Environ. Sci. Technol. 2007, 41, 2688-2695
Hexachlorocyclohexanes in the Canadian Archipelago. 1. Spatial Distribution and Pathways of r-, β-, and γ-HCHs in Surface Water T . F . B I D L E M A N , * ,† H . K Y L I N , ‡,§ L. M. JANTUNEN,† P. A. HELM,| AND R. W. MACDONALD⊥ Centre for Atmospheric Research Experiments, Science and Technology Branch, Environment Canada, 6248 Eighth Line, Egbert, Ontario L0L 1N0, Canada, Norwegian Institute for Air Research, Polar Environmental Centre, NO-9296 Tromsø, Norway, Department of Environmental Assessment, Swedish University of Agricultural Sciences, P.O. Box 7050, SE-705-07, Uppsala, Sweden, Department of Chemical Engineering and Applied Chemistry, University of Toronto, 200 College Street, Toronto, Ontario M5S 3E5, Canada, and Department of Fisheries and Oceans, Institute of Ocean Sciences, 9860 West Saanich Road, Sidney, British Columbia V8L 4B2, Canada
Hexachlorocyclohexanes (HCHs) in the surface water of the Canadian Archipelago and south Beaufort Sea were measured in summer, 1999. Overall concentrations of HCH isomers were in order of abundance: R-HCH (3.5 ( 1.2 ng L-1) > γ-HCH (0.31 ( 0.07 ng L-1) > β-HCH (0.10 ( 0.03 ng L-1). Concentrations and ratios of R-HCH/γ-HCH decreased significantly (p < 0.001 to 0.003) from west to east, but there was no significant variation in R-HCH/ β-HCH. The (+) enantiomer of R-HCH was preferentially degraded, with enantiomer fractions (EFs) ranging from 0.432-0.463 and increasing significantly (p < 0.001) from west to east. Concentrations also varied latitudinally for R-HCH and γ-HCH (p < 0.002) but not for β-HCH. Principal component analysis with variables R-HCH and γ-HCH concentrations, EF, latitude, and longitude accounted for 71% (PC 1) and 16% (PC 2) of the variance. Mixing in the eastern Archipelago was modeled by assuming three end members with characteristic concentrations of R-HCH and γ-HCH. The model accounted for the observed concentrations and higher EFs of R-HCH at the eastern stations.
Introduction Hexachlorocyclohexane (HCH) has been used as an insecticide for over 50 years and has contaminated oceans worldwide (1-10). Technical HCH is a mixture of several isomers, the most abundant being R-HCH (60-70%), β-HCH (5-12%), and γ-HCH (10-15%) (4). The technical product * Corresponding author phone: (705)458-3322; fax: (705)458-3301; e-mail:
[email protected]. † Environment Canada. ‡ Norwegian Institute for Air Research. § Swedish University of Agricultural Sciences. | University of Toronto. Current address: Environmental Monitoring and Reporting Branch, Ontario Ministry of the Environment, 125 Resources Road, West Wing, Toronto, ON M9P 3V6, Canada. ⊥ Institute of Ocean Sciences. 2688
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 8, 2007
was deregistered in Canada in 1971 and the U.S.A. in 1976 (11), phased out in Europe during the 1970s and 1980s (12), and replaced by lindane (pure γ-HCH). Usage of technical HCH continued in China until 1983 and the former Soviet Union until 1990, and as of 1996 had not been banned in India and some countries in Africa and South America (11). Although the major usage of technical HCH was in Asian countries (11, 13), concentrations of R-HCH in ocean surface waters of the Northern Hemisphere measured after the late 1980s increase with latitude (4, 14) and are an order of magnitude higher in the Arctic Ocean and its regional seas compared to temperate and subtropical North Pacific waters. This is due to more favorable air-to-water partitioning at cold temperatures (4, 14) and slower rates of chemical degradation (15). Heaviest loadings of HCHs to the Arctic Ocean occurred during the 1960s through the 1980s by atmospheric transport and air-water gas exchange, precipitation and riverine input, and migration through northflowing ocean currents (16). Relative loadings by these pathways varied over time and differed for the eastern and western sides of the Arctic Ocean, which have been termed the North American Arctic Ocean (NAAO) and Eurasian Arctic Ocean (EAO) (see Figure 1 in ref 16). Cumulatively, gas exchange and ocean currents supplied ∼80% of the R-HCH that entered the entire Arctic Ocean between 1945 and 2000 (16). Gas exchange dominated until the early 1980s, but dramatic drops in HCH emissions and atmospheric concentrations throughout the 1980s and 1990s (17-19) led to declining atmospheric loadings and after 1990 the main input of R-HCH was through ocean currents (16, 20). In recent times the Arctic Ocean is experiencing a net loss of R-HCH, largely by water outflow from the NAAO and microbial degradation (16, 20). Concentrations of R-HCH are 2-4 times higher in the NAAO than the EAO, as reported in a number of studies (3, 5, 6, 9, 13, 16, 20, 21) and summarized in Figure 4.29 in the 2002 Arctic Assessment Report (1). Higher concentrations in the NAAO are thought to have arisen from historical atmospheric transport of technical HCH across the Pacific Ocean from Asian countries where use of technical HCH was high during the 1970s and early 1980s. Atmospheric deposition into the North Pacific followed by ocean current transport into the NAAO through the Bering Strait was also an important pathway, especially for β-HCH which lagged R-HCH in making its peak appearance in the NAAO (22). Within the NAAO, R-HCH concentrations are highest in the Beaufort Gyre and central Canadian Archipelago and lower in the Chukchi Sea and northern Canada Basin (1, 5-7, 16, 21-27). All three HCH isomers bioaccumulate. Differences in the relative body burdens of R-HCH and β-HCH among species result from selective metabolism and varying concentration distributions in Arctic Ocean water masses (13). Levels of ΣHCHs in ringed seal blubber were higher in the Beaufort Sea and Canadian Archipelago compared to eastern Greenland and Svalbard, reflecting the higher water concentrations in the Archipelago (28, 29). Near Barrow, Alaska, R-HCH was the most abundant isomer in ringed seal, whereas β-HCH predominated in polar bear (30). Bowhead whales reversed the proportion of R-HCH/β-HCH in their blubber during their annual migrations between the Bering and Beaufort seas (13, 31). Proportions of β-HCH/ΣHCHs or β-HCH/ R-HCH in the Canadian Arctic have increased in birds and marine mammals since the 1970s and 1980s (13). Through measurements taken predominantly in the NAAO during mid-1980s and 1990s by various investigators (5, 6, 21, 23-27), the upper interior ocean in the western Arctic 10.1021/es062375b CCC: $37.00
2007 American Chemical Society Published on Web 03/14/2007
FIGURE 1. Map of the Canadian Archipelago and eastern Beaufort Sea. Top: TNW-99 sampling stations and surface water flow pathways. Insets show depth profiles of r-HCH at the Ice Island in 1986 (20, 24, 26) and the southern Beaufort Sea in 1993 (20). Bottom: topography and vertical distribution of water masses. Representative concentrations of ΣHCHs are shown as bars at the top. has been revealed to contain a dynamic reservoir of HCH stationed between the Bering Sea and the Canadian Archipelago. HCH has been lost from the reservoir through degradation and outflow through the Archipelago (13, 16, 20, 21) with downstream consequences in Baffin Bay and the North Atlantic Ocean (32). With the exception of a few measurements of HCH made in 1992-1993 near Resolute Bay (23, 25) there are no data from which to evaluate how the projected large transports of HCH are processed while passing through the vast channels of the Archipelago. The Swedish Tundra Northwest 1999 (TNW-99) expedition provided an opportunity for a survey of HCH concentrations and isomeric composition in surface water and air across the Canadian Archipelago and the estimation of air-water gas exchange. The spatial distribution and pathways of R-, β-, and γ-HCHs in surface water from TNW-99 are reported here, and gas exchange studies are covered in a companion paper (33).
Experimental Section Sample Collection and Preparation. The TNW-99 expedition was made on the CCGS Louis S. St-Laurent during JulySeptember, 1999. Following a route generally outlined by the sequence of stations in Figure 1A, the ship traveled from Nuuk, Greenland across the Davis Strait to Iqaluit, Canada (stations 1 and 2), then traversed the Archipelago from Hudson Strait (stations 3 and 4) to Barrow Strait (station 9), along a southern route to Tuktoyaktuk (stations 10 and 14) and the southern Beaufort Sea (stations 11-13 and 15), and returned along a northern route to Barrow Strait near Resolute Bay (station 19), Ellef Ringnes Island (stations 21 and 22), over Devon Island through Jones Sound (station 24), past the mouth of Lancaster Sound (stations 25-27), and through Baffin Bay and the Davis Strait (stations 28-30). Samples on the outbound route between Iqaluit and Resolute Bay (stations 1-9) were collected by Environment Canada (EC) personnel, while those between Resolute Bay-Beaufort Sea VOL. 41, NO. 8, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2689
and on the return trip were collected by personnel from the Swedish University for Agricultural Sciences (SLU). In total, 153 small volume (SV) surface samples of 4-20 L were taken for R- and γ-HCHs (replicates of 3-8 samples per station), and 17 single large volume (LV) surface samples of 80-200 L were taken for β-HCH and other organochlorine pesticides (not reported here). Surface water was collected with a submersible pump through polytetrafluoroethylene (PTFE) tubing surrounded by a metal mesh into stainless steel cans. Locations are listed in Table S1 of Supporting Information and shown in Figure 1A. The samples collected by EC and SLU were processed separately by the two groups using the methods of Jantunen et al. (8) with variations by SLU as noted below. After spiking the cans of water with R-HCH-d6 and γ-HCH-d6 surrogates (SV samples) or only R-HCH-d6 (LV samples) (Cambridge Isotope Laboratories, Andover, MA), the water was passed through a glass fiber filter followed by a 200 mg ENV+ cartridge (IST Isolute, Biotage, Charlottesville, WV) (SV samples), a glass column containing 50 g of XAD-2 resin (20-60 mesh, Supelco, Bellefonte, PA) (LV samples collected by EC), or a column containing 1 g ENV+ (LV samples collected by SLU). After use, the ENV+ cartridges were placed in sealed plastic bags and frozen. XAD-2 was placed in 100mL glass bottles and refrigerated at 4 °C. Sorbent preparation and further details of sampling and extraction methods used by EC for SV and LV samples, and SLU for SV samples, are described by Jantunen et al. (8). A slight change was made by using 12 mL of dichloromethane followed by 12 mL of acetone (instead of dichloromethane alone) to elute ENV+ cartridges, to improve recovery. Extracts were cleaned up on a 1-g column of neutral alumina (0.0630.30 mm, EM Science, deactivated with 6% water) topped with sodium sulfate. Extract volumes were brought to 1 mL (quantitative analysis) or 100 µL (enantiomer analysis) in isooctane under a nitrogen stream. Additional cleanup was done before chiral analysis by vortex mixing the extracts in 1 mL of isooctane with ∼0.5 mL of 18 M sulfuric acid, centrifuging, washing the solvent layer with deionized water, and then adjusting the volume. The LV samples collected by SLU on ENV+ were eluted using larger volumes of solvents than the SV cartridges and then cleaned up by gel permeation chromatography and fractionated on deactivated silica (34). An aliquot of the eluate from the silica gel column was treated with 15% fuming sulfuric acid for the determination of the HCHs. Analysis. Analysis was done by capillary gas chromatography-electron capture negative ion mass spectrometry using an Agilent 6890 GC-5973 MSD. Quantitative analysis was done on a DB-5 or DB-5MS column (J&W, Agilent Technologies, Palo Alto, CA) as reported by Jantunen et al. (8). The enantiomers of R-HCH were separated by EC on three columns: BGB-172 (BGB Analytik AG, Switzerland), Betadex120 (Supelco, Bellefonte PA), and Rtx-βDEXcst (Restek, Bellefonte, PA). SLU used Betadex-120 only. Descriptions of these columns and typical operating conditions are given by Jantunen et al. (8) and Shen et al. (32). Results are reported as the enantiomer fraction, EF ) areas of (+)/[(+) + (-)] peaks (35). Unsupervised Principal Component Analysis (PCA) was performed using the nonlinear iterative partial least-squares (NIPALS) algorithm (software provided by Chemometrics Clinic, Seattle, WA). The data, which included R-HCH concentration and EF, γ-HCH concentration, latitude, and longitude as variables, were autoscaled (mean centered and variance normalized). Quality Control. Blanks for HCHs averaged 1-2 pg L-1, well below sample concentrations (Table 1). Recoveries of deuterated surrogates R-HCH-d6 and γ-HCH-d6 were 84% and 79% from SV samples and 68% (R-HCH-d6) for LV 2690
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 8, 2007
samples. Concentrations in Table 1 were adjusted for mean recoveries. Relative percent standard deviations (%RSD) for replicate SV samples (3-8 samples at a station) ranged from 0.5 to 25% and averaged 5.2% for R-HCH. The %RSDs for γ-HCH ranged from 5.3 to 48% and averaged 18%. Racemic standards were repeatedly injected on the three chiral columns by EC to determine the reproducibility of measuring the EFs. Average EF values for R-HCH were as follows: 0.499 ( 0.004 (n ) 22) on BGB-172, 0.500 ( 0.003 (n ) 27) on the Betadex-120, and 0.500 ( 0.001 (n ) 12) on the RTx-βDEXcst. The criterion used for peak purity in samples was in agreement with the target/qualifying ion ratio within (5% of the standard values. Confirmational analysis for enantiomers was done by comparing EFs for samples obtained from the two or three columns. The average absolute difference between the columns was 1.2% (30 comparisons). The %RSDs for EFs in replicate samples ranged from 0.4 to 2.4% and averaged 1.1%. Collections by the two research groups were made at stations in the central Archipelago (Figure 1), though at different times, allowing a limited comparison of results. Average concentrations measured at stations 7-9 (total n ) 5) by EC were as follows: R-HCH 3.1 ( 0.47, γ-HCH 0.32 ( 0.07, and EF of R-HCH 0.445 ( 0.0013. Measurements by SLU at stations 19 and 20 (total n ) 10) were as follows: R-HCH 3.4 ( 0.05, γ-HCH 0.38 ( 0.05, and EF of R-HCH 0.441 ( 0.005. Means in each case are not significantly different (p > 0.05). Single measurements of β-HCH were made at stations 7 (EC) and 19 (SLU) and were 0.099 ng L-1 in each case.
Results and Discussion Trends in Concentrations and R-HCH Enantiomer Fractions. Concentration and EF results are summarized in Table 1 and given for each station in Table S1 (Supporting Information). Surface water concentrations (ng L-1) of HCHs in the southern Beaufort Sea and Archipelago ranged from 1.1 to 5.4 (R-HCH), 0.056 to 0.16 (β-HCH), and 0.19 to 0.45 (γ-HCH). The tendency toward higher R-HCH concentrations and higher ratios of R-HCH/γ-HCH in the southern Beaufort Sea and western Archipelago and lower ones in the centraleastern Archipelago is a dominant feature of these data (Figure 2). EFs of R-HCH ranged from 0.432 to 0.463, indicating preferential degradation of the (+) enantiomer. EFs were lowest in the southern Beaufort Sea and increased across the Archipelago from west to east (Figure 2). Linear regressions vs longitude (Tables 1 and S2) were significant (p < 0.001-0.003) for R-HCH, β-HCH, γ-HCH, EF of R-HCH and R-HCH/γ-HCH, but not for R-HCH/β-HCH. Regressions vs latitude (Tables 1 and S3) were significant (p < 0.0010.049) for R-HCH, γ-HCH, EF, and R-HCH/β-HCH. PCA using the variables R-HCH and γ-HCH concentrations, R-HCH EF, latitude, and longitude resulted in the first two PCs accounting for 71% and 16% of the variance (Figure 3). Samples clustered into four main groups, representing the Davis Strait and lower Baffin Bay (stations 1-6), upper Baffin Bay and Lancaster Sound (stations 25-28), central, north, and west Archipelago (stations 9 and 17-14), and the southern Beaufort Sea (stations 10-15). Transition stations 7, 8, 16, 29, and 30 fall in between these groups. The variables plot (vectors in Figure 3) show that R-HCH is most strongly associated with longitude and γ-HCH with latitude. Surface water concentrations of R-HCH and γ-HCH in the central Archipelago stations 7-9, 19, and 20 averaged 3.2 and 0.36 ng L-1 (Table 1). Sampling in Barrow Strait during 1992-1993 showed R-HCH ranging from 3.6-4.7 and γ-HCH ) 0.44-0.52 ng L-1 (Table 1) (23, 25). Falconer et al. (25) found no vertical gradients of either HCH isomer between 2 and 100 m and an average EF (0.482) higher than measured during TNW-99. Hargrave et al. (23) noted that surface water
TABLE 1. HCHs in Surface Water of the Canadian Archipelago and Adjoining Seas r-HCH (ng L-1)
γ-HCH (ng L-1)
β-HCH (ng L-1)
EF r-HCH
TNW-99 location
stationsa
range
mean
SD
range
mean
SD
range
mean
SD
range
mean
SD
east archipelago central archipelago north archipelago west archipelago Beaufort Sea
1-6, 25-30 7-9. 19,20 21-24 16-18 10-15
1.1-3.1 2.8-3.5 4.1-4.8 4.6-5.4 3.9-5.4
2.4 3.2 4.4 4.9 4.7
0.49 0.28 0.32 0.44 0.58
0.19-0.27 0.29-0.41 0.32-0.38 0.39-0.45 0.26-0.36
0.24 0.36 0.36 0.42 0.31
0.03 0.05 0.03 0.03 0.05
0.056-0.12
0.079 0.10 0.11 0.12 0.14
0.023
0.441-0.463 0.440-0.444 0.437-0.449 0.433-0.438 0.432-0.439
0.452 0.444 0.441 0.436 0.436
0.006 0.003 0.006 0.003 0.003
literature reports
years
VOL. 41, NO. 8, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
central archipelago Baffin Bay Beaufort Sea northern Canada Basin Lincoln Sea East Greenland Current Labrador Sea western North Atlantic
a
See Figure 1a.
b
r-HCH (ng L-1) range mean
1992-1993
3.3-6.1
1998 1986-1993 1994
4.5-7.1 2.1-2.7
2.4
1993 1994-2005
3.1-4.3 0.15-0.69
2005 1998
0.14-0.24
SD
4.2 1.1
γ-HCH (ng L-1) range mean 0.31-0.65
0.07
0.075-0.12 0.10-0.14 0.12-0.16 β-HCH (ng L-1) range mean
SD
SD
0.03 0.03 0.02
range
EF r-HCH mean
0.48
0.482
0.2
0.01
0.2
0.61-0.81 0.36-0.70
0.49
0.11
3.4 0.37
0.6 0.21
0.39-0.60 0.04-0.23
0.50 0.12
0.07 0.08
0.08 0.20
0.04
0.029-0.034
0.02 0.033
0.005
0.07
0.01
0.12-0.26
0.18
0.06
0.15-0.26 0.013-0.016
0.20 0.014
0.06 0.002
0.451-0.480
ref 23,25
0.45
0.01
0.470
0.009
36 21,24,26 5, 6 21b 5, 6, 9, 43
0.010 0.460-0.472
0.466
TNW-99 relationships
r2
p
R-HCH vs longitude β-HCH vs longitude γ-HCH vs longitude EF of R-HCH vs longitude R-HCH/γ-HCH vs longitude R-HCH/β-HCH vs longitude R-HCH vs latitude β-HCH vs latitude γ-HCH vs latitude EF of R-HCH vs latitude R-HCH/γ-HCH vs latitude R-HCH/β-HCH vs latitude
0.73 0.71 0.27 0.69 0.48