Orographic Cold-Trapping of Persistent Organic Pollutants by

Environmental Health Program,. Concordia University College of Alberta, 10537-44 Street,. Edmonton, Alberta, Canada, T6A 1W1. DAVID W. SCHINDLER...
1 downloads 4 Views 194KB Size
Research Orographic Cold-Trapping of Persistent Organic Pollutants by Vegetation in Mountains of Western Canada DEBORAH A. DAVIDSON,* ANDREW C. WILKINSON, AND JULES M. BLAIS Program for Chemical and Environmental Toxicology, Department of Biology, University of Ottawa, P.O. Box 450, Station A, Ottawa, Ontario, Canada, K1N 6N5 LYNDA E. KIMPE Department of Biology, University of Ottawa, P.O. Box 450, Station A, Ottawa, Ontario, Canada, K1N 6N5 KAREN M. MCDONALD Environmental Health Program, Concordia University College of Alberta, 10537-44 Street, Edmonton, Alberta, Canada, T6A 1W1 DAVID W. SCHINDLER Department of Biological Sciences, University of Alberta, CW 405 Biological Sciences Center, Edmonton, Alberta, Canada, T6G 2E9

Conifer needles from mountain areas of Alberta and British Columbia, Canada, were collected from sites that ranged in altitude from 770 to 2200 masl and were analyzed for polychlorinated biphenyls (PCBs) and organochlorine pesticides (OCs) to determine if they are progressively concentrated in colder, more elevated mountain areas, where temperatures decrease as elevation increases. Concentrations of OCs in needles ranged from 43 to 2430 pg g-1, 55-17500 pg g-1, and 11-2930 pg g-1 (dry weight), for total hexachlorocyclohexanes (HCHs), PCBs, and endosulfans, respectively. The more volatile OCs, with subcooled liquid vapor pressures (PL) > 0.1 Pa at 25 °C, increased at higher altitudes, whereas the less volatile OCs were either unrelated or inversely correlated with altitude. These spatial patterns were similar for species of spruce (Picea engelmannii and glauca) and pine (Pinus contorta and albicaulis). Back trajectories revealed that air masses arriving at these sites traveled over Asia and the Pacific Ocean before reaching the Rocky Mountains. Results from this study demonstrate that alpine ecosystems accumulate these chemicals to the same degree that is observed in polar environments that are known to receive contaminants by long-range transport.

TABLE 1. Sample Sizes for Each Species of Vegetation Sampled in the Rocky Mountains site Donald Station Dixon Dam Vermilion Lakes Wapta Lake Lower Kananaskis Lake Bow Lake Rock Isle

altitude Engelmann White Lodgepole Whitebark (masl) Spruce Spruce Pine Pine 770 948 1380 1590 1667

2 2 20 12

1975 2200

20 18

10 12 24 6

2

8 2 22 20 12 20 2

16

tionation (1). Compounds prone to enrichment in cold environments are typically those with subcooled liquid vapor pressures between 0.01 and 1.0 Pa at 25 °C and include several of the organochlorine (OC) compounds (2). They accumulate in northern food chains to levels that sometimes exceed human consumption guidelines (3). Soil and vegetation are dominant reservoirs for these chemicals and are vectors through which persistent chemicals may enter terrestrial food chains (4). POP concentrations in various media have been shown to increase at higher latitudes. For example, concentrations of R-hexachlorocyclohexane (R-HCH), γ-HCH, hexachlorobenzene (HCB), and pentachloroanisole in bark samples increase significantly in northern regions (5, 6). Similar relationships have been observed for dichlorodiphenyltrichloroethane (DDT) (7) and polycyclic aromatic hydrocarbons (PAHs) (8) in fish. Air concentrations of semivolatile POPs show an inverse relationship with latitude (9), particularly for heavier, less volatile compounds. As contaminated air masses move to colder, higher latitudes, these compounds condense and partition to aerosols, soil, vegetation, and water from a lowering of vapor pressure at lower temperatures. Chemical distillation toward colder regions can occur in mountainous areas, where temperatures decrease as elevation increases. In the Canadian Rockies, concentrations of many OC compounds in snowpacks were correlated with site elevation (10), and similar patterns were reported for OC compounds in fish from Canada (11) and Europe (12) and in soils from New Hampshire, USA (13), and the Canary Islands (14). Because only a small portion of POPs deposited by rain in this area reaches streamwater (15), most of the POPs inventory in this area, resulting from precipitation and gas phase deposition, is likely stored in terrestrial vegetation and soil. It is hypothesized that cold temperatures at higher elevations promote the deposition and accumulation of POPs, particularly the more volatile compounds which tend to evaporate in warmer areas and be transported by air currents and deposited in colder alpine and arctic areas (1, 2). In this study, a large data set of OC pesticide concentrations in vegetation from the mountain regions of Alberta and British Columbia, Canada, was compiled to determine altitudinal trends and to test this hypothesis.

Introduction

Experimental Section

Regions with cold climates are susceptible to the enrichment of persistent organic pollutants (POPs) through global frac-

Sample Collection. New growth of Engelmann spruce (Picea engelmannii; n ) 74), white spruce (Picea glauca; n ) 54), lodgepole pine (Pinus contorta; n ) 86), and whitebark pine (Pinus albicaulis; n ) 16) needles were collected throughout the summers of 1999 and 2000 from seven sites in the Canadian Rockies (Table 1, Figure 1). Needles were taken at

* Corresponding author phone: (613)736-3532; fax: (613)736-3630; e-mail: [email protected]. Present address: Pest Management Regulatory Agency, 2720 Riverside Drive, A.L. 6606E, Ottawa, ON, Canada, K1A 0K9. 10.1021/es020605q CCC: $25.00 Published on Web 12/07/2002

 2003 American Chemical Society

VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

209

FIGURE 1. Location of sampling sites. Points A. Vancouver Island; B. Vancouver; C. Calgary; and D. Edmonton are for reference only. Samples were collected at E. Bow Lake; F. Donald Station, G. Wapta Lake; H. Vermilion Lakes; I. Rock Isle; J. Lower Kananaskis Lake; and K. Dixon Dam. heights between five and seven feet and removed from outer whorls of the tree with a pair of solvent-rinsed scissors. Samples were wrapped in clean aluminum foil, stored in resealable plastic bags, and sent to the University of Ottawa for storage at -20 °C until extraction. Chemicals. Florisil PR grade 60/100 mesh (Supelco, Bellefonte, PA) was activated in a muffle furnace at 600 °C for 6 h and stored at 130 °C. All field surrogates, lab surrogates, and pure standards used at the University of Ottawa were obtained from the National Laboratory for Environmental Testing (NLET) in Burlington, ON, Canada. Field surrogates contained 1,3,5-tribromobenzene (1,3,5-TBB), 1,2,4,5-tetrabromobenzene (1,2,4,5-TTBB), and δ-hexachlorocyclohexane (δ-HCH) at approximately 100 pg µL-1 in methanol. Laboratory OC surrogates included 1,3-dibromobenzene (1,3-DBB) and endrin ketone at approximately 50 pg µL-1 in isooctane. Octachloronaphthalene (OCN), at approximately 50 pg µL-1, in isooctane was also used as a laboratory surrogate. Lab surrogates were diluted 10-fold with isooctane before use. Mirex, used as an internal standard, was obtained from Ultra Scientific (North Kingstown, RI) at 100 ng µL-1 in methanol and further diluted to 2000 pg µL-1 in isooctane. Pure reference standard solutions were used for instrument calibration, recovery evaluation, and analyte identification and quantification. Extraction. A 5 g subsample of needles removed from the twig was placed in a clean 40 mL amber glass vial, spiked with field surrogates, and sonicated for 30 min in a 20:80 mixture of acetone:hexane covering the vegetation in an ultrasonic cleaner. The solvent was decanted and placed into a clean separatory funnel, which allowed for removal of water and emulsion through liquid-liquid extraction with Omnisolv water. The liquid-liquid extraction was performed in triplicate before the extract was further dried with sodium sulfate. The extract was then evaporated down to a few milliliters using a rotary evaporator with a water bath at 30 °C. Following lipid determination (see below), the extract was concentrated to 2 mL with a gentle stream of ultrahigh purity nitrogen. After spiking with laboratory surrogates, the extract was cleaned up and fractionated on a Florisil column packed as follows: glass wool, 8 g of 1.2% deactivated Florisil, 1 g of 210

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003

sodium sulfate. Florisil was deactivated by vortexing 7.9 g of the magnesium silicate adsorbent with 96 µL of Omnisolv water for 2 h. The packed column was pre-rinsed with 50 mL of hexane that was drained to the top of the sodium sulfate before exposing the sodium sulfate to the air, and the eluate was discarded. The extract was placed on top of the column and allowed to pass through the column prior to fractionation. Fraction 1 (37 mL of hexane) contained HCH, heptachlor, HCB, dichlorodiphenyldichloroethylene (DDE), and DDT; fraction 2 (38 mL of a 15:85 mixture of dichloromethane: hexane) contained HCH, chlordane, and dichlorodiphenyldichloroethane (DDD); and fraction 3 (52 mL of dichloromethane) contained endosulfan, dieldrin, endrin, heptachlor epoxide, and methoxychlor. Each fraction was drained to the top of the column, solvent-exchanged into isooctane, and evaporated down to approximately 0.5 mL by rotary evaporation followed by nitrogen blow-down. Mirex was added as an internal standard, and the extract was topped up to 1 mL with isooctane. Each fraction was stored at -20 °C until analysis. The moisture and lipid content of each sample were determined so that concentrations could be normalized to either lipid weight or dry weight, as desired. Prior to the first nitrogen evaporation, the extract was topped up to 12 mL with hexane, and a 10% aliquot of the extract was placed in a preweighed aluminum weigh dish. After drying in a desiccator for several days, the dish was re-weighed, and the percent lipid of vegetation fresh weight was calculated. For moisture determination, a 2 g subsample was weighed into a 40 mL amber glass vial, dried at 95 °C for 24 h, and reweighed, and the percent moisture was calculated. Analytical. Vegetation extracts were analyzed by a HewlettPackard 6890 gas chromatograph equipped with a split/ splitless injector, a 30 m × 250 µm i.d. HP-5 capillary column (5% phenyl, 95% dimethylpolysiloxane) with a 0.25 µm thickness, and a 63Ni electron capture detector. One microliter of extract was injected in the splitless mode with an initial injector temperature of 250 °C. Helium was used as the carrier gas and nitrogen as the makeup gas. The temperature program conditions were 80 °C held for two minutes, ramped at 10 °C per minute to 110 °C and then at 3 °C per minute

TABLE 2. Results for the Correlation between POP Concentration and Altitude for Each Speciesa spruce

pine

compound

r

m

r

m

PL (Pa)b

β-endosulfan heptachlor HCB R-HCH γ-HCH dieldrin R-endosulfan p,p′-DDE γ-chlordane R-chlordane heptachlor epoxide endrin p,p′-DDD methoxychlor o,p′-DDT p,p′-DDT

0.281 0.418c 0.288c 0.379c -0.251c 0.302 0.186 0.258c -0.244 -0.227 0.041

2.34 × 10-4 4.28 × 10-4 2.82 × 10-4 3.09 × 10-4 -2.34 × 10-4 2.88 × 10-4 1.71 × 10-4 2.13 × 10-4 -4.51 × 10-4 -3.47 × 10-4 5.37 × 10-5

0.582c 0.394c 0.331c 0.222c 0.092 -0.304 0.023 -0.070 -0.460 -0.318 -0.330c

6.95 × 10-4 6.35 × 10-4 3.80 × 10-4 1.90 × 10-4 9.95 × 10-5 -3.42 × 10-4 2.55 × 10-4 -5.91 × 10-5 -1.33 × 10-3 -4.26 × 10-4 -5.51 × 10-4

3.94 × 10-1 2.67 × 10-1 2.45 × 10-1 1.00 × 10-1 2.74 × 10-2 1.60 × 10-2 8.00 × 10-3 3.72 × 10-3 2.65 × 10-3 2.65 × 10-3 2.56 × 10-3

0.073 -0.835 0.208 0.478 -0.131

9.20 × 10-5 -3.75 × 10-4 2.28 × 10-4 3.11 × 10-4 -9.52 × 10-4

-0.196 0.938 -0.489 0.059 -

-4.27 × 10-4 7.31 × 10-4 -1.66 × 10-4 1.07 × 10-4 -

1.32 × 10-3 6.93 × 10-4 5.46 × 10-4 1.72 × 10-4 1.35 × 10-4

a Given are the Pearson’s correlation coefficients (r), the slopes for the lines of best fit (m) for the correlation between the log concentration (pg g-1 dry weight) and site elevation (masl) for samples of spruce and pine foliage, and the subcooled liquid vapor pressure (PL) at 25 °C. Compounds are listed in order of decreasing vapor pressure from highest to lowest volatility. The correlation between the concentration of p,p′-DDT in pine foliage and elevation could not be calculated due to the small sample size. b Values for the subcooled liquid vapor pressure (PL) at 25 °C were obtained from Mackay et al. (32), with the exception of heptachlor epoxide, which was quoted from Howard (33). c Correlations significant at the 95% confidence level are indicated in bold.

FIGURE 2. Correlations between POP concentrations in vegetation and altitude. Points represent the mean concentration (pg g-1 dry weight) at each elevation (masl) and error bars represent the standard error. Lines of best fit were drawn if analyses revealed a significant correlation between the original concentrations and elevation, as in Table 2. Samples in which analytes were not detected were omitted from analyses and mean calculations, giving rise to the following sample sizes for each site: Donald Station, n ) 20; Dixon Dam, n ) 14; Vermilion Lakes, n ) 48; Lower Kananaskis Lakes, n ) 46; Wapta Lake, n ) 24; Bow Lake, n ) 40; Rock Isle, n ) 38. to 280 °C, and held for five minutes. The detector temperature was 350 °C. Chromatographic analysis and quantification of sample extracts were performed using HP Chemstation software (Rev.

A.06.03, Hewlett-Packard, Palo Alto, CA). Sample extracts were screened for 16 OC pesticides. Analytes were considered present if the sample peak and its corresponding reference peak eluted within a retention time window of 0.04 min. A VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

211

five-point calibration curve was prepared using reference standards between 1.9 and 530 pg µL-1. Calibration was based on a linear curve forced through the origin with linear weighting to all five points. The coefficients of determination for all calibration curves were 0.99 or higher. Mid-level standards for the OC pesticides were analyzed at intervals throughout the sample analysis and used to recalibrate the instrument after every 15 injections. Sample peak quantification was based on the relative response of the internal standard against the target analytes. Data analysis was performed using either SYSTAT, version 9.01 (SPSS Inc., Chicago, IL) or SPSS for Windows Student Version, release 9.0.1 (SPSS Inc., Chicago, IL). Quality Control. Aluminum foil used for storage of vegetation samples was rinsed with ACS grade acetone and hexane before heating for 12 h at 200 °C. All glassware was washed with industrial grade detergent, rinsed in triplicate with tap water followed by deionized water, rinsed with ACS grade acetone and hexane, and heated for 12 h at 200 °C. Glass Pasteur pipets used for transfer of solvents and sample extracts were also heated for 12 h at 200 °C. To monitor potential laboratory contamination, procedural blanks were processed after every 10 vegetation extractions. All data were blank-corrected prior to analysis by subtracting the mean blank concentration from the extract concentration and corrected for loss of analyte in the 10% aliquot used for lipid determination. Field surrogates were recovered with 83.5 ( 0.8% efficiency, while the recovery of OC laboratory surrogates were 104.1 ( 1.3%. Blanks contained detectable levels of R-HCH, HCB, heptachlor epoxide, R-endosulfan, R-chlordane, dieldrin, p,p′-DDE, and endrin at 16.9 ( 1.1%, 64.2 ( 8.2%, 1.0 ( 0.3%, 15.8 ( 1.9%, 1.1 ( 0.2%, 7.2 ( 1.2%, 1.6 ( 0.1%, 138.2 ( 35.7%, respectively, of mean concentrations in vegetation samples. Back Trajectories. Five day back trajectories were computed for the sampling area for July, 2000, at pressures of 850, 725, and 500 mbar using a Lagrangian model based on meteorological data objectively analyzed every 6 h. Detailed descriptions of the methods used to compile these trajectories can be found in Cheng et al. (16) and McDonald et al. (17).

TABLE 3. Calculation of ∆H Based on Regressions of the log Lipid-Normalized Concentration (pg g-1 Lipid) against 1/T ( the Standard Errora compound OC pesticides RHCH HCB R-endosulfan β-endosulfan methoxychlor polychlorinated biphenyls Tetra CB Penta CB Hexa CB Sum CB

212

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003

∆H (19)

22 ( 7 46 ( 15 33 ( 17 65 ( 19 91 ( 42

68.4 68.5 80.4 82.4

30 ( 11 22 ( 11 25 ( 11 26 ( 9

79.0 83.6

a (See text). Experimentally determined ∆H values from ref 19 are provided for comparison.

TABLE 4. Global Distribution of DDT/DDE Ratios medium

geographic region

vegetation Africa

Results and Discussion The most ubiquitous compounds present in the vegetation were R-HCH, γ-HCH, and HCB, occurring in over 97% of the samples. Heptachlor epoxide and methoxychlor were the most concentrated OC pesticides detected, with respective mean concentrations of 3100 ( 1100 pg g-1 dry weight and 2600 ( 870 pg g-1 dry weight (error estimates represent the standard error of the mean). Average concentrations of p,p′DDD and p,p′-DDE were 72 ( 9.6 pg g-1 dry weight and 55 ( 4.0 pg g-1 dry weight, making these compounds the most dilute of the OC pesticides detected. Concentrations of endosulfan and HCB detected in this study were comparable to those reported for vegetation samples collected from northern Russia and Alaska (5). Levels of DDT were also comparable to those reported in northern Norway, and dieldrin concentrations were within the range detected in vegetation from northern Russia (5). These northern areas have been investigated for years as potential sinks for global emissions of POPs. To examine the potential for POP distillation in this mountain region, the log of the analyte concentrations in vegetation was correlated with altitude. Samples were grouped into two classes, with ‘spruce’ consisting of samples of Engelmann spruce and white spruce, and ‘pine’ consisting of samples of whitebark pine and lodgepole pine. Several of the more volatile OC pesticides (PL > 0.1 Pa at 25 °C) correlated significantly with altitude, while the others were unrelated or inversely related to elevation (Table 2, Figure 2). This trend

∆H (kJ mol-1) ( SE (this study)

air

location

Burkina Faso Sierra Leone Ghana Benin Guinea Ivory Coast Kenya Asia Nepal India Jordan Indonesia Japan Mauritius Europe Italy Netherlands Greece Austria Finland Czech Canada Rocky Mountains Mexico Mexico Central America Guatemala South America Venezuela South America Chile Canada Alert Canada Tagish Russia Dunai USA/Canada Great Lakes USA Alabama Russia Irkutsk Lake Baikal Central America Belize

ratio DDT/DDE ref 10.22 7.83 6.67 5.10 4.57 2.06 0.95 7.37 3.76 0.98 2.15 1.27 6.77 1.63 1.15 0.76 0.63 0.50 0.41 1.88 ( 0.22a 5.38 2.64 9.27 1.09 0.33 0.31 0.39 0.60 0.71 2.50 1.20 2.30

34 35 6 6 6 6 6 6 34 34 34 34 35 36 36 34 36 36 36 34 6 6 34 37 37 37 38 39 40 41 41

a The error estimate for the DDT/DDE ratio obtained in this study represents the standard error of the mean for a sample size of 20.

was observed for both genera, with the exception of β-endosulfan in spruce not showing a significant relationship with elevation. Pseudo enthalpies were calculated from the slope of the lipid-normalized concentrations in vegetation vs 1/T (∆H ) SR ln 10), where S is the slope of the regression line, and R is 8.314 J K-1 mol-1. Pseudo enthalpies in this study ranged from 22 to 91 kJ mol-1 for OCs, similar to other pseudo enthalpies calculated in the Laurentian Great Lakes region (18), but lower than experimentally derived values (19): 68-93 kJ mol-1 (Table 3). It has been proposed that when ∆H values measured in the field are much lower than experimentally determined ∆H values, long-range transport of these substances is likely the dominant factor affecting OCs distributions, and local volatilization from surface to air is less important (19). However, given the variability in air concentrations and fluctuations in temperature, it is unlikely

FIGURE 3. Back trajectories for the sampling area. Five-day back trajectories were computed every 6 h using a Lagrangian model at 850 mbar, 725 mbar, and 500 mbar for Banff, Alberta, a central point in the study area. Pressures of 850, 725 and 500 mbar represent average elevations of 1200, 2300 and 5500 masl. that a stable partitioning equilibrium is ever achieved between vegetation and air (20). Relationships observed in this study likely represent non steady-state conditions. Linear regression between temperature and elevation revealed a drop of 5 °C for every 1000-meter rise. Although plant lipid content increased significantly throughout the summer season (Pearson’s r ) 0.284, p < 0.01) at a rate of 3.42 × 10-5 g lipid g-1 needle day-1, dilution due to increased lipid content was insignificant, lowering the concentration from growth dilution by only 1.86%. Neither lipid content nor moisture content in vegetation was correlated with elevation. However, the magnitudes of the slopes for the regression lines were not dependent on the concentration of analyte detected in the vegetation samples. Thus, the relative increase in chemical concentration with elevation for different compounds is due to their relative volatilities and not their absolute levels in the environment. The relatively volatile γ-HCH, or lindane, showed an inverse relationship with elevation, likely due to application of this pesticide at low altitudes in nearby southern Alberta and British Columbia, as this compound is currently used in Canada (21, 22). An inverse relationship between ΣHCH concentration in spruce needles and altitude was also reported in Austria and was caused by the proximity of low altitude sites to agriculture where γ-HCH is applied (23). Ratios of both R-HCH to γ-HCH and DDT to DDE can be used as indicators of contaminant source age (6, 24). A high R/γ-HCH ratio suggests older or more distant sources because

R-HCH is no longer used in North America and the γ-isomer may be photolytically degraded to R-HCH (25). Furthermore, γ-HCH is more water-soluble than R-HCH, making it more prone to removal from air through precipitation or by deposition to water (26). The average R/γ-HCH ratio in this study was 2.04 ( 0.14 and reached as high as 19 in one sample from Rock Isle, the most elevated site. Levels of γ-HCH were highest in vegetation collected from Dixon Dam, a site near farming activity where application of this compound continues. The ratio of R-HCH to γ-HCH increased significantly with elevation (Pearson’s r ) 0.292, p < 0.001), suggesting that higher sites are receiving older or more distant sources of this compound, possibly because of lower air pressures that offer less resistance to air movement (Figure 3). Ratios of R-HCH to γ-HCH in lakes increase with latitude, indicating long-range transport of aged HCH to northern lakes (26). Once transported to colder regions, R-HCH is less likely to be transported further due to the low temperatures that hinder its evaporation (26). Low DDT/DDE ratios indicate aged sources because DDT is converted primarily to DDE (24). When both DDT and DDE were detected in vegetation samples, ratios ranged from 0.05 to 3.8, with a mean of 1.88(0.22, suggesting recent sources of DDT to this region. All but four of the 20 extracts that contained both DDT and its metabolite were sampled from Vermilion Lakes at 1380 masl. The distribution of species was fairly uniform across study sites with samples of pine and spruce collected at each elevation, except for Dixon Dam VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

213

at which spruce foliage dominated the samples (Table 1). It is not clear why DDT was detected most frequently at Vermilion Lakes. This site is on the eastern slope of the Rockies, along with most of the other sites and is located at mid-altitude. The parent isomer p,p′-DDT was detected only in samples from Rock Isle, Vermilion Lakes, and Wapta Lake collected during July and August of the 1999 sampling season. Because most of the p,p′-DDT was detected during a brief period, it is possible that contaminated air masses were arriving to the region at this time. Given the relatively low ratios reported previously throughout North America and Europe compared to other parts of the world (Table 3), the bulk of the DDT present in the Rocky Mountains could be arising from continued application of the parent DDT isomers in Asia, and possibly Africa, and South America. Arctic air shows higher levels of total DDT when air masses reaching the arctic have lingered over Asia (27). It has also been shown that North America receives pollutants from Eurasia via mid-latitude westerly winds that travel over the Pacific Ocean (28). Five-day back trajectories for air masses arriving at the sampling area for the summer months of 2000 reveal that air masses are traveling over the Pacific Ocean and over parts of Asia before reaching the Rocky Mountains (Figure 3). Donald et al. (11) performed five-day back trajectories from Bow Lake at 700 mbar for 1990, 1991, and 1992. The back-trajectories revealed that air arriving at Bow Lake originated from the Pacific and Arctic Oceans 60% of the time, while the United States, Canada, and Siberia were sources 21%, 11%, and 8% of the time, respectively (11). Higher sites are also more likely to receive air masses from farther distances than lower sites due to lower pressures that offer less resistance to the movement of air masses (Figure 3). Lower sites also tend to receive air from the east to a greater extent than higher sites, although the bulk of the air masses arrive from the west. Cotham and Bidleman (29) showed that HCB and dieldrin are evaporating from the Arctic Ocean, and air over the North Pacific Ocean has been shown to contain POPs (28). Chlorinated hydrocarbons, which are transported to oceans primarily through atmospheric transport (30), can evaporate fairly rapidly from water resulting in short half-lives in this medium (31). Thus, Asia and the Pacific Ocean are possible suppliers of fresh DDT to this region arriving via long-range transport. The enrichment of persistent chemicals seen here in vegetation from the Canadian Rockies suggests that cold condensation and fractionation are occurring in elevated areas and are not exclusive to remote, polar environments. The sampling sites in this study span only two degrees in latitude and four degrees in longitude. Thus mountain environments provide the conditions necessary to observe true chemical fractionation on a relatively small scale where terrestrial and aquatic ecosystems are exposed to similar regional sources and global air masses. Fractionation observed previously in global studies (5, 6) might have been influenced by dissimilarities in POP emissions between both countries and continents where regulations, restrictions, and economic conditions vary greatly. In light of these results, it is evident that alpine ecosystems warrant similar attention provided to arctic environments when it comes to POP transport and possible bioaccumulation in food webs. The high variability in concentrations suggest that other factors that we did not consider, such as precipitation, relative humidity, wind speed, slope and aspect might also affect contaminant deposition and accumulation in conifer needles. Furthermore, the assumption often made while modeling chemical partitioning between air and vegetation, that only soluble plant cuticular lipids are involved in the storage and distribution of these chemicals in plant tissue, may not adequately describe the actual processes involved in chemical 214

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003

uptake. For a better explanation of chemical exchange between air and vegetation, the effect of different tissues needs to be considered, particularly the influence of volatile constituents, such as toluene and terpenes, and other lipophilic constituents, such as lignins and tannins, which may influence the lipophilic nature of plant tissue. In order for this to occur, better analytical methods for determining chemical partitioning in different plant tissues will need to be employed.

Acknowledgments We are grateful to John Smol, Michael McLachlan, and David Donald for their comments on earlier drafts of this manuscript. Jeff Ridal provided advice for G.C. method development. The work of Jason Young in sample collection is greatly appreciated. This manuscript was improved by the suggestions of two anonymous reviewers. This research was supported by an NSERC Research Grant to J.M.B., an NSERC Strategic Grant to D.W.S., and an NSERC Post-Graduate Scholarship to D.A.D.

Literature Cited (1) Wania, F.; Mackay, D. Ambio 1993, 22, 10-18. (2) Wania, F.; Mackay, D. Environ. Sci. Technol. 1996, 30, 390A396A. (3) Kidd, K. A.; Schindler, D. W.; Muir, D. C. G.; Lockhart, W. L.; Heisslen, R. H. Science 1995, 269, 240-242. (4) Bo¨hme, F.; Welsch-Pausch, K.; McLachlan, M. S. Environ. Sci. Technol. 1999, 33, 1805-1813. (5) Simonich, S. L.; Hites, R. A. Science 1995, 269, 1851-1854. (6) Calamari, D.; Bacci, E.; Focardi, S.; Gaggi, C.; Morosini, M.; Vighi, M. Environ. Sci. Technol. 1991, 25, 1489-1495. (7) Hofer, R.; Lackner, R.; Kargl, J.; Thaler, B.; Tait, D.; Bonetti, L.; Vistocco, R.; Flaim, G. Water, Air, Soil Pollut. 2001, 125, 189200. (8) Escartı´n, E.; Porte, C. Environ. Sci. Technol. 1999, 33, 406-409. (9) Agrell, C.; Okla, L.; Larsson, P.; Backe, C.; Wania, F. Environ. Sci. Technol. 1999, 33, 1149-1156. (10) Blais, J. M.; Schindler, D. W.; Muir, D. C. G.; Kimpe, L. E.; Donald, D. B.; Rosenberg, B. Nature 1998, 395, 585-588. (11) Donald, D. B.; Stern, G. A.; Muir, D. C. G.; Fowler, B. R.; Miskimmin, B. M.; Bailey, R. Environ. Sci. Technol. 1998, 32, 1391-1397. (12) Grimalt, J. O.; Fernandez, P.; Berdie, L.; Vilanova, R. M.; Catalan, J.; Psenner, R.; Hofer, R.; Appleby, P. G.; Rosseland, B. O.; Massabuau, J. C.; Battarbee, R. W. Environ. Sci. Technol. 2001, 35, 2690-2697. (13) Smith, W. H.; Hale, R. C.; Greaves, J.; Huggett, R. J. Environ. Sci. Technol. 1993, 27: 2244-2246. (14) Ribes, A.; Grimalt, J. O.; Torres Garcia, C. J.; Cuevas, E. Environ. Sci. Technol. 2002, 36, 1879-1885. (15) Blais, J. M.; Schindler, D. W.; Muir, D. C. G.; Sharp, M.; Donald, D.; Laferniere, M.; Braekevelt, E.; Strachan, W. M. J.; Comba, M.; Backus, S. Ambio 2001, 36, 410-415. (16) Cheng, L.; McDonald, K. M.; Angle, R. P.; Sandhu, H. S. Atmos. Environ. 1998, 32, 673-681. (17) McDonald, K. M.; Cheng, L.; Olson, M. P.; Angle, R. P. Atmos. Environ. 1996, 30, 2969-2980. (18) Hoff, R. M.; Brice, K. A.; Halsall, C. J. Environ. Sci. Technol. 1998, 32, 1793-1798. (19) Hinckley, D. A.; Bidleman, T. F.; Foreman, W. T.; Tuschal, J. R. J. Chem. Eng. Data 1990, 35, 232-237. (20) Morosini, M.; Schreltmu ¨ ller, J.; Reuter, U.; Ballschmiter, K. Environ. Sci. Technol. 1993, 27, 1517-1523. (21) Bidleman, T. F. Water, Air, Soil Pollut. 1999, 115, 115-166. (22) www.nrtee-trnee.ca/eng/programs/Current_Programs/Health/ LINDANE_e.htm. (23) Weiss, P.; Lorbeer, G.; Scharf, S. Chemosphere 2000, 40, 11591171. (24) de March, B. G. E.; de Wit, C. A.; Muir, D. C. G. Chapter 6: Persistent organic pollutants. In AMAP Assessment Report: Arctic Pollution Issues; Arctic Monitoring and Assessment Programme, Oslo, Norway, 1998; pp 183-371. (25) Oehme, M. Ambio 1991, 20, 293-297. (26) Law, S. A.; Diamond, M. L.; Helm, P. A.; Jantunen, L. M.; Alaee, M. Environ. Toxicol. Chem. 2001, 20, 2690-2698.

(27) Bailey, R.; Barrie, L. A.; Halsall, C. J.; Fellin, P.; Muir, D. C. G. J. Geophys. Res. 2000, 105, 11805-11811. (28) Wilkening, K. E.; Barrie, L. A.; Engle, M. Science 2000, 290, 6567. (29) Cotham, W. E. Jr.; Bidleman, T. F. Chemosphere 1991, 22, 165188. (30) Harvey, G. R.; Steinhauer, W. G. Atmos. Environ. 1974, 8, 777782. (31) Mackay, D.; Wolkoff, A. W. Environ. Sci. Technol. 1973, 7, 611614. (32) Mackay, D.; Shiu, W.; Ma, K.-C. Physical-Chemical Properties and Environmental Fate Handbook (CD-ROM); Chapman & Hall/CRCnetBase: Boca Raton, FL, 2000. (33) Howard, P. H. Handbook of Environmental Fate and Exposure Data for Organic Chemicals. Pesticides. Vol. III; Lewis Publishers: Chelsea, MI, 1991. (34) Calamari, D.; Tremolada, P.; Notarianni, V. Environ. Sci. Technol. 1995, 29, 2267-2272. (35) Tremolada, P.; Calamari, D.; Gaggi, C.; Bacci, E. Chemosphere 1993, 11, 2235-2252.

(36) Calamari, D.; Tremolada, P.; Di Guardo, A.; Vighi, M. Environ. Sci. Technol. 1994, 28, 429-434. (37) Halsall, C. J.; Bailey, R.; Stern, G. A.; Barrie, L. A.; Fellin, P.; Muir, D. C. G.; Rosenberg, B.; Rovinsky, F. Y.; Kononov, E. Y.; Pastukhov, B. Environ. Pollut. 1998, 102, 51-62. (38) Cortes, D. R.; Basu, I.; Sweet, C. W.; Brice, K. A.; Hoff, R. M.; Hites, R. A. Environ. Sci. Technol. 1998, 32, 1920-1927. (39) Jantunen, L. M. M.; Bidleman, T. F.; Harner, T.; Parkhurst, W. J. Environ. Sci. Technol. 2000, 34, 5097-5105. (40) McConell, L. L.; Kucklick, J. R.; Bidleman, T. F.; Ivanov, G. P.; Chernyak, S. M. Environ. Sci. Technol. 1996, 30, 2975-2983. (41) Alegria, H. A.; Bidleman, T. F.; Shaw, T. J. Environ. Sci. Technol. 2000, 34, 1953-1958.

Received for review February 20, 2002. Revised manuscript received October 11, 2002. Accepted October 17, 2002. ES020605Q

VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

215