Diurnal Variation and Chemical Characteristics of Atmospheric Aerosol Particles and Their Source Fingerprints at Xiamen Bay

Tsung-Chang Li1, Chung-Yi Wu1, Wei-Hsiang Chen1, Chung-Shin Yuan1,2*, Shui-Ping Wu2, Xin-Hong Wang2, Ke Du3

1 Institute of Environmental Engineering, National Sun Yat-sen University, Kaohsiung, Taiwan 2 State Key Laboratory of Marine Environmental Science, Xiamen University, Xiamen, China 3 Institute of Urban Environment, Chinese Academy of Sciences, Xiamen, China

ABSTRACT

This study investigated the diurnal variation of mass concentration and chemical composition of atmospheric aerosol particles sampled at Xiamen Bay, located on the west bank of the Taiwan Strait. Atmospheric PM10 samples were collected at ten particulate matter (PM) sampling sites at Xiamen Bay, including five sites at the Kinmen Islands and five sites in urban Xiamen, at both daytime and nighttime during the regular and intensive sampling periods. Regular sampling was conducted to collect PM10 with high-volume samplers three times a month from April 2009 to April 2010, while intensive sampling was conducted to collect PM2.5 and PM2.5–10 with dichotomous samplers in the spring and winter of 2009 and 2010. This study further selected ten major emission sources (e.g., stone processing, power plants, soil dusts, and biomass burning) at Xiamen Bay to collect fugitive particulate samples which were then resuspended in a self-designed resuspension chamber to collect PM2.5 and PM2.5–10 with two separate dichotomous samplers for further chemical analysis. The results from PM10 sampling indicated that atmospheric aerosol particles tended to be accumulated in Xiamen Bay all year round, but especially in spring and winter. A significant diurnal variation of PM10 was observed, with higher PM10 concentrations in the daytime during the regular sampling periods. The chemical analysis results showed that the major 2– – + chemical components of PM10 were SO4 , NO3 , NH4 , OC, EC, and crustal elements (Ca, Mg, Fe, and Al), which were usually higher in the daytime than at night at Xiamen Bay. The differences were most pronounced at night, where the concentrations of most anthropogenic elements (Ni, Cu, As, and V) were higher than those in the daytime. The elemental composition of PM emitted from stone processing and the cement industry were dominated by crustal elements, particularly Ca, whereas the profile of top-soil mainly contained Al and Ca. The profiles of industrial sources were dominated by secondary inorganic aerosols and EC. Moreover, construction and road dusts contained large amounts of Fe and Al, while 2– biomass burning released large amounts of K, OC, and SO4 .

Keywords: Xiamen Bay; Atmospheric particulates; Diurnal variation; Resuspension chamber; Source fingerprints.

INTRODUCTION seven Air Quality Zones (AQZs) in Taiwan. Unlike Kinmen Islands, the stationary sources in metro Xiamen were Xiamen Bay located at the west bank of Taiwan Strait approximately twenty times higher than those in Kinmen has two major islands, Kinmen and Xiamen Islands, and Islands. Thus, this study also collected resuspended dusts surrounding coastal region as shown in Fig. 1. In recent from major industrial sources around Xiamen Bay for further years, the rapid development of economy and industry along physicochemical analysis of suspended particles. Atmospheric Xiamen Bay results in serious environmental problems, particulates with aerodynamic diameter less than 10 μm particularly poor air quality and visibility impairment. (PM10), especially the fine particle fraction (aerodynamic Ambient air quality monitoring data shows that a high diameter ≤ 2.5 μm), has been found to associate with health percentage (4.6–13.2%) of poor air quality (PSI > 100) in problem, such as mortality asthma (Anderson, et al., 1992; Kinmen Island, which is the worst air quality region among Dockery, et al., 1993; Dockery, et al., 1994). As well as the size distribution, the chemical composition of atmospheric particulates can induce health-related effects. Of particular concern is the fact that toxic trace metals, such as lead * Corresponding author. Tel.: 886-7-5252000 ext. 4409; (Pb), zinc (Zn), arsenic (As), and copper (Cu), are in the Fax: 886-7-5254409 form of fine particles with a size distribution equivalent to E-mail address: [email protected] atmospheric particulates with 1.0 μm or less in diameter

Li et al., Aerosol and Air Quality Research, 13: 596–607, 2013 597

China

Taiwan Taiwan Strait

Fig. 1. Location of ten PM10 sampling sites at Xiamen Bay.

(Dongarrà et al., 2007; Zhang, 2009; Dongarrà et al., 2010; and mobile sources including industrial and automobile Lim et al., 2010). Previous investigators suggest that trace combustion processes. However, the source profiles of local metals distribute widely throughout the lung on fine particles fugitive sources from industrial and agriculture activities such could catalyze and enhance the formation of oxidants, which as road dusts, construction area, farmland dusts, cement in turn produces tissue damage (Fernández et al., 2002). industry, stone processing industry, ceramic tile industry, Previous studies reported that the chemical composition of biomass burning were limited and incomplete. Therefore, atmospheric particulates correlates with ambient air quality, the aim of this study is to collect and chemically characterize particularly the impairment of atmospheric visibility (Yuan PM10 emitted from major fugitive sources in the study area, et al., 2002; Lee et al., 2005; Yuan et al., 2006). Our recent which can be further applied for the source apportionment studies revealed that the main composition of PM10 are of PM10 by using CMB model (Ho et al., 2003; Mugica et 2– – + secondary aerosols (SO4 , NO3 , and NH4 ), crustal elements, al., 2009). and carbons at Xiamen Bay in the seasons of winter and spring (Zhao et al., 2011; Li et al., 2012; Wu et al., 2012). EXPERIMENTAL METHODS Although several previous studies have been performed on filed measurement of atmospheric PMs and their chemical Sampling Protocol of Atmospheric Aerosol Particles composition at Xiamen Bay (Wu et al., 2010; Zhao et al., In this study, ambient air quality monitoring were 2011; Li et al., 2012; Wu et al., 2012), there is no study conducted at ten particulate matter sampling sites at Xiamen focused on source characterization in terms of elemental Bay, including five sites at Kinmen Islands (Lieyu, Jinding, content of PM fractions. The chemical composition of PMs Jinsa, Guchung, and Bortsuen) and five sites at metro Xiamen emitted from various sources presents their fingerprints and (Xiamen, Daderng, Anhai, Jinjing, and Xiangzhi) as shown source profiles which can be potentially used for CMB in Fig. 1 and Table 1. Sampling sites XM and JJ located in models. The source profiles used for previous study on CMB the downtown Xiamen and Jinjing are next to a street and receptor modeling at Xiamen Bay (Li et al., 2012) were likely to be influenced by direct emissions from vehicular mainly obtained from previous researcher’s findings of the exhausts and textile industries. Air pollutants from Zhangzhou chemical composition of PMs emitted from various stationary Harbor, Xiamen Harbor, Songyu power plant and Xiamen

598 Li et al., Aerosol and Air Quality Research, 13: 596–607, 2013

Table 1. General information of particulate matter sampling sites located at Xiamen Bay. Regions Sampling site Site Abbr. Longitude Latitude Lieyu LY 118°14'30" 24°25'50" Jinding JD 118°20'14" 24°26'53" Kinmen Jinsa JS 118°24'43" 24°29'19" Islands Guchung GC 118°26'19" 24°25'18" Bortsuen BT 118°18'45" 24°24'6" Xiamen XM 118°05'25" 24°26'08" Daderng DD 118°19'49" 24°33'33" metro Anhai AH 118°28'11" 24°43'9" Xiamen Jinjing JJ 118°36'10" 24°34'34" Xiangzhi XZ 118°44'58" 24°44'44" metropolitan area would influence the ambient air quality in high-density polyethylene (HDPE) bottles after passing at site XM during the northeastern monsoon periods, through sieves (less than 38 μm mesh size), which were then whereas air pollutants emitted from Houshi power plant homogenized manually and resuspended to the air for PM could be transported to the downwind sites. Site DD sampling and chemical analysis. The resuspended PM2.5 located at an island of Dadeng in northern Xiamen is a and PM2.5–10 samples were collected by two dichotomous tourism township with a cargo harbor. Kinmen Islands are samplers using teflon and quartz filters, respectively. This recognized as a national park and its local emission sources study collected ten major types of industrial dusts in the are well controlled. Thus, sites LY, JD, JS, GC, and BT in region surrounding Xiamen Bay and brought back to the Kinmen Islands are more likely to be influenced by Air Pollution Laboratory in the Institute of Environmental regional air pollution, especially from the upwind northern Engineering at National Sun Yat-sen University, and then and northeastern industrial areas in Jinjiang River and resuspended the dusts in a self-designed resuspension Jinjing during the northeastern monsoon seasons. chamber for sampling PM2.5 and PM2.5–10 with two separate Particulate matter sampling was conducted in the daytime dichotomous samplers (Fig. 2). After sampling, the chemical (8:00–17:00) and at nighttime (17:00–8:00), for both regular composition of atmospheric particulates, including water- and intensive sampling. Regular sampling was conducted soluble ions, metallic elements, and carbonaceous contents to collect PM10 with high-volume samplers trice a month were further analyzed (Yatkin and Bayram, 2008). This from April 2009 to April 2010, while intensive sampling study finally established the database of particulates emitted was conducted to collect fine (PM2.5) and coarse (PM2.5–10) from various industrial sources and then compare their particles with dichotomous samplers for consecutive 5 days physicochemical fingerprints at Xiamen Bay. in the spring and winter of 2009–2010. After sampling, the physicochemical properties of PM, including mass Chemical Analysis concentrations, water-soluble ionic species, metallic elements, After sampling, quartz fiber filters were temporarily stored and carbonaceous contents were further analyzed. at 4°C and then transported back to the Air Pollution Laboratory in the Institute of Environmental Engineering at Sampling of Resuspension Dusts National Sun Yat-Sen University for further conditioning, In recent years, air pollution of Xiamen Bay (including weighing, and chemical analysis. All PM10 sampling filters Kinmen and Xiamen Islands) has become one of the most collected by high-volume samplers were cut into four serious environmental problems accompanying with rapid identical parts. One part of the quartz fiber filter was analyzed urbanization and economic development. In order to for water-soluble ionic species. The filter analyzed for ionic understand the physicochemical fingerprints of particulate species was put inside a 15-mL PE bottle for each sample. matter emitted from various PM emission sources at Xiamen Distilled de-ionized water (D.I. H2O) was added into each Bay, this study selected ten major emission sources (tile kiln bottle and vibrated in an ultrasonic process for approximately stoves, power plants, soil dusts, biomass burning, and etc.) 60 mins. An ion chromatographer (DIONEX, Model DX- at Xiamen Bay to collect their emitted particulate samples 120) was used to analyze the concentration of major anions – – 2– – + + + which were initially sieved with Tyler 400 mesh (dp < 38 (F , Cl , SO4 , and NO3 ) and cations (NH4 , K , Na , μm) to obtain dusts of 5 g. As shown in Table 2, descriptions Ca2+and Mg2+). of major PM fugitive sources were selected in this study Another part of the quartz fiber filter analyzed for metallic surrounding the Xiamen Bay. Two dichotomous samplers elements was initially digested in a 20 mL mixed acid in the resuspension chamber were used to collect PM2.5 and solution (HNO3:HClO3 = 3:7) at 150–200°C for 2 hrs, and PM2.5–10 from the resuspended dusts obtained from major then diluted to 25 ml with distilled de-ionized water (D.I. fugitive PM sources. The sampling of PM in the chamber H2O). During the digestion, D.I. H2O was added to the continued for at least 30 mins to ensure adequate PM residual solution two or more times in order to eliminate the deposition on the filters. In addition to industrial sources, acid content of the digestion solution. The metallic elements fugitive sources including top-soils, coal piles, and coal ashes of the PM10, including Na, Ca, Al, Fe, Mg, K, Zn, Cr, Ti, were further tested. The soil and ash samples were stored Mn, Ba, Sr, Ni, Pb, and Cu, were then analyzed with an

Li et al., Aerosol and Air Quality Research, 13: 596–607, 2013 599

Table 2. Descriptions of major fugitive PM sources surrounding the Xiamen Bay. Sources Description Stone processing industry Dusts emitted from stone segment processing in metro Xiamen Cement industry Dusts emitted from cement production in metro Xiamen Ceramic industry Dusts emitted from ceramic production in metro Xiamen Ceramic tile industry Dusts emitted from ceramic tile production in metro Xiamen Coal pile Fugitive dusts emitted from coal piles in metro Xiamen Coal ash Fugitive dusts emitted from bottom ash piles in coal-fired power plants in metro Xiamen Farmland dusts Fugitive dusts emitted from the top-soil of farmlands in Kinmen Islands Biomass burning Ashes emitted from burning the stalk of sorghum and weeds in Kinmen Islands Road dusts Fugitive dusts emitted from paved and unpaved roads in Kinmen Islands Construction dusts Fugitive dusts emitted from construction areas in Kinmen Islands

Quality Assurance and Quality Control The quality assurance and quality control (QA/QC) for both PM sampling and chemical analysis were conducted in this study. Prior to conducting PM10 sampling, the flow rate of each PM10 sampler was carefully calibrated with an orifice calibrator. Quartz fiber filters were then carefully handled and placed on the PM10 samplers to prevent potential cracking during the sampling procedure. After sampling, aluminum foil was used to fold the quartz fiber filters, which were then temporarily stored at 4°C and transported back for further chemical analysis. The sampling and analytical procedure was similar to that described in various previous studies (Cheng and Tsai, 2000; Yuan et al., 2006; Tsai et al., 2008; Tsai et al., 2010, 2011, 2012). Both field and transportation blanks were undertaken for PM10 sampling, while reagent and filter blanks were applied for chemical analysis. The determination coefficient (R2) of the calibration curve for each chemical species was required to be higher than 0.995. Background contamination was routinely monitored Fig. 2. Schematic diagram of the resupension chamber. by using operational blanks (unexposed filters), that were proceeded simultaneously with field samples. The background interference was found to be insignificant and can inductively coupled plasma-atomic emission spectrometer, thus be ignored in this study. At least 10% of the samples ICP-AES (Perkin Elmer, Model Optima 2000DV). were analyzed by spiking with a known amount of metallic Two parts of the quartz fiber filters were further used to and ionic species to determine their recovery efficiencies. measure the carbonaceous contents of PM10. The carbonaceous contents, including elemental, organic, and Transportation Routes of Air Masses total carbons (OC, EC, and TC), were measured with an The mass concentrations measured at the stations varied elemental analyzer (Carlo Erba, Model 1108). Before with seasons and wind directions, higher values were recorded sampling, all quartz fiber filters were pre-heated to 900°C in winter (northeasterly wind) than in summer (southwesterly for 1.5 hr to expel potential carbon impurities from the wind). In order to trace the air mass, backward trajectories fiber filters. The preheating procedure could minimize the from a receptor site are commonly used to identify air background carbon in the quartz fiber filters and matrix, pollution source regions and specific sources (Man and Shih, which might cause interference with the analytical results, 2001; Kong et al., 2010; Zhu et al., 2011). The Hybrid leading to an overestimation of the carbonaceous content Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) of PM10. The elemental analyzer (EA) was operated using is a widely used model that plots the trajectory of a single the procedure of oxidation at 1020°C and that of reduction air parcel from a specific location and height above ground at 500°C, with continuous heating for 15 mins. Additionally, over a period of time. The 96-hour backward trajectories of one-eighth of the quartz fiber filters was heated in advance air parcels that arrived at Xiamen Bay on four different by hot nitrogen gas at 340–345°C for at least 30 mins to days. The level of atmospheric PM10 was affected by expel the organic carbon (OC) fraction, after which the meteorological condition, thus PM10 concentrations in spring amount of elemental carbon (EC) was determined. Another and winter was much higher than those in fall and summer. one-eighth of the quartz fiber filter was analyzed without Results obtained from backward trajectories showed that heating to determine total carbon (TC). The amount of OC the concentrations of PM10 blown from the north were was then estimated by subtracting EC from TC. generally higher than those from the south.

600 Li et al., Aerosol and Air Quality Research, 13: 596–607, 2013

RESULTS AND DISCUSSION of PM10 concentration at Xiamen Bay due to long-range transportation. Among ten sampling sites, the average Mass Concentration of PM10 and Mass Ratio of Fine and concentrations of 24-hr PM10 frequently violated the ambient Coarse Particles air quality standard of 125 μg/m3 in winter and spring. The The regular sampling of PM10 concentrations at ten highest 24-hr PM10 concentrations mostly occurred in sampling sites at Xiamen Bay was conducted from April winter and spring, while the lowest generally occurred in 2010 to April 2011. As illustrated in Fig. 3, the highest PM10 summer. The ratio of PM10 concentrations in the daytime to concentrations in the daytime were in spring (XM: 144.42 nighttime (D/N) were generally higher than unity at 3 3 μg/m ), summer (AH: 141.74 μg/m ), fall (AH: 150.08 Xiamen Bay. The average ratio of PM10 concentrations in 3 3 μg/m ), and winter (JJ: 161.31 μg/m ); and the highest PM10 the daytime to those at nighttime (D/N) were 1.30, 1.44, concentrations at nighttime were in spring (AH: 102.30 1.48, and 1.52, respectively. Additionally, the average D/N 3 3 μg/m ), summer (LY: 106.70 μg/m ), winter (AH: 161.20 ratio of PM10 concentrations were 1.54 at Kinmen Islands, μg/m3), and fall (AH: 161.87 μg/m3), respectively, at Xiamen which were generally higher than those (1.30) in metro Bay. The daytime and nighttime wind roses are illustrated Xiamen. Seven sampling sites (A1, A2, B1, B2, B3, B4 in Fig. 4. Fig. 4 illustrates the prevailing wind direction and and B5) located at the center of Xiamen Bay had relatively speed both daytime and nighttime at Xiamen Bay. It shows higher PM10 concentration than other three sampling sites that the prevailing winds were respectively blown from the (A3, A4 and A5) located outside the Xiamen Bay. Our northeastern monsoons (from fall to early spring) and the previous study concluded that a superimposition phenomenon southwestern monsoons (from late spring to summer). The was regularly observed during the air pollution episodes at prevailing winds varied frequently from the northeastern Xiamen Bay. This study further revealed that local emissions monsoons to the southwestern monsoons and vice versa from the Xiamen Bay were generally more significant than during the sampling periods of this study. As illustrated in long-rang transportation from the Northeastern Monsoons. Fig. 3 and Fig. 4, the wind directions were commonly stable, Besides, high PM10 concentrations might also result from but the PM10 concentration generally varied relatively more human activities in the daytime than those at nighttime. frequently in the daytime and at nighttime. The differences Regular PM10 sampling results showed a significant diurnal of PM10 concentration in the daytime and at nighttime variation with higher PM10 concentrations in the daytime suggested that local fugitive emissions dominated PM10 compared to those at nighttime in all seasons. However, an concentration level in summer and fall. However, in winter opposite trend was observed at site DD in spring and fall. and spring, the northeastern monsoons brought highly Results from PM sampling indicated that atmospheric polluted air masses from the far urban and industrial areas particulates had a tendency to accumulate at Xiamen Bay to the downwind sites. It thus caused a significant increase all year round. A seasonal variation of PM10 was commonly

Fig. 3. Variation of PM10 concentration during the sampling periods at Xiamen Bay.

Li et al., Aerosol and Air Quality Research, 13: 596–607, 2013 601

Day Night Day Night

(a) Spring (c) Fall

Day Night Day Night

(b) Summer (d) Winter Fig. 4. Wind roses of regular sampling periods: (a) Spring; (b) Summer; (c) Fall; (d) Winter.

observed with higher PM10 concentration occurred in spring the warm sector ahead, while coarse particles dominated in and winter. In fall, winter, and spring, the sites DD and/or the frontal passage at Xiamen Bay. In this study, high PM10 AH neighboring northern Xiamen Island generally had the concentrations were generally dominant in coarse particle highest PM10 concentrations at Xiamen Bay. High PM10 mode. The results indicated that high PM10 concentrations concentrations were always observed at the sites adjacent were mainly attributed to local upwind fugitive sources (i.e., to industrial areas along the northern coast of Xiamen Bay. industrial areas along the northern coast of Xiamen Bay) of The northern winds could transport atmospheric particles the sampling sites during the intensive sampling periods. from the upwind emission sources to the downwind sites at Particularly, the northern winds transported atmospheric Kinmen Islands, resulting in a significant increase of particles from the upwind fugitive sources to the downwind atmospheric PM10 concentrations at Kinmen Islands. During Kinmen Islands. According to previous study at Xiamen Bay, the intensive sampling periods, both PM2.5 and PM2.5–10 were the local emissions at Xiamen Bay were as important as long- simultaneously sampled at site JD and JS with dichotomous range transportation from the Northeastern Monsoons, and samplers. The mass ratios of fine and coarse particles are thus a superimposition phenomenon was regularly observed presented in Table 3. During the intensive sampling period during the air pollution episodes at Xiamen Bay (Li et al., in winter, the highest PM10 and PM2.5 concentration were 2012). As shown in Table 3, both high PM2.5/PM10 ratios 3 182.77 and 107.30 μg/m in the daytime, and were 207.12 and high PM2.5 concentrations were generally observed at and 165.19 μg/m3 at nighttime, respectively. The mass ratios high PM episodes whenas the winds were prevailingly blown of PM2.5 to PM10 (PM2.5/PM10) ranged from 26.32 to 77.22% from the northeast. This phenomenon was probably caused and 29.47 to 79.75% in the daytime and at nighttime, by long-rang transportation from the northern China. Besides, respectively. Generally speaking, fine particles dominated in low PM2.5/PM10 ratios and high PM2.5–10 concentrations

Table 3. Mass ratio of fine and coarse particles during the intensive sampling periods. Sampling date Jinding (JD) Jinsa (JS) and period PM2.5–10 PM2.5 PM10 PM2.5/PM10 PM2.5–10 PM2.5 PM10 PM2.5/PM10 Day 19.16 65.00 84.16 77.23% 21.45 93.93 115.38 81.41% Dec. 20 Night 96.87 53.53 150.41 35.59% 41.94 165.19 207.12 79.75% Day 86.21 71.83 158.04 45.45% 75.47 107.30 182.77 58.71% Dec. 21 Night 55.43 68.56 123.99 55.30% 83.61 56.13 139.74 40.17% Day 89.63 42.92 132.56 32.38% 95.21 55.83 151.04 36.96% Dec. 22 Night 64.83 49.90 114.73 43.49% 113.82 47.55 161.37 29.47% Day 119.08 24.25 143.33 16.92% 88.93 44.11 133.04 33.16% Dec. 23 Night 129.01 73.06 202.07 36.15% 131.40 68.29 199.69 34.20% Day 118.76 53.30 172.06 30.98% 89.13 31.84 120.96 26.32% Dec. 24 Night 33.61 18.23 51.84 35.17% 28.49 17.62 46.12 38.22%

602 Li et al., Aerosol and Air Quality Research, 13: 596–607, 2013 were generally observed as the wind directions changed Table 4. Physical characteristics of particulates emitted from dramatically at Xiamen Bay, indicating that high PM10 various sources at Xiamen Bay. concentrations were dominated by coarse particles emitted Emission sources PM2.5/PM10 (%) from local fugitive sources. This can be further proven by Stone processing industry 20.83 Table 4, the fugitive dusts except biomass burning were Cement industry 30.04 dominant by coarse particles (PM2.5–10). Ceramic industry 33.63 Ceramic tile industry 37.40 Chemical Properties of PM10 Coal 25.52 This study revealed that the concentration of ionic species Coal ash 14.12 in the daytime were generally higher than those at nighttime. 2– – + Farmland dusts 23.87 Secondary inorganic aerosols (i.e., SO4 , NO3 , and NH4 ) Biomass burning 89.83 was abundant on PM10, accounting for about 85% of total Construction 30.55 ionic species (Fig. 5), and the most possible chemical species Road dusts 29.17 of PM10 were secondary inorganic aerosols (i.e., ammonium sulfate ((NH4)2SO4) and ammonium nitrate (NH4NO3)) as well as organic carbons in this study. The percentage of total metal of Pb is the metallic marker emitted from oil burning ionic species accounted for approximately 48% of PM10 in of heavy vessels. The concentrations of Ca and Fe at metro this study. Xiamen in the rural open lands and construction sites were The concentrations of crustal elements (e.g., Ca, Mg, Fe, higher than those at other regions. Al and Ca are the main and Al) were generally higher than anthropogenic elements metallic elements in the earth’s crust and particles emitted (e.g., Zn and Pb) as shown in Fig. 6. Most of the time, the from the cement industry, and the wind-blown dusts and concentrations of metallic elements in the daytime were frictional works from construction sites where increased the generally higher than those at nighttime. However, in spring atmospheric loading of dust particles (Huang et al., 1994). In and summer, some metals (Pb, Cu, Zn, As) at nighttime addition, there are many stone processing plants located at were higher than those in daytime. In fall and winter, the the north of site JJ. Moreover, other anthropogenic metallic concentrations of metals in the daytime were generally higher elements could be also transported to Xiamen Bay from the than those at nighttime. The highest metallic concentrations southeastern coast of China through long-range transportation were appeared at site DD, suggesting that anthropogenic by the Northeastern Monsoons. The metallic concentrations metallic elements of PM10 were highly affected by the of PM10 in the daytime were generally higher than those in surrounding emission sources at site DD. Pollution emissions the nighttime. The daytime and nighttime PM10 concentration from municipal and industrial incinerators as well as heavy ratios (D/N) for Mg, K, Ca, Cr, Mn, Fe, Zn, Al, Cu, As, and vessels are very heavy around Xiamen Bay, resulting in V were in the same order of magnitude, however, the D/N high Zn and Pb emissions. High Pb emissions may came ratios of Cd, Pb, Ni, and Ti in spring and summer varied from incinerator, leaded-gasoline combustion. High Pb even higher than an order of magnitude, indicating that the concentrations could be contributed form dense heavy emission sources of PM were different in the daytime and vessels which were quite busy around Xiamen Bay. The at nighttime.

100 100 (a) Kinmen (a) Kinmen Spring Daytime Summer Daytime Nighttime ) Nighttime 3

10 ) 10 3 g/m μ g/m μ

1 1 Concentration ( Concentration Concentration ( Concentration 0.1 0.1

0.01 0.01 F- Cl- NO3- SO42- Na+ NH4+ K+ Mg2+ Ca2+ F- Cl- NO3- SO42- Na+ NH4+ K+ Mg2+ Ca2+

100 100 (a) Kinmen (a) Kinmen Fall Daytime Winter Daytime Nighttime

) Nighttime

3 10 ) 10 3 g/m μ g/m μ

1 1 Concentration ( Concentration 0.1 ( Concentration 0.1

0.01 0.01 F- Cl- NO3- SO42- Na+ NH4+ K+ Mg2+ Ca2+ F- Cl- NO3- SO42- Na+ NH4+ K+ Mg2+ Ca2+

Fig. 5. Variation of water-soluble ionic concentrations of PM10 during the sampling periods.

Li et al., Aerosol and Air Quality Research, 13: 596–607, 2013 603

10 10 KinmenSpring daytime KinmenSummer daytime nighttime nighttime ) ) 3 3 g/m g/m 1 1 μ μ

0.1 0.1 Concentration ( Concentration Concentration ( Concentration

0.01 0.01 MgKCaCrMnFeZnAlCdPbNiCuAsTiV MgK CaCrMnFeZnAlCdPbNiCuAsTiV

10 10 KinmenFall daytime KinmenWinter daytime nighttime nighttime ) ) 3 3 g/m g/m 1 1 μ μ

0.1 0.1 Concentration ( Concentration ( Concentration

0.01 0.01 MgK CaCrMnFeZnAlCdPbNiCuAsTiV Mg K Ca Cr Mn Fe Zn Al Cd Pb Ni Cu As Ti V

Fig. 6. Variation of metallic elements concentration of PM10 during the sampling periods.

The carbonaceous contents of PM10 sampled at Xiamen (PM2.5–10), while biomass burning was dominant in fine Bay are illustrated in Fig. 7. Elemental carbon (EC), which particles (PM2.5). Fig. 8 illustrates the ionic species, metallic has a chemical structure similar to impure graphite, originated elements, and carbonaceous contents of various industrial primarily from direct emissions of fuel combustion. Organic sources. carbon (OC) is emitted from primary anthropogenic sources The fingerprints of stone processing industry were Ca, Al, 2– + and secondary organic aerosols formed by chemical reactions and Fe. Cement industry was abundant of Ca, SO4 , and K . in the atmosphere. In this study, the OC concentrations of Ceramic industry was abundant of Al, K, Mg. Ceramic tile PM10 were always higher than EC for all seasons at each industry was abundant of Al, Fe, and Ca. Construction dusts – sampling site. The average OC/EC ratios of PM10 ranged were abundant of Fe, Al, and NO3 . Farmland dusts were + from 1.27 to 1.77. It indicated that OC was the major species abundant of Fe, Al, and NH4 . Road dusts were abundant of carbonaceous contents of PM10 at Xiamen Bay. of Al, Fe, and Zn. Biomass burning was abundant of K, 2– OC, and SO4 . Coal was abundant of EC, Al, and Fe. Coal – Physicochemical Fingerprints of Atmospheric Particulates ash was abundant of Al, Ca, and NO3 . Overall, the water- Emitted from Various Sources at Xiamen Bay soluble ionic components were dominated in secondary 2– – + This study revealed that the mass ratio of PM2.5 and inorganic aerosols (SO4 , NO3 , and NH4 ), but oceanic + – PM10 (PM2.5/PM10) was 89.83% for biomass burning, and materials (Na and Cl ) were also observed in the industrial PM2.5/PM10 for other industrial sources were 14.12–37.40% dusts. Metallic elements were dominant in Ca, Al, Mg, and (Table 4). The results showed that all industrial dusts except K, while the power plants and road dusts contained As, Cr, for biomass burning were dominant in coarse particles and other trace toxics. It suggested that the resuspended

20 Daytime Kinmen Nighttime )

3 15 g/m μ

Concentration ( Concentration 5

0 EC OC EC OC EC OC EC OC Spring Summer Fall Winter

Fig. 7. Variation of EC and OC of PM10 during the sampling periods.

604 Li et al., Aerosol and Air Quality Research, 13: 596–607, 2013

(a) Stone processing industry, (b) Cement industry, (c) Ceramic industry, (d) Ceramic tile industry, (e) Coal, (f) Coal ash, (g) Farmland dusts, (h) Biomass burning, (i) Road dusts, (j) Construction Fig. 8. Chemical fingerprints of particulates emitted from various sources at Xiamen Bay. dusts consisted of anthropogenic metallic elements emitted transportation routes as shown in Fig. 9 (Chen et al., 2012). from industrial sources adjacent to the Xiamen Bay. The Continental–marine–continental (CMC), trajectories were mass ratios of Fe and Al (Fe/Al) for ceramics and ceramic originated from the Asian continent, outflowed to the East tile industries were in the range of 0.20–0.26 which were China Sea and then turned back to the continent to reach lower than Taylor’s Fe/Al ratios of crustal materials. the Xiamen Bay. Coastal Continental (CC) represented two types of trajectories from the East China coastal regions Transportation Routes of Air Masses Observed at Kinmen that moved along the coastline to reach the Xiamen Bay. Islands These two types were basically Asian continental types Air masses arriving Kinmen islands included three typical predominantly occurred in fall, winter, and spring. Southwest

Li et al., Aerosol and Air Quality Research, 13: 596–607, 2013 605

Continental–Marine–Continental Coastal Continental

Southwest Marine

Fig. 9. Three typical transportation routs of air masses observed at Xiaman Bay with HYSPLIT model.

marine (SWM) trajectories were originated from the The comparison of PM10 concentrations in the daytime and southwest sea, hundreds of kilometers away from the Xiamen at nighttime for three different air trajectories is shown in Bay. PM10 concentrations associated with the CMC type Fig. 10. The average concentrations of 24-hr PM10 ranged trajectory were the highest in three types of trajectories. Under from 93.67 to 114.32 μg/m3 and from 76.20 to 93.55 μg/m3 this condition, large amounts of anthropogenic aerosols in the daytime and at nighttime, respectively. The average emitted from local industries and vehicles over the Asian concentrations of 24-hr PM10 were 106.55±35.10 and 93.55 continent were transported to the Xiamen Bay. The mass ± 49.65 μg/m3 in the air trajectories of CMC in the daytime concentrations associated with the SWM type trajectories and at nighttime, respectively. The average concentrations were the lowest among the continental types as they were of 24-hr PM10 were 114.32 ± 47.26 and 91.55 ± 46.38 diluted by the marine air along the coast of South China Sea. μg/m3 in the air trajectories of CC in the daytime and at

606 Li et al., Aerosol and Air Quality Research, 13: 596–607, 2013

200 Day Night )

3 150 g/m μ

100 Concentration ( Concentration 10

PM 50

0 CMC CC SWM Air Trajectories

Fig. 10. Comparison of PM10 concentrations for different air trajectories reaching the Kinmen Islands.

nighttime, respectively. The average concentrations of 24- chemical compositions of ammonium sulfate ((NH4)2SO4) 3 hr PM10 were 93.67 ± 33.78 and 76.20 ± 41.61 μg/m in the and ammonium nitrate (NH4NO3). The concentrations of air trajectories of SWM in the daytime and at nighttime, crustal elements (e.g., Ca, Mg, Fe, and Al) were generally respectively. The average PM10 concentrations in the air higher than anthropogenic elements (e.g., Zn and Pb). In trajectories of CC were generally the highest, while the lowest spring and summer, some metals (e.g., Pb, Cu, Zn, As) at occurred in the air trajectories of SWM. The results indicated nighttime were higher than those in the daytime. In fall and that the level of atmospheric PM10 was highly influenced by winter, the concentrations of metals in the daytime were meteorological condition. The PM10 concentrations observed generally higher than those at nighttime. The average in spring and winter were generally higher than those in OC/EC ratios of PM10 ranged from 1.27 to 1.77, indicating fall and summer. Results from backward trajectories showed that OC was the major species of carbonaceous contents of that the concentrations of PM10 blown from the north were PM10 at Xiamen Bay. The physicochemical fingerprints of much higher than those from the south. The air trajectories atmospheric particulates emitted from various sources at of CMC and CC at Xiamen Bay were dominated by the Xiamen Bay showed that all industrial dusts except for Northeastern Monsoons, which blew air pollutants from biomass burning were dominant in coarse particles (PM2.5–10), the southeastern coast of China to the Xiamen Bay, causing while biomass burning was dominant in fine particles (PM2.5). a significant increase in the concentration of atmospheric According to transportation routes of air masses observed at particle at Kinmen Islands. The air trajectories of SEM were Kinmen Islands, the mass concentrations of PM10 associated dominated by the Southwest Monsoons blowing from the with the SWM type trajectories were the lowest among the South China Sea, resulting in the lowest PM10 concentration continental types as they were diluted by the marine air at Kinmen Islands. along the coast of South China Sea. PM10 concentrations associated with the CMC type trajectory were the highest CONCLUSIONS in three types of trajectories. CMC trajectories were originated from the hundreds of kilometers away from the The average concentrations of 24-hr PM10 frequently Xiamen Bay. violated the ambient air quality standard of 125 μg/m3 in winter and spring at Xiamen Bay. The highest 24-hr PM10 ACKNOWLEDGMENTS concentrations mostly occurred in winter and spring, while the lowest generally occurred in summer. The ratio of PM10 This study was performed under the auspices of concentrations in the daytime to those at nighttime (D/N) were Environmental Protection Bureau of Kinmen Government. generally higher than unity in summer and fall, indicating that The authors are grateful to Professor Shui-Ping Wu and PM10 had a tendency to accumulate at Xiamen Bay in the Xin-Hong Wang in the State Key Laboratory of Marine daytime than those at nighttime. Results from PM sampling Environmental Science, Xiamen University for their indicated that atmospheric particulates had a tendency to assistance in sampling and analysis for this study. accumulate at Xiamen Bay all year round. A seasonal variation of PM10 was commonly observed with higher REFERENCES PM10 concentration occurred in spring and winter. Secondary 2– – + inorganic aerosols (i.e., SO4 , NO3 , and NH4 ) was abundant Anderson, K.R., Avol, E.L., Edwards, S.A., Shamoo, D.A., on PM10, accounting for about 85% of total ionic species Pen, R.C., Linn, W.S. and Hackney, J.D. (1992). Controlled and the percentage of total ionic species were accounting Exposures of Volunteers to Respirable Carbon and for about 48% of PM10, suggesting that the most possible Sulphuric Acid Aerosols. J. Air Waste Manage. Assoc.

Li et al., Aerosol and Air Quality Research, 13: 596–607, 2013 607

42:770–776. (2008). Effects of Sea-land Breezes on the Spatial and Chen, B., Du, K., Wang, Y., Chen, J.S., Zhao, J.P., Wang, Temporal Distribution of Gaseous Air Pollutants around K., Zhang, F.W. and Xu, L.L. (2012). Emission and the Coastal Region of Southern Taiwan. J. Environ. Eng. Transport of Carbonaceous Aerosols in Urbanized Coastal Manage. 18: 387–396. Areas in China. Aerosol Air Qual. Res. 12: 371–378. Tsai, H.H., Yuan, C.S., Hung, C.H. and Lin, Y.C. (2010). Cheng, M.T. and Tsai, Y.I. (2000). Characterization of Comparing Physicochemical Properties of Ambient Visibility and Atmospheric Aerosols in Urban, Suburban, Particulate Matter of Hot Spots in a Highly Polluted Air and Remote Areas. Sci. Total Environ. 263: 101–114. Quality Zone. Aerosol Air Qual. Res. 10: 331–344. Dockery, D.W., Pope III, C.A., Xu, X., Spengler, J.D., Ware, Tsai, H.H., Yuan, C.S. Hung, C.H. and Lin, C. (2011). J.H., Fay, M.E., Ferris, B.G. and Speizer, F.E. (1993). An Physicochemical Properties of PM2.5 and PM2.5–10 at Inland Association between Air Pollution and Mortality in Six and Offshore Sites over Southeastern Coastal Region of U.S. Cities. New Engl. J. Med. 329: 1573–1759. Taiwan Strait. Aerosol Air Qual. Res. 11: 664–678. Dockery, D.W. and Pope, C.A. (1994). Acute Respiratory Tsai, J.H., Lin, J.H., Yao, Y.C. and Chiang, H.L. (2012). Effects of Particulate Air Pollution. Auun. Rev. Publ. Size Distribution and Water Soluble Ions of Ambient Health 15: 107–132. Particulate Matter on Episode and Non-episode Days in Dongarrà, G., Mannoa, E., Varricaa, D. and Vultaggio, M. Southern Taiwan. Aerosol Air Qual. Res. 12: 263–274. (2007). Mass Levels, Crustal Component and Trace Wu, S.P., Qian, R.R., Li, T.C., Wang, X.H., Hong, H.S., Elements in PM10 in Palermo, Italy. Atmos. Environ. 41: and Yuan, C.S. (2012). Seasonal Variation for the Ratio 7977–7986. of BaP to BeP at Different Sites in Great Xiamen Bay. J. Dongarrà, G, Manno, E., Varrica, D., Lombardo, M., and Environ. Monit. 14: 1221–1230. Vultaggio, M. (2010). Study on Ambient Concentrations Yatkin, S. and Bayram, A. (2008). Determination of Major of PM10, PM10–2.5, PM2.5 and Gaseous Pollutants. Trace Natural and Anthropogenic Source Profiles for Particulate Elements and Chemical Speciation of Atmospheric Matter and Trace Elements in Izmir, Turkey. Chemosphere Particulates. Atmos. Environ. 44: 5244–5257. 71: 685–696. Fernández, A.J., Ternero, M., Barragán, F.J. and Jiménez, Yuan, C.S., Lee, C.G., Liu, S.H., Yuan, C., Yang, H.Y. and J.C. (2002). A Chemical Speciation of Trace Metals for Chen, C.T. (2002). Developing Strategies for Improving Fine Urban Particles. Atmos. Environ. 36: 773–780. Urban Visual Air Quality. Aerosol Air Qual. Res. 2: 9–22. Ho, K.F., Lee, S.C., Chow, J.C. and Watson, J.G. (2003). Yuan, C.S., Sau, C.C., Chen, M.C., Hung, M.H., Chang, Characterization of PM10 and PM2.5 Source Profiles for S.W. and Lin Y.C. (2004). Mass Concentration and Size- Fugitive Dusts in Hong Kong. Atmos. Environ. 37: 1023– resolved Chemical Composition of Atmospheric Aerosols 1032. Sampled at Pescardores Islands during Asian Dust Storm Kong, S.F., Han, B., Bai, Z.P., Chen, L., Shi, J.W. and Xu, Periods in the Years of 2001 and 2002. Terr. Atmos. Ocean. Z. (2010). Receptor Modeling of PM2.5, PM10 and TSP Sci. 15: 857–879. in Different Seasons and Long-range Transport Analysis Yuan, C.S., Lee, C.G., Liu, S.H., Chang, J.C., Yuan, C. and at a Coastal Site of Tianjin, China. Sci. Total Environ. Yang, H.Y. (2006). Correlation of Atmospheric Visibility 408: 4681–4694. with Chemical Composition of Kaohsiung Aerosols. Lee, C.G., Yuan, C.S., Chang, J.C. and Yuan, C. (2005). Atmos. Res. 82: 663–679. Effects of Aerosol Species on Atmospheric Visibility in Zhang, H. (2005). An Assessment of Heavy Metals Kaohsiung City, Taiwan. J. Air Waste Manage. Assoc. Contributed by Industry in Urban Atmosphere from 55: 1031–1041. Nanjing, China. Environ. Monit. Assess. 154: 451–458. Li, T.C, Yuan, C.S., Wu, S.P. and Wang, X.H. (2013). Zhao, J.P., Zhang, F.W., Xu, Y., Chen, J.S., Yin L.Q., Shang Physicochemical Characteristics and Source Apportionment X.S. and Xu, L.L. (2011). Chemical Characteristics of of Atmospheric Aerosol Particles in Kinmen-Xiamen Particulate Matter during a Heavy Dust Episode in a Airshed. Aerosol Air Qual. Res., In Press. Coastal City, Xiamen, 2010. Aerosol Air Qual. Res. 11: Lim, J.M., Lee, J.H., Moona, J.H., Chung, Y.S. and Kim, 299–308. K.H. (2010). Airborne PM10 and Metals from Multifarious Zhao, J.P., Zhang, F.W., Xu, Y. and Chen, J.S. (2011). Sources in an Industrial Complex Area. Atmos. Res. 96: Characterization of Water-soluble Inorganic Ions in Size- 53–64. segregated Aerosols in Coastal City, Xiamen. Atmos. Res. Man, C.K. and Shih, M.Y. (2001). Identification of Sources 99: 546–562. of PM10 Aerosols in Hong Kong by Wind Trajectory Zhu, L., Huang, X., Shi, H., Cai, X. and Song, Y. (2011). Analysis. J. Aerosol Sci. 32: 1213–1223. Transport Pathways and Potential Sources of PM10 in Mugica, V., Ortiz, E., Molina, L., Vizcaya-Ruiz, A.D., Nebot, Beijing. Atmos. Environ. 45: 594–604. A., Quintana, R., Aguilar, J. and Alcantara, E. (2009). PM Composition and Source Reconciliation in Mexico City. Atmos. Environ. 43: 5068–5074. Received for review, August 3, 2012 Tsai, H.H., Ti, T.H. Yuan, C.S., Hung, C.H. and Lin, C. Accepted, November 19, 2012