Article pubs.acs.org/est
Cite This: Environ. Sci. Technol. XXXX, XXX, XXX-XXX
Updated Polychlorinated Biphenyl Mass Budget for Lake Michigan Jiehong Guo,† Kevin Romanak,† Stephen Westenbroek,‡ An Li,§ Russell G. Kreis, Jr.,∥ Ronald A. Hites,† and Marta Venier*,† †
School of Public and Environmental Affairs, Indiana University, Bloomington, Indiana 47405, United States U.S. Geological Survey, Wisconsin Water Science Center, Middleton, Wisconsin 53562, United States § School of Public Health, University of Illinois at Chicago, Chicago, Illinois 60612, United States ∥ United States Environmental Protection Agency, Office of Research and Development, Grosse Ile, Michigan 48138, United States ‡
S Supporting Information *
ABSTRACT: This study revisits and updates the Lake Michigan Mass Balance Project (LMMBP) for polychlorinated biphenyls (PCBs) that was conducted in 1994−1995. This work uses recent concentrations of PCBs in tributary and open lake water, air, and sediment to calculate an updated mass budget. Five of the 11 LMMBP tributaries were revisited in 2015. In these five tributaries, the geometric mean concentrations of ∑PCBs (sum of 85 congeners) ranged from 1.52 to 22.4 ng L−1. The highest concentrations of PCBs were generally found in the Lower Fox River and in the Indiana Harbor and Ship Canal. The input flows of ∑PCBs from wet deposition, dry deposition, tributary loading, and air to water exchange, and the output flows due to sediment burial, volatilization from water to air, and transport to Lake Huron and through the Chicago Diversion were calculated, as well as flows related to the internal processes of settling, resuspension, and sediment−water diffusion. The net transfer of ∑PCBs is 1240 ± 531 kg yr−1 out of the lake. This net transfer is 46% lower than that estimated in 1994−1995. PCB concentrations in most matrices in the lake are decreasing, which drove the decline of all the individual input and output flows. Atmospheric deposition has become negligible, while volatilization from the water surface is still a major route of loss, releasing PCBs from the lake into the air. Large masses of PCBs remain in the water column and surface sediments and are likely to contribute to the future efflux of PCBs from the lake to the air.
■
INTRODUCTION
To improve water quality and to protect Lake Michigan’s ecosystem, the U.S. EPA initiated the Lake Michigan Mass Balance Project (LMMBP) in 1994−1995.15 Through two years’ field data collection, the concentrations of 65−110 individual PCB congeners were measured in air, water, sediment, and biota samples from all around the Lake.2 One objective of the LMMBP was to estimate the PCB loadings from different environmental media and to construct mass budgets and mass balance models.15 Two mass balance models, MICHTOX and LM2-Toxic, were developed. MICHTOX is a screening-level model and provided preliminary assessment of PCB fate, while LM2-Toxic is a more sophisticated model that looked at transport and fate of PCBs, coupling the mass balance of organic carbon solids and PCB dynamics.16 In these models, detailed PCB mass flow rates, phase distributions, and mass inventories in different compartments of the lake were computed. The primary conclusions of the first LMMBP effort were: (a) air−water exchange was the major process for PCB movement in the system and (b) reducing the atmospheric input of PCB to Lake Michigan was a critical step in the overall reduction of PCB levels in the lake.16,17 Before the LMMBP,
Polychlorinated biphenyls (PCBs) were first synthesized in 1881 and were widely used in commercial products starting in the 1930s.1 Because of their detrimental impact on human health and environmental persistence, PCB production and open uses in the U.S. ended in 1977.2 The estimated total U.S. production and use of commercial PCBs from 1930 to 1977 was between 500 000 and 642 000 t.3−5 In the Great Lakes region, PCBs were first found in fish (lake trout and bloater chubs from Lake Michigan) in 1968.6 PCBs in the Great Lakes were coming from paper mills, tanneries, foundries, power plants, and machine shops.7 For example, Green Bay in Lake Michigan was heavily contaminated with PCBs because of numerous paper mills in that area. Some of these mills specialized in recycling carbonless copy paper, which was manufactured with PCBs.8 PCB concentrations in Great Lakes air and sediment reached their peak between the 1950s and 1970s and began declining after restrictions were put in place in the late 1970s.9,10 PCB contamination of water and fish from the Great Lakes have also been decreasing,11,12 but concentrations in these media remain a public health and ecological concern.13 For example, PCBs that had previously accumulated in sediment are now re-equilibrating back into the water and fish, and PCBs in water are being released to the air.9,14 © XXXX American Chemical Society
Received: June 6, 2017 Revised: September 20, 2017 Accepted: September 25, 2017
A
DOI: 10.1021/acs.est.7b02904 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
Figure 1. Sampling sites for air, open lake water, sediment, and tributary water in Lake Michigan. Each polygon shape shows the drainage area of each tributary.
Sampling and Sample Pretreatment. The sampling sites for the tributaries, air, open lake water, and sediments are shown in Figure 1. Tributary water samples, including both the dissolved and particle phases, were collected every 3 weeks in 2015 from the same sites in the Grand, Kalamazoo, St. Joseph, and Lower Fox Rivers and from the Indiana Harbor and Ship Canal. A total of 59 samples were collected from these five tributaries by two USGS field offices (Middleton, Wisconsin, and Lansing, Michigan) between April and December of 2015. These five tributaries were selected because they showed the highest loads of PCBs in the 1994−1995 Lake Michigan Mass Balance Project (LMMBP).19 For each sampling, approximately 160 L of water was pumped through a glass fiber filter (Whatman GF/F, nominal pore size: 0.7 μm) to collect the particle phase and then through a glass column packed with precleaned XAD-2 resin to collect the dissolved phase. This sampling procedure was similar to that in the LMMBP, where 80−160 L of water were pumped through a 0.7 μm filter.19 The dissolved phase (XAD-2) and the particle phase (glass fiber filters) were analyzed separately. Briefly, the XAD-2 resin or filter was loaded in a Soxhlet extractor, spiked with the surrogate standards (PCB-14, PCB-65, and PCB-166) and extracted for 30 h using 800 mL of a 1:1 (v:v) mixture of hexane and acetone. The extract was then concentrated by rotary evaporation to a volume of about 100 mL, and the water and hexane layers were separated. The water layer was back extracted with 100 mL of hexane and then with 75 mL of hexane twice. The four hexane extracts were then combined and further concentrated. Each concentrated sample was fractionated on a 3.0% water deactivated silica gel column using 35 mL of hexane. Further concentration of the extract was completed using nitrogen blowdown to about 1 mL, at which point, PCB-30 and PCB-204 were added as internal standards. Additional details about the sampling, extraction, and analyte isolation methods can be found elsewhere.20,21
the MICHTOX model was also applied to estimate PCB loadings in 1970s. According to the pre-LMMBP model results, atmospheric deposition was the major input of PCBs to Lake Michigan and air−water exchange was controlled by gas absorption.18 This study revisits and updates the Lake Michigan PCB mass budget study conducted in 1994−1995.16 In this study, 5 of the 11 original LMMBP tributaries were revisited in 2015. PCB concentrations were measured as 85 individual congeners and reported as the sum of all measured congeners (∑PCB). Data from these five tributaries were then combined with the most recent concentrations of PCBs in air, open lake water, and sediment to calculate an updated mass budget. The specific objectives of this study were to (a) measure tributary water concentrations of PCBs and compare differences among tributaries; (b) estimate ∑PCB flows (mass loadings) associated with atmospheric deposition, tributary input, sediment burial, sediment−water interaction, and water movement, as well as the inventories of ∑PCBs in the lake’s water and sediment; (c) compare ∑PCB flows from this study to the 1994−1995 mass balance results and to the 1970s MICHTOX results; and (d) characterize the implied health effects of PCBs based on the updated mass budget for Lake Michigan.
■
MATERIALS AND METHODS
Chemicals. All solvents were HPLC or Optima grade. Silica gel (100−200 mesh, 75−150 μm, grade 644) and granular anhydrous sodium sulfate (Na2SO4) were purchased from Fisher Scientific (Pittsburgh, PA). They were baked at 300 °C for >12 h and cooled in a desiccator. Silica gel was deactivated with 3.0% water (by weight) 12 h before use. PCB standards were custom-made and were from AccuStandard (New Haven, CT). This study analyzed 85 PCB congeners, which are listed in the Supporting Information (SI). B
DOI: 10.1021/acs.est.7b02904 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
Table 1. Comparison of ∑PCB Concentrations and Number of Samples (n) between 2010−2015 (Geometric Mean ± One Standard Error) and 1994-1995 (Arithmetic Mean ± One Standard Error) In Lake Michigan and at Sleeping Bear Dunes and at Chicago 2010−2015 n 11 11 12 12 13
22.4 1.52 5.49 1.66 19.4
air, gas (pg/m3) Chicago Sleeping Bear Dunes
30 29
± ± ± ± ±
6.5 0.30 0.33 0.32 4.2
conc. ± ± ± ± ±
39 47 38 33 15b
53.0 2.36 22.9 2.90 76.0
282 ± 60c 36.5 ± 5.0c
19d (18)e 15d (18)e
2600 ± 436d (2600 ± 448)e 380 ± 142d (145 ± 4)e
9.08 ± 1.18f 3.12 ± 0.37f
19d (18)e 15d (18)e
91 ± 11d (85 ± 11)e 18 ± 5d (12 ± 1)e
10
1340 ± 175c 263 ± 44g
17 16
16000 ± 6791 1300 ± 220
1 1 3 4 3
0.19i 0.30i 0.37 ± 0.07i 0.36 ± 0.05i 51.5 ± 2.2j
8 11 336 347 119
0.12 0.30 0.19 0.19 50.5
air, particle (pg/m3) Chicago Sleeping Bear Dunes
water, open lake dissolved (ng/L) Straits of Mackinac (64a)h Chicago (popu)h sites not including Chicago (pop1, 50b and pop2)h whole lake (pop1, 50b, pop2 and popu)h sediment (ng/g dw)
n
conc.
water, tributaries (ng/L)a Lower Fox River Grand River Kalamazoo River St. Joseph River Indiana Harbor & Ship Canal
precipitation (pg/L) Chicago Sleeping Bear
1994−1995
± ± ± ± ±
4.18 0.11 1.59 0.19 5.9b
0.02 0.08 0.08 0.01 4.0k
a
Dissolved plus particle. bFrom Grand Calumet. cFrom Integrated Atmospheric Deposition Network (IADN) data of 2015. dFrom LMMBP report;2 From Miller et al.;51 fEstimated from concentration ratio between gas and particle phases in Miller et al.51 and the gas concentration in IADN data of 2015. gEstimated from concentrations in mass balance report2 and half-life reported for Chicago;10 hSampling sites in Figure 1. iFrom Venier et al.13 with slight modification: replacing the nondetected samples with one-third of minimum detected concentrations for each congener. jFrom Great Lakes Sediment Surveillance Program (0−1 cm). kFrom LMMBP (0.5 or 1.5 cm thickness). e
internal standards. PCBs in sediment samples were measured on an Agilent 7890 gas chromatograph coupled with an Agilent 7001B triple quadrupole mass spectrometer with an electron impact ionization source. The multimode injection port was operated in the solvent vent mode with 60 μL (20 μL × 3) of total injection per run. A DB-5MS (J&W Scientific) 60-m column (250 μm i.d., 0.25 μm film thickness) was used for separation. Data were acquired in multiple reaction monitoring mode. Detailed information about instrument conditions and ions is given in the Supporting Information (SI). Quality Control. Procedural blanks and field blanks were processed along with samples. For tributary water samples, there were 15 laboratory procedural blanks and two field blanks (one per sampling team). The levels of most PCBs in the procedural blanks were
