P1020240

(a) Arctic sampling sites    (b) Antarctic sampling sites

FIGURĘ 1. Locations of passive air sampling sites at the Korean polar research stations: (a) Dasan station (78 55'N, 11 56'E) at Ny-Alesund, Spitsbergen, Norway and (b) King Sejong station (62'13'S, 58 47'W) on the Barton Peninsula, King George Island, Antarctica.

Accordingly, POPs are believed to remain within the linear uptake phase of the PAS during a one-year deployment, and air concentrations (pg*m~³) can be estimated semiąuanti-tatively from the amounts of Chemicals seąuestrated by the PAS (pg*PAS^-1) usinga sampling ratę of 0.52 m³• day¹ • PAS^-1 (73). This PAS has been used for year-round monitoring of PCBs and OCPs at various sites, including sonie in the Canadian Arctic [17,20,25). Prior to PAS deployment, XAD resin was Soxhlet extracted with acetone for 24 h, followed by two extractions with toluene. The XAD resin (60 mL for each PAS) was dried in a clean desiccator and transferred to a precleaned stainless Steel mesh Container and sealed in an airtight stainless Steel tubę with Teflon stoppers. These PAS columns were stored at -4 °C until being shipped to sampling sites. Details on sampler preparation and sampling proce-dures can be found elsewhere (73, 25).

Sampling Sites. "Dasan”, operating sińce April 2002 and located in Ny-Alesund on the high Arctic island of Spitsbergen, Svalbard, Norway (78°55'N, 11°56'E), is part of an interna-tional research station involving France, Germany, Italy, Japan, Korea, Norway, and the United Kingdom. "King Sejong”, located on the Barton Peninsula, King George Island, Antarctica (62°13'S, 58°47'W), has been operating as a permanent research station sińce February 1988. These two stations are supported by the Korea Polar Research Institute. The PAS were deployed for one year at three sites (A, B, C) in Ny-Alesund (August 7, 2005-August 1, 2006) and three sites (D, E, F) on King George Island (December 18, 2004-December5,2005). In orderto investigatetheinflucnce of local pollution, the sampling sites varied in terms of their distances from the main buildings (Figurę 1). After one year of deployment, theXAD-filIed mesh cylinders were retrieved using airtight stainless Steel tubes with Teflon stoppers and were transported to the laboratory along with field-blank samples (77). The retrieved samples were stored at -4 °C until extraction.

Meteorological Parameters. Hourly data for wind speed, wind direction, and air temperaturę were measured by automated weather Systems operating in the vicinity of the sampling sites. Temporal changes in air temperaturę and wind speed were plotted, and wind-rose diagrams were constructed to find the dominant wind directions and to investigate the relationship between air concentrations and the location of the sampling sites. To identify the origin of air masses arriving at the Arctic and Antarctic stations, five-day baekward trajectories with a startingheightof 50 m were calculated once a day at 0 UTC (Coordinated Universal Time) for the entire sampling period using HYSPLIT 4 (http:// www.arl.noaa.gov/ready/hysplit4.html).

Chemical Analysis. A fuli description of the analytical procedurę is provided in the Supporting Information. In summary, XAD 2-resin samples were analyzed in balches of six, each corisisting of two procedural blanks, a field blank, and three real samples. The resin samples were Soxhlet extracted, and the extracts were cleaned up by silica gel chromatography and activated alumina chromatography. Ali 209 PCB congeners and the OCPs (HCH, aldrin, dieldrin, heptachlor, octachlorostyrene, endosulfan (I and II), en-dosulfan sulfate, chlordane, nonachlor, DDE, DDD, DDT, and mirex) were analyzed by an Agilent-6890 gas chromato-graph coupled to a Micromass Ultima (Micromass, U.K.) higli-resolution mass spectrometer. Recoveries of interna! sur-rogate standards for both PCBs and OCPs were between 65 and 110%, and the accuracy of determining PCBs and OCPs in spiked samples was between 15 and 20%.

Multimedia Environmental Model. The measured PCB data were compared with simulation results obtained with the zonally averaged global multimedia environmen tal model Globo-POP. The Arctic and Antarctic regions are two of 10 climate zones in this model. Each climate zonę is represented by four vertical air layers, different types of soils (cultivated, uncultivated, coniferous and deciduous forest soils), forest canopies (coniferous and deciduous), ffesh water, fresh water sediment, and the surface ocean. A fuli description of the model and a history of modifications can be found elsewhere [26-29). Model-input properties for 11 selected congeners (PCB-8, 28, 31, 52, 101, 105, 118, 138. 153, 180, 194) have been described elsewhere (4). Historical emission data were obtained from the updated global PCB emission imrentory (3). The high emission scenario in ref 3 was used because it was previously found to produce levels and pattems of PCBs in the Arctic atmosphere that are comparable with measured data (4). Whereas the uncertainty of predicted absolute concentrations would be high, predicted congener pattems are believed to be reliable because these patterns are mostly determined by the fairly well established compositions of technical PCB mixtures.

Results and Discussion

Meteorological Conditions. Time series of air temperaturę and wind speed for the sampling periods are depicted in Figurę SI in the Supporting Infonnation. The average temperaturę and wind speed at the Arctic station were 3.8 m/s and -4.3 °C, respectively; those at the Antarctic station were 7.4 m/s and -1.9 °C, indicating relatively mild meteorological conditions at both sites. The daily maximum wind speeds were generally less than 10 m/s at the Arctic station and less than 15 m/s at the Antarctic station. Wind tunnel experiments suggested that wind speeds up to 15 m/s had little effect on the sampling ratę of the XAD resin-based PAS [13).

Wind-rose diagrams and five-day baekward trajectories were considered to evaluate potential pollution within the stations and the influence of LRAT (Figurę 2). Ali monthly trajectory plots are presented in Figurę S2. Southeasterly winds were prevailing at Dasan, while Northern and Northwestern winds were dominant at King Sejong. The wind-rose diagrams in combination with the site map (Figurę 1) suggest that the Antarctic sampling sites are located down-wind from the main building. Those sites (D, E, F) may therefore have been influencetl by emissions from buildings or other potential local sources within the station. This is not the case for the Arctic sites.

The trajectory analyses show the influence of the polar easterlies at Dasan and the prevailing westerlies at King Sejong (Figurę S2). There are no large seasonal differences in trajectory patterns at either station. Considering these patterns, the levels of PCBs at the Arctic station may be considerably affected by emissions from Russia and Northern

7126 ■ ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOl. 42, NO. 19, 2008