Physicochemical Properties of PM 2 . 5 and PM 2 . 5 – 10 at Inland and Offshore Sites over Southeastern Coastal Region of Taiwan Strait

This study investigates the effects of sea-land breezes (SLBs) and northeastern monsoon (NEM) on the physicochemical properties of particulate matter (PM) in the atmosphere over southeastern coastal region of Taiwan Strait. The intensive PM sampling protocol was consecutively conducted for forty-eight hours. During the sampling periods, PM2.5 and PM2.5–10 were simultaneously measured with dichotomous samplers at four sites (two inland and two at offshore sites) and PM10 was measured with beta-ray monitors at these same four sites. Strong SLBs were regularly observed in the coastal region of southern Taiwan during the SLBs periods, while significant northeastern monsoons appeared during the NEM periods. The mass ratios of PM2.5/PM10 during the NEM periods were always higher than the SLBs periods. The most abundant ionic species of PM were SO4, NO3, and NH4. The most common chemical compounds of PM in southern Taiwan were ammonium sulfate ((NH4)2SO4) and ammonium nitrate (NH4NO3). Carbon contents of PM during the NEM periods were higher than during the SLBs periods. The organic-to-elemental-carbon ratio (OC/EC) of PM2.5 ranged from 1.05 to 4.39 with an average of 2.26. The order of major metallic elements of PM2.5 in the SLBs and NEM periods is Fe > Ca > K > Al > Mg > Zn > Pb and Ca > Fe > Al > K > Mg > V > Ni, respectively, and of PM2.5–10 is Ca > K > Al > Fe > Mg > Zn and Fe > Ca > Al > K > Mg > V > Ni, respectively. This study reveals that the accumulation of PM offshore, due to land breezes, influences the tempospatial distribution of PM at the coastal region in southern Taiwan. Moreover, the nss-[SO4 ]/[Na] ratio regarded as a PM pollution index, is more suitable than the [NO3]/[Na] ratio.


INTRODUCTION
Atmospheric particulate matter (PM) is a complex mixture of elemental carbon (EC), organic carbon (OC), ammonium (NH 4 + ), nitrates (NO 3 -), sulfates (SO 4 2-), mineral trace elements, and water (Turnbull and Harrison, 2000;Lee and Kang, 2001;Lin, 2002;Hueglin et al., 2005).Among them, SO 4 2-, NO 3 -and NH 4 + are the most abundant components of atmospheric PM in the atmosphere.NO 3 -may be present in the gas phase as nitric acid vapor, however, SO 4 2-is almost exclusively found in the particle phase (Wyers and Duyzer, 1997).Organic carbon (OC) and elemental carbon (EC) particles are released mainly from incomplete combustion of carbonaceous fuels.EC is essentially a primary pollutant, emitted directly from the combustion processes, while OC has both primary and secondary origins.Primary OC is emitted in particulate form, while secondary OC is formed in the atmosphere through the gas-to-particle conversion processes of volatile organic compounds (VOCs) (Pandis et al., 1992;Pankow, 1994).
Sea-land breezes (SLBs) and northeastern monsoon (NEM) play important roles in transporting air pollutants to and from urban area at the coastline in southern Taiwan (Ding et al., 2004;Tsai et al., 2008).Several studies have analyzed air mass quality and meteorological datasets in order to interpret atmospheric aerosol levels in European and Asian cities (Hien et al., 2002;Yang, 2002;Oanh et al., 2006).Smith et al. (2001) used 3-year PM 10 data sets from three monitoring sites, in combination with relevant air trajectory and meteorological data, to identify factors influencing particulate matter levels in Greater London.Previous researches have investigated the effects of SLBs on the spatial distribution and transport of ambient air pollutants, particularly for ozone episodes (Nester, 1995;Venkatesan et al., 2002;Ding et al., 2004;Levin et al., 2005;Evtyugina et al., 2006;Tsai et al., 2008).Only a few investigations have focused on the influence of SLBs on atmospheric aerosols (Rodríguez et al., 2002;Viana et al., 2005).Their studies investigated about the origin of high PM 10 and TSP concentrations at coastal sites in eastern Spain.Atmospheric coastal dynamics can exert a significant influence on the levels and chemical composition of atmospheric PM (Viana et al., 2005).Sea breezes may penetrate deep inland and cause ozone episodes by early afternoon, resulting in higher concentration of marine aerosols in the daytime than in the nighttime (Ding et al., 2004).Sea-salt particles are important in both the chemistry and radiative transfer occurring in the lower troposphere (O'Dowd et al., 1997).Moreover, Levin et al. (2005) revealed that land breezes over a lake in the nighttime bring newly emitted particulate matter to the southeast of the Dead Sea, which is responsible for the formation of lower haze layer in the early morning.
High PM 10 episodes frequently occur over southern Taiwan, a highly industrialized region (Tsai et al., 2010).In the fall, shallow northeasterly winds prevail after a frontal passage and are diverted by the Central Mountain Ridge because of its mean altitude of about 2,600 m.Numerical results indicate that anthropogenic emissions from the north of metro Kaohsiung contribute as much as 41% of ozone to the metro Kaohsiung and 24% to the inland rural area during the northerly flow (Lin et al., 2007).After a sea breeze develops, strong onshore flows transport significant amounts of air masses with preformed ozone and/or its precursors to the inland rural areas, resulting in high ozone episodes that frequently occur over southern Taiwan during the fall season.Both numerical and field sampling studies have investigated on the air-pollution episodes in northern, central, southwestern, and southern Taiwan (Cheng, 2002;Yu and Chang, 2000;Liu et al., 2002;Tsai and Chen, 2006).These investigations mainly focused on the origin of ozone and the influence of SLBs on the tempospatial distribution of ozone during the O 3 episodes.However, no investigation has been conducted to study the effects of SLBs on the physicochemical properties of PM during the PM 10 episodes in southern Taiwan.
Previous researches have investigated about the seasonal and spatial distribution of ambient air pollutants by using field measurement data obtained from the inland stations of Taiwan Air Quality Monitoring Network (TAQMN) of Taiwan Environmental Protection Administration (TEPA).So far, offshore sampling of air pollutants has not been carried out in southern Taiwan.A previous study has revealed that PM has a tendency to be transported back and forth across the coastline of southern Taiwan during the PM 10 episodes (Tsai et al., 2010).Moreover, the accumulation of air pollutants in the offshore region due to SLBs might cause local hot spots in the highly polluted region.

Sampling Protocol
In this study, particulate matter was simultaneously collected both inland and offshore during five sampling periods: August 16-18 (I), November 2-4 (II), 2006, January 24-26 (III), March 6-8 (IV), and May 2-4 (V), 2007.Inland sampling was conducted at two sites associated with fourteen stations of TAQMN, while offshore sampling was conducted at Hsiau-Liou-Chiou Island (HLC) (approximately 14 km offshore) and on a mobile air quality monitoring boat.The sampling periods (I) (II) (V) were selected to investigate the influence of SLBs on the spatial and temporal variation of PM in the summer season, while the sampling periods (III) (IV) were selected to investigate the influences of both SLBs and NEM on the spatial and temporal variation of PM in the winter season.The location of the inland and offshore sites over southeastern coastal region of Taiwan Strait is shown in Fig. 1.
Inland PM sampling was conducted at fourteen sites, including twelve stations of TAQMN and two others selected for this particular study.The two sites during the SLBs periods were located at the National Kaohsiung University (NKU; 22.732°N, 120.285°E) and Fu-In College (FIC; 22.604°N, 120.390°E), while the three sites for the NEM period were located at the Zao-In Junior High School (ZIJ; 22.678°N, 120.295°E),Chi-Shen Junior High School (CSJ; 22.631°N, 120.287°E), and National Kaohsiung First University (NKFU; 22.758°N, 120.337°E).At these sampling sites, mobile air quality monitoring vehicles were used to simultaneously collect atmospheric PM (PM 2.5 , PM 2.5-10 , and PM 10 ) with dichotomous samplers (Anderson, Model Series 241).The sampling flow rate of the dichotomous sampler was 16.7 L/min, and the filters used in this study were 37 mm quartz filter.Offshore PM sampling was conducted at two sites located in the Taiwan Strait, including an offshore island, Hsiau-Liu-Chiou (HLC; 22.350°N, 120.368°E), approximately 14 km away from the southwest coast of Kaohsiung City, and a mobile air quality monitoring boat navigated about 4-12 kilometers away from the coastline of Kaohsiung City.Atmospheric PM mixed with marine aerosols (PM 2.5 , PM 2.5-10 , and PM 10 ) were sampled at the top of the boat with a dichotomous sampler to prevent the interferences from the exhaust gases of the boat itself as well as the seawater sprays emitted from the ocean.The routes of the boat were along the coast of southern Taiwan for offshore sampling of marine aerosols during the five sampling periods.
In order to investigate the influence of SLBs and NEM on the tempo-spatial distribution of atmospheric PM, a PM sampling protocol was consecutively conducted for fortyeight hours (from 8:00 am to 8:00 am) in each sampling period.During the sampling periods, PM 2.5 and PM 2.5-10 were collected by a dichotomous sampler at each sampling site.PM 10 was measured by dichotomous samplers at four sites (including inland and offshore sites) and -ray monitors at twelve ambient stations of TAQMN.Offshore PM sampling on the boat having traveled along the coast of southern Taiwan was conducted to collect marine aerosols by a dichotomous sampler and a -ray monitor during the sampling periods.Quartz filters of 37 mm were used to collect atmospheric aerosols for further analysis of their chemical composition.

Chemical Analysis
After sampling, quartz filters were temporarily stored at 4°C and transported back to the Air Pollution Laboratory in the Institute of Environmental Engineering at the National Sun Yat-Sen University for further conditioning, weighing, and chemical analysis.The quartz filters were initially cut into four identical parts.One part was analyzed for ionic species, two parts were used for the analysis of carbonaceous content, and the rest was used for analysis of metallic content.The filters for ionic species were put into 15-mL PE bottles.Distilled, de-ionized water was added into each bottle prior to ultrasonic process for approximately 60 min.Ion chromatography (DIONEX DX-120) was used to analyze major anions (F -, Cl -, SO 4 2-, and NO 3 -) and cations (NH 4 + , Ca 2+ , Na + , K + , and Mg 2+ ) (Yuan et al., 2004).A quarter of the filter analyzed for metals was digested in 20 mL concentrated nitric acid at 150-200°C for 2 h, and then diluted to 25 mL with distilled de-ionized water (D.I.H 2 O).For the analysis of metallic content, the filters were initially digested in a 30 mL mixed acid solution (HNO 3 :HClO 4 = 3:7).During the digestion, D.I.
H 2 O was added into the residual solution twice or more in order to eliminate the acid content of the digestion solution.The residual solution was then diluted to 25 mL using 0.5 N HNO 3 and stored in PE bottles.The metallic elements of PM included Ti, As, Fe, Cu, Mn, Ca, Mg, K, Al, Pb, Cd, Ni, Cr, Zn, and V, and were further analyzed with an inductively coupled plasma-atomic emission spectrometer, ICP-AES (Perkin Elmer, Optima 2000DV).Moreover, total and elemental carbons (TC and EC) of PM were determined with an elemental analyzer, EA (Carlo Erba EA 1108).A quarter of each filter was heated in an oxygen-free environment at 340°C for 100 min to expel OC and then fed into the elemental analyzer to obtain the EC content.Another quarter of each filter was fed directly into the elemental analyzer without pre-heating to obtain the TC content (Lin, 2002).OC could then be determined by subtracting EC from total carbons TC.

Backward Trajectory
Backward trajectories were simulated with an Air Quality Trajectory Model (version 1.1) developed by Taiwan EPA.
The assimilation method, which incorporates Barnes objective method to interpolate spatial values and the variation-kinematical model, was adopted to correct the effects of complex terrain and produce the hourly wind fields data by using 24 surface stations over southern Taiwan, including two meteorological stations from the Central Weather Bureau of Taiwan (CWBT) and 22 ambient air quality monitoring stations from the Taiwan Air Quality Monitoring Network (TAQMN).By utilizing the hourly wind fields over surface, backward trajectories were simulated for 10 h initiating at seven selected air quality monitoring stations.The trajectories were constructed using only the horizontal wind components, had a segment resolution of 1 h, and the interpolation was linear in both time and space.The initial time of backward trajectories for each monitoring station was set whenas the maximum hourly PM 10 concentration was observed.

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 PM sampling, the sampling flow rate of each PM sampler was carefully calibrated with a film flowmeter (MCH-01 SENSIDYNE Inc.).Both field and transportation blanks were conducted for PM sampling, while reagent and filter blanks were undertaken for chemical analysis.The determination coefficient (R 2 ) of the calibration curve for each chemical species was required to be higher than 0.995.

Meteorological Data
In this study, we observed the diurnal variation of meteorological parameters around the coastal region of southern Taiwan during the intensive sampling periods (Table 1).Three sampling periods (I) and (V) were influenced by SLBs, while two other periods (III) and (IV) were dominated by NEM.During the sampling period (I), sea breezes from directions of 300°-350° blew at 10:00-23:00, while land breezes (2-4 m/sec) were regularly observed at nighttime.The sampling period (II) was influenced by both SLBs and NEM blown in the directions of 300°-350° and 250°-40°, respectively.The sampling periods (III) and (IV) with typical winter weather were dominated by NEM, whereas the winds were blown in the directions of 300°-30° with wind speeds of 2-6 m/sec.Taiwan, located in the East Asian subtropical region, is predominantly influenced by the northeasterly monsoon from late fall to spring and by the southwesterly monsoon from summer to mid-fall.Sea breezes are superimposed on these monsoons when meteorological conditions are suitable for producing a daytime northwesterly or southwesterly flow.The surface flows are generally weak, but increase in speed from almost calm up to about 2 m/sec and remain steady for 5-6 hours throughout the afternoon, with wind directions between 270° and 300°.Similar results were found by Kambezidis et al. (1998).They found that the summer southwesterly flows showed stronger sea breezes, with the wind speeds increasing from < 2 to 4 m/sec and the directions of southwesterly flow confined to the sector 200°-270° (Kambezidis et al., 1998).Air temperature contrasts between the land and the sea drive sea breezes when the land is strongly heated under cloudless skies and calm conditions (Byun et al., 2007).Strong land breezes were regularly observed at nighttime during the monitoring periods (I), (II), and (V), respectively.However, no significant SLBs appeared in the coastal region of southern Taiwan during the sampling periods (III) and (IV).

PM Concentration
The concentration of PM 2.5 , PM 2.5-10 , and PM 10 collected at the seven selected sampling sites is summarized in Table 1.Atmospheric PM levels at the inland sites (NKU and FIC) were generally higher than those at the offshore sites (OSB and HLC) since industrial and mobile sources in the inland region were the major particle emission sources during the SLBs periods (Table 2).The daytime concentrations of PM 2.5 and PM 2.5-10 at the inland sites were higher than those at nighttime except for the OSB sites.Moreover, the averaged mass ratios of PM 2.5 to PM 10 (PM 2.5 /PM 10 ) at the inland sites were higher than those at the offshore sites.These results concurred with previous researches, indicating that PM concentration is relatively lower during the land breeze periods than during the sea breeze periods (Tsai et al., 2008).The highest PM 10 level up to 90.79 ± 41.12 g/m 3 was observed in daytime at NKU where it had been previously claimed, incorrectly, as having the best air quality zone in metro Kaohsiung.The PM 2.5 /PM 10 ratios observed at the inland sites ranged from 53.14% to 57.36% with an average of 55.09%, while the PM 2.5 /PM 10 ratio at the offshore sites ranged from 45.36% to 49.73% with an average of 47.38%.These results indicated that the inland sites had a higher fraction of fine particles (PM 2.5 ), whereas the offshore sites had a higher fraction of coarse particles (PM 2.5-10 ).These phenomena were attributed to the fact that marine aerosols are generally abundant in the coarse particles (Yuan et al., 2004).
Table 1 also summarizes the average and standard deviation of PM concentrations at the inland and offshore sites during the NEM periods.The averaged concentrations of PM 2.5 and PM 2.5-10 at nighttime were higher than daytime at inland sites.Among four sites, ZIJ had the highest average PM 2.5 concentration (60.99 ± 17.20 g/m 3 ) and OSB had the highest average PM 10 concentration (96.79 ± 48.34 g/m 3 ).During the NEM periods, the PM 2.5 /PM 10 ratios at the inland sites were higher than the offshore sites.The concentrations in the fall and winter seasons with NEM were higher than those in the summer with SLBs in the coastal region, as illustrated in Fig. 2.
The PM 2.5 /PM 10 ratios during the NEM periods were always higher than those during the SLBs periods.During the NEM periods, the PM 10 concentrations at the inland sites were higher than those at the offshore site.On the contrary, high PM 10 concentrations observed during the NEM periods were mainly influenced by the northerly winds, which transported PM 10 from the northern region to metro Kaohsiung.

Chemical Composition of PM
The ionic species, carbonaceous and metallic contents of PM 2.5 and PM 2.5-10 sampled in the coastal region of southern Taiwan during the sampling periods are summarized in Table 3.The most abundant chemical components of PM were SO 4 2-, NO 3 -, NH 4 + , and Cl -.It indicated that the SO 4 2-concentration during the NEM periods was higher than that during the SLBs periods.These results strongly suggest that SO 4 2-coming from both northern and local anthropogenic pollution sources was still the major contributor to atmospheric PM except that its contribution during the SLBs periods was much less than the NEM periods.The mass concentration of PM 2.5 was generally higher than that of PM 2.5-10 during the PM 10 episodes in this study.The concentration of Cl -and Na + for PM 2.5-10 at offshore sites was higher than those at the inland sites.These results suggest that sea salts from oceanic spray was still the major contributor to PM 2.5-10 .The replacement of chloride from sea-salt particles is caused by the accumulation of sulfate and nitrate by atmospheric particles, and nitric acid prefers to react with sodium chloride to form stable sodium nitrate (Wall et al., 1988, Fang et al., 2000).
Excess sulfate content of PM (i.e.non-sea salt sulfate, nss-[SO 4 2-]) can be estimated by subtracting the amount of SO 4 2-in sea salts from atmospheric SO 4 2-.The amount of marine sulfate was estimated from the SO 4 2-and Na + ratio of bulk seawater.The Na + in the atmospheric PM is assumed to originate from sea salt particles only (Colbeck and Harrison, 1984).Hence, excess SO 4 2-can be derived by the following equation.

nss-[SO
Table 4 shows the ionic species, carbonaceous contents and PM correlation matrix.Within the chemical composition of PM 2.5 , nss-SO 4 2-and nitrate (NO 3 -) were strongly correlated with ammonium (NH 4 + ) during the SLBs periods.The most likely chemical compounds in metro Kaohsiung would be ammonium sulfate ((NH 4 ) 2 SO 4 ) and ammonium nitrate (NH 4 NO 3 ).Secondary PM, mainly characterized as sulfate and nitrate, are believed to be mostly responsible for the scattering of visible light causing the degradation of visibility (Yuan et al., 2006).However, the ratio of chloride to sodium for PM 2.5x10 stayed almost constant for the SLBs and NEM periods.For atmospheric PM, the concentration of crustal ions (K + and Ca 2+ ) in the NEM periods was higher than those in the SLBs periods.However, a strong northern monsoon blew crustal ions of PM from north to south.
Carbonaceous contents of PM 2.5 and PM 2.5-10 are listed in Table 2 and the ratios of OC to EC concentrations (OC/EC) are shown in Fig. 3.The OC/EC ratios of PM 2.5 were generally higher than of PM 2.5-10 during the SLBs and NEM periods.Moreover, the OC/EC ratios at the inland sites were much higher than at the offshore sites during the SLBs and NEM periods.Carbonaceous contents of atmospheric PM at daytime during the SLBs periods higher than during the NEM periods, while an opposite trend was observed in nighttime.The OC/EC ratios of PM 2.5 ranged from 1.05 to 3.56 with an average of 2.26.It is interesting to note that the OC/EC ratio of PM 2.5 was 3.56 at inland sites during NEM periods, which is close to that in gasoline vehicle emissions (3.2) (Turpin and Huntzicker, 1995).The OC/EC ratio higher than a threshold of 2.0-2.2indicates the potential formation of secondary aerosols (Chow et al., 1994;Turpin and Huntizicker, 1995).Cao et al. (2009) reported that annual mean concentrations of OC and EC in Hangzhou, China were 24.41 and 4.06 g/m 3 , respectively.In this study, the OC/EC ratios were lower than those measured in Xi'an (3.2) (Cao et al., 2005) and Hangzhou (6.0) (Cao et al., 2009), China, however, they were higher than those in Beijing, China (2.0) (Zhang et al., 2009) and in Seoul, Korea (1.3) (Park et al., 2002).Particularly, the average OC/EC ratios in daytime were generally higher than the threshold.The results suggest that secondary aerosols could be commonly formed during daytime in southern Taiwan.Frequent clear days and high solar intensity in southern Taiwan initiate a photochemical reaction during the NEM periods.Favorable meteorological conditions, coal combustion and VOCs emitted from industrial, power plants and mobile sources both in the northern and local regions resulted in higher OC/EC ratios during the NEM periods and were higher than those during the SLBs periods (Cao et al., 2003).Moreover, a strong northerly wind blew a huge amount of VOCs from the north to the south.
Fig. 4 illustrates the metallic contents of PM 2.5 and PM 2.5-10 at inland and offshore sites in southern Taiwan.The correlation matrix of metallic contents and atmospheric PM is shown in Table 5.The most abundant metallic elements in the atmospheric PM 2.5-10 were crustal elements, including Ca, Fe, and Al (Fig. 4) during the SLBs and NEM periods.Other metals, such as Mg, Pb, V, and Zn were also enriched in the atmospheric PM.Regardless of the periods (SLBs or NEM), a high concentration of Mg was consistent with ionic species (Mg 2+ ) of PM 2.5 and PM 2.5-10 at offshore sites.The most abundant metallic elements were Ca and Fe for both fine and coarse particles, respectively.An increase of Ni and V concentration is typically observed in the fuel burning sources (Hung et al., 1994;Lee et al., 1994) is due to the relative weight of this source emission during the NEM periods.The concentration of Zn and Pb in the SLBs periods was always higher than those in the NEM periods.This is true for Zn, Pb, and Fe typically associated with traffic emission (Weckwerth, 2001;Sternbeck et al., 2002;Chellam et al., 2005).Although the distribution of metallic elements was similar to some extent, the differences between the SLBs and NEM periods were identified.The order of major metallic elements of atmospheric PM 2.5 during the SLBs and NEM periods is Fe > Ca > K > Al > Mg > Zn > Pb and Ca > Fe > Al > K > Mg > V > Ni, respectively.The composition of the atmospheric PM 2.5-10 during the SLBs and NEM periods is Ca > K > Al > Fe > Mg > Zn and Fe > Ca > Al > K > Mg > V > Ni, respectively.
The analysis of enrichment factors (EF) relative to earth's crust composition can be used to identify the origins of elements from crustal or anthropogenic sources (Taylor &McLennan, 1995).The chemical species of field measured particulates were compared with a reference element (Al) as shown in below,    where C element is the concentration of any elements, and C refercence is the concentration of the reference element.Typically, Al, Si, Fe was chosen as the reference element, of which Al was used in this study.As shown Fig. 5, comparative analysis of EF values was however limited on PM data sets; this is because all the computation is conducted to compare the concentration ratios in all airborne particles against those ratios representing earth's crust.If the computed EF values for a given element exceeds far above unity (e.g., EF > 10), it can imply that the extent of enrichment is significant for that element compared to its crustal composition (Cao et al., 2009;Alleman et al., 2010).It was seen that several toxic metals like Pb, Ni, As, Zn, Cu, Ti, V, and Cd showed EF values above a few tens or hundreds, suggesting the possibly important role of anthropogenic sources.However in accord with the general expectation, most of crustal components showed much smaller values of < 10(Ca, Fe, Mn, Mg, Cr and K).In addition, K element was influenced not only by crustal (soil dust) but also by anthropogenic sources (biomass burning) (Cao et al., 2009).Similarly to our researches, Cao et al. (2009) also observed consistently low EF values for crustal components by conducting field measurements of PM 10 from five distinctive urban locations  in Hangzhou, China.The previous studies made in sampling sites around the other countries are also highly compatible with the results of this study (Alleman et al., 2010).

Effects of Sea-land Breeze and Northeastern Monsoon
A backward trajectory model is used to determine the transportation routes of PM before they arrived at the sampling locations.Backward trajectory modeling involves tracing the transportation of air parcel through the meteorological wind field.Prior to modeling, the meteorological data (wind velocity and direction) of the transportation of air mass were gathered from the Taiwan Environmental Protection Administration and Taiwan Central Weather Bureau.In addition, a three-dimension wind field, including wind speed and direction, was used to determine the transportation routes of suspended particles in the backward trajectory model.The backward trajectories of the air parcel transported toward the inland air quality monitoring sites around the coastal region of southern Taiwan during the SLBs and NEM periods are illustrated in Fig. 6.It showed that, during the SLBs periods, sea breezes blown in the early morning could transport the offshore PM back inland in southern Taiwan, which resulted in relatively high PM concentration at inland sites in the afternoon.On the contrary, high PM concentrations observed during the NEM periods were mainly blown by NEM which transported the cross-boundary PM from north to south.Backward trajectories also indicated that sea breezes could further blow offshore cross-boundary PM east of metro Kaohsiung.
A previous study revealed that HNO 3 prefers to react with NaCl to form stable NaNO 3 in the coarse mode (Wall et al., 1988).The regression between Na + with nss-[SO 4 2-] and NO 3 -concentrations obtained in the present study is shown in Fig. 6 PM pollutant index can differentiate PM collected at offshore and inland sites.The regression slope of [NO 3 -] to [Na + ] for PM 2.5 and PM 2.5-10 at inland sites is higher than at the offshore sites during the SLBs periods.Moreover, the regression slopes are different between the inland and offshore sites, indicating that atmospheric PM was not emitted from the same sources.

CONCLUSIONS
The effects of SLBs and the northeastern monsoon on the physicochemical properties of atmospheric aerosols over southeastern coastal region of Taiwan Strait were investigated in this study.Strong SLBs were regularly observed during the monitoring periods (I), (II), and (V).However, no significant SLBs appeared in the coastal region of southern Taiwan during the sampling periods (III) and (IV).Meteorological and model simulation results showed that PM could be transported back and forth across the coastline in the investigation region.The PM 2.5 /PM 10 ratios during the NEM periods were always higher than those during the SLBs periods.The most abundant ionic species of PM were: SO 4 2-, NO 3 -and NH 4 + .The most likely chemical compounds at Kaohsiung region would be ((NH 4 ) 2 SO 4 ) and (NH 4 NO 3 ).The carbon content of atmospheric aerosol particles during the NEM periods was higher than during the SLBs periods.The OC/EC ratio of PM 2.5 ranged from 1.05 to 4.39 with an average of 2.26.The order of major metallic elements of atmospheric PM 2.5 for the SLBs and NEM periods is Fe > Ca > K > Al > Mg > Zn > Pb and Ca > Fe > Al > K > Mg > V > Ni, respectively, and atmospheric PM 2.5-10 is Ca > K > Al > Fe > Mg > Zn and Fe > Ca > Al > K > Mg > V > Ni, respectively.The results of the nss-[SO 4 2-]/[Na + ] ratio was more suitable for a PM pollution index than was the [NO 3 -]/[Na + ] ratio.This study revealed that the accumulation of particulate matter in the near-ocean region due to sea-land breeze had a regular influence on the physicochemical properties of PM in the coastal region of southern Taiwan.The results indicated that the atmospheric PM might not be emitted from the same source during the sampling periods.

ACKNOWLEDGMENTS
This study was performed under the auspices of Environmental Protection Bureau of Kaohsiung Municipal Government.The authors are grateful to the CENPRO Technology Limited Company, the KELEE Environmental Consultant Corporation, the ENVIMAC Technology and Consultant Corporation, and the SOUTHERN TAIWAN Environmental Consultant Corporation for their assistance in sampling PM for this study.

Fig. 1 .
Fig. 1.Location of the PM sampling sites around the coastal region of southern Taiwan and the mobile air quality monitoring boat for the offshore monitoring protocol(x-and y-axis legends are UTM in kilometer).

TimeFig. 2 .
Fig. 2. Variation of PM 10 concentration in the coastal region of southern Taiwan during the sampling periods.

Fig. 3 .
Fig. 3. OC/EC ratio of PM 2.5 and PM 2.5-10 during daytime and nighttime at the sampling sites in the SLBs and NEM periods.

Fig. 4 .
Fig. 4. Metallic content and percentage of PM 2.5 and PM 2.5-10 during daytime and nighttime at the sampling sites in the SLBs and NEM periods.

Fig. 5 .
Fig. 5. Enrichment factors (EF) of PM at during the SLBs and NEM periods.

Fig. 6 .
Fig. 6.Backward trajectories of air parcel transported toward the inland sites in the coastal region of southern Taiwan during (a) SLBs and (b) NEM periods (x-and y-axis legends are UTM in meter).

Table 1 .
Meteorological parameters measured during the intensive air quality monitoring periods.

Table 2 .
The mass concentration and the ratios of different fraction of PM in the coastal region of southern Taiwan during the SLBs and NEM periods.Sampling sites for the SLBs periods ; b Sampling sites for the NEM periods. a

Table 4 .
Correlation matrix of ionic species, carbonaceous contents, PM 2.5 and PM 2.5-10 in the coastal region of southern Taiwan.

Table 5 .
Correlation matrix of metallic contents, PM 2.5 and PM 2.5-10 in the coastal region of southern Taiwan.