Comparative study on PCDD/F pollution in soil from the Antarctic, Arctic and Tibetan Plateau

  • Published on

  • View

  • Download

Embed Size (px)


<ul><li><p>Science of the Total Environment 497498 (2014) 353359</p><p>Contents lists available at ScienceDirect</p><p>Science of the Total Environment</p><p>j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenvComparative study on PCDD/F pollution in soil from the Antarctic, Arcticand Tibetan PlateauShenglan Jia a, Qiang Wang b, Li Li a, Xuekun Fang a,c, Yehong Shi d, Weiguang Xu a, Jianxin Hu a,a State Key Joint Laboratory of Environmental Simulation and Pollution Control, College of Environmental Sciences and Engineering, Collaborative Innovation Center for Regional Envi-ronmental Quality, Peking University, Beijing 100871, PR Chinab School of Chemistry and Chemical Engineering, Central South University, Changsha 410083, PR Chinac Norwegian Institute for Air Research, P.O. Box 100, 2027 Kjeller, Norwayd Beijing General Research Institute of Mining &amp; Metallurgy, Beijing 100160, PR China</p><p>H I G H L I G H T S</p><p> The level of PCDD/Fs in soil at the Three Poles was reported. PCDD/F congener profiles in soil from the Three Poles were compared. Potential local and regional PCDD/F sources were conducted by FLEXPART simulation. Corresponding author. Tel.: +86 10 62756593.E-mail address: (J. Hu).</p><p> 2014 Elsevier B.V. All rights reserved.a b s t r a c ta r t i c l e i n f oArticle history:Received 24 April 2014Received in revised form 21 July 2014Accepted 28 July 2014Available online xxxx</p><p>Editor: Adrian Covaci</p><p>Keywords:PCDD/FsSoilThree PolesFLEXPART modelThe concentrations of polychlorinated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) in 35 soil samplescollected from Fildes Peninsula in the Antarctic, Ny-lesund in the Arctic, and Zhangmu-Nyalam in the TibetanPlateau were reported in this study. A comparison of the total concentration and TEQ of PCDD/Fs at the ThreePoles was conducted. Both the total concentration and TEQ of PCDD/Fs demonstrates a decreasing trend inthe order of Zhangmu-Nyalam (mean: 26.22 pg/g, 0.37 pg I-TEQ/g) N Ny-lesund (mean: 9.97 pg/g,0.33 pg I-TEQ/g) N Fildes Peninsula (mean: 2.18 pg/g, 0.015 pg I-TEQ/g) (p b 0.05). In all samples, the congenerand homologue profiles dominated with higher (seven and eight) chlorinated PCDD/Fs (more than 85% of thetotal mass percentage of PCDD/Fs) at the Three Poles. Finally, a FLEXPART backward simulation was used topreliminarily identify the potential local and regional anthropogenic sources of PCDD/Fs. The results implythat the air masses passing over surrounding regions with significant PCDD/F emissions might contribute tothe occurrence of PCDD/Fs in both the Arctic and Tibetan Plateau.</p><p> 2014 Elsevier B.V. All rights reserved.1. Introduction</p><p>Polychlorinated dibenzo-p-dioxins and polychlorinated dibenzo-furans (PCDD/Fs) are persistent organic pollutants (POPs) regulatedby the Stockholm Convention ( They are uninten-tional byproducts of combustion and chlorinating processes duringboth natural and human activities, such as wildfire and waste incin-eration (USEPA, 2000). Among all 210 congeners, 17 PCDD/Fs withchlorines at the 2, 3, 7, and 8 positions are dioxin-toxic, and pose ahigh risk to ecological systems and human health due to severe ab-normality inductive, mutagenic, and carcinogenic effects (Wikoffet al., 2012). The Antarctic, the Arctic, and the Tibetan Plateau arecollectively referred to as Three Poles of the earth. They share thecommon characteristics of extremely harsh climate and fragile eco-systems. With few anthropogenic activities, the Three Poles wereonce thought to be free from man-made chemical contamination.There have been few reports of PCDD/Fs at the Three Poles, due tothe extreme environment and the expense of sampling. However,the available data suggest that PCDD/Fs occur at the Three Poles:PCDD/Fs in soil (2.484.30 pg/g dry weight) from the eastern TibetanPlateau were reported by Pan et al. (2013); PCDD/Fs were detectedin the blubber of ringed seals from the Arctic (0.6108.2 pg/g lipidweight) (Savinov et al., 2011); and PCDD/Fs in Antarctic penguinswere measured (772 pg/g wet weight) (Corsolini et al., 2007). Aslisted above, most available reports of PCDD/F levels were based onthe monitoring of biota and conducted to assess the ecological risk.However, because abiotic media are the transfer media and main re-ceptors of POPs, they can be used to investigate the potential sourcesand transportation pathways of POPs (Hale et al., 2006). Among thevarious abiotic environmental media, soil is one of the first points</p><p>;domain=pdf image image</p></li><li><p>354 S. Jia et al. / Science of the Total Environment 497498 (2014) 353359of entry of PCDD/Fs into terrestrial ecological systems (Carballeiraet al., 2006). We therefore selected soil to monitor PCDD/Fs in theThree Poles in this study. And the investigation of PCDD/F levels in soilfrom these regions could provide fundamental data for a range of aca-demic applications, such as assessing the local ecological risk of PCDD/Fs, as well as the effects of human disturbances on the environment.</p><p>The identification of the potential PCDD/F sources of the Three Polesis essential for understanding the transport of PCDD/Fs. Previously,many studies have highlighted the importance of global long-rangetransport as the main source of POPs at the Three Poles (AMAP, 1998;Pienitz et al., 2004). For instance, Wania and his colleagues proposedthe theory of global fractionation, which posited that remote regionstended to accumulate POPs through long-range transfer (Wania andMackay, 1993;Wania, 2003). However, due to the increasing human ac-tivities at the Three Poles and the surrounding regions, the contribu-tions of those local and regional sources to the PCDD/F occurrencecan't be overlooked; thus, identifying the localized anthropogenicsources of PCDD/Fs at the Three Poles is needed. Simple backward tra-jectory models were often used to describe the neighboring sources ofPOPs at a simplified way (e.g., detailed atmospheric transport andchemical processes were ignored). The FLEXible PARTicle dispersionmodel (FLEXPART) considered not only atmospheric turbulence andconvection, but also chemical removal of POPs by hydroxyl (OH) radi-cals during the backward simulation. In the previous study, it has beensuccessfully applied in investigating the transport behavior ofpolychlorinated biphenyl (PCB) (Eckhardt et al., 2009).</p><p>The soil samples from Fildes Peninsula in the Antarctic, Ny-lesundin the Arctic and Zhangmu-Nyalam in the Tibetan Plateau were collect-ed in this study. The concentrations of PCDD/Fs and the congener profileat the Three Poles were compared to maximize spatial variation over aglobal geographic range. The potential local and regional anthropogenicsources of PCDD/Fs in the sampling sites were then investigated using aFLEXPART simulation.2. Materials and methods</p><p>2.1. Sample collection</p><p>The sampling locations were Fildes Peninsula in the Antarctic, Ny-lesund in the Arctic, and Zhangmu-Nyalam in the Tibetan Plateau(Fig. 1).Fig. 1. Locations of sampling sites in (a) Ny-lesund,Fildes Peninsula is located onKingGeorge Island-the largest island ofthe South Shetland Islands and hosts nine scientific stations within theAntarctic Treaty System. The local temperature ranges from 20.3 to8.5 C with an average of 1.5 C. From December 2007 to January2008, fifteen topsoil (upper 5 cm) samples (62 S, 58 W) werecollected.</p><p>Ny-lesund is the northernmost permanent settlement, and isinhabited by ca. 30 scientists and technicians. The region is influencedby polar cyclones and the North Atlantic Ocean circulation. The aver-age annual temperature is 5.1 C. From July to August in 2008, wecollected topsoil samples from 20 sampling sites (78 N, 12 E).</p><p>There are approximately 6000 residents of the Zhangmu-Nyalamregion (28 N, 86 E), which includes both a warm and wet subtropicaltemperate zone, and a cold frigid zone. In September 2011, ninesampling sites were established from Zhangmu (located 2636 mabove sea level, on the south side of the Himalayas, with an annual av-erage temperature of 18 C) to Nyalam (located 5129m above sea level,on the north side of theHimalayas, with an average annual temperatureof 2.1 C, and a range of 19.1 to 22.1 C).2.2. Sample extraction/cleanup</p><p>The soil samples were kept in aluminum foil and stored at 20 Cbefore extraction. Twenty-gram samples were freeze-dried, ground,passed through 80-mesh filters, and spiked with labeled standards asa recovery standard (13C-PCDD/Fs, Wellington Laboratories, Canada).They were then extracted five times by pressurized liquid extractionusing toluene at 150 C and 1500 psi (ASE-300, Dionex, USA). Thecrude extracts were concentrated by rotary evaporation, and thencleaned up with multi-step procedures, which included acid washing(98% H2SO4), multi-layer silica gel (from top to bottom in turn: 6 gNa2SO4, 0.9 g silica gel, 3 g basic silica gel with 3% KOH, 4.5 g acid silicagel with 44% H2SO4, 6 g acid silica gel with 22% H2SO4, 0.9 g silica geland 3 g silica gel with 10% AgNO3), gel permeation chromatography(Bio-Beads SX3, Bio-Rad, USA), and an activated carbondiatomitecolumn. The final volume of the samples was adjusted to 20 L byadding the internal standards (13C-1,2,3,4-TeCDD, Wellington Labo-ratories, Canada) before instrumental analysis.</p><p>The reagents used in this study were pesticide-grade toluene,acetone, methylene chloride, decane, and hexane. Decane was ob-tained from Wako Pure Chemical Industries, Ltd. (Osaka, Japan),(b) Fildes Peninsula, and (c) Zhangmu-Nyalam.</p><p>image of Fig.1</p></li><li><p>355S. Jia et al. / Science of the Total Environment 497498 (2014) 353359and the other reagents were purchased from Thermo Fisher Scientific(Waltham, MA, USA).</p><p>A high-resolution gas chromatography (HP6890, Agilent Tech-nologies, USA) coupled with a high-resolution mass spectrometer(JMS 700D, JEOL, Tokyo, Japan) was used, and was equipped with aDB-5MS column (60 m 0.25 mm I.D., 0.25-m thickness: J &amp; WScientific, Folsom, CA, USA) for the determination of TeCDD/Fs toHxCDD/Fs [the inlet temperature was 260 C. The column temperaturewas programmed from120 C (1min) to 200 C at 20 C/min, 2 C/minto 260 C (10 min)]. A DB-17MS column (15m 0.32mm I.D., 0.25-mthickness: J &amp;WScientific)was used for the determination of HpCDD/Fsand OCDD/Fs [the inlet temperature was 280 C and the column tem-perature program: 100 C (1 min), 20 C/min to 280 C (10 min)].The flow rate was 1.2 mL/min, and the injection volume was 2 L. MSoperating conditionswere as follows: ionization energy, 38 eV; ion cur-rent, 600 A; accelerating voltage, 7.0 kV; detection mode, single ionmonitoring (SIM); and mass resolution, 10,000.</p><p>2.3. QA/QC</p><p>One field blank and one method blank were analyzed for each batchof samples, andno background level higher than limit of detection (LOD)was recorded. The LODs of the congeners ranged from 0.02 to 0.40 pg/g.The recovery efficiencies ranged from 54.47% to 104.35%, which waswithin the requirement of US EPA Method 1613B (EPA, 1994).</p><p>2.4. Data analysis</p><p>Prior to statistical analysis, a Levene analysis (p N 0.05) was used todetermine the homogeneity of variance of the dataset. KruskalWallisH-tests were then used to evaluate datasets that did not conform to anormal distribution (p N 0.05). All statistical analyses were conductedusing IBM SPSS v-20.0. A value of p b 0.05 was taken to indicatestatistical significance.</p><p>The Lagrangian particle dispersion model FLEXPART v-9.02 (Stohlet al., 1998, 2005; was used to establish theatmospheric sourcereceptor (sampling areas) relationships. As the po-tential emission sources outnumbered the sampling areas (receptors),backward simulation instead of forward simulation was used inthis study. The backward model calculations were detailed by previousstudies (Stohl et al., 2003; Seibert and Frank, 2004). Briefly, thebackward simulation was driven with operational three-hourlymeteorological data of 1 1 horizontal resolution and 91 verticallevels, from the European Center for Medium Range Weather forecast(ECMWF). Forty thousand particles (not real particle maters, butso-called ideal tracer parcels) were released every 3 h at ca. 20 mabove model ground at each sampling site and traced backward fora twenty-day time to calculate the emission sensitivity (ES). Eachgrid cell has an ES value (in units of ns/kg). Providing a unit sourcestrength (1 kg/s) in each grid cell, the ES value equals to the simulatedconcentration that the cell would produce at the receptor. The higherthe ES value, the greater the contribution of the unit emission fluxdensity from the source to the concentration at the receptor.</p><p>FLEXPART is suitable for simulating atmospheric transport of POPsthat are not strongly adsorbed on airborne particles (log KOA b 11 andlog KAW N 5) (Meyer et al., 2005; Eckhardt et al., 2009). PCDD/Fs arepresent in both the gas phase and particle phase in the atmosphere.The four to five chlorinated PCDD/Fs are present mainly in the gasphase (N50%), and have the above properties (i.e., log KOA b 11 andlog KAW N5). However, six to eight chlorinated congeners are mainlyin the particle phase (N50%) and have log KOA N 11 (Wagrowski andHites, 1998; Harner et al., 2000). Besides, for six to eight chlorinatedcongeners, primary emission sources are not the only effective sourcesto the atmosphere, and the secondary sources (six to eight chlorinatedcongeners contaminated soils, water, etc.) might also influencethe measurement locations. Thus, the FLEXPART simulation of thesourcereceptor relationships is applicable for four to five chlorinat-ed PCDD/Fs, but not six to eight chlorinated PCDD/Fs. However, mostanthropogenic PCDD/F sources emit both four to five and six to eightchlorinated congeners (Table 2) (EPA, 1998); the identified local and re-gional emission areas for four to five chlorinated PCDD/Fs could be sim-ilar to primary emission sources for six to eight chlorinated PCDD/Fs.</p><p>In this study, an air tracer was used as the released parcel for thisbackward simulation. The outputs of the FLEXPART backward simula-tion were 0.5 0.5 3-h ES maps for each sampling site. And the 3-hESs for each site were averaged over 1 year in this study. It should benoted that emission contributions were not calculated (multiplicationof the footprint ES with the emission flux densities derived from anemission inventory) because of the lack of appropriate emission profileinventories. Therefore, the ESmaps covered only the preliminary poten-tial emission sources that were identified. These ESs were derived bysimulating the atmospheric tracer concentrations above the ground ateach sampling site, instead of the concen...</p></li></ul>


View more >