Environ. Sci. Technol. 2009, 43, 796–803
Seasonally Resolved Concentrations of Persistent Organic Pollutants in the Global Atmosphere from the First Year of the GAPS Study K A R L A P O Z O , † T O M H A R N E R , * ,† SUM CHI LEE,† FRANK WANIA,‡ DEREK C. G. MUIR,§ AND KEVIN C. JONES| Atmospheric Science & Technology Directorate, Environment Canada, 4905 Dufferin Street, Toronto, Ontario, Canada M3H 5T4, Department of Physical and Environmental Sciences, University of Toronto Scarborough, 1265 Military Trail, Scarborough, Ontario, Canada, M1C 1A4, Water Science & Technology Directorate, Environment Canada, Burlington, Ontario, Canada L7R 4A6, and Centre for Chemicals Management, Lancaster Environment Centre, Lancaster University, Lancaster, UK LA1 4YQ
Received July 28, 2008. Revised manuscript received October 21, 2008. Accepted October 22, 2008.
Concentrations of persistent organic pollutants (POPs) in air are reported from the first full year of the Global Atmospheric Passive Sampling (GAPS) Network. Passive air samplers composed of polyurethane foam disks (PUF-disk samplers) were deployed over four consecutive three-month periods in 2005 to measure seasonal concentrations of POPs at a variety of site types on a global scale, with an emphasis on background/ remote locations. Samples for the last three quarters are reported here for the first time. Annual geometric mean (GM) concentrations in air (pg · m-3) were highest for endosulfan, a currently used pesticide (GM ) 82), and polychlorinated biphenyls (PCBs) (GM ) 26). Other chemicals regularly detected included R- and γ-hexachlorocyclohexane (HCH), chlordanes, heptachlor, heptachlor epoxide, dieldrin, p,p’-DDE and polybrominated diphenyl ethers (PBDEs). With the exception of lower concentrations during the first quarter, no seasonal patterns were observed on a global basis. In contrast, some distinct seasonal patterns were observed on a site-specific basis. For instance, endosulfans exhibited strong seasonality with highest concentrations during the summer periods, especially at or near agricultural sites. The latitudinal distribution of target chemicals reflected the estimated spatial variability of global emissions, with highest concentrations observed in the midlatitudes of the northern hemisphere. In the case of PCBs, the GAPS data reflected and were well correlated with global emission estimates, with highest concentrations in developed and industrialized regions. Data provided through the GAPS Network establish global baseline values, and * Corresponding author fax: 416-739-4281; e-mail: tom.harner@ ec.gc.ca. † Atmospheric Science & Technology Directorate, Environment Canada. ‡ University of Toronto Scarborough. § Water Science & Technology Directorate, Environment Canada. | Lancaster University. 796
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 3, 2009
continuation of the time series will contribute to the effectiveness evaluation of global treaties on POPs (e.g., Stockholm Convention). Globally resolved data will also foster the development and validation of global transport models for POPs, and the investigation of seasonal and interannual trends in concentrations of POPs in the global atmosphere.
Introduction Persistent organic pollutants (POPs) are characterized by their toxicity, potential to bioaccumulate, environmental persistence, and ability to be distributed globally, often accumulating in remote and sensitive environments (1). The Stockholm Convention on POPs, under the United Nations Environment Programme (UNEP) (2), identified an initial set of these chemicalss“the dirty dozen”sto be taken out of circulation worldwide. The dirty dozen includes nine organochlorine pesticides (OCPs), and some industrial compounds such as the polychlorinated biphenyls (PCBs), dibenzodioxins, and dibenzofurans. There also exists a process for adding other chemicals that exhibit POP-like behavior, such as the polybrominated diphenyl ethers (PBDEs) (3). To achieve the objectives of the Stockholm Convention, member countries are obliged to implement control measures to eliminate release of POPs to the environment. A plan to measure the “effectiveness” of these control measures (as stipulated under Article 16 of the Convention) relies on a global monitoring program of key environmental media: initially air and human tissues. Rapid response makes the atmosphere a suitable sentinel for changes in the global emissions of POPs. Measurements of POPs in air will be useful in assessing the regional and global transport of POPs. However, meeting these objectives is challenged by the scarcity of spatially resolved data and longterm air monitoring programs for POPs. In some regions there exists no information at all on concentrations of the dirty dozen chemicals in air. This lack can largely be attributed to the high cost and technical requirements of conventional high-volume air sampling and instrumental analysis for POPs. The Global Atmospheric Passive Sampling (GAPS) Network addresses and resolves these issues. Initiated in December 2004 as a pilot study at more than 50 sites, GAPS employs simple sampling devices that are inexpensive, small, and free of electricity requirements (4). Passive air samplers (PAS) have been used in numerous spatial studies at local, regional, and continental scales for monitoring of POPs (4-8). The samplers consist of a polyurethane foam (PUF) disk housed in a stainless steel chamber (4-7). Simplicity and cost-effectiveness are the main reasons for the success of this sampling approach. GAPS relies on the support of international participants who volunteer their time to deploy samplers at global sites. In many instances, existing monitoring stations and operators are usedssuch as those affiliated with the Global Atmospheric Watch program of the World Meteorological Organization, the European Monitoring and Assessment Program, and the Arctic Monitoring and Assessment Program, as well as the National Oceanic and Atmospheric Administration and the Geological Survey of the United States. The small size of the sampler and sampling medium allows for inexpensive transport by courier or mail. The samplers are easily mounted and require changing of the sample media only 4 times each year. GAPS is coordinated through the Hazardous Air Pollutants Laboratory, Environment Canada, in Toronto, Canada, where all samples are prepared and received after deployment for analysis of POPs. The GAPS Network has contributed, and continues to 10.1021/es802106a CCC: $40.75
2009 American Chemical Society
Published on Web 12/10/2008
FIGURE 1. Map of sampling sites operated under GAPS in 2005. Schematic and photograph of the “flying saucer” PUF-disk sampler deployed at Bukit Kototabang, Indonesia.
contribute, global-scale data to investigate the regional and global transport of POPs. With continued monitoring, the data may be useful for investigating temporal trends (to observe expected decreases in POPs concentrations in air) and to assess the effectiveness of international control strategies. In this article, results from the first full year of sampling under GAPS, over four consecutive sampling periods in 2005, are presented.
Materials and Methods Sample Collection. The PUF disk sampler consists of two stainless steel domes (approximately 30 cm in diameter), resembling a flying saucer (Figure 1). The spacing between the upper and lower chambers allows air to enter the device and flow over the foam disk while providing protection from high winds, direct precipitation, sunlight, and coarse particle deposition. Polyurethane foam is the same material used in many conventional high-volume, pumped samplers and has a relatively high retention capacity for gas-phase POPs. Prior to exposure, PUF disks were precleaned by Soxhlet extraction for 24 h using acetone and then for another 24 h using petroleum ether (4). Before they were sent to the study sites in amber glass jars, PUF disks were fortified with depuration compounds to assess site-specific sampling rates. These covered a wide range of volatility and included d6γ-HCH, and polychlorinated biphenyl (PCB) congeners 3, 9, 15, 30, 107, and 198, as described elsewhere (6). More details of sample extraction, clean up, recovery tests, and calculations based on depuration compounds are presented elsewhere (6). PUF disks were changed every 3 months over the period December 2004 to December 2005. GAPS sites are mostly in regional/continental background locations, away from local emissions of OCPs and other POPs. Some sites are in urban and agricultural regions to allow comparison of the levels of the target compounds in remote areas with those in typical source regions. The samplers were mounted at least 1.5 m above ground level and detailed, illustrated protocols, including tools required for changing samplers were sent to each site. Sites operated under GAPS study in 2005 are shown in Figure 1 and site-specific information is presented in the Supporting Information (SI) Table S1. One field blank was deployed at each site by inserting and removing a PUF disk from the sampler and then storing and treating it as a sample.
Deriving Concentrations in Air. The uptake of POPs by PUF disks has been previously characterized (4-6). The samplers trap mainly gas-phase contaminants (5, 6) with an air sampling rate of approximately 3-4 m3 day-1, equivalent to ∼300 m3 over a 3-month period. Site-specific air sampling rates are determined by adding known amounts of depuration compounds to the PUF disks prior to deployment. These are either isotopically labeled compounds or other chemicals not present in the atmosphere. By analyzing the loss of these chemicals during the deployment period it is possible to estimate the average air-side controlled sampling rate which is similar for all POPs. This approach is described in detail by Pozo et al. 4, 6 and Gouin et al. (9). Chemical Analysis. PUF disk extracts were analyzed for a suite of target compounds that included OCPs, PCBs, and PBDEs. Samples were screened for 48 PCB congeners, 17 PBDEs, and 19 OCPs: R-, β-, γ-, δ-HCHs, aldrin, heptachlor, heptachlor epoxide, cis-chlordane, trans-chlordane, transnonachlor, endosulfan 1, endosulfan 2, endosulfan sulfate, o,p’-DDE, p,p’-DDE, o,p’-DDD, p,p’-DDD, o,p’-DDT, p,p’DDT (Ultra Scientific, North Kingstown, RI). Results are only reported for OCPs that were detected consistently. PCB concentrations are reported as the sum of the 48 congeners (∑48) (6). Analysis of PUF disk extracts was carried out by gas chromatography-mass spectrometry (GC-MS) on a HewlettPackard 6890 GC-5973 MS for PCBs, using electron impact (EI), and for OCPs and PBDEs by negative chemical ionization (NCI). Conditions for NCI and EI analysis and selection of target/qualifier ions are described elsewhere (6). Limit of detection (LOD) was defined as the average field blank (n ) 30) plus three standard deviations (SD). When target compounds were not detected in blanks, the 1/2 instrumental detection limit (IDL) value was substituted for LOD. For data values that fell below the LOD, 1/2 LOD was used for calculating means. All qualified data (i.e., those exceeding the LOD) were blank corrected.
Results and Discussion Quality Assurance/Quality Control. Method recoveries for PCBs and OCPs were generally >85%. For PBDEs, recoveries had previously been assessed in the same laboratory (4-6). Surrogate recoveries of (94 ( 15)%, (80 ( 22)%, and (70 ( 20)% for 13C-PCB-105, d6-R-HCH, and d8-p,p’-DDT, respectively, were used to confirm analytical integrity. Sample values VOL. 43, NO. 3, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
797
are not corrected for recovery. The instrumental detection limits (IDL) were determined by assessing the injection amount that corresponded to a signal/noise value of 3:1. For PCBs, IDL values ranged from 0.025 pg for dichlorobiphenyls to 0.31 pg for octachlorobiphenyls. For OCPs, IDLs ranged from 0.02 to 0.99 pg. Results were reported only if the signal exceeded three times the baseline noise. Blank levels were assessed from 30 field and 39 laboratory blanks. For the OCPs, blanks were often less than the IDL, with the exception of those for R-HCH, TC, and endosulfan 1; their blank values were also low, however, and ranged from 0.4 to 3% of the sample amounts. Blank levels for p,p’DDE and endosulfan 2 were typically below the IDL. PCBs congeners that are typically found in air have not been detected in field blanks with the exception of di- (PCB-8), tri(PCB-37), penta- (PCB-87, -123), hexa- (-137), and heptachlorinated (PCB-170) congeners. Solvent (method) blank values for individual PCB congeners were low and not detectable for most higher molecular weight PCBs. The LODs for PBDEs were relatively high because field blank results were often of the same magnitude as sample values. PBDE results for congeners 47, 99, and 100 are shown in Figure S1 in the Supporting Information. Many of the PBDE data did not exceed the LOD and were reported as below detection limit (BDL) in Table S1. IDLs for PBDEs were (ng)
