Long-range atmospheric transport of PAHs, PCBs and PBDEs to the central and eastern Mediterranean and changes of PCB and PBDE congener patterns in summer 2010

Atmospheric Environment 111 (2015) 51e59

Contents lists available at ScienceDirect

Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv

Long-range atmospheric transport of PAHs, PCBs and PBDEs to the central and eastern Mediterranean and changes of PCB and PBDE congener patterns in summer 2010  a, Marie D. Mulder a, Angelika Heil b, Petr Kuku
cka a, Jan Kuta a, Petra P
ribylova a a , c, * Roman Proke
s , Gerhard Lammel a b c

Masaryk University, Research Centre for Toxic Compounds in the Environment, Brno, Czech Republic Max Planck Institute for Chemistry, Atmospheric Chemistry Dept., Mainz, Germany Max Planck Institute for Chemistry, Multiphase Chemistry Dept., Mainz, Germany

h i g h l i g h t s
PCBs, DDXs, PBDEs, penta- and hexachlorobenzene were measured shipbourne in summer.
Transport paths can explain air samples' composition and concentration levels.
PCB degraded along transport by reaction with the OH radical.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 5 November 2014 Received in revised form 13 March 2015 Accepted 22 March 2015 Available online 1 April 2015

The central and eastern Mediterranean is a receptor area for persistent organic pollutants (POPs) emitted in western, central and eastern Europe, particularly during summer. Atmospheric concentrations of PCBs, DDXs, PBDEs, penta- and hexachlorobenzene were measured during a ship-borne survey in the summer of 2010. The concentration of PCBs (sum of 7 congeners) was 3.61 (2.08e7.72) pg m3, of which 6.7% was associated with the particulate phase. The mean concentration of DDT isomers and their metabolites, DDE and DDD, was 2.60 (0.46e7.60) pg m3 (particulate mass fraction q ¼ 0.097), of penta- and hexachlorobenzene 0.22 (<0.39e2.80) pg m3 and 6.29 (2.48e24.16) pg m3, respectively, and of PBDEs (sum of 8 congeners) 7.31 (2.80e19.89) pg m3. The air masses studied had been transported mostly across central Europe, some crossing western Europe. The observed changes of PCB congener patterns along transport routes are in agreement with the perception that the reaction with the OH radical is dominating PCB atmospheric lifetime, and indicate an overestimation of the second order gas-phase reaction rate coefficient of PCB153 with OH by structure-activity relationship. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Long-range transport Marine environment Persistent organic pollutants Mediterranean

1. Introduction The marine atmospheric environment is a receptor area for most legacy persistent organic pollutants (POPs) as no significant marine sources exist. In some sea regions, however, accumulation in surface waters due to atmospheric deposition has caused flux reversal (Bidleman et al., 1995; Lakaschus et al., 2002; Mulder et al., 2014). Polychlorinated biphenyls (PCBs) and polybrominated diphenylethers (PBDEs) are mostly emitted in urban areas. While primary

* Corresponding author. Masaryk University, Research Centre for Toxic Compounds in the Environment, Brno, Czech Republic. E-mail address: [email protected] (G. Lammel). http://dx.doi.org/10.1016/j.atmosenv.2015.03.044 1352-2310/© 2015 Elsevier Ltd. All rights reserved.

sources of PCBs are still important (Gasic et al., 2010; Schuster et al., 2010), the majority of sources are secondary (e.g. polluted soils; Lammel and Stemmler, 2012). In Italy, PCBs were produced from 1958 to 1983 with a cumulative production of approximately 31,092 tons (Breivik et al., 2007). In Greece, approximately 400 transformers and 15,000 capacitors containing PCBs were still in use in 1993 (Mandalakis et al., 2002). The phasing out process was planned to end in 2005 in Italy, and in 2010 in Greece (CLEEN, 2005). PBDEs occur in three technical mixtures: pentaBDE, octaBDE and decaBDE. The penta mix formulation consists of ten isomers dominated by the tetraBDEs BDE47 and BDE99 (>70%). The octaBDE mixture contains BDE183 as major congener and the decaBDE mixture consists primarily of the fully brominated BDE209 (la Guardia et al., 2006). The

52

M.D. Mulder et al. / Atmospheric Environment 111 (2015) 51e59

pentaBDE and octaBDE mixtures were banned by the European Union in 2004, while the use of decaBDE was restricted in 2008 (EU, 2003 and EU, 2008). Current major PBDE sources are waste processing and separation, sewage and sludge releases, and landfill leachate (Darnerud et al., 2001). Similarly, almost no primary sources should exist, solely combustion and re-volatilisation should distribute hexachlorobenzene (HCB), pentachlorobenzene (PeCB) and dichlorodiphenyltrichloroethane (DDT) which mainly entered the environment applied as pesticides. Primary sources of polycyclic aromatic hydrocarbons (PAHs) are combustion processes (power plants, biomass burning and road transport; at sea: ship exhaust, see Donateo et al., 2014; Lammel, 2015). Earlier studies of the central and eastern Mediterranean air, reported PCB at Finokalia and Iraklion, Crete (Mandalakis and Stephanou, 2002; Iacovidou et al., 2009), in and near Athens (Mandalakis et al., 2002), and in Bari, southern Italy (Estellano et al., 2014); PBDEs in Finokalia and Iraklion, Crete (Iacovidou et al., 2009); PAHs in the Bursa area, northwestern Turkey (Esen et al., 2008); PAHs (Cetin et al., 2007; Bozlaker et al., 2008a, 2008b; Kaya et al., 2012; Aydin et al., 2014), PCBs (Sofuoglu et al., 2004; Cetin et al., 2007; Bozlaker et al., 2008a,b; Kaya et al., 2012; Aydin et al., 2014), and PBDEs (Cetin and Odabasi, 2008) in the Aliaga and Izmir area, western Turkey (and other areas not directly relevant for this study). Most POPs are semivolatile (vapour pressures at 298 K in the range 10-6e102 Pa) and, hence, partition between the gas and particulate phases, influenced by temperature, particulate phase chemical composition and particle size. Due to significant lipophilicity (Kow ¼ 104e107) and bioaccumulation, these substances pose a hazard for marine ecosystems. This paper reports concentrations in air of PCBs, DDT including its metabolites DDE and DDD (substance group called DDX), PBDEs, penta- and hexachlorobenzene sampled in the summer of 2010 aboard the RV Urania cruise. To increase our understanding of PCBs and OCPs long-range transport (LRT) and cycling in the marine environment, we identify air mass origin and potential source areas influencing the samples and test 3 gas-particle partitioning models. Mulder et al., 2014 reported from the same RV Urania campaign, the PAH concentrations in air and water, gas-particle partitioning and a simulation of re-volatilisation of retene (RET). Tropospheric LRT to the central and eastern Mediterranean arriving from the north to north-west is reported to be the most frequent, followed by transport from the east to northeast (Mihalopoulos et al., 1997; Kouvarakis et al., 2000; Lelieveld et al., 2002; Duncan and Bey, 2004). During summer, a strong eastewest surface pressure difference (>20 hPa) is generated between the Azores' high and Asian monsoon low-pressure regimes, and causes northerly flow in the lower troposphere over the Mediterranean (Lelieveld et al., 2002). Subsidence is prevailing due to the branch of the Hadley circulation descending above the Mediterranean Sea. LRT from urban and industrial sources on land in the region as well as, in western, central and eastern Europe are the predominant sources of PAHs and halogenated persistent organics in the central and eastern Mediterranean (Gogou et al., 1996; Mandalakis et al., 2003; Tsapakis et al., 2003; Esen et al., 2008; Iacovidou et al., 2009; Berrojalbiz et al., 2014). Emissions from wildfires affecting the region carry PAHs (Mulder et al., 2014) and, potentially, also halogenated persistent organics (Eckhardt et al., 2007). POPs' photochemical degradation in air, albeit slow when €pffer and Wagner, 2007), is excompared to other organics (Klo pected to contribute significantly to the change of congener patterns during LRT from source to remote areas, at least in the case of PCBs (Wania and Daly, 2002; Lammel and Stemmler, 2012). Direct evidence of photochemical degradation of PCBs was observed in the eastern Mediterranean (Mandalakis et al., 2003). A wide

spectrum of photochemical degradation rates is expected for PBDEs (Eriksson et al., 2004). The objective of this study was to identify potential source regions, pathways and types of emissions contributing to the pollution of the marine atmospheric environment by POPs, and to track congener pattern change during LRT. To this end shipbourne measurements in the central and eastern Mediterranean Sea were combined with a Lagrangian particle dispersion model. 2. Methods 2.1. Sampling and chemical analysis The gas- and particulate-phase air samples of selected POPs and meteorological parameters were collected during the RV Urania cruise, 27 August e 12 September 2010 (Fig. 1). In total 15 samples (8e36 h; 230e1060 m3 of air, see Supplementary Material (SM), Table S1) were collected with high volume (HV) samplers (Digitel), containing for the gaseous phase one quartz fibre filter (Whatman) and for the particulate phase two polyurethane foam (PUF) plugs (Molitan, density 0.030 g cm3, 110 mm diameter, 50 mm depth, cleaned by extraction in acetone and dichloromethane, 8 h each, placed in a glass cartridge). Samples extraction, fractionation for the analysis of polycyclic aromatic hydrocarbons separate from halogenated compounds, and identification and quantification by gas chromatography coupled to mass spectrometry is similar to previous studies  et al., 2009) and explained in detail in the SM, S1.2. (Kl anova The targeted compounds include seven PCB congeners (28, 52, 101, 118, 153, 138, 180), ten PBDE congeners (28, 47, 66, 100, 99, 85, 154, 153, 183, 209), PeCB, HCB, and DDT and its metabolites (o,p0 DDE, p,p0 -DDE, o,p0 -DDD, p,p0 -DDD, o,p0 -DDT, p,p0 -DDT). The limit of quantification (LOQ) is calculated as the mean of field blanks plus three times the standard deviation. For PBDEs in the particulate phase too few field blanks (2) were available. Therefore, field blanks were used from a campaign using the identical sampling and analysis methods conducted 2 years later in the same region (BDE 47, 99, 100, 209). For other BDE congeners, in lack of field blanks LOD/2 was considered as the field blank and the relative standard deviation of BDE 100 was used to derive LOQ. PM mass fractions were determined from filter gravimetry (Sprovieri, personal communication). 2.2. Analysis of long-range atmospheric transport The Lagrangian particle dispersion model FLEXPART (Stohl et al., 1998, 2005) is used to identify air mass origins, pathways and map residence times. The meteorological data (0.5  0.5 resolution, 3hourly) is used from the Global Forecast System (GFS, NCEP, USA). For every sampe (8e36 h duration), particles were released every simulated 6 h (100,000 particles were released during 2 h, randomly between 0 and 100 m altitude) at the ship's position. The particles were followed for 5 days backward in time within a large region (30-60 N/10 W-40 E; see Fig. 1b). The model output was a measure of the concentration-weighted time that particles resided in grid cells. To this end, residence times of particles in individual grid cells were added up. For 2D plots particles in grid cells were added up in one direction, vertically (Fig. 1b and Fig. 3a, Fig. S1) or zonally (Fig. 3b). Transport times between locations and boundary layer heights were calculated using trajectories generated by the HYSPLIT model (Draxler and Rolph, 2003). Back trajectories were run every 2 h. For boundary layer heights the average during 4 h preceding arrival of back-trajectories was used. Furthermore, synoptic weather charts were used for transport analysis. The daily time series fire-related PM2.5 emissions as provided by

M.D. Mulder et al. / Atmospheric Environment 111 (2015) 51e59

53

Fig. 1. Spatial coverage of air samples (blue and green, sample numbers shown) (a) and mapped concentration-weighted residence times (s) of air masses during the 5 d period prior to arrival at the ship position (b). for individual samples shown in the SM, Fig. S1. Projection of sum over 15 km, 0.5  0.5 grid cells over the total measurement period. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

the Global Fire Assimilation System (GFASv1.0; Kaiser et al., 2012) were interpolated to hourly values using a Gaussian curve according to Kaiser et al., 2009. These were used to produce the estimated contribution of fires to the sampled PAH levels and profiles by multiplication with the residence times obtained from FLEXPART (Sections 3.3.3, S2.3). 2.3. Models of gas-particle partitioning The PCB data is used to test three gas-particle partitioning models, i.e. an adsorption model and two absorption models. The models are described in the SM, S1.3. 3. Results and discussion 3.1. Concentrations and phase partitioning Concentrations and particulate mass fractions are listed in Table 1, except for BDE 66, which was not detected in any of the samples, and BDE 183, which was excluded because its concentration exceeded LOQ in only 6 out of 15 samples. Concentrations of the PAHs were recently published elsewhere (Mulder et al., 2014). The pollutants were found to be significantly correlated with each other (r ¼ 0.57e0.95, t-test, p > 0.05), except for the pair HCB and PBDEs (r ¼ 0.41). The decabrominated ether (BDE 209) accounted for 37% and congeners related to pentaBDE for 62% of the PBDEs. These percentages are comparable to what was found in bulk atmospheric deposition, namely 25% (15e35) related to pentaBDE and 73% (63e83) to decaBDE at mountain sites in western and central Europe 2004e07 (Arellano et al., 2014), and 35% related to pentaBDE and 48% to decaBDE in rural northern Italy 2005 (Vivens et al., 2007). The levels of total (¼ gaseous þ particulate) concentrations were similar throughout the cruise, for most pollutant classes within one standard deviation of the campaign's mean concentration. Only the samples #8, #13 and #15 exceeded one standard deviation of the mean (Fig. 2). Sample #8 represents the maximum concentrations of the cruise for most substances (most PCBs, HCB, PeCB, high concentrations of DDX and the highest S8 PBDEs). Air mass analyses revealed that this sample captured air that previously had resided over the Aliaga-Izmir industrial area (western Turkey). Sample #15 picked up pollution in the vicinity of the harbour of Palermo (Sicily, Italy; see SM, Fig. S1c). Moreover, these three samples were taken at

Table 1 Concentrations in air, ctot ¼ cg þ cp (pg m3) and particulate mass fraction, q, of (a.) PCBs, (b.) OCPs (PeCB, HCB, DDX) and (c.) PBDE as time-weighted mean (minemax). nLOQ ¼ number of samples with both cg > LOQ and cp > LOQ (out of 15). For calculation of the mean, values < LOQ were ignored.

a. PCB 28 PCB 52 PCB 101 PCB 118 PCB 153 PCB 138 PCB 180 S7PCBs b. PeCB HCB o,p0 -DDE p,p0 -DDE o,p0 -DDD p,p0 -DDD o,p0 -DDT p,p0 -DDT SOCPs c. BDE 28 BDE 47 BDE 100 BDE 99 BDE 85 BDE 154 BDE 153 BDE 209 S8PBDEs

nLOQ

ctot (mean (minemax))

15 15 15 15 15 15 15

0.58 0.69 0.58 0.32 0.57 0.48 0.32 3.61

5 15 15 14 15 15 15 15

0.22 (<0.39e2.80) 6.29 (2.48e24.16) 0.12 (0.028e0.32) 1.06 (<0.35e3.70) 0.090 (0.043e0.25) 0.18 (0.061e0.53) 0.39 (0.13e1.07) 0.76 (0.19e1.72) 9.11 (2.94e31.48)

15 15 15 15 11 15 15 15

0.088 (0.016e0.46) 2.82 (0.57e7.49) 0.28 (0.081e1.10) 1.21 (0.36e4.73) 0.016 (<0.0085e0.065) 0.050 (0.019e0.21) 0.047 (0.017e0.14) 2.79 (0.717e9.11) 7.42 (2.87e20.056)

(0.17e1.63) (0.40e1.99) (0.33e1.23) (0.13e0.54) (0.29e1.18) (0.22e0.89) (0.15e0.55) (2.08e7.72)

q (mean) 0.06 0.03 0.02 0.10 0.05 0.07 0.14

0.01 0.07 0.04 0.03 0.17 0.10 0.05 0.13

<0.01 0.04 0.12 0.19 0.33 0.34 0.29 0.13

night, when a low wind speed and shallow planetary boundary layer (on average 550, 370 and 430 m, during sample #8, #13 and #15, respectively) caused accumulation of pollutants (e.g., Salmond and McKendry, 2005; Gasi c et al., 2010). In case of sample #13 mixing seems to be the only explanation. Few of the fifteen samples have been collected in areas, which had been studied before, such that concentrations can be compared with previous measurements. The concentrations of most compounds found in this study are remarkably low. Samples #3e4 were taken near the harbour of Taranto, for which Estellano et al., 2014, reported a factor ten higher PCB concentrations in summer 2009 for the same seven congeners as reported here (38 pg m3 vs.

54

M.D. Mulder et al. / Atmospheric Environment 111 (2015) 51e59

Fig. 2. Time series of samples during the cruise. Top: temperature ( C), wind speed (m s1), PM10 (mg m3) and air mass origins (green colour coding; see SM, Table S4); bottom: total concentration (gaseous þ particulate) of 25 PAHs (ng m3), 7 PCBs, 8 PBDEs, DDX and HCB (pg m3), with possible urban influence (Urban, grey). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.7 pg m3 and 3.5 pg m3 in our data set). A possible explanation for the low concentrations measured end of August 2010 is that most of the industrial processes restarted later, early September (i.e., after the end of the summer break, as being common practice in the region). The PCB concentrations of samples #7e9, S7PCB ¼ 3.05, 7.72 and 4.49, respectively, are lower than found in the central and southern Aegean in July 2000 (background site near Athens 19 pg m3 of S7PCB except PCB180; Mandalakis et al., 2002) and lower than measured in 1999e2001 at Finokalia, (13.29 pg m3 for S7PCB except PCB52; Mandalakis and Stephanou, 2002). Iacovidou et al., 2009, also reports PCB concentrations for Finokalia in summer 2006, namely S5PCB ¼ 8.4 pg m3 (S7PCB except PCB28 and PCB138). In a heavily polluted area, Aliaga, Turkey, Kaya et al., 2012, found S41PCB ¼ 134 pg m3 e 231 ng m3. The DDX concentration observed in May 2003 near Izmir (Sofuoglu et al., 2004) was 20 ± 8, 5 ± 4 and 29 ± 21 pg m3 for p,p0 DDE, p,p0 -DDT and p,p0 -DDD, respectively, i.e., a factor of 10e40 higher than in our data set (0.72, 0.14 and 0.72 pg m3 in sample #7 and 2.00, 0.26, 1.07, respectively, in sample #8). The PBDE concentrations in samples #7e8, taken near Crete, S8BDE ¼ 11.8 pg m3 and 19.9 pg m3, respectively, comparable to earlier observations from the Izmir area i.e., S8BDE ¼ 4.5e63.6 pg m3 at suburban, urban and industrial sites (sampled in September 2004 and June 2005; Cetin and Odabasi, 2008), from urban sites in Athens (22.52 pg m3 in autumn 2006) and Iraklion, Crete (13.53 pg m3 in spring 2007; Mandalakis et al., 2009), and from the background site Finokalia, Crete (summer 2006, S6BDE ¼ 3.3 pg m3; Iacovidou et al., 2009). Even though the concentrations reported in this study are mostly lower than found in earlier studies from the Eastern and Central Mediterranean, a decreasing long-term trend cannot be concluded: Much more data would be needed to discriminate against the uncertainty caused by different sampling protocols, seasonality of emissions and the variation in air mass origins. Nevertheless, emission reductions might have contributed. Furthermore, the PCB secondary emissions from the European continent and sea surfaces are expected to decline since decades (Lammel and Stemmler, 2012). A discussion of the PAH levels can be found in Mulder et al., 2014.

The time-weighted mean (minemax; n ¼ 15) of the particulate mass fraction, q, is 0.05 (<0.01e0.16) for S7PCB, 0.01 (<0.01e0.02) for HCB and PeCB, 0.07 (<0.01e0.20) for DDX compounds and 0.10 (0.04e0.36) for PBDEs. The mean (minemax; n ¼ 15) PM2.5 and PM10 concentrations were 10.9 (1.8e42.5) and 23.2 (9.7e51.4) mg m3, respectively. The correlation of q with PM10 was not significant (r ¼ 0.47e0.48, p < 0.01). For PCBs gas-particle partitioning models tested mostly underpredict partitioning to the particulate phase (Table S3). Due to lack of composition (e.g. fraction of organic matter) and humidity data, and a possible sampling artefact (underestimated q; Mandalakis and Stephanou, 2002) these results are not really conclusive (see discussion in S2.3). Higher brominated BDEs are not expected to be in phase equilibrium as relaxation to disturbances from deposition (or, eventual, re-volatilisation) is slow (Li et al., 2015). Therefore, the particulate mass fraction, q, of BDEs with >4 Br atoms should be lower in air masses recently influenced by emissions. In our data set, BDE209 in samples influenced by urban air (samples #7e9, 15) was indeed less associated with PM (q ¼ 0.11 ± 0.11) than by average over all samples (q ¼ 0.16 ± 0.13). 3.2. Long-range atmospheric transport analysis The large scale advection of the air masses, which influenced the samples was mostly across central Europe, but also partly across western Europe (Fig. 1). 3.2.1. Pathways of air masses and sample attributes By the pathways they took, the air masses can be categorised in 5 groups (Fig. 2, Table S4). These are detailed and residence time maps for individual samples are given in the SM, Fig. S1. According to long-range advection and atmospheric features, we investigated the possible influence of (A) night-time (low mixing eventually leading to high concentrations; samples #8, #13, and #15), (B) PM2.5 and PM10 concentrations, and (C) possible urban influence as indicated by air mass passage over a metropolitan area e.g., Athens (samples #7e9 and #15). The samples with attribute A reflect the highest concentration levels (sum of all measured substances), and are most pronounced for PBDEs. Local sources were suspectedly more important than LRT for A samples

M.D. Mulder et al. / Atmospheric Environment 111 (2015) 51e59

(various air mass categories), as low wind speeds (2.7, 2.8 and 4.8 m s1 for samples #8, 13 and #15, respectively), and a relatively low atmospheric mixing layer prevailed (Gogou et al., 1996). Sample #8 contains the highest pollutant levels with regard to PAHs, PCBs, PBDE, HCB and PeCB (but not DDX) found on the cruise. Sample #15 contains the highest levels of DDX and the concentration levels of the other pollutant groups rank second compared to sample #8. Air mass analyses revealed that the air sampled by sample #8 resided a significant amount of time above a high emission area, the Aliaga industrial area, north of Izmir, western Turkey, more investigated below. Similarly, sample #15 was influenced by the urban (and industrial) areas of Palermo and Rome. The pollutant concentrations of sample #13 are about 1.5 times higher than of the previous and the following sample, #12 and #14. This seems to reflect the mixing layer depth change from z600 m (sample #12) to z400 m (sample #13) and back to z600 m (sample #14). PM (B) levels were not significantly correlated with most pollutants (r  0.44), except for DDX (r ¼ 0.60, t-test, p > 0.05). The low pollutant levels found in sample #10 can be explained by a high wind speed and a large distance to land. The low PM10 level of sample #12 reflects a pathway of the air mass without recent influence from land based emissions (see Fig. S2). The fact that PCB sources are predominantly located in urban areas (C), including in the Mediterranean (Mandalakis et al., 2002; Gasic et al., 2010), is reflected in our data set as two samples (samples #9 and #15) among the three with highest PCB content (samples #8e9 and #15) had been influenced by urban areas as shown by the back trajectories. Moreover, the pathways of air masses that correspond to samples #8 and #15, passed over central Europe, which has been remaining a region of major secondary PCB  et al., 2009). PCB and PBDE sources (Pistocchi, 2008; Dvorska congener patterns in such samples (C) are discussed below (Section 3.3.2). 3.3. Pollution sources influencing the samples collected at sea 3.3.1. Turkish west coast The polluted air collected in sample #8, with the highest concentrations of PAH, PeCB, HCB, PCB, PBDE and among the highest of DDX during the cruise (Fig. 2), and to a lesser extent in samples #7 and #9 had passed over the Izmir and Aliaga industrial area, western Turkey. This area hosts major industrial pollutant sources including a large petroleum refinery and petrochemical complex, shipbreaking/scrap iron dockyards, scrap processing ironesteel plants, scrap storage and classification sites, steel rolling mills, related dense a harbour, heavy road and rail traffic, and a natural gasefired power plant. The majority of the emitted PAHs and PCBs is expected to be deposited in the area (Kaya et al., 2012). The transport model indicates residence times in the Aliaga area within the boundary layer of z2e3 h, z30 h and z5e6 h for the samples #7, #8 and #9, respectively. Fig. 3 shows for sample #8, the five day backward simulation and the vertical (zonally summed) north to south cross section. In particular the right panel shows that before arriving at the ship's location, the air was influenced by emissions in the Aliaga area: The boundary layer depth during these z30 h ranged 280e630 m. The residence time distributions for samples #7 and #9 (see Fig. S2) also show that the air had passed over the Aliaga area, however, mostly at a much higher altitude (2200e4600 m and 3000e4200 m for samples #7 and #9, respectively, i.e., above the boundary layer), such that surface emissions were not picked up. In this study, the residence time analyses proofs to be an indispensable tool, to explain air sample composition and concentration levels. It is concluded that the Izmir-Aliaga area is also a source of PeCB, HCB and PBDE for the

55

region (apart from PAH and PCB, reported earlier; Sofuoglu et al., 2004; Cetin et al., 2007; besides others). Photochemistry during transport influencing sample #8 is studied below, Section 3.3.2. 3.3.2. Photochemistry of PCBs and PBDEs during transport from continental sources to the open sea The gas-phase reaction with the hydroxyl radical and the direct photolysis in the particulate phase may provide sinks to PCBs and € derstro €m et al., PBDEs, respectively (Mandalakis et al., 2003; So 2004):


ð2Þ Dci =Dt ¼  qi ji þ ð1  qi ÞkiOH cOH ci where ci and cOH (molec cm3) are the PCB or PBDE and OH radical 3 1 1 concentrations, k(2) s ) is the PCB or PBDE second iOH (cm molec order gas-phase reaction rate coefficient, ji is the PBDE heterogeneous photoloysis rate coefficient (set to zero for PCBs), qi is the PCB or PBDE particulate mass fraction. By comparison of observed and predicted PCB and PBDE congener concentration ratios, (ci/cj)obsd and (ci/cj)pred, respectively (i and j denoting different congeners), at the ship's position, we test the hypotheses that (a) for PCBs the reaction of the gaseous molecule with the hydroxyl radical was the only sink while direct photolysis is negligible (j set to zero; Mandalakis et al., 2003), and (b) for PBDEs heterogeneous photolysis of BDE209 was the dominant sink. Dry gaseous deposition is assumed to be negligible. This can be expected to hold for PCBs, as these were recently found to be close to phase equilibrium, and hence, are being subject to very low absolute vertical fluxes, at least in the eastern Mediterranean (Berrojalbiz et al., 2014). Simultaneous measurements are not available. Therefore, (ci/cj)obsd is based on previous data from the same season, with the assumption that the observed PCB patterns were conserved in the source areas. This assumption is justified, as the concentration ratios in air are dominated by the ratio of the emission fluxes (which are determined by air and soil temperatures and are similar for the same season) and ratio of the soil burdens (changing very slowly; Lammel and Stemmler, 2012). For PBDEs, we consider heteroge€ derstro € m et al., 2004) and reaction neous photolysis of decaBDE (So of decaBDE and pentaPDE with the OH radical (estimated based on the AOPWIN model; Schenker et al., 2008). PentaBDE's photolysis is neglected, because it is much less efficient as indicated by studies on photolysis in solvents (Eriksson et al., 2004; Wei et al., 2013). To adjust q to its value at travelling altitude we assume a doubling of the PBDE particulate mass fraction per 13 K temperature decrease, similar to what was observed for PAHs (Lammel et al., 2010), as no such data are available for PBDEs. Predicted congener concentration ratios are calculated as:

h

      ð2Þ ci cj pred ¼ ci cj src exp  qi ji þ ð1  qi ÞkiOH cOH  qj jj  i   ð2Þ þ 1  qj kjOH cOH Dt where (ci/cj)src and (ci/cj)pred are the observed and predicted concentration ratios of PCB and PBDE congeners i and j, respectively in the source region and at the ship's position, respectively; and Dt (s) is the daylight travelling time between source area and ship's position. Transport times and altitudes are based on back trajectories (HYSPLIT model). Only daylight times are considered for photochemistry. Climatological input data are used for day-time OH concentration (July, 36 N; Spivakovsky et al., 2000) and temperature at transport altitude (summer 2010 over Athens; Philandras et al., 2014; USSA, 1976). Experimental (PCB28; Anderson and Hites, 1996) or estimated (based on structure-activity relationship for all other congeners, AOPWIN model of the EPI suite; USEPA,

56

M.D. Mulder et al. / Atmospheric Environment 111 (2015) 51e59

Fig. 3. Concentration-weighted residence times of air mass collected in sample #8: (a) mapped 15 km columns of 0.5  0.5 grid cells, and (b) vertical, zonally summed cross section.

2012; Beyer et al., 2003) temperature dependent reaction rate coefficients, k(2) iOH (T), are used. Previously observed PCB data were adopted for the source areas Athens (samples #7) and Aliaga-Izmir (samples #7e8). Such data are not or not available or are insufficient for Rome (sample #15) and for the Po Valley (sample #6). Simultaneously observed PCB data for the source region central Europe (samples #4e6 and #15) were adopted from Ko
setice, Czech Republic (own data, unpublished), a regional background site  et al., 2009). Recent PBDE data were (Holoubek et al., 2007; Dvorska available for Izmir (Cetin and Odabasi, 2008), but not for the other urban areas, which may have influenced the samples. The uncertainty of (ci/cj)pred is dominated by the uncertainty of the transport time, Dt (based on the back-trajectory analyses, specified in Table 2), while DT, DcOH (both functions of Dh) and measurement uncertainties D(ci/cj)src (Dci ¼ ±10% and ±15% for i ¼ PCB and PBDE, respectively, based on the results of side by side active air sampling; own unpublished data) contribute to a lesser extent. After LRT from central Europe and the Athens and Izmir urban areas the observed PCB congener ratios cPCB28/cPCB52, cPCB28/cPCB101 and cPCB28/cPCB153 agree with prediction (Table 2aec), in some cases very well. This confirms the perception that the reaction with OH is dominating the atmospheric lifetime of these PCBs. The exceptions are slight overpredictions of cPCB28/cPCB101 and cPCB28/cPCB153 after transports from central Europe (Table 2b) and cPCB28/cPCB153 after transport from Izmir, and overprediction of cPCB28/cPCB153 after transport from Athens (Table 2c). The latter could be caused by a slight change of (cPCB28/cPCB153)src during the last decade ((cPCB28/ cPCB153)src is based on measurements in the year 2000) related to particularly slow depletion of PCB153 in soils of Athens, or an 13 overestimate of k(2) cm3 molec1 s1 rather PCB153OH (<1.2  10 13 3 1 1 than k(2) ¼ 1.6  10 cm molec s at 298 K; USEPA, PCB153OH 2012) (or a combination of both). As (cPCB28/cPCB153)pred is slightly overpredicted for transports from Izmir (samples #7e8, Table 2c), the latter explanation appears more plausible. After LRT from the Izmir urban area (cBDE209/cBDE100)obsd is in agreement with predicted values (Table 2d). This may suggest that jBDE100 is inefficient in ambient air and that the photolysis rates of BDE209 (when sorbed on sand and sediment) are spanning the reactivity on ambient aerosol surfaces. Degradation of PBDE congeners along transport cannot be constrained any further due to a lack of kinetic data €derstro € m et al., 2004; Schenker et al., 2008; Wei et al., 2013). (So 3.3.3. Emissions from fires and the sea surface RET, which is a marker for soft wood burning (e.g., Ramdahl, 1983), was found to be net-volatilisational along the cruise (Mulder et al., 2014). This could be explained by the combination of

emissions in on-going and previous fires in the region using fire observations from satellite (Mulder et al., 2014). Wildfires in eastern Europe had affected the region until shortly before the cruise (Diapouli et al., 2014), but were not relevant for our samples. The mean time the air travelled over sea before arriving at the ship's location is longest for samples #13, #1, #14 and #12 (88 h, 87 h, 82 h and 80 h, respectively, while the average is 51 h). The fraction of total RET among PAHs is correlated (r ¼ 0.46) and highest for samples #12, #9, #11 and #6 (0.032, 0.028, 0.027 and 0.020, respectively, while it is 0.013 on average). Thus, volatilisation from the sea surface may explain RET enrichment in sample #12. Another potential indicator for RET enrichment, the ratio of total RET over PM2.5, is negatively correlated with travel time over the sea (r ¼ 0.59, p < 0.05) and is highest for three samples #15, #5 and #4 (0.0087, 0.0051 and 0.0042, respectively, while it is 0.0021 on average). The negative correlation can be explained by the large contribution sea-salt emissions make to PM2.5 (Mamane et al., 2008). We mapped the potential influence of fire-related RET on the samples as the product of fire-related PM2.5 mass flux (kg m2 s1) and residence times of air masses (shown in Fig. S3 for the entire measurement period). This product ranges 11.4 (0.2e56.1) g m2 for individual samples (summed over the region, see Fig. S3), with maximum values found for samples #1 (56.1 g m2), #5 (30.1 g m2), #2 (20.1 g m2) and #6 (17.5 g m2). Thus, primary, fire-related emissions may explain RET enrichment in samples #5 and #6. In conclusion, out of 7 samples enriched in RET (among PAHs or in PM2.5), 2 are suggested to be related to primary and 1 to secondary RET emissions (and others, eventually, explained by a combination of primary and secondary emissions). 4. Conclusions The observations indicate that the Izmir-Aliaga area is a source of PeCB, HCB and PBDE for the region, apart from PAHs and PCBs (reported earlier; Sofuoglu et al., 2004; Cetin et al., 2007; besides others). By combining fire distributions and air mass histories, the significance of the source open fires for RET levels could be confirmed for individual samples. Changes of PCB congener patterns along transport routes are in agreement with the perception that the reaction with the OH radical is dominating PCB atmospheric lifetime (Mandalakis et al., 2003), and possibly indicate an overestimation of the second order gas-phase reaction rate coefficient of PCB153 with the OH radical by structure-activity relationship (AOPWIN model; USEPA, 2012). The data availability is insufficient to test PBDE photochemistry along transport by observational data. For many PCB and all of the PBDE congeners

M.D. Mulder et al. / Atmospheric Environment 111 (2015) 51e59

57

Table 2 (2) Photochemically induced changes of selected (a-c) PCB congener total concentration ratios based on cOH (climatological; Spivakovsky et al., 2000), k(2) iOH (T) and kjOH (T) € derstro €m et al., 2004), k(2) (experimental or estimated data; Anderson and Hites, 1996; USEPA, 2012), and (d) PBDE congener total concentration ratios based on jBDE209 (So iOH and k(2) jOH (estimated data; USEPA, 2012; Schenker et al., 2008) along transport from source area ((ci/cj)source) to the ship's position ((ci/cj)pred, (ci/cj)obsd) for selected air masses (corresponding to samples#4e5,#7e8). Daylight transport duration, Dt, and transport altitude, h, of air masses (minemax). Sample#

Dt (h)

h (km)

Source area (cPCB28/cPCB52)src

(cPCB28/cPCB52)pred

(cPCB28/cPCB52)obsda

a. PCB28/PCB52 4 5 7 7 8

6e20 14e36 10e36 14e34 14e16

0.9e2.1 1.0e2.0 0.6e1.4 0.2e1.4 0.3e0.5

Central Europe 0.6e2.4b

0.5e2.3 0.5e2.3 0.7e1.3 0.4e2.4 0.5e2.4

0.9e1.1 0.8e1.0 0.9e1.1

Sample#

Dt (h)

h (km)

Source area (cPCB28/cPCB101)src

(cPCB28/cPCB101)pred

(cPCB28/cPCB101)obsda

b. PCB28/PCB101 4 5 7 7 8

6e20 14e36 10e36 14e34 14e16

0.9e2.1 1.0e2.0 0.6e1.4 0.2e1.4 0.3e0.5

Central Europe 1.7e2.1b

1.6e2.1 1.5e2.0 1.1e1.8 1.6e3.8 1.8e3.7

1.1e1.4 1.0e1.3 1.6e1.9

Sample#

Dt (h)

h (km)

Source area (cPCB28/cPCB153)src

(cPCB28/cPCB153)pred

(cPCB28/cPCB153)obsda

c. PCB28/PCB153 4 5 7 7 8

6e20 14e36 10e36 14e34 14e16

0.9e2.1 1.0e2.0 0.6e1.4 0.2e1.4 0.3e0.5

Central Europe 2.6e3.5b

2.3e3.4 2.3e3.3 3.4e6.0 2.8e8.2 3.9e8.2

1.8e2.2 1.6e2.0 2.0e2.4

Athens 1.2e1.5 Izmir 0.5e2.7d

c

Athens 1.7e2.0c Izmir 2.0e4.1d

Athens 5.5e6.7c Izmir 4.5e9.2d

0.8e1.0

1.5e1.9

2.8e3.4

Sample#

Dt (h)

h (km)

Source area (cBDE209/cBDE100)src

(cBDE209/cBDE100)pred

(cBDE209/cBDE100)obsde

d. BDE209/BDE100 8

14e16

0.3e0.5

Izmir 8.3e11.7f

7.4e10.0g

9.4e13

Range based on measurement uncertainty Dci ¼ ±10%, derived based on the results of side by side sampling (unpublished data). b Kosetice, Czech Republic: range of PCB congener patterns measured 4e5 days before and 1e2 days after passage of air mass (own data, unpublished). c Athens: PCB pattern July 2000, relative standard deviation of ci src, cj src was 35e55% (Mandalakis et al., 2002). d Range spanned by means measured at urban and industrial sites in Aliaga (30 km N of Izmir, gaseous PCB only, mean of 4 seasons 2009e10; Kaya et al., 2012) and at a residential-rural site near Urla in summer 2012 (20 km W of Izmir; A. Sofuoglu, Izmir Institute of Technology, personal communication), relative standard deviation of ci src, cj src 110e230% (residential-rural site near Urla) or even higher (urban and industrial sites in Aliaga). e Range based on measurement uncertainty of Dci ¼ ±15%, derived based on the results of side by side sampling in summer in the Mediterranean (own data, unpublished). f For industrial and urban sites measured in summer 2004, relative standard deviation of ci src, cj src was 25e50% (Cetin and Odabasi, 2008). g €derstro € m et al., 2004) taken as lower and upper estimates of reactivity for the particulate mass fraction, Photolysis rates of BDE209 sorbed to sand and sediment (So respectively. a

occurring in the atmospheric environment, gas-phase rate coefficients of the reaction with the hydroxyl radical, k(2) iOH, are needed. Acknowledgements We thank Francesca Sprovieri, CNR, Rende, and the RV Urania crew for scientific support during the cruise and PM data, and Sabine Eckhardt, NILU, for model training. This research was supported by the Granting Agency of the Czech Republic (GACR project no. P503/11/1230) and by the European Commission (European Structural Funds project no. CZ.1.05/2.1.00/01.0001, CETOCOEN), and it was co-funded by the European Social Fund and the Ministry of Education of the Czech Republic (LM2011028 and LO1214). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.atmosenv.2015.03.044. References Anderson, P.N., Hites, R.A., 1996. Hydroxyl radical reactions: the major pathway for polychlorinated biphenyls from the atmosphere. Environ. Sci. Technol. 30, 1756e1763.  pez, J.F., Rose, N.L., Nickus, U., Thies, H., Stuchlik, E., Arellano, L., Fern andez, P., Lo Camarero, L., Catalan, J., Grimalt, J.O., 2014. Atmospheric deposition of polybromodiphenyl ethers in remote mountain regions of Europe. Atmos. Chem. Phys. 14, 4441e4457. Aydin, Y.M., Kara, M., Dumanoglu, Y., Odabasi, M., Elbir, T., 2014. Source

apportionment of polycyclic aromatic hydrocarbons (PAHs) and polychlorinated biphenyls (PCBs) in ambient air of an industrial region in Turkey. Atmos. Environ. 97, 271e285. nez, J., Mariani, G., Wollgast, J., Hanke, G., Dachs, J., 2014. Berrojalbiz, N., Castro-Jime Atmospheric occurrence, transport and deposition of polychlorinated biphenyls and hexachlorobenzene in the Mediterranean and Black Seas. Atmos. Chem. Phys. 14, 8947e8959. Beyer, A., Wania, F., Gouin, T., Mackay, D., Matthies, M., 2003. Temperature dependence of the characteristic travel distance. Environ. Sci. Technol. 37, 766e771. Bidleman, T.F., Jantunen, L., Falconer, R.L., Barrie, L., Fellin, P., 1995. Decline of hexachlorocyclohexane atmosphere and reversal of air-sea exchange. Geophys. Res. Lett. 22, 219e222. Bozlaker, A., Muezzinoglu, A., Odabasi, M., 2008a. Atmospheric concentrations, dry deposition and air-soil exchange of polycyclic aromatic hydrocarbons (PAHs) in an industrial region in Turkey. J. Hazard Mater. 153, 1093e1102. Bozlaker, A., Odabasi, M., Muezzinoglu, A., 2008b. Dry deposition and soil-air gas exchange of polychlorinated biphenyls (PCBs) in an industrial area. Environ. Pollut. 156, 784e793. Breivik, K., Sweetman, A., Pacyna, J.M., Jones, K.C., 2007. Towards a global historical emission inventory for selected PCB congeners d a mass balance approach. 3. An update. Sci. Total Environ. 377, 296e307. Cetin, B., Odabasi, M., 2008. Atmospheric concentrations and phase partitioning of polybrominated diphenyl ethers (PBDEs) in Izmir, Turkey. Chemosphere 71, 1067e1078. Cetin, B., Yatkin, S., Bayram, A., Odabasi, M., 2007. Ambient concentrations and source apportionment of PCBs and trace elements around an industrial area in Izmir, Turkey. Chemosphere 69, 1267e1277. CLEEN e Chemical Legislation Enforcement Network, 2005. EuroPCB: Inventory PCB Enforcement in Member States. Part 1. Internet: http://www.cleen-europe.eu/ file/download/71/EuroPCB_part_I_final.pdf (last access: 24.10.2014).  hannesson, T., Larsen, P.B., Viluksela, M., 2001. PolDarnerud, P.O., Eriksen, G.S., Jo ybrominated diphenylethers: occurrence, dietary exposure, and toxicology. Environ. Health Perspect. 109, 49e68. Diapouli, E., Popovicheva, O., Kistler, M., Vratolis, S., Persiantseva, N., Timofeev, M.,

58

M.D. Mulder et al. / Atmospheric Environment 111 (2015) 51e59

Kasper-Giebl, A., Eleftheriadis, K., 2014. Physicochemical characterization of aged biomass burning aerosol after long-range transport to Greece from large scale wildfires in Russia and surrounding regions, summer 2010. Atmos. Environ. 86, 393e404. Donateo, A., Gregoris, E., Gambaro, A., Merico, E., Roberto, R., Nocioni, A., Contini, D., 2014. Contribution of harbour activities and ship traffic to PM2.5, particle number concentrations and PAHs in a port city of the Mediterranean Sea (Italy). Environ. Sci. Pollut. Res. 21, 9415e9429. Draxler, R.R., Rolph, G.D., 2003. HYSPLIT (HYbrid Single-particle Lagrangian Integrated Trajectory) Model Access via NOAA ARL READY. NOAA Air Resources Laboratory, Silver Springs, USA. Available from: http://www.arl.noaa.gov/ready/ hysplit4.html. Dvorsk a, A., Lammel, G., Holoubek, I., 2009. Recent trends of persistent organic pollutants in air in Europe - air monitoring in combination with air mass trajectory statistics as a tool to study the effectivity of regional chemical policy. Atmos. Environ. 43, 1280e1287. Duncan, B.N., Bey, I., 2004. A modeling study of the export pathways of pollution from Europe: seasonal and interannual variations (1987e1997). J. Geophys. Res. 109, D08301. €, S., Stohl, A., 2007. Record high peaks in polyEckhardt, S., Breivik, K., Mano chlorinated biphenyl concentrations in the Arctic atmosphere due to longrange transport of biomass burning emissions. Atmos. Chem. Phys. 7, 4527e4536. Eriksson, J., Green, N., Marsh, G., Bergman, A., 2004. Photochemical decomposition of 15 polybrominated diphenyl ether congeners in methanol/water. Environ. Sci. Technol. 38, 3119e3125. Esen, F., Tasdemir, Y., Vardar, N., 2008. Atmospheric concentrations of PAHs, their possible sources and gas-particle partitioning at a residential site of Bursa, Turkey. Atmos. Res. 88, 243e255. Estellano, V.H., Pozo, K., Silibello, C., Mulder, M.D., Efstathiou, C., Tomasino, M.P., Funaro, F., Donadio, I., Focardi, S., 2014. Characterization of urban pollution in two cities of the Puglia region in Southern Italy using field measurements and air quality (AQ) model approach. Atmos. Pollut. Res. 5, 34e41. EU, 2003. Directive 2003/11/EC of the European parliament and of the Council of 6 February 2003 amending for the 24th time council directive 76/769/EEC relating to restrictions on the marketing and use of certain dangerous substances and preparations (pentabromodiphenyl ether, octabromodiphenyl ether). Off. J. Eur. Union L 42 (45). Internet: http://eur-lex.europa.eu/legalcontent/EN/TXT/PDF/?uri¼CELEX:32003L0011&from¼EN (last access: 24.10.2014). EU, 2008. V. Court proceedings, court of justice. Official Journal of the European Union. Off. J. Eur. Union 51. C116/02, 9 May 2008 Internet: http://eur-lex.europa. eu/LexUriServ/LexUriServ.do?uri¼OJ:C:2008:116:0002:0003:EN:PDF. nova , J., Scheringer, M., Ili Gasi c, B., MacLeod, M., Kla c, P., Lammel, G., Pajovi c, A., Breivik, K., Holoubek, I., Hungerbühler, K., 2010. Quantification of sources of PCBs to the atmosphere in urban areas: a comparison of cities in North America, Western Europe and former Yugoslavia. Environ. Pollut. 158, 3230e3235. Gogou, A.N., Stratigakis, M., Kanakidou, E.G., Stephanou, 1996. Organic aerosols in Eastern Mediterranean: components source reconciliation by using molecular markers and atmospheric back trajectories. Org. Geochem 25, 79e96. Holoubek, I., Kl anov a, J., Jarkovský, J., Kohoutek, J., 2007. Trends in background levels of persistent organic pollutants at Kosetice observatory, Czech Republic. Part I. Ambient air and wet deposition 1988e2005. J. Environ. Mon. 9, 557e563. Iacovidou, E., Mandalakis, M., Stephanou, E.G., 2009. Occurrence and diurnal variation of polychlorinated biphenyls and polybrominated diphenyl ethers in the background atmosphere of Eastern Mediterranean. Chemosphere 77, 1161e1167. Kaiser, J.W., Heil, A., Andreae, M.O., Benedetti, A., Chubarova, N., Jones, L., Morcrette, J.J., Razinger, M., Schultz, M.G., Suttie, M., van der Werf, G.R., 2012. Biomass burning emissions estimated with a global fire assimilation system based on observed fire radiative power. Biogeosciences 9, 527e554. Kaiser, J., Flemming, J., Schultz, M.G., Suttie, M., Wooster, M.J., 2009. The MACC Global Fire Assimilation System: First Emission Products (GFASv0). In: ECMWF Techn. Memo. No. 596. European Centre for Medium-range Weather Forecasts, Reading, UK, p. 20. Internet: http://old.ecmwf.int/publications/library/ ecpublications/_pdf/tm/501-600/tm596.pdf (last time accessed: 3 Nov 2014). Kaya, E., Dumanoglu, Y., Kara, M., Altiok, H., Bayram, A., Elbir, T., Odabasi, M., 2012. Spatial and temporal variation and airesoil exchange of atmospheric PAHs and PCBs in an industrial region. Atmos. Pollut. Res. 3, 435e449.
nov , P., Kare
s, R., Tomsej, T., Kla a, J., Cupr, P., Holoubek, I., Bor uvkov a, J., P
ribylova Ocelka, T., 2009. Monitoring of persistent organic pollutants in africa. part 1: passive air sampling across the continent in 2008. J. Environ. Monit. 11, 1952e1963. €pffer, W., Wagner, B.O., 2007. Atmospheric Degradation of Organic Substances: Klo Data for Persistence and Long-range Transport Potential. Wiley-VCH, Weinheim, Germany, p. 2007. Kouvarakis, G., Tsigaridis, K., Kanakidou, M., Mihalopoulos, N., 2000. Temporal variations of surface regional background ozone over Crete Island in the southeast Mediterranean. J. Geophys. Res. 105, 4399e4407. la Guardia, M.J., Hale, R.C., Harvey, E., 2006. Detailed polybrominated diphenyl ether (PBDE) congener composition of the widely used penta-, octa-, and deca-PBDE technical flame-retardant mixtures. Environ. Sci. Technol. 40, 6247e6254. Lakaschus, S., Weber, K., Wania, F., Schrems, O., 2002. The air-sea equilibrium and time trend of hexachlorocyclohexane in the Atlantic ocean between the Arctic and Antarctic. Environ. Sci. Technol. 36, 138e145.

Lammel, G., 2015. Polycyclic aromatic compounds in the atmosphere e a review identifying research needs. Polycycl. Aromat. Compd. 35 http://dx.doi.org/ 10.1080/10406638.2014.931870. Lammel, G., Stemmler, I., 2012. Fractionation and current time trends of PCB congeners: evolvement of distributions 1950e2010 studied using a global atmosphere-ocean general circulation model. Atmos. Chem. Phys. 12, 7199e7213. nova , J., Ili Lammel, G., Kla c, P., Kohoutek, J., Gasi c, B., Kovaci c, I., Laki c, N., Radi c, R., 2010. Polycyclic aromatic hydrocarbons on small spatial and temporal scales e I. Levels and variabilities. Atmos. Environ. 44, 5015e5021. Lelieveld, J., Berresheim, H., Borrmann, S., Crutzen, P.J., Dentener, F.J., Fischer, H., Feichter, J., Flatau, P.J., Heland, J., Holzinger, R., Korrmann, R., Lawrence, M.G., Levin, Z., Markowicz, K.M., Mihalopoulos, N., Minikin, A., Ramanathan, V., de Reus, M., Roelofs, G.J., Scheeren, H.A., Sciare, J., Schlager, H., Schultz, M., Siegmund, P., Steil, B., Stephanou, E.G., Stier, P., Traub, M., Warneke, C., Williams, J., Ziereis, H., 2002. Global air pollution crossroads over the Mediterranean. Science 298, 794e799. Li, Y.F., Ma, W.L., Yang, M., 2015. Prediction of gas/particle partitioning of polybrominated diphenyl ethers (PBDEs) in global air: a theoretical study. Atmos. Chem. Phys. 15, 1669e1681. Mamane, Y., Perrino, C., Yossef, O., Catrambone, M., 2008. Source characterization of fine and coarse particles at the East Mediterranean coast. Atmos. Environ. 42, 6114e6130. Mandalakis, M., Stephanou, E.G., 2002. Study of atmospheric PCB concentrations over the eastern Mediterranean Sea. J. Geophys. Res. 107, 4716. Mandalakis, M., Tsapakis, M., Tsoga, A., Stephanou, E.G., 2002. Gas-particle concentrations and distribution of aliphatic hydrocarbons, PAHs, PCBs and PCDD/Fs in the atmosphere of Athens (Greece). Atmos. Environ. 36, 4023e4035. Mandalakis, M., Berresheim, H., Stephanou, E.G., 2003. Direct evidence for destruction of polychlorinated biphenyls by hydroxyl radicals in the subtropical atmosphere. Environ. Sci. Technol. 37, 542e547. Mandalakis, M., Besis, A., Stephanou, E.G., 2009. Particle size distribution and gasparticle partitioning of atmospheric polybrominated diphenylethers in urban areas of Greece. Environ. Pollut. 157, 1227e1233. Mihalopoulos, N., Stephanou, E., Kanakidou, M., Pilitsidis, S., Bousquet, P., 1997. Tropospheric aerosol ionic composition in the Eastern Mediterranean region. Tellus B 49, 314e326. , J., Kuta, J., Proke
s, R., Sprovieri, F., Mulder, M.D., Heil, A., Kuku
cka, P., Kl anova Lammel, G., 2014. Airesea exchange and gaseparticle partitioning of polycyclic aromatic hydrocarbons in the Mediterranean. Atmos. Chem. Phys. 14, 8905e8915. Philandras, C.M., Nastos, P.T., Kapsomenakis, I.N., Repapis, C.C., 2014. Climatology of upper air temperature in the Eastern Mediterranean region. Atmos. Res. 152, 29e42. Pistocchi, A., 2008. A GIS-based approach for modeling the fate and transport of pollutants in Europe. Environ. Sci. Technol. 42, 3640e3647. Ramdahl, T., 1983. Retene d a molecular marker of wood combustion in ambient air. Nature 306, 580e582. Salmond, J.A., McKendry, I.G., 2005. A review of turbulence in the very stable nocturnal boundary layer and its implications for air quality. Prog. Phys. Geogr. 29, 171e188. Schenker, U., Soltermann, F., Scheringer, M., Hungerbühler, K., 2008. Modeling the environmental fate of polybrominated diphenyl ethers (PBDEs): the importance of photolysis for the formation of lighter PBDEs. Environ. Sci. Technol. 42, 9244e9249. Schuster, J.K., Gioia, R., Sweetman, A.J., Jones, K.C., 2010. Temporal trends and controling factors for PCBs in the UK atmosphere (1991-2008). Environ. Sci. Technol. 44, 8068e8074. €derstro €m, G., Sellstro €m, U., de Wit, C.A., Tyskling, M., 2004. Photolytic debromiSo nation of decabromodiphenylether 209 (BDE209). Environ. Sci. Technol. 38, 127e132. Sofuoglu, A., Cetin, E., Bozacioglu, S.S., Sener, G.D., Odabasi, M., 2004. Short-term variation in ambient concentrations and gas/particle partitioning of organochlorine pesticides in Izmir, Turkey. Atmos. Environ. 38, 4483e4493. Spivakovsky, C.M., Logan, J.A., Montzka, S.A., Balkanski, Y.J., Foreman-Fowler, M., Jones, D.B.A., Horowitz, L.W., Fusco, A.C., Brenninkmeijer, C.A.M., Prather, M.J., Wofsy, S.C., McElroy, M.B., 2000. Three-dimensional climatological distribution of tropospheric OH: update and evaluation. J. Geophys. Res. 105, 8931e8980. Stohl, A., Hitzenberger, M., Wotawa, G., 1998. Validation of the Lagrangian particle dispersion model FLEXPART against large scale tracer experiments. Atmos. Environ. 32, 4245e4264. Stohl, A., Forster, C., Frank, A., Seibert, P., Wotawa, G., 2005. Technical note: the Lagrangian particle dispersion model FLEXPART version 6.2. Atmos. Chem. Phys. 5, 2461e2474. Tsapakis, M., Stephanou, E.G., Karakassis, I., 2003. Evaluation of atmospheric transport as a non-point source of polycyclic aromatic hydrocarbons in marine sediments of the Eastern Mediterranean. Mar. Chem. 80, 283e298. USEPA, 2012. Estimation Programs Interface Suite™ for Microsoft® Windows, v. 4.11. United States Environmental Protection Agency, Washington, DC, USA. USSA, 1976. U.S. Standard Atmosphere 1976. U.S. Government Printing Office, Washington. Internet: http://www.pdas.com/atmos.html (last access: 10 Oct 2014). Vivens, I., Canuti, E., Castro-Jimenez, J., Christoph, E.H., Eisenreich, S.J., Hanke, G., Huber, T., Mariani, G., Mueller, A., Skejo, H., Umlauf, G., Wollgast, J., 2007. Occurrence of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs),

M.D. Mulder et al. / Atmospheric Environment 111 (2015) 51e59 polychlorinated biphenyls (PCBs) and polybrominated diphenyl ethers (PBDEs) in Lake Maggiore (Italy and Switzerland). J. Environ. Mon. 9, 589e598. Wania, F., Daly, G.L., 2002. Estimating the contribution of degradation in air and deposition to the deep sea to the global loss of PCBs. Atmos. Environ. 36,

59

5581e5593. Wei, H., Zou, Y.H., Li, A., Christensen, E.R., Rockne, K.J., 2013. Photolytic debromination pathway of polybrominated diphenyl ethers in hexane by sunlight. Environ. Pollut. 174, 194e200.