Environ. Sci. Technol. 1997, 31, 1811-1816
Temporal and Spatial Trends in a Long-Term Study of Gas-Phase PCB Concentrations near the Great Lakes BARBARA R. HILLERY,† ILORA BASU,† CLYDE W. SWEET,‡ AND R O N A L D A . H I T E S * ,† School of Public and Environmental Affairs and Department of Chemistry, Indiana University, Bloomington, Indiana 47405, and Illinois State Water Survey, Champaign, Illinois 61820
Atmospheric concentrations of gas-phase polychlorinated biphenyls (PCBs) were measured at three sites near the Great Lakes as part of the Integrated Atmospheric Deposition Network (IADN). Air samples were collected every 12 days beginning in November 1990. Mean yearly values ranged from 89 to 370 pg/m3. Concentrations at the more urban site near Lake Erie (Sturgeon Point) were about two times higher than concentrations at the two more remote sites near Lake Superior (Eagle Harbor) and Lake Michigan (Sleeping Bear Dunes). Multiple regression analysis was used to relate atmospheric PCB concentrations to meteorological conditions. As expected, atmospheric temperature had a very significant effect on gas-phase PCB concentrations. Including wind speed and wind direction in the regression analysis improved the predicted values only slightly. Atmospheric levels of PCBs have remained unchanged as a function of time near Lake Superior but have declined slightly over time near Lakes Michigan and Erie. The rate constants at the latter two sites indicate an environmental half-life in the atmosphere of approximately 6 yr, which agrees well with PCB half-lives in other environmental compartments.
Introduction Polychlorinated biphenyls (PCBs) enter the Great Lakes by deposition from the atmosphere. Eisenreich et al. (1) concluded that atmospheric deposition accounted for 85% of the total PCB input to the relatively clean waters of Lake Superior. Swackhamer et al. (2) found PCBs in the air, water, and sediments of Siskiwit Lake, an isolated lake on Isle Royale, an island in Lake Superior. The only reasonable explanation for the presence of these organic compounds in this insular water was atmospheric transport and deposition. However, the relative importance of atmospheric deposition to the overall contaminant loadings to the Great Lakes remains to be determined. Several attempts have been made to estimate the relative importance of atmospheric deposition (3-6). In each case, it was noted that insufficient information existed to make this determination. Specifically, it was noted that there were insufficient data and that sampling and analytical methods needed to be standardized. In an attempt to rectify this situation, the United States and Canada established the Integrated Atmospheric Deposition Network (IADN), designed * Corresponding author e-mail:
[email protected]. † Indiana University. ‡ Illinois State Water Survey.
S0013-936X(96)00990-X CCC: $14.00
1997 American Chemical Society
to collect regional data representative of the atmosphere near the lakes (5, 7, 8). The goals of IADN include long-term monitoring and research to determine loadings and temporal and spatial trends of persistent toxic substances in the Great Lakes System (9). Production of PCBs in the United States stopped in 1977, but due to their chemical stability, they have lingered in the environment. Considerable effort has been spent in determining their fate and clearance time. Concentrations have been decreasing in soils, sediment, peat, vegetation, and water (10-18). Studies in biota have shown concentrations to be decreasing with a half-life of approximately 6-15 yr, though occasionally no change was seen (19-30). Atmospheric PCB concentrations, however, seem to have remained constant over this time (31-37), which is anomalous given that the atmosphere is the transport medium to the other environmental compartments. Atmospheric concentrations of PCBs on a relatively longterm basis have been studied by several groups (32-37). Unfortunately, seasonal variability has obscured the longterm trends, if any. In addition, differences in the frequency and duration of previous studies have complicated attempts to understand temporal trends. Previous studies were also geographically diverse and thus may not be relevant to the Great Lakes. This paper will report on gas-phase PCB data obtained from three U.S. IADN sites: Eagle Harbor near Lake Superior, Sleeping Bear Dunes near Lake Michigan, and Sturgeon Point near Lake Erie. As much as 5 yr worth of data now exist for these sites, allowing us to investigate temporal and spatial trends in atmospheric PCB concentrations.
Experimental Methods Sampling Sites. IADN consists of five master sampling sites, located primarily on remote shorelines of the Great Lakes in the United States and Canada, as shown in Figure 1. Two of these sites are operated by Canada and three by the United States. We report here on data collected from the three U.S. sites. Sample collection began at Eagle Harbor, MI, near Lake Superior (latitude 47°27′47′′; longitude 88°08′59′′) in November 1990. Sample collection at Sleeping Bear Dunes, MI, near Lake Michigan (latitude 44°45′38′′; longitude 86°03′30′′) and at Sturgeon Point, NY, near Lake Erie (latitude 42°41′34′′; longitude 79°03′20′′) began in November 1991. The IADN sites were chosen to be representative of the atmosphere near the lakes and to minimize the influence of local sources. Sampling equipment was placed within 1 km of the shoreline. Sampling Protocol. Details of the sampling protocols and analytical procedures can be found elsewhere (8, 38). Briefly, air samples were collected using high-volume samplers (Graseby General Metal Works) fitted with automatic filter covers to prevent passive loading (Andersen SamplSaver). Samples were collected every 12 days for a period of 24 h. A calibrating orifice device (Graseby Calibration Kit G2535, Graseby General Metal Works) was taken to the sites every 3 months to check flow rates and adjust them as necessary. Rates were set at 34 m3/h, with a permanent manometer installed in each sampler to verify normal operation between calibrations. A quartz fiber filter was installed for the collection of particulates. Through May 4, 1992, gas-phase organics were collected on polyurethane foam (PUF; Olympic Foam Products). After that date, gas-phase organics were collected on Amberlite XAD-2 resin (Sigma Chemical Co.). All adsorbent materials were precleaned by Soxhlet extraction using a series of solvents of varying polarity. Each site was equipped with a 10-m tower for the collection of meteorological data. Instruments included wind speed and direction sensors (Met-One) installed at 10 m height and
VOL. 31, NO. 6, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1811
FIGURE 1. Locations of the five sampling sites of the Integrated Atmospheric Deposition Network. temperature, relative humidity, and solar radiance sensors (Campbell Scientific) installed at 2 m height. Data were recorded automatically every 6 s using a data-logger (Campbell Scientific) and output as mean hourly values. Analytical. PCBs and other compounds were removed from the adsorbent material by 24-h Soxhlet extraction using a 1:1 mixture of acetone and hexane (OmniSolv, EM Science). The extract was concentrated by rotary evaporation, and the solvent was exchanged to hexane. Silica gel (Aldrich Chemical) was used to fractionate the concentrated hexane extract, with the PCB fraction eluting in hexane. Samples were again concentrated, this time by nitrogen blow-down, before adding the quantitation standards and injecting them onto the gas chromatograph (GC). Chromatographic separation was performed on a Hewlett Packard 5890 GC. Initially, chromatography was done on 30 m × 250 µm i.d. (df ) 0.25 µm) DB-5 columns, but by 1994, all chromatography was done on 60 m × 250 µm i.d. (df ) 0.10 µm) DB-5 columns (both from J&W Scientific). The GC was equipped with an electron capture detector and an autosampler (Hewlett Packard 7673A). Injection of 1 µL was performed in the splitless mode. The carrier gas was hydrogen, and the detector makeup gas was nitrogen. The temperature program was 100 °C for 1 min, 1 °C/min to 240 °C, and 10 °C/min to 280 °C with a 20-min hold. The injector temperature was 250 °C, and the detector temperature was 350 °C. Data Manipulation. Approximately 100 individual PCB congeners were quantified by a Hewlett-Packard 3365 Series II Chemstation system using internal standards (PCB 30 or PCB 204) that had been added just prior to chromatographic analysis. The mass of each individual congener was calculated with relative response factors obtained using a standard mixture of Aroclors 1232, 1248, and 1262 in a ratio of 25:18: 18, as analyzed by Mullin (39). Individual masses were transformed to atmospheric concentrations using the actual air volume sampled. Each congener concentration was converted to a partial pressure using the ideal gas law, thus adjusting for the molecular weight of the compound and for the average atmospheric temperature during the sampling period. The partial pressures of the individual congeners were then summed to give a value for total PCBs. A productmoment correlation analysis indicated that the individual congener partial pressures were strongly correlated with the total PCB value; therefore, total partial pressures have been used for the investigation of trends and geographic comparisons. Average values are based on a meteorological year, where winter is December-February, spring is March-May, and so forth. Rather than have a year containing only 1 month, the initial November data (1990 for Eagle Harbor, 1991 for Sleeping Bear Dunes and Sturgeon Point) were incorporated into the subsequent year. Meteorological data were averaged over the 24-h period corresponding to each sampling event.
1812
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 6, 1997
FIGURE 2. Yearly mean gas-phase concentrations and standard errors of total PCB concentrations (in pg/m3) at the three U.S. IADN sampling sites. Quality Assurance. Extensive quality control (QC) sampling was used for this project. All sites had a second highvolume sampler for obtaining either field blanks or duplicate samples. Laboratory blanks and spiked matrix samples were run with every extraction batch (approximately 10 samples/ batch). Surrogate standards (PCBs 14, 65, and 166) were added to every sample prior to extraction to monitor recovery. All reported results were well within required QC limits (40). Field blanks averaged only a few pg/m3 for total gas-phase PCBs; therefore, data have not been corrected for blank values. Surrogate recoveries averaged about 94% for PCB 14 (an early eluter) and about 98% for PCB 166 (a late eluter); therefore, data have not been corrected for recoveries. When duplicate samples existed, both points were included in the data set rather than taking an average value.
Results and Discussion Concentrations. Yearly arithmetic average concentrations for each of the three sites are shown in Figure 2. Overall, yearly average values (( standard error) ranged from a low of 89 ( 11 pg/m3 at Eagle Harbor in 1991 to a high of 370 ( 51 pg/m3 at Sturgeon Point in 1993. These concentrations are well within the range of values reported for gas-phase PCB concentrations near the Great Lakes (2, 6, 32, 34, 37, 41, 42). Some spatial variation was seen in the results. The average values over all years were 128 ( 12 pg/m3 at Eagle Harbor, 160 ( 10 pg/m3 at Sleeping Bear Dunes, and 315 ( 20 pg/m3 at Sturgeon Point. Although the PCB concentrations at Eagle Harbor and Sleeping Bear Dunes are statistically indistinguishable, Sturgeon Point concentrations are always significantly higher, usually by a factor of 2-3. We had expected it to be even higher. PCBs are a product of anthropogenic activity, and Sturgeon Point is within 20 km of a major urban center (Buffalo, NY) with a population of well over 1 million people. In contrast, Eagle Harbor is situated in Keweenaw County, MI, with a population of fewer than 2000 people, and Sleeping Bear Dunes is situated in Leelanau County, MI, with a population of fewer than 20 000 people (1990 Census figures). Temperature Effects. Gas-phase concentrations of PCBs exhibit a seasonal trend (2, 31, 32, 34). This trend is largely the result of cyclical temperature variations, with lower concentrations occurring in the colder, winter months and higher concentrations in the warmer, summer months. To demonstrate the seasonality of the IADN PCB data, individual concentrations were plotted versus time, along with appropriate atmospheric temperatures for comparison (see Figures 3-5). Clearly, concentrations are cyclic. This can perhaps best be seen in the Eagle Harbor data (Figure 3), where upswings in the data correspond to approximately day numbers 150, 550, 1300, and 1600. However, there are
FIGURE 3. Fluctuation in gas-phase concentrations (in pg/m3) of total PCBs over time and daily mean atmospheric temperature (in K) for Eagle Harbor near Lake Superior.
FIGURE 5. Fluctuation in gas-phase concentrations (in pg/m3) of total PCBs over time and daily mean atmospheric temperature (in K) for Sturgeon Point near Lake Erie.
FIGURE 4. Fluctuation in gas-phase concentrations (in pg/m3) of total PCBs over time and daily mean atmospheric temperature (in K) for Sleeping Bear Dunes near Lake Michigan. considerable variations, with many of the higher concentrations at all three sites occurring in the cooler, spring and fall months. Thermodynamically, the gas-phase behavior of PCBs can be described by the Clausius-Clapeyron equation:
ln P )
1 + const (-∆H R )(T)
(1)
where P is the partial pressure in atm, T is the temperature in K, ∆H is the heat of vaporization in J/mol, and R is the gas constant. Clausius-Clapeyron plots were developed for each site to further examine the temperature dependence of atmospheric PCB concentrations. Each graph in Figure 6 shows the natural logarithm of the partial pressure for total PCBs (based on the summation of the partial pressures of individual congeners) versus the reciprocal of the average atmospheric temperature for the appropriate 24-h sampling period. These data cover the period from inception of sampling at each site to November 1995.
FIGURE 6. Clausius-Clapeyron plots relating the natural logarithm of the partial pressure of total PCBs (in atm) to the inverse temperature (in K). Clearly, Figure 6 shows that PCB partial pressures increase with temperature. A simple linear regression (eq 1 in Table 1) applied to these data gives the lines shown, with linear correlation coefficients (r) that are all significant at the 99% confidence level or higher. The heats of vaporization, calculated from the slopes of the regression lines, are 37 kJ/ mol at Eagle Harbor, 37 kJ/mol at Sleeping Bear Dunes, and 38 kJ/mol at Sturgeon Point. These values are consistent
VOL. 31, NO. 6, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1813
TABLE 1. Regression Parameters and Standard Deviations for Eqs 1-5a Eagle Harbor
Sleeping Bear Dunes
Sturgeon Point
Equation 1: ln P ) a0 + a1(1/T) a0 -17 ( 10% -16 ( 10% -15 ( 9% a1 -4501 ( 10% -4471 ( 10% -4603 ( 8% -r 0.625 0.641 0.710
a0 a1 a2 -r
Equation 2: ln P ) a0 + a1(1/T) + a2 ln WS -16 ( 15% -17 ( 10% -15 ( 9% -4542 ( 11% -4323 ( 12% -4383 ( 8% -0.01 ( 960% -0.10 ( 163% -0.39 ( 26% 0.609 0.642 0.744
a0 a1 a2 a3 -r
Equation 3: ln P ) a0 + a1(1/T) + a2 sin WD + a3 cos WD -17 ( 12% -16 ( 11% -16 ( 9% -4297 ( 13% -4452 ( 11% -4343 ( 9% -0.09 ( 84% -0.14 ( 57% 0.22 ( 29% -0.15 ( 71% -0.02 ( 325% -0.17 ( 45% 0.623 0.652 0.743
a0 a1 a2 -r
Equation 4: ln P ) a0 + a1(1/T) + a2 time -17 ( 10% -15 ( 11% -14 ( 9% -4429 ( 11% -4580 ( 10% -4742 ( 8% 0.000007 ( 1600% -0.00033 ( 36% -0.00027 ( 41% 0.622 0.664 0.724
Equation 5: ln P ) a0 + a1(1/T) + a2 ln WS + a3 sin WD + a4 cos WD + a5 time a0 -17 ( 11% -16 ( 11% -15 ( 9% a1 -4221 ( 13% -4426 ( 11% -4450 ( 8% a2 -0.01 ( 100% -0.18 ( 89% -0.32 ( 35% a3 -0.09 ( 90% -0.20 ( 39% 0.16 ( 40% a4 -0.17 ( 65% -0.05 ( 158% -0.08 ( 102% a5 0.00009 ( 130% -0.00037 ( 33% -0.00031 ( 34% -r 0.625 0.686 0.773 a Coefficients with a standard deviation less than 50% are shown in boldface.
with one another, but they are about half the previously reported heats of vaporization for PCBs (34, 36, 43). Falconer and Bidleman (43) have determined the vapor pressure of individual PCB congeners as a function of temperature under controlled laboratory conditions using gas chromatography. These experiments gave heats of vaporization ranging from 75 kJ/mol for trichlorinated PCBs to 100 kJ/mol for octachlorinated PCBs. Because atmospheric PCBs in the gas phase are predominantly the tri-, tetra-, and pentachlorinated homologues, environmental heats of vaporization will be near the lower end of the range reported by Falconer and Bidleman (43). In addition, Hoff et al. (34) reported PCB atmospheric concentrations at Egbert, Ontario, that gave a PCB heat of vaporization of 74 ( 10 kJ/mol, very close to the results of the controlled laboratory experiments. Panshin and Hites (36) reported a heat of vaporization of 63 ( 16 kJ/mol for PCBs sampled in Bloomington, IN. Both of these data sets were obtained over land. The current IADN data were obtained on the shores of the Great Lakes, and it seems likely that factors related to air-water exchange of PCBs could be responsible for the lower heats of vaporization. The heat of vaporization is an indication of the energy required to overcome attractive forces and convert 1 mol of PCBs from the liquid phase to the gas phase. At the air-water interface, the phase transfer process is complicated by issues such as the aqueous solubility of the compound. PCBs have very low aqueous solubilities in the range of 10-5-10-11 mol/L (44). This rejection by the water could account for the lower energy required for water-air phase transfer relative to land-air phase transfer. Wind Effects. While temperature is obviously an important parameter controlling the atmospheric gas-phase concentrations of semivolatile compounds, temperature alone fails to fully predict atmospheric concentrations of PCBs. The plots shown in Figure 6 give coefficients of determination (r2)
1814
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 6, 1997
on the order of 40%. Statistically, this implies that only about half of the variance in partial pressure is influenced by temperature, with the rest being attributable to other factors (including the analytical variability of about (30%). Because wind has been shown to have some correlation to atmospheric concentrations of selected semivolatile organic compounds (35, 41, 42, 45, 46), we looked at wind speed to see if this could be a controlling parameter. Starting with the known parameters of the Clausius-Clapeyron equation, we added wind speed as a factor, giving eq 2:
ln P ) a0 + a1
(T1) + a ln WS 2
(2)
where WS is the wind speed in mph. A multiple regression analysis applied to these data gave the results shown in Table 1 for eq 2. This analysis again shows the predominant effect of temperature. While incorporating wind speed gave slight improvements in some of the correlations, these improvements are probably due to little more than the increased number of parameters in the equation. Table 1 shows the very large standard deviations associated with the wind speed parameter (a2) at Eagle Harbor ((960%) and at Sleeping Bear Dunes ((163%). Conversely, the standard deviation at the more urban site, Sturgeon Point, is only (26%. Thus, local wind speed may be an important variable in near-urban areas, where short-range transport of atmospheric pollutants is more likely to affect gas-phase concentrations. Wind direction may also be a controlling parameter for total PCB partial pressures. Given that wind direction is measured as an angle relative to true north (0°), mean direction was determined by using trigonometric relations to determine the direction of the resultant of individual wind vectors. These data were incorporated into the multiple regression, giving eq 3:
(T1) + a sin WD + a cos WD
ln P ) a0 + a1
2
3
(3)
where WD is the wind direction in degrees. The results for this equation are shown in Table 1, eq 3. Again, there are geographical differences. Wind direction seems to have no significance at Eagle Harbor, where both the sine and the cosine terms have coefficients (a2 and a3) with standard deviations of about 80%. This is similar to the situation found by Hoff et al. (45), who reported that air trajectory data had little relevance to PCB concentrations at Egbert, Ontario. While it may seem obvious that concentrations would be higher downwind from sources, this simplification only applies to transport over short distances and time periods. PCBs are environmentally stable, and thus over the long term, their transport will be affected by their tendency to continually recycle through the atmosphere. The rates of volatilization and deposition are dependent on the chemical parameters of the PCBs and environmental conditions. At Sleeping Bear Dunes, the cosine term (a3) is still insignificant, having a coefficient with a standard deviation of over (300% (see Table 1, eq 3). In this case, however, the standard deviation for the coefficient for the sine term (a2) drops to about (57%. This coefficient has a negative value. Because the sine of angles between 180° and 360° are negative, we conclude that winds coming from the west might influence atmospheric PCB concentrations at Sleeping Bear Dunes. Winds from the west generally would be winds coming across Lake Michigan from Green Bay. At Sturgeon Point, wind direction becomes still more relevant. In this case, the sine term (a2) is positive with a standard deviation of only (29%, and the cosine term (a3) is negative with a standard deviation of (45%. Considering both terms reduces the significant wind direction to a single quadrant, which corresponds to the angles between 90° and
TABLE 2. Half-Lives of PCBs in Various Environmental Compartments compartment
half-life
ref
air, Lake Superior air, Bermuda air, Bloomington water, Lake Superior water, Lake Michigan vegetation, U.K. moss, Norway soil, U.K. mussels, Mediterranean bloaters, Lake Michigan trout, Lake Michigan trout, Lake Michigan trout, Lake Ontario pike, Baltic Sea spot-tail shiners, Great Lakes fish, Lake Michigan gull eggs, Lake Ontario gull eggs, Lake Erie gull eggs, Lake Superior gull eggs, Lake Ontario Arctic ringed seals Arctic polar bears
no change (in 10 yr) no change (in 23 yr) no change (in 7 yr) 2.5-6.3 9 4 6-14 3-17 9 8 4 6.3 10 9 5 7 6 no change (in 10 yr) g18 g21 6 15
Baker and Eisenreich (33) Panshin and Hites (35) Panshin and Hites (36) Jeremiason et al. (17) Pearson et al. (18) Jones et al. (16) Lead et al. (15) Alcock et al. (10) Sole et al. (26) Hesselberg et al. (21) Devault et al. (20) Miller et al. (23) Borgmann and Whittle (22) Moilanen et al. (19) Suns et al. (24) Stow (25) Irwin and Lageroos (27) Irwin and Lageroos (27) Smith (28) Smith (28) Addison et al. (29) Norstrom et al. (30)
180°. Thus, the winds having the greatest influence on atmospheric PCB concentrations at Sturgeon Point seem to be coming from the southeast, across land rather than across Lake Erie. Temporal Trends. One of the objectives of IADN is to identify temporal trends in concentrations of persistent substances such as PCBs and to determine if atmospheric levels have been declining since PCBs were banned in 1976. The atmosphere, however, is a dynamic system, subject to both cyclical and non-cyclical fluctuations. There are major fluctuations in temperature, for instance, and this is responsible for seasonal trends. Thus, any attempt to determine temporal trends must consider atmospheric temperature. Starting with the relationship between the natural logarithm of the partial pressure and the inverse temperature, we added time as a parameter to our multiple regression, giving eq 4:
ln P ) a0 + a1
(T1) + a time 2
(4)
Notice that a2 is a first-order rate constant. The results for eq 4 are shown in Table 1. Not surprisingly, temperature is still the predominant parameter, but the coefficients for the time term (a2) are interesting. For Eagle Harbor, this term has a standard deviation of about (1600%; however, note that the standard deviation for this term at Sleeping Bear Dunes is (36% and at Sturgeon Point is (41%. At both Sleeping Bear Dunes and Sturgeon Point, the time coefficients are about the same and negative, indicating a decline in partial pressures over time. To make sure that the results were not an artifact caused by including the 1991 data only for Eagle Harbor, the regression was repeated for all sites using only values from 1992 onward. To make sure that the results were not an artifact caused by switching from PUF to XAD resin for sample collection in May 1992, the regression was repeated for all sites using only data from this date onward. In both cases, there were slight differences in the numbers obtained, but no differences in the trends. Therefore, results are presented using the entire data set. It is interesting to note that a simple display of average concentrations as a function of time, such as shown in Figure 2, does not reveal this temporal trend. The effect of temperature is so strong that it must be included as a regression parameter in order to determine atmospheric changes as a function of time. Slight differences in annual
average temperature can obscure temporal trends if the data are simply averaged on this scale. After investigating these parameters stepwise, it was important to combine all of the parameters into one equation, giving eq 5:
ln P ) a0 + a1
(T1) + a ln WS + a sin WD + 2
3
a4 cos WD + a5 time (5) The overall statistical evaluation does not change much; see Table 1 for eq 5. Temperature (see the a1 term) remains the dominant effect. Wind speed (see the a2 term) continues to exert an influence only at Sturgeon Point. Wind direction is less significant at Sturgeon Point, with the standard deviation of the cosine term (a4) increasing from (45% to (102%. Most importantly, the time parameter (see the a5 term) does not change much for Sleeping Bear Dunes and Sturgeon Point. The average of the time coefficients for the two sites is -0.00034 versus -0.00030 for eq 4, and the standard deviation has improved slightly. Based on an average rate constant of 0.00034 day-1, the environmental half-life of PCBs in the atmosphere near Lake Michigan and Lake Erie is about 6 yr. This differs considerably from previously reported atmospheric PCB data; see the first three rows of Table 2. Panshin and Hites (36) sampled air in Bloomington, IN, in 1993 and compared the results to the 1986-1987 study of Hermanson and Hites (32). They found no evidence that atmospheric concentrations were changing. Panshin and Hites (35) also sampled air coming in off the ocean to Bermuda. These results were compared to 20 years of literature data, and again no decline in atmospheric PCB concentrations could be detected. These later experiments focused on background atmospheric PCB levels, and these authors avoided any local influence by sampling only when the wind was blowing over the open ocean. Thus, the situation in Bermuda may be similar to the situation at Eagle Harbor, where there is no statistical evidence that atmospheric PCB levels are changing. This result is similar to that of Baker and Eisenreich (33), who reported that atmospheric PCB concentrations over Lake Superior remained unchanged during a 10-yr period. The half-life of about 6 yr for atmospheric PCBs at Sleeping Bear Dunes and Sturgeon Point compares favorably with trends observed in other environmental compartments; see Table 2. PCBs have been declining in water with half-lives
VOL. 31, NO. 6, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1815
of 3-9 yr (17, 18). A half-life of 4 yr can be estimated for PCB concentrations in vegetation (16), 6-14 yr for moss (15), and 3-17 yr for soil (10). PCBs in mussels in Fangar Bay in the western Mediterranean declined with a half-life of about 9 yr, though mussels in another area (Alfacs Bay) of the region showed no decline (26). The average half-life of PCBs in fish is about 7 yr (19-25), but in gull eggs, it seems to vary from about 6 to over 25 yr (27, 28). For both fish and gull eggs, determining temporal trends in PCBs is complicated by internal lake processes such as changes in prey dynamics (28). The half-life of PCBs in Arctic ringed seals (29) is about 6 yr, and in polar bears (30) it is about 15 yr. This latter value may be higher due to the paucity of data available for analysis of temporal trends, or it may be attributable to the higher trophic level of these bears. In contrast to these clear changes in the water and biota PCB concentrations, the lack of any observable change (until now) in the atmospheric PCB concentrations was peculiar. It now seems likely that long-term changes were hidden by large seasonal (temperature) effects. In addition, most longterm comparisons were based on data obtained by different researchers, using different analytical methods and different time periods, making direct comparisons inconclusive. This problem was one of the reasons for the creation of IADN. The atmosphere is a dynamic system, and it is impossible to account for its normal variations without long-term study. While this study is long compared to other such studies, it is not yet long enough to sufficiently account for normal climatic variability. Thus, it is important to continue to make measurements of PCBs (and other compounds) at these sites.
Acknowledgments The authors thank the U.S. Environmental Protection Agency’s Great Lakes National Program Office for funding (Grant GL995656).
Literature Cited (1) Eisenreich, S. J.; Looney, B. B.; Thornton, J. D. Environ. Sci. Technol. 1981, 28, 30-38. (2) Swackhamer, D. L.; McVeety, B. D.; Hites, R. A. Environ. Sci. Technol. 1988, 22, 664-672. (3) Strachan, W. M. J.; Eisenreich, S. J. Mass balancing of toxic chemicals in the Great Lakes: The role of Atmospheric Deposition; International Joint Commission: Ontario, Canada, 1988. (4) Eisenreich, S. J.; Strachan, W. M. J. Estimating atmospheric deposition of toxic substances to the Great Lakessan update; Report of the Gray Freshwater Biological Institute; University of Minnesota: Navarre, MN, 1992. (5) Hoff, R. M.; Strachan, W. M. J.; Sweet, C. W.; Chan, C. H.; Shackleton, M.; Bidleman, T. F.; Brice, K. A.; Burniston, D. A.; Cussion, S.; Gatz, D. F.; Harlin, K.; Schroeder, W. H. Atmos. Environ. 1996, 30, 3505-3527. (6) Sweet, C. W.; Murphy, T. J.; Bannasch, J. H.; Kelsey, C. A.; Hong, J. J. Great Lakes Res. 1993, 19, 109-128. (7) Hillery, B. R.; Hoff, R. M.; Hites, R. A. In Atmospheric Deposition of Contaminants to the Great Lakes and Coastal Waters; Baker, J. E., Ed.; SETAC Press: Pensacola, FL, 1997. (8) Gatz, D. F.; Sweet, C. W.; Basu, I.; Vermette, S.; Harlin, K.; Bauer, S. Great Lakes Integrated Atmospheric Deposition Network Data Report 1990-1992; Illinois State Water Survey: Champaign, IL, 1994. (9) Great Lakes Water Quality Agreement of 1978, as amended 1987. International Joint Commission, United States and Canada, 1994. (10) Alcock, R. E.; Johnston, A. E.; McGrath, S. P.; Berrow, M. L.; Jones, K. C. Environ. Sci. Technol. 1993, 27, 1918-1923. (11) Pavoni, R.; Sfriso, A.; Marcomini, A. Mar. Chem. 1987, 21, 2535. (12) Eisenreich, S. J.; Capel, P. D.; Robbins, J. A.; Bourbonniere, R. Environ. Sci. Technol. 1989, 23, 1116-1126.
1816
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 6, 1997
(13) Kjeller, L.; Rappe, C. Environ. Sci. Technol. 1995, 29, 346-355. (14) Rapaport, R. A.; Eisenreich, S. J. Environ. Sci. Technol. 1988, 22, 931-941. (15) Lead, W. A.; Steinnes, E.; Jones, K. C. Environ. Sci. Technol. 1996, 30, 524-530. (16) Jones, K. C.; Sanders, G.; Wild, S. R.; Burnett, V.; Johnston, A. E. Nature 1992, 356, 137-140. (17) Jeremiason, J. D.; Hornbuckle, K. C.; Eisenreich, S. J. Environ. Sci. Technol. 1994, 28, 903-914. (18) Pearson, R. F.; Hornbuckle, K. C.; Eisenreich, S. J., Swackhamer, D. L. Environ. Sci. Technol. 1996, 30, 1429-1436. (19) Moilanen, R.; Pyysalo, H.; Wickstroem, K.; Linko, R. Bull. Environ. Contam. Toxicol. 1982, 29, 334-340. (20) DeVault, D. S.; Clark, J. M.; Lahvis, G. J. Great Lakes Res. 1988, 14, 23-33. (21) Hesselberg, R. J.; Hickey, J. P.; Nortrup, D. A.; Willford, W. A. J. Great Lakes Res. 1990, 16, 121-129. (22) Borgmann, U.; Whittle, D. M. J. Great Lakes Res. 1991, 17, 368381. (23) Miller, M. A.; Madenjian, C. P.; Masnado, R. G. J. Great Lakes Res. 1992, 18, 742-754. (24) Suns, K. R.; Hitchin, G. H.; Toner, D. J. Great Lakes Res. 1993, 19, 703-714. (25) Stow, C. A. Environ. Sci. Technol. 1995, 29, 522-527. (26) Sole, M.; Porte, C.; Pastor, D.; Albaiges, J. Chemosphere 1994, 28, 897-903. (27) Irwin, R. J.; Lageroos, D. Toxic Air Pollution in the Great Lakes Basin: A Call for Action; Sierra Club, Madison, WI, 1988. (28) Smith, D. W. Environ. Sci. Technol. 1995, 29, 740-750. (29) Addison, R. F.; Zinck, M. E.; Smith, T. G. Environ. Sci. Technol. 1986, 20, 253-256. (30) Norstrom, R. J.; Simon, M.; Muir, D. C. G.; Schweinsburg, R. E. Environ. Sci. Technol. 1988, 22, 1063-1071. (31) Manchester-Neesvig, J. B.; Andren, A. W. Environ. Sci. Technol. 1989, 23, 1138-1148. (32) Hermanson, M. H.; Hites, R. A. Environ. Sci. Technol. 1989, 23, 253-1258. (33) Baker, J. E.; Eisenreich, S. J. Environ. Sci. Technol. 1990, 24, 342352. (34) Hoff, R. A.; Muir, D. C. G.; Grift, N. P. Environ. Sci. Technol. 1992, 26, 266-275. (35) Panshin, S. Y.; Hites, R. A. Environ. Sci. Technol. 1994, 28, 20012007. (36) Panshin, S. Y.; Hites, R. A. Environ. Sci. Technol. 1994, 28, 20082013. (37) Monosmith, C. L.; Hermanson, M. H. Environ. Sci. Technol. 1996, 30, 3464-3472. (38) Basu, I. Analysis of PCBs and Pesticides in Air and Precipitation Samples, IADN Project Standard Operating Procedure; Indiana University: Bloomington, IN, 1995. (39) Mullin, M. D. PCB Workshop; U.S. EPA Large Lake Research Station: Grosse Ile, MI, 1985, updated 1994. (40) Basu, I.; Hillery, B. R.; Hites, R. A. Quality Assurance Project Plan v. 3, Integrated Atmospheric Deposition Network; Indiana University: Bloomington, IN, 1995. (41) Hornbuckle, K. C.; Sweet, C. W.; Pearson, R. F.; Swackhamer, D. L.; Eisenreich, S. J. Environ. Sci. Technol. 1995, 29, 869-877. (42) Hornbuckle, K. C.; Jeremiason, J. D.; Sweet, C. W.; Eisenreich, S. J. Environ. Sci. Technol. 1994, 28, 1491-1501. (43) Falconer, R. L.; Bidleman, T. F. Atmos. Environ. 1994, 28, 547554. (44) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M. Environmental Organic Chemistry; John Wiley and Sons: New York, 1993; p 77. (45) Hoff, R. A.; Muir, D. C. G.; Grift, N. P. Environ. Sci. Technol. 1992, 26, 276-283. (46) Burgoyne, T. W.; Hites, R. A. Environ. Sci. Technol. 1993, 27, 910-913.
Received for review November 27, 1996. Revised manuscript received January 25, 1997. Accepted February 3, 1997.X ES960990H X
Abstract published in Advance ACS Abstracts, April 15, 1997.
