Environ. Sci. Technol. 2002, 36, 5051-5056
Gas-Phase Polychlorinated Biphenyl and Hexachlorocyclohexane Concentrations near the Great Lakes: A Historical Perspective STEPHANIE S. BUEHLER, ILORA BASU, AND RONALD A. HITES* Environmental Science Research Center, School of Public and Environmental Affairs and Department of Chemistry, Indiana University, Bloomington, Indiana 47405
The Integrated Atmospheric Deposition Network (IADN) has been measuring gas-phase, polychlorinated biphenyl (PCB) concentrations at sites near Lakes Michigan and Superior for over a decade. Data through 2000 were used in this study to investigate PCB temporal trends in the Great Lakes atmosphere. Decreasing trends were found at both sites, and half-lives of approximately 20 yr were calculated using IADN data. However, when these data were supplemented by historical data for Lakes Michigan and Superior dating back to 1977, half-lives dropped to 10 and 6 yr, respectively. These latter half-lives agreed well with half-lives in other environmental compartments. Exponential curves fitted to the historical and IADN data indicated little decline in PCB concentrations in the basin since the mid-1990s. A similar historical analysis of Rand γ-hexachlorocyclohexane (HCH) data indicated that IADN data were the best predictor of trends, resulting in halflives of around 4 yr for both compounds. γ-HCH concentrations, however, have shown little decline in recent years, most likely because of its continuing use. PCB and R-HCH temporal trends indicated that bans on these substances have helped to remove them from the atmosphere. This work also showed that decades of data may be necessary to properly interpret long-term temporal trends in gas-phase organochlorine concentrations.
Introduction The regulation of persistent organic pollutants (POPs) in the Great Lakes basin started with the the elimination of point sources. As these sources were eliminated, scientists turned their attention to other mechanisms by which pollutants were transported into the Lakes, and these studies indicated that atmospheric deposition was a major source of POP contamination (1-4). Thus, to better understand the role of the atmosphere as a cause of Great Lakes contamination, the Integrated Atmospheric Deposition Network (IADN) was created under Annex 15 of the 1987 amended version of the 1978 Great Lakes Water Quality Agreement (5). IADN is a joint effort of the Canadian and U.S. governments, with multiple air sampling stations on each of the Great Lakes. The network was established to provide long-term measurements of atmospheric concentrations of POPs, and IADN uses these data to determine loadings and spatial and temporal trends of these chemicals to the Great Lakes. This * Corresponding author e-mail:
[email protected]. 10.1021/es0207091 CCC: $22.00 Published on Web 11/02/2002
2002 American Chemical Society
network has been in operation for 12 yr, and it has obtained atmospheric data on 27 POPs from 14 sites. IADN now provides a unique, long-term, continuous database for the region. Previous IADN studies have used this extensive database to explore temporal trends in gas-phase polychlorinated biphenyl (PCB) concentrations in the Great Lakes atmosphere (6, 7). With as much as 10 yr of data now available at the IADN sites, this paper will extend the work of Hillery et al. (6) and Simcik et al. (7) with the addition of three years of data (1998-2000). We will investigate gas-phase total PCB temporal trends at two U.S. sites: Sleeping Bear Dunes on Lake Michigan and Eagle Harbor on Lake Superior. This paper will not only look at trends based solely on IADN data but will also put the IADN data in a historical context by looking at total PCB gas-phase concentrations as far back as 1977. This approach will also be used to investigate gas-phase Rand γ-hexachlorocyclohexane (HCH) concentrations near Lake Michigan. By examining PCBs and HCHs on a longer time scale than before, we will gain a better understanding of their behavior in the Great Lakes basin. We will also be able to determine how regulations have reduced the atmospheric concentrations of PCBs and HCHs.
Experimental Section Sampling Methodology. The sampling site at Sleeping Bear Dunes, MI (latitude 44°45′38′′; longitude 86°03′30′′), is approximately 1 km from the shore of Lake Michigan. Sampling at this station on the northwest coast of Michigan’s Lower Peninsula began in December 1991. The Lake Superior site, approximately 50 m from the shore, is located at Eagle Harbor, MI (latitude 47°27′47′′; longitude 88°08′59′′) on the Keweenaw Peninsula. Sampling began there in November 1990. Both stations are located in rural areas, well removed from urban influences. A brief synopsis of the sampling and analytical procedures will be presented here, but full details can be found elsewhere (6, 8). Air samples were collected using a modified Anderson high-volume sampler (General Motor Works, model GS 2310) every 12 days for 24 h. Air was pulled through the sampler at a rate of 34 m3/h. Particles were collected on quartz fiber filters (Whatman QM-A), and gas-phase compounds were collected on 40 g of XAD-2 resin (Sigma, Amberlite 20-60 mesh). Prior to May 4, 1992, gas-phase compounds were collected on polyurethane foam. Hourly averages for air temperatures were collected at 2 m height. Analytical Methodology. The adsorbent material was Soxhlet extracted for 24 h with a mixture of 50% acetone in hexane. The extracts were then concentrated and solventexchanged to hexane by rotary evaporation. Concentrated extracts were fractionated on 3.5% (w/w) water deactivated silica gel, with the PCB fraction eluting with hexane and the HCHs eluting with 50% dichloromethane in hexane. The extracts were further concentrated under a stream of nitrogen and spiked with PCB-30, PCB-65, PCB-155, and PCB-204 as quantitation standards. Both PCBs and HCHs were analyzed on a Hewlett-Packard 5890 gas chromatograph with a 63Ni electron capture detector. A 60 m by 250 µm i.d. (film thickness 0.10 µm) DB-5 column was used. Extensive quality control measures were implemented throughout the analytical procedure (9). Field and laboratory blanks as well as duplicate samples were collected regularly. Laboratory blank values were less than 5% of actual sample values, while field blank values were on average less than 10% of sample values, eliminating the need for blank correction. Field and lab duplicates were on average within VOL. 36, NO. 23, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5051
30% and 10% of each other, respectively. For every batch of 10 samples, a matrix spike was included to monitor extraction efficiencies, as were surrogate spikes in each sample. Recoveries ranged from 77% to 107% (9) for both PCBs and HCHs. Data Analysis. The amount of data available from the two sites varied. Gas-phase data from January 1991 to December 2000 were used for Eagle Harbor on Lake Superior, while gas-phase data from January 1992 to December 2000 were used for Sleeping Bear Dunes on Lake Michigan. Sample concentrations were determined by dividing the measured mass of each analyte by the volume of air sampled. Results for duplicate samples were averaged to give a single value, which was then used throughout all subsequent calculations. Concentration data can be obtained by submitting a data request through the IADN website at http://www.mscsmc.ec.gc.ca/iadn/data/form/form_e.html. Partial pressures were calculated for each sample using the ideal gas law, which incorporated the compound’s concentration and the average air temperature during sampling. Approximately 100 individual PCB congeners are analyzed by our procedure, but total PCBs (∑PCB) are used in this paper because we have found that individual congeners are strongly correlated to their total (6, 10). To obtain a partial pressure for ∑PCB, individual congener partial pressures were summed together. Temperature has a large effect on the atmospheric partial pressures of these compounds (6, 7, 11-13). Much of this effect can be modeled using the Clausius-Clapeyron equation:
ln P )
-∆H 1 + const R T
()
(1)
where P is the partial pressure of the compound in atmospheres, ∆H is a phase transition energy of the compound (in kJ/mol), R is the gas constant, and T is the air temperature (in Kelvin) averaged over the 24-h sampling period. This temperature effect can be removed by adjusting the partial pressures to a reference temperature of 288 K by using the following equation:
P288 ) P exp
1 1 [-∆H R (288 T)]
(2)
where P288 is the compound’s partial pressure at the reference temperature of 288 K and -∆H/R is taken from eq 1. This approach provides results similar to and not significantly different from the multiple linear regression technique used in previous papers (6, 7, 12, 14). Thus, P288 values will be used throughout this paper to allow for comparisons between IADN and historical data. Historical Data. All PCB and HCH data prior to 1991 for Lake Superior and 1992 for Lake Michigan are considered historical data, and these data were taken from previously published papers (2, 13, 15-25). The analytical methods used in these studies were similar to those used by IADN. The same calculations used for the IADN data were used on the historical data with a few adjustments. ∑PCB concentrations were taken directly from the literature. Most of the papers did not list the concentrations of individual PCB congeners, so ∑PCB values were the only measurements available; these totals differed slightly from study to study. To calculate partial pressures for these concentrations, a weighted molecular weight average for ∑PCB, based on individual congeners measured by IADN, was used. This average was 276 g/mol. This method provides a reasonable estimation of molecular weight since the congeners summed by IADN and individual researchers are similar. Thus, partial pressures were calcu5052
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 23, 2002
lated for ∑PCB and not summed from individual congener partial pressures as with the IADN data. Much of the temperature data needed to calculate partial pressures were not available in the original studies. In these instances, data from the National Data Buoy Center (NDBC), part of the National Weather Service of the National Oceanic and Atmospheric Administration, were used. Buoys located close to the study’s sampling sites or those that had archived data were used for Lakes Michigan and Superior. NDBC temperature data were averaged over the sampling time given in each study to obtain a single temperature measurement for each sample. Temperature-corrected partial pressures, P288, were calculated for each historical sample using the slope (-∆H/R) obtained from the IADN data. Annual averages of the historical data represent all data found in that year, which is not necessarily all from one study. Data from only Lakes Superior and Michigan were used here because historical atmospheric concentration data for POPs for the other Great Lakes have not been published.
Results and Discussion IADN ∑PCB Temporal Trends. Hillery et al. (6) and Simcik et al. (7) looked at IADN gas-phase ∑PCB trends through 1995 and 1997, respectively. As a continuation of these studies, this paper will look at IADN ∑PCB data through 2000. Given that Hillery et al. (6) summed different congeners than Simcik et al. (7) to find ∑PCB and that Simcik et al. (7) separated the ∑PCB data into over-water and over-land samples before temporal analysis and only used data through August of 1997, all results presented here have been recalculated. The individual congeners summed to calculate ∑PCB in this study are the same as those used by Simcik et al. (7), and all time periods include data through December of a given year. Any differences between values reported here and in previous studies are a result of these different summation strategies. By plotting the natural logarithm of the temperaturecorrected ∑PCB partial pressure (P288) against time, the temporal trends at each IADN site can be examined. Figure 1 shows such a plot for Sleeping Bear Dunes on Lake Michigan. Trends for ∑PCB data through 1995, through 1997, and through 2000 are represented by the three separate lines. These periods represent the time frames of the studies by Hillery et al. (6), Simcik et al. (7), and this study, respectively. The environmental half-lives reported here represent the offsets between the elimination of gas-phase pollutants through atmospheric oxidation reactions and sediment and soil burial, on one hand, and the continuous release of these pollutants from reservoirs in water, soil, and biota, on the other. Half-lives were calculated by dividing the slope of each line into the natural logarithm of 2. Lake Michigan gas-phase ∑PCB for all time periods shows statistically significant declines over time, but these trends vary greatly depending on how much of the IADN data are used. For data through 1995, ∑PCB has a half-life of 3 yr. When two more years of data are added, the half-life increases to 5 yr. When ∑PCB data through 2000 are added, the halflife quadruples to 20 yr. The increase in half-life from 1995 to 1997 is most likely due to some high ∑PCB concentrations at the end of 1997. The unusually high half-life for data through 2000 is likely due to the apparent increase in ΣPCB concentrations in 1998 and 1999, represented by the higher values in Figure 1 from 2900 to 3800 days. ∑PCB concentrations in Lake Superior also had varying half-lives over time, although these data showed much more scatter than that for Lake Michigan. ∑PCB through 1995 had no statistically significant declining trend. With the addition of data through 1997, a half-life of 9 yr was observed, almost twice as long as the same period at our Lake Michigan site. Analysis of all IADN ∑PCB data through 2000 at Lake Superior
FIGURE 1. Natural logarithm of gas-phase ∑PCB temperature-corrected partial pressures as a function of time for the IADN sampling station at Sleeping Bear Dunes on Lake Michigan. The black, red, and blue lines represent the temporal trends for data through the end of 1995, 1997, and 2000, respectively. All trends are significant at the 95% confidence level or better. Correlation coefficients (r) and half-lives (t1/2) with standard errors are given for each trend line. gave a half-life of 19 yr, similar to that found at Lake Michigan for the same period. An apparent increase in ∑PCB concentrations was also observed at this site, with levels beginning to rise in late 1999 and maximizing in 2000. Since PCBs have been banned for almost 25 yr, there should be no new anthropogenic sources of these compounds. With no sources, one would expect ∑PCB concentrations to continue to decline as their supply in terrestrial and aquatic surfaces diminishes. Thus, the apparent increase in ∑PCB atmospheric concentrations near the Great Lakes in 1998-1999 is anomalous, and we explored several hypotheses to explain this phenomenon. Lake water levels in the basin in recent years have been very low, possibly exposing contaminated near-shore sediments and allowing PCBs to volatilize into the atmosphere (26-30). One of the strongest El Ninos on record hit the United States in 1998, producing less ice cover than normal on the Great Lakes (31, 32), with the potential of increasing atmospheric PCB levels. We were unable to statistically link any of these events to the apparent increase in ∑PCB concentrations in the Great Lakes atmosphere in 1998-1999. The previous analyses by Hillery et al. (6) and Simcik et al. (7) relied solely on IADN data to infer long-term trends in atmospheric concentrations of ∑PCB. Given the abnormal behavior of recent trends from this study, it seemed necessary to examine the IADN data from a different perspective in order to gain a better understanding of how ∑PCB are behaving in the Great Lakes atmosphere. To this end, the data presented above were analyzed in a broader, historical context by incorporating ∑PCB data that had been collected previously by other researchers. Historical ∑PCB Temporal Trends. Table 1 shows average annual ∑PCB concentrations and sources for historical data as well as all IADN data before temperature correction. Eight years of gas-phase ∑PCB data from 1978 to 1990 were added to Lake Superior atmospheric data collected by IADN, while seven years of data from 1977 to 1991 were found to supplement the Lake Michigan IADN data. All of these data were plotted as annual average, temperature-corrected, partial pressures (P288) versus time. The plots for ∑PCB for Lakes Michigan and Superior are shown in Figure 2. With ∑PCB data as far back as 1977, a trend can be tracked from the time this substance was banned until present. Not only
does this approach present a more complete picture of gasphase ∑PCB contamination in the Great Lakes atmosphere, but these graphs provide a better indication of where the IADN data fit into the overall picture of ∑PCB trends. The data for both lakes (see Figure 2) follow exponential decreases showing first-order rate loss processes for ∑PCB. Both lakes showed large declines in ∑PCB concentrations from the late 1970s to the mid-1990s, when concentrations started to become relatively steady. Although the IADN data themselves do show some fluctuations, the overall exponential curve indicates that, since the production of PCBs was banned in 1977, levels in the Great Lakes basin have decreased by a factor of 7-10 with annual concentrations dropping from around 1000 to approximately 100 pg/m3. Using the rate constants from the curves in Figure 2, halflives for ∑PCB in air near each lake can be calculated. ∑PCB in Lake Michigan’s atmosphere has a half-life of 10 ( 1.8 yr, while Lake Superior’s data indicate a half-life of 5.7 ( 0.7 yr. These values are in contrast to the half-lives of approximately 20 yr found using IADN data through 2000. Given that these two lakes have different physical characteristics, we would expect PCBs to behave differently at each site, resulting in different half-lives over each lake, and indeed, there are different half-lives: 6 yr for Lake Superior and 10 yr for Lake Michigan. However, given the much greater volume and lower water temperature of Lake Superior as compared to Lake Michigan, we would expect that the half-life for Lake Superior would be greater than for Lake Michigan, which is the opposite of what we observe. We conclude that the precision of these half-life estimates is such that the values for the two lakes cannot be distinguished from one another. The half-lives of 6-10 yr for gas-phase ∑PCB at Lakes Superior and Michigan compare well with half-lives found in other media at both lakes, as shown in Table 2. PCB halflives in gull eggs, trout, and smelt in Lake Superior are on average 11 ( 2 yr (34, 35); half-lives for PCBs in water and settling particles in the lake are approximately 3 yr (34). Smith (34) found a half-life of 4 yr for PCBs in the air over Lake Superior using mostly pre-IADN data. PCBs in Lake Michigan water also have a half-life of 4 yr (35). PCBs in fish and trout in Lake Michigan have a half-life of approximately 6 yr (36, 37). All of these rates are similar to the rate at which ∑PCB in the atmosphere are declining in this study. When recent VOL. 36, NO. 23, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5053
TABLE 1. Annual Average Atmospheric Concentrations for Lake Superior and Lake Michigan ∑PCB and Lake Michigan HCHsa year
concn (pg/m3)
1978 1979 1980 1984 1985 1986 1988 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
1500 900 1000 470 210 1300 350 190 88 89 110 92 120 58 66 60 100 100
year
concn (pg/m3)
1977 1984 1985 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
870 470 210 160 630 280 750 170 240 120 89 85 91 190 210 80
year
r-HCH (pg/m3)
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
210 140 100 190 130 140 85 64 64 69 48 48
Lake Superior ∑PCB standard error 220 140 350 80 52 210 52 33 12 14 16 27 21 6.9 12 8.7 12 22 Lake Michigan ∑PCB standard error 130 80 52 35 88 210 160 21 29 14 13 14 25 34 33 12
n
source
13 8 8 10 4 5 3 7 28 23 24 20 24 26 31 24 25 29
2 15 15 16 16 17 18 18 IADN IADN IADN IADN IADN IADN IADN IADN IADN IADN
n
source
7 10 4 9 31 2 12 22 28 24 27 28 23 22 29 30
20 16 16 21 22, 23 13 13, 24 IADN IADN IADN IADN IADN IADN IADN IADN IADN
Lake Michigan HCHs standard γ-HCH standard error (pg/m3) error 19 44 15 18 11 14 4.6 4.3 5.4 10 4.5 5.0
25 42 51 50 58 40 20 29 24 40 23
6.9 11 11 13 15 10 5.2 5.8 5.8 11 6.4
n
source
11 17 12 25 29 21 27 26 28 21 30 31
25 13, 25 13, 24 IADN IADN IADN IADN IADN IADN IADN IADN IADN
a Sources are given for historical data. All other data were taken from IADN measurements.
FIGURE 2. Annual average temperature-corrected partial pressures (P288) of gas-phase ∑PCB in femtoatmospheres (10-15 atm) as a function of time for Lakes Michigan (A) and Superior (B). Averages in red indicate IADN data. Averages in green indicate the sum of the 10 most abundant PCB congeners in the IADN data at each site. All other points are historical data (see Table 1). Error bars represent the standard error for each annual average. Correlation coefficients (r) and half-lives (t1/2) (with asymptotic errors) are given for each exponential curve and are significant at the 95% confidence level or greater.
TABLE 2. PCB Half-Lives in Other Environmental Compartments for Lakes Superior and Michigan compartment
lake
t1/2
source
Gull Eggs (Granite Island) Gull Eggs (Agawa Rock) F&Oa lake trout EPAb lake trout Michiganc lake trout Wisconsinc lake trout lake trout F&Oa smelt air water settling particles
Superior Superior Superior Superior Superior Superior Superior Superior Superior Superior Superior
9.9 7.7 23 8.7 9.9 9.9 8.4 8.7 4.1 2.8 2.7
34 34 34 34 34 34 33d 34 34 34 34
fish trout lake trout water
Michigan Michigan Michigan Michigan
7.0 6.3 4.0 4.0
37 36 33d 35
a Sampling was done by the Canadian Fisheries and Oceans. Sampling was done by the U.S. Environmental Protection Agency. c Sampling was done by the states of Michigan and Wisconsin. d Data from ref 33 supplemented by data from Hickey (38). We did the nonlinear curve fitting; see Figure 3. b
lake trout data from Hickey (38) for Lakes Michigan and Superior were added to older data from De Vault et al. (33), declining trends and exponential curves similar to those found for gas-phase ∑PCB were observed (see Figure 3). It is interesting to note that in neither case are the concentrations monotonically decreasing; compare Figures 2 and 3. It remains to be determined why concentrations increase in some years. In this historical context, the apparent increases in ∑PCB shown in the IADN data from 1998 to 1999 are most likely temporary, and in fact, the average ∑PCB concentration 5054
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 23, 2002
in 2000 decreased from that in 1999. In an effort to determine if the increase in ∑PCB concentrations was a product of analytical error related to summing all congeners, Figure 2 also shows the annual average concentrations of the sum of the 10 most abundant congeners in the IADN PCB data at each sampling site. It is
FIGURE 3. ∑PCB concentrations (µg/g wet weight) in Lake Michigan (A) and Lake Superior (B) lake trout. Error bars are for 95% confidence intervals. For Lake Michigan, the curve represents the fit for 19741998. Data through 1992 are from De Vault et al. (33), with the remaining data from Hickey (38). Both fits are significant at the 95% confidence level or better. Half-lives (t1/2) (with asymptotic standard error) and correlation coefficients (r) are given for each curve. clear from these data that the congener concentrations closely track the ∑PCB trends over time. This includes an increase in the 10 congener sums, which tracks the increase in ∑PCB. In addition, an investigation of the annual averages of the individual congeners in this sum indicated that they followed the same trend as the sum of 10 congeners. These relationships indicate that ∑PCB is an accurate indicator of congener trends and that the increase in IADN ∑PCB data during 19971999 is real. HCH Temporal Trends. IADN also measures the concentration of many organochlorine pesticides in the air near the Great Lakes, including R- and γ-HCH. Both HCH isomers were components of technical HCH, a broad spectrum insecticide introduced in 1942. This technical mixture was composed of 55-70% R-HCH and 10-18% γ-HCH, with other HCH isomers rounding out the mixture. By 1978, around the time that the production of PCBs had been banned in the United States, all U.S. registrations of technical HCH had also been canceled. This eliminated the use of R-HCH, but γ-HCH (also called lindane) continued to be used. In fact, it is still widely used, particularly in Canada. IADN R- and γ-HCH gas-phase data along with some historical data for Lake Michigan are presented here as a comparison to the ∑PCB data described above. Since the use of R-HCH was banned in the United States at about the same time as PCBs, we expected to see similarities in the temporal trends of the two compounds. However, since γ-HCH has current sources, we expected to see a different long-term trend than that of ∑PCB and R-HCH. Figure 4 shows plots of temperature-corrected partial pressures (P288) of R- and γ-HCH as a function of time. The data represent both IADN measurements and two or three years of historical data. Unfortunately, only limited historical data for R- and
FIGURE 4. Lake Michigan annual average r-HCH (A) and γ-HCH (B) temperature-corrected partial pressures (P288) in femtoatmospheres (10-15 atm) as a function of time. Data in red are from IADN measurements. All other data are from historical measurements. Black curves represent the fit to the IADN and historical data together. Red curves represent the fit to the IADN data only. Correlation coefficients (r) and half-lives (t1/2) (with asymptotic standard error) corresponding to each curve are given in matching colors. All fits are significant at the 99% confidence level or greater except the γ-HCH fit to all data, which is not significant (NS). Error bars represent the standard error of each annual average. γ-HCH could be found and only for Lake Michigan. Thus, Figure 4 is much more limited in its scope than Figure 2. Exponential curves have been fitted to the IADN data only (red lines) and to the IADN and historical data together (black lines). Curves fitted to the IADN and historical data together indicate a half-life of about 5.7 ( 1.3 yr for R-HCH and no significant declining trend for γ-HCH. This half-life for R-HCH is at least twice as long as the half-life of 2.7 yr found in a previous study of IADN data through 1996 (12). The lack of a statistically significant trend for γ-HCH would seem to be an indication of its current use. The IADN data, however, tell a different story. Curves fitted to only IADN data indicate statistically significant declining trends for both compounds with a half-life of 3.5 ( 0.4 yr for R-HCH and 4.4 ( 0.7 yr for γ-HCH. It is also clear that these trends provide a better fit to the data. The half-life for R-HCH is similar to that found in a previous study, while that for γ-HCH is twice as long as previously thought (12). Looking at Figure 4, it is easy to see why the half-lives of these pesticides have or have not changed with the addition of more data. After 1996, R-HCH concentrations continued to decline at a similar rate, giving us a half-life similar to what Cortes et al. (12) found. Concentrations of γ-HCH, however, have remained relatively steady since 1996. This slowing in decline is most likely due to the current use of this pesticide. Given that γ-HCH was a part of the technical HCH mixture, we would expect concentrations to significantly decrease as bans on this mixture were made worldwide. Current uses of γ-HCH, however, could help to buffer its recent concentrations in the atmosphere, resulting in the VOL. 36, NO. 23, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5055
relatively steady concentrations observed since 1996. Incidentally, no R- or γ-HCH biota data were available for comparison. Clearly, the behavior of HCHs in the atmosphere can be understood using IADN data only, unlike with ∑PCB. This could be because of the measurement technology for these compounds. As discussed earlier, ∑PCB is the sum of measurements made on approximately 100 individual and coeluting chromatographic peaks. On the other hand, R- and γ-HCH are each represented by a single gas chromatographic peak, resulting in much less analytical error than for ∑PCB. Even though R-HCH and PCBs were banned around the same time, there are some differences between the ∑PCB plots in Figure 2 and the R-HCH plots in Figure 4. The IADN R-HCH data have been consistently declining since sampling began. IADN ∑PCB data over Lake Michigan showed little decline and seemed to indicate almost constant gas-phase concentrations when viewed against the historical data. Nevertheless, both R-HCH and ∑PCB temporal trends indicate that the bans on these two substances in the United States have worked. These data show that regulations banning chemicals believed to be harmful really do have a positive impact on the environment, helping to cleanse the atmosphere of such pollutants over time. To see these positive effects, however, long-term monitoring is needed. Programs such as IADN and others like it play a vital role in understanding pollutant trends. However, it is not enough to have only years of data for some compounds; long-term monitoring may need to go on for decades if we are to fully understand how these pollutants degrade in the environment.
Acknowledgments We thank Team IADN, the U.S. Environmental Protection Agency’s Great Lakes National Program Office for funding (GL995656), and Dr. James P. Hickey of the USGS/Great Lakes Science Center for generously providing recent lake trout data.
Literature Cited (1) Eisenreich, S. J.; Looney, B. B.; Thornton, J. D. Environ. Sci. Technol. 1981, 15, 30-38. (2) Eisenreich, S. J.; Hollod, G. J.; Johnson, T. C. In Atmospheric Pollutants in Natural Waters; Eisenreich, S. J., Ed.; Ann Arbor Science: Ann Arbor, MI, 1981; pp 425-444. (3) Murphy, T. J. In Toxic Contaminants in the Great Lakes; Nriagu, J. O., Simmons, M. S., Eds.; John Wiley and Sons: New York, 1984; pp 53-79. (4) Strachan, W. M. J.; Eisenreich, S. J. Mass Balancing of Toxic Chemicals in the Great Lakes: The Role of Atmospheric Deposition; International Joint Commission: Windsor, ON, Canada, 1988. (5) Great Lakes Water Quality Agreement of 1978, as amended 1987. International Joint Commission, United States and Canada, 1994. (6) Hillery, B. R.; Basu, I.; Sweet, C. W.; Hites, R. A. Environ. Sci. Technol. 1997, 31, 1811-1816. (7) Simcik, M. F.; Basu, I.; Sweet, C. W.; Hites, R. A. Environ. Sci. Technol. 1999, 33, 1991-1995.
5056
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 23, 2002
(8) Basu, I. Analysis of PCBs, Pesticides, and PAHs in Air and Precipitation Samples, IADN Project Sample Preparation Procedure; Indiana University: Bloomington, IN, 1999. (9) Basu, I.; Lee, A. D. Quality Control Report 2000; Indiana University: Bloomington, IN, 2001. (10) Buehler, S. S.; Basu, I.; Hites, R. A. Environ. Sci. Technol. 2001, 35, 2417-2422. (11) Halsall, C. J.; Gevao, B.; Howsam, M.; Lee, R. G. M.; Ockenden, W. A.; Jones, K. C. Atmos. Environ. 1999, 33, 541-552. (12) Cortes, D. R.; Basu, I.; Sweet, C. W.; Brice, K. A.; Hoff, R. M.; Hites, R. A. Environ. Sci. Technol. 1998, 32, 1920-1927. (13) Monosmith, C. L.; Hermanson, M. H. Environ. Sci. Technol. 1996, 30, 3464-3472. (14) Cortes, D. R.; Hoff, R. M.; Brice, K. A.; Hites, R. A. Environ. Sci. Technol. 1999, 33, 2145-2150. (15) Eisenreich, S. J. In Sources and Fates of Aquatic Pollutants; Hites, R. A., Eisenreich, S. J., Eds.; American Chemical Society: Washington, DC, 1987; pp 393-469. (16) Manchester-Neesvig, J. B.; Andren, A. W. Environ. Sci. Technol. 1989, 23, 1138-1148. (17) Baker, J. E.; Eisenreich, S. J. Environ. Sci. Technol. 1990, 24, 342-352. (18) Hornbuckle, K. C.; Jeremiason, J. D.; Sweet, C. W.; Eisenreich, S. J. Environ. Sci. Technol. 1994, 28, 1491-1501. (19) Eisenreich, S. J.; Looney, B. B.; Hollod, G. J. In Physical Behavior of PCBs in the Great Lakes; Mackay, D., Paterson, S., Eisenreich, S. J., Simmons, M. S., Eds.; Ann Arbor Science: Ann Arbor, MI, 1983; pp 115-125. (20) Doskey, P. V.; Andren, A. W. J. Great Lakes Res. 1981, 7, 15-20. (21) Cotham, W. E.; Bidleman, T. F. Environ. Sci. Technol. 1995, 29, 2782-2789. (22) Achman, D. R.; Hornbuckle, K. C.; Eisenreich, S. J. Environ. Sci. Technol. 1993, 27, 75-87. (23) Hornbuckle, K. C.; Achman, D. R.; Eisenreich, S. J. Environ. Sci. Technol. 1993, 27, 87-98. (24) Pirrone, N.; Keeler, G. J.; Holsen, T. M. Environ. Sci. Technol. 1995, 29, 2123-2132. (25) McConnell, L. L.; Cotham, W. E.; Bidleman, T. F. Environ. Sci. Technol. 1993, 27, 1304-1311. (26) Chiarenzelli, J.; Scrudato, R.; Arnold, G.; Wunderlich, M.; Rafferty, D. Chemosphere 1996, 33, 899-911. (27) Cullen, A. C.; Vorhees, D. J.; Altshul, L. M. Environ. Sci. Technol. 1996, 30, 1581-1588. (28) Chiarenzelli, J.; Scrudato, R.; Wunderlich, M. Environ. Sci. Technol. 1997, 31, 597-602. (29) Chiarenzelli, J.; Scrudato, R.; Wunderlich, M.; Oenga, G. N.; Lashko, O. P. Chemosphere 1997, 34, 2429-2436. (30) Bushart, S. P.; Bush, B.; Barnard, E. L.; Bott, A. Environ. Toxicol. Chem. 1998, 17, 1927-1933. (31) Nicholls, K. H. Limnol. Oceanogr. 1998, 43, 715-719. (32) Nicholls, K. H.; Hopkins, G. J.; Standke, S. J.; Nakamoto, L. J. Great Lakes Res. 2001, 27, 402-422. (33) De Vault, D. S.; Hesselberg, R.; Rodgers, P. W.; Feist, T. J. J. Great Lakes Res. 1996, 22, 884-895. (34) Smith, D. W. J. Great Lakes Res. 2000, 26, 152-163. (35) Offenberg, J. H.; Baker, J. E. J. Great Lakes Res. 2000, 26, 196208. (36) Miller, M. A.; Madenjian, C. P.; Masnado, R. G. J. Great Lakes Res. 1992, 18, 742-754. (37) Stow, C. A. Environ. Sci. Technol. 1995, 29, 522-527. (38) Hickey, J. P. USGS, personal communication, 2001.
Received for review April 26, 2002. Revised manuscript received September 11, 2002. Accepted September 16, 2002. ES0207091
