of the development of effective control strategies to reduce the impacts of atmospheric PM including yellow sand and long-range transported aerosols, the Korean EPA in 2007 launched an intensive air pollution monitoring research program known as the Supersite program, which is similar to the U.S. EPA Supersite program (Solomon and Sioutas, 2008). The primary objectives of the Korean Supersite program are to monitor the physicochemical characteristics of the yellow sand and long-range transported aerosols, estimate the domestic background concentrations of the aerosol particles at remote background sites, provide unprecedented physicochemical characterization of ambient PM in urban areas, and evaluate semi-continuous measurement methods of PM and its chemical speciation data. This study was carried out at a South Area Supersite in Gwangju, Korea, which began in 2009. The South Area air pollution supersite study was designed to develop an extended set of time-resolved and compositionally resolved PM data for use in apportioning sources and resolving their local and regional contributions. The objectives of this study was to c

ompare results between filter-based inte
ompare results between filter-based integrated and semi-continuous measurements of mass and its major chemical constituents from data gathered for approximately two months in 2011 at a South Area Supersite in Gwangju, Korea. In addition to evaluating the near continuous methods, an attempt was made to elucidate the key chemical characteristics of two PM2.5 haze pollution episodes revealed by time-resolved measurement methods. EXPERIMENTAL Description of the South Area Supersite at Gwangju The South Area Supersite (latitude 35.23°N, longitude 126.85°E) is located in the southern part of Korea, as depicted in Fig. 1. The site is located in a semi-urban commercial and residential area which is surrounded by agricultural lands and traffic roads, and is ~20 km northwest of downtown Gwangju and northeast of two industrial complexes. Emissions from these areas to the site could be resolved during typical western and southern flow regimes. The closest major traffic road lies approximately 0.3 km southeast of the site, and the Honam express highway is located 1.5 km west of the site. Meteorological data were monitored at the Gwangju regional meteorological station, located 7 km south of the site. Gwangju has a population of about of about 1.4 million in a 501.4 km area, with 350,000 motor vehicles, of which gasoline vehicles account for ~70%. In addition, ~85% of total air pollution emissions in the city are attributed to vehicular sources. Air quality in Gwangju has been reported to be influenced by local pollution and long-range transported pollution from polluted regions of China (Kim et al., 2001; Park et al., 2005a, 2006b, 2010, 2012; et al., 2010; Jung and Kim, 2011; Park and Cho, 2011). High levels of PM pollution are also observed due to the burning of agricultural wastes in farming areas nearby the Supersite (Ryu et al., 2004, 2007; Park et al., 2006b; Jung and Kim, 2011). Thus, the South Area Supersite located in Gwangju is an excellent choice for studying the properties of local, regional, and biomass burning aerosol emissions and their effect on urban air quality, in order to investigate our hypothesis regarding time-resolved measurements and aerosol age. Highly time-resolved measurements of PM, organic and elemental carbon (OC and EC), ionic species, and elemental species have been made at the Gwangju Supersite at 1-hr intervals since January 2010 using commercial semi-continuous instruments. However, in the present study, only the measurement data recorded between February 1 and March 31, 2011 were utilized due to frequent malfunction of some instruments. Hourly ambient temperature ranged from –7.4 to 16.7°C (a mean of 3.1°C) during February and from –2.9 to 19.7°C (a mean of 5.8°C) during March. Average relative humidity (RH) during February and March was 64.8% (18.2–99.9%) and 55.5% (16.9–99.5%). 24-hr PM Mass and Speciation Measurements For comparison of the semi-continuous measurements, 24-hr integrated filter-based PM samples were collected between midnight and midnight with three sets of sequential samplers (PMS103, APM Korea) and then used for determining the total mass concentration, the amounts of carbonaceous species (OC and EC), eight ionic species (Cl, SO, Na, NH, and Mg), and 29 elemental species (Si, S, K, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Br, Rb, Sr, Mo, Cd, Sn, Sb, Te, Cs, Ba, Hg, Pb, and Bi). During the filter-based sampling of aerosol particles, the artifacts that can arise due to particle-to-particle interactions, gas-particle interactions and the dissociation loss of semi-volatile species can change the collected aerosol composition. Positive and negative artifacts may have been created in this study because the sequential sampler used did not have a denuder to remove gases prior to sampling Park et al.Aerosol and Air Quality ResearchFig. 1. Area map of Gwangju PM Supersite. particles or a backup filter to collect vapor from particles collected on the filter. Issues arising from the sampling

artifacts are discussed below. The Tef
artifacts are discussed below. The Teflon filters (Teflo membrane, 2.0-m pore size, Gelman Science, USA) were weighed before and after sample collections with a microbalance of 1 g sensitivity and analyzed for elemental components using X-ray fluorescence (XRF). The filters were conditioned for about 48hr in a clean chamber maintained at RH of 40% and temperature of 20°C. Aerosol samples collected on 47-mm diameter quartz fiber filters were analyzed for OC and EC using the NIOSH thermal-optical transmittance (TOT) method. In our protocol, OC and EC were determined as follows. OC was evolved under a stream of high purity He (99.999% higher). The sample was heated in five steps to a final temperature of 800°C in order to volatilize OC and EC in a helium atmosphere. To evolve EC and pyrolyzed OC, which were removed in the second part of the analysis, the sample was first cooled to 550°C, and then heated under a mixture of 2% O/98% He, in three temperature steps, to a final temperature of 800°C. The TOT analyzer utilized laser transmission to correct for OC charring. EC was determined as the carbon evolved after the filter transmittance returned to its initial value. The detection limits of OC and EC, which are defined as twice the average of the field blanks, were 0.15 and 0.03 , respectively. The precision of the OC and EC measurements was 4.0 and 7.5%, respectively. Zefluor filters (Zefluor, 2 m pore size, Gelman Science, USA) from a sequential sampler were analyzed for eight ionic species (Cl, NH, Mg). In order to extract the ionic species from the sampled filters, each filter was first put into a 125-mL HDPE vial, and wetted in 0.2 mL of HPLC-grade COH and then 30 mL of distilled deionized water. The solution was extracted using ultra-sonication for 30 min and mechanically extracted again for 60 min. All the extracts were filtered using 0.4-Park et al.Aerosol and Air Quality ResearchTeflon filters and analyzed by ion chromatography (Metrohm Modula IC, model 818 IC pump/819 IC detector). Filter blank values were also applied for background subtraction for each of the filters. The IC system consisted of an anion IC (Metrohm Modula with a suppressor, equipped with a Metrohm Metrosep A Supp-5, 4 × 150 mm, anion column) and a cation IC (Metrohm Modula without a suppressor, equipped with a Metrohm Metrosep C4, 4 × 150 mm, cation column). The eluants were 3.5 mM sodium carbonate )/1.0 mM sodium bicarbonate (NaHCO) for the anion IC and 4.0 mM nitric acid (HNO)/0.7 mM dipicolinic flow rate for the anion and cation was maintained at 0.7 and 0.9 mL/min, respectively. The measured detection limits for Cl, and Mg were 12.3, 12.1, 14.9, 11.3, 4.7, 8.5, 12.4, and 4.7 ng/m, respectively. Semi-Continuous PM Measurements Mass Concentration mass concentration was determined continuously every one hour with a beta-attenuation monitor (BAM1020, MetOne Instrument Inc., USA). Integrated 24-h averages of 1-hr PM data were compared with those from 24-h integrated PM speciation sampler. The corresponding regression relationships: PM2.5BAMg/m) = (1.02 ± 0.03) g/m) + (1.50 ± 1.27), R = 0.97, indicated that the continuous BAM1020 measurements provided excellent agreement with 24-h integrated filter-based PM mass. Organic and Elemental Carbon Measurements Concentrations of OC and EC in PM were measured using a model-4 semi-continuous OC-EC field analyzer (Sunset Laboratory Inc., USA) with 1-hr time resolution. This analyzer is based on the NIOSH method 5040 (NIOSH, 1996). In the carbon analyzer, OC and EC were determined as follows: OC was evolved under a stream of ultrahigh purity He (99.999% higher) while the sample was heated in two temperature steps (600°C for 80 s and 840°C for 90 s). The evolved carbon was oxidized to CO in an oxidizing oven. The CO is then swept out of the oxidizing oven by a helium stream and measured directly by a self-contained non-dispersive infrared (NDIR) detector system. To evolve EC and pyrolyzed OC, the sample was cooled to

650°C, and then heated under a mixture o
650°C, and then heated under a mixture of He/O in two temperature steps (650°C for 35 s and 880°C for 105 s). The analyzer accounted for this pyrolysis by defining the split between the OC and EC as the point when the analysis transmittance is equal to the transmittance before analysis. The EC is determined as the carbon evolved after the filter transmittance returned to its initial value. Ambient air was drawn at 8.0 L/min through a PM sharp-cut cyclone. A carbon impregnated multi-channel parallel-plate diffusion denuder was used to remove any semi-volatile organic vapors absorbed on the quartz filter media (Turpin et al., 2000). However, the use of a denuder also disturbs the gas/particle equilibrium at the quartz fiber filter, which can cause OC to volatilize from the filter (Turpin et al.2000). During the 2-month study period, a total of 1,338 1-hr measurements, out of a possible 1,415, were recorded, yielding a data capture efficiency of 95.0%. However, 89.0% of these data were flagged as good data and the remaining 11% as invalid because of filter replacement, maintenance, regular check-up, filter blowouts during sampling, power failure, etc. To determine the degree of equivalence, 24-hr integrated OC and EC concentrations (52 measurements available) were regressed against 24-hr averages of the 1-hr OC and EC measurements from the semi-continuous carbon analyzer. In calculating the 24-hr averages, regression analyses were made using only valid OC and EC data. Data pairs were used in the regressions only if 30% of the 24 1-hr measurements were missing. Regressions were made for 43 data pairs during the study period. Fig. 2 represents the relationship between 24-hr averages of the 1-hr data and 24-hr integrated data for OC and EC measurements. Sunset OC concentrations were ~27% lower than 24-hr integrated OC concentrations [OCsunset= (0.73 ± 0.04)OCfilterg C/m) + (0.05 ± 0.20), R = 0.94]. Sunset EC data were also ~27% lower than 24-hr integrated EC concentrations [ECSunsetg C/m) = (0.73 ± 0.03)ECfilterg C/m) + (0.12 ± 0.05), R = 0.90]. The difference in the OC data could have been due to positive and negative artifacts between the semi-continuous carbon analyzer (operated with the denuder) and the PM speciation sampler without the denuder (Park et al., 2005b; Polidori et al., 2006), to the different face velocities of air, or to the different TOT temperature programs (Schauer et al., 2003). The face velocity of air passing through the filters was 20.1 and 111.1 cm/s for the speciation sampler and semi-continuous carbon analyzer, respectively. Therefore, the OC data from the PM speciation sampler were expected to be clearly higher than the semi-continuous measurements. The large discrepancy for the EC data was attributed to the high detection limit of EC in the semi-continuous carbon analyzer at the 1–2 hr time resolution (Lim and Turpin, 2002) and differences in the temperature programs. A previous study has indicated that EC measurements are sensitive to operating conditions despite the use of the same analytical method, i.e., TOT (Schauer et al., 2003). In this work, the original OC and EC data were used to test the mass balance closure of PM2.5 and to estimate the relative contribution of carbonaceous particles to the PMIonic Species Measurements Hourly concentrations of eight ionic species in PM2.5 were measured using an ambient ion monitor (AIM, URG9000D, URG Corporation). The AIM monitor drew air in through a 2.5 sharp-cut cyclone at a volumetric-flow controlled rate of 3 L/min to remove the large particles from the air stream. In order to minimize the loss of particles during sampling, a Teflon-coated aluminum pipe was set up in the vertical direction. The sample was then drawn through a liquid diffusion denuder where the interfacing acid and alkaline gases were removed. In order to achieve high collection efficiencies, the particle-laden air stream then entered the aerosol super-saturation steam chamber to enhance the parti

cle growth. Grown particles, i.e., dropl
cle growth. Grown particles, i.e., droplets, were collected every hour by an inertial particle separator and were then injected into the ion chromatograph. Both particles and Park et al.Aerosol and Air Quality ResearchFig. 2. Comparison between filter-based and semi-continuous measurements for PM major chemical components. gases were collected and injected using syringe pumps, PEEK valves, Teflon® tubing and PEEK tubing. A detailed description of the monitor can be found elsewhere (http://www.urgcorp.com). The 24-hr averages of hourly Cl, NH were regressed against 24-hr integrated ionic species data measured with a PMS103 speciation sampler. Comparison results between the two methods for SO concentrations are shown also in Fig. 2. Very strong correlations were found between the two methods with R of 0.90–0.97. AIM SO concentrations were about 10% lower than the filter-based SO measurements with 24-hr average filter-based K conc (g/m0.00.40.81.21.62.02.42.83.23.64.024-hr average AIM K conc (0.00.40.81.21.62.02.42.83.23.64.0AIM vs. filter-based KLinear fit: y = 090x - 0.02, R=0.9724-hr average filter-based OC conc (g C/m0246810121416182024-hr average Sunset OC conc (g C/m24612161820Sunset carbon monitor vs filter-based OCLinear fit: y = 0.73x + 0.05, R=0.9424-hr average filter-based SO conc (g/m03691215182124273024-hr average AIM SO conc (g/m036152124AIM vs filter-based SOLinear fit: y = 0.89x + 0.49, R=0.9024-hr average filter-based NO conc (g/m048121620242832364024-hr average AIM NOg/m0482024323640AIM vs. filter-based NOLinear fit: y = 0.76x + 0.28, R=0.9524-hr average filter-based EC conc (g C/m0.00.51.01.52.02.53.03.54.04.55.024-hr average Sunset EC conc (g C/m0.51.01.52.53.04.04.55.0Sunset carbon monitor vs. filter-based ECLinear fit: y = 0.73x + 0.12, R24-hr average filter-based NH conc (0246810121416182024-hr average AIM NH conc (g/m02481820AIM vs. filter-based NHLinear fit: y = 1.00x + 0.00, R=0.91Park et al.Aerosol and Air Quality Researchan R of 0.90. The SO from AIM has been reported to exhibit a positive interference at SO concentration exceeding 30 ppbv and negative readings at SO levels exceeding 20 (Wu and Wang, 2007). Although not shown here, however, no SO interference was observed in the present study. The observed AIM SO data were corrected based on the regression slope between the two methods to check the mass balance closure of PM2.5 concentrations were strongly correlated with the filter-based NOmeasurements with a slope of 0.76 and an R of 0.95, indicating that the AIM NO concentrations were about 24% lower than those from the filter measurements. The higher NO concentrations on the filter samples may have been due to absorption of HNO on the filter, causing positive bias. The sampling artifacts of nitrate and ammonium also occurred due to evaporation of semi-volatile NH from particles collected on the filter (Koutrakis et al., 1992; Zhang and McMurry, 1992), leading to a negative bias which can be substantial (45–75%) at high temperature and at low aerosol loading (Nie et al., 2010; Saarnio et al., 2010). However, the negative artifact of NO particles in this study may not have been significant because of low ambient temperatures and very high aerosol loading. Low temperature in winter favors the partition of ammonium nitrate in the particulate phase. Hence, the sampling artifacts probably were not a major problem in winter. AIM hourly NO data were corrected using a regression relationship (slope and intercept) between the 24-hr averages of AIM hourly data and 24-hr integrated speciation data and used to investigate the characteristics of ionic species observed during two haze pollution episodes, which are discussed below. Measurements of Elemental Constituents Hourly concentrations of elemental constituents (K, Ca, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Ag, Cd, Sn, Sb, Ba, Ag, Tl, and Pb) in PM were measured with an ambient metals monitor (model: Xact 620, Cooper Enviro

nmental Services LLC, USA). The Xact 620
nmental Services LLC, USA). The Xact 620 monitoring system uses reel-to-reel filter tape sampling and nondestructive XRF (Ed- this acronym has already been defined above) analysis to monitor ambient air. The air was sampled at a flow rate of 16.7 L/min through a PM inlet and drawn through a filter tape. The resulting PM deposit was automatically advanced and analyzed by XRF for selected metals as the next sample was being collected. Sampling and analysis were performed continuously and simultaneously, except during advancement of the tape (~20 sec) and during daily automated quality assurance checks. The measurement method of the metal species in ambient air particles was based on EPA Method IO 3.3 (“Determination of metals in ambient PM using XRF). The minimum detection limits of K, Ca, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Cd, Sn, Sb, Ba, Tl, and Pb for time resolution of 1-hr were 0.81, 0.32, 0.18, 0.14, 0.11, 0.07, 0.08, 0.05, 0.04, 0.06, 0.04, 0.03, 0.03, 1.35, 2.54, 0.67, 0.40, 0.05, and 0.05 ng/m, respectively. The 24-hr averages of hourly elemental species measurements (K, Ca, Fe, Mn, Cu, Zn, As, and Pb) were compared with those derived from 24-hr integrated filter sample (results from other elemental species are not provided here because of insufficient paired data). Agreement between elemental species concentrations measured by the two methods showed correlation coefficients ) greater than 0.80, with the exception of Fe (R = 0.70). Slopes between the measured concentrations from the two methods ranged from 0.74 (Zn) to 1.80 (Ca), with the majority in the range of 0.83–1.15, suggesting high accuracy of continuous XRF measurements for these species. RESULTS AND DISCUSSION Reconstructed PM Mass Balance Closure Hourly-based chemical mass balance closure was assessed by comparing BAM1020 PM mass concentration. The reconstructed PM mass was computed by summing the concentrations of organic matter (OM), EC, SO, crustal material, and trace metals. OM was obtained by applying a multiplicative factor of 1.6 (i.e., urban aerosols) to OC to account for the oxygen and hydrogen associated with OC (Turpin and Lim, 2001). We estimated the concentration of crustal material by using the following equation reported by Malm et al. (1996): Crustal material (g/m) = 2.20[Al] + 2.49[Si] + 1.63[Ca] + 2.42[Fe] + 1.94[Ti]. The sum of the concentrations of all the other elements is reported here as the trace metals value. In this study, because Al and Si data were not available, the contribution of these elements was not included in the estimated concentrations of the crustal material, thus resulting in an underestimation of the crustal material in the reconstructed fine mass. Fig. 3 compares the BAM and reconstructed PM concentrations. As shown, the reconstructed PM was highly correlated with the measured BAM PM concentrations with an R of 0.92. However, the reconstructed PM was about 19% lower than the measured PM. As discussed before, the reconstructed mass concentrations may have been underestimated due to low contributions of crustal materials, OM, and EC to PM mass and particle water content. Descriptions of Haze Pollution EpisodesAs shown in Fig. 3, it is apparent that PM mass was highly variable on a time scale of 1 hr, whereas the use of 24-hr average largely reduces this variability. For example, the hourly averaged PM mass concentration on February 4 reached 159.0 g/m, i.e., approximately 1.3 times greater than the 24-hr average mass of 122.8 g/m. Over the two-month study period, the average 1-hr PM mass concentration was 44.4 ± 30.0 µg/m with a maximum of 159.0 , and the 24-hr averaged BAM PM mass ranged from 14.6–122.8 g/m. The 24-hr averaged Korean NAAQS of 50.0 g/m was exceeded on 20 days during the study period. These daily PM mass concentration ranged from 50.4 g/mon March 20 to 122.8 g/m on February 4. These pollution episodes were associated with regional haze sulfate and nitrate events, biomass burning emissions, and local traff

ic emissions under air stagnation and lo
ic emissions under air stagnation and low-level inversions conditions. Out of 20 days exceeding the 24-hr averaged Korean PM2.5 NAAQS, two distinct haze episodes are discussed herein: February 1–8 Park et al.Aerosol and Air Quality ResearchFig. 3. Comparison of hourly measured and reconstructed PM mass concentrations. (episode I) and March 11–12 (episode II), 2011. Two haze episodes were classified based on CO/NO ratios and Kconcentrations. Haze episode I which developed between February 01 and 08, was strongly associated with regional haze pollution along with wildfire smoke emissions over southern China. This episode was further characterized by elevated afternoon SO and O concentrations (O = 46–60 ppb on February 01–05). Episode II was from 14:00 March 11–17:00 March 12 during which hourly PM increased from a background level of 23 g/m to a maximum of 116 g/m. This episode was characterized by locally produced pollution with low CO/NO ratios and broad peaks for OC, EC, and NO concentrations. During episode I, local wind speeds were generally weak to moderate averaging 1.1 m/s (range: 0.1 to 3.7 m/s). Ambient temperatures ranged from –7.4 to 9.9°C (mean: 2.7°C). Relative humidity (RH) had a range sufficient to cause hygroscopic growth (26.1 to 96.6%; mean = 67.1 ± 19.6%), and ranged from� 75–80% in the evening/morning hours. During episode II, temperatures were in the range of –0.1 to 16.2°C (mean = 7.6°C). Winds ranged from 0.1 to 6.3 m/s (mean = 2.2 m/s), and were weak during the night (0.1–1.6 m/s). Relative humidity ranged from 19.8–92.9% with a mean of 56.3%. Daytime RH was low (0%), but&#x 44.;退 was 75% during the night, especially on the evening/morning. To examine the differences in the chemical composition of PM observed during the two episodes, we analyzed MODIS images of two types, i.e., Aqua & Terra MODIS image (http://earthdata.nasa.gov/firms) and MODIS true color images (http://lance-modis.eosdis.nasa.gov/cgi-bin/imagery/ realtime.cgi), and transport pathways of the air mass, as illustrated in Fig. 4. The MODIS images show that during episode I, numerous fire counts were observed and a thick haze layer lingered over northern and southern China, and the Korean peninsula. While during episode II, a clear sky was observed over northern China and the Korean peninsula. The transport pathways of the air mass prior to arriving at the site were determined using the Hybrid Single-Particle Lagrangian Integrated Trajectories (HYSPLIT 4.5) model (Draxler and Ralph, 2012). Isentropic five-day backward trajectories were computed for every 12-hr (0000 UTC and 1200 UTC time) interval for each day of the two episodes, at three altitudes (500, 1000, and 1500 m) above ground level (AGL). Two of the trajectory results are also presented in Fig. 4, one for episode I (1200 UTC February 04) and the other for episode II (1200 UTC March 11). For episode I, air masses originating from the polluted regions of northern China (at 500 and 1000 m AGL) or regions of Southern China (at 1500 m AGL), passed over the Yellow Sea and reached the site. Air mass trajectories for episode I Measurement period (2011)02/01 02/08 02/15 02/22 03/01 03/08 03/15 03/22 03/29 mass concentration (g/m100120140160180200MetOne BAM PMReconstructed PM2.5MetOne BAM PM conc (020406080100120140160180200Reconstructed PM2.5 conc (g/m100120140160180200Reconstructed versus BAM PMLinear fit: y = 0.81x - 2.3, R=0.91Park et al.Aerosol and Air Quality Research Fig. 4. MODIS images and back trajectories of air mass arriving at the site for the two haze episodes. Symbols in : 500 m AGL, : 1000 m AGL, : 1500 m AGL. Park et al.Aerosol and Air Quality Researchindicated a combination of transport directions of northern and southern China and the air mass was classified as a mixture of wild-fire plumes, regionally transported and local pollution. Similar to the transport pathway for episode I, the air mass for episode II originated in the regions of

northeastern China and arrived at the s
northeastern China and arrived at the site, but without any observable influence of wild-fire plumes. In this study, ratio of CO to NO (= NO + NO) was also used to examine the evolution of the aerosol chemical composition with respect to the age of the air masses as a proxy for proximity to major pollution sources and atmospheric processing (Mogan et al., 2010; Freney et al.2011). Morgan et al. (2010) classified CO/NO ratios of 5–15, 10–50, and� 50 as urban, near-source and aged regional conditions, respectively. Temporal variation in the CO/NOratio is shown in Fig. 5. The mixing ratios of CO and NOfor episodes I and II were 896 (475–1319 ppb) and 32 (6–139 ppb), and 482 (247–688 ppb) and 40 (7–97 ppb), respectively, with greater impact of CO during episode I. The CO/NO ratios ranged from 9 to 159 with an average value of 41 ± 26 in episode I and from 6 to 44 with an average of 18 ± 11 in episode II, respectively, confirming that the air masses arriving at the site during episode I were associated with more aged regional components than those during episode II. Fig. 6 shows the mass factions of EC, OM, SO and NO to the PM mass against the CO/NOratio for the two episodes. As shown in Fig. 6, the significant amounts of EC, OM, SO and NO observed during episode I were associated with high CO/NO �ratio ( 50). Higher EC and OM contributions to the PM favor low ratios arising from presumably local emission sources. However, their absolute concentrations were not greatly affected by the CO/NO ratio (not shown here). That is, it is found that high concentrations of EC, OC, SO were also related to either the boundary layer air or regional aged air masses. Table 1 summarizes the PM concentrations and the chemical components for episodes I and II. As shown in Table 1, as much as 50% of the PM mass is attributed to secondary SO particles, with NO being most significant fraction of the PM mass for the two episodes. The large difference in chemical compositions between the two haze episodes is due to the NO and Kconcentrations. Fig. 7 shows temporal profiles of PMOC, EC, K, and SO concentrations for the two pollution episodes. Fig. 8 shows correlation relationships between OC and K, and between K/OC ratio for the period of February 04–05. Temporal profiles of NO, NO and Oconcentrations, ambient temperature, and RH are shown in Fig. 9. Characteristics of Carbonaceous Particles OC and EC concentrations were 6.9 (2.6–14.0 g C/mand 2.2 (0.5–4.5 g C/m) for episode I and 6.3 (2.0–10.3 C/m) and 2.0 (0.7–5.1 g C/m) for episode II, respectively. OM and EC contributed 13.1 ± 2.7% and 3.3 ± 0.7% to the measured PM2.5 for episode I, and 14.2 ± 3.8% and 3.1 ± 0.8% for episode II, respectively. The EC and OC concentrations were strongly correlated with the following regression relationships; for episode I: OC = 2.76 EC + 0.95, R = 0.71, for episode II: OC = 3.01 EC + 0.46, R = 0.83, suggesting common source emissions (i.e., vehicles) during episodes I and II. The OC/EC ratio was 3.2 ± 0.6 (1.8–5.4) and 3.3 ± 0.6 (2.1–4.8) for episodes I and II, respectively. Diurnal OC and EC peak concentrations over the two episodes occurred mainly during morning and evening rush-hour periods, indicating the significant impact of traffic. As the boundary layer rose throughout the morning and early afternoon, the OC and EC concentrations tended to decrease. However, close inspection of Fig. 7 revealed high background OC and EC levels during episode I with values of 3.0–4.0 and ~2.0 g C/m, respectively, probably due to long-range transport of polluted air masses (shown in Fig. 4) and air stagnation conditions. The OC and EC background levels during episode II were approximately 1.0 and 0.5 g C/mrespectively. According to MODIS satellite images and air mass back trajectory analysis (Fig. 4), huge forest wildfires occurred in the southern regions of China on February 01–02, and the air mass passing over these wildfire regions reached th

e Fig. 5. Variations of CO/NO ratio f
e Fig. 5. Variations of CO/NO ratio for the two haze episodes. March 11~12Measurement period (2011)03/11 0003/11 0803/11 1603/12 0003/12 0803/12 1603/13 00CO/NO ratio (-)06080100120160180200CO/NOFebruary 1~8Measurement period (2011)02/01 02/02 02/03 02/04 02/05 02/06 02/07 02/08 02/09 ratio (-)04080100140160200CO/NO ratioPark et al.Aerosol and Air Quality ResearchFig. 6. Fractions of EC, OM, SO, and NO to the PM against CO/NO ratio for the two episodes. sampling site on February 04. Therefore, relationships among , and the K/OC ratio from February 04–05, when the K concentration significantly increased, were examined to evaluate the effect of the forest fire plumes on carbonaceous particles observed at the site (Fig. 8). It has been known that biomass burning emissions contribute significantly to atmospheric OC, EC, and K concentrations (Andreae, 1983; et al., 1995; Andreae and Merlet, 2001; Reid et al.2005; Park et al., 2006b; Ram and Sarin, 2010, 2011). Vehicles and fossil fuel emissions are the dominant sources February 01~08 ratio020406080100120140160180200EC/PM2.5 (%)012392.5 vs. CO/NO ratioMarch 11~12 ratio0102030405060708090100EC/PM2.5 (%)0134672.5 vs. CO/NO ratio020406080100120140160180200Organic mass (OM)/PM2.5 (%)03OM/PM2.5 vs. CO/NO ratio ratio0102030405060708090100Organic mass (OM)/PM2.5 (%)02.5 vs. CO/NO ratioFebruary 01~08CO/NO ratio020406080100120140160180200/PM2.5 (%)07 vs. CO/NO ratioMarch 11~12CO/NO ratio0102030405060708090100/PM2.5 (%)076370/PM vs. CO/NOMarch 11~12CO/NO ratio0102030405060708090100/PM2.5 (%)07/PM vs. CO/NO ratioFebruary 01~08CO/NO ratio020406080100120140160180200/PM2.5 (%)07/PM2.5 vs. CO/NO ratioPark et al.Aerosol and Air Quality ResearchTable 1. Concentration summary of chemical species in PM for the two episodes. Parameter Unit Episode IIMean Range Mean Range g/m 88 27–159 78 23–116 2.2 0.5–4.5 2.0 0.7–5.1 6.9 2.6–14.0 6.3 2.0–10.3 g/m 18.5 3.1–48.4 16.5 8.4–27.7 g/m 22.3 6.2–53.1 32.3 5.7–49.5 g/m 1.7 0.3–4.8 1.5 0.3–3.0 g/m 0.9 0.6–1.1 0.8 0.7–0.9 g/m 8.9 4.0–15.1 6.3 2.8–8.2 g/m 1.8 0.3–5.6 0.5 0.2–0.7 OC/EC - 3.2 1.8–5.4 3.3 2.1–4.8 /OC - 0.26 0.07–0.63 0.10 0.06–0.20 - 0.55 0.25–0.76 0.50 0.37–0.71 g/m 2.0 0.3–5.9 0.9 0.4–1.3 g/m 0.27 0.03–0.74 0.26 0.14–0.41 g/m 0.46 0.09–0.92 0.54 0.24–1.31 Ti ng/m 17 2–40 15 8–28 V ng/m 8 3–29 8 2–13 Cr ng/m 9 2–16 13 4–28 Mn ng/m 39 10–87 56 21–130 Ni ng/m 6 3–12 8 2–14 Cu ng/m 23 6–54 28 11–64 g/m 0.13 0.05–0.32 0.30 0.10–0.60 As ng/m 17 3–34 16 6–28 Se ng/m 8 2–13 6 2–8 Cd ng/m 6 1–46 23 2–57 Sb ng/m 73 23–111 86 46–121 Ba ng/m 97 11–325 36 17–61 Pb ng/m 130 28–308 107 41–151 Note) Episode I covers the period of February 01–08, 2011; Episode II covers the period of March 11–12, 2011; + SO). of OC and EC, but have little impact on K concentrations. In this study, OC was strongly correlated with EC (R = 0.82) during the 2-days period, but the OC/EC ratio (range: 2.8–5.3, mean: 3.4) did not increase, which was significantly lower than those reported for Savannah burning, forest fires, and agricultural waste burning emissions (Andreae, 1983; et al., 2005b, 2006b; Reid et al., 2005; Ram and Sarin, 2011; Ram et al., 2012a, 2012b; Zhang et al., 2012b), or slightly lower than that (5.7) from emissions of forest fires in the Indochina (Chuang et al., 2012). Other studies have shown even when biomass materials are burned, a low OC/EC ratio is possible because it depends greatly on the type of materials burned and combustion conditions (Turn et al., 1997; Saarnio et al., 2010; Bae et al., 2012; Zhang et al., 2012a). For example, Turn et al. (1997) indicated that for wood fuels such as Walnut prunings, Almond prunings, and Ponderosa pine slash, the OC/EC ratio in PM was in the range of 1.7–2.1. It was also found that when pine needles, cherry trees, and acacia trees are combusted, their OC/EC ratios in PM were 0.83, 0.20, and 0.21, respectively (Bae et al., 2012). As shown in Fig. 7, the

concentration of Kan indicator of bioma
concentration of Kan indicator of biomass burning emissions (Simoneit et al.2004; Park et al., 2006b; Zhou et al., 2009; Saarnio et al.2010; Chuang et al., 2012), increased from 1.5 µg/m at 00:00 to 5.6 µg/m at 12:00 on February 04, and remained high for 2 days. K concentration accounted on average 2.9% of the PM (1.2–4.8%), which was comparable to that (2.5%) reported by a recent study of Chuang et al.(2012). In addition to the biomass burning, the abundance of K in urban air is attributed to soil dust. To assess the origin of K in this study, the relationship between total K was examined. Previous studies indicated quite low /K ratio of 0.1–0.2 in geological material (Gertler et al.1995) and 0.01–0.02 in Mongolia soil dust (Kim et al.2010). Water-soluble K and total K were strongly correlated (slope = 0.82; R = 0.96) for episode I (not shown here), suggesting these two constituents were clearly derived from a similar emission source, most likely forest fire emissions. This was evidenced by the MODIS satellite image, in which numerous fire counts were observed during episode I, and transport pathway of air masses. Scatter plots between OC – and K – K/OC for high K episode (February 04–05) were strongly correlated (R of 0.77 and 0.77, respectively) (Fig. 8), suggesting an emission source from biomass burning et al., 2012b). The K/OC ratio for the 2-days period showed a range of 0.15–0.63 with a mean of 0.38, which was significantly higher than those reported for biomass burning aerosols. Previous studies have reported KPark et al.Aerosol and Air Quality ResearchFig. 7. Temporal trends of PM, OC, EC, K and SO concentrations for the two haze episodes. ratios from 0.08–0.10 for savanna burning (Echalar et al.1995), 0.04–0.13 for agricultural waste burning (Andreae and Merlet, 2001), 0.19–0.21 for wheat straw burning/fertilizer (Duan et al., 2004), 0.02-0.14 for biomass burning (Ram and Sarin, 2010, 2011; Ram et al., 2012a), and 0.04–0.24 for wheat and/or stalk burnings (Zhang et al., 2012b). The high K/OC ratio found in this study may be attributed to traffic and fossil fuels emissions, as well as biomass burning emissions. Accordingly, even though low OC/EC and high /OC ratios were observed for the February 04–05 episode, relatively high K concentrations and the significant linear relationships between EC – OC (R = 0.82), OC – K (R = 0.77), and K – K = 0.77) suggest contribution of carbonaceous aerosols from long-range transport of biomass burning plumes occurring in the southern regions of China during episode I, as well as the traffic emissions. Sulfate Concentration The concentrations of SO observed during episodes I and II were 18.5 ± 8.7 g/m (3.1–48.4 g/m) and 16.5 ± March 11~12Measurement period (2011)03/11 0003/11 0803/11 1603/12 0003/12 0803/12 1603/13 00OC concentration (g C/m02414161820EC concentration (g C/m012789February 01~08Measurement period (2011)02/01 02/02 02/03 02/04 02/05 02/06 02/07 02/08 02/09 OC concentration (g C/m026814161820EC concentration (g C/m01345789EC Februry 01~08 Measurement period (2011)02/01 02/02 02/03 02/04 02/05 02/06 02/07 02/08 02/09 2.5 concentration (04080100PM2.5 time-series plotMarch 11~12Measurement period (2011)03/11 0003/11 0803/11 1603/12 0003/12 0803/12 1603/13 002.5 concentration (g/m100120140160180200PM2.5 time-series plotFebruary 01~08Measurement period (2011)02/01 02/02 02/03 02/04 02/05 02/06 02/07 02/08 02/09 concentration (0618243042485460 concentration (01345789 (right "y" axis)March 11~12Measurement period (2011)03/11 0003/11 0803/11 1603/12 0003/12 0803/12 1603/13 00 concentration (062430 concentration (0.20.81.0 (right "y" axis)Park et al.Aerosol and Air Quality ResearchFig. 8. Regression relationships between K and OC (left) and between K/OC and K (right) for February 04–05. 5.7 g/m (8.4–27.7 g/m), respectively, contributing 20.7 % (6.4–53.4%) and 25.2 % (9.4–53.5%) to the PM. The aerosols in fine pa

rticles are formed through gas phase rea
rticles are formed through gas phase reactions, in-cloud processes and aerosol droplet processes (Guo et al., 2010). The sulfur oxidation ratio [SOR = SO + SO)] has been used as a good indicator to assess the atmospheric transformation of SO to SO (Kaneyasu et al., 1995; Sahu et al., 2009). SOR in winter is expected to be low due to the low temperature and low photochemical flux, lowering the possibility of gas-phase oxidation of SOto SO. Non-sea salt SO was calculated by subtracting the sea-salt SO from the measured SO. SO in seawater was estimated from the SO ratio in seawater. Previously reported 24-hr measurements made at a Gwangju urban site of Korea, which is 10 km away from our site, indicated an SOR of 0.51 in summer and 0.14 in winter (Park et al., 2010, 2012). The calculated SOR was 0.55 ± 0.11 (0.25–0.76) and 0.50 ± 0.09 (0.37–0.71) for episodes I and II, respectively, which were substantially high even in winter. Urban SO in PM is composed of urban and regional components. The regional SO contribution can be evaluated by its diurnal behavior and background levels. For episode I, the temporal profile of SO concentration (Fig. 7) exhibited a continuous increase in the SO background level from 3.0 g/m on February 1 to 20.0 g/m on February 5 suggest a greater regional contribution, and are probably due to long range transport and regional sulfate formation, or air stagnation. An obvious diurnal pattern was observed with maximum SO in the afternoon and the SOtransient increased gradually from 26.6 g/m on February 1 to 48.4 g/m on February 4. The maximum hourly O concentrations on February 1, 2, 3, 4, and 5 were 46 and 13 ppb, 50 and 16 ppb, 58 and 14 ppb, 52 and 19 ppb, and 48 and 12 ppb, respectively, which were relatively high even in winter. The SO concentration was highly correlated with SO with an R of 0.64 (data not shown here). This implies that local SO emissions were also an important source of SOformation, and that slow gas-phase oxidation of SO probably played a critical role in the SO production. During episode I, the temporal variation of sulfate concentrations and air mass trajectories suggested that some of the sulfate excess over background could be attributed to sources in the local region. As a result, the high SOR and highly elevated SO concentration observed during episode I suggest that in addition to local production, SOwas probably caused by long-range secondary atmospheric processing. For episode II, SO concentrations continued to increase from 14:00 on March 11 until 19:00 to a maximum of 23.6 g/m, accounting for 33.7% of the PM, after which the dropped to 9.3 g/m at 07:00 on March 12. At this time it increased again and peaked at 13:00 hr with a maximum of 27.7 g/m, constituting 40.2% of the PM2.5then decreased gradually. Hourly O and SO concentrations were in the ranges of 47–50 and 5–11 ppb between 14:00 and 19:00 on March 11, respectively. They subsequently decreased to 17 and 2 ppb at 07:00 hr of March 12 and again increased to 66 and 13 ppb at 13:00 hr, respectively. Hourly temperature and RH were 8.2–13.0°C and 20–43% between 14:00 and 19:00 on March 11, and 4.6–15.0°C and 28–75% between 07:00 and 13:00 on March 12, respectively. The SO peaks occurring in the afternoon hours on both March 11 and 12 coincided with maximum ozone concentrations, thus, both SO peaks were driven by photochemistry. Throughout episode II, the SOconcentration was strongly correlated with SO (R = 0.85) (not shown here). Local wind directions in the afternoon hours were between 230 and 250° (not shown here) where industrial sources are located. This suggests a number of industrial sources located in southwest direction from the site may have contributed to the observed SO concentration. Consequently, the pattern of these SO excursions in relation to wind direction and the strong correlation between and SO suggests local SO emissions were likely an important source of SO at the site during episode II. Nitrate Concentration C

oncentrations of NO particles during epi
oncentrations of NO particles during episodes I and II February 04~05OC concentration (g C/m02468101214161820Water-soluble K concentration (g/m0124579 vs. OCLinear fit: y = 0.46x- 0.14, R=0.77February 04~05 Water-soluble K concentration (g/m0.00.71.42.12.83.54.24.95.66.37.0/OC ratio (-)0.00.10.20.30.40.50.60.70.80.91.0/OC vs. KLinear fit: y = 0.09x + 0.07, R=0.77Park et al.Aerosol and Air Quality Researchwere 22.3 ± 9.7 g/m (6.2–53.1) and 32.3 ± 13.3 g/m (5.7–49.5), respectively, accounting for 24.2% (9.5–40.4) and 40.9% (19.0–66.3%) of the PM2.5 mass. Atmospheric nitrate particles are mainly formed through gas-phase oxidation process of NO and OH during daytime and heterogeneous reactions involving nitrate radical and N during nighttime (Smith et al., 1995; Wittig et al., 2004; Park et al., 2005c). The nitrate ambient concentration is also driven by gas-to-particle partitioning of ammonium nitrate and boundary layer dynamics (Seinfeld and Pandis, 2006). For episode I, the NO concentration of 7.3 g/m at 10:00 on February 01 was gradually increased to reach a maximum of 53.1 g/m at 12:00 on February 04, accounting for 34.1% of the PM mass. As shown in Fig. 9, during this period, the pollution event exceeding 20 g/m lasted about 5 days and relatively high NO background levels were also maintained, which could be attributed to the thick haze layers lingering over southern and northern China and the Korean peninsula, and to transport pathways of air masses arriving at the site (Fig. 4). Some studies have indicated that a strong diurnal pattern in PM nitrate concentration was exhibited with a maximum in the morning and minimum in the afternoon (Wittig et al., 2004; Park et al., 2005c). Low nitrate concentrations during the daytime are typically because the nitrate in PM is mostly in the gas phase as nitric acid due to the higher temperature and lower RH compared to the morning and evening. However, as shown in Fig. 9, morning peaks and afternoon minimums in the nitrate concentrations were not readily observed at the site. Instead, the daytime nitrate transients with very high levels over background were often found around 12:00–13:00. Temperatures, RH, and wind speeds in the afternoon were 4.0–10.0°C, 26–59%, 1.5–2.5 m/s, respectively. The nitrate peaks were more enhanced in the daytime than in the nighttime. The nitrate daytime peak concentrations on February 2, 3, 4, 5 and 6 were 42.0, 40.8, 53.1, 38.0, and 21.4 g/m, accounting for 40.4, 40.0, 34.1, 27.0, and 24.3% of the PM, respectively. Periods showing these peaks were reasonably consistent with those showing high concentrations. Concentrations of NO and O at these periods were 34 and 41, 15 and 47, 21 and 48, 11 and 48, and 11 and 60 ppb, respectively. These O levels were comparable to the highest O concentrations on these days, which were observed around 15:00–17:00, with levels of 50, 58, 54, 48, and 60 ppb, respectively. During the daytime excursion periods, the concentrations of NO and carbonaceous species (OC and EC) were reduced to their background levels on the corresponding day. Typically the afternoon peak concentrations of fine nitrate particles have been attributed to the photochemical formation of nitric acid due to the increased ozone concentration, especially in summertime (Seinfeld and Pandis, 2006). However, the present situation, in which ozone levels of 40-60 ppb were observed in winter, facilitated the formation of nitric acid through gas-phase oxidation of NO. These daytime nitrate transients during episode I were probably associated with transports of severe haze lingering over the upwind locations of China (Fig. 4) to the sampling site, as well as photochemical secondary aerosol production of nitrate and low temperatures of 4.0–10°C. This hypothesis is further supported by the relatively high CO/NO ratio, as illustrated in Fig. 5. A closer look at the CO/NO ratio for episode I shows that the ratios at the times of the elevated NO levels were 46–80,

44–78, 42–91, 81–93, and 72–91, suggesti
44–78, 42–91, 81–93, and 72–91, suggesting that the air masses arriving at the site were in regional aged conditions. Evening NO peaks were observed four times, at 20:00 on each of 02, 04, 05, and 07 February, with levels of 26.9, 38.3, 22.1, and 24.9 g/m24.1, 25.1, and 26.3% of the PM, respectively. NOconcentrations at the times of these nighttime peaks were 26, 25, 35, and 43.3 ppb, respectively. The increases in in the evening were probably related to the lower mass concentrations of sulfate at night, which allowed excess to react with HNO (Seinfeld and Pandis, 2006). These evening peaks coincided with the enhanced NOconcentrations, but the ambient concentrations of the nitrate were not necessarily proportional to the concentrations of precursor emissions because the rates at which they form and their gas-to-particle partitioning are controlled by meteorological parameters other than the concentration of the precursor gas. A similar result was found elsewhere (Park et al., 2005c). After its evening peaks, nitrate decreased for some time, but was maintained at high levels due to continuous haze phenomena and poor atmospheric dispersion conditions at night. Because the RH during the nighttime (70–93%) was high enough to form aqueous nitric acid, the nitrate observed at night may have accumulated in wet aerosol particles. Thus, the significantly elevated concentrations of particles observed during the nighttime may have been related to low ambient temperatures, high RH, and the regional aged air masses. In summary, the high NO values during episode I probably reflect long-range secondary atmospheric processing of nitrate particles, in addition to local emission sources. Episode II was characterized by a rapid increase of nitrate starting at 14:00 on March 11 and persisting until 12:00 on March 12, with a small drop in concentration between 01:00 and 07:00 on March 12. As shown in Figs. 7 and 9, OC, EC, and NO concentrations peaked in the morning of March 11 prior to the onset of episode II, and were probably associated with morning rush hour traffic patterns. The enhancement of morning NO concentration, to a maximum of 18.4 g/m at 09:00, was attributed to high RH and low temperatures. RH and temperatures between 02:00 and 09:00 on March 11 were 84–93% and 0.0–1.0°C, respectively. After 09:00, the NO concentration decreased to 4.6 g/m at 13:00, started to increase again until 20:00–21:00, with a maximum of 49.5 g/m, accounting for 46.3% of the PM2.5and then decreased to about 36.0 g/m, which was still high. From 14:00 to 21:00 on March 11, as shown in Fig. 7, the temporal behaviors of OC, EC, and SO concentrations were similar to that of NO, probably due to local emissions. Moreover, NO increased from 12 to 57 ppb, and O was maintained at 40–50 ppb until 20:00 hr, but then was sharply reduced to 10–20 ppb (Fig. 9). Hourly ambient temperature and RH were 7.8–13.0°C and 24–48%, which is below the deliquescence RH (DRH) of NH (61% @25°C), respectively. Hourly wind speed was 2.2–4.6 m/s. Based on Park et al.Aerosol and Air Quality ResearchFig. 9. Temporal profiles of NO concentrations along with NO, temperature, and relative humidity for the two haze the concentrations of precursor gases and meteorological conditions, the continuous increase of NO concentration between 14:00 and 20:00 on March 11 was probably due to gas-phase production of HNO via the reaction of NO and OH at low RH (%), with the majority of the HNOresiding in the particle phase. In addition to the photochemical production of HNO, air masses transporting from regions of northern China to the site could have partially contributed to the enhanced NO concentrations in the afternoon (see Fig. 4). As shown in Fig. 9, the nitrate concentrations remained high between 22:00 March 11 and 07:00 March 12, ranging from 36.2 to 46.9 g/m. This NO excursion was associated with low NO/high NO, and high RH, following high afternoon O levels of about 50 ppb, presumably due to nighttime ra

dical chemistry. During these periods, t
dical chemistry. During these periods, the wind speed was 0.1–2.2 m/s with a mean of 1.0 m/s, and RH and the temperature were 72–79% and 3.4–8.4°C, respectively. The mixing ratios of NO and NOat 22:00 were 29 and 73 ppb, respectively. Heterogeneous reactions involving nitrate radical and N, which are March 11~12Measurement period (2011)03/11 0003/11 0803/11 1603/12 0003/12 0803/12 1603/13 00Ambient temperature (deg C)-9-6-361821Relative humidity (%)100110120Ambient temperatureRelative humidityMarch 11~12Measurement period (2011)03/11 0003/11 0803/11 1603/12 0003/12 0803/12 1603/13 00 and O concentrations (ppb)0203050708090100O3February 01~08Measurement period (2011)02/01 02/02 02/03 02/04 02/05 02/06 02/07 02/08 02/09 Ambient temperature (deg C)-9-6-391518Relative humidity (%)203040508090100110Ambient temperatureRelative humidityFebruary 01~08Measurement period (2011)02/01 02/02 02/03 02/04 02/05 02/06 02/07 02/08 02/09 and O concentrations (ppb)100O3February 01~08Measurement period (2011)02/01 02/02 02/03 02/04 02/05 02/06 02/07 02/08 02/09 and NO conc (g/m06182436425460-March 11~12Measurement period (2011)03/11 0003/11 0803/11 1603/12 0003/12 0803/12 1603/13 00and NO conc (g/m0618244854-Park et al.Aerosol and Air Quality Researchformed during nighttime (Smith et al., 1995), are a potential source of the nighttime nitrate. Nitric acid formed through the heterogeneous hydrolysis of N at night can retain its molecular form in the gas phase, or be transferred to the particulate phase. The NO3– observed between 22:00 and 07:00 may have existed in the aqueous phase because of the RH of 72–79%. The consequently elevated NO, which induced the high PM levels during episode II, was attributed to gas-phase photochemical reactions during the daytime and heterogeneous hydrolysis during the nighttime. Possible Sources of Crustal and Trace Metals in Ambient Particles One of the major anthropogenic sources in urban air is vehicle emissions. Local sources have a greater effect on EC concentrations. Variations in local diesel engines or combustion sources could affect EC at a site. Like EC concentrations, ambient concentrations of metals were dominated by local sources. Road dust is generally re-suspended by the turbulent passage of motor vehicles over local roads. The road dust profile contains crustal elements (Al, Si, Ca, Fe), along with metal constituents such as Mo, Sb, Cu, Zn, Ba, Mn, Cr, and Pb, reflecting enrichment from other sources and urban pollutants (Schauer et al.These metals are known to be emitted from road traffic abrasion processes or are important constituents of road dust (Harrison et al., 2003; Schauer et al., 2006; Bukowiecki et al.2009; Amato et al., 2010). Brake wear emissions have been reported to contain significant amounts of Ba, Zn, Cu, and Sb, as well as Fe and other crustal elements (Garg et al.2000; Sternbeck et al., 2002; Thorpe and Harrison, 2008). Previous studies indicated also that Sb in urban environments is associated with traffic emissions, and more specifically with emissions emerging from brake wear (Weckwerth, 2001; et al., 2005; Gomez et al., 2005; Johansson et al.2009). As shown in Table 1, the concentration differences of Sb and the other brake wear-related elements (Fe, Cr, Mn, Cu, and Pb) between episodes I and II were not significant. The concentration of Sb was 73 ± 16 ng/m (22.5–111.0) and 86 ± 19 ng/m (46–121) for episodes I and II, respectively. The mean concentration of Fe, one of the most abundant measured elements, was similar during the two episodes, ranging from 0.09 to 0.92 µg/m with a mean of 0.46 µg/mand from 0.24 to 1.31 µg/m with a mean of 0.54 µg/mrespectively. However, Ba, another marker of brake wear, was about three times higher in episode I (97 ng/m) than in episode II (36 ng/m), while Zn was two times higher in ) than in episode I (0.13 µg/msuggests that in addition to the road dust, there were relevant local emis

sions of these elements at the site. Tak
sions of these elements at the site. Taking EC, Fe, and Sb as measures of road traffic influence, correlation analyses among EC, crustal elements and other road traffic-related metals were performed and their results are presented in Table 2. During episode I, Sb was correlated with Fe (R= 0.71), Ca (R = 0.67), Cr (R = 0.64), and Mn (R = 0.53) and to a lesser extent with Zn (R = 0.42), Pb (R = 0.41), Cu (R = 0.34), and Ba (R = 0.26), which is suggestive of other sources such as road dust for these elements. During episode II, Sb had strong correlations with Cr, Ca, Ba, Zn, Pb, Mn, Fe, and Cu (R = 0.88, 0.86, 0.83, 0.77, 0.76, 0.76, 0.64, and 0.46, respectively). Ca, Fe, Cr, Mn and Pb are typically present at significant levels in crustal material and soils, but the relatively high correlations between Sb and the metals indicated that the concentrations of these metals must be affected by local sources other than soil. Also EC had moderate correlations with the metals during episodes I and II. Because re-suspended road dust is generally too large to dominate PM, the good relations of Sb and EC with the other metals indicate that tailpipe and brake wear emissions were important sources for these elements at the site, with a greater effect in episode II. Sb and Cd were very weakly correlated (R = 0.04) during episode II, indicating that other sources, rather than vehicle-related emissions, dominated the concentration of Cd at the site. SUMMARY AND CONCLUSION Hourly measurements of PM mass, elemental and organic carbon (EC and OC), ionic species, and elemental species were made at the South Area Supersite at Gwangju, Korea, using commercial semi-continuous instruments. During the 2 months between February 01 and March 31, 2011, a total of 20 days wherein daily PM mass concentrations exceeding 24-hr average Korean NAAQS of 50.0 were identified. Two haze episodes out of them were discussed; the episode I (February 01–08) was associated with regional pollution along with severe haze layers and wildfires emissions, and characterized further by high CO/NO ratios and high K concentrations. The episode II (March 11–12) was characterized by locally produced pollution with low CO/NO ratios, short-term SO excursions, and broad peaks in OC, EC, and NO concentrations. The EC and OC concentrations were strongly correlated with R of 0.71 and 0.83 for episodes I and II, respectively, suggesting common source emissions (i.e., vehicles). Their respective OC/EC ratio was 3.2 and 3.3. However, the concentration of K, an indicator of biomass burning emissions, was increased for February 04–05 during the episode I and reached an hourly maximum of 5.6 g/maccounting for ~93% of the total K concentration. This enhancement was clearly evidenced by the MODIS satellite image, in which numerous fire counts were observed over the southern regions of China, and transport pathways of air masses arriving at the site. Abnormally, however, the OC/EC ratio (mean: 3.4) did not increase for the 2-days period. Previous studies suggested that a low OC/EC ratio is possible because it depends greatly on the type of materials burned and combustion conditions. Accordingly, although low OC/EC ratio was observed when relatively high K concentrations occurred, the significant linear relationships between EC – OC (R = 0.82), OC – K (R = 0.77), and K = 0.77) between February 04 and 05 suggest contribution of carbonaceous aerosols from long-range transport of biomass burning plumes occurring in the southern regions of China during episode I, as well as the traffic emissions. The good correlations between SO and SO (R = 0.64 and 0.85 for episodes I and II) and local wind directions Park et al.Aerosol and Air Quality ResearchTable 2. Correlation coefficients among EC, crustal elements, and metals. (a) Episode I Component EC Ca Fe Ti Cr Mn Cu Zn Sb Ba Pb EC - 0.44 0.64 0.56 0.58 0.36 0.51 0.62 0.51 0.33 0.74 0.44 - 0.76 0.91 0.80 0.70 0.63 0.54 0.82 0.67 0.71 0.

64 0.76 - 0.81 0.87 0.84 0.61 0.82 0.84
64 0.76 - 0.81 0.87 0.84 0.61 0.82 0.84 0.54 0.74 0.51 0.82 0.84 0.79 0.80 0.73 0.58 0.65 - 0.51 0.64 0.33 0.67 0.54 0.46 0.82 0.64 0.90 0.45 0.51 - 0.66 (b) Episode II Component EC Ca Fe Ti Cr Mn Cu Zn Cd Sb Ba Pb EC - 0.52 0.63 0.62 0.65 0.64 0.55 0.72 0.35 0.66 0.76 0.69 0.52 - 0.74 0.92 0.88 0.81 0.46 0.74 0.15 0.93 0.81 0.68 0.63 0.74 - 0.88 0.91 0.97 0.50 0.85 0.64 0.80 0.89 0.68 0.66 0.93 0.80 0.89 0.94 0.87 0.68 0.88 0.20 - 0.91 0.87 0.76 0.81 0.89 0.90 0.94 0.91 0.60 0.88 0.35 0.91 - 0.78 All figures are significant at p . suggest that local SO emissions for the two episodes were an important source of SO formation, and that the gas-phase oxidation of SO due to high O probably played a critical role in the SO production. The time-series plot of SOthe SO-to-SO relationship, local wind directions, and the air mass trajectories indicated that regional contributions during episode I probably dominated increased SOconcentrations, while the local urban contribution dominated the SO concentration during episode II. For NO particles, the daytime NO concentration was often significantly elevated over background levels during episode I, which suggests that these particles can be attributed to the transport of severe haze lingering over the Chinese locations upwind of the sampling site, as well as the photochemical production of nitric acid and low temperatures of 4.0–10°C. Similar to the formation processes for episode I, the elevated NOleading to high PM during episode II was attributed to gas-phase photochemical reactions during the daytime and heterogeneous hydrolysis during the nighttime. To investigate the influence of road traffic on the concentrations of elemental constituents observed in PMduring the two episodes, correlation analyses among crustal elements (Ca, Fe, and Mn) and road traffic-related metals (Sb, Cu, Zn, Ba, Mn, Cr, and Pb) were performed. Good correlations of Sb, a specific marker for brake wear, with crustal elements and other metals revealed brake wear emission to be an important source for these elements during episodes I and II, with greater road traffic influence in episode II than in episode I. These metal correlation analyses support the conclusion that the enrichment of crustal and trace elements at sites close to the roadside can be mainly attributed to vehicle wear products rather than tailpipe exhaust emissions. ACKNOWLEDGEMENTS This work was supported by project “Investigation on chemical characteristics of PM2.5 and its formation processes by areas” funded from the National Institute of Environmental Research. This work was also partially supported by the General Researcher Program through a NRF grant funded by the Korea government (MEST) (No. 2011-0007222). The authors also gratefully acknowledge the NOAA Air Resources Laboratory (ARL) for the provision of the HYSPLIT transport and dispersion model and/or READY website (http://www. arl.noaa.gov/ready.php) used in this publication. REFERENCES Amato, F., Nava, S., Lucarelli, F., Querol, X., Alastuey, A., Baldasano, J.M. and Pandolfi, M. (2010). Comprehensive Assessment of PM Emissions from Paved Roads: Real-world Emission Factors and Intense Street Cleaning Trials. Sci. Total Environ. 408: 4309–4318. Andreae, M.O. (1983). Soot Carbon and Excess Fine Potassium: Long Range Transport of Combustion-Derived Aerosols. Science 220: 1148–1151. Andreae, M.O. and Merlet, P. (2001). Emission of Trace Gases and Aerosols from Biomass Burning. Global Biogeochem. Cycles 15: 955–966. Bae, M.S., Shin, J.S., Lee, K.Y. and Kim, Y.J. (2012) Characteristics of Carbonaceous Aerosols Analyzed by the Analysis of Organic Molecular Markers at Gosan Supersite, Proce. 54th Meeting of KOSAE, p. 30. Bukowiecki, N., Lienemann, P., Hill, M., Figi, R., Richard, A., Furger, M., Rickers, K., Falkenberg, G., Zhao, Y., Cliff, S.S., Prevot, A.S.H., Baltensperger, U., Buchmann, B. and Gehrig, R. (2009). Real-world Emission Factors for Antimony and Other Brake Wear

Related Trace Elements: Size-Segregated
Related Trace Elements: Size-Segregated Values for Light and Heavy Duty Vehicles. Environ. Sci. Technol. 43: 8072–8078. Chameides, W.L., Yu, H., Liu, S.C., Bergin, M., Zhou, X., Mearns, L., Wang, G., Kiang, C.S., Saylor, R.D., Luo, C., Haung, Y., Steiner, A. and Giorgi, F. (1999). Case Study of the Effects of Atmospheric Aerosols and Regional Haze on Agriculture: An Opportunity to Enhance Crop Yields in China through Emission Controls? Proc. Nat. Acad. Sci. U.S.A. 96: 13626–13633. Charlson, R.J., Lovelock, J.E., Andreae, M.O. and Warren, S.G. (1987). Oceanic Phytoplankton, Atmospheric Sulfur, Park et al.Aerosol and Air Quality ResearchCloud Albedo and Climate. Nature 326: 655–661. Chen, L.W.A., Chow, J.C., Doddridge, B.G., Dickerson, R.R., Ryan, W.F. and Mueller, P.K. (2003). Analysis of a Summertime PM and Haze Episode in the Mid-Atlantic Region. J. Air Waste Manage. Assoc. 53: 946–956. Chuang, M.T., Chou, C.C.K., Sopajareepom, K., Lin, NH., Wang, J.L., Sheu, G.R., Chang, Y.J. and Lee, C.T. (2012) Characterization of Aerosol Chemical Properties from Near-Source Biomass Burning in the Northern Indochina during 7-SEAS/Dongsha Experiment. Atmos. Environ.Press, http://dx.doi.org/10.1016/j.atmosenv.2012. 06.056. Draxler, R.R. and Rolph, G.D. (2012). HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) Model Access via NOAA ARL READY Website (http://ready. arl.noaa.gov/HYSPLIT.php), NOAA Air Resources Laboratory, Silver Spring, MD.Duan, F., Liu, X., Yu, T. and Cachier, H. (2004). Identification and Estimate of Biomass Burning Contribution to the Urban Aerosol Organic Carbon Concentrations in Beijing. Atmos. Environ. 38: 1275–1282. Echalar, F., Gaudichet, A., Cachier, H. and Artaxo, P. (1995). Aerosol Emissions by Tropical Forest and Savanna Biomass Burning: Characteristic Trace Elements and Fluxes. Geophys. Res. Lett. 22: 3034–3042. Freney, E.J., Sellegri1, K., Canonaco, F., Boulon1, J. Hervo1, M., Weigel, R., Pichon, J.M., Colomb, A., Prevot, A.S.H. and Laj, P. (2011). Seasonal Variations in Aerosol Particle Composition at the Puy-de-Dome Research Station. Atmos. Chem. Phys. Discuss. 11: 27139–27170. Furuta, N., Iijima, A., Kambe, A., Sakai, K. and Sato, K. (2005). Concentrations, Enrichment and Predominant Sources of Sb and Other Trace Elements in Size Classified Airborne Particulate Matter Collected in Tokyo from 1995 to 2004. J. Environ. Monit. 7: 1155–1161. Garg, B.D., Cadle, S.H., Mulawa, P.A., Groblicki, P.J., Laroo, C. and Parr, GA. (2000). Brake Wear Particulate Matter Emissions. Environ. Sci. Technol. 34: 44634469. Gertler, A.W., Lowenthal, D.A. and Coulombe, W.G. (1995). PM Source Apportionment Study in Bullhead City, Arizona. J. Air Waste Manage. Assoc. 45: 75–82. Gomez, D.R., Gine, M.F., Bellato, A.C.S. and Smichowski, P. (2005). Antimony: A Traffic-Related Element in the Atmosphere of Buenos Aires, Argentina. J. Environ. Monit.7: 1162–1168. Guo, S., Hu, M., Wang, Z.B., Slanina, J. and Zhao, Y.L. (2010). Size-Resolved Water-Soluble Ionic Compositions in the Summer of Beijing: Implication of Regional Secondary Formation. Atmos. Chem. Phys. 10: 947–959. Harrison, R.M., Tilling, R., Romero, M.S.C., Harrad, S. and Jarvis, K. (2003). A Study of Trace Metals and Polycyclic Aromatic Hydrocarbons in the Roadside Environment. Atmos. Environ. 37: 2391–2402. Johansson, G., Norman, M. and Burman, L. (2009). Road Traffic Emission Factors for Heavy Metals. Atmos. 43: 4681–4688. Jung, J. and Kim, Y.J. (2011). Tracking Sources of Severe Haze Episodes and Their Physicochemical and Hygroscopic Properties under Asian Continental Outflow: Long-Range Transport Pollution, Post-harvest Biomass Burning, and Asian Dust. J. Geophys. Res. 116: D02206. Jung, J., Kim, Y.J., Lee, K.Y., Cayetano, M.G., Batmunkh, T., Koo, J.H. and Kim, J. (2010). Spectral Optical Properties of Long-Range Transport Asian Dust and Pollution Aerosols over Northeast Asia in 2007 and 2008.Atmos. Chem. Phys. 10: 5391–5408. Kaneyasu, N., Ohta, S. and Murao, N. (199

5). Seasonal Variation in the Chemical C
5). Seasonal Variation in the Chemical Composition of Atmospheric Aerosols and Gaseous Species in Sapporo, Japan. Atmos. 29: 1559–1568. Kang, C.M., Lee, H.S., Kang, B.W., Lee, S.K. and Sunwoo, Y. (2004). Chemical Characteristics of Acidic Gas Pollutants and PM Species during Hazy Episodes in Seoul, South Korea. Atmos. Environ. 38: 4749–4760. Kim, D.R., Kim, J.S. and Ban, S.J. (2010). A Study on the Characteristics of Soil in the Asian Dust Source Regions of Mongolia. J. Korean Soc. Atmos. Environ. 26: 606–615 (in Korean). Kim, K.W., Kim, Y.J. and Oh, S.J. (2001). Visibility Impairment during Yellow Sand Periods in the Urban Atmosphere of Kwangju, Korea. Atmos. Environ. 35: 5157–5167. Kim, K.W., Kim, Y.J. and Bang, S.Y. (2008). Summer Time Haze Characteristics of the Urban Atmosphere of Gwangju and the Rural Atmosphere of Anmyon, Korea. Environ. Monit. Assess. 141: 189–199. Koutrakis, P., Thompson, K.M., Wolfson, J.M., Spengler, J.D., Keeler, G.J. and Slater, J.L. (1992). Determination of Aerosol Strong Acidity Losses due to Interactions of Collected Particles: Results from Laboratory and Field Studies. Atmos. Environ. 26A: 987–995. Lee, Y.L. and Sequeira, R. (2001). Visibility Degradation across Hong Kong: Its Components and Their Relative Contributions. Atmos. Environ. 34: 5861–5872. Lim, H.J. and Turpin, B.J. (2002). Origins of Primary and Secondary Organic Aerosol in Atlanta: Results of Time-Resolved Measurements during the Atlanta Supersite Experiment. Environ. Sci. Technol. 36: 4489–4496. Malm, W.C., Molenar, J.V., Eldred, R.A. and Sisler, J.F. (1996). Examining the Relationship among Atmospheric Aerosols and Light Scattering and Extinction in the Grand Canyon Area. 101: 19251–19265. Menon, S., Hansen, J., Nazarenko, L. and Luo, Y. (2002). Climate Effects of Black Carbon Aerosols in China and India. Science 297: 2250–2253. Morgan, W.T., Allan, J.D., Bower, K.N., Highwood, E.J., Liu, D., McMeeking, G.R., Northway, M.J., Williams, P.I., Krejci, R. and Coe, H. (2010). Airborne Measurements of the Spatial Distribution of Aerosol Chemical Composition across Europe and Evolution of the Organic Fraction. Atmos. Chem. Phys. 10: 4065–4083. National Institute of Occupational Safety and Health (NIOSH) (1996). Method 5040 Issue 1: Elemental Carbon (Diesel Exhaust), NIOSH Manual of Analytical Methods, 4th ed., Cincinnati, OH. Nie, W., Wang, T., Gao, X., Pathak, R.K., Wang, X., Gao, R., Zhang, Q., Yang, L. and Wang, W. (2010). Comparison among Filter-Based, Impactor-Based and Continuous Techniques for Measuring Atmospheric Fine Sulfate and Park et al.Aerosol and Air Quality ResearchAtmos. Environ. 44: 4396–4403. Oanh, N.T.K. and Leelasakultum, K. (2011). Analysis of Meteorology and Emission in Haze Episode Prevalence over Mountain-Bounded Region for Early Warning, Sci. Total Environ. 409: 2261–2271. Okada, K., Ikegami, M., Zaizen, Y., Makino, Y., Jensen, J.B. and Gras, J.L. (2001). The Mixture State of Individual Aerosol Particles in the 1997 Indonesian Haze Episode. J. Aerosol Sci. 32: 1269–1279. Park, S.S., Bae, M.S., Schauer, J.J., Ryu S.Y., Kim, Y.J., Cho, S.Y. and Kim, S.J. (2005a). Evaluation of the TMO and TOT Methods for OC and EC Measurements and Their Characteristics in PM at an Urban Site of Korea during ACE-Asia. 39: 5101–5112 Park, S.S., Harrison, D., Pancras, P. and Ondov, J.M. (2005b). Time Resolved Elemental and Organic Carbon Measurements at the Baltimore Supersite in 2002. Geophys. Res. 110: D07S06, doi: 10.1029/2004JD004610. Park, S.S., Ondov. J.M., Harrison, D. and Nair, N.P. (2005c). Seasonal and Shorter-term Variations in Particulate Atmospheric Nitrate in Baltimore. Atmos. Environ. 39: 2011–2020. Park, S.S., Kleissl, J., Harrison, D., Kumar, V., Nair, N.P., Adam, M., Ondov, J. and Parlange, M. (2006a) Characteristics of PM Episodes Revealed by Semi-Continuous Measurements at the Baltimore Supersite at Ponca St. Aerosol Sci. Technol. 40:845–860. Park, S.S., Bae, M.S., Schauer, J.J., Kim, Y.J., Cho, S.Y. and Kim

, S.J. (2006b). Molecular Composition of
, S.J. (2006b). Molecular Composition of PMOrganic Aerosol Measured at an Urban Site of Korea during the ACE-Asia Campaign. Atmos. Environ. 40: 4182–4198. Park, S.S., Cho, S.Y. and Kim, S.J. (2010). Chemical Characteristics of Water-Soluble Components in Fine Particulate Matter at a Gwangju Area. Korean J. Chem. 48: 20–26 (in Korean). Park, S.S. and Cho, S.Y. (2011). Tracking Sources and Behaviors of Water-Soluble Organic Carbon in Fine Particulate Matter Measured at an Urban Site in Korea. Atmos. Environ. 45: 60–72. Park, S.S., Kim, J.H. and Jeong, J.U. (2012). Abundance and Sources of Hydrophilic and Hydrophobic Water-Soluble Organic Carbon at an Urban Site in Korea in Summer. J. Environ. Monit. 14: 224–232. Polidori, A., Turpin, B.J., Lim, H.J., Cabada, J.C., Subramanian, R., Pandis, S.N. and Robinson, A.L. (2006). Local and Regional Secondary Organic Aerosol: Insights from a Year of Semi-Continuous Carbon Measurements at Pittsburgh. Aerosol Sci. Technol. 40: 861–872. Ram, K. and Sarin, M.M. (2010). Spatio-Temporal Variability in Atmospheric Abundances of EC, OC and WSOC over Northern India. J. Aerosol Sci. 41: 88–98. Ram, K. and Sarin, M.M. (2011). Day-Night Variability of EC, OC, WSOC and Inorganic Ions in Urban Environment of Indo-Gangetic Plain: Implications to Secondary Aerosol Formation. Atmos. Environ. 45: 460–468. Ram, K., Sarin, M.M. and Tripathi, S.N. (2012a). Temporal Trends in Atmospheric PM, Elemental Carbon, Organic Carbon, Water-Soluble Organic Carbon, and Optical Properties: Impact of Biomass Burning Emissions in the Indo-Gangetic Plain. Environ. Sci. Technol. 46: 686–695. Ram, K., Sarin, M.M., Sudheer, A.K. and Rengarajan, R. (2012b). Carbonaceous and Secondary Inorganic Aerosols during Wintertime Fog and Haze over Urban Sites in the Indo-Gangetic Plain. Aerosol Air Qual. Res. 12: 359–370. Reid, J.S., Koppmann, R., Eck, T.F. and Eleuterio, D.P. (2005). A Review of Biomass Burning Emissions Part II: Intensive Physical Properties of Biomass Burning Atmos. Chem. Phys. 5: 799–825. Ryu, S.Y., Kim, J.E., Kim, Y.J. and Gang, G.U. (2004). Chemical Composition of Post-Harvest Biomass Burning Aerosols in Gwangju, Korea. J. Air Waste Manage. Assoc.54: 1124–1137. Ryu, S.Y., Kwon, B.G., Kim, Y.J., Kim, H.H. and Chun, K.J. (2007). Characteristics of Biomass Burning Aerosol and Its Impact on Regional Air Quality in the Summer of 2003 at Gwangju, Korea. 84: 362–373. Saarnio, K., Aurela, M., Timonen, H., Saarikoski, S., Teinila, K., Makela, T., Sofiev, M., Koskinen, J., Aalto, P.P., Kulmala, M., Kukkonen, J. and Hillamo, R. (2010). Chemical Composition of Fine Particles in Fresh Smoke Plumes from Boreal Wild-land Fires in Europe. Sci. Total 408: 2527–2542. Sahu, L.K., Kondo, Y., Miyazaki, Y., Kuwata, M., Koike, M., Takegawa, N., Tanimoto, H., Matsueda, H., Yoon, S.C. and Kim, Y.J. (2009). Anthropogenic Aerosols Observed in Asian Continental Outflow at Jeju Island, Korea, in Spring 2005. J. Geophys. Res. 114: D03301, doi: 10.1029/2008JD010306. Schauer, J.J., Mader, B.T., Deminter, J.T., Heidemann, G., Bae, M.S., Seinfeld, J.H., Flagan, R.C., Cary, R.A., Smith, D., Huebert, B.J., Bertram, T., Howell, S., Kline, J.T., Quinn, P., Bates, T., Turpin, B., Lim, H.J., Yu, J.Z., Yang, H. and Keywood, M.D. (2003). ACE-Asia Intercomparison of a Thermal-Optical Method for the Determination of Particle-Phase Organic and Elemental Carbon. Environ. Sci. Technol. 37: 993–1001. Schauer, J.J., Lough, G.C., Shafer, M.M., Christensen, W.F., Arndt, M.F., DeMinter, J.T. and Park, J.S. (2006). Characterization of Metals Emitted from Motor Vehicles, Health Effect Institute. Schichtel, B.A., Husar, R.B., Falke, S.R. and Wilson,W.E. (2001). Haze Trends over the United States, 1980–1995. Atmos. Environ. 35: 5205–5210. Seinfeld, J.H. and Pandis, S.N. (2006). Atmospheric Chemistry and Physics: From Air Pollution to Climate Change Ed., John Wiley & Sons, Inc. Simoneit, B.R.T., Kobayashi, M., Mochida, M., Kawamura, K., Lee, M., Lim, H.J., Turpin, B.

J. and Komazaki, Y. (2004). Composition
J. and Komazaki, Y. (2004). Composition and Major Sources of Organic Compounds of Aerosol Particulate Matter Sampled during the ACE-Asia Campaign. J. Geophys. Res. 109: D19S10. Smith, N., Plane, J.M.C., Nien, C. and Solomon, P.A. (1995). Nighttime Radical Chemistry in the San Joanquin Atmos. Environ.29: 2887–2897. Solomon, P.A. and Sioutas, C. (2008). Continuous and Semicontinuous Monitoring Techniques for Particulate Park et al.Aerosol and Air Quality ResearchMatter Mass and Chemical Components: A Synthesis of Findings from EPA’s Particulate Matter Supersites Program and Related Studies. J. Air Waste Manage. Assoc.58: 164–195. Sternbeck, J., Sjodin, A. and Andreasson, K. (2002). Metal Emissions from Road Traffic and the Influence of Resuspension – Results from Two Tunnel Studies. Atmos. 36: 4735–4744. Sun, Y.L., Zhuang, G.S., Tang, A.H., Wang, Y. and An, Z.S. (2006). Chemical Characteristics of PM and PMin Haze–Fog Episodes in Beijing. Environ. Sci. Technol.40: 3148–3155. Thorpe, A. and Harrison, R. M. (2008). Sources and Properties of Non-Exhaust Particulate Matter from Road Traffic: A Review. Sci. Total Environ. 400: 270–283. Turn, S.Q., Jenkins, B.M., Chow, J.C. and Pritchett, L.C. (1997). Elemental Characterization of Particulate Matter Emitted from Biomass Burning: Wind Tunnel Derived Source Profiles for Herbaceous and Wood Fuels. Geophys. Res. 102(D3): 3683–3699. Turpin, B.J., Saxena, P. and Andrews, E. (2000). Measuring and Simulating Particulate Organics in the Atmosphere: Problems and Prospects. Atmos. Environ.34: 2983–3013. Turpin, B.J. and Lim, H.J. (2001). Species Contributions to Mass Concentrations: Revisiting Common Assumptions for Estimating Organic Mass. Aerosol Sci. 35: 602–610. Watson, J.G. (2002). Visibility: Science and Regulation. Air Waste Manage. Assoc. 52: 628–713. Watson, J.G. and Chow, J.C. (2002). A Wintertime PMEpisode at the Fresno, CA, Supersite. Atmos. Environ. 36: 465–475. Weber, R., Bergin, M., Kiang, C.S., Chameides, W., Orsini, D., St, J.J., Chang, M., Bergin, M., Carrico, C., Lee, Y.N., Dasgupta, P., Slanina, J., Turpin, B., Edgerton, E., Hering, S., Allen, G. and Solomon, P. (2003). Short-Term Temporal Variation in PM Mass and Chemical Composition during the Atlanta Supersite Experiment. Air Waste Manage. Assoc. 53: 84–91. Weckwerth, G. (2001). Verification of Traffic Emitted Aerosol Components in the Ambient Air of Cologne (Germany). Atmos. Environ. 35: 5525–5536. Wittig, A.E., Takahama, S., Khlystov, A.Y., Pandis, S.N., Hering, S., Kirby, B. and Davidson, C. (2004). Semi-continuous PM Inorganic Composition Measurements during the Pittsburgh Air Quality Study. Atmos. Environ.38: 3201–3213. Wu, W.S. and Wang, T. (2007). On the Performance of a Semi-continuous PM Sulphate and Nitrate Instrument under High Loadings of Particulate and Sulphur Dioxide. Atmos. Environ.41: 5442–5451. Yadav, A.K., Kumar, K., Kasim, A., Singh, M.P., Parida, S.K. and Sharan, M. (2003). Visibility and Incidence of Respiratory Diseases during the 1998 Haze Episode in Brunei Darussalam. Pure Appl. Geophys. 160: 265–277. Zhang, X.Q. and McMurry, P.H. (1992). Evaporative Loss of Fine Particulate Nitrates during Sampling. Atmos. 26A: 3305–3312. Zhang, Y., Shao, M., Lin, Y., Luan, S., Mao, N., Chen, W. and Wang, M. (2012a). Emission Inventory of Carbonaceous Pollutants from Biomass Burning in the Pearl River Delta Region, China. Atmos. Environ., In , doi: 10.1016/j.atmosenv.2012.05.055. Zhang, R., Tao, J., Ho, K.F., Shen, Z., Wang, G., Cao, J., Liu, S., Zhang, L. and Lee, S.C. (2012b). Characterization of Atmospheric Organic and Elemental Carbon of PMin a Typical Semi-Arid Area of Northeastern China. Aerosol Air Qual. Res. 12: 792–802. Zhou, Y., Wang, T., Gao, X., Xue, L., Wang, X., Wang, Z., Gao, J., Zhang, Q. and Wang, W. (2009). Continuous Observations of Water-Soluble Ions in PM at Mount Tai (1534 m a.s.l.) in Central-Eastern China. J. Atmos. Chem. 64: 107–127. Received for review, July 12, 2012 Accepted, December