Time Trend Analysis of Atmospheric POPs Concentrations in the Great

Oct 8, 2010 - Trends in the levels of halogenated flame retardants in the Great Lakes atmosphere over the period 2005–2013. Liang-Ying Liu , Amina S...
0 downloads 6 Views 276KB Size
Download PDF
Environ. Sci. Technol. 2010, 44, 8050–8055

Time Trend Analysis of Atmospheric POPs Concentrations in the Great Lakes Region Since 1990 MARTA VENIER AND RONALD A. HITES* School of Public and Environmental Affairs, Indiana University, Bloomington, Indiana 47405

Received May 14, 2010. Revised manuscript received August 18, 2010. Accepted September 6, 2010.

Using a multiple linear regression model of the concentrations of several persistent organic pollutants in the atmospheric vapor and particle phases and in precipitation, we have analyzed a data set of about 700,000 values to determine the rate at which these concentrations are decreasing. These concentrations were measured as part of the Integrated Atmospheric Deposition Network (IADN), which has operated several sites near the North American Great Lakes since 1991. The pollutants measured include 83 polychlorinated biphenyl congeners, 17 polycyclic aromatic hydrocarbons, and 24 organochlorine pesticides. In the approach used here, for each of the three phases, the concentrations of a specific chemical at all the sites were combined and fitted with a regression incorporating the sine and cosine of the Julian Day (relative to 1 January 1990 and with a periodicity of one year) and the population living and working within a 25-km radius of the sampling site. Partial residuals were then calculated for each datum, all of the residuals for the three phases were combined, and an overall halving time was calculated from them. This relatively simple approach indicated that the concentrations of PCBs in air around the Great Lakes are decreasing with an overall halving time of 17 ( 2 years, which is slow for a substance that was banned about 35 years ago. Phenanthrene, chrysene, and endosulfan showed halving times on the order of 10 years. The concentrations of several organochlorine pesticides were decreasing more rapidly; for example Rand γ-HCH (lindane) have halving times of about 3.5 years.

Introduction Over the last 40 years, the production and use of many persistent organic pollutants (POPs) have been severely restricted in North America. DDT was banned as part of the Environmental Protection Act in 1970; polychlorinated biphenyls (PCBs) were banned as part of the Toxic Substances Control Act in 1976; chlordane was banned as a termiticide in 1988; and lindane was banned in Canada in 2004 and in the U.S. in 2009. Nevertheless, these so-called “legacy” pollutants are still with us and, among other ecosystems, are contaminating the North American Great Lakes (1). For example, there are still several Great Lakes fish consumption advisories in place for many fish species (2). This paper focuses on measuring progress in reducing the input of POPs to the Great Lakes by way of the atmosphere, now the most important delivery route to the lakes (1). * Corresponding author e-mail: [email protected]. 8050

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 21, 2010

The Integrated Atmospheric Deposition Network (IADN) is a long-term program that has been measuring the atmospheric concentrations of POPs since 1991 at several sites near the shores of the Great Lakes (3). IADN measures organochlorine pesticides, PCBs, and polycyclic aromatic compounds (PAHs) in the atmospheric vapor and particle phases every 12 days and in precipitation every month. In the past, we have analyzed the concentration data for each phase (particulate, vapor, and precipitation) separately, using different approaches (4). For example, vapor phase concentrations were first converted to partial pressures using the ideal gas law, the partial pressures were then converted to those at a standard temperature (288 K) using the Clausius-Clapeyron equation, and time trends were determined from a linear regression of the natural logarithm of these standardized partial pressures as a function of time. Logarithms of particle phase and precipitation concentrations were fitted with a nonlinear equation to determine seasonal changes and changes over time. These regression analyses were carried out for each chemical in each phase at each site, generating many different outcomes, which were frequently difficult to interpret and to generalize (5-10). In this paper, our goal is to clearly answer one fundamental question: At what rate are the atmospheric concentrations of a specific chemical decreasing (if at all) around the Great Lakes region? To meet this goal, we present here a new approach for the analysis of time trends of the atmospheric concentrations of POPs. This approach analyzes the IADN atmospheric concentrations in all three phases (particle, vapor, and precipitation) and at all sampling sites and fits these data with the same model. This allows us to integrate the data from all phases and from all sites, giving us a better overall depiction of the behavior of these chemicals in the Great Lakes region.

Experimental Section Sampling and Analytical Methodology. Samples were collected at the six United States IADN sites over the following time periods: Brule River, Wisconsin (1996-2002), Eagle Harbor, Michigan (1990-2007), Sleeping Bear Dunes, Michigan (1992-2007), Chicago, Illinois (1996-2007), Cleveland, Ohio (2003-2007), and Sturgeon Point, New York (1992-2007) (see http://www.msc-smc.ec.gc.ca/iadn/index_e.html for more details on the sites). Details of the sample collection, extraction, and analysis procedures can be found elsewhere (11, 12), and only a brief description will be presented here. A modified Anderson highvolume air sampler (General Metal Works, model GS2310) was used to collect air samples for 24 h every 12 days at a flow rate giving a total sample volume of about 820 m3. The gas phase organics were collected on Amberlite XAD-2 resin (Supelco, 20-60 mesh) held in a stainless steel cartridge, and the particle phase was collected on 20.3 × 25.4 cm quartz fiber filters. MIC automated wet-only samplers (MIC Co., Thornhill, ON, Canada) were used to collect precipitation samples. Precipitation was collected with a 46 × 46 cm stainless steel funnel connected to a 30-cm long by 1.5-cm i.d. glass column (ACE Glass, Vineland, NJ) packed with XAD-2 resin. The XAD resin, through which the precipitation flowed, collected both particle and dissolved phase organic compounds. Precipitation samples were integrated over each month. The precipitation volume was recorded to calculate concentrations. After sampling, the XAD and the filters were separately Soxhlet extracted for 24 h with 1:1 (v/v) acetone/hexane. Prior to extraction, recovery standards were spiked into the 10.1021/es101656u

 2010 American Chemical Society

Published on Web 10/08/2010

sample. The extract was reduced in volume by rotary evaporation, the solvent was exchanged to hexane, and this solution was fractionated on a column containing 3.5% (w/ w) water deactivated silica gel. The column was eluted with 25 mL of hexane (fraction 1) and 25 mL of 1:1 (v/v) hexane/ dichloromethane (fraction 2). After N2 blow down, the samples were spiked with internal standards. PCBs and pesticides were analyzed by gas chromatography on HewlettPackard 5890 and 6890 instruments equipped with 63Ni electron capture detectors and DB-5 and DB-1701 (J&W Scientific) 60-m columns (250-µm i.d. and 0.1-µm film thickness). PAHs were analyzed by gas chromatographic mass spectrometry (GC/MS) on an Agilent 6890 GC with a 5973 mass spectrometer, using a DB-5 column (J & W Scientific, 30-m long, 250-µm i.d., and 0.25-µm film thickness). Quantitation was done using the internal standard method. Surrogate standards were used to estimate recoveries of each compound in each sample. Quality Control and Quality Assurance. Quality control and quality assurance procedures were followed to ensure data accuracy. The detailed procedures are described in the IADN Quality Assurance Program Plan and in the IADN Quality Control Project Plan (13). In this paper, “PCBs” represents a suite of 83 congeners that includes the toxicologically relevant ones, and “PAHs” represents the sum of 17 compounds. The concentrations of 24 organochlorine pesticides are reported, either as single compounds (R- and γ-HCH) or as groups (“DDT” ) p,p′-DDT + p,p′-DDD + p,p′-DDE; “chlordanes” ) R- + γ-chlordane + transnonachlor; and “endosulfans” ) endosulfan I + II + endosulfan sulfate).

Results and Discussion Assumptions. To simplify the model, we have made the following assumptions: First, we assumed that each compound would show the same seasonal behavior in a given phase at all sites; this implies that, if the concentrations were fitted with a waveform function, they would show similar amplitudes (which represent the ratio between maximum and minimum concentration) and similar dates of maximum concentration. This assumption is supported by previous studies of IADN data, in which analyses were performed on a site by site basis; for example, the concentration of PAHs in the particle phase maximized around the end of January at all sites (14). Second, we assumed that the population living and working within a 25-km radius of the site was constant across the time period covered by our samples, which is approximately 17 years. The values were calculated using the 2002 population probabilistic estimates from the Oak Ridge National Laboratory (ORNL) Landscan according to a method previously described (15). This assumption seemed reasonable; the population distribution of the Great Lakes region has been relatively stable over the last 20 years. For example, Chicago’s population went from 2,783,726 in 1990 to 2,896,010 in 2000, which corresponds to a 4% increase (16). Furthermore, any population changes that did occur would be of the same order of magnitude as the error of the Landscan’s data set. Third, we assumed that the concentration of a given compound in a specific phase in the atmosphere changed at the same rate at all sites, resulting in similar halving times. The assumption is based on the idea that the main pathways for elimination of these compounds from the atmosphere are their reactions with hydroxyl radicals, the concentrations of which are not expected to vary much across the Great Lakes region (17). Our previous analyses of IADN data confirmed that this assumption is acceptable for most of the compounds; for example, for R-HCH in the vapor phase, the halving times were about 3-4 years at all sampling sites (14).

In some instances, such as for total PCBs, some differences in halving times were observed among the sites, but these differences might have been related to the number and quality of the data. Data Analysis. In the first step of this analysis, the concentration data (in pg/m3 or pg/L) for each chemical at all of the sites were merged together into three filessone for each phasesand the natural logarithm of each concentration was calculated. These data were fitted using the following harmonic regression equation (see the Supporting Information for details on the derivation of this expression): ln C ) a0 + a1sin(zt) + a2cos(zt) + a3log2(pop) + a4t + ε (1) where C is the concentration (gas, particulate, or precipitation), t is time expressed in Julian days starting from 1 January 1990, z ) 2π/365.25 (which fixes the periodicity at one year), a0 is an intercept, a1 and a2 are harmonic coefficients that describe the seasonal variation of the concentration with time, a3 (unitless) describes the change of the concentration as a function of population (a surrogate for human activities), a4 is a first-order rate constant (in days-1), and ε is the residual. This multiple linear regression curve fitting was done using Minitab 15, which gave the coefficients, their errors, and the sum-of-squares associated with each term. The dates corresponding to the maximum concentrations were obtained by calculating the first derivative of the above equation and setting it equal to zero a1cos(ztmax) - a2sin(ztmax) + a4 /z ) 0

(2)

This equation can be solved numerically using the Solver feature of Excel or analytically. The error associated with tmax can be calculated from an error propagation using the errors for a1, a2, and a4 given by the regression (eq 1). The Supporting Information gives details on the analytical solution of eq 2 and on the error calculation associated with tmax. The halving time was calculated from the time coefficient, a4 t1/2 )

ln 2 365.25a4

(3)

This regression was applied to all the compounds measured by the IADN project including 83 PCB congeners (plus the total), 17 PAHs (plus the total), and 24 organochlorine pesticides (plus three totals: DDTs, chlordanes, and endosulfans). The coefficients and their standard errors from the regression using eq 1 for all of the compounds measured by IADN are reported in Table S1 in the Supporting Information. In addition, the dates of maximum concentration and halving times are reported in Table S1. In the second step, the effects of seasonality and population were removed from the data using the coefficients calculated from the regression described above to obtain so-called partial residuals ε′ ) ε + a′4t

(4)

where ε is the residual from the multiple regression (eq 1) and ε′ is the partial residual. For each chemical, the partial residuals for each phase separately and for the three phases combined were determined; a linear regression curve was fitted through these data; and the corresponding halving times were calculated from the slopes of these regression lines. This last step was applied to a selection of the most representative and important compounds measured by IADN; these compounds were representative of each category and included total PCBs, five pesticides, and two PAHs. This partial residual analysis is a useful technique that allows one to identify the relationship between a given VOL. 44, NO. 21, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8051

TABLE 1. Halving Times (In Years) of the Main Compounds and Compound Groups in the Three Different Phases and the Weighted Averages of These Times in All Phases Combined Together (The Latter Regressions Are Shown in Figure 1) compound

particle phase

total PCBs -phenanthrene 8.3 ( 0.6 chrysene 9.0 ( 0.8 total endosulfans 11.1 ( 1.2 total DDTs 5.8 ( 0.6 total chlordanes 5.7 ( 0.4 γ-HCH 5.1 ( 0.4 R-HCH 4.0 ( 0.7

vapor phase

precipitation

combined phases

14.9 ( 1.1 18.5 ( 2.2 9.7 ( 0.9 13.1 ( 1.9 8.2 ( 0.4 11.2 ( 0.8 4.0 ( 0.1 3.4 ( 0.1

NS 11.1 ( 1.6 11.5 ( 1.9 14.5 ( 3.8 14.9 ( 4.6 3.9 ( 0.3 2.7 ( 0.2 2.9 ( 0.1

17.0 ( 1.6 11.0 ( 0.6 9.67 ( 0.51 13.5 ( 1.3 8.61 ( 0.44 6.39 ( 0.25 3.48 ( 0.07 3.26 ( 0.05

independent variable (in this case, time expressed as Julian Days) and the response variable (in this case, the natural logarithm of the concentration) given that other independent variables have been included in the model. In other words, in this process, the data are put on the same scale by removing the variability of the local human population, the seasonal factors, and the intercept. Furthermore, the additional step of combining the partial residuals for the three different phases allows us to calculate a weighted average of the halving times using the number of data points in each phase (N in Table S1) as weights. The results of this process are halving times for each phase and for the three phases combined, all of which describe the temporal trends of specific compounds in the atmosphere of the Great Lakes region. Table 1 presents the halving times for each phase and for the three phases combined. Polychlorinated Biphenyls (PCBs). PCB concentrations generally showed the slowest rate of decline among all of the chemicals measured by IADN. The halving time of PCBs in the vapor phase was 14.9 ( 1.1 years, and the halving time in precipitation was not statistically significant. (PCBs were not measured in the particle phase at any site, and their measurements in precipitation were stopped at nonurban sites after 2005 because of their low abundances.) The combined halving time at all IADN sites (Figure 1A) was 17.0 ( 1.6 years, which is statistically the same as in the vapor phase. In a previous study of IADN PCB data, the concentrations were found to be declining in the vapor phase with average halving times of 14 ( 3 years (5), which is the same as that observed here. In the case of the combined data, the correlation coefficient for the regression is equal to 0.22, which is relatively low considering the large number of data points available, but this regression is still significant at P < 0.001. These findings suggest that, overall, PCB concentrations have not changed much over the last 20 years. This is a remarkable finding, considering that the manufacture, import, export, distribution, and use of PCBs was banned in the U.S. over 30 years ago. Apparently, there are still large amounts of PCBs in transformers, capacitors, and other electrical equipment and in storage and disposal facilities that are slowly leaking into the atmosphere. An interesting feature of the data shown in Figure 1A is the slight relative increase in concentrations between 1998 and 2001, peaking around July 1999. This trend was more pronounced at the remote sites of Eagle Harbor and Sleeping Bear Dunes and almost absent at the urban site of Chicago. This atypical feature is also apparent in some temperature corrected PCB partial pressure vs time plots [see for example Figure 2 of reference 5]. When this feature was first noted, explanations related to atmospheric circulation phenomena such as North Atlantic Oscillations (NAO) or El Nin ˜ o events were suggested (18). The idea was that anomalously high temperatures associated with the NAO or El Nin ˜ o events caused increased volatilization of PCBs from reservoirs such 8052

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 21, 2010

as soils or water, where they had accumulated previously. To us, this seems to be an unlikely explanation because similar behavior was not observed for any other compound; on the other hand, we do not have a better explanation to offer. We do note that after 2001, the PCB concentrations resumed their former values. Polycyclic Aromatic Hydrocarbons (PAHs). Among the PAHs, two representative compounds were chosen for detailed analysis: phenanthrene, a three-ring compound, and chrysene, a four-ring compound. These two compounds were chosen based on their relatively high concentrations, compared with other PAHs, and their presence in all three phases. As can be seen from Table 1 and Figure 1B and C, these compounds showed halving times of 8.3 ( 0.6 and 9.0 ( 0.8 years in the particle phase, 18.5 ( 2.2 and 9.7 ( 0.9 years in the vapor phase, 11.1 ( 1.6 and 11.5 ( 1.9 years in precipitation, and 11.0 ( 0.6 and 9.67 ( 0.51 years in the three phases combined, respectively. While the high value for phenanthrene in the vapor phase is an exception, all of the other values are about 10 years. The two most recent studies on the temporal trends of PAH data from IADN data (where each phase and each site were analyzed separately) showed extremely variable halving times for both compounds. For phenanthrene, the halving times ranged between 9 and 13 years in the vapor phase and between 5 and 10 years in the particle phase (7). For chrysene, the halving times ranged between 4 and 7 years in the vapor phase and 4 and 11 years in the particle phase (7). Previously, few significant long-term trends were detected in precipitation for PAHs; only the urban site at Chicago showed significant halving times of about 3 years for both phenanthrene and chrysene (10). Given the scatter of these values, it had been difficult to conclude much of anything about the overall behavior of PAHs in the atmosphere in the Great Lakes region. The results from the study presented here clearly show that PAH concentrations are slowly decreasing and that these concentrations are decreasing more rapidly than are those of PCBs. The major sources of PAHs to the atmosphere are motor vehicle emissions, accounting for 35% of the annual total; aluminum production and forest fires, each accounting for 17% of the total; and residential wood combustion, coke manufacturing, power generation, and incineration, accounting for 12, 11, 6, and 3% of the total, respectively (19). Emissions of PAHs in Europe (including the 27 Member States plus Iceland, Liechtenstein, Norway, Switzerland, and Turkey) have decreased by 63% between 1990 and 2007 as a result of decreased residential use of coal, improvements in abatement technologies for metal refining and smelting, and stricter road transport regulations (20). Although a similar comprehensive report does not exist for the U.S., the decreasing trends observed in this study could be attributed to a general nationwide effort toward cleaner air, as a result of the 1970 Clean Air Act. For example, the Clean Cities Program, which promotes the use of alternative fuels and cleaner vehicles, reported an annual decrease in the use of gasoline of about 15% in cities (21). If PAH emissions followed this trend, the halving time would be 4.3 years. In addition, over the period 1990 to 2002, a reduction of about 50% in the emission of PAHs from motor vehicles was estimated based on a study using the EPA Mobile 6.2 tool (22). If PAH emission followed this trend, the halving time would be 12 years. The EPA Toxic Release Inventory (TRI) reported a significant decrease of “PAH total on-site air emissions” over the past 10 years (23). Based on these estimates, an observed halving time of phenanthrene and chrysene of about 10 years is reasonable. Organochlorine Pesticides. The endosulfans showed the slowest rate of decline among the pesticides, with halving times ranging from 11 to 14 years (Table 1 and Figure 1D).

FIGURE 1. Partial residuals versus sampling date for the three phases combined for (A) PCBs, (B) phenanthrene, (C) chrysene, (D) endosulfans, (E) DDTs, (F) chlordanes, (G) γ-HCH or lindane, and (H) r-HCH. In addition, the overall regression’s correlation coefficient was relatively low, although significant (r ) 0.18, P < 0.001), suggesting that the atmospheric concentrations of these compounds are decreasing slowly around the Great Lakes. This result is consistent with endosulfan’s usage history. In the U.S., endosulfan is currently registered only for agricultural uses, and it is still applied to a large variety of crops, including fruits, vegetables, flowers, cotton, and tobacco for insect control. In 2002, the EPA estimated its average total annual use to be around 600 t of active ingredient (24). Although data on usage trends are scarce, in 2002, the EPA noticed that endosulfan’s agricultural applications (especially in California) had been slowly declining, especially for small grains and soybeans. Hence, the slow decrease we observe here is likely a reflection of endosulfan’s current use, and we do not expect a significant variation in this trend until this insecticide is actually phased out, which the EPA has decided to do sometime in the next few years (25). DDTs showed significant decreasing trends with halving times of 5.8 ( 0.6, 8.2 ( 0.4, 14.9 ( 4.6, and 8.61 ( 0.44 years in the particle, vapor, precipitation, and combined phases, respectively (Table 1 and Figure 1E). It is not clear why the precipitation value is so high, but the other values suggest a halving time of about 7 years. The correlation coefficient of the partial residual regression for the three phases combined was 0.31 (P < 0.001). This relatively high value suggests that the long-term trend is significant, and the passage of time is an important factor in explaining the observed concentrations. These findings are consistent with the most recent studies of the IADN data; Sun et al. reported generally decreasing trends for the various compounds in the DDT family with halving times in the range of 9-16 years in the vapor phase and no significant change as a function

of time in the particle phase (8). DDT was banned in the U.S. in 1970. Its concentrations are now reaching low levels at most sites and are continuing to decline. Chlordane concentrations are also decreasing steadily, with halving times of 5.7 ( 0.4, 11.2 ( 0.8, 3.9 ( 0.3, and 6.39 ( 0.25 years in the particle, vapor, precipitation, and combined phases, respectively (Table 1 and Figure 1F). While the value for the vapor phase is high compared to the other three, the correlation coefficient for the regression of the three phases combined was significant (r ) 0.38, P < 0.001), indicating that the elapse of time played an important role in explaining the observed atmospheric concentrations in all three phases. Two recent studies of the IADN data that included samples collected up until 2003, showed halving times on the order of 10 years for the gas and particle phases and no significant decrease in these concentrations in precipitation (8, 9). The addition of four more years of data, as given here, revealed a somewhat faster rate of decline. This finding is consistent with the banning of chlordane in the United States in 1988, after it had been in production for 40 years. Currently, although there are no active sources of chlordane, there are several reservoirs resulting from past uses, which could continue to contaminate the atmosphere. The most important such reservoir is probably soil around older houses, which had been treated with chlordane to control termites. It is interesting to note that the partial residuals for chlordane in precipitation lie below the regression line (Figure 1F). Since the intercept of these plots is fixed at zero, this observation suggests that chlordane in precipitation decreased at an even faster rate than it did in the other two phases. Table 1 shows that the halving time for precipitation only was 3.9 ( 0.3 years, which is significantly different from VOL. 44, NO. 21, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8053

this time in the other phases. Although we do not have an explanation for this finding, it may suggest that the precipitation measurements for these compounds are less reliable than the vapor or particle measurements. Among the chemicals included in this analysis, R- and γ-HCH showed the fastest rates of decline, with halving times of 3-5 years in all phases (Table 1 and Figure 1G and H). In addition, the correlation coefficients were particularly high in both cases (r ) 0.77 and 0.66, respectively), indicating that the passage of time was the most important factor in explaining the atmospheric levels of these two chemicals. HCH was commercialized as two predominant products. The first was the technical mixture, which contained 60-70% R-HCH, 5-12% β-HCH, and 10-15% γ-HCH. The second product was the purified γ-isomer, which was the only biologically active isomer; this was called lindane. In the U.S., technical HCH was first registered in the 1940s as an insecticide for a variety of applications, but its use was banned in 1978 because of environmental concerns. It was replaced by lindane, the use of which was gradually restricted, and by 2002, it was used only for the treatment of barley, corn, oats, rye, sorghum, and wheat. Lindane was finally deregistered in the U.S. on October 4, 2006, and by October 1, 2009 it was not used in the U.S. any more (26). It is somewhat surprising that, 40 years after it was banned, R-HCH is still present in the environment despite a relatively rapid elimination rate. Clearly, technical-HCH was heavily used in its heyday. The rapid halving times observed for R-HCH and γ-HCH (lindane) are consistent with the relatively high volatilities of these chemicals. The residual plot for γ-HCH shows some periodicity, especially in the precipitation concentrations (Figure 1G). This is surprising given that annual seasonality has been removed by the sine and cosine terms in eq 1. We note that the second cycle, as shown in the partial residuals, maximizes on April 4, which is 88 days earlier than the seasonal cycle determined from the sine and cosine coefficients. This additional seasonal maximum seems to be a characteristic of γ-HCH, and may well be due to its use as a fungicidal seed treatment in Canada during the spring planting season (27). These two periodic causes of variation of the atmospheric concentrations of γ-HCH, one depending on the temperature and one on its agricultural use, were observed previously in our laboratory by Cortes et al. (27). Overall, despite the simplicity of this approach and the assumptions that were made, this analysis provided interesting results that could be useful, particularly for legislators and regulators, and provides interesting insights into the atmospheric behavior of these compounds on a regional level. It is clear that this model showed a few limits (for example, the residual periodicity in the precipitation phase for γ-HCH) and raised a few questions (for example, the faster removal rate of chlordanes in precipitation vs the other phases). This latter result is somewhat unexpected and warrants further analysis.

Acknowledgments We thank Ilora Basu and team IADN at Indiana University for data acquisition and the U.S. Environmental Protection Agency’s Great Lakes National Program Office for funding (Grant GL00E76601-0, Todd Nettesheim, project officer).

Supporting Information Available Details on the derivation of eq 1 and on the analytical solution of eq 2 and a table with the coefficients for all regressions using eq 1. This information is available free of charge via the Internet at http://pubs.acs.org. 8054

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 21, 2010

Literature Cited (1) Carlson, D. L.; Vault, D. S. D.; Swackhamer, D. L. On the rate of decline of persistent organic contaminants in lake trout (Salvelinus namaycush) from the Great Lakes, 1970-2003. Environ. Sci. Technol. 2010, 44 (6), 2004–2010. (2) Michigan Department of Community Health, Division of Environmental Health. 2010 Michigan Fish Advisory. http:// www.michigan.gov/documents/FishAdvisory03_67354_7.pdf. (3) Buehler, S. S.; Hites, R. A. The Great Lakes’ Integrated Atmospheric Deposition Network. Environ. Sci. Technol. 2002, 36 (17), 354A–359A. (4) Venier, M.; Hites, R. A. Regression model of partial pressures of PCBs, PAHs, and organochlorine pesticides in the Great Lakes atmosphere. Environ. Sci. Technol. 2009, 44 (2), 618–623. (5) Sun, P.; Basu, I.; Blanchard, P.; Brice, K. A.; Hites, R. A. Temporal and spatial trends of atmospheric polychlorinated biphenyl concentrations near the Great Lakes. Environ. Sci. Technol. 2007, 41 (4), 1131–1136. (6) Sun, P.; Basu, I.; Hites, R. A. Temporal trends of polychlorinated biphenyls in precipitation and air at Chicago. Environ. Sci. Technol. 2006, 40 (4), 1178–1183. (7) Sun, P.; Blanchard, P.; Brice, K. A.; Hites, R. A. Trends in polycyclic aromatic hydrocarbon concentrations in the Great Lakes atmosphere. Environ. Sci. Technol. 2006, 40 (20), 6221–6227. (8) Sun, P.; Blanchard, P.; Brice, K.; Hites, R. A. Atmospheric organochlorine pesticide concentrations near the Great Lakes: Temporal and spatial trends. Environ. Sci. Technol. 2006, 40 (21), 6587–6593. (9) Sun, P.; Backus, S.; Blanchard, P.; Hites, R. A. Temporal and spatial trends of organochlorine pesticides in Great Lakes precipitation. Environ. Sci. Technol. 2006, 40 (7), 2135–2141. (10) Sun, P.; Backus, S.; Blanchard, P.; Hites, R. A. Annual variation of polycyclic aromatic hydrocarbon concentrations in precipitation collected near the Great Lakes. Environ. Sci. Technol. 2006, 40 (3), 696–701. (11) Basu, I.; Lee, A. D. Collection of Air and Precipitation Samples: IADN Project Sample Procedure; Indiana University: Bloomington, IN, 2000. (12) Carlson, D. L.; Basu, I.; Hites, R. A. Annual variations of pesticide concentrations in Great Lakes precipitation. Environ. Sci. Technol. 2004, 38 (20), 5290–5296. (13) Basu, I.; Arnold, K. Quality control report 2002, 2003; http:// www.indiana.edu/∼hiteslab/Qapjp06.pdf. (14) Sun, P.; Basu, I.; Blanchard, P.; Backus, S. M.; Brice, K. A.; Hulting, M. L.; Hites, R. A. Temporal and spatial trends of atmospheric toxic substances near the Great Lakes: IADN results through 2003; 2007; http://www.epa.gov/glnpo/monitoring/air2/iadn/IADN_ Trend_Report_final.pdf. (15) Venier, M.; Ferrario, J.; Hites, R. A. Polychlorinated dibenzop-dioxins and dibenzofurans in the atmosphere around the Great Lakes. Environ. Sci. Technol. 2009, 43 (4), 1036–1041. (16) U.S. Census Bureau. American Fact Finder; http://factfinder. census.gov/servlet/SAFFPopulation?_submenuId)population_ 0&_sse)on (04/23/2010). (17) Prinn, R. G.; Weiss, R. F.; Miller, B. R.; Huang, J.; Alyea, F. N.; Cunnold, D. M.; Fraser, P. J.; Hartley, D. E.; Simmonds, P. G. Atmospheric trends and lifetime of CH3CCl3 and global OH concentrations. Science 1995, 269 (5221), 187–192. (18) Ma, J.; Hung, H.; Blanchard, P. How do climate fluctuations affect persistent organic pollutant distribution in North America? Evidence from a decade of air monitoring. Environ. Sci. Technol. 2004, 38 (9), 2538–2543. (19) Khalili, N. R.; Scheff, P. A.; Holsen, T. M. PAH source fingerprints for coke ovens, diesel and gasoline-engines, highway tunnels, and wood combustion emissions. Atmos. Environ. 1995, 29 (4), 533–542. (20) APE 006 Specification - EEA32 Persistent organic pollutant (POP) emissions. European Environment Agency, http://themes.eea. europa.eu/IMS/IMS/ISpecs/ISpecification20090814122935/ full_spec#Data. (21) Johnson, C.; Bergeron, P. Clean Cities Annual Metrics Report 2008; Technical Report, NREL/TP-540-46424; September 2009; http://www1.eere.energy.gov/cleancities/pdfs/46424.pdf. (22) Cook, R.; Driver, L.; Mullen, M. Trends in emissions of air toxics from highway mobile sources, 1990 to 2002. In 13th International Emission Inventory Conference “Working for Clean Air in Clearwater”; Clearwater, FL, 2004.

(23) U. S. EPA. Toxics Release Inventory (TRI) Program; April 2010; http://www.epa.gov/tri/. (24) U.S. EPA. Re-registration eligibility decision for endosulfan (EPA 738-R-02-013), 2002; http://www.epa.gov/oppsrrd1/REDs/ endosulfan_red.pdf. (25) U.S. EPA. EPA Action to Terminate Endosulfan; http://www. epa.gov/pesticides/reregistration/endosulfan/endosulfan-cancl-fs. html. (26) The North American Regional Action Plan (NARAP) on Lindane and Other Hexachlorocyclohexane (HCH) Isomers; Commission

for Environmental Cooperation, November 30, 2006. http:// www.cec.org/files/PDF/POLLUTANTS/LindaneNARAP-Nov06_ en.pdf. (27) Cortes, D. R.; Hoff, R. M.; Brice, K. A.; Hites, R. A. Evidence of current pesticide use from temporal and Clausius-Clapeyron plots: A case study from the Integrated Atmospheric Deposition Network. Environ. Sci. Technol. 1999, 33 (13), 2145–2150.

ES101656U

VOL. 44, NO. 21, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8055