Contribution of marine and continental aerosols to the content of major ions in the precipitation of the central Mediterranean

Science of the Total Environment 370 (2006) 441 – 451 www.elsevier.com/locate/scitotenv

Contribution of marine and continental aerosols to the content of major ions in the precipitation of the central Mediterranean Aleksandra Mihajlidi-Zelić a , Ivana Deršek-Timotić c , Dubravka Relić a , Aleksandar Popović a , Dragana Đorđević b,⁎ a

Department of Chemistry, University of Belgrade, Studentski trg 12-16, 11000 Belgrade, Serbia and Montenegro b ICTM-Centre of Chemistry, Njegoševa, 12, 11000 Belgrade, Serbia and Montenegro c Republic Hydrometeorological Service of Serbia, Kneza Višeslava 66, 11000 Belgrade, Serbia and Montenegro Received 12 April 2006; received in revised form 5 July 2006; accepted 6 July 2006 Available online 17 August 2006

Abstract The region of the investigated receptor is situated in the southern part of the Adriatic Sea in the Mediterranean. The measuring station is located on the seashore, which, being considered as a border area, is representative for the qualitative and quantitative estimation of the influence of marine and continental aerosols on the content of major ions in precipitation. In the sampling period, precipitation in the region of the investigated receptor was more abundant during the summer and autumn than during the winter and spring. The most frequent precipitation heights were up to 20 mm, while high precipitation came exclusively from the continental region. The results of the measurements of ions readily soluble in water were used for the differentiation of marine from continental contributions of primary and secondary aerosols to their content in the precipitation. Using PCA, it was shown that main contribution of Cl−, Na+ and Mg2+ came from primary marine aerosols, while the contribution from continental sources was − + 2+ dominant for the content of SO2− in the precipitation. The continental origin of Ca2+ was from a primary 4 , NO3 , NH4 and Ca 2− − + source, while SO4 , NO3 and NH4 were representatives of secondary aerosols produced by reactions between acid oxides and − alkaline species in the atmosphere, but SO2− 4 and NO3 also exist in the precipitation as free acids. The origin of the trace elements Cd, Cu, Pb and Zn in the precipitation came from anthropogenic emission sources. The results obtained in this work are based on experimental data from 609 samples collected during the period 1995–2000. © 2006 Elsevier B.V. All rights reserved. Keywords: Precipitation; Major ions; Trace elements; Principal component analysis

1. Introduction The atmosphere in the Mediterranean is affected by air masses coming from Sahara desert or from polluted regions of North and East Europe. Strong influence of these two major sources (Saharan dust and pollution ⁎ Corresponding author. Tel.: +381 11 333 6682; fax: +381 11 636 061. E-mail address: [email protected] (D. Đorđević). 0048-9697/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2006.07.017

aerosols originating from Europe) on the chemical composition of the precipitation in the Mediterranean is well established (Koçak et al., 2004; Glavas and Moschonas, 2002; Avila and Alarcon, 1999; AlMomani et al., 1995; Samara and Tsitouridou, 2000). Additionally, chemical composition of rainwater varies from site to site due to the influence of local sources (Kulshrestha et al., 2003). The acidity of the precipitation depends on the availability of the acid precursors and alkaline species. In the regions which are exposed to

442

A. Mihajlidi-Zelić et al. / Science of the Total Environment 370 (2006) 441–451

strong influence of SO2 and NOx gases, originating mostly from combustion of fossil fuels and industrial processes, and where no natural cleansing mechanism of the atmosphere exists, low pH values are expected. The neutralization of acidity occurs, even in areas which have high SO2, and NOx emissions, when atmosphere is highly loaded with alkaline species, such as CaCO3 from airborne dust (Saxena et al., 1996; Glavas and Moschonas, 2002; Al-Momani, 2003) and/or NH3 originating from agricultural, industrial and even natural activities (Schuurkes et al., 1988; Zunckel et al., 2003). The neutralization of the precipitation acidity is a characteristic for the Mediterranean region due to CaCO3, originating from Saharan dust and calcareous soil of the Mediterranean Sea coasts (Avila and Alarcon, 1999; Glavas and Moschonas, 2002). Sulphate, nitrate and ammonium ions are formed in the atmosphere by gas-particle conversion processes. Sulphate and nitrate aerosol are mainly formed from the reaction of sulphuric acid and nitric acid with alkaline species in the atmosphere, e.g. ammonia, sea salt or dust. Ammonium salts obtained by reaction with NH3 are the most important under environmental conditions (Lee and Atkins, 1994; Langford et al., 1992). The sulphates in the atmosphere can also originate from other sources: e.g. marine sea spray aerosol, gypsum CaSO4·2H2O from (re-suspended) dust from the Sahara (Avila and Alarcon, 1999; Glavas and Moschonas, 2002), etc. Sulphuric acid is obtained by the oxidation of SO2, which was either directly emitted to the atmosphere (burning of fossil fuels, industrial processes, volcanoes, combustion of biomasses) or by the oxidation of lower oxidation state sulphur compounds, mainly DMS, which is the dominant source of SO2 in marine atmosphere (Seinfeld and Pandis, 1998). The precursors of nitric acid are NOx, primarily originating from combustion of fossil fuels in high temperature processes (traffic, power plants, industry, domestic fireboxes), soil (microbiological activity), combustion of biomass, lightning, etc. (Seinfeld and Pandis, 1998). The main natural sources of NH3 are the decay of vegetation in soils, wild animals and oceans, while the anthropogenic sources are global livestock farming, the employment of fertilizers in agriculture, and the combustion of biomass (Seinfeld and Pandis, 1998). Far away from land, the only significant source of NH3 is the ocean where it is obtained by the decay of N-containing organic compounds and of the secretion of zooplanktons. The primary pollutants SO2 and NOx and their secondary products H2SO4, HNO3, SO42−, NO3−, as well as their subsequent reaction products (NH4)2SO4 and

NH4NO3 are continuously removed from the atmosphere by dry and wet deposition processes. Although wet deposition is very efficient, it is sporadic by nature. The residence time of SO2 and NOx in the troposphere is 1–3 days (Glavas and Moschonas, 2002), while the residence time of sulphates and nitrates is somewhat longer: for nitrates it is 3–9 days (Seinfeld and Pandis, 1998), and it can even be 10 days for sulphates during dry periods in the Mediterranean region (Luria et al., 1996). Na+, K+, Ca2+, Mg2+ originate mainly from natural sources: from marine aerosol, rocks and soil resuspension and forest fires, while most of the Cd2+, Pb2+, Zn2+ and Cu2+ in precipitation comes from anthropogenic emission sources (traffic, industry) but natural sources, such as volcanic eruptions and resuspension, also contribute (Đorđević et al., 2004, 2005). Determination of chemical composition of rainwater, provides an understanding of the source types that contribute to rainwater chemistry, and enhances the understanding of the local and regional dispersion of pollutants and their potential impacts on ecosystems through deposition processes (Zunckel et al., 2003). The main goal of this study was the differentiation of marine and continental aerosol contribution to the content of anions and cations in precipitation in the Mediterranean region. The differentiation was based on analysis of precipitation samples and processing of the obtained database by Principal Component Analysis (PCA). Study area. The investigated receptor is situated in the coastal belt of a region of the Mediterranean (Fig. 1). The coastal belt is a narrow zone separated from the continental region by a mountain range and has a typical Mediterranean climate. Precipitation samples were collected at a measuring site located in the area of the Hydrometeorological station Herceg Novi (18°33′ N, 42°27′ E). The Hydrometeorological station is situated in the eastern outskirts of Herceg Novi on the coast of the south Adriatic Sea (Đorđević et al., 2004, 2005). The

Fig. 1. Location of the Sampling Station.

A. Mihajlidi-Zelić et al. / Science of the Total Environment 370 (2006) 441–451

Hydrometeorological station is in the framework of the International programme MED-POL. The details regarding the study area are presented in Đorđević et al. (2004). 2. Methodology 2.1. Precipitation sampling A bottle and funnel with a small net between them, the function of which was to prevent contamination of the samples (leaves, small branches, insects…) was used for sampling. The samplers were at a height of 1.5 m above the ground. The following meteorological data were available for each sample: wind velocity, wind direction and temperature. All these data were measured at the same hydrometeorological station. Parallel samples of the precipitation (one sample for the major ions, and the other for the trace elements) were collected daily from 8 am to 8 am (meteorological time) the following day in the period from January 1995 until December 2000. Field blanks were also taken by pouring deionised water through the funnel of the sampler. For determination of major ions, prior to the analysis the sample bottles and plastic equipment used for sampling (PE bottles and funnels) were first submerged in water for 24 h, then they were washed with detergent in warm water to mechanically remove any dirt and then they were rinsed with deionised water until the electrical conductivity of the washings was less than 2 μScm− 1. The sample bottles and plastic equipment (PE bottles and funnels) used for the collection of the precipitation for trace elements analysis were washed with detergent, rinsed with deionised water, submerged in HNO3 for 48 h and finally rinsed with deionised water until the electrical conductivity of the washings was less than 2 μScm− 1. 2.2. Chemical analysis In the context of the MED-POL programme, the following parameters of the precipitation samples were determined: pH, electrical conductivity and the concentrations of SO42−, NO3−, Cl−, NH4+, Na+, K+, Ca2+, Mg2+, Cd2+, Pb2+, Zn2+ and Cu2+. After pH and conductivity measurements, samples were filtered through 0.45 μm membrane filter. Filtered samples used for trace element determinations were conserved by addition of concentrated HNO3 (supra pure) until a pH b 2 was obtained. The precipitation samples were stored in a refrigerator at 4 °C.

443

The amounts of anions in precipitation were determined spectrophotometrically as follows: Cl− by the mercury thiocyanate-iron method, NO3− by the Griess method and SO42− by the barium perchlorateThorin method. The amount of NH4+ was determined spectrophotometrically by the indophenol blue method. The amount of Na+ and K+ in precipitation was determined by the FAES method, the amounts of Mg2+ and Ca2+ by the FAAS method, and Cd, Cu, Pb and Zn by the GF-AAS method. The standard calibration curve for metal determination was obtained using a multi-element standard in which the ratio of the concentration of the metals was analogous to their ratio in the samples. The working standards used in FAAS, FAES and GF-AAS analysis were prepared daily from commercial stock solutions (Varian). Interlaboratory check for analytical quality was provided by participation in laboratory intercomparisons. In 1999 our laboratory participated in the second EMEP's analytical intercomparison of heavy metals in precipitation (EMEP, 1999) and in the 2000 laboratory participated in the eighteenth intercomparison of analytical methods within EMEP (EMEP, 2001). Of the 40 determinations reported (ten parameters in four samples) in the eighteenth intercomparison of analytical methods within EMEP, only for 3 results deviated more than 2 standard deviations of mean value. 2.3. Statistical analysis Principal component analysis (PCA) is a multivariate statistical method which transforms a set of (possibly) correlated variables into a (smaller) set of uncorrelated variables, called principal components, which could explain the variability of most of the original data. The principal components with eigenvalues greater than 1 were subjected to a varimax rotation, which maximizes the variance to obtain a pattern of loadings for each factor which is as diverse as possible, thus lending itself to easier interpretation. After rotation of the factor-loading matrix, the factors were interpreted as origins or common sources (Astel et al., 2004), assuming that the intercorrelations among the original variables were generated by some smaller number of unobserved factors (Kessler et al., 1992). The datasets were processed using the SPSS 10.0 statistical program. 3. Results and discussions The results of the investigations presented in this study form a comprehensive database based on 609 precipitation sample. The basic meteorological parameters, including the air temperature and the amounts of

444

A. Mihajlidi-Zelić et al. / Science of the Total Environment 370 (2006) 441–451

precipitation obtained at the same measuring station, were provided by the Hydrometeorological station for the days with precipitation during the study period. Histograms showing the frequency of the daily amount of precipitation, the pH of the precipitation in the measuring period and the correlation between these variables are presented in Fig. 2a, b and c, respectively. The average daily temperature for the days with precipitation ranged from 5 °C to 26 °C, with a frequency maximum at around 12 °C. The maximum

Fig. 3. The seasonal variations of a) the precipitation height and the electrical conductivity and b) the Cl/Na ratio.

Fig. 2. Histograms of the frequency of: a) the precipitation height, b) the pH of the precipitations in the measurement campaign, and c) the correlation between the precipitation height and the pH.

frequency of the precipitation height was at about 10 mm and the most frequent pH value was about 6.5. The pH values ranged from 4.10 to 9.15, with an average value of 6.26 ± 0.73. Histogram of the pH distribution (Fig. 2b) shows that ≤ 20% samples had a pH b 5.6, and that more than 70% of the precipitation could be characterized as alkaline. Previous studies showed that ∼25% of the precipitation in the centraleastern Mediterranean was acid events pH b 5, and a large portion were alkaline events, pH N 6 (Samara et al., 1992; LeBolloch and Guerzoni, 1995; Guelsoy et al., 1999; Turkan and Cemal Saydam, 2000). Acidic and alkaline rains also occurred in Israel (Maman and Gottlieb, 1995), as well as in the western Mediterranean in Spain (Avila and Alarcon, 1999). They all showed that, in the Mediterranean, marine aerosol and Saharan dust plays a substantial role in neutralizing the acidic constituents. The correlation coefficient (r) between the precipitation height and the pH value was negative and equal to − 0.268, with a confidence level higher than 99%. Large amounts of precipitation (N 80 mm), the circled values in Fig. 2c, came exclusively from the continental direction (northeast segment), while low values (b10 mm) came from all directions, with the ESE continental direction dominant. pH values lower than 5.6 of the abundant precipitation (N80 mm) (Fig. 2c) coming from the continental direction indicate an excess of acids e.g., H2SO4 and HNO3 originating from acid

A. Mihajlidi-Zelić et al. / Science of the Total Environment 370 (2006) 441–451 Table 1 Values of the parameters measured in precipitation during the study campaign 1995–2000 Minimum Maximum Mean Temperature (0C) Precipitation (mm) Conductivity (μScm− 1) pH H+(μeq L− 1) −1 SO2− 4 (μeq L ) NO−3 (μeq L− 1) NH+4(μeq L− 1) Na+(μeq L− 1) Mg2+(μeq L− 1) Ca2+(μeq L− 1) Cl−(μeq L− 1) K+(μeq L− 1) Cd (μg L− 1) Pb (μg L− 1) Cu (μg L− 1) Zn (μg L− 1)

St. Deviation VWM

4.5

26.4

15.0

4.9

0.5

220.2

16.7

22.0

78

99

6

1250

4.10 10−3 3.1 2.9 1.4 6.5 1.7 4.0 12.7 0.3 b0.01 b0.12 b0.12 b0.14

9.20 79.4 2219.6 342.6 156.0 2611.6 533.1 1741.5 1234.1 483.4 23 69 195 40

6.26 2.40 132.78 47.98 43.94 235.94 58.62 147.41 138.23 21.68 0.4 6.2 13.6 2.6

0.73 6.78 149.47 28.15 38.07 286.48 70.54 198.59 111.16 37.23 1.8 10.5 30.4 5.1

5.51 3.12 92.42 40.49 30.70 210.80 50.63 91.37 134.38 13.63

oxides mainly emitted by the combustion of fossil fuels and some other high temperature processes. The neutralizing agents, of local origin (alkaline dust) are extracted during the first stages of precipitation by below cloud scavenging processes. Once this neutralizing agents are removed from the atmosphere (usually in the 5 first mm), in the cloud scavenging processes are dominant supplying SO42−, NO3−. The seasonal variations of the precipitation heights and conductivity in the studied interval (1995–2000) are shown in Fig. 3. The abundance of precipitation in measurement campaign was larger in the summer and autumn period than during the winter and spring, and the electrical conductivity was, as expected, inversely proportional to the amount of precipitation. Fig. 3b) shows the seasonal evolution of the Cl−/Na+ ratio.

445

The results of the measurements of the average daily temperature, the daily amount of precipitation, the electrical conductivity and the values of the concentrations of the investigated ions in the precipitation, given as maximum, minimum, average values and standard deviations, are presented in Table 1. The presented results show that Na+ (235.94 μeqL− 1), 2+ Ca (147.41 μeqL− 1), Cl− (138.23 μeqL− 1) and SO42− (132.78 μeqL− 1) were the dominant components in the precipitation. The ionic species can be arranged according to decreasing amount in precipitation: Na+ N Ca2+ N Cl− N SO42− N Mg2+ N NO3− N NH4+ N K+. The results of this investigation of the contents of ionic species in the precipitation in the receptor region (Fig. 1) are in agreement with the results obtained in other locations in the region of the Mediterranean (Glavas and Moschonas, 2002; Samara and Tsitouridou, 2000; Avila and Alarcon, 1999; Al-Momani et al., 1995; LeBolloch and Guerzoni, 1995). The only significant discrepancies were the concentrations of Na+, Mg2+ and Cl−. Namely, our results, and also the results from the central and east Mediterranean were several times larger than those from the west Mediterranean. Differences between marine concentrations in rain between west and east Mediterranean may be related with meteorology: rain in the west Mediterranean is usually associated with the passage of cold fronts from the Atlantic (Avila and Alarcon, 1999). Seasonal variations of concentrations of the investigated ionic species: Ca2+, Mg2+, Na+, K+, NH4+, Cl−, SO42− and NO3− are shown in Fig. 4. The same seasonal trend can be observed for Mg2+, Na+ and Cl− on one hand and for SO42− and NH4+ on the other. Temporal variation of average monthly concentrations of major inorganic ionic species in precipitation is shown in Fig. 5. In this study it was shown that ratios of Cl−/Na+ around 0.5 were the most frequent, which can be an indicator of a chloride deficiency in the marine aerosol in the border area between marine and continental regions,

Fig. 4. The seasonal variations of the ionic species concentrations (average value ± standard deviation) in the measurement campaign.

446

A. Mihajlidi-Zelić et al. / Science of the Total Environment 370 (2006) 441–451

Fig. 5. Time series of average monthly concentrations (μeq L− 1).

where the investigated receptor is situated (Saxena et al., 1996; Chandra Mouli et al., 2005) The average ratios of Na+/Cl− and (SO42− + NO3− )/(NH4+ + Ca2+ + Mg2+ + K+) are 1.090 and 0.904 respectively. The rose of winds presented in Fig. 6a clearly shows that the strongest wind comes from the west and east

segments. The temperature rose indicate (Fig. 6b) that the days with precipitation having the highest temperatures are influenced by the south segment. The most abundant precipitation (Fig. 6c) comes from the east segment (NE/ E/SSE). The rose of pH, presented in Fig. 6d, indicates that the highest pH values are characteristic for

A. Mihajlidi-Zelić et al. / Science of the Total Environment 370 (2006) 441–451

447

Fig. 6. Roses of a) winds, b) temperature, c) precipitations and d) pH.

precipitation which are under the influence of sea aerosol (S-SW), but there is another influence coming from the N direction. The concentration roses (VWM) of SO42−NO3−, NH4+, − Cl , Na+, Mg2+, Ca2+ and K+ are shown in Fig. 7. The largest contribution to the content of SO42− and NH4+ in the precipitation comes from the east segment. Contributions to the SO42− content from the west and south segment may be from both sea aerosol and from long range transport of Saharan dust. The CaSO4 in the atmosphere may be related to the reaction between calcite from African dust with sulphur compounds (SO2, H2SO4, HSO4−, SO42−). Long-range transport of Saharan dust has already been proved by many authors (Kubilay et al., 2000; Rodríguez et al., 2002; Koçak et al., 2004; Đorđević et al., 2004, 2005; Alastuey et al., 2005). The roses of the concentrations of K+, Ca2+, and Cl- have the largest contributions from the south segment. It is already known that Saharan dust can be a source of Ca2+ in the atmospheric aerosols of the Mediterranean region (Al-Momani et al., 1995; Avila and Alarcon, 1999;

Samara and Tsitouridou, 2000). Concentration roses of Na+and Mg2+, which are extremely similar, are also similar with Cl− concentration rose. The graphical presentation of the correlations is given in Fig. 8. The best correlations are for the following pairs: Na+–Mg2+, Mg2+–Cl− and Na+–Cl− which indicate their common sources. It seems that there are two correlations patterns between Mg2+ and Na+ with Cl−. This may be related with the volatilization of Cl− in summer by the reaction of sea spray (i.e. NaCl) with acidic species. PCA was applied on the database and on the basis of the existing correlations, a new set of variables was formed — 13 components were extracted. The extracted components indicate that there are 13 potential different sources of contributions of the investigated ionic species and chemical elements in the precipitation. Realistically, a few sources of the main contributions could be assessed. According to the Kaiser Criterion, the first 5 components, with initial eigenvalues larger than 1 for the available database, represent the main influences. These 5 components comprise 69.064% of the total variance of

Fig. 7. Roses of the ionic species concentrations (VWM).

448

A. Mihajlidi-Zelić et al. / Science of the Total Environment 370 (2006) 441–451

Fig. 8. The correlations between the ionic species at the level of significance p N 99%.

A. Mihajlidi-Zelić et al. / Science of the Total Environment 370 (2006) 441–451

the investigated set of data, which also presents their total contribution. The varimax rotation was applied on the 5 significant factors with an initial eigenvalue N 1. The obtained varimax rotated factor loadings are shown in Table 2. Component 1 (F1), with the highest variance of 27.6%, has a high loading for ionic species present in large amounts in seawater: Na+ (0.846), Mg2+ (0.861) and Cl− (0.863) and K+ (0.588). Hence, it can be termed the marine component because it presents the contribution of marine aerosol (Fig. 9). For component 2 (F2), which explains 14.8% of the total variance of the investigated data set, high loadings values for NH4+ (0.840) and SO42− (0.749) and moderate values for K+ (0.443) and Ca2+ (0.489) and for NO3− (0.403) are characteristic (Fig. 9). The representatives of secondary aerosols (NH4+ and SO42−) correlate well with component 2 (F2), hence this component can be considered as the representative of the contributions of secondary aerosols. It is known that combustion of fossil fuels presents one of the main anthropogenic sources of SO2 as the precursor of SO42− which is obtained by oxidation reactions in the atmosphere (Migliavacca et al., 2004; Zunckel et al., 2003). The coal-fired power plant Pljevlje, which is situated 50 km to the north of the sampling site, could be a significant source of SO2, but SO2 also originates from local traffic and combustion in individual furnaces (Đorđević et al., 2005). One of the potential sources of SO2, the further oxidation of which in atmosphere produces SO42−, could also be volcanic eruptions in south Italy, such as, for example, the permanently active volcano Stromboli (Đorđević et al., 2004). Marine aerosol and Saharan dust, namely the reaction of calcite from Saharan dust with sulphur Table 2 Varimax rotated factor loadings of the measured species Component

H+ SO2− 4 NO−3 NH+4 Na+ Mg2+ Ca2+ Cl− K+ Conductivity Cd Pb Cu Zn

1

2

3

4

5

− 0.029 0.405 0.085 − 0.004 0.846 0.861 0.353 0.863 0.588 0.659 − 0.039 0.083 − 0.070 0.013

−0.183 0.749 0.403 0.840 0.173 0.040 0.443 0.048 0.489 0.465 0.019 0.161 0.059 −0.171

0.058 − 0.073 − 0.285 − 0.007 − 0.108 0.029 0.121 − 0.033 0.170 0.055 0.010 0.787 0.753 0.731

0.861 0.147 0.529 −0.088 0.040 −0.083 −0.189 0.061 −0.135 0.152 0.043 0.058 −0.114 −0.006

0.111 0.065 − 0.150 − 0.070 − 0.132 − 0.007 0.106 0.006 − 0.006 0.069 0.960 − 0.204 0.113 0.067

449

Fig. 9. The factors representing the associations of ionic species.

compounds, also have an influence on the content of sulphates in the precipitation in the region of the investigated receptor. In the region of the investigated receptor, primary marine aerosol has an influence on the content of SO42− ions in the precipitation, but they also originated from secondary aerosols. The loading factors of NH4+ and K+ in the same component could suggest that the contributions came from combustion processes. Forest fires, which are frequent in the investigated region, could be their source (Saxena et al., 1996), but also the use of fertilizers in the agriculture sector, as well as the livestock industry, i.e. animal wastes, excretion, as sources of NH4+ ions should not be neglected (Saxena et al., 1996; Zunckel et al., 2003). Generally, it can be considered that the profile of the sources in component 2 (F2) corresponds to emissions from the continental sector. The third component (F3) accounts for 10.2% of the total variance and has high loadings of Pb (0.790), Cu (0.754) and Zn (0.729) can be connected with continental sources (Fig. 9). The sources of these elements can be high temperature processes, such as the ironworks Nikšić, which is located 50 km north and the slag heap at Mojkovac located about 100 km northeast of the receptor (Đorđević et al., 2005), traffic (Al-Momani, 2003; Huang et al., 1994; Fernández et al., 2004) and also long range transport. In the fourth component (F4), which comprises 8.55% of the total variance, H+ has high (0.861) and NO3− a moderate (0.529) loading, hence it can be considered as an acidic component (Fig. 9). The fifth component (F5), with 7.8% of the total variance, is characterized by a high Cd loading (0.963). It has already been shown that the most significant source of Cd emission in the region of this receptor is from a source of plant for the repairs of boats located 5 km SSE of the investigated receptor, on the peninsula Luštica (Đorđević et al., 2005).

450

A. Mihajlidi-Zelić et al. / Science of the Total Environment 370 (2006) 441–451

The electrical conductivity is assigned to factor 1, which practically means that the association of the ionic species Na+–Mg2+–Cl−–K+ makes the highest contribution to the electrical conductivity. Nevertheless, in the bi-variant correlations of electrical conductivity with the ionic species, the highest value was obtained between electrical conductivity and SO42−(r = 0.597). Analysis of the factors showed that Ca2+ does not have a high loading in factor F1 which is associated with the influence of marine aerosol (Fig. 9). The eigenvalue for Ca2+ has moderate loadings in F1 and in F2, the latter component, which is associated with continental influences, showed a higher contribution. 4. Conclusion The investigation of the content of ionic species in precipitation can contribute to an understanding of the influence of different emission sources and their contribution to the quality of the air of some region. The results of this investigation of the contents of ionic species in the precipitation in the receptor region are in agreement with the results obtained in other locations in the region of the Mediterranean. The presented results show that Na+, Ca2+, Cl− and SO42− were the dominant components in the precipitation. The best correlations are between Na+ and Mg2+, Na+ and Cl− and Mg2+ and Cl− indicating their common origin, e.g. marine aerosol. Two correlations patterns Na+ and Cl− and Mg2+ and Cl− may be related with the volatilization of Cl− in summer by the reaction of NaCl from sea spray with acidic species. Of the ionic species present in the precipitation the highest individual contribution to the electrical conductivity comes from SO42−, while analysis of the main components (PCA) showed that the main carriers of the electrical conductivity of the precipitation were Mg2+, Na+, Cl− and K+, cumulatively. Ca2+ originates mainly from continental sources. The Cl−/Na+ ratio in precipitation significantly deviated from the Cl−/ Na+ ratio in seawater, which could indicate a chlorine deficiency in the atmospheric aerosol of the border region between marine and continental areas. The contribution of marine aerosol to the content of the major ionic species in the precipitation is dominant in the region of the investigated receptor, but sources from the continental directions also have an influence. Acknowledgements The authors are grateful to the Ministry of Science and Environment of the Republic of Serbia for its financial support. We would also like to thank the

Hydrometeorological Services of the Republic of Serbia and of the Republic of Montenegro for the meteorological data used in this paper. References Alastuey A, Querol X, Castillo S, Escudero M, Avila A, Cuevas E, et al. Characterisation of TSP and PM25 at Izaña and Sta Cruz de Tenerife (Canary Islands, Spain) during a Saharan dust episode (July 2002). Atmos Environ 2005;39:4715–28. Al-Momani IF. Trace elements in atmospheric precipitation at Northern Jordan measured by ICP-MS: acidity and possible sources. Atmos Environ 2003;37:4507–15. Al-Momani IF, Tuncel S, Eler U, Ortel E, Sirin G, Tuncel G. Major ion composition of wet and dry deposition in the eastern Mediterranean basin. Sci Total Environ 1995;164:75–85. Astel A, Mazerski J, Polkowska Z, Namiesnik J. Application of PCA and time series analysis in studies of precipitation in Tricity (Poland). Adv Environ Res 2004;8:337–49. Avila A, Alarcon M. Relationship between precipitation chemistry and meteorological situations at a rural site in NE Spain. Atmos Environ 1999;33:1663–77. Chandra Mouli P, Venkata Mohan S, Jayarama Redd S. Rainwater chemistry at a regional representative urban site: influence of terrestrial sources on ionic composition. Atmos Environ 2005;39: 999-1008. Đorđević D, Vukmirović Z, Tosić I, Unkasević M. Contribution of dust transport and resuspension to particulate matter levels in the Mediterranean atmosphere. Atmos Environ 2004;38:3637–45. Đorđević D, Mihajlidi-Zelić A, Relić D. Differentiation of the contribution of local resuspension from that of regional and remote sources on trace elements content in the atmospheric aerosol in the Mediterranean area. Atmos Environ 2005;39:6271–81. EMEP. Analytical intercomparison of heavy metals in precipitation 1999. EMEP/CCC-Report 8/2000, NILU, Norway. http://www. nilu.no/projects/ccc/reports/cccr8-2000.pdf. Cited 12 Mar 2006. EMEP. The eighteenth intercomparison of analytical methods within EMEP. EMEP/CCC-Report 11/2001, NILU, Norway http://www. nilu.no/projects/ccc/reports/cccr11-2001.pdf. Cited 12 Mar 2006. Fernández AJ, Ternero M, Fernández F. Source characterisation of fine urban particles by multivariate analysis of trace metals speciation. Atmos Environ 2004;38:873–86. Glavas S, Moschonas N. Origin of observed acidic–alkaline rains in a wet-only precipitation study in a Mediterranean coastal site, Patras, Greece. Atmos Environ 2002;36:3089–99. Guelsoy G, Tayanc M, Ertuerk F. Chemical analysis of the major ions in the precipitation of Istanbul, Turkey. Environ Pollut 1999;105: 273–80. Huang X, Olmez I, Aras NK. Emissions of trace elements from motor vehicles: potential marker elements and source composition profile. Atmos Environ 1994;28:1385–91. Kessler CJ, Porter TH, Firth D, Sager TW, Hemphill MW. Factor analysis of trends in Texas acidic deposition. Atmos Environ 1992;26:1137–46. Koçak M, Kubilay N, Mihalopoulos N. Ionic composition of lower tropospheric aerosols at northeastern Mediterranean site: implications regarding sources and long-range transport. Atmos Environ 2004;38:2067–77. Kubilay N, Ničković S, Moulin C, Dulac F. An illustration of the transport and deposition of mineral dust onto the eastern Mediterranean. Atmos Environ 2000;34:1293–303.

A. Mihajlidi-Zelić et al. / Science of the Total Environment 370 (2006) 441–451 Kulshrestha UC, Kulshrestha MJ, Sekar R, Sastry GSR, Vairamani M. Chemical characteristics of rainwater at an urban site of South Central India. Atmos Environ 2003;37:3019–26. Langford AO, Fehsenfeld FC, Zachariassen J, Schimel DS. Gaseous ammonia fluxes and background concentrations in terrestrial ecosystems in the United States. Glob Biogeochem Cycles 1992;6: 459–83. LeBolloch O, Guerzoni S. Acid and alkaline deposition in precipitation on the western coast of Sardinia, central Mediterranean. Water Air Soil Pollut 1995;85:2155–60. Lee DS, Atkins DHF. Atmospheric ammonia emissions from agricultural waste combustion. Geophys Res Lett 1994;21:281–4. Luria M, Peleg M, Sharf G, Tov-Alper DS, Spitz N, Ben Ami Y, et al. Atmospheric sulphur over east Mediterranean region. J Geophys Res 1996;101:25917–30. Maman Y, Gottlieb J. Ten years of precipitation in Haifa, Israel. Water Air Soil Pollut 1995;82:549–58. Migliavacca D, Teixeira EC, Pires M, Fachel J. Study of chemical elements in atmospheric precipitation in South Brazil. Atmos Environ 2004;38:1641–56. Rodríguez S, Querol X, Alastuey A, Plana F. Sources and processes affecting levels and composition of atmospheric aerosol in the western Mediterranean. J Geophys Res 2002;107:12-1-12-14.

451

Samara C, Tsitouridou R. Fine and coarse ionic aerosol composition in relation to wet and dry deposition. Water Air Soil Pollut 2000;120: 71–88. Samara C, Tsitouridou R, Balafoutis C. Chemical composition of rain in Thessaloniki, Greece, in relation to meteorological conditions. Atmos Environ 1992;26B:359–67. Saxena A, Kulshrestha UC, Kumar N, Kumari KM, Srivastava SS. Characterization of precipitation at Agra. Atmos Environ 1996;30: 3405–12. Seinfeld JH, Pandis, SN. Atmospheric Chemistry and Physics From Air Pollution to Climate Change, John Wiley and Sons, Inc, New York 1998, pp. 55–74, 363–381. Schuurkes JAR, Maenen MJ, Roelofs JM. Chemical characteristics of precipitation in NH3-affected areas. Atmos Environ 1988;22: 1689–98. Turkan Ö, Cemal Saydam A. Acidic and alkaline precipitation in the Cilician Basin, north-eastern Mediterranean Sea. Sci Total Environ 2000;253:93-109. Zunckel M, Saizar C, Zarauz J. Rainwater composition in northeast Uruguay. Atmos Environ 2003;37:1601–11.