Post-1990 Temporal Trends of PCBs and Organochlorine Pesticides

---

Article pubs.acs.org/est

Post-1990 Temporal Trends of PCBs and Organochlorine Pesticides in the Atmosphere and in Fish from Lakes Erie, Michigan, and Superior Amina Salamova,† James J. Pagano,‡ Thomas M. Holsen,§ and Ronald A. Hites†,* †

School of Public and Environmental Affairs, Indiana University, Bloomington, Indiana 47405, United States Department of Chemistry, State University of New York at Oswego, Oswego, New York 13126, United States § Department of Civil and Environmental Engineering, Clarkson University, Potsdam, New York 13699, United States ‡

S Supporting Information *

ABSTRACT: We have analyzed concentration data sets covering the period 1992−2010 from the Great Lakes Integrated Atmospheric Deposition Network and from the Great Lakes Fish Monitoring and Surveillance Program to determine and compare pollutant time trends in the atmosphere and in fish. The analytes of interest were polychlorinated biphenyls, DDTs, chlordanes, dieldrin, and α- and γ-hexachlorocyclohexane (HCHs), and the sites of interest were Lakes Erie, Michigan, and Superior. Overall, we found no significant differences between the atmospheric and fish temporal trends for any of these compounds in any of the lakes. Polychlorinated biphenyl concentrations are decreasing in both the atmosphere and in the fish with halving times of 14 ± 2 years. The halving times for DDTs, chlordanes, and dieldrin are 8.7 ± 0.4 years for both the atmosphere and the fish. The most rapid temporal trend was observed for α- and γ-HCH concentrations, which are decreasing in both the atmosphere and in fish with halving times of 3.3 ± 0.4 years. The practical implications of these results are discussed.

■

INTRODUCTION The atmospheric deposition of pollutants to lakes and their subsequent bioaccumulation in fish was first noted in the Great Lakes. Specifically, Swain et al.1 reported that trout from Siskiwit Lake on Isle Royale (the largest lake on the largest island in the largest of the Great Lakes) were contaminated with several organochlorine pollutants. This came as a surprise given the remoteness of this lake. Swain pointed out that “Siskiwit Lake [is] a deep, cold lake on Isle Royale, well removed from the direct influences of man.” He observed that concentrations of “several organic residues in the flesh of fish from Siskiwit Lake were significantly higher than corresponding fish from Lake Superior. Polychlorinated biphenyl [concentrations] are double the Lake Superior mean value (34 vs. 15 ppm lipid), and p,p′-DDE showed a more than 10-fold increase (68 vs. 4 ppm lipid) in Siskiwit Lake.” In fact, the atmosphere is now an important transport pathway for the input and removal of toxic and persistent organic pollutants (POPs) into and out of the Great Lakes. Although atmospheric concentrations can respond rapidly to changes in emissions, it is not known how rapidly the lakes and the fish in them respond because of what we might call “environmental hysteresis.” The Integrated Atmospheric Deposition Network (IADN) was begun in 1990 to address this linkage. The goals of IADN are to “determine the atmospheric loadings and trends (both spatial and temporal) of priority toxic chemicals to the Great Lakes and its basin on, at least, a biennial basis [and to] acquire quality-assured air and precipitation concentration measurements, with attention to © 2013 American Chemical Society

continuity and consistency of those measurements, so that trend data are not biased by changes in network operations or personnel.”2 The pollutants on which the IADN focuses include (but are not limited to) polychlorinated biphenyls (PCBs), several chlorinated pesticides, many halogenated flame retardants, and about 20 polycyclic aromatic hydrocarbons. IADN is a joint program of the United States and Canada, and it has been measuring the concentrations of these pollutants in atmospheric samples collected on the shores of the five Great Lakes every 12 days since 1990. In parallel, the Great Lakes Fish Monitoring and Surveillance Program (GLFMSP) has been measuring a similar suite of pollutants in Great Lakes top predator fish species (walleye in Lake Erie and lake trout in the other four lakes) since 1970. This program samples fish annually at one of two sites on each of the Great Lakes; the two sites rotate between odd and even years. Both of these programs are managed by the United States Environmental Protection Agency’s Great Lakes National Program Office.3 The IADN and the GLFMSP have provided long-term data sets of organic pollutants, and data from each program have been widely reported along with temporal trend analyses for these contaminants. Unfortunately, the IADN and GLFMSP trend analyses have used different data set time periods and Received: Revised: Accepted: Published: 9109

April 30, 2013 July 16, 2013 July 24, 2013 August 1, 2013 dx.doi.org/10.1021/es401895g | Environ. Sci. Technol. 2013, 47, 9109−9114

Environmental Science & Technology

Article

Details on the sample sites, the sample collection methods, and the analysis procedures of the atmospheric samples can be found elsewhere,2,21,22 and only a brief description is presented here. A modified Anderson high-volume air sampler (General Metal Works, model GS2310) was used to collect air samples for 24 h every 12 days with a total sample volume of ∼820 m3. The vapor phase was collected on Amberlite XAD-2 resin (Supelco, Bellefonte, PA; 20−60 mesh) held in a stainless steel cartridge. The compounds reported in this study are found almost exclusively in these vapor phase samples. The XAD cartridges were shipped to the Indiana University laboratory and stored at less than −20 °C until extraction. The returned XAD resin was Soxhlet extracted for 24 h with 1:1 (v/v) acetone in hexane. Prior to extraction, recovery standards were spiked into the sample. The extract was reduced in volume by rotary evaporation, exchanged into hexane, and fractionated on a column containing 3.5% w/w water deactivated silica gel. The column was eluted with 25 mL of hexane (called fraction 1) and 25 mL of 1:1 (v/v) dichloromethane in hexane (called fraction 2). After N2 blow down, the sample was spiked with the internal standards. PCBs (eluting in fraction 1) and pesticides (eluting in fraction 2) were analyzed by gas chromatography on Hewlett-Packard 5890 and 6890 gas chromatographs equipped with 63Ni electron capture detectors and with DB-5 and DB-1701 columns (J&W Scientific, 30 m × 250 μm i.d. with a 0.25 μm liquid phase thickness). Quantitation was done using the internal standard method. The treatment and analysis of the atmospheric samples have been consistent throughout IADN. Fish Samples. Fish samples were collected at the following GLFMSP sampling sites: Middle Bass Islands (even years) and Dunkirk (odd years) on Lake Erie; Saugatuck (even years) and Sturgeon Bay (odd years) on Lake Michigan; Apostle Islands (even years) and Keweenaw Point (odd years) on Lake Superior. Each site was alternatively sampled in odd and even years during 1992−2010. The details on sampling procedures are given elsewhere.23,24 Briefly, lake trout (600−700 mm) from Lakes Michigan and Superior and walleye (400−500 mm) from Lake Erie were collected following the GLFMSP sampling protocols.23 Each year, 50 fish from each site were composited by size into 10 groups of 5 fish each. Each group of 5 fish was ground together and homogenized. The resulting 10 composite samples were frozen and shipped to the laboratory and stored at less than −20 °C until extraction. Extraction and cleanup methods for the fish samples were different for the periods of 1990−1997, 1998−2003, and 2004−2010 and have been described previously.16−20,25 During the two earlier periods, the ground and homogenized whole fish samples were mixed with anhydrous Na2SO4 and Soxhlet extracted. Lipids were removed by acid catalyzed precipitation followed by either centrifugation; by gel permeation chromatography; or by Florisil, alumina, or silica liquid−solid chromatography. PCBs and organochlorine pesticides were quantitated using gas chromatography with an electron capture detector and gas chromatographic mass spectrometry in the electron capture negative ionization mode, respectively. The samples collected during 2004−2010 were extracted using accelerated solvent extraction (ASE, Dionex) with dichloromethane. Surrogate standards were added to the samples before extraction to assess extraction efficiency. The lipids were removed by gel permeation chromatography (GPC, Waters), and the samples were fractionated on 4% water deactivated

different statistical methods. For example, atmospheric concentration data have been analyzed separately for each chemical at each site by converting the vapor phase concentrations to partial pressures and normalizing these partial pressures to a standard temperature (288 K). Temporal trends of these normalized partial pressures were determined by a linear regression of their natural logarithms as a function of sampling date.4−9 More recently, temporal trends have been determined using multiple harmonic regression techniques to identify long-term trends and to precisely determine the annual periodicity of these atmospheric concentrations.10−14 Fish concentration data have, most recently, been analyzed using one- and two-segment exponential models.15−20 The twosegment model has demonstrated that the temporal trends since about 1990 are considerably slower than they were prior to this time. There have been no comprehensive time trend comparisons of Great Lakes atmospheric and fish concentrations over the same time periods and using the same statistical methods. The specific goal of this study is to compare the temporal trends in the concentrations of PCBs, DDTs, chlordanes, dieldrin, and α- and γ-hexachlorocyclohexanes (HCHs) in the atmosphere and in fish from Lakes Erie, Michigan, and Superior using IADN and GLFMSP concentration data sets. The period used in this study was 1992−2010, the years when these two data sets overlapped, when reliable data from IADN began to be obtained, and when the rapid decreases in concentrations of these compounds in fish had slowed. This is the first comprehensive study comparing the temporal trends in the concentrations of these chemicals in the atmosphere and in fish in the Great Lakes region using long-term and overlapping time periods for the two data sets. In addition, similar data analysis methods (linear regressions) were applied to both the atmospheric and fish data sets, allowing a direct comparison of the temporal trends. Only IADN and GLFMSP data were used because of the strict quality control and quality assurance procedures followed by these two environmental measurement programs and because both data sets were available to us. The goal of this work is not mechanistic. The air−water−fish pollutant transfer process is complex and is dependent on many exogenous factors such as air and water temperature, a pollutant’s concentrations in both the dissolved and particle aquatic phases, trophic levels of the fish, and more. In addition, the air−water transfer process is a two-way street. Pollutants can enter the water from the air some of the time, and they can enter the air from the water at other times. In this work, we are asking a simple question: Do the rates of decrease measured in the air near Lakes Erie, Michigan, and Superior track these rates in higher tropic level fish from these lakes? This question is of interest given that the production, sale, and use of these compounds in the United States and Canada have been restricted for 10−40 years. Because of environmental hysteresis, it is not obvious if the pollutant concentrations in air and fish would track each other, if one would lag behind the other, or if, for example, metabolism in the fish or reactions in the atmosphere would cause these rates to be significantly different.

■

EXPERIMENTAL SECTION Atmospheric Samples. Atmospheric samples were collected over the period 1992−2010, inclusive, at three of the U.S. IADN sites. These sites were: Sturgeon Point, New York, at Lake Erie; Sleeping Bear Dunes, Michigan, at Lake Michigan; and Eagle Harbor, Michigan, at Lake Superior. 9110

dx.doi.org/10.1021/es401895g | Environ. Sci. Technol. 2013, 47, 9109−9114

Environmental Science & Technology

Article

silica eluted with hexane (called fraction 1) followed by 1:1 (v/ v) hexane in dichloromethane (called fraction 2). PCBs eluted in fraction 1, and pesticides eluted in fraction 2. These compounds were analyzed using gas chromatography with an electron capture detector (Agilent 7890A) on a DB-XLB capillary column (60 m × 250 μm i.d. with a 0.25 μm liquid phase thickness). Quantitation was performed using the internal standard method. In this work, total PCB (denoted as ΣPCB) concentrations are the sum of 80−100 congener concentrations commonly measured in both the atmospheric and fish samples. DDT concentrations (ΣDDT) are the sum of p,p′-DDT, p,p′-DDD, and p,p′-DDE concentrations; and chlordane concentrations (Σchlordane) are the sum of α- and γ-chlordane and transnonachlor concentrations. The concentrations of dieldrin and α- and γ-hexachlorocyclohexane (HCH) are reported individually. Quality Control and Quality Assurance. Quality control and quality assurance (QA/QC) procedures were followed to ensure data accuracy in both the IADN and GLFMSP programs. The detailed QA/QC procedures for the air samples are described in the IADN Quality Assurance Project Plan.26 The GLFMSP followed an extensive QA/QC program using laboratory blanks and National Institute of Standards and Technology lake trout Standard Reference Material, as well as several interlaboratory studies. The results of these studies can be found elsewhere.18,19

Figure 1 shows an example of this data analysis procedure for the atmospheric and fish ΣPCB concentrations for Lake

Figure 1. Regressions of atmospheric ΣPCB concentrations (in pg/ m3) vs. time at Sleeping Bear Dunes and of lake trout ΣPCB concentrations (in ng/g wet weight) vs. time for Lake Michigan; see the SI for the regression details.

■

RESULTS AND DISCUSSION Over the period of interest, IADN has accumulated several thousand measurements of PCB and chlorinated pesticide concentrations in the atmosphere. These measurements have been made at 12 day intervals, and thus, there are about 30 measurements of each compound per year. However, the fish measurements are made on an annual basis (with replication). Thus to make the comparison of these two data sets tractable and to give a balanced statistical design, we first calculated the geometric means of the atmospheric concentrations of each compound for each year at each site. The geometric mean was used here because it is known that these concentrations are lognormally distributed. In both cases, we fitted the data using a simple first-order rate model, ln(conc) = a0 + a1t

Michigan (see SI for all other regression data). Both of these regressions are highly significant at p < 0.001. From these regressions, we note that the concentrations of ΣPCB have been decreasing exponentially in both the atmosphere and in the fish from Lake Michigan with halving times of 12.2 ± 3.0 and 9.8 ± 1.1 years, respectively. The Student’s t-test was used for the comparison of the halving times in the atmosphere vs those in the fish. The tvalues were calculated using t=

−ln(2) a1

1/2

2 2 (stderrair + stderrfish )

(3)

where rate is the halving time in either the air or fish, and stderr is the standard error of the rate in either the air or fish. The number of degrees of freedom is the sum of the total number of points used in the two regressions minus 2. Using the calculated t-values and the number of degrees of freedom, the probability of the two rates not being different can be obtained using the 2-tail t-distribution function. For example for the total PCB data shown in Figure 1, the regression was based on 19 data points, so the degree of freedom is 36. The difference of the halving-times is 2.4, and the standard error of this difference is 3.2. Thus, the t-value is 0.740, which corresponds to a probability of 46%. In other words, the probability that the two rates are not different is high (about 50%). Table 1 shows all halving times with their standard errors, the associated t-values, and their probabilities. In Lake Erie, these probabilities are in the range of 35−55% (averaging 44%). These probabilities indicate that the halving times for a given compound are not significantly different between air and fish and that the concentrations of these compounds in air and fish sampled from Lake Erie are decreasing at the about same rates. In fact, none of these probabilities are lower than the normal 5% cutoff for

(1)

where conc is the contaminant’s concentration in the air or fish, t is the year of sampling, ranging between 1992 and 2010, a0 is the intercept that rectifies the units, and a1 is a first-order rate constant with units of years−1. These regressions were done using Sigma Plot 11.0, which gave the regression coefficients (a0 and a1), their standard errors, and the sum-of-squares associated with each term and with the regression. Any measurements at or below the analytical detection limits were empty cells. This regression was applied to ΣPCB, ΣDDT, Σchlordanes, dieldrin, and α- and γ-HCH. All of the input data and the regression parameter results are given in the Supporting Information, SI. From a1, we calculated the time for the concentrations to decrease by one-half, which we call a halving time, using, t1/2 =

rateair − rate fish

(2) 9111

dx.doi.org/10.1021/es401895g | Environ. Sci. Technol. 2013, 47, 9109−9114

Environmental Science & Technology

Article

relatively high probabilities indicate that the halving times for a given compound are not significantly different and that the concentrations of these compounds in air and fish sampled near and from Lake Superior are decreasing at the about same rates. The two exceptions are ΣPCBs and γ-HCH (lindane), for which the probabilities are about 90%. This strongly indicates that the halving times in the air and in the fish in Lake Michigan are very close to one another for these compounds. Among all the chemicals in this study, α- and γ-HCH showed the fastest rates of decline in both the atmosphere and fish for all three lakes. Technical HCH consisted of 60−70% α-HCH, 2−12% β-HCH, and 10−15% γ-HCH. Technical HCH was used as an insecticide, but its use in North America was phased out in 1970s. At that time, it was replaced by a product consisting of the γ-isomer (lindane), which had wide agricultural, pharmaceutical, residential, and commercial uses in the United States until it too was phased out in 2009. When significant, the halving times for the atmospheric and fish αand γ-HCH concentrations were 3−4 years over the period of 1999−2010 (Table 1). Regressions for α-HCH concentrations in fish from Lake Michigan and the α- and γ-HCH concentrations in fish from Lake Erie were not significant. These results are probably caused by the relatively low detection frequency of one or both of these compounds in Lake Michigan and Erie fish. The partial r2 values for the regressions of atmospheric α- and γ-HCH concentrations as a function of time were particularly high, ranging from 0.67 to 0.97. This suggests that time is the most significant factor influencing the decline of these concentrations. As discussed above, the air−water−fish pollutant transfer process is complex. But the data presented here indicate that the rates of decrease measured in the air near Lakes Erie, Michigan, and Superior are not significantly different than these rates in higher trophic level fish. Despite the complexity of the mechanisms involved in the air−water−fish transport process, it is clear that pollutant concentrations in the two end members of this system are changing at the same ratesat least for the compounds and for the three Great Lakes discussed here. All of the rates are negative, meaning that environmental regulations restricting the use of these compounds in the United States and Canada have been effectiveespecially for the hexachlorocyclohexanes. What are the practical implications of these results? The most obvious suggestion is that both air and fish measurement programs are not needed to keep track of temporal trends of these legacy compounds. While the results presented here indicate that these two measurement programs are complementary, we believe that both programs provide different types of information and should both be continued. Although total PCB concentrations are changing at the same rate in air and fish, these two media have different PCB congener profiles, suggesting that there is an intermediate step in accumulation of PCBs in Great Lakes fish. This step is most likely related to the different lipophilicities of the different congeners. For example, in Lake Michigan in 2011, the average chlorination level in air was 3.5 chlorines/biphenyl, but in fish it was 5.1 chlorines/biphenyl, indicating the preferential accumulation of the more chlorinated and more lipophilic congeners. Overall PCB concentrations may now be changing at the same rate in air and fish but for different reasons, and air and fish trends may diverge in the future. (Congener-specific PCB measurements in fish did not start until 2000, so congener specific time trend trends could not be examined.) Since

Table 1. Halving Times (in Years) with Their Standard Errors for the Atmospheric and Fish Concentrations of ΣPCB, ΣDDT, Σchlordane, Dieldrin, and α- and γ-HCHs at Lakes Michigan, Erie, and Superiora

compound

atmospheric halving times at Sleeping Bear Dunesb

ΣPCB ΣDDT Σchlordane dieldrin α-HCHd γ-HCHd

22.5 8.9 8.1 7.4 2.9 2.5

± ± ± ± ± ±

9.1 0.8 0.7 1.0 0.3 0.1

ΣPCB ΣDDT Σchlordane dieldrin α-HCHd γ-HCHd

12.2 8.4 8.6 8.0 3.4 2.6

± ± ± ± ± ±

3.0 1.1 1.2 1.3 0.4 0.2

ΣPCB ΣDDT Σchlordane dieldrin α-HCHd γ-HCHd

13.0 10.4 10.0 8.3 3.9 3.1

± ± ± ± ± ±

2.8 2.0 1.6 1.2 0.3 0.2

fish tissue halving times throughout the lakeb

Lake Erie 15.2 ± 7.0 13.4 ± 4.8 9.7 ± 2.2 9.8 ± 2.5 NS NS Lake Michigan 9.8 ± 1.1 6.2 ± 0.5 10.1 ± 2.4 7.4 ± 1.3 NS 3.0 ± 0.8 Lake Superior 13.7 ± 6.6 7.7 ± 2.0 7.7 ± 2.0 7.0 ± 1.1 4.9 ± 0.8 3.2 ± 0.9

student’s t-value

probability (%)

0.633 0.936 0.706 0.908

53.1 35.6 48.5 37.0

0.740 1.820 0.558 0.330

46.4 7.7 58.1 74.3

0.403

69.0

0.102 0.957 0.895 0.742 1.109 0.153

91.9 34.5 37.7 46.3 27.9 88.0

The t-values associated with the differences between the air and fish halving times and their errors are also given. The probabilities are the odds that the halving times in air vs. those in fish are not different. bAll errors are given as standard errors. cNS: not significant at P < 0.05. d The regression was applied to 1999−2010 atm and fish concentration data. a

significance. It is interesting to note that the halving times for ΣDDTs, Σchlordanes, and dieldrin in air are less (but not significantly so) than those in fish. It is not yet clear why this might be so. In Lake Michigan, these probabilities are all in the range of 45−75% (averaging 62%) with one low exception. These high probabilities indicate that the halving times for a given compound are not significantly different between air and fish. The overall conclusion is that the concentrations of these compounds in air sampled near Lake Michigan and in fish sampled from this lake are decreasing at the about same rates. The exception is ΣDDT (P = 7.7%). In this case, the fish concentrations are decreasing more rapidly (but not significantly so, assuming a significance cut off of 5%) than the air concentrations. It is not obvious why this might be the case, but a few words on DDT are warranted. DDT was banned in the U.S. over 40 years ago, and its concentrations in the environment decreased rapidly in the 1970s and 1980s. Our data show a continuing decline in ΣDDT concentrations, and the rate of this decline is statistically significant for both the atmospheric and the fish samples from all three lakes. The ΣDDT halving times in the atmosphere and in the fish were 8.4 ± 1.1 and 6.2 ± 0.5 years, respectively. This observation suggests that ΣDDT concentrations are decreasing at similar rates in those two matrices. In Lake Superior, these probabilities are all in the range of 30−45% (averaging 37%) with two high exceptions. These 9112

dx.doi.org/10.1021/es401895g | Environ. Sci. Technol. 2013, 47, 9109−9114

Environmental Science & Technology

Article

Figure 2. Concentrations of ΣPCBs at Chicago (in red) and at Sleeping Bear Dunes (in green). The data are given as the natural logarithms of the concentrations in pg/m3. The curves were fitted using ln(C) = a0 + a1 sin(zt) + a2 cos(zt) + a3 t, where z = 2π/365.25 and t = time in days. The regression results are all significant at P < 0.001. In both cases, the concentrations maximize around July 1 of each year.

of many pollutants at Middle Base Islands (in the western part of the lake) were higher than those at Dunkirk (in the eastern part of the lake), except for oxychlordane and DDTs.19 This latter difference is likely caused by the urban centers along the Detroit River impacting the western end of Lake Erie. Clearly, knowing where a pollutant’s concentrations are highest is useful when it comes to suggesting remediation activities. Because lipophilic pollutants are bioconcentrated into fish over the lifetime of the fish, these samples are an excellent resource for the identification of nontargeted contaminants. These concentrations would be higher in fish tissue, and almost by definition, they would be persistent organic pollutants. For example, recent analysis of these Great Lakes fish samples has identified tetraphenyl tina contaminant that was not suspected to be present in the Great Lakes ecosystem.27 It is also relatively simple to freeze and store samples of the fish tissue composites to provide an archive that one could use after the identification of a new and unsuspected pollutant to assess retrospectively its historical inputs into the environment. For example, it was this frozen tissue resource that was first used to quantitate, as a function of time, the levels of polybrominated flame retardants in Great Lakes fish.15 It is more difficult to reliably archive extracts of air samples.

different congeners have different toxicological properties, changes in ecosystem impacts due to the changes in concentrations we are seeing are unknown. The air and fish measurement programs have very different time constants. The fish program has a time constant of years, and the air program has a time constant of days. In fact, the fish have been in the water integrating pollutants over their entire lifetimes (which for lake trout and walleye is about 10−20 years), but the air samples represent a snapshot of the concentrations over a one day period every 12 days. This idea is illustrated in Figure 2, which shows the same atmospheric data set as Figure 1, except the individual air measurements are shown instead of the annual averages. Clearly, there are relatively high-frequency variations (fine structure) in these data that cannot be detected with once-a-year fish samples. For example, a strong seasonal effect and an anomalous elevation in these concentrations in 1998−2000 are obvious in these air data. Incidentally, subtle annual oscillations in fish concentrations were also observed and may be attributed to trophodynamic, meteorological, or climatological changes affecting the lake environment, and in fish, an anomalous minimum concentration was found in 2007−2008 in all three lakes.19 The air sampling is associated with a specific location, but the fish sampling is, by intention, designed to sample the open lake populations. Air is collected at one place during the entire time period, and that location is known exactly. However, fish are collected over several square kilometers of the lake, and the fish themselves have traveled for unknown distances before being collected (or as our mothers used to say, “Don’t eat that fish you don’t know where it’s been!”). The pragmatic result of this difference is that it is easier to detect the influence of population centers on air concentrations,10−14 than on fish concentrations. For example, Figure 2 includes air data for total PCBs at both Sleeping Bear Dunes (a remote site) and at Chicago; note the factor of 10 offset in these concentrations. The impact of urban centers near Lake Michigan was not seen in lake trout concentrations; the levels of all pollutants in the Saugatuck (in the southern part of the lake) and Sturgeon Bay (in the northern part of the lake) fish subsamples were not statistically significant. However, concentrations at sampling sites in Lake Erie did show the effect of urban centers; the levels

■

ASSOCIATED CONTENT

S Supporting Information *

All of the annual atmospheric and fish concentration data (given as their natural logarithms) for all of the compounds in Lakes Erie, Michigan, and Superior are given along with the regression parameters and their standard errors and the t-values and their probabilities. This material is free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest. 9113

dx.doi.org/10.1021/es401895g | Environ. Sci. Technol. 2013, 47, 9109−9114

Environmental Science & Technology

■

Article

the corner in Great Lakes trout 1980−2009. Environ. Sci. Technol. 2012, 46, 9890−9897. (19) Chang, F.; Pagano, J. J.; Crimmins, B. S.; Milligan, M. S.; Xia, X.; Hopke, P. K.; Holsen, T. M. Temporal trends of polychlorinated biphenyls and organochlorine pesticides in Great Lakes fish, 1999− 2009. Sci. Total Environ. 2012, 439, 284−290. (20) Carlson, D. L.; De Vault, D. S.; 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, 2004−2010. (21) Basu, I.; Bays, J. C. Collection of air and precipitation samples. IADN Project Standard Operating Procedure; Indiana University: Bloomington, IN, 2010. (22) Basu, I.; Arnold, K. A. Analysis of PCBs, pesticides, PAHs, and PBDEs in air and precipitation samples. IADN Project Sample Preparation Procedure; Indiana University: Bloomington, IN, 2010. (23) Murphy, E. Quality Assurance Project Plan for Sample Collection Activities; Great Lakes Fish Monitoring Program. U.S. Environmental Protection Agency Great Lakes National Program Office: Chicago, IL, 2004. (24) Xia, X. Y.; Crimmins, B. S.; Hopke, P. K.; Pagano, J. J.; Milligan, M. S.; Holsen, T. M. Toxaphene analysis in Great Lakes fish: A comparison of GC-EI/MS/MS and GC-ECNI-MS, individual congener standard and technical mixture for quantification of toxaphene. Anal. Bioanal. Chem. 2009, 395, 457−463. (25) De Vault, D. S.; Hesselberg, R.; Rodgers, P. W.; Feist, T. J. Contaminant trends in lake trout and walleye from the Laurentian Greta Lakes. J. Great Lakes Res. 1996, 22, 884−895. (26) U.S. Environmental Protection Agency Quality Assurance Project Plan; Integrated Atmospheric Deposition Network. 2011 (27) Milligan, M. S.; Richards, D.; Crimmins, B.; Xia, X.; Holsen, T.; Hopke, P.; Pagano, J. Non-targeted and targeted identification of emerging contaminants in Great Lakes fish and fish eggs using GC × GC TOF mass spectrometry. Presented at the Annual Meeting of the International Association for Great Lakes Research, Purdue University; West Lafayette, IN, June 2−6, 2013.

ACKNOWLEDGMENTS We thank the IADN and GLFMSP teams for site operations and for sample analysis and the U.S. Environmental Protection Agency Great Lakes National Program Office of the U.S. Environmental Protection Agency for funding (IADN Grant No. GL00E76601-0, Todd Nettesheim, project manager; GLFMSP Grant No. GL96594201-1, Elizabeth Murphy, project manager).

■

REFERENCES

(1) Swain, W. R. Chlorinated organic residues in fish, water, and precipitation from the vicinity of Isle Royale, Lake Superior. J. Great Lakes Res. 1978, 4, 398−407. (2) Environment Canada. Integrated Atmospheric Deposition Network (IADN) https://www.ec.gc.ca/rs-mn/default.asp?lang= En&n=BFE9D3A3-1. Accessed on July 12, 2013. (3) United States Environmental Protection Agency. About the Great Lakes National Program Office (GLNPO). http://www.epa.gov/ aboutepa/glnpo.html. Accessed on November 29, 2012. (4) Sun, P.; Backus, S.; Blanchard, P.; Hites, R. A. Annual variations of polycyclic aromatic hydrocarbon concentrations in precipitation collected near the Great Lakes. Environ. Sci. Technol. 2006, 40, 696− 701. (5) Sun, P.; Basu, I.; Hites, R. A. Temporal trends of polychlorinated biphenyls in precipitation and air at Chicago. Environ. Sci. Technol. 2006, 40, 1178−1183. (6) 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, 2135−2141. (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, 6221−6227. (8) Sun, P.; Blanchard, P.; Brice, K. A.; Hites, R. A. Atmospheric organochlorine pesticide concentrations near the Great Lakes: Temporal and spatial trends. Environ. Sci. Technol. 2006, 40, 6587− 6593. (9) 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, 1131−1136. (10) Venier, M.; Hites, R. A. Regression model of partial pressures of PCBs, PAHs, and organochlorine pesticides in the Great Lakes’ atmosphere. Environ. Sci. Technol. 2010, 44, 618−623. (11) Venier, M.; Hung, H.; Tych, W.; Hites, R. A. Temporal trends of persistent organic pollutants: A comparison of different time series models. Environ. Sci. Technol. 2012, 46, 3928−3934. (12) Venier, M.; Hites, R. A. Time trend analysis of atmospheric POPs concentrations in the Great Lakes region since 1990. Environ. Sci. Technol. 2010, 44, 8050−8055. (13) Salamova, A.; Hites, R. A. Discontinued and alternative brominated flame retardants in the atmosphere and precipitation from the Great Lakes basin. Environ. Sci. Technol. 2011, 45, 8698− 8706. (14) Salamova, A.; Hites, R. A. Dechlorane Plus in the atmosphere and precipitation near the Great Lakes. Environ. Sci. Technol. 2011, 45, 9924−9930. (15) Zhu, L. Y.; Hites, R. A. Temporal trends and spatial distributions of brominated flame retardants in archived fishes from the Great Lakes. Environ. Sci. Technol. 2004, 38, 2779−2784. (16) Hickey, J. P.; Batterman, S. A.; Chernyak, S. M. Trends of chlorinated organic contaminants in Great Lakes trout and walleye from 1970 to 1998. Arch. Environ. Contam. Toxicol. 2006, 50, 97−110. (17) Batterman, S.; Chernyak, S.; Gwynn, E.; Cantonwine, D.; Jia, C.; Begnoche, L.; Hickey, J. P. Trends of brominated diphenyl ethers in fresh and archived Great Lakes fish (1979 − 2005). Chemosphere 2007, 69, 444−457. (18) Crimmins, B. S.; Pagano, J. J.; Xia, X.; Hopke, P. K.; Milligan, M. S.; Holsen, T. M. Polybrominated diphenyl ethers (PBDEs): Turning 9114

dx.doi.org/10.1021/es401895g | Environ. Sci. Technol. 2013, 47, 9109−9114