INTER-RELATIONSHIPS AND SEASONAL VARIATIONS OF ...

INTER-RELATIONSHIPS AND SEASONAL VARIATIONS OF INORGANIC COMPONENTS OF PM10 IN A WESTERN PACIFIC COASTAL CITY SUK FUN KAN and PETER A. TANNER∗ Department of Biology and Chemistry, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong SAR, P.R. China (∗ author for correspondence, e-mail: [email protected], Tel: 2788 7840, Fax: 27887406)

(Received 23 August 2004; accepted 22 March 2005)

Abstract. Five year (1995–1999) datasets for inorganic major and trace components in PM10 from three monitoring sites in Hong Kong are presented. The concentrations of most chemical constituents show a high degree of seasonality, with high concentrations in winter but low ones in summer, resulting from the combined effects of summer monsoon scavenging and the influx of the continental air mass of the winter monsoon. High correlations exist within crustal components, as well as within sea-salt components. Non-seasalt sulphate and ammonium ion contributions to the PM10 mass are similar at all sites and are greatest in the winter months, showing the importance of non-local contributions. The concentrations of chemical components in PM10 from this study are compared with those in wet deposition in Hong Kong, as well as with PM10 results from other countries. In general, the Asian countries experience more elevated concentrations of PM10 . Keywords: aerosol, Asia, East Asian monsoon, particulate matter

1. Introduction The large population and fossil fuel consumption increases in Pacific Rim region of Asia, especially in China, have produced a rapid growth of anthropogenic emissions. The 1995 estimates for SO2 and NOx emissions in China are 25 Mt and 12 Mt, respectively (Streets and Waldhoff, 2000). The anthropogenic aerosols produced by gas-to-particle conversion are thus huge in amount and ubiquitous. Aerosols affect human health, since prolonged exposure to PM10 can result in increased mortality, morbidity and incidence of respiratory diseases (Dockery and Pope III, 1994). Long term monitoring data in the Western Pacific coastal area have seldom been presented and analyzed. Qin et al. (1997) and Cheng et al. (2000) analyzed the chemical properties of PM10 in Hong Kong from 1990–1994 and 1995–1996 respectively. A seasonal cycle was found for the major ions, with a summer minimum and a winter maximum. The sources of particulate matter in Hong Kong are partly local, and also due to emissions from the Asian continent (Man and Shih, 2001). Recently, the major components of PM10 and PM2.5 have been compared from a three-month study in Hong Kong (Ho et al., 2003). Concentrations of organic carbon Water, Air, and Soil Pollution (2005) 165: 113–130

C

Springer 2005

114

S. F. KAN AND P. A. TANNER

and elemental carbon in Hong Kong have been compared with those at other locations by Ho et al. (2002). The objective of our present study focuses on analyzing recent PM10 datasets over a five-year period from three monitoring stations in the Western Pacific coastal area: Hong Kong. We have focused upon the correlation between the inorganic chemical components of the aerosols, seasonal variations, and the relationships between components in PM10 with those of wet deposition. The comparison of our results with those from other locations is also given.

2. Setting and Meteorology Hong Kong, with land area 1076 km2 , is located at about 22◦ N 114◦ E, on the south China coast, Figure 1(a). With a population of about 7 million, many industrial activities have largely migrated to the south of China, so that the major air pollution sources are power stations and the 0.5 million vehicles (Government Printer, 1996). Most industries are small-to-medium enterprises and there is a considerable degree of intermixing of residential and industrial areas (Chan et al., 2001). To the north of Hong Kong lies the Shenzhen Special Administrative Region, which is a burgeoning region with diverse industries. The Pearl River Delta area, with the major city Guangzhou, has been identified as a regional source of pollution (Kok et al., 1997).

Figure 1. (a) Setting of Hong Kong and (b), (c) the PM10 sampling sites and other locations in Hong Kong. (PM10 sampling sites are numbered 1, 5 and 6: 1 is Central-Western, CW; 5 is Sham Shui Po, SSP; and 6 is Yuen Long, YL. Two wet deposition monitoring stations are at 3. Kwun Tong, KT and 1. CW. Weather stations are at: 4. King’s Park; and 7. Hong Kong Observatory, HKO. The other location referred to in the text is 2. City University of Hong Kong).

115

INTER-RELATIONSHIPS AND SEASONAL VARIATIONS

TABLE I Meteorological data for Hong Kong in 1999 (from HKO, 1999) Total rainfall (mm)a

Mean daily global solar radiation (MJ m−2 )b

Mean wind speed and direction (km h−1 ;◦ )c

Month

Mean temperature (mean range) (◦ C)a

January

17.3 (19.6–15.5)

4.5

10.14

24.9 (060)

February March April May June July August September October

18.7 (21.2–16.5) 20.4 (22.7–18.4) 24.3 (26.9–22.4) 24.9 (27.2–23.1) 28.9 (31.5–26.8) 29.2 (31.6–27.2) 28.3 (30.8–26.2) 27.8 (30.4–25.9) 26.2 (28.6–24.6)

Trace 23.6 176.9 177.8 197.4 203.8 892.0 365.7 38.8

13.33 9.35 14.03 13.36 16.38 15.30 12.56 13.39 14.51

24.9 (070) 25.1 (070) 27.2 (070) 25.1 (070) 20.0 (190) 20.9 (220) 19.7 (220) 21.7 (010) 28.3 (080)

November December

22.2 (24.5–20.3) 16.8 (19.1–14.8)

15.7 32.9

13.03 10.80

29.1 (070) 28.0 (010)

Measured at a Hong Kong Observatory (7 in Figure 1(c)); b King’s Park (4 in Figure 1(c)); c Waglan Island (Figure 1(b)).

The climate of Hong Kong has been described in detail elsewhere (Heywood, 1953). Meteorological data for the months of the year 1999 in Hong Kong are shown in Table I and exhibit the salient features. The rainfall occurs mainly in the summer and the short autumn period. Hong Kong is under the influence of the East Asian monsoon system, with north/north easterly winds prevalent during the winter months and south/south westerly winds prevalent in summer months. The land-sea breeze also plays a considerable role during weak synoptic conditions, due to the mountainous coastal topography in Hong Kong. Tropical cyclones occur frequently in the period from May to November over the western North Pacific and South China Sea, and about 3 or 4 make land each year in the vicinity of Hong Kong. The tropical cyclone generally transports aerosol with high concentrations of seasalt components (Tanner and Wong, 1997).

3. Data Sources and Analysis The compositions of PM10 samples were investigated by analyzing the five-year datasets (1995–1999) from the Hong Kong Environmental Protection Department (HK EPD) CD-ROM: The Air Quality in Hong Kong, and subsequent private communications. In this study, we mainly focus upon data from the 3 monitoring stations Central Western (CW: residential), Sham Shui Po (SSP: mixed residential,

116


commercial, industrial) and Yuen Long (YL: residential with rapid urban development) which span the length of the Territory. The site locations (marked 1, 5 and 6, respectively) are shown in Figures 1(b), (c). PM10 was collected for 24 hours at the monitoring stations by high volume air samplers (Andersen Samplers, Inc) which operate on a 6-day cycle. Routine calibration was performed using a certified orifice transfer standard (Air Services Laboratory, 2003). The Air Services Laboratory of the HK EPD has been accredited by the Hong Kong Laboratory Accreditation Scheme. From 1995 onwards, quartz microfibre filters (Whatman QM-A), comprising high purity silica microfibres, were used. These have low levels of alkaline earth metals so that sulphate and nitrate artifacts due to SO2 and NOx reactions are virtually eliminated (Tsai and Perng, 1998). Other artifacts arise from deliquescence, which can produce ion-replacement reactions; and the loss of volatile species such as NH4 Cl and NH4 NO3 (Appel et al., 1984; Bai et al., 1995). The filters were conditioned at 30–50% relative humidity (with a variation of less than 5%) for gravimetric measurement, before and after sampling. The inorganic components (Al, As, Ba, Br, Ca, Cd, Cl, Cr, Cu, Fe, K, Mg, Mn, Na, NH4 , Ni, NO3 , Pb, SO4 , V, Zn) were analyzed by the Hong Kong Government Laboratory (the charges of the chemical species are omitted for clarity throughout this work). Water-soluble anions were determined by ion chromatography, whereas the cations were determined by atomic spectrometric techniques following digestion with concentrated acids. Some anions, such as O2− , HCO3 − , CO3 2− , OH− and water-soluble organic anions were not included in the chemical analysis. pH measurements were performed on samples collected at City University [location 2, Figure 1(b), (c)]. Relative to the blank (water) sample (pH = 5.52), the measured mean pH values for aqueous extracts of TSP, PM10 and PM2.5 collected on various days were 6.9(N = 15), 6.4(N = 15) and 5.6(N = 2), respectively. This shows the increase in acidity when going from coarse to fine particulate matter. The non-seasalt (nss-) concentrations of components were based upon the relative abundance of the species in seawater, assuming that Na in the sample originates completely from seawater. Wet deposition samplers were deployed at rooftop level in the air monitoring stations, with no obstacles to restrict free rainfall or dust collection. The samplers collected wet deposition on a weekly basis in 286 mm diameter white high-density polyethylene buckets. The unit used a conductive rain sensor which switched a motor-driven cover to seal the collection bucket during dry conditions. Herein we compare the concentrations of inorganic components in wet deposition at Central Western with those in PM10 at the same station. The daily average surface wind speed and direction data were measured at the meteorological station at Waglan Island, Figure 1(b), and are more representative of the synoptic scale airflow over Hong Kong, thereby avoiding any local modification of this information owing to the complex topography of Hong Kong. The data of rainfall amount was extracted from the Hong Kong Observatory, labeled 7 in Figure 1(c).


117

Statistical analyses were carried out using Sigmastat version 2.01 and GraphPad Prism version 2.01. Data were tested for normality (Kolmogorov-Smirnov test) prior to use and were log-transformed if they are not normally distributed. The Pearson Product Moment correlation test was used if the data set was normally distributed; otherwise, the Spearman Rank Order test was used (Zar, 1984).

4. Results and Discussion 4.1. MEAN CONCENTRATIONS OF INORGANIC IN PM 10 AND INTER- RELATIONS

COMPONENTS

There were 50–60 samples of PM10 during each year (1995–1999) for each site. At each site, the dataset for each of the species listed in Section 3 did not pass the Kolmogorov-Smirnov test for normal distribution, within each of the 5 years. Table II shows the means, standard deviation (SD) and the number of samples (N) for inorganic components of PM10 for the period 1995–1999. The major inorganic species in PM10 comprise Cl, Na, NH4 and SO4 . Anthropogenic inorganic species contribute most to the mass of PM10 (K, NH4 , SO4 , NO3 : between 9–46%), followed by crustal species (Al, Mn, Fe, Ca and Ba: between 1–18%), then marine species (Mg, Na, Cl: 0.6–51%), and finally, trace species (As, Cd, Ni, Pb, Cr, V, Zn, Cu, and Br: 0.2–2.9%). nss-SO4 contributes totally >97% to the total concentration of SO4 , and 17% by mass to the total concentration of PM10 . Sea-salt species, Na, Cl and Mg show strong inter-correlations (R > 0.50; P < 0.0001, N > 255). The correlations between Na and Cl at CW and SSP (R = 0.75, P < 0.0001, N = 273; R = 0.74, P < 0.0001, N = 271, respectively) are stronger than that at the more inland site, YL (R = 0.57, P < 0.0001, N = 255). The crustal element, Al is highly correlated (R > 0.85, P < 0.0001, N > 255) with Fe, Mn and Ca at each of the three sites. Considering the anthropogenic species, NH4 and SO4 were found to be strongly correlated (R > 0.93, P < 0.0001, N > 255), and in addition, with a similar seasonal trend at each of the three sites. It has generally been found that SO4 or NH4 show fairly constant concentrations over wide areas (Pakkanen et al., 2001). Following Adams et al. (1999), the source of particulate ammonium is from the condensation of gaseous ammonia, whilst those of particulate SO4 are from condensation of gaseous sulfuric acid, and SO2 to sulfate conversion (Barth et al., 2000). The major source (>73%) of SO2 in Hong Kong is from power generation in the western part of the Territory (CH2M Hill, 2002). Fortunately, the prevailing winds are generally easterly. The SO2 emissions in Hong Kong are 5 times smaller than those of adjacent Guangdong County, in southern China: 0.15 Mt y−1 versus 0.72 Mt y−1 (Streets and Waldhoff, 2000). High sulphur-content fuel oils were banned in Hong Kong in 1990, and regulations were put in place for the sulphur content of automotive diesel not to exceed 0.2% by mass (April 1995) and 0.05%

118


TABLE II Mean concentration (mean, ng m−3 ), standard deviation (SD, ng m−3 ), number of samples (N), of inorganic species in PM10 at three sites in HK from 1995—1999a Mean (SD) N Species Site

1995

1996

1997

1998

1999

Al

CW 421 (586) 47 326 (262) 47 219 (132) 59 190 (180) 61 245 (232) 59 SSP 344 (353) 46 318 (260) 51 224 (183) 55 215 (226) 59 242 (206) 60 YL 572 (479) 23 403 (340) 53 291 (186) 59 256 (196) 61 282 (258) 59

As

CW 5.1 (5.9) 47 5.1 (4.5) 47 4.4 (4.5) 59 3.9 (4.2) 61 3.7 (3.1) 59 SSP 4.9 (5.9) 46 4.8 (4.5) 51 3.9 (4.6) 55 5.4 (11.9) 59 4.0 (6.4) 60 YL 9.2 (8.3) 23 7.1 (6.7) 53 5.8 (5.8) 59 5.2 (5.4) 61 4.8 (4.0) 59

Ba

CW 13.6 (6.6) 47 12.2 (4.8) 47 11.5 (6.0) 59 10.7 (5.3) 61 12.0 (6.8) 59 SSP 13.8 (5.1) 46 14.3 (8.1) 51 16.5 (8.8) 55 16.2 (9.6) 59 19.7 (11.7) 60 YL 17.0 (8.3) 23 17.0 (11.9) 53 14.8 (9.1) 59 14.1 (8.6) 61 15.5 (9.0) 59

Br

CW 13.2 (7.7) 47 14.5 (6.1) 47 13.2 (7.7) 59 11.3 (6.4) 61 9.0 (3.6) 59 SSP 10.3 (5.2) 46 12.0 (6.6) 51 11.6 (5.4) 55 9.9 (4.5) 59 9.0 (6.6) 60 YL 12.7 (8.7) 23 13.2 (7.9) 53 11.8 (7.8) 59 10.6 (6.4) 61 9.0 (3.7) 59

Ca

CW 1113 (1344) 47 944 (599) 47 713 (347) 59 660 (562) 61 820 (721) 59 SSP 920 (749) 46 891 (545) 51 697 (456) 55 761 (637) 59 928 (619) 60 YL 1549 (1334) 23 1085 (853) 53 814 (520) 59 860 (655) 61 1010 (868) 59

Cd

CW 1.2 (1.0) 47 1.3 (0.9) 47 1.6 (1.9) 59 1.3 (1.7) 61 1.4 (1.3) 59 SSP 1.1 (0.8) 46 1.2 (1.0) 51 1.1 (0.9) 55 1.2 (1.2) 59 1.6 (1.8) 60 YL 1.9 (1.3) 23 1.6 (1.5) 53 1.6 (2.0) 59 1.6 (1.9) 61 1.8 (1.7) 59

Cl

CW 2182 (1528) 47 2157 (1219) 47 1476 (1166) 59 1824 (1457) 61 1671 (1388) 59 SSP 1105 (943) 46 1144 (892) 51 1265 (1202) 55 1108 (1261) 59 1209 (1059) 60 YL 734 (822) 23 1051 (748) 53 857 (744) 59 1088 (993) 61 870 (1012) 59

Cr

CW 1.9 (1.8) 47 2.1 (1.8) 47 1.7 (1.4) 59 1.4 (1.0) 61 1.4 (0.8) 59 SSP 1.9 (1.7) 46 2.7 (1.9) 51 2.0 (1.4) 55 1.5 (1.2) 59 1.7 (1.5) 60 YL 3.0 (2.4) 23 2.5 (1.6) 53 2.1 (1.5) 59 1.9 (1.3) 61 1.8 (1.2) 59

Cu

CW 44.2 (29.5) 47 34.5 (18.8) 47 31.5 (16.7) 59 25.6 (10.7) 61 29.0 (16.0) 59 SSP 36.4 (20.4) 46 43.3 (24.6) 51 38.0 (14.2) 55 25.1 (7.6) 59 31.7 (26.7) 60 YL 47.9 (17.3) 23 32.0 (19.4) 53 28.4 (10.6) 59 32.5 (14.9) 61 38.9 (29.1) 59

Fe

CW 686 (800) 47 546 (354) 47 400 (210) 59 360 (298) 61 446 (359) 59 SSP 633 (497) 46 589 (384) 51 467 (280) 55 463 (354) 59 499 (306) 60 YL 913 (712) 23 674 (495) 53 524 (299) 59 524 (351) 61 582 (405) 59

K

CW 728 (657) 47 689 (566) 47 632 (476) 59 567 (523) 61 620 (469) 59 SSP 687 (643) 46 633 (565) 51 562 (528) 55 560 (502) 59 734 (905) 60 YL 1312 (910) 23 1047 (982) 53 808 (647) 59 776 (749) 61 792 (668) 59 (Continued on next page)

119


TABLE II (Continued ) Mean (SD) N Species Site

1995

1996

1997

1998

1999

Mg

CW 430 (468) 47 345 (146) 47 266 (117) 59 275 (140) 61 334 (215) 59 SSP 279 (183) 46 277 (143) 51 241 (120) 55 255 (167) 59 272 (143) 60 YL 280 (234) 23 232 (116) 53 191 (88) 59 204 (96) 61 243 (181) 59

Mn

CW 31.8 (34.5) 47 21.1 (14.6) 47 18.4 (11.9) 59 14.8 (12.3) 61 17.3 (11.1) 59 SSP 27.3 (30.4) 46 19.9 (14.7) 51 17.7 (12.9) 55 16.6 (14.2) 59 17.1 (12.4) 60 YL 45.0 (40.0) 23 22.2 (17.4) 53 17.5 (11.6) 59 17.6 (14.2) 61 21.6 (15.4) 59

Na

CW 2353 (956) 47 2247 (902) 47 1866 (855) 59 1940 (982) 61 2078 (1076) 59 SSP 1542 (912) 46 1695 (930) 51 1614 (955) 55 1592 (1036) 59 1574 (967) 60 YL 915 (620) 23 1185 (746) 53 1084 (705) 59 1137 (755) 61 1197 (939) 59

NH4

CW 2500 (2227) 47 2375 (1810) 47 3298 (2505) 59 2298 (2263) 61 2273 (1713) 59 SSP 2348 (1701) 46 2190 (1386) 51 2737 (1849) 55 2227 (1783) 59 2162 (1989) 60 YL 3090 (1860) 23 2961 (2195) 53 3589 (2785) 59 2873 (2645) 61 2595 (2002) 59

Ni

CW 3.4 (2.4) 47 2.9 (1.6) 47 4.6 (4.2) 59 2.7 (1.7) 61 2.8 (1.9) 59 SSP 3.8 (2.8) 46 4.4 (2.4) 51 4.8 (3.5) 55 4.3 (5.4) 59 3.5 (3.6) 60 YL 3.5 (2.2) 23 3.1 (1.8) 53 3.6 (2.6) 59 3.2 (2.2) 61 3.2 (1.8) 59

NO3

CW 3670 (2599) 47 3424 (2057) 47 3460 (2676) 59 3028 (2719) 61 3226 (2178) 59 SSP 2911 (1985) 46 2896 (1995) 51 3249 (2135) 55 2997 (1907) 59 3254 (3186) 60 YL 3742 (3641) 23 4062 (3351) 53 3960 (3501) 59 3705 (3317) 61 3502 (2428) 59

Pb

CW 70.8 (60.9) 47 69.9 (57.9) 47 60.4 (55.9) 59 51.7 (57.8) 61 58.7 (57.1) 59 SSP 73.7 (59.5) 46 69.3 (57.5) 51 52.5 (43.0) 55 54.4 (53.6) 59 71.6 (102) 60 YL 119 (72.8) 23 111 (111) 53 77.5 (64.0) 59 74.9 (78.0) 61 80.8 (80.1) 59

SO4 †

CW 10.7 (5.1) 47 9.8 (5.4) 47 11.6 (6.6) 59 8.8 (5.5) 61 9.5 (5.0) 59 SSP 9.7 (5.1) 46 9.4 (4.6) 51 9.7 (5.1) 55 8.8 (5.5) 59 9.0 (6.0) 60 YL 11.5 (5.1) 23 9.9 (6.1) 53 11.0 (6.6) 59 9.3 (6.1) 61 9.5 (5.6) 59

V

CW 7.8 (5.4) 47 5.1 (2.6) 47 7.5 (7.2) 59 6.0 (5.1) 61 6.1 (4.8) 59 SSP 8.6 (6.9) 46 6.9 (5.6) 51 6.5 (5.8) 55 6.8 (6.2) 59 7.8 (10.2) 60 YL 7.4 (2.6) 23 5.2 (3.5) 53 6.2 (5.7) 59 6.2 (5.0) 61 5.8 (3.6) 59

Zn

CW 124 (76.4) 47 152 (116) 47 162 (113) 59 128 (107) 61 129 (95.4) 59 SSP 106 (68.4) 46 114 (86.2) 51 128 (87.3) 55 120 (102) 59 152 (170) 60 YL 187 (114) 23 150 (139) 53 146 (114) 59 153 (149) 61 204 (189) 59

a

There is a measurement bias at YL in 1995 since samplings commenced during the second half of the year. † Unit µg m−3 .

by mass (April 1997). At present, little is known about the emissions inventory of NH3 in Hong Kong, and possible contributing factors such as vehicle emissions have not been quantified. The high correlation between these species results from the formation of compounds of the type (NH4 )2−x Hx SO4

120


(Pathak et al., 2004). The annual equivalent:equivalent ratios [SO4 ]/[NH4 ] are between 1.3 and 1.5 at the three monitoring sites, with the lower values being at YL. The arithmetic annual mean [SO4 ]/[NH4 ] ratio at a quasi-rural site in Hong Kong was measured as 1.5 (Lee and Sequeira, 2002). The similar ratio found at different sites in Hong Kong shows that much of the SO4 and its precursors may come from outside Hong Kong. This is particularly the case in winter, when Hong Kong is influenced by northerly winds. Since the droplet (0.57 µm) mode dominates the size distributions of these ions, the atmospheric residence time is several days (Zhuang et al., 1999a). Some anthropogenic species are highly correlated with Pb: for example, Cd (R > 0.91, P < 0.0001, N > 255) and Zn (R > 0.87, P < 0.0001, N > 255). The introduction of unleaded gasoline in Hong Kong took place in 1991, and leaded gasoline was totally banned in 1999. From Pb:Br ratios in PM10 , we have inferred that Pb levels in Hong Kong represent the regional background (Wai and Tanner, 2002). The strong correlations with Cd and Ni may indicate regional sources of these species. 4.2. S EASONAL

VARIATIONS OF COMPONENTS IN

PM 10

The seasonal variabilities of the categories of species: marine, crustal, anthropogenic and trace, at the 3 sites have been analyzed, and representative species are shown in Figure 2. In Figure 2(a) and (b), Na and Cl show similar seasonal trends, with higher concentrations for the sites nearer to the sea coast. Hong Kong is influenced by southerly winds in summer which transport marine aerosols from the South China Sea. The marine species, however, do not exhibit summer maxima and we ascribe this to the washout process by summer monsoon rains which lowers their ambient concentrations. Thus, the mass ratios with respect to total PM10 do exhibit summer maxima. The April spike for Na and several other species at CW monitoring station (Figure 2(a,b)) arises from the more favourable transport of these species on a Territorial basis, due to the more elevated wind speeds in April (than in the adjacent months) for 3 out of the 5 years. Figures 2(c–f, h) show the marked seasonal variability of anthropogenic species. Overall, most of the species show higher concentrations in winter, but lower in summer. This can be attributed to several factors. The mixing layer is higher and the rainfall more abundant in summer, whereas in winter there is the continental outflow, increased occurrence of inversions and reduction in vertical diffusion (Lee et al., 1994). Figure 2(g) for V does not show a summer minimum, attributed to increased electricity production, and fuel consumption in the hot summer. The crustal species, for example Fe in Figure 2(i), do not show a significant inter-site variation in their seasonal behaviours, and similar trends have been found in urban and also rural areas of Britain for Al, Sc and Fe (Lee et al., 1994). The concentrations of nss-SO4 and NO3 are high in winter and low in summer, as also found by Zhuang et al. (1999b), and indicative of continental outflow.

Figure 2. Seasonal variations of selected inorganic species in PM10 at CW, SSP and YL monitoring stations.


121

122


The seasonal variation of the nss-SO4 /NO3 ratio has been used as an index to point out the sources of sulfur and nitrogen compounds in the atmosphere (Carmichael et al., 1997). Figure 3(b) shows that in Hong Kong the equivalent:equivalent ratio ranges from 2.5–7.6, with a maximum in September. Lee et al. (2001) investigated the characteristics of long-range transport of air pollutants

Figure 3. Seasonal variation of the equivalent:equivalent ratios for (a) Cl/Na, (b) nss-SO4 /NO3 and (c) (nss-SO4 + NO3 )/NH4 at CW, SSP and YL monitoring stations.


123

in northeastern Asia by comparing the molar ratio of ambient SO2 to NOx in Korea (0.5) with the (nss-SO4 )/(total-NO3 ) molar ratio (2.8). It was then concluded that the majority of nss-SO4 and NO3 aerosols were not local. However, there are many other contributing factors (including meteorology) to be considered when considering these ratios, as discussed for bulk deposition by Tanner (1999). In the case of PM10 , the presence of sampling artifacts for these species, as well as the temperature-dependent dissociation of NH4 NO3 further complicate the discussion. Figure 3(a) shows the seasonal variation of the equivalent ratio Cl/Na, and chloride depletion occurs since nearly all values are less than that in seawater (1.17). The September maximum ratio for the nss-SO4 /NO3 ratio, Figure 3(b), coincides with the period of maximum chloride depletion, Figure 3(a), showing the importance of the reaction of H2 SO4 with coarse mode NaCl particles (Wolff, 1984). Note that the maximum of the seasonal variation of (nss-SO4 + NO3 )/NH4 (Figure 3(c): frequently referred to as “acidity”) in July does not coincide with the period of maximum chloride depletion. It has previously been found that in coarse particles in Hong Kong, more chloride was substituted by nitrate than by sulphate (Yao et al., 2003). 4.3. COMPARISON

OF THE COMPOSITION OF PARTICULATE MATTER WITH THAT OF WET DEPOSITION

Table III shows the relative abundance of species, using either Al or Na as reference, in wet deposition (WD) and PM10 at CW station, from 1995–1999. The seasalt components Na and Cl, and also the anthropogenic species V and K, are observed to be relatively more abundant in wet deposition than in PM10 . Two crustal species Ca and Mn show similar relative mass concentrations (WD/PM10 ∼ 0.9) to that of the normalizing crustal species, Al. The anthropogenic species NO3 , SO4 and NH4 are relatively less abundant in WD than in PM10 . The relative concentrations depend upon many factors. Previous size-selective measurements of aerosol particles in Hong Kong have shown that the mean aerodynamic diameter of NH4 is