Influence of an Asian Dust Storm and Southeast Asian Biomass Burning on the Characteristics of Seashore Atmospheric Aerosols in Southern Taiwan

This study explores the effects of an Asian dust storm (ADS) and Southeast Asian biomass burning on the composition of atmospheric aerosols in the coastal area of southern Taiwan in spring 2010. Coarse and fine particles were collected using two manual dichotomous samplers (Dichots) that were equipped with Teflon and Quartz filters. The results reveal that the concentrations of PM10 and PM2.5 in the ADS period were about twice those before and after this. More than half of water-soluble ions in coarse particles (PM2.5–10) were Cl and Na (sea salt), while 70% of water-soluble ions of fine particles were SO4 and NH4 (secondary aerosols). The OC/EC ratios of PM2.5 and PM2.5–10 were all above 2.0. Over 98% of the metals in coarse and fine particles were crustal elements (98.1–99.1%), and over 60% of the metals in PM10 were in coarse particles (PM2.5–10). The water-soluble ion, EC, OC, and metal contents in PM2.5 all exceeded those in PM2.5–10. The ADS was associated with higher concentrations of water-soluble ions in PM2.5 and PM2.5–10 than was the non-ADS period, except for K in PM2.5–10. ADS most increased the Ca content and concentration, causing a rise in the average concentrations of Ca in coarse and fine particles by factors of 3.0 and 3.2, respectively. In the ADS period, although the OC and EC concentrations of coarse and fine particles increased, the OC and EC contents in these particles decreased. The ADS period was associated with higher metal contents in coarse and fine particles than the non-ADS period, and the increase in PM2.5–10 (more than 2%) exceeded that in PM2.5. In PM2.5–10 and PM2.5, the Ca and Ni contents increased the most, respectively. In addition, the concentrations of the water-soluble ions, EC, OC, and metals in PM2.5–10 and PM2.5 increased, but their contents decreased, while the contents of the other constituents (= total – (water-soluble ions + EC + OC + metals)) in PM2.5–10 and PM2.5 increased. It is also found that Southeast Asian biomass burning is related to the deterioration of the air quality of southern Taiwan.


INTRODUCTION
Many studies have demonstrated the effects of atmospheric suspended particulates on human health.Atmospheric particles with diameters of under 10 μm can enter human lungs by respiration, causing or worsening respiratory and cardiovascular diseases (Sardar et al., 2005;Brook, 2007;Franck et al., 2011).Additionally, studies of the in-vitro testing of atmospheric particles have shown that the components of PMs (such as water-soluble ions and polycyclic aromatic hydrocarbons) may be toxic to cells (Chen et al., 2006;Lin et al., 2008;Tsai et al., 2011).
Over 90% of aerosol in nature comes from desert/loess areas or marine areas (Pueschel, 1995).Sand and dust particles can be found throughout the world's atmosphere owing to long-distance transmission under various weather conditions (including dust storms).A dust storm is a natural climatic condition that is most commonly observed in areas that have undergone desertification (with soft soil surfaces, dryness, and without vegetation) (Tsai et al., 2008).Some climatic conditions, such as surface wind speed, vertical airflow, rain or snow, may increase the probability of dust storms.Dust particles that are lifted off the ground by wind, and have large diameters typically settle rapidly on the surface of the ground under gravity in neighboring areas; dust particles with small diameters may be lifted 1,500 to 3,000 m above the ground, and travel further (Kim et al., 2003;Tan et al., 2012).This phenomenon is commonly observed in deserts in central Eurasia, from which lifted sand and dust particles may travel eastward to Japan, Korea, the North Pacific and even 10,000 km to Hawaii.When such particles move southward, they can affect the air quality in Taiwan, Hong Kong, and the entire South China Sea region.Therefore, a vast area is affected by dust storms (Wang et al., 2011;Chuang et al., 2012).
The island of Taiwan is situated in a unique position in East Asia when observing pollution outflows from Southeast Asia and the Asian continent.The winter monsoon often drives Asian continental outflows originating from inland China to the Pacific region.On the other hand, southwesterly winds that prevail from late spring to summer (about April to August) bring abundant moisture and warm marine air to the island.Because of the unique geographical location and meteorological conditions, it has been demonstrated that the environmental quality of Taiwan can be influenced by East Asian atmospheric pollution events, such as acid deposition, dust storm, and biomass burning (Chen et al., 2004;Cheng et al., 2009;Sheu et al., 2010;Yang et al., 2012).Also, long-range transport of air pollutants may reduce visibility in air and be associated with an increase in amounts of organic toxic pollutants in regional PM.For example, Chi et al. (2008) observed that the amounts of dioxins (PCDD/Fs) and polycholorinated biphenyls (PCBs) at the northern coast of Taiwan and in Taipei city in an ADS period in 2006 were double to triple the normal amounts.
The characteristics of marine aerosols also affect regional air quality.Moreover, it is important to understand the effects of dust storms on the compositions of atmospheric particles.However, little attention has been paid to the characteristics of coastal aerosols and the effect of ADS thereon.In this study, coastal atmospheric coarse and fine particles were collected using two Dichots in the spring of 2010, when Taiwan was affected by an ADS, to investigate the characteristics of coastal aerosols and the impact of ADS on the compositions of atmospheric aerosols in the coastal areas of Taiwan.In this study, the data of water-soluble ions, carbon compositions, and metallic species in the particulate samples are quite useful to further picture and understand marine aerosols.Our findings provide crucial reference information regarding the background data of natural aerosol sources.

Instruments and Sampling Methods
Two dichotomous samplers (Graseby Andersen G241) were used simultaneously to collect samples of atmospheric coarse particles (PM 2.5-10 ) and fine particles (PM 2.5 ), at the seaside in Pintung, Taiwan (Fig. 1).These two samplers, equipped with Teflon and Quartz filters, respectively, were operated at a total flow rate of 16.7 L/min.Sampling was performed for 24 hours (from 9:00 a.m. until 9:00 a.m. the following day).The sampling periods were from March 14 to 22 and from April 12 to April 19, 2010.During these periods, Taiwan was affected by an ADS on March 21 and 22, so these two days were the ADS period; the other 17 sampling days constituted the non-ADS period.The particulate samples that were collected using the Teflon filters were analyzed to measure the amounts of eight water-soluble ions (Na + , K + , Mg 2+ , Ca 2+ , NH 4 + , Cl -, NO 3 -, and SO 4 2- ) and 18 metals (Na, Mg, Al, K, Ca, Fe, Ti, Cr, Mn, Ni, Cu, Zn, Sr, Cd, Sb, Ba, Pb, and Ag) in the PM.The PM that was collected using Quartz fiber filters was analyzed to measure the EC and OC constituents.
Before sampling, the quartz fiber filters were heated for 2.5 h at 900°C to reduce the filter or matrix background concentration which might influence the analysis.Before and after field sampling, the Teflon and quartz filters were dried for 24 h in a desiccator at 25 ± 3°C and a relative humidity of 40 ± 5%; they were then weighed on an electronic balance (A&D Co. Ltd., Japan; Model: HM-202) with a precision of 0.01 mg to determine the mass concentration.The mass concentration of particles was determined by dividing the particle mass by the sampled air volume.The sample filters that were collected from the field were stored in a refrigerator at 4°C before they were chemically analyzed to limit loss of volatile components.

Analysis of Water-Soluble Ions
Before water-soluble ions were analyzed, quarter sections of each Teflon filter with collected particles were extracted using 10 mL of ultra-pure water (specific resistance ≥ 18.3 MΩm).The water-soluble ions were extracted for 120 minutes using an ultrasonic bath.Each extraction solution was filtered using a cellulose acetate filter (pore size: 0.2 μm) and stored in a plastic vial in a refrigerator at 4°C before chemical analysis using ion chromatography (IC) (DIONEX ICS-3000).

Carbon Analysis and Quality Control
The carbon contents (elemental carbon (EC) and total carbon (TC)) of the particles that were collected by Quartz filters were analyzed using a total organic carbon analyzer (TOC-5000A; Shimadzu Corp., Japan) that was equipped with a suspended solid measuring (SSM) instrument.To make the carbon measurements, the samples were placed in a sample boat, and were then manually pushed into a 900°C burner that was filled with oxygen to ensure complete combustion.After CO 2 and H 2 O had been formed, the H 2 O was separated using a draining device, and the CO 2 content was determined using a non-dispersive infrared (NDIR) gas analyzer.Finally, data processing and calculations were conducted to determine the carbon content of the sample.One quarter of each filter was heated in an oven at 350°C for 100 min to expel the OC content and was then placed in the elemental analyzer to determine the EC content; another quarter of each filter was fed directly into the elemental analyzer without pretreatment to measure the TC content (Lin, 2002).The OC value was obtained by subtracting the EC content from the TC content.An organic analytical standard (OAS) (model B2038; Elemental Microanalysis Ltd., U.K.) that consisted of purified urea was used as the routine working standard for determining the carbon content.
The detection limit for measuring carbon content (in μg/m 3 and for a sampling duration of 24 h) was 0.014 μg/m 3 , and the recovery efficiencies of carbon measurements were 84.2-112.3%(average = 92.3%).Both field and laboratory blank samples were prepared and analyzed for each sampling and analysis.All data were blank corrected.

Metal Analysis and Quality Control
Before the particle-bound metals had been chemically analyzed, three quarters of each Teflon filter with collected particles were (microwave) digested for 25 minutes using a 10 mL 0.8 M HNO 3 solutions.The microwave process comprised two cycles; each cycle was operated at two operating powers -1% of 300 W (twice for 2.5 min per time) and 100% of 1200 W (three times for 2.5 min per time) based on the power series of 1200-300-1200-300-1200.Then, the digested solution was diluted to a volume of 25 mL using ultra-pure water (specific resistance ≥ 18.3 MΩcm) for the analysis of 18 metals (Na, Mg, Al, K, Ca, Fe, Ti, Cr, Mn, Ni, Cu, Zn, Sr, Cd, Sb, Ba, Pb, and Ag) by inductively coupled plasma-mass spectrometry (ICP-MS) (Agilent, 7500 series).The calibration was performed using multi-element (metal) standards (certified reference materials (CRMs); Spex, Metuchen, USA) in a 1% (v/v) HNO 3 solution.Every tenth sample was spiked using the liquid standards that contained known amounts of the metal elements that were being analyzed.The CRMs were also used as quality control standards.

Concentrations of Atmospheric Coarse and Fine Particles
Table 1 presents the mass concentrations and chemical compositions of PM 2.5 and PM 2.5-10 collected using Quartz and Teflon filters during Non-ADS and ADS periods.The PM 2.5 /PM 10 ratios were in the range 0.26-0.54(average = 0.41 ± 0.08), indicating that most Pingtung coastal atmospheric particles were coarse.The particle concentrations of PM 2.5 and PM 2.5-10 (collected from the two Dichot samplers that were used in simultaneous sampling) were very similar, despite the use of different filters (Quartz and Telfon).Moreover, the PM 2.5 and PM 2.5-10 concentrations were both highest during the ADS period (March 21-22).The ADS/ non-ADS concentration ratios of PM 2.5 and PM 2.5-10 were 2.12 and 2.00, respectively.The comparison of PM 2.5 , PM 10 , and PM 2.5 /PM 10 ratios between this research and several related studies is shown in Table 2.This study and the compared studies all observed that the concentrations of PM 2.5 , PM 10 , and TSP in ADS periods were significantly higher than those in Non-ADS periods (Mori et al., 2003;Chen et al., 2004;Lee et al., 2009;Hong et al., 2010;Bian et al., 2011;Li et al., 2011;Wang et al., 2011;Zhao et al., 2011;Chuang et al., 2012); however, we also found that the PM 2.5 /PM 10 value decreased in the ADS period because of the increase of dust (PM 2.5-10 ) in this study.This phenomenon is similar to those in several studies, indicating that ADS increased the regional concentrations of atmospheric particles (especially that of coarse particulate matter (PM 2.5-10 )) greatly (Chen et al., 2004;Cheng et al., 2005;Cheng et al., 2007;Cheng et al., 2009).

Concentrations of Water-Soluble Ions in Particles
Table 1 shows that water-soluble ions made up approximately 40-50% of the mass of coastal atmospheric coarse and fine particles in southern Taiwan.The ions with three highest concentrations in PM 2.5 were SO 4 2-, NH 4 + , and Na + , in that order, which accounted for 76% (in average) of total water-soluble ions (ΣIons).Around 70% of the water-soluble ions in PM 2.5 were the secondary aerosol species of SO 4 2-and NH 4 + .In contrast, the ions with highest there concentrations in PM 2.5-10 were Cl -, Na + , and NO 3 -in that order.Approximately 53% of total watersoluble ions were those of sea salt (Cl -and Na + ).For PM 2.5 , most of the A/C ratios were less than one (average = 0.86) but those for PM 2.5-10 were close to one (average = 1.01).The concentrations of all water-soluble ions in coarse and fine particles, except for K + in PM 2.5-10 , were higher in the ADS period than in the non-ADS period.Among the water-soluble ions, the Ca 2+ had the highest ADS/non-ADS concentration ratio and content ratio (3.1 and 1.4, respectively).
The mean concentrations of Ca 2+ in PM 2.5 and PM 2.5-10 in the non-ADS period were 0.10 ± 0.03 and 0.27 ± 0.15 μg/m 3 , respectively (Fig. 2).The mean concentrations of PM 2.5 and PM 2.5-10 -bound Ca 2+ increased remarkably (by 0.30 and 0.86 μg/m 3 , respectively) in the ADS period (March 21 and 22).Consequently, the ADS/non-ADS Ca 2+ concentration ratios in PM 2.5 and PM 2.5-10 were 3.0 and 3.2, respectively.These results were similar to those obtained elsewhere (Choi et al., 2001;Chen et al., 2004;Cheng et al., 2009).The A/C (summation of anion equivalents to summation of cation equivalents) ratios were calculated for ionic balance.During the ADS period, the A/C values for both coarse and fine particles were increased by an average factor of 1.1 (Table 1).This finding is similar to previous observations that more MgSO 4 , CaSO 4 , Ca(NO 3 ) 2 and Mg(NO 3 ) 2 were present in coarse and fine particles during ADS periods (Tsitouridou et al., 2003;Tsai and Chen, 2006;Zhang et al., 2011).Furthermore, the amounts of crustrelated constituents (such as MgSO 4 , CaSO 4 , Ca(NO 3 ) 2 , and Mg(NO 3 ) 2 ) in PM also noticeably increased (Zhang et al., 2003;Cao et al., 2008;Shen et al., 2009).

Carbon Constituents of Particles
During the Non-ADS and ADS periods, the EC and OC contents in PM 2.5 were 1.4 times those in PM 2.5-10 ; additionally, the average OC/EC ratios in coarse and fine particles exceeded 2.0 (Table 1), implying the possible contribution of secondary aerosols to both coarse and fine particles (Turpin et al., 1991;Turpin and Huntzicker, 1995;Castro et al., 1999).The ADS/non-ADS ratio of EC concentrations in PM 2.5 was 1.94, while the corresponding ratio of OC concentrations was 1.79.The ADS/non-ADS EC/PM 2.5 and OC/PM 2.5 ratios were 0.90 and 0.82, respectively.For PM 2.5-10 , the ADS/non-ADS EC and OC concentration ratios were 1.80 and 1.66, respectively, and the EC/PM 2.5-10 and OC/PM 2.5-10 ratios were 0.90 and 0.84, respectively.Therefore, the concentrations of OC and EC in coarse and fine particles increased during the ADS period (ADS/non-ADS = 1.8-1.9 on average) but the OC and EC contents in these particles decreased (ADS/non-ADS = 0.8-0.9 in average).These variations in OC and EC concentrations and contents may be associated with an increase in the amounts of crustal elements (such as Ca and Mg), which are two of primary species in storm dust (Yuan et al., 2006;Cao et al., 2008).

Metal Concentrations of Particles
Table 1 also lists the concentrations of 18 metals (Na, Mg, Al, K, Ca, Fe, Ti, Cr, Mn, Ni, Cu, Zn, Sr, Cd, Sb, Ba, Pb, and Ag) in PM 2.5 and PM 2.5-10 during the non-ADS and ADS periods.Among the 18 metallic species, Na had the highest concentration in both PM 2.5 and PM 2.5-10 .Na also had the greatest percentage (29.1-53.1%) of ΣMetals in both coarse and fine particles, followed by Ca (13.4-27.0%)and K (7.12-23.4%).These three metals were mostly from sea water and the crust, while Ti, Cr, Ni, Zn, Sr, Ba, Pb, Mn, Cu, Sb, Cd, and Ag were mainly anthropogenic emissions, and the sums of the percentages of these 12 metals in PM ΣMetals were 0.94-1.50%.According to Table 1, crustal elements (Na, Mg, Al, K, Ca, and Fe) accounted for over 98% (98.1-99.1%) of the total concentrations of metals in both PM 2.5 and PM 2.5-10 in the coastal area of southern Taiwan, so the contribution of anthropogenic emissions to the total metal concentrations was slight (< 2%), in spite of the influence of ADS.Additionally, the percentages of ΣMetals in PM 2.5-10 (11.6 ± 0.67% and 13.8% in the non-ADS and ADS periods, respectively) were higher than in PM 2.5 .For the non-ADS period, the ΣMetals concentrations as percentages in PM 2.5 and PM 2.5-10 (Metals/PM) were 8.47% ± 0.67% and 11.6 ± 0.67%, respectively, while those in the ADS period were 10.9% and 13.8%.The metal contents in both coarse and fine particles in the ADS period were higher than in the non-ADS period; furthermore, the metal percentage increased (by more than 2%) in PM 2.5-10 more than in PM 2.5 .The Ca content in PM 2.5-10 increased more than any other metal contents (followed by Fe and Al) in ADS, whereas the Ni content increased the most in PM 2.5 (followed by Al and Ca).In the ADS period, Na had the highest concentration (followed by Ca) in both PM 2.5 and PM 2.5-10 ; among anthropogenic metals, Zn had the highest concentration (followed by Mn).Fig. 3 shows the ratio of metal contents in non-ADS to ADS periods.The increase in Ca content in ADS was higher than that of other metal species ((C X ) A /(C X ) N-A = 1.99) ((CX)A: the concentration of X metal element in ADS period; (CX)N-A: the concentration of X metal element in non-ADS period)) for PM 2.5-10 , followed by Fe (1.45) and Al (1.36).In PM 2.5 , the metal elements whose contents increased the most in ADS were Ni,Al,and Ca (1.63,1.60,and 1.59,respectively).These results reveal that in the ADS period, the crustal element contents increased in both PM 2.5 and PM 2.5-10 .Wang et al. (2005) also found that, in atmospheric TSP that were collected in Beijing, China, Ca 2+ , Al and Ca were the species that exhibited the greatest increases in content during dust and super dust episodes.

Compositions of Particles
During the non-ADS period, sea salt (Cl -and Na + ), secondary aerosols (SO 4 2-, NO 3 -and NH 4 + ), crustal elements (K + , Mg 2+ and Ca 2+ ), EC, OC, and metals accounted for 78% and 63% of PM 2.5 and PM 2.5-10 , respectively; in the ADS period, the corresponding values were 70% and 54%, respectively (Fig. 4).Hence, the sum (percentage) of these contents was higher in PM 2.5 than in PM 2.5-10 in the atmosphere of coastal region southern Taiwan, despite the difference between the non-ADS and ADS periods.The concentrations of water-soluble ions, EC, OC, and metals in coarse and fine particles in the ADS period were all higher than in the non-ADS period (but the contents of these components decreased); the concentrations of other constituents (= total -(water-soluble ions + EC + OC + metals)) increased.The content of other constituents in PM 2.5 increased from 22% in the non-ADS period to 30% in the ADS period, and that in PM 2.5-10 increased from 37% to 46%.These results suggest that the main constituents of the sampled particles were crustal elements (such as Si, Ca, Mg, Al, K, and Fe) (Cheng et al., 2005;Yuan et al., 2006;Cao et al., 2008).

The Influence of Southeast Asian Biomass Burning
Fig. 5 presents the variations of CO, NO x , O 3 , and PM 2.5 concentrations obtained from the Hengchuen air quality monitoring station in April 10-19, 2010.It is inferred that the higher concentrations of CO, NO x , O 3 , and PM 2.5 measured by Hengchuen air quality monitoring station (located in southern Taiwan) should be influenced by the biomass burning in Southeast Asia in April 12-13, 2010.The maximum hourly concentrations of CO, NO x , O 3 , and PM 2.5 were about 2 times that of yearly average.This phenomenon might be affected by the Southeast Asian biomass burning.According to the global fire maps (Fig. 6) during April 10-20, 2010 from the satellite images of MODIS (Moderate Resolution Imaging Spectrometer), biomass burning actually occurred in Thailand, Vietnam,  and Philippines during the sampling period of this study (in Taiwan).Furthermore, as can be seen from the 5-day meteorological back trajectories of Hengchuen air quality monitoring station in April 12-13, 2010 (Fig. 7), the atmosphere of Hengchuen in April 12-13, 2010 was probably influenced by the air parcel from low altitude of Philippines.Several researches have pointed out that March and April are referred to the spring farming season in Southeast Asia, and the biomass burning from the agricultural activities usually emits lots of gaseous (such as NO x , CO, and hydrocarbons) and particulate pollutants.These pollutants not only have severe impact to the local atmospheric environment (Pochanart et al., 2001;Gadde et al., 2009), but also influence the air of South China Sea regions and  Southwest China coastal areas (Hsu et al., 2003;Deng et al., 2008), and even deteriorate the air quality of Taiwan (Lin et al., 2010;Yang et al., 2012).However, long-range transport of air pollutants are often interfered with regional pollutants and it is difficult to clarify the degree of such influence.Several studies based on mountain stations or sky balloons (at least 2 km high) have suggested that high altitude air quality of Taiwan, Hong Kong, and southeastern China may be affected by Southeast Asian biomass burning (Chan et al., 2003;Lee et al., 2011;Yang et al., 2012).In this study, the measurements were conducted from the Hengchuen air quality monitoring station (about 70 m high) in southwestern Taiwan.There are no known point emission (including biomass burning) sources around the station.Furthermore, we observed that Southeast Asian biomass burning affected the coastal atmosphere of southern Taiwan based on low altitude (70 m) measurements.This finding has not yet been addressed in literature, and is valuable because the low altitude (70 m) aerosols are more concerned for health than (high altitude) mountain aerosols.In this study, the data of water-soluble ions, carbon compositions, and metallic species in the particulate samples are quite useful to further picture and understand marine aerosols.Our findings provide crucial reference information regarding the background data of natural aerosol sources.

CONCLUSIONS
In this study, atmospheric coarse and fine particles were sampled in Pingtung coastal area during ADS and non-ADS periods in spring 2010 to investigate the impact of ADS on the compositions of coastal atmospheric particles in southern Taiwan.The results indicate that the PM 2.5 and PM 2.5-10 concentrations were highest in the ADS period (about double those before and after the ADS period), as determined using different types (Quartz and Telfon) of filters for Dichot sampling.About half of the water-soluble ions of PM 2.5-10 in the Pintung coastal atmosphere were from sea salt (Cl -and Na + ) in both non-ADS and ADS periods.Secondary aerosols (SO 4 2-and NH 4 + ) were the primary constituents of PM 2.5 , accounting for 70% of all water-soluble ions.Among all of the ions, Ca 2+ exhibited the greatest increase in concentration and content in ADS.The mean concentrations of Ca 2+ in PM 2.5 and PM 2.5-10 in the ADS period were 3.0 and 3.2 times those in the non-ADS period, respectively.The OC/EC ratios were greater than 2 for both PM 2.5 and PM 2.5-10 .Crustal elements (Na, Mg, Al, K, Ca, and Fe) were more than 98% (98.4-99.1%) of the metals in both coarse and fine particles.In the ADS period, the increase in metallic content in coarse particles was greater than in fine particles; for PM 2.5-10 , Ca exhibited the highest increase in content, followed by Fe and Al in that order, whereas in PM 2.5 , Ni exhibited the greatest increase in content, followed by Al and Ca in that order.The concentrations of water-soluble ions, EC, OC and metals in coarse and fine particles increased in the ADS period but the contents of these constituents decreased.The amounts of other constituents in coarse and fine particles also increased during the ADS period.It is also found that Southeast Asian biomass burning is related to the deterioration of the air quality of southern Taiwan.

Fig. 2 .
Fig. 2. Daily variations of Ca 2+ concentrations in ADS and non-ADS periods.

Fig. 3 .
Fig. 3. Metal contents in PM 2.5 and PM 2.5-10 in ADS and non-ADS periods.

Table 1 .
Mass concentrations and chemical compositions of PM 2.5 and PM 2.5-10 collected using Quartz and Teflon filters during Non-ADS and ADS periods.