&EPA
United States
Environmental Protection
Agency
Results of the Lake Michigan Mass
Balance Study: Polychlorinated
Biphenyls and frans-Nonachlor Data
Report
April 2004
-------
U.S. Environmental Protection Agency
Great Lakes National Program Office (G-17J)
77 West Jackson Boulevard
Chicago, IL 60604
EPA905R-01-011
-------
Results of the Lake Michigan Mass Balance Study:
Poly chlorinated Biphenyls and /rara-Nonachlor Data Report
Prepared for:
US EPA Great Lakes National Program Office
77 West Jackson Boulevard
Chicago, IL 60604
Prepared by:
Harry B. McCarty, Ph.D.
Judy Schofield,
Ken Miller, and
Robert N. Brent, Ph.D.
DynCorp - A CSC Company
6101 Stevenson Avenue
Alexandria, Virginia 22304
and
Patricia Van Hoof, Ph.D.
and Brian Eadie, Ph.D.
Great Lakes Environmental Research Laboratory
National Oceanic and Atmospheric Administration
2205 Commonwealth Boulevard
Ann Arbor, Michigan 48105
April 2004
-------
Acknowledgments
This report was prepared under the direction of Glenn Warren, Project Officer, and Louis Blume, Work
Assignment Manager and Quality Assurance Program Manager, USEPA Great Lakes National Program
Office. The report was prepared by Harry B. McCarty, Judy Schofield, Ken Miller, and Robert N. Brent,
with DynCorp's Science and Engineering Programs, and Patricia Van Hoof and Brian Eadie, with the
Great Lakes Environmental Research Laboratory, National Oceanic and Atmospheric Administration,
with significant contributions from the LMMB Principal Investigators for PCBs and fra«s-nonachlor and
Molly Middlebrook of DynCorp. GLNPO thanks these investigators and their associates for their
technical support in project development and implementation. Additional assistance and graphics for
Chapter 5 were provided by Michael Mullin and Ronald Rossmann of the USEPA Large Lakes Research
Station at Grosse lie, Michigan. GLNPO thanks Bill Sukloff of Environment Canada and Syd Allan for
use and implementation of the Research Data Management Quality System (RDMQ).
GLNPO also thanks the following reviewers of the draft report for their comments and observations:
Dr. Chris Collins, Imperial College, London, U.K., and Dr. Joseph K. Comeau, Alburg, VT.
LMMB Principal Investigators for PCB and fraws-Nonachlor
Clyde Sweet (atmosphere)
Illinois State Water Survey
2204 Griffith Drive
Champaign, IL 61820
Ron Kites and Ilora Basu (atmosphere)
Indiana University
1005 East Tenth Street
Bloomington, IN 47405
Steve Eisenreich (atmosphere)
University of Minnesota
Gray Freshwater Biological Institute, and
Rutgers University
New Brunswick, NJ 08901
William Sonzogni (tributary)
University of Wisconsin
Wisconsin State Lab of Hygiene
465 Henry Mall
Madison, WI 53706
Eric Crecelius (open lake)
Battelle Marine Science Laboratories
1529 West Sequim Bay Road
Sequim,WA 98382
Patricia Van Hoof and Brian Eadie (sediment)
Great Lakes Environmental Research Laboratory
National Oceanic and Atmospheric Administration
2205 Commonwealth Boulevard
Ann Arbor, MI 48105
Deborah Swackhamer (lower pelagic food web)
University of Minnesota
School of Public Health
Box 807 Mayo
420 Delaware St.
Minneapolis, MN 55455
Robert Hesselberg and James Hickey (fish)
United State Geological Survey
Biological Resources Discipline
(formerly National Biological Service)
1451 Green Road
Ann Arbor, MI 48105
April 2004
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Portions of this document were developed using information presented in the following publications and
internal reports:
Franz, T. P., S. J. Eisenreich, and T. Holsen. 1998. Dry deposition of parti culate poly chlorinated
biphenyls andpolycyclic aromatic hydrocarbons to Lake Michigan. Environ. Sci. Technol.
32(23): 3681-3688.
Green, M. L., J. V. DePinto, C. W. Sweet, and K. C. Hornbuckle. 2000. Regional Spatial and Temporal
Interpolation of Atmospheric PCBs: Interpretation of Lake Michigan Mass Balance Data.
Environ. Sci. Technol. 34(9): 1833-1850.
Holson, T. M., G. J. Keeler, K. N. Noll, G. Fang, W. Lee, and J. Lin. 1993. Dry Deposition and Particle
Size Distributions Measured during the Lake Michigan Urban Air Toxics Study. Environ. Sci.
Technol. 27(7): 1327-1333.
Hornbuckle, K. C., J. V. DePinto, M. L. Green, S. M. Miller, and J. J. Bogdan. 2001. Atmospheric
Deposition of Persistent Organic Pollutants: Results of the Lake Michigan Mass Balance Study.
Green Bay, Wisconsin, International Association for Great Lakes Research: Abstracts, June 10 -
14,2001.
Hornbuckle, K. C., C. W. Sweet, R. F. Pearson, D. L. Swackhamer, and S. J. Eisenreich. 1995.
Assessing annual water-air fluxes ofpolychlorinated biphenyls in Lake Michigan. Environ. Sci.
Technol. 29(4): 869.
Hornbuckle, K. C., D. R. Achman, and S. J. Eisenreich. 1993. Over-Water and Over-Land
Polychlorinated Biphenyls in Green Bay, Lake Michigan. Environ. Sci. Technol. 27(1): 87-98.
Madenjian, C. P., T. J. DeSorcie, R. M. Stedman, E. H. Brown, Jr., G. W. Eck, L. J. Schmidt, R. J.
Hesselberg, S. M. Chernyak, and D. R. Passino-Reader. 1999. Spatial Patterns in PCS
Concentrations of Lake Michigan Lake Trout. J. Great Lakes Res. 25(1): 149-159.
Madenjian, C. P., L. J. Schmidt, S. M. Chernyak, R. F. Elliott, T. J. DeSorcie, R. T. Quintal, L. J.
Begnoche, and R. J. Hesselberg. 1999. Variation in Net Trophic Transfer Efficiencies among 21
PCB Congeners. Environ. Sci. Technol. 33(21): 3768-3773.
Madenjian, C. P., R. F. Elliot, L. J. Schmidt, T. J. DeSorcie, R. J. Hesselberg, R. T. Quintal, L. J.
Begnoche, P. M. Bouchard, M. E. and Holey. 1998. Net Trophic Transfer Efficiency of PCBs to
Lake Michigan Coho Salmon from Their Prey. Environ. Sci. Technol. 32(20): 3063-3067.
Miller, S. M. 1999. Spatial and Temporal Variability of Organic and Nutrient Compounds in
Atmospheric Media Collected During the Lake Michigan Mass Balance Study. M.S. thesis.
University of New York at Buffalo, Buffalo, New York. 181 pp.
Simcik, M. F., R. M. Hoff, W. M. J. Strachan, C. W. Sweet, I. Basu, and R. A. Kites. 2000. Temporal
Trends of Semivolatile Organic Contaminants in Great Lakes Precipitation. Environ. Sci.
Technol. 34(3): 361-367.
Sweet, C. 2000. Sampling of Atmospheric PCBs in the Lake Michigan Mass Balance Study (LMMB),
1994-1995. Internal document. Great Lakes National Program Office.
April 2004
-------
Totten, L. A., C. L. Gigliotti, and S. J. Eisenreich. 2001. Re-evaluation of Air-water Exchange Fluxes of
PCBs in Green Bay and Southern Lake Michigan in AEOLOS. Green Bay, Wisconsin.
International Association for Great Lakes Research: Abstracts, June 10 - 14, 2001.
Trowbridge, A. G. and D. L. Swackhamer. 2001. An Analysis of Poly chlorinated Biphenyl
Concentrations in Lower Trophic Level Organisms of the Lake Michigan Foodweb. Green Bay,
Wisconsin. International Association for Great Lakes Research: Abstracts, June 10-14, 2001.
Van Hoof, P. 2000. PCBs in Lake Michigan Surficial Sediments. Internal document. Great Lakes
National Program Office.
April 2004
-------
Table of Contents
Acknowledgments i
Executive Summary ES-1
Chapter 1 Project Overview .
1.1 Background
1.2 Description .
1.3 Scope
1.3.1 Modeled Pollutants
1.3.1.1 Poly chlorinated Biphenyls
1.3.1.2 fra«5-Nonachlor
1.3.1.3 Atrazine
1.3.1.4 Mercury
1.3.2 Other Measured Parameters
1.3.3 Measured Compartments . .
1.4 Objectives
1.5 Design . . .
-1
-1
-1
-2
-2
-3
-3
-4
-4
-6
-7
-9
-9
-9
-9
1.5.3 Workgroups 1-10
1.5.4 Information Management 1-11
1.5.4.1 Data Reporting 1-11
1.5.4.2 Great Lakes Environmental Monitoring Database 1-13
1.5.4.3 Public Access to LMMB Data 1-13
1.5.5 Quality Assurance Program 1-13
1.6 Project Documents and Products 1-15
Chapter 2 PCB/fra«s-Nonachlor Study Overview 2-1
2.1 PCB Introduction 2-1
1.5.1 Organization
1.5.2 Study Participants
. 1 Physical/Chemical Properties 2-1
.2 History of PCB Production 2-2
.3 Regulatory Background 2-3
.4 Fate and Effects 2-3
2. .5 Biological Transformations 2-4
2. .6 Toxicity 2-4
2.2 fra«5-Nonachlor Introduction 2-5
2.2.1 Physical/Chemical Properties 2-5
2.2.2 fra«5-Nonachlor Production 2-6
2.2.3 Regulatory Background 2-6
2.2.4 Fate and Effects 2-7
2.2.5 Toxicity 2-7
2.3 Study Design 2-7
2.3.1 Description 2-7
2.3.2 Scope 2-7
2.3.3 Organization/Management 2-7
2.4 Sampling Locations 2-8
2.4.1 Atmospheric Components 2-8
2.4.2 Tributaries 2-11
2.4.3 Open Lake 2-13
2.4.4 Sediment 2-14
April 2004 v
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
2.4.5 Lower Pelagic Food Web Organisms 2-14
2.4.6 Fish 2-15
2.5 Sampling Methods 2-16
2.5.1 Atmospheric Components 2-16
2.5.1.1 Vapor Fraction 2-16
2.5.1.2 Particulate Fraction 2-17
2.5.1.3 Precipitation Fraction 2-17
2.5.1.4 Dry Deposition 2-17
2.5.2 Tributaries 2-18
2.5.3 Open Lake 2-18
2.5.4 Sediment 2-19
2.5.5 Lower Pelagic Food Web Organisms 2-19
2.5.6 Fish 2-19
2.6 Analytical Methods 2-20
2.6.1 General Analytical Considerations 2-20
2.6.2 Data Presented in this Report 2-23
2.6.3 Atmospheric Components 2-24
2.6.4 Tributaries and Open Lake 2-24
2.6.5 Sediment 2-25
2.6.6 Lower Pelagic Food Web Organisms 2-25
2.6.7 Fish 2-25
2.7 Quality Implementation and Assessment 2-25
Chapter 3 PCBs/fra«s-Nonachlor in Atmospheric Components 3-1
3.1 Results 3-1
3.1.1 Vapor Fraction 3-5
3.1.1.1 Temporal Variation 3-12
3.1.1.2 Geographical Variation 3-15
3.1.2 Particulate Fraction 3-19
3.1.2.1 Temporal Variation 3-23
3.1.2.2 Geographical Variation 3-26
3.1.3 Precipitation Fraction 3-29
3.1.3.1 Temporal Variation 3-34
3.1.3.2 Geographical Variation 3-38
3.1.4 Dry Deposition 3-42
3.2 Quality Implementation and Assessment 3-43
3.3 Data Interpretation 3-54
3.3.1 Atmospheric Sources 3-54
3.3.2 Atmospheric Concentrations 3-56
3.3.3 Seasonality 3-58
3.3.4 Regional Considerations 3-58
Chapter 4 PCBs/frans-Nonachlor in Tributaries 4-1
4.1 Results 4-1
4.1.1 Temporal Variation 4-3
4.1.2 Geographical Variation 4-9
4.2 Quality Implementation and Assessment 4-19
4.3 Data Interpretation 4-25
4.3.1 Comparison to Historical Studies 4-25
4.3.2 Regional Considerations 4-27
4.3.3 Other Interpretations and Perspectives 4-27
vi April 2004
-------
Table of Contents
Chapter 5 PCBs/fra«s-Nonachlor in Open-lake Water 5-1
5.1 Results 5-1
5.1.1 Temporal Variation 5-4
5.1.2 Geographical Variation 5-4
5.2 Quality Implementation and Assessment 5-26
5.2.1 Sample Collection 5-26
5.2.2 Data Assessments 5-26
5.2.3 Evaluation of Blanks 5-31
5.3 Data Interpretation 5-38
5.3.1 Comparison to Historical Studies 5-38
5.3.2 Regional Considerations 5-45
5.3.3 Other Interpretations and Perspectives 5-48
Chapter 6 PCBs/frans-Nonachlor in Sediment 6-1
6.1 Results 6-1
6.1.1 Geographical Variation 6-3
6.1.2 Frequency Distributions 6-5
6.2 Quality Implementation and Assessment 6-7
6.2.1 Sample Collection 6-8
6.2.2 Data Assessments 6-8
6.3 Data Interpretation 6-12
6.3.1 Comparison to Historical Studies 6-12
6.3.2 Other Interpretations and Perspectives 6-14
6.3.2.1 Relationship of PCB Congeners and trans-Nonachlor with Sediment
Organic Carbon 6-14
6.3.2.2 Contour Maps 6-18
Chapter 7 PCBs/fra«s-Nonachlor in the Lower Pelagic Food Web 7-1
7.1 Results 7-1
7.1.1 Sample Type and Species Variation 7-3
7.1.2 Seasonal Variation 7-7
7.1.3 Geographical Variation 7-8
7.1.4 Bioaccumulation 7-12
7.2 Quality Implementation and Assessment 7-13
7.3 Data Interpretation 7-17
7.3.1 Comparison to Historical Studies 7-17
7.3.2 Seasonal Variation 7-18
7.3.3 Bioaccumulation and Trophic Transfer 7-19
7.3.4 Fractionation 7-21
Chapter 8 PCBs/fra«s-Nonachlor in Fish 8-1
8.1 Results 8-1
. 1 Species Variation 8-2
.2 Factors Affecting Contaminant Concentrations 8-8
.3 Geographical Variation 8-11
.4 Seasonal Variation 8-13
.5 Bioaccumulation 8-14
8.2 Quality Implementation and Assessment 8-15
April 2004 vii
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
8.3 Data Interpretation 8-20
8.3.1 Comparison to Fish Advisory Levels 8-20
8.3.2 Comparison to Historical Studies 8-21
8.3.3 Regional Considerations 8-21
8.3.4 Factors Affecting Contaminant Concentrations 8-22
8.3.5 Bioaccumulation and Trophic Transfer 8-22
Chapter 9 Cross-Media Interpretations 9-1
9.1 Summary of PCB and fra«s-Nonachlor Concentrations in Lake Michigan
Compartments 9-1
9.2 Bioaccumulation and Biomagnification 9-7
9.3 Fractionation 9-10
9.4 Toxic PCB Congeners 9-12
References R-l
viii April 2004
-------
Table of Contents
List of Tables
Table 1-1. Characteristics of Lake Michigan Mass Balance Modeled Pollutants 1-5
Table 1-2. Lake Michigan Mass Balance Study Parameters 1-6
Table 2-1. Numbers of Congeners in Each Level of Chlorination 2-1
Table 2-2. U.S. Domestic Production of Commercial Aroclor Mixtures from 1957 to 1977 2-2
Table 2-3. Typical Composition (%) of the Five Aroclor Mixtures with Greatest U.S. Production 2-3
Table 2-4. World Health Organization Toxic PCB Congeners 2-5
Table 2-5. Components Sampled by Principal Investigators 2-8
Table 2-6. Site Classifications, Planned Frequencies, and Types of Vapor-phase and
Particulate-phase Samples Collected at Shoreline and Out-of-basin Stations 2-9
Table 2-7. Watershed Characteristics for Tributaries Monitored in the LMMB Study 2-12
Table 2-8. Open-lake Cruise Dates and Number of Stations Occupied 2-13
Table 2-9. Number of Fish Collected by Species and Location 2-15
Table 2-10. Number of Fish Collected by Technique 2-16
Table 2-11. PCB Congeners that Contribute at Least 1% of the Mass of the Mullin Mix 2-22
Table 2-12. Maximum Number of PCB Congeners Reported, by Laboratory 2-23
Table 3-1. Number of Atmospheric Samples Collected and Analyzed for PCB Congeners
and Total PCBs 3-2
Table 3-2. Number of Atmospheric Samples Collected and Analyzed for fra«s-Nonachlor 3-4
Table 3-3. Monthly Composite Concentrations of Vapor-phase PCB 33 Measured in Samples
Collected Around Lake Michigan from April 1994 to October 1995 3-7
Table 3-4. Monthly Composite Concentrations of Vapor-Phase PCB 118 Measured in Samples
Collected Around Lake Michigan from April 1994 to October 1995 3-8
Table 3-5. Monthly Composite Concentrations of Vapor-Phase PCB 180 Measured in Samples
Collected Around Lake Michigan from April 1994 to October 1995 3-9
Table 3-6. Monthly Composite Concentrations of Vapor-Phase Total PCBs Measured in Samples
Collected Around Lake Michigan from April 1994 to October 1995 3-10
Table 3-7. Monthly Composite Concentrations of Vapor-Phase fra«s-Nonachlor Measured in
Samples Collected Around Lake Michigan from April 1994 to October 1995 3-11
Table 3-8. Mean Monthly Composite Concentrations of Vapor-Phase Total PCBs, PCB 118,
and fra«5-Nonachlor at LMMB Study Sampling Stations in and around Lake
Michigan from April 1994 to October 1995 3-16
Table 3-9. Monthly Composite Concentrations of Particulate-Phase PCB 33 Measured in
Samples Collected Around Lake Michigan from April 1994 to October 1995 3-20
Table 3-10. Monthly Composite Concentrations of Particulate-phase PCB 118 Measured in
Samples Collected Around Lake Michigan from April 1994 to October 1995 3-21
Table 3-11. Monthly Composite Concentrations of Particulate-Phase PCB 180 Measured in
Samples Collected Around Lake Michigan from April 1994 to October 1995 3-21
Table 3-12. Monthly Composite Concentrations of Particulate-phase Total PCBs Measured in
Samples Collected Around Lake Michigan from April 1994 to October 1995 3-22
Table 3-13. Monthly Composite Concentrations of Particulate-Phase fra«s-Nonachlor Measured
in Samples Collected Around Lake Michigan from April 1994 to October 1995 .... 3-23
Table 3-14. Mean Monthly Composite Concentrations of Particulate-phase Total PCBs, PCB 33,
and fra«5-Nonachlor at LMMB Study Sampling Stations in and around Lake
Michigan between April 1994 and October 1995 3-27
Table 3-15. Monthly Composite Concentrations of PCB 33 Measured in Precipitation Samples
Collected around Lake Michigan from April 1994 to October 1995 3-30
Table 3-16. Monthly Composite Concentrations of PCB 118 Measured in Precipitation Samples
Collected around Lake Michigan from April 1994 to October 1995 3-31
April 2004
IX
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 3-17. Monthly Composite Concentrations of PCB 180 Measured in Precipitation Samples
Collected around Lake Michigan from April 1994 to October 1995 3-32
Table 3-18. Monthly Composite Concentrations of total PCBs Measured in Precipitation Samples
Collected around Lake Michigan from April 1994 to October 1995 3-33
Table 3-19. Monthly Composite Concentrations of fra«s-Nonachlor Measured in Precipitation
Samples Collected around Lake Michigan from April 1994 to October 1995 3-34
Table 3-20. Mean Precipitation Concentrations of Total PCB, PCB 33 and fra«s-Nonachlor at
LMMB Study Sampling Stations in and around Lake Michigan between
March 1994 and October 1995 3-39
Table 3-21. Monthly Composite Concentrations of PCBs and fra«s-Nonachlor Measured in
Dry Deposition 3-43
Table 3-22. Field Sample Flags Applied to Select PCB Congeners and fra«s-Nonachlor Results
in Atmospheric Samples Analyzed at Illinois Water Survey 3-46
Table 3-23. Field Sample Flags Applied to Select PCB Congener and fra«s-Nonachlor Results
in Atmospheric Samples Analyzed at Indiana University 3-47
Table 3-24. Field Sample Flags Applied to Select PCB Congener and fra«s-Nonachlor Results
in Dry Deposition Atmospheric Samples 3-48
Table 3-25. Data Quality Assessment for PCB 33 in Atmospheric Samples 3-52
Table 3-26. Data Quality Assessment for PCB 118 in Atmospheric Samples 3-52
Table 3-27. Data Quality Assessment for PCB 180 in Atmospheric Samples 3-53
Table 3-28. Data Quality Assessment for fra«s-Nonachlor in Atmospheric Samples 3-53
Table 4-1. Number of Tributary Samples Analyzed for Dissolved and Particulate PCB
Congeners and Total PCBs 4-1
Table 4-2. Number of Tributary Samples Analyzed for Dissolved and Particulate
fra«5-Nonachlor 4-2
Table 4-3. Tributary Classifications Relative to Responsiveness to Precipitation Events 4-3
Table 4-4. Concentrations of PCB Congener 33 Measured in Tributaries 4-15
Table 4-5. Concentrations of PCB Congener 118 Measured in Tributaries 4-16
Table 4-6. Concentrations of PCB Congener 180 Measured in Tributaries 4-17
Table 4-7. Concentrations of Total PCBs Measured in Tributaries 4-18
Table 4-8. Concentrations of fra«s-Nonachlor Measured in Tributaries 4-19
Table 4-9. Summary of Routine Field Sample Flags Applied to Select PCB Congeners and
fra«5-Nonachlor in Tributary Samples 4-22
Table 4-10. Data Quality Assessment for Select PCB Congeners and fra«s-Nonachlor in
Tributary Samples 4-24
Table 4-11. Correlation of Particulate PCB and fra«s-Nonachlor Concentrations with
Chlorophyll a in the Fox and Milwaukee Rivers 4-27
Table 4-12. Correlation of Particulate Total PCB Concentrations with Chlorophyll a and
Total Solids in Lake Michigan Tributaries 4-28
Table 5-1. Numbers of Open-lake Samples Analyzed for Dissolved and Particulate
PCB Congeners and Total PCBs 5-2
Table 5-2. Number of Open-lake Samples Analyzed for Dissolved and Particulate
fra«5-Nonachlor 5-3
Table 5-3. Concentrations of Dissolved PCB Congener 33 Measured in Open-lake Samples .... 5-5
Table 5-4. Concentrations of Particulate PCB Congener 33 Measured in Open-lake Samples .... 5-6
Table 5-5. Concentrations of Dissolved PCB Congener 118 Measured in Open-lake Samples . . . 5-7
Table 5-6. Concentrations of Particulate PCB Congener 118 Measured in Open-lake Samples ... 5-8
Table 5-7. Concentrations of Dissolved PCB Congener 180 Measured in Open-lake Samples . . . 5-9
Table 5-8. Concentrations of Particulate PCB Congener 180 Measured in Open-lake Samples . . 5-10
Table 5-9. Concentrations of Dissolved Total PCBs Measured in Open-lake Samples 5-11
Table 5-10. Concentrations of Particulate Total PCBs Measured in Open-lake Samples 5-12
April 2004
-------
Table of Contents
Table 5-11. Concentrations of Dissolved fra«s-Nonachlor Measured in Open-lake Samples
Table 5-12. Concentrations of Particulate fra«s-Nonachlor Measured in Open-lake Samples ....
Table 5-13. Summary of Routine Field Sample Flags Applied to Select PCB Congeners and
fra«s-Nonachlor in Open-lake Samples
Table 5-14. Data Quality Assessment for Select PCB Congeners and fra«s-Nonachlor in
Open-lake Water Samples
Table 5-15. Criteria Used to Evaluate Data to be Included in the Estimation of the Mean
Concentrations of PCB Congeners
Table 5-16. Comparison of Summary Statistics for LMMB Open-lake PCB Congener Results
after Removal of Sample Results associated with Contaminated Blanks
Table 5-17. Results of North/South Comparisons of Open-lake Concentrations of PCBs and
fra«s-Nonachlor
Table 5-18. Results of Depth Comparisons of Open-lake Concentrations of PCBs and
fra«s-Nonachlor
Table 6-1. Summary Statistics for Surficial Sediments
Table 6-2. Summary of Routine Field Sample Flags Applied to Select PCB Congeners and
fra«s-Nonachlor in Sediment Samples
Table 6-3. Data Quality Assessment for Select PCB Congeners and fra«s-Nonachlor in
Sediment Samples
Table 6-4. Linear Regression Parameters of Log PCB and fra«s-Nonachlor versus Log OC
Content in Surficial Sediments in Lake Michigan
Table 6-5. Significant Differences between Linear Regression Parameters among Lake
Michigan Basins for Log Analyte versus Log Organic Carbon
Table 7-1. Number of Lower Pelagic Food Web Samples Analyzed for PCB Congeners and
fra«s-Nonachlor
Table 7-2. Mean Concentrations of PCBs and fra«s-Nonachlor Measured in the Lower
Pelagic Food Web
Table 7-3. Mean Concentrations of Total PCBs Measured in the Lower Pelagic Food Web
at Various Sampling Locations
Table 7-4. Mean Concentrations of fra«s-Nonachlor Measured in the Lower Pelagic Food Web
at Various Sampling Locations
Table 7-5. Bioaccumulation Factors for PCBs and fra«s-Nonachlor in the Lower Pelagic
Food Web
Table 7-6. Biomagnification Factors for PCBs and fra«s-Nonachlor between Primary
Producers and Primary Consumers
Table 7-7. Summary of Routine Field Sample Flags Applied to Select PCB Congeners
and fra«s-Nonachlor in the Lower Pelagic Food Web
Table 7-8. Data Quality Assessment for Select PCB Congeners and fra«s-Nonachlor in
Lower Pelagic Food Web Samples
Table 8-1. Number of Fish Samples Analyzed for PCB Congeners and fra«s-Nonachlor
Table 8-2. Mean Concentrations of Specific PCB Congeners in Lake Michigan Fish
(Wet-weight Basis)
Table 8-3. Mean Concentrations of Total PCBs and fra«s-Nonachlor in Lake Michigan Fish
(Wet-weight Basis)
Table 8-4. Mean Concentrations of Total PCBs and fra«s-Nonachlor in Lake Michigan Fish
(Dry-weight Basis)
Table 8-5. Mean Concentrations of Total PCBs and fra«s-Nonachlor in Lake Michigan Fish
(Lipid-weight Basis)
Table 8-6. Correlation Between Log-transformed Total PCBs and fra«s-Nonachlor
Concentrations in Lake Michigan Fish and Fish Length and Lipid Content
5-13
5-14
5-29
5-30
5-32
5-33
5-45
5-47
. 6-4
6-10
6-11
6-17
6-18
. 7-2
. 7-5
. 7-9
7-11
7-12
7-13
7-14
7-16
. 8-2
. 8-3
. 8-4
. 8-6
. 8-8
. 8-9
April 2004
XI
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 8-7. Results of Multiple Regression Significance Test for Effects of Fish Length and Lipid
Content on Concentrations of Total PCBs and fra«s-Nonachlor Concentrations in
Lake Michigan Fish 8-10
Table 8-8. Bioaccumulation Factors for PCBs and fra«s-Nonachlor in Lake Michigan Fish .... 8-15
Table 8-9. Summary of Routine Field Sample Flags Applied to Select PCB Congeners and
fra«5-Nonachlor in Fish 8-17
Table 8-10. Data Quality Assessment for Select PCB Congeners and fra«s-Nonachlor
in Fish Samples 8-19
Table 9-1. Summary of Samples from each Ecosystem Compartment with Detectable
Levels of PCBs and fra«s-Nonachlor 9-1
Table 9-2. Toxic PCB Congeners and Toxicity Equivalency Factors (TEF) 9-12
Table 9-3. Toxic Equivalent Concentrations (TEQs) for Dioxin-like PCB Congeners in
Lake Michigan Fish 9-14
XII
April 2004
-------
Table of Contents
List of Figures
Figure 1-1. Simplified Mass Balance Approach 1-2
Figure 1-2. Lake Michigan Mass Balance Study Sampling Locations 1-8
Figure 1-3. Flow of Information in the Lake Michigan Mass Balance Study 1-12
Figure 2-1. Generalized 2-D Structure of Biphenyl 2-1
Figure 2-2. Generalized 2-D Structure of fra«s-Nonachlor 2-5
Figure 2-3. Generalized 2-D Structure of fra«s-Chlordane 2-6
Figure 2-4. Atmospheric Sampling Stations 2-10
Figure 2-5. Tributary Sampling Stations 2-11
Figure 2-6. Open-Lake Water Column Sampling Stations 2-13
Figure 2-7. Sediment Sampling Stations 2-14
Figure 2-8. Sediment Trap Locations 2-14
Figure 2-9. Sampling Stations for Lower Pelagic Food Web Organisms and Fish 2-15
Figure 3-1. Temporal Variations of Total Vapor-Phase PCB and fra«s-Nonachlor Concentrations
Measured at Lake Michigan Shoreline and Out-of-Basin Stations from April 1994
to October 1995 3-13
Figure 3-2. Seasonal Differences in Vapor-phase Total PCB Concentrations Measured at Lake
Michigan Shoreline and Out-of-basin Stations from April 1994 and October 1995 . . 3-14
Figure 3-3. Seasonal Differences in fra«s-Nonachlor Concentrations Measured at Lake Michigan
Shoreline and Out-of-Basin Stations from April 1994 to October 1995 3-15
Figure 3-4. Vapor-phase PCB 118 Concentrations Measured at Shoreline and Out-of-basin
Sampling Stations around Lake Michigan from April 1994 to October 1995 3-17
Figure 3-5. Vapor-phase fra«s-Nonachlor Concentrations Measured at Sampling Stations
around Lake Michigan from April 1994 to October 1995 3-18
Figure 3-6. Concentrations of PCB 118 in Vapor Measured in Over-water Samples Collected
in the Northern and Southern Areas of Lake Michigan 3-19
Figure 3-7. Temporal Variations in Particulate-phase PCB 118 Concentrations Measured at Lake
Michigan Shoreline and Out-of-basin Stations from April 1994 to October 1995 .... 3-24
Figure 3-8. Seasonal Differences in Particulate-phase Total PCB Concentrations Measured at Lake
Michigan Shoreline and Out-of-basin Stations from April 1994 to October 1995 .... 3-25
Figure 3-9. Seasonal Differences in Particulate-phase fra«s-Nonachlor Concentrations Measured
at Lake Michigan Shoreline and Out-of-basin Stations from April 1994 to
October 1995 3-26
Figure 3-10. Particulate-phase PCB 118 Concentrations Measured at Lake Michigan Shoreline
and Out-of basin Stations from April 1994 to October 1995 3-28
Figure 3-11. Particulate-phase fra«s-Nonachlor Concentrations Measured at Lake Michigan
Shoreline and Out-of-basin Stations from April 1994 to October 1995 3-28
Figure 3-12. Temporal Variation in Precipitation PCB 33 and fra«s-Nonachlor Concentrations
Measured at Lake Michigan Shoreline and Out-of-basin Stations from
March 1994 to October 1995 3-35
Figure 3-13. Seasonal Patterns of PCB 33 Concentrations in Precipitation Measured at
Lake Michigan Shoreline and Out-of-basin Stations from March 1994
to October 1995 3-36
Figure 3-14. Percent of Precipitation PCB 33 Sample Results Reported as Zero, by Season 3-37
Figure 3-15. Seasonal Patterns of fra«s-Nonachlor Concentrations in Precipitation Measured
at Lake Michigan Shoreline and Out-of-basin Stations from March 1994
to October 1995 3-37
Figure 3-16. Percent of Precipitation fra«s-Nonachlor Sample Results Reported as Zero,
by Season 3-38
April 2004 xiii
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 3-17. Precipitation-phase Total PCB Concentrations Measured at Lake Michigan
Shoreline and Out-of-basin Stations from March 1994 to October 1995 3-40
Figure 3-18. Precipitation-phase fra«s-Nonachlor Concentrations Measured at Lake Michigan
Shoreline and Out-of-basin Stations from March 1994 to October 1995 and the
Percent of Sample Results Reported as Zero for each Sampling Location 3-41
Figure 3-19. Mean Percentage of Individual PCB Congener's Contribution to Total PCB
Concentrations 3-55
Figure 3-20. Percentages of PCBs 33 and 180 in Precipitation Samples Reported as Zero,
Below the Detection Limits, and Above the Detection Limits 3-56
Figure 3-21. Proportion of High Molecular Weight PCB Congeners in the Vapor Phase,
Particulate Phase, and Precipitation 3-57
Figure 4-1. Temporal Variation in Total Dissolved PCB Concentrations Measured in
Lake Michigan Tributaries 4-4
Figure 4-2. Temporal Variation in Total Particulate PCB Concentrations Measured in
Lake Michigan Tributaries 4-5
Figure 4-3. Temporal Variation in Total Dissolved and Particulate fra«s-Nonachlor
Concentrations Measured in Lake Michigan Tributaries 4-7
Figure 4-4. Temporal Variation in Total Dissolved and Particulate fra«s-Nonachlor
Concentrations Measured in Lake Michigan Tributaries without the
Winter Mean for the Manistique River 4-8
Figure 4-5. Mean Dissolved and Particulate Concentrations of PCB 33 in
Lake Michigan Tributaries 4-9
Figure 4-6. Mean Dissolved and Particulate Concentrations of PCB 180 in
Lake Michigan Tributaries 4-10
Figure 4-7. Mean Dissolved Total PCB Concentrations in Lake Michigan Tributaries 4-11
Figure 4-8. Mean Particulate Total PCB Concentrations in Lake Michigan Tributaries in 1994
and 1995 4-12
Figure 4-9. Mean Dissolved fra«s-Nonachlor Concentrations in Lake Michigan Tributaries .... 4-13
Figure 4-10. Mean Particulate fra«s-Nonachlor Concentrations in Lake Michigan Tributaries
in 1994 and 1995 4-14
Figure 4-11. Mean Dissolved and Particulate Total PCB Concentrations in
Lake Michigan Tributaries 4-27
Figure 5-1. Open-lake Sampling Stations 5-1
Figure 5-2. Dissolved Total PCB by Cruise 5-4
Figure 5-3. Dissolved Total PCB by Cruise 5-4
Figure 5-4. Concentrations of Dissolved PCB Congener 33 Measured in Open-lake Samples ... 5-16
Figure 5-5. Concentrations of Dissolved PCB Congener 118 Measured in Open-lake Samples . . 5-17
Figure 5-6. Concentrations of Dissolved PCB Congener 180 Measured in Open-lake Samples . . 5-18
Figure 5-7. Concentrations of Particulate PCB Congener 33 Measured in Open-lake Samples . . . 5-19
Figure 5-8. Concentrations of Particulate PCB Congener 118 Measured in Open-lake Samples . . 5-20
Figure 5-9. Concentrations of Particulate PCB Congener 180 Measured in Open-lake Samples . . 5-21
Figure 5-10. Concentrations of Dissolved fra«s-Nonachlor Measured in Open-lake Samples 5-22
Figure 5-11. Concentrations of Particulate trans-Nonachlor Measured in Open-lake Samples .... 5-23
Figure 5-12. Concentrations of Dissolved Total PCBs Measured in Open-lake Samples 5-24
Figure 5-13. Concentrations of Particulate Total PCBs Measured in Open-lake Samples 5-25
Figure 5-14. Summary Statistics for LMMB Open-lake PCB 33 Results after Removal of
Sample Results associated with Contaminated Blanks 5-35
Figure 5-15. Summary Statistics for LMMB Open-lake PCB 118 Results after Removal of
Sample Results associated with Contaminated Blanks 5-36
Figure 5-16. Summary Statistics for LMMB Open-lake PCB 180 Results after Removal of
Sample Results associated with Contaminated Blanks 5-37
xiv April 2004
-------
Table of Contents
Figure 5-17. Concentrations of Total PCBs Measured in Open-lake Samples in September 1980 . . 5-40
Figure 5-18. Concentrations of Total PCBs Measured in Open-lake Samples in September 1981 . . 5-41
Figure 5-19. Concentrations of Total PCBs Measured in Open-lake Samples in September 1991 . . 5-42
Figure 5-20 Concentrations of Total PCBs Measured in LMMB Open-lake Samples in
September 1995 5-43
Figure 5-21. Concentrations of Total PCBs Measured in All LMMB Open-lake Samples 5-44
Figure 6-1. Sediment Sampling Stations 6-1
Figure 6-2. Sediment Trap Locations 6-2
Figure 6-3. Lake Michigan Basins 6-3
Figure 6-4. Frequency Distributions of Total PCBs, fra«s-Nonachlor, and Organic Carbon 6-5
Figure 6-5. Frequency Distributions of Total PCBs by Basin 6-6
Figure 6-6. Frequency Distributions of fra«s-Nonachlor by Basin 6-7
Figure 6-7. Comparison of Measured Historical Total PCB Results in Surficial Sediment
and Interpolated Total PCB Results from Contours of the 1994-1995
LMMB Collections 6-13
Figure 6-8. Relationships of PCB/fra«s-Nonachlor Concentrations and Organic Carbon
Content for the Southern Basin 6-15
Figure 6-9. Relationships of PCB/frans-Nonachlor Concentrations and Organic Carbon
Content for the Central Basins 6-15
Figure 6-10. Relationships of PCB/fra«s-Nonachlor Concentrations and Organic Carbon
Content for the Northern Basin 6-16
Figure 6-11. Relationships of PCB/frans-Nonachlor Concentrations and Organic Carbon
Content for the Straits Region 6-16
Figure 6-12. Contour Plot of Organic Carbon Content in Lake Michigan Sediments 6-18
Figure 6-13. Contour Plot of Total PCBs in Lake Michigan Sediments 6-20
Figure 6-14. Contour Plot of PCB 28+31 in Lake Michigan Sediments 6-20
Figure 6-15. Contour Plot of PCB 118 in Lake Michigan Sediments 6-20
Figure 6-16. Contour Plot of PCB 180 in Lake Michigan Sediments 6-20
Figure 6-17. Contour Plot of fra«s-Nonachlor in Lake Michigan Sediments 6-21
Figure 6-18. Organic Carbon Content versus Depth in Lake Michigan 6-21
Figure 6-19. Plot of PCB 28+31 Concentrations versus PCB 180 Concentrations in
Lake Michigan Sediments 6-22
Figure 7-1. Total PCB and fra«s-Nonachlor Concentrations in the Lower Pelagic Food Web 7-4
Figure 7-2. Normalized (by Lipid Content) Total PCB and fra«s-Nonachlor Concentrations
in the Lower Pelagic Food Web 7-6
Figure 7-3. Seasonal Variation of Total PCB Concentrations Measured in the Lower Pelagic
Food Web of Lake Michigan 7-7
Figure 7-4. Seasonal Variation in fra«s-Nonachlor Concentrations Measured in the Lower
Pelagic Food Web of Lake Michigan 7-8
Figure 7-5. Lower Pelagic Food Web Structure and Biomagnification Factors for Total PCBs
(BMFp) and fra«s-Nonachlor (BMFT) 7-21
Figure 7-6. Mean Percentage of Individual PCB Congener Contribution to
Total PCB Concentrations 7-22
Figure 7-7. Relative Enrichment or Depletion of PCB Congener Homolog Groups between
Compartments 7-23
Figure 8-1. Total PCB and fra«s-Nonachlor Concentrations in Lake Michigan Fish
(Wet-weight Basis) 8-5
Figure 8-2. Total PCB and fra«s-Nonachlor Concentrations 8-7
Figure 8-3. Relationship between Fish Length and Total PCB and fra«s-Nonachlor
Concentrations in Lake Michigan Lake Trout 8-10
April 2004 xv
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 8-4. Total PCB and fra«s-Nonachlor Concentrations in Fish from Three Biological
Sampling Stations in Lake Michigan 8-11
Figure 8-5. Total PCB and fra«s-Nonachlor Concentrations in Lake Michigan Fish during
Spring, Summer, and Autumn 8-14
Figure 8-6. Percentage of Lake Michigan Coho Salmon and Lake Trout Samples within
each PCB Fish Advisory Category 8-20
Figure 9-1. Concentrations of Total PCBs in the Atmosphere, Tributaries, Water Column,
and Sediments of Lake Michigan 9-3
Figure 9-2. Concentrations of fra«s-Nonachlor in the Atmosphere, Tributaries, Water Column,
and Sediments of Lake Michigan 9-4
Figure 9-3. Lake Michigan Bathymetry 9-6
Figure 9-4. Total PCB and fra«s-Nonachlor Concentrations in Various Components of
the Lake Michigan Ecosystem 9-7
Figure 9-5. Biomagnification Factors for Total PCBs (BMFP) and fra«s-Nonachlor (BMFT)
in a Simplified Lake Michigan Food Web 9-9
Figure 9-6. Octanol-water Partition Coefficients (Kow) for PCB Congeners 9-10
Figure 9-7. Bioaccumulation Factors in the Lake Michigan Food Web versus
Log Octanol-Water Partition Coefficients for Individual PCB Congeners 9-11
Figure 9-8. Biomagnification Factors in the Lake Michigan Food Web versus
Log Octanol-Water Partition Coefficients for Individual PCB Congeners 9-12
Figure 9-9. Concentration of Toxic PCB Congeners in the Lake Michigan Ecosystem 9-13
Figure 9-10. Bioaccumulation Factors (BAFs) for Toxic PCB Congeners and Total PCBs in
Lake Michigan Biota 9-14
xvi April 2004
-------
Executive Summary
The U.S. Environmental Protection Agency's Great Lakes National Program Office (GLNPO) and its
partners instituted the Lake Michigan Mass Balance (LMMB) Study to measure and model the
concentrations of representative pollutants within important compartments of the Lake Michigan
ecosystem. The goal of the LMMB Study was to develop a sound, scientific base of information to guide
future toxic load reduction efforts at the Federal, State, Tribal, and local levels. Objectives of the study
were to:
1. Estimate pollutant loading rates,
2. Establish a baseline to gauge future progress,
3. Predict the benefits associated with load reductions, and
4. Further understand ecosystem dynamics.
The LMMB Study measured the concentrations of mercury, poly chlorinated biphenyls (PCBs), trans-
nonachlor, and atrazine in the atmosphere, tributaries, lake water, sediments, and food webs of Lake
Michigan. This document summarizes the PCB and fra«s-nonachlor data collected as part of the LMMB
Study, and is one in a series of data reports that documents the project.
PCBs are synthetic organic chemicals that are chemically inert, nonflammable, and do not transmit
electrical current. These properties, combined with high melting and boiling points, made PCBs useful in
a wide variety of industrial applications, particularly as dielectric fluids in electrical transformers and
capacitors. In the United States, the Monsanto Company produced commercial mixtures of PCBs by
chlorinating biphenyl and sold the mixtures under the trade name Aroclor. The parent compound,
biphenyl, consists of two six-membered aromatic carbon rings joined by a single carbon-carbon bond.
Chlorination of biphenyl attaches one or more chlorine atoms to carbon atoms in the two ring structures.
PCB production and export in the U.S. was halted in October 1977 under the auspices of the Toxic
Substances Control Act (TSCA).
There are 209 possible arrangements of chlorine atoms, and each of the arrangements is referred to as a
PCB "congener." The formal chemical name of each congener identifies the specific positions and the
total number of chlorine atoms in the congener. However, individual PCB congeners are often referred to
simply by "congener number," e.g., PCB 1 to PCB 209. This report focuses on the results for three
congeners, PCB 33, PCB 118, and PCB 180 in all of the samples except sediments. For sediments
(Chapter 6), the report focuses on the results for the sum of PCBs 28 and 31, PCB 118, and PCB 180.
fra«5-Nonachlor is the common name for 1,2,3,4,5,6,7,8,8-nonachloro-3a,4,7,7a-tetrahydro-4,7-
methanoindan, a member of the class of cyclodiene pesticides. fra«s-Nonachlor was not produced as a
pure compound, but was a major component of the pesticide "technical chlordane," which is a mixture of
at least 140 related compounds. Chlordane mixtures were first produced in the U.S. in 1948 and various
formulations of chlordane were widely used as pesticides on food crops and lawns, and for termite control
from 1948 to 1988. In April 1988, EPA canceled all commercial uses of chlordane in the U.S.
Study Design
In the LMMB Study, PCBs and fra«s-nonachlor were measured in atmospheric, tributary, open-lake
water column samples, sediments, lower pelagic food web organisms, and fish. From March 1994
through October 1995, over 1000 samples were collected from locations in and around Lake Michigan
(see Figure 1-2 in Chapter 1) and analyzed by gas chromatography with either electron capture or mass
spectrometry detectors. Atmospheric vapor, particulate, and precipitation samples were collected from
eight stations surrounding Lake Michigan and three background stations outside the Lake Michigan basin.
Tributary samples were collected from 11 rivers that flow into Lake Michigan. Open-lake water column
April 2004 ES-1
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
samples were collected from 38 sampling stations in Lake Michigan, 2 stations in Green Bay, and 1
station in Lake Huron. Sediment samples were collected from over 100 stations in Lake Michigan and
Green Bay. Samples of participate matter were collected in sediment traps deployed at eight stations in
Lake Michigan. Samples of phytoplankton and zooplankton were collected from 12 stations in Lake
Michigan. Samples of My sis and Diporeict were collected at 10 stations in Lake Michigan. Specimens of
lake trout, coho salmon, bloater chub, alewife, smelt, deepwater sculpin, and slimy sculpin were collected
from three to 79 locations around the lake, depending on the species.
PCBs and fraws-Nonachlor in Atmospheric Components
Vapor-phase PCB congeners were detected in the vast majority of the samples collected from all LMMB
Study stations. Monthly composite concentrations of vapor-phase total PCBs ranged from 0 pg/m3 at
Beaver Island and Brule River stations to 6300 pg/m3 at the IIT Chicago station. Vapor-phase PCB
results exhibited a seasonal trend, with higher concentrations occurring in summer months and lower
concentrations occurring in winter months and may be a result of the interaction of the vapor pressures of
the PCBs and the increased temperatures during summer months. Vapor-phase PCB congener and total
PCB concentrations varied by sampling station. Urban and urban-influenced sites had higher mean
monthly composite concentrations for the duration of the study period than rural sites.
Vapor-phase fra«s-nonachlor was detected much less frequently than PCB congeners. Vapor-phase
fra«5-nonachlor was not detected in the samples from two over-water stations. Of the 28 sampling
stations, 13 stations had 13% to 50% of the individual samples below detection limits. Only one sample
was collected at Stations 380 and 19M and each had a result of zero. Concentrations of vapor-phase
fra«5-nonachlor ranged from 0 pg/m3 at over-water stations 380 and 19M to 118 pg/m3 at Bondville.
Non-zero mean monthly composite concentrations of fra«s-nonachlor for each sampling station ranged
from 2.1 pg/m3 measured at Brule River to 43 pg/m3 measured at Bondville. Vapor-phase trans-
nonachlor results showed an even stronger seasonal variation than the vapor-phase PCB results. All of
the sites exhibited similar trends in vapor phase fra«s-nonachlor concentrations, with higher
concentrations generally occurring in the summer and lower concentrations in the winter, despite
differences between sites of an order of magnitude or more. For the urban site IIT Chicago,
concentrations of vapor-phase fra«s-nonachlor were 1.5 pg/m3 in February 1995 and were 50 times higher
in July 1995, at 80 pg/m3. For the rural Bondville station, vapor-phase fra«s-nonachlor was 2 pg/m3 in
February 1995 and was 60 times higher in July 1995 at 120 pg/m3.
Particulate-phase PCB congeners were detected in the majority of the samples collected from all LMMB
Study stations. Concentrations of particulate-phase total PCBs ranged from 0 pg/m3 at the Beaver Island
station to 250 pg/m3 at the IIT Chicago station. Particulate-phase PCB congener and total PCB
concentrations varied by sampling station. Urban and urban-influenced sites had higher mean monthly
composite concentrations for the duration of the study period than rural sites, consistent with the
hypothesis that urban and urban-influenced areas contain significant sources of particulate-phase PCBs.
Particulate-phase fra«s-nonachlor was detected much less frequently than PCB congeners in the samples.
Except for the samples collected at the Empire Michigan station, fra«s-nonachlor was reported as being
below the sample-specific detection limit in 20 - 100% of the particulate-phase samples from the other 16
stations. Concentrations of particulate-phase fra«s-nonachlor ranged from 0 pg/m3 at 12 stations to 2.6
pg/m3 at Bondville. Mean monthly composite concentrations of fra«s-nonachlor for each sampling
station ranged from 0.16 pg/m3 measured at GB24M to 1.2 pg/m3 measured at IIT Chicago, with a
concentration of 1.8 pg/m3 for the only sample collected at over-water Station 1.
PCB congeners were detected in many of the precipitation samples collected from the LMMB Study
stations. However, the overall frequency of occurrence of PCBs in the precipitation samples was lower
ES-2 April 2004
-------
Executive Summary
than for the vapor-phase and particulate-phase samples. For total PCBs, the mean concentrations in
precipitation ranged from 290 pg/L at the Eagle Harbor station to 16,000 pg/L at the IIT Chicago station.
fra«5-Nonachlor was detected even less frequently than the PCB congeners in the precipitation samples.
Except for the samples collected at the IIT Chicago station, trans-nonachlor was reported as being below
the sample-specific detection limit in 75 to 100% of the precipitation samples from all stations. The
concentrations of trans-nonachlor in the precipitation samples ranged from 0 pg/L at every site to a high
of 630 pg/L at the Chiwaukee Prairie site.
PCBs and fraws-Nonachlor in Tributaries
The dissolved total PCB concentrations ranged from not detected in four tributaries to 48 ng/L in the
Grand Calumet, while particulate total PCB concentrations ranged from not detected in four tributaries to
120 ng/L in the Sheboygan River. Mean dissolved total PCB concentrations ranged from 0.43 ng/L in the
Pere Marquette River to 35 ng/L in the Grand Calumet, while mean particulate concentration ranged from
0.25 ng/L in the Muskegon River to 55 ng/L in the Sheboygan River.
The concentrations of dissolved and particulate total PCBs exhibited a seasonal trend for many of the
tributaries, with higher mean concentrations occurring in summer months and lower mean concentrations
occurring in winter months. There were significant differences between seasons for the dissolved total
PCB concentrations in nine of the eleven tributaries, and significant differences between season for the
particulate total PCB concentration in six of the eleven tributaries. However, the trend was not consistent
across all of the tributaries. The mean seasonal concentrations of dissolved and particulate total PCBs
across all 11 tributaries span at least two orders of magnitude.
Concentrations of dissolved trans-nonachlor ranged from not detected in seven tributaries to 0.19 ng/L in
the Manistique River, while particulate fra«s-nonachlor ranged from not detected in five tributaries to
0.38 ng/L in the Manistique River. Mean dissolved trans-nonachlor concentrations ranged 0.0033 ng/L
in the Menominee River to 0.026 ng/L in the St. Joseph River, while mean particulate trans-nonachlor
concentrations ranged from 0.0028 ng/L in the Menominee River to 0.074 ng/L in the St. Joseph River.
The mean concentrations of dissolved and particulate trans-nonachlor show fewer significant differences
than the total PCB results. Eight of the eleven tributaries exhibit no statistically significant differences in
mean dissolved trans-nonachlor concentrations among the seasons. Of the other three tributaries, the
mean dissolved trans-nonachlor in the Kalamazoo River is never the lowest in spring or summer, and
never the highest in autumn, while in the Sheboygan River, mean dissolved trans-nonachlor is never the
lowest in the summer, or the highest in the winter. The dissolved trans-nonachlor results for the
Manistique River are characterized by a very high mean concentration in the winter which is significantly
different from the other three seasons, which in turn, are not significantly different from one another. The
very high winter mean concentration is repeated in the particulate trans-nonachlor results in this tributary.
PCBs and fraws-Nonachlor in Open-lake Water
The concentrations of dissolved PCB congeners are generally lowest in the far northern areas of the lake
that are removed from urban influences. The highest dissolved concentrations generally are found in the
southwest area of the lake, near the urban areas of Chicago and Milwaukee.
The particulate PCB concentrations are highest in Green Bay, at Station GB 17, with much lower
particulate PCB concentrations in the remainder of the lake. The particulate concentrations of PCBs 118
and 180 show a slight increase in the southeast portion of the lake, in the area between the mouths of the
April 2004 ES-3
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
St. Joseph and Kalamazoo Rivers. However, the concentrations of participate PCBs 118 and 180 in that
area are still 2 to 5 times lower than in the upper reaches of Green Bay.
The dissolved concentrations of fra«s-nonachlor are similar to those of the dissolved PCB congeners,
with an apparent increase in concentration in the southwest portion of the lake, near Chicago. The
particulate concentrations of fra«s-nonachlor are similar to those of the particulate PCB congeners, with
the highest concentrations in Green Bay, near Station GB 17. However, particulate fra«s-nonachlor
concentrations appear to increase in areas of the lake adjacent to most of the major urban areas around the
lake. Similar increases occur near the discharges of the Manistique and Pere Marquette Rivers, which are
not associated with urban areas, suggesting that the increases near the urban area may be a function of
river-borne sources of particulate fra«s-nonachlor, including resuspension of contaminated sediments.
PCBs and fraws-Nonachlor in Sediments
The mean concentrations of the PCB congeners and total PCBs exhibit a general trend of decreasing
concentrations from south to north, with the lowest concentrations in the Straits region. In contrast, the
mean concentrations of trans-nonachlor, while lowest in the Straits region, exhibit no south to north
trend. The organic carbon content data exhibit a pattern similar to that for fra«s-nonachlor.
Total PCBs exhibit a wide range in concentration. Although not a true bimodal distribution, two distinct
groups are evident within the total PCB distribution. The first group of samples, from nondepositional
and transitional stations, has very low concentrations that exponentially decline in number with increasing
concentration (0 - 30 ng/g). The second group of samples taken mainly from depositional sites (35 - 225
ng/g) is more normally distributed, though tailing toward higher levels is evident.
Concentrations of PCBs and trans-nonachlor in surficial sediments increase with increasing organic
carbon (OC) content. The spread in the data suggest that the southern basin stations were significantly
higher in contamination. This finding was expected as hydrophobic compounds are known to partition
strongly to organic matter. The southern basin the relationships of the PCB congeners and organic carbon
content exhibited an additional feature not seen in other basins. A peak in PCB concentration was
observed at -25 mg/g OC for PCB 28+31 and at -35 mg/g for PCB 118 and PCB 180. The relationship
of fra«5-nonachlor with OC did not follow the same pattern as any particular PCB congener.
PCBs and fraws-Nonachlor in Lower Pelagic Food Web Organisms
PCB and trans-nonachlor concentrations measured in the lower pelagic food web differed significantly
among phytoplankton, zooplankton, Mysis spp., and Diporeia spp. Concentrations of total PCBs and
trans-nonachlor were highest in samples of Diporeia spp., followed by Mysis spp., zooplankton, and
phytoplankton, respectively. Total PCB concentrations were 9 times higher in Diporeia spp. than in
phytoplankton, averaging 420, 250, 170, and 49 ng/g dry weight in Diporeia spp., Mysis spp.,
zooplankton, and phytoplankton samples, respectively. Trans-Nonachlor concentrations were 19 times
higher in Diporeia spp. than in phytoplankton, averaging 32, 25, 16, and 1.7 ng/g dry weight in Diporeia
spp., Mysis spp., zooplankton, and phytoplankton samples, respectively.
A portion of the difference in PCB and fra«s-nonachlor concentrations among lower pelagic food web
sample types is likely due to variations in the lipid content of the samples. Hydrophobic organic
contaminants such as PCBs and fra«s-nonachlor preferentially concentrate in the fatty tissues of
organisms, so those organisms with higher lipid content will likely concentrate more of these
contaminants. The differences in lipid content among the sample types, however, explained only a
quarter to less than half of the variability in total PCB and fra«s-nonachlor concentrations. Even when
total PCB and fra«s-nonachlor concentrations were normalized by lipid content, the trends in PCB and
ES-4 April 2004
-------
Executive Summary
trans-nonachlor concentrations among the sample types were almost always the same. Normalized total
PCB and fra«s-nonachlor concentrations in Diporeia spp. and Mysis spp. were significantly higher than in
zooplankton and phytoplankton, and normalized fra«s-nonachlor concentrations in zooplankton were
significantly higher than in phytoplankton. Normalized total PCB concentrations in zooplankton,
however, were not significantly different than in phytoplankton.
PCBs and fraws-Nonachlor in Fish
PCB and fra«s-nonachlor concentrations differed significantly among species. Significantly higher levels
of total PCBs and fra«s-nonachlor were observed in Lake trout, a top predator in the Lake Michigan
pelagic food web, than in any other fish species. Mean wet-weight concentrations of total PCBs and
fra«5-nonachlor in lake trout were 3.6 and 2.9 times higher than for any other species. This trend was
similar for dry-weight basis PCB and fra«s-nonachlor concentrations. Mean dry-weight basis total PCB
concentrations in lake trout were from 1.2 to 16 times higher than in other species, and mean dry-weight
basis fra«5-nonachlor concentrations were 2.4 to 34 times higher in lake trout than in other species.
When PCB and fra«s-nonachlor concentrations were compared among fish species on a lipid-normalized
basis, lake trout still contained higher levels of contamination than all other species with the exception of
adult coho salmon. Mean lipid-normalized total PCB and fra«s-nonachlor concentrations were highest in
adult coho salmon and second highest in lake trout. Lipid-normalized total PCB and fra«s-nonachlor
concentrations in these two top predator fish species were significantly higher than in any of the forage
fish species. The higher mean concentrations of lipid-normalized contaminants in adult coho salmon
were due to the relatively low lipid content in this species. Lipid content in adult coho salmon averaged
only 4%, compared to 16% in lake trout. Of the species analyzed in this study, only smelt contained
lower lipid content (3.6%) than adult coho salmon.
The lowest total PCB and fra«s-nonachlor concentrations on a wet-weight, dry-weight, or lipid-weight
basis were consistently found in hatchery and yearling coho salmon. This species is raised in hatcheries
and annually stocked in Lake Michigan. Hatchery samples consisted of immature coho collected directly
from the Platte River hatchery, and yearling samples consisted of immature coho collected in Lake
Michigan. The reduced contamination in these sample types most likely reflects both the young age of
the fish and reduced contaminant exposure from hatchery food and water sources.
The Great Lakes Fish Consumption Advisory Task Force has set a fish advisory category of "no
consumption" at PCB levels above 2000 ng/g, and established four lesser consumption categories ranging
form unrestricted consumption to no more than 6 meals per year. Of the Lake Michigan fish analyzed in
the LMMB Study, only lake trout contained PCBs above the 2000 ng/g level. In fact, 56% of lake trout
samples exceeded this tolerance level, and the mean total PCB concentration for Lake Michigan lake trout
was 3000 ng/g (or 3 ppm), which is 50% above the 2000 ng/g tolerance level. No coho salmon or lake
trout samples fell into the unrestricted consumption category. Coho salmon primarily fell into the 1
meal/mo and 6 meals/yr categories. These categories contained 46% and 44% of coho salmon samples,
respectively, with only 9% of coho salmon samples falling into the 1 meal/wk category. Lake trout
primarily fell into the no consumption category (56%), with only 0.4%, 17%, and 26% in the 1 meal/wk,
1 meal/mo, and 6 meals/yr categories, respectively.
Mass Balance and Modeling Efforts
The data collection and quality assurance efforts described in this report were designed to support the
LMMB Study and related efforts to model the concentrations of pollutants in the Lake Michigan
ecosystem. However, the mass balance itself and the associated modeling efforts are beyond the scope of
this data report but will be described in later documents from GLNPO.
April 2004 ES-5
-------
Chapter 1
Project Overview
The U.S. Environmental Protection Agency's Great Lakes National Program Office (GLNPO) and its
partners instituted the Lake Michigan Mass Balance (LMMB) Study to measure and model the
concentrations of representative pollutants within important compartments of the Lake Michigan
ecosystem. Concentrations of poly chlorinated biphenyls (PCBs), fra«s-nonachlor, atrazine, and mercury
in the atmosphere, tributaries, lake water, sediments, and food webs of Lake Michigan. This document
summarizes the PCB and fra«s-nonachlor data collected as part of the LMMB Study.
1.1 Background
The Great Lakes, which contain 20% of the world's freshwater, are a globally important natural resource
that are currently threatened by multiple stressors. While significant progress has been made to improve
the quality of the lakes, pollutant loads from point, non-point, atmospheric, and legacy sources continue
to impair ecosystem functions and limit the attainability of designated uses of these resources. Fish
consumption advisories and beach closings continue to be issued, emphasizing the human health concerns
from lake contamination. Physical and biological stressors such as invasion of non-native species and
habitat loss also continue to threaten the biological diversity and integrity of the Great Lakes.
The United States and Canada have recognized the significance and importance of the Great Lakes as a
natural resource and have taken steps to restore and protect the lakes. In 1978, both countries signed the
Great Lakes Water Quality Agreement (GLWQA). This agreement calls for the restoration and
maintenance of the chemical, physical, and biological integrity of the Great Lakes by developing plans to
monitor and limit pollutant flows into the lakes.
The GLWQA, as well as Section 118(c) of the Clean Water Act, required the development of Lake-wide
Management Plans (LaMPs) for each Great Lake. The purpose of these LaMPs is to document an
approach to reducing inputs of critical pollutants to the Great Lakes and restoring and maintaining Great
Lakes integrity. To assist in developing these LaMPs and to monitor progress in pollutant reduction,
Federal, State, Tribal, and local entities have instituted Enhanced Monitoring Plans. Monitoring is
essential to the development of baseline conditions for the Great Lakes and provides a sound scientific
base of information to guide future toxic load reduction efforts.
The LMMB Study is a part of the Enhanced Monitoring Plan for Lake Michigan. The LMMB Study was
a coordinated effort among Federal, State, and academic scientists to monitor tributary and atmospheric
pollutant loads, develop source inventories of toxic substances, and evaluate the fates and effects of these
pollutants in Lake Michigan. A mass balance modeling approach provides the predictive ability to
determine the environmental benefits of specific load reduction scenarios for toxic substances and the
time required to realize those benefits. This predictive ability will allow Federal, State, Tribal, and local
agencies to make more informed load reduction decisions.
1.2 Description
The LMMB Study used a mass balance approach to evaluate the sources, transport, and fate of
contaminants in the Lake Michigan ecosystem. A mass balance approach is based on the law of
conservation of mass, which states that the amount of a pollutant accumulating in a system is equal to the
amount entering the system, less the amount of that pollutant leaving or chemically changed in the system
(Figure 1-1).
April 2004 1-1
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Mnssi,
If the system is defined as the Lake Michigan/Green
Bay water column, then pollutants may enter the
system via tributaries, direct runoff, the atmosphere
(wet deposition, dry deposition, and sorption from the
vapor phase), the sediment, and the Straits of
Mackinac. Pollutants may leave the system through
volatilization to the atmosphere, loss to the sediment,
or discharge through the Straits of Mackinac and the
Chicago water diversion. The law of conservation of
mass also can be applied to other systems such as
biota, sediment, or air.
The LMMB Study measured contaminant
concentrations in various inputs and ecosystem
compartments over spatial and temporal scales.
Mathematical models that track the transport and fate
of contaminants within Lake Michigan are being
developed and calibrated using these field data. The
LMMB Study is the first lake-wide application of a
mass balance determination for toxics in the Great
Lakes and will serve as the basis of future mass budget/mass balance efforts.
What is mass balance?
System
Mass,
'SlOITll
Ma-iXj,, = MasStrnqfonaed
Maxxollt
Figure 1-1. Simplified Mass Balance Approach
1.3 Scope
1.3.1 Modeled Pollutants
When EPA published the Water Quality Guidance for the Great Lakes System (58 FR 20802), the Agency
established water quality criteria for 29 pollutants. Those criteria are designed to protect aquatic life,
terrestrial wildlife, and human health. PCBs, trans-nonachlor, and mercury are included in the list of 29
pollutants. The water quality criteria and values proposed in the guidance apply to all of the ambient
waters of the Great Lakes system, regardless of the sources of pollutants in those waters. The proposed
criteria provide a uniform basis for integrating Federal, State, and Tribal efforts to protect and restore the
Great Lakes ecosystem.
The number of pollutants that can be intensively monitored and modeled in the Great Lakes system is
limited by the resources available to collect and analyze thousands of samples, assure the quality of the
results, manage the data, and develop and calibrate the necessary models. Therefore, the LMMB Study
focused on constructing mass balance models for a limited group of pollutants. PCBs, fra«s-nonachlor,
atrazine, and mercury were selected for inclusion in the LMMB Study because these pollutants currently
or potentially pose a risk to aquatic and terrestrial organisms (including humans) in the Lake Michigan
ecosystem. These pollutants also were selected to cover a wide range of chemical and physical properties
and represent other classes of compounds which pose current or potential problems. Once a mass budget
for selected pollutants is established and a mass balance model calibrated, additional contaminants can be
modeled with limited data and future resources can be devoted to activities such as emission inventories
and dispersion modeling.
1-2
April 2004
-------
Project Overview
1.3.1.1 Polychlorinated Biphenyls
PCBs are a class of man-made, chlorinated, organic chemicals that include 209 congeners, or specific
PCB compounds. The highly stable, nonflammable, non-conductive properties of these compounds have
made them useful in a variety of products including electrical transformers and capacitors, plastics,
rubber, paints, adhesives, and sealants. PCBs were produced for such industrial uses in the form of
complex mixtures under the trade name "Aroclor" and were commercially available from 1930 through
1977, when EPA banned their production due to environmental and public health concerns. PCBs also
may be produced by combustion processes, including incineration, and can be found in stack emissions
and ash from incinerators.
Seven Aroclor formulations were included in the Priority Pollutant List developed by the EPA Office of
Water under the auspices of the Clean Water Act because they were found by EPA in the effluents from
one or more wastewater treatment facilities. Aroclors may have entered the Great Lakes through other
means, including spills or improper disposal of transformer fluids, contaminated soils washing into the
watershed, or discharges from ships. The PCBs produced by combustion processes may be released to
the atmosphere, where they are transported in both vapor and particulate phases and enter the lakes
through either dry deposition or precipitation events (e.g., rain).
The stability and persistence of PCBs, which made them useful in industrial applications, have also made
these compounds ubiquitous in the environment. PCBs do not readily degrade and thus accumulate in
water bodies and aquatic sediments. PCBs also bioaccumulate, or buildup, in living tissues. Levels of
PCBs in some fish from Lake Michigan exceed U.S. Food and Drug Administration tolerances,
prompting closure of some commercial fisheries and issuance offish consumption advisories. PCBs are a
probable human carcinogen, and human health effects of PCB exposure include stomach, kidney, and
liver damage, liver and biliary tract cancer, and reproductive effects, including effects on the fetus after
exposure of the mother.
PCB congeners exhibit a wide range of physical and chemical properties (e.g., vapor pressures,
solubilities, boiling points), are relatively resistant to degradation, and are ubiquitous. These properties
make them ideal surrogates for a wide range of organic compounds from anthropogenic sources.
In the LMMB Study, PCBs were selected as a model for conservative organic compounds (USEPA,
1997a).
13.12 trans-Nonachlor
fra«5-Nonachlor is a component of the pesticide chlordane. Chlordane is a mixture of chlorinated
hydrocarbons that was manufactured and used as a pesticide from 1948 to 1988. Prior to 1983,
approximately 3.6 million pounds of chlordane were used annually in the U.S. In 1988, EPA banned all
production and use of chlordane in the U.S.
Like PCBs, chlordane is relatively persistent and bioaccumulative. fra«s-Nonachlor is the most
bioaccumulative of the chlordanes. fra«s-Nonachlor is a probable human carcinogen. Other human
health effects include neurological effects, blood dyscrasia, hepatoxicity, immunotoxicity, and endocrine
system disruption.
Historically, trans-nonachlor may have entered the Great Lakes through a variety of means related to the
application of chlordane, including improper or indiscriminate application, improper cleaning and
disposal of pesticide application equipment, or contaminated soils washing into the watershed.
April 2004 1-3
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
In the LMMB Study, trans-nonachlor was selected as a model for the cyclodiene pesticides (USEPA,
1997a).
1.3.1.3 Atrazine
Atrazine is a herbicide based on a triazine ring structure with three carbon atoms alternating with three
nitrogen atoms. Atrazine is the most widely used herbicide in the U.S. for corn and sorghum production.
Atrazine has been used as an agricultural herbicide since 1959 and 64 to 75 million pounds of atrazine are
used annually in the U.S. Atrazine is extensively used in the upper Midwest, including the Lake Michigan
watershed, where it is primarily associated with corn crops.
Unlike PCBs and trans-nonachlor, atrazine is not extremely persistent or bioaccumulative. Atrazine is
moderately susceptible to biodegradation, with a half-life in soils of about 60 - 150 days. Atrazine may
persist considerably longer in water and is relatively non-reactive in the atmosphere. Atrazine rarely
exceeds the maximum contaminant level (MCL) set by USEPA as a drinking water standard, but
localized peak values can exceed the MCL following rainfall events after atrazine application. Atrazine
can cause human health effects such as weight loss, cardiovascular damage, muscle and adrenal
degeneration, and congestion of heart, lungs, and kidneys. Atrazine is also toxic to aquatic plants.
In the LMMB Study, atrazine was selected as a model for reactive, biodegradable compounds in current
use (USEPA, 1997A).
1.3.1.4 Mercury
Mercury is a naturally-occurring toxic metal. Mercury is used in battery cells, barometers, thermometers,
switches, fluorescent lamps, and as a catalyst in the oxidation of organic compounds. Global releases of
mercury in the environment are both natural and anthropogenic (caused by human activity). It is
estimated that about 5,500 metric tons of mercury are released annually to the air, soil, and water from
anthropogenic and natural sources (USEPA 1997b). These sources include combustion of various fuels
such as coal; mining, smelting and manufacturing activities; wastewater; agricultural, animal and food
wastes; chlor-alkali plants; and pulp and paper mills.
As an elemental metal, mercury is extremely persistent in all media. Mercury also bioaccumulates with
reported bioconcentration factors in fish tissues in the range of 63,000 to 100,000. Mercury is a
neurotoxin and possible human carcinogen and causes the following human health effects: stomach, large
intestine, brain, lung, and kidney damage; blood pressure and heart rate increase, and fetus damage.
In the LMMB Study, mercury was selected as a model for bioaccumulative metals (USEPA, 1997a).
1-4 April 2004
-------
Project Overview
Table 1-1. Characteristics of Lake Michigan Mass Balance Modeled Pollutants
Pollutant
PCBs
trans-
Nona-
chlor3
Atrazine
Mercury
Sources
• Waste incinerators
(unintentional
byproducts of
combustion)
• Industrial
dischargers
• Electrical power
• Application to
crops and gardens
• Application to
crops
• Waste disposal
• Manufacturing
processes
• Energy production
• Ore processing
• Municipal and
medical waste
incinerators
• Chlor-alkali
factories
• Fuel combustion
Uses
• Electrical
transformers and
capacitors
• Carbonless copy
paper
• Plasticizers
• Hydraulic fluids
• Pesticide on corn
and citrus crops
• Pesticide on lawns
and gardens
• Herbicide for corn
and sorghum
production
• Battery cells
• Barometers
• Dental fillings
• Thermometers
• Switches
• Fluorescent lamps
Toxic Effects
• Probable human carcinogen
• Hearing and vision
impairment
• Liver function alterations
• Reproductive impairment and
deformities in fish and wildlife
• Probable human carcinogen
• Nervous system effects
• Blood system effects
• Liver, kidney, heart, lung,
spleen, and adrenal gland
damage
• Weightless
• Cardiovascular damage
• Muscle and adrenal
degeneration
• Congestion of heart, lungs,
and kidneys
• Toxic to aquatic plants
• Possible human carcinogen
• Damage to brain and kidneys
• Adverse affects on the
developing fetus, sperm, and
male reproductive organs
Biocon-
centration
Factor1
1,800 to
180,000
4,000 to 40,000
2 to 100
63,000 to
100,000
EPA
Regulatory
Standards2
MCL = 0.5 |jg/L
CCC = 14ng/L
HH = 0.17ng/L
MCL = 2 |jg/L
CMC = 2.4 |jg/L
CCC = 4.3 ng/L
HH = 2.1 ng/L
MCL = 3 M9/L
CMC4 = 350 |jg/L
CCC4 = 12|jg/L
MCL = 2 |jg/L
CMC = 1.4|jg/L
CCC = 0.77 |jg/L
HH = 50 ng/L
FWA5 = 2.4 |jg/L
FWC5=12ng/L
Wildlife6 = 1.3 ng/L
1 From: USEPA. 1995a. National Primary Drinking Water Regulations, Contaminant Specific Fact Sheets, Inorganic Chemicals, Technical
Version. EPA811/F-95/002-T. U.S. Environmental Protection Agency, Office of Water, Washington, D.C.; and USEPA. 1995b. National
Primary Drinking Water Regulations, Contaminant Specific Fact Sheets, Synthetic Organic Chemicals, Technical Version. EPA 811/F-
95/003-T. U.S. Environmental Protection Agency, Office of Water, Washington, DC.
2 MCL = Maximum Contaminant Level for drinking water. CMC = Criterion Maximum Concentration for protection of aquatic life from acute
toxicity. CCC = Criterion Continuous Concentration for protection of aquatic life from chronic toxicity. HH = Water quality criteria for
protection of human health from water and fish consumption. Data from: USEPA. 1999. National Recommended Water Quality Criteria-
Correction. EPA 822/Z-99/001. U.S. Environmental Protection Agency, Office of Water, Washington, DC.
3 Characteristics presented are for chlordane. frans-Nonachlor is a principle component of the pesticide chlordane.
4 Draft water quality criteria for protection of aquatic life. From: USEPA. 2001a. Ambient Aquatic Life Water Quality Criteria for Atrazine.
U.S. Environmental Protection Agency, Office of Water, Washington, DC.
5 FWA = Freshwater acute water quality criterion. FWC = Freshwater chronic water quality criterion. From National Toxics Rule (58 FR
60848).
6 Wildlife criterion. From the Stay of Federal Water Quality Criteria for Metals (60 FR 22208), 40 CFR131.36 and the Water Quality Guidance
for the Great Lakes System (40 CFR 132).
April 2004
1-5
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
1.3.2 Other Measured Parameters
In addition to the four chemicals modeled in the LMMB Study, many other chemicals and parameters
were measured in the LMMB Study as part of the Enhanced Monitoring Program. A survey of these
chemicals and parameters will aid in understanding the overall ecological integrity of Lake Michigan.
These additional parameters include various biological indicators, meteorological parameters, and
organic, metal, and conventional chemicals in Lake Michigan. A complete listing of all parameters
included in this study is provided in Table 1-2.
Table 1-2. Lake Michigan Mass Balance Study Parameters
Organics
acenaphthene
acenaphthylene
aldrin
anthracene
atrazine
a-BHC
(3-BHC
5-BHC
Y-BHC (Lindane)
benzo [a] anthracene
benzo [g,h,i\ perylene
benzo [b] fluoranthene
benzo [k\ fluoranthene
benzo [e] pyrene
benzo [a] pyrene
a-chlordane
y-chlordane
chrysene
coronene
p,p'-DDE
p,p'-DDD
p,p'-DDT
endosulfan sulfate
endosulfan I
endosulfan II
endrin
endrin aldehyde
endrin ketone
fluoranthene
fluorene
heptachlor
heptachlor epoxide
hexachlorobenzene (HCB)
indeno[1,2,3-cd] pyrene
mirex
frans-nonachlor
oxychlordane
PCB congeners
phenanthrene
pyrene
retene
toxaphene
Metals
aluminum
arsenic
calcium
cadmium
chromium
cesium
copper
iron
mercury
potassium
magnesium
manganese
sodium
nickel
lead
selenium
thorium
titanium
vanadium
zinc
1-6
April 2004
-------
Project Overview
Table 1-2. Lake Michigan Mass Balance Study Parameters
Conventionals
alkalinity
ammonia
bromine
chloride
chlorine sulfate
conductivity
dissolved organic carbon
dissolved oxygen
dissolved phosphorous
dissolved reactive silica
dry weight fraction
elemental carbon
nitrate
ort/io-phosphorous
particulate organic carbon
percent moisture
PH
phosphorous
silica
silicon
temperature
total Kjeldahl nitrogen
total organic carbon
total phosphorous
total suspended particulates
total hardness
turbidity
Biologicals
fish species
fish age
fish maturity
chlorophyll a
fish lipid amount
fish weight
fish length
fish taxonomy
fish diet analysis
primary productivity
Meteorological
air temperature
relative humidity
barometric pressure
weather conditions
wind direction
wind speed
visibility
wave height and direction
1.3.3 Measured Compartments
In the LMMB Study, contaminants were measured in the following compartments:
Open-Lake Water Column - The water column in the open lake was sampled and analyzed for the
modeled pollutants.
• Tributaries - Tributary water columns were sampled and analyzed for the modeled pollutants.
• Fish - Top predators and forage-base species were sampled and analyzed for diet analysis and
contaminant burden. Fish were not analyzed for atrazine because atrazine is not bioaccumulative.
Lower Pelagic Food Web - Lower pelagic food web organisms were sampled and analyzed for
species diversity, taxonomy, and contaminant burden. Individual samples of the lower pelagic food
web included mixed phytoplankton, mixed zooplankton, Diporeia spp., and My sis spp. The lower
pelagic food web was not analyzed for atrazine because atrazine is not bioaccumulative.
• Sediments - Cores were collected and trap devices were used to collect sediment for determination of
contaminants and sedimentation rates. Sediments were not analyzed for atrazine because atrazine is
relatively water soluble, degradable, and does not generally accumulate in sediments.
• Atmosphere - Vapor-, particulate-, and precipitation-phase samples were collected and analyzed for
the modeled pollutants.
April 2004
1-7
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
For the modeled pollutants, more than 20,000 samples were collected and analyzed, including more than
9000 quality control (QC) samples, at more than 300 sampling locations (Figure 1-2). Field data
collection activities were initially envisioned as a one-year effort. However, it became evident early into
the project that a longer collection period would be necessary to provide a full year of concurrent
information on contaminant loads and ambient concentrations for modeling purposes. Therefore, field
sampling occurred from April 1994 to October 1995.
Figure 1-2. Lake Michigan Mass Balance Study Sampling Locations
Beaver Island
;
Manitowoc
Sheboygan River
Milwaukee
Milwaukee River
Chiwaukee Prairie
NT Chicago
Chicago SWFP Intake
Grand Calumet Harbor
Pere Marquette
River
Muskegon River
Muskegon
Grand River
Kalamazoo River
South Haven
St. Joseph River
Benton Harbor
Indiana Dunes
—i
A Atmospheric Station
Q Tributary Station
Biota Station
• Sediment Station
^ Water Column Station
1-8
April 2004
-------
Project Overview
1.4 Objectives
The goal of the LMMB Study was to develop a sound, scientific base of information to guide future toxic
load reduction efforts at the Federal, State, Tribal, and local levels. To meet this goal, the four following
LMMB Study objectives were developed:
Estimate pollutant loading rates - Environmental sampling of major media will allow estimation of
relative loading rates of critical pollutants to the Lake Michigan Basin.
• Establish baseline - Environmental sampling and estimated loading rates will establish a baseline
against which future progress and contaminant reductions can be gauged.
Predict benefits associated with load reductions - The completed mass balance model will provide
a predictive tool that environmental decision-makers and managers may use to evaluate the benefits
of specific load reduction scenarios.
• Understand ecosystem dynamics - Information from the extensive LMMB monitoring and
modeling efforts will improve our scientific understanding of the environmental processes governing
contaminant cycling and availability within relatively closed ecosystems.
1.5 Design
1.5.1 Organization
The Great Lakes National Program Office proposed a mass balance approach to provide coherent,
ecosystem-based evaluation of toxics in Lake Michigan. GLNPO served as the program sponsor for the
LMMB Study. GLNPO formed two committees to coordinate study planning, the Program Steering
Committee and the Technical Coordinating Committee. These committees were comprised of scientists
from Federal, State, academic, and commercial institutions (see Section 1.5.2, Study Participants). The
committees administered a wide variety of tasks including: planning the project, locating the funding,
designing the sample collection, coordinating sample collection activities, locating qualified laboratories,
coordinating analytical activities, assembling the data, assuring the quality of the data, assembling skilled
modelers, developing the models, and communicating interim and final project results. The National
Health and Environmental Effects Research Laboratory (NHEERL)/Mid-Continent Ecology Division
(MED)/Large Lakes and Rivers Forecasting Research Branch (LLRFRB) at Gross lie, Michigan, in
cooperation with the National Oceanic and Atmospheric Administration (NOAA) Great Lakes
Environmental Research Laboratory (GLERL) and the Atmospheric Sciences Modeling Division are
supporting the modeling component of the mass balance study by developing a suite of integrated mass
balance models to simulate the transport, fate, and bioaccumulation of the study target analytes.
1.5.2 Study Participants
The LMMB Study was a coordinated effort among Federal, State, academic, and commercial institutions.
The following agencies and organizations have all played roles in ensuring the success of the LMMB
Study. Except for the three organizations indicated with an asterisk (*), all of the participants were
members of the LMMB steering committee.
April 2004 1-9
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Federal and International
*• USEPA Great Lakes National Program Office (Program Sponsor)
•> USEPA Region 5 Water Division
»• USEPA Region 5 Air Division
•> USEPA Office of Research and Development (ORD) NHEERL/MED/LLRFRB
*• USEPA Office of Research and Development National Exposure Research Laboratory
•> U.S. Department of Interior (USDOI) U.S. Geological Survey (USGS) Water Resources Division
*• USDOI USGS Biological Resources Division Great Lakes Science Center (GLSC)
•> U.S. Fish and Wildlife Service (USFWS)
*• U.S. Department of Commerce NOAA/GLERL
•> USEPA Office of Air and Radiation*
•> USEPA Office of Water*
»• U.S. Department of Energy, Battelle Northwest
*• Environment Canada*
State
*• Illinois Department of Natural Resources
»• Illinois Water Survey
*• Indiana Department of Environmental Management
*• Michigan Department of Environmental Quality (MDEQ)
*• Wisconsin Department of Natural Resources
»• Wisconsin State Lab of Hygiene
Academic and Commercial
*• Indiana University
*• Rutgers University
*• University of Maryland
*• University of Michigan
*• University of Minnesota
*• University of Wisconsin
*• Grace Analytical
1.5.3 Workgroups
Eleven workgroups were formed to provide oversight and management of specific project elements. The
workgroups facilitated planning and implementation of the study in a coordinated and systematic fashion.
The workgroups communicated regularly through participation in monthly conference calls and annual
"all-hands" meetings. Workgroup chairs were selected and were responsible for managing tasks under
the purview of the workgroup and communicating the status of activities to other workgroups. The
workgroups and workgroup chairs are listed below.
* Program Steering Committee - Paul Horvatin (USEPA/GLNPO)
*• Technical Coordinating Committee - Paul Horvatin (USEPA/GLNPO)
* Modeling Workgroup - William Richardson (USEPA/ORD/NHEERL/MED/LLRFRB)
> Air Monitoring Workgroup - Jackie Bode (USEPA/GLNPO)
•> Biota Workgroup - Paul Bertram (USEPA/GLNPO) and John Gannon (USDOI/USGS/GLSC)
•> Chemistry Workgroup - David Anderson (USEPA/GLNPO)
•> Data Management Workgroup - Kenneth Klewin and Philip Strobel (USEPA/GLNPO)
1-10 April 2004
-------
Project Overview
> Lake Monitoring Workgroup - Glenn Warren (USEPA/GLNPO)
»• Tributary Monitoring Workgroup - Gary Kohlhepp (USEPA Region 5 Water Division) and Robert
Day (Michigan Department of Environmental Quality)
»• Quality Assurance Workgroup - Louis Blume and Michael Papp (USEPA/GLNPO)
*• Sediment Monitoring Workgroup - Brian Eadie (NOAA/GLERL)
1.5.4 Information Management
As program sponsor, GLNPO managed information collected during the LMMB Study. Principal
investigators (Pis) participating in the study reported field and analytical data to GLNPO. GLNPO
developed a data standard for reporting field and analytical data and a database for storing and retrieving
study data. GLNPO also was responsible for conducting data verification activities and releasing verified
data to the study modelers and the public. The flow of information is illustrated in Figure 1-3.
15.4.? Data Reporting
More than twenty organizations produced LMMB data through the collection and analysis of more than
20,000 samples. In the interest of standardization, specific formats (i.e., file formats and codes to
represent certain data values) were established for reporting LMMB data. Each format specified the
"rules" by which data were submitted, and, in many cases, the allowable values by which they were to be
reported. The data reporting formats were designed to capture all pertinent sampling and analytical
information from the field crews and laboratory analysts. Data reporting formats and the resulting Great
Lakes Environmental Monitoring Database (GLENDA, see Section 1.5.4.2,) were designed to be
applicable to projects outside the LMMB as well. For the LMMB Study, special conditions were applied
for reporting analytical results. Because the data were being used for input to study models, principal
investigators were asked to report analytical results as measured, even when measurements were below
estimated detection limits. The quality assurance program discussed in Section 1.5.5 included identifying
(i.e., flagging) all analytical results that were below estimated detection limits.
Principal investigators (including sampling crews and the analytical laboratories) supplied sample
collection and analysis data following the standardized reporting formats if possible. LMMB data were
then processed through an automated SAS-based data verification system, the Research Data
Management and Quality Control System (RDMQ), for quality assurance/quality control checking. After
verification and validation by the PI, the data sets were output in a form specific for upload to GLENDA.
Finally, these data sets were uploaded to GLENDA.
April 2004 1-11
-------
Figure 1-3. Flow of Information in the Lake Michigan Mass Balance Study
Principal
Investigator (PI)
Collect and Analyze
Samples
Report Field and
Analytical Data
(According to LMMB Data
Standard)
GLNPO Data
Management
Workgroup
Receive, Store, and
Transmit Data
GLNPO QA
Workgroup
Conduct Data
Verification (Merge Field
and Analytical Data using
RDMQ)
Store, Transmit, and
Upload Data to
GLENDA
Produce final verified
data file and provide to
PI for review and
approval
Produce Final Verified
Data File and Transmit
(in GLENDA-compatible
Format)
LMMB Study
Modelers
Input Data to Study
Models
External Parties
1-12
-------
Project Overview
1.5.4.2 Great Lakes Environmental Monitoring Database
Central to the data management effort is a computerized database system to house LMMB Study and
other project results. That system, the Great Lakes Environmental Monitoring Database (GLENDA), was
developed to provide data entry, storage, access and analysis capabilities to meet the needs of mass
balance modelers and other potential users of Great Lakes data.
Development of GLENDA began in 1993 with a logical model based on the modernized STORET
concept and requirements analysis. GLENDA was developed with the following guiding principles:
True multi-media scope - water, air, sediment, taxonomy, fish tissue, fish diet, and meteorology
data can all be housed in the database
• Data of documented quality - data quality is documented by including results of quality control
parameters
Extensive contextual indicators - ensures data longevity by including enough information to allow
future or secondary users to make use of the data
• Flexible and expandable - the database is able to accept data from any Great Lakes monitoring
project
National compatibility - GLENDA is compatible with STORET and allows ease of transfer between
these large databases
In an effort to reduce the data administration burden and ensure consistency of data in this database,
GLNPO developed several key tools. Features including standard data definitions, reference tables,
standard automated data entry applications, and analytical tools are (or will soon be) available.
15.4.3 Public Access to LMMB Data
All LMMB data that have been verified (through the QC process) and validated (accepted by the PI) are
available to the public. Currently, GLNPO requires that written requests be made to obtain LMMB data.
The data sets are available in several formats including WK1, DBF, and SD2. More information about
the data sets is available on the LMMB web site at: http://www.epa.gov/glnpo.
The primary reason for requiring an official request form for LMMB data is to keep track of requests.
This allows GLNPO to know how many requests have been made, who has requested data, and what use
they intend for the data. This information assists GLNPO in managing and providing public access to
Great Lakes data and conducting public outreach activities. As of November 2000, 38 requests for
LMMB data have been made: 8 from EPA, 5 from other federal agencies, 5 from state agencies, 5 from
universities, 10 from consultants, 3 from international agencies, and 2 from non-profit or other groups. In
the future, after all data are verified and validated, GLNPO intends to make condensed versions of the
data sets available on the LMMB web site for downloading. This will allow easy public access to LMMB
data.
Additional details of the information management for the LMMB Study can be found in The Lake
Michigan Mass Balance Study Quality Assurance Report (USEPA, 200 Ib).
1.5.5 Quality Assurance Program
At the outset of the LMMB Study, managers recognized that the data gathered and the models developed
from the study would be used extensively by decision makers responsible for making environmental,
economic, and policy decisions. Environmental measurements are never true values and always contain
some level of uncertainty. Decision makers, therefore, must recognize and be sufficiently comfortable
April 2004 1-13
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
with the uncertainty associated with data on which their decisions are based. In recognition of this
requirement, LMMB Study managers established a QA program goal of ensuring that data produced
under the LMMB Study would meet defined standards of quality with a specified level of confidence.
The QA program prescribed minimum standards to which all organizations collecting data were required
to adhere. Data quality was defined, controlled, and assessed through activities implemented within
various parameter groups (e.g., organic, inorganic, and biological parameters). QA activities included the
following:
QA Program - Prior to initiating data collection activities, plans were developed, discussed, and
refined to ensure that study objectives were adequately defined and to ensure that all QA activities
necessary to meet study objectives were considered and implemented.
• QA Workgroup - EPA established a QA Workgroup whose primary function was to ensure that the
overall QA goals of the study were met.
• QA Project Plans (QAPPs) - EPA worked with Pis to define program objectives, data quality
objectives (DQOs), and measurement quality objectives (MQOs) for use in preparing QAPPs.
Principal investigators submitted QAPPs to EPA for review and approval. EPA reviewed each QAPP
for required QA elements and soundness of planned QA activities.
Training - Before data collection activities, Pis conducted training sessions to ensure that individuals
were capable of properly performing data collection activities for the LMMB Study.
• Monthly Conference Calls and Annual Meetings - EPA, Pis, and support contractors participated
in monthly conference calls and annual meetings to discuss project status and objectives, QA issues,
data reporting issues, and project schedules.
• Standardized Data Reporting Format - Principal investigators were required to submit all data in a
standardized data reporting format that was designed to ensure consistency in reporting and facilitate
data verification, data validation, and database development.
• Intercomparison Studies - EPA conducted studies to compare performance among different Pis
analyzing similar samples. The studies were used to evaluate the comparability and accuracy of
program data.
• Technical Systems Audits - During the study, EPA formally audited each Pi's laboratory for
compliance with their QAPPs, the overall study objectives, and pre-determined standards of good
laboratory practice.
• Data Verification - Pis and EPA evaluated project data against pre-determined MQOs and DQOs to
ensure that only data of acceptable quality would be included in the program database.
Statistical Assessments - EPA made statistical assessments of the LMMB Study data to estimate
elements of precision, bias, and uncertainty.
• Data Validation - EPA and modelers are evaluating the data against the model objectives.
Comparability of data among Pis participating in the LMMB Study was deemed to be important for
successful completion of the study. Therefore, measurement quality objectives (MQOs) for several data
attributes were developed by the Pis and defined in the QAPPs. MQOs were designed to control various
phases of the measurement process and to ensure that the total measurement uncertainty was within the
ranges prescribed by the DQOs. MQOs were defined in terms of six attributes:
Sensitivity/Detectability - The determination of the low-range critical value that a method-specific
procedure can reliably discern for a given pollutant. Sensitivity measures included, among others,
method detection limits (MDLs) as defined at 40 CFR Part 136, system detection limits (SDLs), or
instrument detection limits (IDLs).
• Precision - A measure of the degree to which data generated from replicate or repetitive
measurements differ from one another. Analysis of duplicate samples was used to assess precision.
1-14 April 2004
-------
Project Overview
• Bias - The degree of agreement between a measured and actual value. Bias was expressed in terms
of the recovery of an appropriate standard reference material or spiked sample.
Completeness - The measure of the number of samples successfully analyzed and reported compared
to the number that were scheduled to be collected.
• Comparability - The confidence with which one data set can be compared to other data sets.
• Representativeness - The degree to which data accurately and precisely represent characteristics of a
population, parameter variations at a sampling point, a process condition, or an environmental
condition.
The Pi-defined MQOs also were used as the basis for the data verification process. GLNPO conducted
data verification through the LMMB QA Workgroup. The workgroup was chaired by GLNPO's Quality
Assurance Manager and consisted of quality control coordinators that were responsible for conducting
review of specific data sets. Data verification was performed by comparing all field and QC sample
results produced by each PI with their MQOs and with overall LMMB Study objectives. If a result failed
to meet predefined criteria, the QC Coordinator contacted the PI to discuss the result, verify that it was
correctly reported, and determine if corrective actions were feasible. If the result was correctly reported
and corrective actions were not feasible, the results were flagged to inform data users of the failure.
These flags were not intended to suggest that data were not useable; rather they were intended to caution
the user about an aspect of the data that did not meet the predefined criteria. Data that met all predefined
requirements were flagged to indicate that the results had been verified and were determined to meet
applicable MQOs. In this way, every data point was assigned one or more validity flags based on the
results of the QC checks. GLNPO also derived data quality assessments for each LMMB Study data set
for a subset of the attributes listed above, specifically sensitivity, precision, and bias. The LMMB Study
modelers and the Large Lakes Research Station Database Manager also perform data quality assessments
prior to inputting data into study models. Such activities include verifying the readability of electronic
files, identifying missing data, checking units, and identifying outliers. A detailed description of the
quality assurance program is included in The Lake Michigan Mass Balance Study Quality Assurance
Report (USEPA, 200 Ib). A brief summary of quality implementation and assessment is provided in each
of the following chapters.
1.6 Project Documents and Products
During project planning, LMMB participants developed study tools including work plans, a methods
compendium, quality assurance project plans, and data reporting standards. Through these tools, LMMB
participants documented many aspects of the study including information management and quality
assurance procedures. Many of these documents are available on GLNPO's web site at:
http: //www. epa.gov/glnpo/lmmb.
LMMB Work Plan
Designers of the LMMB Study have documented their approach in a report entitled Lake Michigan Mass
Budget/Mass Balance Work Plan (USEPA, 1997a). The work plan describes the essential elements of a
mass balance study and the approach used to measure and model these elements in the Lake Michigan
system. This document was developed based upon the efforts of many Federal and State scientists and
staff who participated in the initial planning workshop, as well as Pis.
Quality Assurance Program/Project Plans
The Lake Michigan Mass Balance Project Quality Assurance Plan for Mathematical Modeling, Version
3.0 (USEPA, 1998) documents the quality assurance process for the development and application of
LMMB models, including hydrodynamic, sediment transport, eutrophication, transport chemical fate, and
food web bioaccumulation models.
April 2004 1-15
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
The Enhanced Monitoring Program Quality Assurance Program Plan
The Enhanced Monitoring Program Quality Assurance Program Plan (USEPA, 1997c) was developed in
1993 to ensure that data generated from the LMMB Study supports its intended use.
LMMB Methods Compendium
The Lake Michigan Mass Balance Project (LMMB) Methods Compendium (USEPA, 1997d, 1997e)
describes the sampling and analytical methods used in the LMMB Study. The entire three volumes are
available on GLNPO's web site mentioned above.
LMMB Data Reporting Formats and Data Administration Plan
Data management for the LMMB Study was a focus from the planning stage through data collection,
verification, validation, reporting, and archiving. The goal of consistent and compatible data was a key to
the success of the project. The goal was met primarily through the development of standard formats for
reporting environmental data. The data management philosophy is outlined on the LMMB web site
mentioned above.
Lake Michigan LaMP
"Annex 2" of the 1972 Canadian-American Great Lakes Water Quality Agreement (amended in 1978,
1983, and 1987) prompted development of Lakewide Area Management Plans (LaMPs) for each Great
Lake. The purpose of these LaMPs is to document an approach to reducing input of critical pollutants to
the Great Lakes and restoring and maintaining Great Lakes integrity. The Lake Michigan LaMP calls for
basin-wide management of toxic chemicals.
GLENDA Database
Central to the data management effort is a computerized data system to house Lake Michigan Mass
Balance and other project results. That system, the Great Lakes Environmental Monitoring Database
(GLENDA), was developed to provide data entry, storage, access and analysis capabilities to meet the
needs of mass balance modelers and other potential users of Great Lakes data.
LMMB Data Reports
This report is one in a series of data reports that summarize the data from monitoring associated with
EPA's Lake Michigan Mass Balance Study. In addition to this data report on PCBs and fra«s-nonachlor,
data reports are being published for atrazine (USEPA, 200Ic) and mercury (USEPA, 2004).
Future Documents and Products
Following the completion of modeling efforts associated with the LMMB Study, GLNPO anticipates
publishing reports summarizing the modeling results. In 2005, GLNPO also anticipates conducting a
reassessment of Lake Michigan to calibrate and confirm modeling results with data collected 10 years
after the initial LMMB sampling.
1-16 April 2004
-------
Figure 2-1. Generalized 2-D Structure of
Biphenyl
Chapter 2
PCB/trans-Nonachlor Study Overview
2.1 PCB Introduction
2.1.1 Physical/Chemical Properties
Polychlorinated biphenyls (PCBs) are a class of synthetic
organic chemicals characterized by two six-membered
aromatic carbon rings joined by a single carbon-carbon bond
that constitute the parent biphenyl. One or more chlorine
atoms are attached to carbon atoms in the two ring structures.
There are 209 possible arrangements of chlorine atoms, and
each of the arrangements is referred to as a PCB "congener."
The carbon atoms on both of the aromatic rings are numbered
1 through 6, to indicate their position around each ring. The
superscript prime (') is used to distinguish similar positions on the two rings. The two carbon atoms linking
the rings are numbered 1 and 1'. Each of the remaining 10 carbon atoms (numbered 2 to 6 and 2' to 6') can
bond with one hydrogen or chlorine atom. The formal chemical name of each of the 209 congeners identifies
the specific positions and the total number of chlorine atoms in the congener, e.g., 2,2',4,5'-
tetrachlorobipheny 1.
5'
Table 2-1. Numbers of Congeners in Each Level of Chlorination
There are 3 possible PCB congeners that
contain a single chlorine atom. The two
ring structures are symmetrical, such that
the structures 2-monochlorobiphenyl and
2'-monochlorobiphenyl are identical. In
addition, the two rings can rotate along
the carbon bond that joins them, such
that the 2 and 6 positions on each ring
are equivalent, as are the 3 and 5
positions. By convention, the equivalent
position with the lower number is used to
describe the compound. Thus, a
monochlorobiphenyl with the chlorine
attached in the 2 or 2', or 6 or 6' position
is named 2-monochlorobiphenyl. The
monochlorobiphenyl with the chlorine
attached in the 3 or 3', 5 or 5' position is named 3-monochlorobiphenyl. The monochlorobiphenyl with the
chlorine attached in the 4 or 4' position is named 4-monochlorobiphenyl.
PCBs are often described in terms of the "level of chlorination," which refers to the number of chlorine
atoms attached to the biphenyl ring (e.g., monochlorobiphenyls, dichlorobiphenyls, and trichloro-
biphenyls). The numbers of PCB congeners in each level of chlorination are shown in Table 2-1. The
physical and chemical properties of PCBs are dependent on the number of chlorine atoms and their
respective positions in the two ring structures, thus, they vary with the congener. In general, PCBs with
the same number of chlorine atoms have similar physical and chemical properties.
GLNPO adopted the numbering system developed by Ballschmiter and Zell (1980) that simplifies the
identification of each congener by assigning each possible congener a number 1 through 209. That
notation is used throughout the remainder of this discussion.
No. of Chlorine Atoms
1
2
3
4
5
6
7
8
9
10
Level of Chlorination
monochlorobiphenyl
dichlorobiphenyl
trichlorobiphenyl
tetrachlorobiphenyl
pentachlorobiphenyl
hexachlorobiphenyl
heptochlorobiphenyl
octachlorobiphenyl
nonachlorobiphenyl
decachlorobiphenyl
No. of Congeners
3
12
24
42
46
42
24
12
3
1
April 2004
2-1
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
There is a subset of PCB congeners in which the pattern of chlorine substitution causes "steric hindrance"
that limits the rotation of the two aromatic rings around their common carbon-carbon bond such that the
two rings lie in the same plane. These congeners are called "coplanar" PCBs and as a result of their flat
configuration, they have a greater ability to penetrate the walls of living cells and generally exhibit greater
toxicological effects (see Section 2.3).
In general, PCBs are chemically inert, nonflammable, and do not transmit electrical current. These
properties, combined with high melting and boiling points made PCBs useful in a wide variety of
industrial applications, particularly as dielectric fluids in electrical transformers and capacitors. Flash
points for PCBs are in the range of 170 to 380°C (H. Fiedler, 2001)
PCBs are readily soluble in organic solvents, but have low solubility in water. Water solubility decreases
dramatically as the number of chlorine atoms attached to the parent biphenyl structure increases. For
example, the water solubility of unsubstituted biphenyl has been reported to be within the range of 5.9 to
7.5 mg/L, while the water solubility of PCB 209, containing 10 chlorine atoms, is estimated at 4xlO"6
mg/L, or 4 ng/L (Shiu and Mackay, 1986). The vapor pressures of PCB congeners decrease in a
similarly dramatic fashion with increased chlorination. Dunnivant and Elzerman (1988) report vapor
pressures for several dichlorobiphenyl congeners on the order of 2xlO"6 atmospheres, while the vapor
pressures for some hexachlorobiphenyls are on the order of 3xlO"10 or 3xlO"n atmospheres. In the LMMB
Study, PCBs served as a model for conservative organic pollutants.
2.1.2 History of PCB Production
In the United States, the Monsanto Company
produced commercial mixtures of PCBs by
chlorinating biphenyl and sold the mixtures under
the trade name Aroclor. There were nine Aroclor
mixtures produced in the U.S. and they were
differentiated by a series of four-digit numbers
(e.g., Aroclor 1242). The names of eight of the
nine mixtures begin with the number "12,"
representing the 12 carbon atoms in the parent
biphenyl, and end in a two-digit number that
represents the percentage of chlorine (by weight) in
the mixture. Aroclor 1016 was the only mixture
that violated this naming scheme, because its
number began with "10" instead of "12," and it
contained more than 16% chlorine. The nine
Aroclor mixtures are shown in Table 2-2, along
with their percent chlorine (by weight), and the
approximate percentage of domestic production
that they comprised for the period from 195 7 to 1977.
Aroclors with the greatest U.S. production, in terms
Table 2-2. U.S. Domestic Production of Commercial
Aroclor Mixtures from 1957 to 1977*
Mixture
Aroclor 101 6
Aroclor 1221
Aroclor 1232
Aroclor 1242
Aroclor 1248
Aroclor 1254
Aroclor 1260
Aroclor 1262
Aroclor 1268
Percent Chlorine
by Weight
41
21
32
42
48
54
60
62
68
Percent of U.S.
Production
13
1
<1
52
7
16
11
1
<1
*Adapted from Brown, 1994. Total production percent is
greater than 100%, due to rounding.
Table 2-3 contains the typical compositions of the five
of the percentages of various levels of chlorination.
PCB production facilities were built in Austria, Germany, France, Great Britain, Italy, Japan, Spain, the
USSR, and the U.S. PCB mixtures were produced in Germany under the trade name Clophens, in Japan
under the names Kanechlors and Sanotherm, and in France as Phenoclor and Pyralene. World-wide
production of Aroclors is estimated to have been 1.5 million metric tonnes (3.3 billion pounds)
(Rantanen, 1992; Ivanov and Sandell, 1992).
2-2
April 2004
-------
PCB/trans-Nonachlor Study Overview
Table 2-3. Typical Composition (%) of the Five Aroclor Mixtures with Greatest U.S. Production*
Biphenyls by Level of Chlorination
Monochlorobiphenyls
Dichlorobiphenyls
Trichlorobiphenyls
Tetrachlorobiphenyls
Pentachlorobiphenyls
Hexachlorobiphenyls
Heptachlorobiphenyls
Octachlorobiphenyls
Nonachlorobiphenyls
Decachlorobiphenyl
Aroclor
1016
2
19
57
22
<1%
<1%
<1%
<1%
<1%
<1%
1242
1
13
45
31
10
<1%
<1%
<1%
<1%
<1%
1248
<1%
1
21
49
27
2
<1%
<1%
<1%
<1%
1254
<1%
<1%
1
15
53
26
4
<1%
<1%
<1%
1260
<1%
<1%
<1%
<1%
12
42
38
7
1
<1%
*Adapted from PCBs: Cancer Dose-Response Assessment and Application to Environmental Mixtures, EPA/600/P-96/001F,
September 1996.
2.1.3 Regulatory Background
PCB production and export in the U.S. was halted in October 1977 under the auspices of the Toxic
Substances Control Act (TSCA). Use and import of PCBs were banned in Japan in 1972. In addition to
the production ban instituted under TSCA, EPA regulates PCBs, as Aroclors, under a wide range of
environmental statutes. For example, Aroclors are regulated in effluent guidelines developed under the
Clean Water Act and administered through the National Pollutant Discharge Elimination System
(NPDES). The Office of Water has established water quality criteria (WQC) for freshwater and marine
systems. The freshwater chronic WQC is 0.014 [ig/L of total Aroclor. The marine chronic WQC is 0.030
lig/L. The WQC for human health is 4.5 x 10'5 [ig/L. Under the Safe Drinking Water Act, EPA has
established a maximum contaminant limit (MCL) of 0.50 |^g/L for total Aroclor. Under the auspices of
the Resource Conservation and Recovery Act (RCRA) EPA has placed Aroclors on Appendix VIII
(hazardous substances) and Appendix IX (groundwater monitoring), and has established a Universal
Treatment Standard (UTS) of 1 mg/kg of Aroclors in non-wastewaters and 0.10 mg/L in wastewaters.
PCBs are included in the Toxics Release Inventory (TRI) developed under the Emergency Planning and
Community Right to Know Act (EPCRA).
2.1.4 Fate and Effects
The fate and effects of PCBs in the environment are driven by the physical properties of the individual
PCB congeners. In general, PCBs are hydrophobic and lipophilic. Therefore, they are more likely to be
found in soils, sediments, and tissues, than dissolved in water. PCBs found in water are likely to be
associated with the particulate phase, with some of the lower chlorinated congeners present in the
dissolved phase.
Baker and Eisenreich (1990) examined the behavior of PCBs in Lake Superior, finding that PCBs
volatilize when river inputs to the lake are relatively high in PCBs. They calculated volatilization rates
for PCBs that are approximately equal to the rate of atmospheric deposition into the lake. These findings
support a model of PCB cycling proposed by Mackay and Patterson (1986) in which PCBs dissolved in
rain or sorbed onto air particulates are transported to surface water. This input results in a fugacity
April 2004
2-3
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
gradient that favors volatilization from the water into the atmosphere, where the PCBs are dissolved in
rain water or sorbed onto particulates again, thus repeating the cycle.
2.1.5 Biological Transformations
PCBs can be degraded by microorganisms by several mechanisms, including both aerobic and anaerobic
processes. Although PCBs are generally resistant to aerobic breakdown, Rochkind etal. (1986) found
that microorganisms of the genera Acetobacter, Alcaligenes, and Pseudomonas are capable of degrading
PCBs under aerobic conditions. However, the rates of degradation are greatly influenced by the chlorine
substitution pattern in each congener. Rochkind et al. found that:
higher chlorinated PCBs degrade more slowly under aerobic conditions than those with fewer
chlorine atoms;
• PCBs with chlorines on only one ring are metabolized more rapidly than PCBs in the same level of
chlorination, but with the chlorine atoms distributed on both rings;
• the ring with fewer chlorines will be hydroxylated first, and;
chlorine atoms attached at the ortho position (2 or 6 on the ring) significantly inhibit degradation.
In anaerobic environments, PCBs will undergo reductive dechlorination (loss of chlorine), with the
degradation rate related to the number of chlorine atoms on the PCB. As a result, the more highly
chlorinated congeners are more readily dechlorinated under anaerobic conditions. Fiedler et al. (1994)
found that the chlorines in the meta and para positions (3, 4, and 5 on the ring) were more readily lost by
reductive dechlorination, resulting in an apparent increase in the prevalence of the or^o-substituted PCBs
in the environment.
PCBs are lipophilic, concentrating in fatty tissues of organisms. As a result, PCBs are not readily
excreted from organisms as intact chlorinated biphenyls. The metabolic transformation of PCBs requires
that the molecule undergo hydroxylation to make it more polar, and thus more water soluble. In higher
organisms, PCBs are hydroxylated through a metabolic pathway involving the hepatic monooxygenase
system mediated by the enzyme Cytochrome P-450. As with other transformations of PCBs, the rates are
related to the level of chlorination and the substitution patterns.
As a result of both their lipophilic nature and their low rates of biodegradation, PCBs accumulate, or
bioconcentrate, to higher concentrations in each subsequent trophic level of an ecosystem. Fiedler et al.
(1994) has shown that both the total PCB concentrations and the concentrations of the "toxic" or "dioxin-
like" PCBs (see Section 2.1.6) increase in typical aquatic systems consisting of phytoplankton,
zooplankton, plankton-eating fish, piscivorous fish, and piscivorous birds.
2.1.6 Toxicity
EPA has classified PCBs as carcinogens. Other human health effects include disruption of endocrine
systems. Effects of short-term exposures in humans include chloracne, changes in skin pigmentation,
numbness of the limb and general weakness. Long-term exposures have been associated with changes in
liver function, irritations of the nose, throat and intestinal tract, fertility problems, birth defects, premature
births, and neurological and developmental problems in newborns.
Many of these health effects of PCBs are similar to those related to exposures to poly chlorinated dibenzo-
/•-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs) and are believed to result from the
actions of a subset of the 209 possible PCB congeners known as the "coplanar" PCBs. The coplanar
PCBs have structures that resemble the 2,3,7,8-substituted PCDDs/PCDFs, in which the two aromatic
rings lie in one plane and are believed to be more readily transported across cell membranes. The World
2-4 April 2004
-------
PCB/trans-Nonachlor Study Overview
Health Organization (WHO) has classified 12 PCB congeners as the "toxic" PCBs, or the "dioxin-like"
PCBs, based on structure-activity relationships (see Table 2-4 and Van den Berg etal., 1998). However,
for reasons discussed in Section 2.6, the toxic PCBs were not the primary focus of the LMMB Study.
Table 2-4. World Health Organization Toxic PCB Congeners
Congener Number
77
81
126
169
105
114
118
123
156
157
167
189
PCB Congener
3,3',4,4'-Tetrachlorobiphenyl
3,4,4',5-Tetrachlorobiphenyl
3,3',4,4',5-Pentachlorobiphenyl
3,3',4,4',5,5'-Hexachlorobiphenyl
2,3,3',4,4'-Pentachlorobiphenyl
2,3,4,4',5-Pentachlorobiphenyl
2,3',4,4',5-Pentachlorobiphenyl
2',3,4,4',5-Pentachlorobiphenyl
2,3,3',4,4',5-Hexachlorobiphenyl
2,3,3',4,4',5'-Hexachlorobiphenyl
2,3',4,4',5,5'-Hexachlorobiphenyl
2,3,3',4,4',5,5'-Heptachlorobiphenyl
Structural Group
Non-orf/io substituted PCB
Non-orf/io substituted PCB
Non-orf/io substituted PCB
Non-orf/io substituted PCB
Mono-orf/io substituted PCB
Mono-orf/io substituted PCB
Mono-orf/io substituted PCB
Mono-orf/io substituted PCB
Mono-orf/io substituted PCB
Mono-orf/io substituted PCB
Mono-orf/io substituted PCB
Mono-orf/io substituted PCB
The term "ortho" in Table 2-4 refers to the position of the chlorines attached to the phenyl ring structure,
relative to the carbon-carbon bond between the two rings. Chlorines attached at the 2, 2', 6, or 6'
positions of the biphenyl structure (see Figure 2-1) are in the "ortho" position. A "non-ort/2o" congener
does not have any chlorines in the 2, 2', 6, or 6' positions, while a "mono-ort/2o" congener has a chlorine
attached at one of those positions.
2.2 frans-Nonachlor Introduction
2.2.1 Physical/Chemical Properties
fra«5-Nonachlor is the common name for 1,2,3,4,5,6,7,8,8-
nona-chloro-3a,4,7,7a-tetrahydro-4,7-methanoindan,amember
of the class of cyclodiene pesticides. fra«s-Nonachlor is a
component of the pesticide formulation "technical chlordane,"
comprising about 1.1% of that mixture of at least 140 related
compounds. fra«s-Nonachlor differs from the two structural
isomers of the compound chlordane (cis- and trans -chlordane)
by the addition of one additional chlorine atom, in the number
3 position of the molecule (see Figures 2-2 and 2-3).
Figure 2-2. Generalized 2-D Structure of
frans-Nonachlor
Cl
April 2004
2-5
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Compared to the two isomers of chlordane, there is very little
published physical-chemical data for fra«s-nonachlor. For
example, Mackay et al. (1992) do not list fra«s-nonachlor at
all. The general properties of fra«s-nonachlor can be inferred
from those of chlordane. The addition of one more chlorine
atom to the chlordane structure increases the melting and
boiling points for fra«s-nonachlor, decreases the water
solubility, and decreases the vapor pressure. fra«s-Nonachlor
has low solubility in water, but is readily soluble in organic
solvents. It has a melting point above 100°C and a boiling
point above 175°C.
In the LMMB Study, fra«s-nonachlor serves as a model for the
cyclodiene pesticides.
2.2.2 frans-Nonachlor Production
Figure 2-3. Generalized 2-D Structure of
frans-Chlordane
fra«5-Nonachlor was not produced as a pure compound, but was one of the major components of
technical chlordane mixtures. Chlordane is produced through the chlorination of cyclopentadiene to form
hexachloropentadiene. This intermediate product is condensed to form chlordene, which undergoes
additional chlorination to produce cis- and fra«s-chlordane, plus fra«s-nonachlor, heptachlor, and several
other major components. Chlordane mixtures were first produced in the U.S. in 1948 and various
formulations of chlordane were widely used as pesticides on food crops and lawns, and for termite control
from 1948 to 1988. In April 1988, all commercial uses in the U.S. were banned (USEPA, 1988).
Since 1988, the only U.S. domestic manufacturer of chlordane has been the Velsicol Chemical Company
of Memphis, TN. Production data for chlordane are difficult to obtain, but EPA estimated that 3.5 to 4
million pounds of chlordane were distributed in 1986. Data from the Toxics Release Inventory for 1990
indicate that 100,000 to 1 million pounds of chlordane were produced that year.
EPA estimated that more than 7.5 million pounds of chlordane were used for home, lawn, and garden
purposes in the U.S. in 1974 (USEPA, 1975). Other sources estimate total production in 1974 was on the
order of 21 million pounds, suggesting that over 13 million pounds of chlordane were exported that year
(WHO, 1988).
2.2.3 Regulatory Background
fra«5-Nonachlor has been regulated in the U.S. as a component of chlordane. Beginning in 1975, EPA
ordered a halt to the use of chlordane and the related pesticide heptachlor for most household and
agricultural uses, citing an imminent human cancer hazard. The 1975 action limited the use of chlordane
to underground injection for termite control and treatment of the roots and tops of non-food plants. From
July 1983 to April 1988, EPA further restricted the use of chlordane to underground injection for termite
control. In April 1988, EPA canceled all commercial uses of chlordane in the U.S.
Regulation of fra«s-nonachlor has typically been accomplished by analogy to chlordane. Thus, the
maximum contaminant limit (MCL) established under the Safe Drinking Water Act is 2 i-ig/L for trans-
nonachlor, the same MCL used for chlordane. Similarly, the water quality criteria developed under the
Clean Water Act for fra«s-nonachlor have the same values as were developed for chlordane.
Chlordane has been banned in 47 countries, including the U.S., and 14 additional countries have severely
restricted its use.
2-6
April 2004
-------
PCB/trans-Nonachlor Study Overview
2.2.4 Fate and Effects
All of the components of chlordane have been found to bioaccumulate, with fra«s-nonachlor the most
bioaccumulative of the components. In general, fra«s-nonachlor is hydrophobic and lipophilic.
Therefore, it is more likely to be found in soils, sediments, and tissues than dissolved in water. When
found in water, fra«s-nonachlor is likely to be associated with the particulate phase, rather than actually
dissolved in water.
2.2.5 Toxicity
As with other aspects of fra«s-nonachlor, the majority of the data on toxicity has been determined for
either technical chlordane mixtures or the predominant chlordane isomers. To date, the data from human
studies have not provided sufficient evidence to conclude that either chlordane or fra«s-nonachlor is a
human carcinogen. However, mice fed low levels of chlordane in food developed liver cancer.
Therefore, trans-nonachlor is considered to be a probable human carcinogen. Other human health effects
include neurological effects, blood dyscrasia, hepatoxicity, immunotoxicity, and endocrine system
disruption.
2.3 Study Design
2.3.1 Description
PCBs and fra«s-nonachlor were chosen for analysis in the LMMB Study as representatives of the
persistent, bioaccumulative chlorinated compounds. PCB congeners and fra«s-nonachlor were measured
in vapor, precipitation, particulates, atmospheric dry deposition, water in the open lake, tributaries,
sediment, lower pelagic food web organisms, and fish. The data generated from this study were used to
estimate an overall mass balance of PCBs and fra«s-nonachlor in Lake Michigan (see Section 1.4).
2.3.2 Scope
To develop a mass balance of PCBs and fra«s-nonachlor in Lake Michigan, all significant sources and
stores of PCBs and fra«s-nonachlor in the environment were measured. Significant sources and stores
included tributary inputs, atmospheric inputs from the vapor phase, particulate phase, and precipitation,
sediment, lower pelagic food web organisms, and fish. The specific components that were studied are
shown in Table 2-5.
Field sampling was conducted from February 1994 through October 1995, with an additional sampling
cruise in May 1996 to retrieve sediment traps and collect samples at Stations LM94-11, LM94-17,
LM94-18, LM94-21S and LM94-32.
2.3.3 Organization/Management
The responsibility for collecting and analyzing PCBs and fra«s-nonachlor samples from the various
components was divided among multiple principal investigators (Table 2-5). Each principal investigator
developed a Quality Assurance Project Plan (QAPP) that was submitted to EPA's Great Lakes National
Program Office. The QAPPs detailed the project management, study design, and sampling and analysis
procedures that would be used in the study and the quality control elements that would be implemented to
protect the integrity of the data. The LMMB Quality assurance program is further discussed in Section
2.6, and detailed information on the quality assurance activities and data quality assessment specific to
each ecosystem component are discussed in Chapters 3-5.
April 2004 2-7
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 2-5. Components Sampled by Principal Investigators
Ecosystem Compartment
Atmosphere
Tributary
Open Lake
Sediment
Lower Pelagic
Food Web Organisms
Fish
Component
Vapor
Particulate
Precipitation
Vapor
Particulate
Precipitation
Dry deposition
Dissolved
Total
Dissolved
Total
Surficial
Resuspended
Mysis
Diporeia
Zooplankton
Phytoplankton
Lake Trout
Slimy sculpin
Deepwater sculpin
Coho salmon
Alewife
Bloater chub
Smelt
Principal Investigator
Clyde Sweet, Illinois State Water Survey
(Sleeping Bear Dunes from 4/94 to 7/94 and all other sites)
Ron Hites and Nora Basu, Indiana University
(Sleeping Bear Dunes only, from 8/94 to 10/95)
Steve Eisenreich, University of Minnesota Gray Freshwater Biological
Institute and Rutgers University
William Sonzogni, University of Wisconsin, Wisconsin State Lab of
Hygiene
Eric Crecelius, Battelle Sequim
Patricia Van Hoof and Brian Eadie, Great Lakes Environmental
Research Laboratory, National Oceanic and Atmospheric
Administration
Deborah Swackhamer, University of Minnesota
Robert Hesselberg and James Mickey, United States Geological
Survey, Biological Resources Discipline (formerly National Biological
Service)
2.4 Sampling Locations
2.4.1 Atmospheric Components
Atmospheric samples were collected at eight shoreline sampling stations and 16 open-lake sampling
stations within Lake Michigan and two open-lake sampling stations in Green Bay. In addition, three out-
of-basin land-based sampling stations were established as regional background sites to represent air
coming over Lake Michigan during periods of southwest or northwest prevailing winds. The sampling
locations and sampling frequencies for the LMMB Project were selected through discussions with experts
in the filed during several workshops, including the Great Lakes Mass Balance Planning Workshop in
April 1992 and the LMMB Planning Meeting in September 1993. Site-selection criteria considered
predominant annual wind directions, source areas, and episodic summer events. In general, sites were
selected to be regionally representative of land-use categories and to represent the different potential
sources of pollutants in this study (e.g., releases associated with population centers versus applications of
agricultural activities). The sampling frequencies were designed to capture the expected variability at the
sites (e.g., more frequent sampling in urban areas versus less frequent sampling in remote areas and more
frequent sampling in spring summer, and fall months).
The shoreline atmospheric sampling stations include those specific to the LMMB Study as well as several
that are part of the Integrated Atmospheric Deposition Network (IADN). Samples were collected from
2-8
April 2004
-------
PCB/trans-Nonachlor Study Overview
the land-based IADN stations at the Brule River, Eagle Harbor, and Bondville from April 1994 through
October 1995. Sampling at these IADN stations was governed by study design and quality assurance
programs specific to IADN, but generally similar to those in the LMMB Study, so the data have been
incorporated into the LMMB database. The locations of the shoreline atmospheric PCB sampling stations
within the Lake Michigan basin are shown in Figure 2-4. The site classifications, planned sampling
frequencies, and types of vapor-phase and particulate-phase samples are shown in Table 2-6. These
frequencies were generally followed as sampling schedules permitted and except in cases of sampler
malfunction, lack of precipitation, or when circumstances prevented retrieval of a sample.
Table 2-6. Site Classifications, Planned Frequencies, and Types of Vapor-phase and Particulate-phase
Samples Collected at Shoreline and Out-of-basin Stations
Site Classification
Urban - Major urban sources within
1km
Urban-influenced - Major urban
sources within 10 km
Rural - Urban sources generally 10
to 50 km away, but agricultural
sources within 1 km
Remote - No urban areas or major
sources of air pollutants within 50
km
Planned Frequency
One 24-h composite
collected every 3 days
One 24-h composite
collected every 3 days
One 24-h composite
collected every 6 days
One 24-h composite
collected every 12
days
Site Names
NT Chicago
Chiwaukee Prairie
Indiana Dunes
Manitowoc
Muskegon
Bondville
South Haven
Beaver Island
Brule River
Eagle Harbor
Sleeping Bear
Dunes
Sample Type
Multiple 24-h samples composited
to represent 1 sample per month
Multiple 24-h samples composited
to represent 1 sample per month
Multiple 24-h samples composited
to represent 1 sample per month
Multiple 24-h samples composited
to represent 1 sample per month
Individual 24-h samples analyzed
and mathematically composited to
represent 1 sample per month
PCBs and trans-nonachlor were measured in vapor, particulates, and precipitation samples collected at 16
locations in Lake Michigan during seven cruises of the Research Vessel Lake Guardian between April
1994 and October 1995 and at two stations in Green Bay. Because these open-lake samples were
collected on board the ship, they are single-day samples, and were not composited by month.
Monthly integrated dry deposition samples were collected from the following stations: 68th Street Crib,
Sleeping Bear Dunes, Harrison Crib, IIT Chicago and South Haven. In addition, two 4-day composite
dry deposition samples were collected on the ship in July 1994.
April 2004
2-9
-------
Figure 2-4. Atmospheric Sampling Stations
Atmospheric Stations o IADN Stations ©
Eagle
Harbor
r.
Lake Superior
-,
*
' ///Michigan rT
0 ° .® Sleeping Bear
rnn«
40M
'. Lake
Manitowocc Michigan '•
\
280
27M
O
23M
0
Chwaukee Prairie :
•
0 *V
MB19M'
18M
° ° rV
380!o South
Haven?j_
11
6
_
5 6 \
o o S
IIT-Chicago^ Qi
' O Wna Dun
••
Chicago SWFF ' d O Wna Dunes
Crib Wake O
George Washington HS
2-10
-------
PCB/trans-Nonachlor Study Overview
Figure 2-5. Tributary Sampling Stations
2.4.2 Tributaries
Tributary samples were collected from 11 rivers
that flow into Lake Michigan (Figure 2-5).
These tributaries included the Menoninee, Fox,
Sheboygan, and Milwaukee Rivers in
Wisconsin; the Grand Calumet River in
Indiana; and the St. Joseph, Kalamazoo, Grand,
Muskegon, Pere Marquette, and Manistique
Rivers in Michigan. With the exception of the
Pere Marquette River, these tributaries were
selected for the LMMB Study because of
elevated concentrations of contaminants in
resident fish. The Pere Marquette River was
selected because it has a fairly large and
pristine watershed. Samples collected from the
Pere Marquette River can be used to estimate
loads from the small portion of the Lake
Michigan watershed that was not monitored in
this study. The 11 monitored tributaries
represent greater than 90% of the total river
flow into Lake Michigan and an even higher
percentage of the total tributary load of
pollutants into Lake Michigan.
Table 2-7 describes specific watershed
characteristics and impairment information for
each of the monitored tributaries. Of the 11
tributaries, 6 (the Kalamazoo, Manistique,
Menominee, Fox, Sheboygan, and Grand
Calumet Rivers) are classified as Great Lakes areas of concern (AOCs). Areas of concern are severely
degraded geographic areas within the Great Lakes Basin. They are defined by the US-Canada Great Lakes
Water Quality Agreement (Annex 2 of the 1987 Protocol) as "geographic areas that fail to meet the general
or specific objectives of the agreement where such failure has caused or is likely to cause impairment of
beneficial use or the area's ability to support aquatic life." Most of the eleven tributaries are also listed on
the Clean Water Act Section 303(d) list of impaired water bodies due to contamination from mercury, PCBs,
and other pollutants.
April 2004
2-11
-------
Table 2-7. Watershed Characteristics for Tributaries Monitored in the LMMB Study
Tributary
St. Joseph
Kalamazoo
Grand (lower)
Muskegon
Pere Marquette
Manistique
Menominee
Fox (lower)
Sheboygan
Milwaukee
Grand Calumet
Watershed
area
(mi2)
4685
2047
2003
2686
2644
1464
2306
442
2201
864
1039
Total river
miles in
watershed
3743
1560
2014
1886
1356
1061
1660
700
1699
802
760
Riparian Habitat
Forested
25-50%
25-50%
25-50%
25-50%
25-50%
>75%
>75%
25-50%
25-50%
25-50%
25-50%
Agricultural/
Urban
>50%
>50%
>50%
>50%
>50%
20-50%
20-50%
>50%
>50%
>50%
>50%
IWI Score3
3- less serious problems, low
vulnerability
3- less serious problems, low
vulnerability
5- more serious problems, low
vulnerability
5- more serious problems, low
vulnerability
3- less serious problems, low
vulnerability
1- better quality, low vulnerability
1- better quality, low vulnerability
6- more serious problems, high
vulnerability
5- more serious problems, low
vulnerability
5- more serious problems, low
vulnerability
5- more serious problems, low
vulnerability
Impaired forb
Ł co//, mercury, PCBs, pathogens, macro-
invertebrate community
Mercury, PCBs
PCBs, pathogens
Mercury, PCBs
Mercury, PCBs, pathogens
Dioxin, PCBs, mercury, pathogens
PCBs, organic enrichment, dissolved oxygen
PCBs, mercury
PCBs
PCBs, pesticides, lead, mercury, dissolved
oxygen, cyanide, chlorides, impaired biotic
community, oil and grease, copper
Area of
Concern
X
X
X
X
X
X
"EPA's Index of Watershed Indicators Score for assessing the health of aquatic resources.
bBased on 1998 listing of Clean Water Act Section 303(d) impaired waters.
2-12
-------
PCB/trans-Nonachlor Study Overview
2.4.3 Open Lake
Open-lake water column samples were
collected from 38 sampling locations on Lake
Michigan, 2 sampling locations in Green Bay,
and 1 sampling location on Lake Huron (Figure
2-6). Open-lake samples were collected during
eight cruises of the R/VLake Guardian between
April 1994 and September 1995. Due to field
conditions and other considerations, samples
were not collected at all 41 stations during each
cruise. The dates of the eight cruises and the
total number of stations occupied are shown in
Table 2-8.
Figure 2-6. Open-Lake Water Column Sampling Stations
<$"
MBit 11
43
•1
LH5IM
*
The stations used in the LMMB Study and
shown in Figure 2-6 include many historical
stations used by GLNPO and other programs.
Stations established specifically forthe LMMB
Study are shown in Figure 2-6 with identifiers
that begin with "MB." The identifier for the
station in Lake Huron begins with "LH," as
seen in the upper right corner of Figure 2-6.
The identifiers for the two stations in Green
Bay begin with "GB," as seen in the upper left
portion of Figure 2-6. Some of the stations are
those specified by GLNPO as "Master" stations
that are used for various purposes beyond the
LMMB Study. The identifiers for those master
stations end in "M."
Wisconsin
Ilknois
A O 3SO '
5 !
; i -•
Michigan
Indiana
The first survey occurred in the early spring just after "ice out" in April 1994. The second survey was in
early summer (June 1994) after the onset of stratification and following the spring runoff period of
agricultural chemicals from crop land. The third survey was in late summer (August 1994) during later
stages of stratification. The fourth and fifth surveys, conducted in October 1994 and January 1995,
sampled only a few of the Lake Michigan sites. The sixth survey occurred in March 1995 just after ice
out. The seventh and eighth surveys occurred in August and September 1995, during stratification.
Table 2-8. Open-lake Cruise Dates and Number of Stations Occupied
Cruise Date
April 1994
June 1994
August 1994
October 1994
January 1995
March 1995
August 1995
September 1995
Number of Stations Occupied
39
14
41
37
4
40
19
41
April 2004
2-13
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
2.4.4 Sediment
In 1994 and 1995, 117 sediment samples were collected from Lake Michigan and 6 samples were
collected from Green Bay, by box coring, Ponar dredge, and gravity coring. The sediment sampling
locations were selected to help define the three depositional zones (depositional, transitional, and non-
depositional). The locations and the sampling device used at each location are shown in Figure 2-7.
In addition, sediment traps were deployed at eight locations in Lake Michigan (see Figure 2-8). However,
samples could not be retrieved from the traps at two of those locations. Sample retrieval was successful at
trap locations 1, 2, 5, 6, 7, and 8.
Figure 2-7. Sediment Sampling Stations
Figure 2-8. Sediment Trap Locations
2.4.5 Lower Pelagic Food Web Organisms
Plankton sampling locations were selected by GLNPO and the Pis in advance of sampling. The sites
included eight stations in three biological sampling areas or "biota boxes" (Stations 110, 140, 180, 240,
280, 310, 340, and 380), three master stations (18M, 27M, and 47M), and a fourth biota box centered
around Station 5, near Chicago. The four biota boxes are outlined in red in Figure 2-9. Samples were
collected on seven occasions, from June 1994 to September 1995. In addition, zooplankton were
collected from Station 19M in January 1995 and phytoplankton were collected from Stations 23M and 41
in June 1994. A total of 72 zooplankton and 71 phytoplankton samples were collected during the study.
In addition to the plankton samples, samples ofMysis and Diporeia were to be collected at the stations
where lake trout were collected (near biota boxes 1-3) and at one site NE of Chicago. Mysis and Diporeia
samples were collected from stations 140, 180, 240, 280, 340, 380, 47M and 5. Samples ofMysis also
were collected at stations 27M and 18M. A total of 53 Mysis samples and 39 Diporeia samples were
2-14
April 2004
-------
PCB/trans-Nonachlor Study Overview
collected. Mysis feed on both phytoplankton and zooplankton, while Diporeia feed to detritus and
phytoplankton (see Figure 7-5 in Chapter 7).
2.4.6 Fish
Specimens of lake trout, coho salmon, bloater chub, Figure 2-9. Sampling Stations for Lower Pelagic
alewife, smelt, deepwater sculpin, and slimy sculpin Food Web Organisms and Fish
were collected using various means. Two subsets of
alewife and bloater chub were differentiated, based on
total length (see Table 2-9). Coho salmon were
differentiated into three subsets by age (hatchery,
yearling, and adult). In their adult stages, the lake trout
and Coho salmon are piscivorous fish, while the other
fish specifies collected are forage fish that generally
feed on plankton and detritus (see Figure 7-5 in
Chapter 7).
Where possible, five fish were composited for
analysis. Fish were collected during the spring,
summer, and fall of 1994, and the spring and fall of
1995, from three of the four biota boxes in Figure 2-9
(fish were not collected from the biota box at Station
5, near Chicago).
Additional fish were collected from locations
throughout the lake by means ranging from gill nets to
hook-and-line angling. Table 2-9 contains a summary
of the numbers of individual fish collected, the total
number of locations from which they were collected,
and the total number of samples submitted for
analysis, by species. Table 2-10 contains a summary of the sample collection techniques used, by species.
Table 2-9. Number of Fish Collected by Species and Location
Species
Lake Trout
Slimy sculpin
Deepwater sculpin
Smelt
Coho salmon - adult
Coho salmon - yearling
Coho salmon - hatchery
Alewife > 120 mm
Alewife < 120 mm
Bloater chub > 160 mm
Bloater chub < 160 mm
Total Number of Individual
Fish Collected
1087
315
325
365
238
38
25
347
298
334
348
Number of Locations
3
3
3
3
79
22
1
3
3
3
3
Number of Composite
Samples Created
246
69
74
73
54
8
5
70
60
67
70
April 2004
2-15
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 2-10. Number of Fish Collected by Technique
Species
Lake Trout
Slimy sculpin
Deepwater sculpin
Smelt
Coho salmon - adult
Coho salmon - yearling
Coho salmon - hatchery
Alewife>120 mm
Alewife< 120 mm
Bloater chub > 160 mm
Bloater chub < 160 mm
Number of Fish Collected by Technique
Hook and Line
-
-
-
-
235
29
-
-
-
-
-
Gill Net
1029
25
-
25
3
-
-
-
-
-
-
Bottom Trawl
58
290
325
340
-
-
-
347
298
334
348
Harvest Weir
-
-
-
-
-
9
-
-
-
-
-
Dip Net
-
-
-
-
-
-
25
-
-
-
-
2.5 Sampling Methods
Full details of the sampling methods used in the LMMB Study have been published by EPA in a Methods
Compendium (USEPA, 1997d and USEPA, 1997e). A brief summary is provided below.
2.5.1 Atmospheric Components
Each shoreline site had a 10-meter meteorological tower and a number of meteorological instruments
including wind speed and wind direction sensors at a height of 10 m (Met-One, Grants Pass, OR), a solar
radiation sensor (Ll-Cor, model LI 200S, Lincoln, NE), temperature and relative humidity sensors
(Campbell Scientific, Logan, UT), and a standard Belfort rain gauge (Belfort Instrument, Baltimore, MD)
with a Nipher wind shield. All of the meteorological sensors were automatically recorded every 6
seconds using a datalogger (Campbell Scientific, model 2IX, Logan, UT).
2.5.17 Vapor Fraction
Airborne semivolatile organic contaminants, which include PCBs and trans-nonachlor, were collected
using high-volume (Hi-Vol) samplers for organics, modified to include an aluminum tube behind the
filter holder that accommodated a vapor trap consisting of a stainless steel cartridge of XAD-2® resin.
The samplers were operated for a 24-hour period at a flow rate of 34 m3 per hour. The frequencies of
sampling depended on the nature of the site (e.g., urban versus remote), as described in Section 2.4.1.
The samplers were checked each week for proper functioning, to collect a sample, or to set up for the next
collection. Samples from shoreline sites, except those collected at Sleeping Bear Dunes, were physically
composited to yield one sample per site representing the entire month. Samples from Sleeping Bear
Dunes were not composited, but analyzed separately. The results for each sample at Sleeping Bear Dunes
were mathematically composited to yield one result for the entire month.
The samples collected aboard the R/VLake Guardian were not composited, but were analyzed
individually, because there was at most one sample per station in any given month and not all stations
were sampled on every cruise. In addition, the results for a small number of samples could not be tied to
2-16
April 2004
-------
PCB/trans-Nonachlor Study Overview
specific locations or master stations (e.g., the samples were collected while the R/V Lake Guardian
steamed between stations). The locations of these samples are listed as "unrecorded."
2.5.1.2 Particulate Fraction
Airborne participate organic contaminants were collected on 20.3 x 25.4 cm pre-fired quartz fiber filters
using high-volume samplers for organics. The samplers were operated for a 24-hour period at a flow rate
of 34 m3 per hour. The frequencies of sampling depended on the nature of the site (e.g., urban versus
remote), as described in Section 2.4.1. The samplers were checked each week for proper functioning, to
collect a sample, or to set up for the next collection. Samples from shoreline sites, except those collected
at Sleeping Bear Dunes, were physically composited to yield one sample per site representing the entire
month. Samples from Sleeping Bear Dunes were not composited, but analyzed separately. The results
for each sample at Sleeping Bear Dunes were mathematically composited to yield one result for the entire
month.
The samples collected aboard the R/V Lake Guardian were not composited, but were analyzed
individually, because there was at most one sample per station in any given month and not all stations
were sampled on every cruise. In addition, the results for a small number of samples could not be tied to
specific locations or master stations (e.g., the samples were collected while the R/V Lake Guardian
steamed between stations). The locations of these samples are listed as "unrecorded."
2.5.13 Precipitation Fraction
A MIC sampler (Meteorological Instruments of Canada) with a 0.212 m2 stainless steel catch basin was
used for collecting precipitation samples for analysis of PCBs and fra«s-nonachlor. The sampler was
modified for all-weather operation by enclosing and insulating the space underneath the sampler. The
temperature in the enclosure was maintained at 10 to 15°C during the winter using a small space heater.
The collector also was fitted with a precipitation sensor and a retractable cover. The catch basin remained
covered to prevent evaporation until precipitation was detected by the sensor.
During sample collection, precipitation passed through a column containing XAD-2® resin that adsorbed
PCBs and fra«s-nonachlor in the precipitation sample. Glass wool plugs inserted on either side of the
XAD-2® resin trapped any particles in the sample. The sampler was in operation continuously for four-
week periods in order to collect the required 5 L of precipitation, and was checked each week to ensure
proper functioning and sample collection. The XAD columns were sealed with PTFE caps, transported to
the testing laboratory, and stored in air-tight containers at -18°C until analysis. In the laboratory, the
precipitation collection funnel was rinsed with water and wiped with a piece of clean quartz fiber filter
paper to remove adhering particles. The filter paper and rinsings were included as part of the sample.
2.5.14 Dry Deposition
Dry particle deposition was measured using deposition plates. Strips of Mylar® (approximately 5.7 cm by
1.8 cm) were coated with a layer of Apeizon® L grease about 5 |im thick to keep particles from bouncing
off the surface of the strip. This approach was designed to address concerns that traditional sampling
methods such as high volume samplers may underestimate the deposition of large particles (>10
The strips were attached to plates made from polyvinyl chloride (PVC). The leading edge of each PVC
plate was tapered at about a 10° angle to provide a laminar flow of air over the plate and reduce
turbulence. The deposition plates were mounted in an Eagle II automatic dry deposition collector at each
site. Each collector contained two deposition plates. A wind vane on the collector points the leading
April 2004 2-17
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
edge of the plates into the wind and a moisture sensor activates a motorized cover that protects the plates
from rain or snow. A timer monitored the total time that the plates were exposed.
The plates were exposed to dry deposition for periods ranging from 4 to 97 days. The total length of time
over which samples were collected represented 20 to 30% of the duration of the LMMB Study. The
longer exposures (e.g., 97 days) were required in rural areas where particle deposition rates were low.
Completed sample strips were weighed in the field, transferred to clean wide-mouth jars, frozen, and
shipped to the laboratory for analysis.
2.5.2 Tributaries
The number and timing of sampling events were dependent upon the stability of the tributaries and the
timing of increased flow events. Tributaries with greater stability (i.e., those that are less responsive to
precipitation events) were sampled less frequently than those that were more variable. Sampling was
timed to collect approximately one-third of the samples during base flow conditions (i.e., when flows
were below the 20th percentile of the historic flow regime) and approximately two-thirds of the samples
when flows were above the 20th percentile.
Tributary samples were collected as near to river mouths as possible without being subject to flow
reversals that are common near river mouths in Lake Michigan. Composite samples were obtained using
the USGS quarter-point sampling procedure. In this procedure, the stream is visually divided into three
equal flow areas. At the center of each flow area, samples were collected from 0.2 and 0.8 times the
depth. All six samples were then composited and pumped (using a peristaltic pump) through a 0.7-^m
glass fiber filter. Filters were removed, folded in quarters, and wrapped in aluminum foil for particulate
PCB analysis. The filtrate was then passed through a 250-g XAD-2® resin column at a flow rate of 500-
1000 mL per minute to trap dissolved organics. Samples were chilled and delivered to the testing
laboratory for analysis.
2.5.3 Open Lake
Open-lake samples were collected from various depths depending upon the stratification conditions.
During stratification, open-lake stations were sampled at the mid-epilimnion and mid-hypolimnion.
During non-stratified periods, samples were collected at mid-water column depth and two meters below
the surface. At master stations, during times of non-stratification, were sampled at mid water column, one
meter below the surface, and two meters off the bottom. During times of stratification, master stations
were sampled at one meter below the surface, mid-epilimnion, mid-hypolimnion, and two meters off the
bottom. In addition, Stations 18M and 41 were sampled at the thermocline and 2 meters below the
surface during stratification.
Water samples were collected using an onboard sampling and filtration system on the R/VLake Guardian.
Samples were drawn with a peristaltic pump through a hose held overboard at a specific depth. A volume
of water ranging from 100 to 1000 L was pumped through a "Pentaplate" filtration apparatus holding up
to five glass-fiber filters in parallel. The filtered water was discharged into glass carboys and
subsequently pumped through a column containing 250 g of XAD-2® resin at a flow rate of 500-1000 mL
per minute to trap dissolved organics. The total volume of water passed through the filtration and XAD-
2® systems was measured in the glass carboys and recorded.
After sampling was complete, the filters and the XAD-2® resin column were removed from the sampling
apparatus. The filters were frozen onboard the ship, and transferred to the laboratory for analysis while
still frozen. The resin columns were cooled to 4°C and transferred to the laboratory for analysis.
2-18 April 2004
-------
PCB/trans-Nonachlor Study Overview
2.5.4 Sediment
Sediment samples were collected using three types of equipment. A modified Soutar box corer was used
to retrieve cores approximately 60 cm in length, with well-preserved sediment-water interfaces. Samples
also were collected using a gravity corer and a Ponar dredge, depending on the nature of the lake bottom
at the collection point. Surficial sediment samples were collected from the cores by sectioning the cores
into intervals ranging from 0.5 to 1.5 cm thick. The sections were transferred to glass containers, frozen
onboard the ship, and transferred to the laboratory while still frozen.
Sediment trap samples were collected and split while still wet. The 60-mL trap bottles were allowed to
settle for 25 hours, under refrigeration. The overlying water (-25 mL) was poured off into a beaker, and
the remaining trap sample was poured into the splitter reservoir through a 700-(im screen. The sample
was split into four subsamples and washed through the splitter using the overlying water in the beaker and
distilled water. The subsamples were transferred to glass containers, frozen onboard the ship, and
transferred to the laboratory while still frozen.
2.5.5 Lower Pelagic Food Web Organisms
Phytoplankton were collected using a device called a phytovibe. This device was specially designed and
constructed for GLNPO for collecting large volumes of plankton for analysis of chemical contaminants
such as mercury and PCBs. The phytovibe consists of a pair of inverted pyramids constructed of stainless
steel mesh lined with 10-^m Nitex netting. Water is pumped by a submersible pump through nylon
tubing into the top of the device, which has an opening that is 1 m2. The end of the nylon tubing is
covered with 100-^m netting to remove zooplankton. In order to prevent plugging of the netting with
plankton, the phytovibe is shaken by a motor. The samples were washed down into a detachable
sampling cup with lake water and collected for processing. Sampling times ranged from 6 to 14 hours,
depending on plankton concentration in the water and sample size needed for a particular analysis.
The depth of collection was chosen based on interpretations of the temperature, fluorescence, and
bathymetric profiles from the ship, with the objective of choosing a depth that maximized the occurrence
of phytoplankton that were being grazed. This generally corresponded to the epilimnion or the
subthermocline chlorophyll maximum in stratified conditions.
Zooplankton were collected in nested Nitex nets of two different mesh sizes (102-(im and 500-(im) during
standard vertical tows, from near the bottom to the surface. The 500-^m nets were used to exclude larger
organisms, including small fish, from the zooplankton samples. The number of tows performed was
dependent on the mass of sample collected per tow. Diporeia spp. were collected in benthic tows, and
My sis spp. were collected in vertical and benthic tows.
Plankton samples were transferred to glass containers, frozen onboard the ship, and transferred to the
laboratory while still frozen.
2.5.6 Fish
Whole fish were collected intact, with all body fluids and no incisions, except lake trout, which had their
stomachs removed. Fish were wrapped in aluminum foil, placed in polyethylene bags, tagged, and frozen
onboard the vessel. The fish were aged by checking for coded wire tags on the head and for fin clips.
Fish were then composited by age, location, species, and size range. Samples were homogenized at the
laboratory using a 40-qt vertical cutter mixer for large fish, 12-qt Stephan Machinery vertical cutter for
medium-sized fish, or a high-speed 2-qt Robot Coupe for small fish. Additional information on the
numbers offish collected and the composite samples that were created can be found in Section 2.4.6.
April 2004 2-19
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
2.6 Analytical Methods
Full details of the analytical methods used in the LMMB Study have been published by EPA in a methods
compendium (USEPA, 1997d and USEPA 1997e). The general considerations in the choice of methods
are discussed in Section 2.6.1. The convention for the data presented in this report is discussed in Section
2.6.2, and brief summaries of the specifics of the analyses for each lake component are provided in
Sections 2.6.3 to 2.6.7.
2.6.1 General Analytical Considerations
Three significant, and interrelated aspects of the analytical methods for the PCBs and trans-nonachlor
used in this study were the selection of analytical instrumentation, the selection of specific PCB
congeners of interest, and the provision of a common set of analytical standards to all of the investigators.
The vast majority of the commonly used analytical techniques for PCB analysis employ gas
chromatography as a means of separating the PCB congeners. Traditionally, methods for the analysis of
Aroclor mixtures have employed an electron capture detector (ECD) and produce chromatographic data
that require that the analyst look for patterns of chromatographic peaks that resemble those in an authentic
standard of each Aroclor. However, as a result of the changes that can occur in environmental samples
(e.g., "weathering"), the pattern of peaks present in an environmental sample often changes enough to
make it difficult or impossible to identify a pattern attributable to a specific Aroclor mixture.
Given these difficulties with Aroclor analysis, the most common alternative is to perform analyses for the
individual PCB congeners. At present, there are no gas chromatography columns that can completely
separate all 209 individual PCBs. Many of the congeners pass through the gas chromatograph at the same
time, a phenomenon called "coelution." Using an electron capture detector, it is not possible to
distinguish between two PCB congeners that coelute. Using a mass spectrometer as a detector permits the
analyst to distinguish some coeluting congeners because the mass spectrometer can differentiate the
masses of the individual compounds. Therefore, if two congeners coelute but have different molecular
weights (e.g., a tetrachlorobiphenyl and a pentachlorobiphenyl), the mass spectrometer can distinguish
between them. However, even the combination of gas chromatography and mass spectrometry (GC/MS)
cannot separate all 209 congeners. The costs of GC/MS analyses are substantially higher than those for
GC/ECD. For this reason and others described below, GC/ECD was chosen as the primary analytical
technique for the PCBs. GC/ECD also is applicable to the analysis of fra«s-nonachlor, and the trans-
nonachlor results often were determined from the same analyses as the PCB congeners. The combination
of gas chromatography and negative ion chemical ionization mass spectrometry was used for the analysis
of PCB congeners in fish tissues.
A critical aspect of any analytical procedure is the availability of authentic standards of the analytes of
interest. PCB analyses present a significant challenge in this regard because Aroclors were produced on
an industrial scale as mixtures of PCB congeners, with some degree of variation between production lots
over time. The commercial production of those Aroclors was outlawed over 20 years ago, but EPA has
issued a small number of permits for the manufacture of Aroclors for use as analytical standards.
Although the manufacturers of the standards go to great lengths to ensure that their standards resemble the
original commercial mixtures, some differences are expected, and those differences can affect the end
results of the analyses.
In order to address the concerns about methods and standards for the PCBs, EPA developed an approach
to selecting appropriate PCB congeners for analysis and for providing common analytical standards to all
of the investigators in the LMMB. Michael Mullin of EPA developed a mixture of specific proportions of
Aroclors 1232, 1248, and 1262 that was designed to contain all of the Aroclor-derived congeners that are
2-20 April 2004
-------
PCB/trans-Nonachlor Study Overview
likely to be found in environmental samples. That mixture is known as the "Mullin mix" and was
provided to all of the investigators in the LMMB for use as a common reference standard.
In addition, information on the specific PCB congeners contained in the Mullin mix was used to focus the
LMMB PCB analyses on those congeners that represent a significant portion of mass of Aroclors that
may have been released into the environment. This is a critical aspect of the LMMB Study. Although the
toxic effects of PCBs are certainly of interest in Lake Michigan, the purpose of the LMMB Study is to
develop a mass balance for the anthropogenic pollutants described in Chapter 1. Therefore, the
investigators focused on PCB congeners that occur in relatively high concentrations in the Aroclors that
may have been released into the environment.
The majority of the toxic effects of PCB exposure are associated with the 12 WHO toxic congeners.
However, not all of these congeners can be readily separated by GC methods and most of the toxic
congeners are present in environmental samples at concentrations that require very sensitive analytical
methods (e.g, high resolution mass spectrometry). Data on the composition of congeners in the Mullin
mix were used to select those congeners the represent a significant portion of the mass of Aroclors 1232,
1248, and 1262 and that the investigators were able to detect reliably. Table 2-11 contains a list of the 45
PCB congeners that constitute at least 1% of the mass of the Mullin mix of the three Aroclors.
Collectively, these 45 congeners contribute 77.31% of the mass of the Mullin mix. There are 77
additional congeners that have been identified in the Mullin mix, each of which contributes 0.01% to
0.99% of the total mass. Collectively, these 77 congeners contribute 22.69% of the mass of the Mullin
mix.
Data for the Mullin mix indicate that 10 of the 12 WHO toxic congeners are present in the mix at
concentrations that could be measured. Seven of those 10 toxic congeners can be separated from all the
other congeners and represent 1.17% of the mass of the Mullin mix. PCB 118 is the toxic congener with
the highest concentration, representing 0.6% of the mass in the Mullin mix. Collectively, the 10 toxic
congeners that are present and the other 5 coeluting congeners represent 5.50% of the mass.
The data in Table 2-11 demonstrate why looking for just the toxic PCB congeners would make it difficult
to develop a robust mass balance for Lake Michigan, since such an approach would ignore 95 to 99% of
the mass of PCBs present in Aroclors.
The use of the Mullin mix as a common analytical standard across all of the laboratories provides an
important control mechanism in the analysis of the PCBs and provided a mechanism with which to select
a basic set of PCB congeners of interest for the LMMB Study. Differences in the study requirements and
specific instruments used in each laboratory, matrix-specific sample preparation and extraction
techniques, and other factors enabled each of the laboratories to report results for additional PCB
congeners. However, different numbers of additional congeners or coeluting groups of congeners being
reported in each of the laboratories. The maximum number of congeners reported for each laboratory is
shown in Table 2-12.
The "total" PCB concentrations also were determined for each sample in the study. The totals are simply
the sum of all of the PCB congeners that were found in the sample. However, each PI determined the
total PCB concentrations based on the specific congeners determined in that laboratory and some
adjustments were made when interferences from non-PCB analytes were apparent.
April 2004 2-21
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 2-11. PCB Congeners that Contribute at Least 1% of the Mass of the Mullin Mix
PCB Congener
PCB-008+005
PCB-001
PCB-031+028
PCB-003
PCB-066
PCB-180
PCB-044
PCB-052
PCB-132+153+105
PCB-203+196
PCB-015+017
PCB-018
PCB-056+060
PCB-201
PCB-033
PCB-070+076
PCB-022
PCB-187+182
PCB-163+138
PCB-004+010
PCB-123+149
PCB-174
PCB-041+071
PCB-049
PCB-074
PCB-032
PCB-016
PCB-064
PCB-194
PCB-095
PCB-170+190
Concentration in Mullin Mix (ng/mL)
15
14
12
8.5
7.1
6.6
5.0
5.0
4.8
4.6
4.5
4.4
4.4
4.4
4.2
3.9
3.6
3.6
3.5
3.4
3.3
3.2
2.8
2.8
2.6
2.5
2.4
2.2
2.1
2.0
2.0
% of Total Mass in Mullin Mix
7.71
7.20
6.17
4.37
3.65
3.39
2.57
2.57
2.47
2.36
2.31
2.26
2.26
2.26
2.16
2.00
1.85
1.85
1.80
1.75
1.70
1.64
1.44
1.44
1.34
1.29
1.23
1.13
1.08
1.03
1.03
+ indicates congeners that typically coelute
2-22
April 2004
-------
PCB/trans-Nonachlor Study Overview
Table 2-12. Maximum Number of PCB Congeners Reported, by Laboratory
Laboratory
Battelle, Sequim
Indiana University
University of Wisconsin, Wisconsin State Lab of Hygiene
University of Minnesota
National Oceanic and Atmospheric Administration
Rutgers University
Illinois State Water Survey
United States Geological Survey, Biological Resources Discipline
Maximum Number of PCB Congeners or Groups
of Coeluting Congeners
105
108
65 (78)*
110
105
86
107
80
The figure in parentheses represents the number of congeners reported in analyses conducted after a change in the laboratory's
standard operating procedure that was instituted in November 1994.
2.6.2 Data Presented in this Report
As noted in Section 2.1, there are 209 possible PCB congeners, and the investigators in this study
reported results for 65 to 110 of these congeners, depending on the capabilities of each laboratory (Table
2-12). Given the hundreds of samples that were collected for some media, it is impractical to present all
of the PCB congener results for any of the media in this report, let alone the entire LMMB Study. This
report seeks to strike a balance between the depth of the data presentation and a desire to limit the report
to a manageable size. Therefore, except as noted below, throughout the remainder of this report,
summaries are presented for the results for the following analytes:
• PCB congener 33 (2',3,4-trichlorobiphenyl)
• PCB congener 118 (2,3',4,4',5-pentachlorobiphenyl)
• PCB congener 180 (2,2',3,4,4',5,5'-heptachlorobiphenyl)
• Total PCBs
• fra«5-nonachlor
The three PCB congeners were selected for presentation for the following reasons:
These congeners do not coelute with any other congeners that were abundant in the Aroclors in any of
the LMMB media (thus allowing presentation and data interpretation across all media in the LMMB
report),
• They were reported by all but one of the investigators in the LMMB, thus enabling direct
comparisons across all media (except dry deposition),
• They represent different levels of chlorination of the parent biphenyl structure and thus should
represent a variety of environmental fates for PCBs,
• None were used as surrogates in the analysis,
• All three congeners are present in the "Mullin mix" standard of Aroclors used for the LMMB Study,
and
PCB 118 represents one of the coplanar congeners currently deemed "toxic" because of its biological
activity, while PCBs 33 and 180 currently are not classified in the "toxic" group.
The only exception to this approach occurs in Chapter 6, on the sediment analyses, where the principal
investigator found that PCB 33 coeluted with the organochlorine pesticide heptachlor. As a result, the
presence of heptachlor in the sediment samples lead to high recoveries of PCB 33 in matrix spike
April 2004 2-23
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
samples. Therefore, the principal investigator provided detailed graphics for the coeluting PCB congener
pair 28+31 and not PCB 33. The coelution with heptachlor was not evident in the matrix spike recoveries
from samples of other matrices, and other matrices would not be expected to retain heptachlor from
historical sources to the extent that may occur in sediments.
2.6.3 Atmospheric Components
The preparation and analysis of three of the four fractions of the atmospheric samples followed similar
procedures. Soxhlet extraction was used to extract the PCBs and trans-nonachlor from:
• XAD-2® resin used to collect the vapor phase samples,
Glass fiber filters used to collect the particulate samples, and
• XAD-2® resin used to collect the precipitation samples.
Hexane:acetone (50:50) was used for all three Soxhlet extractions and the extracts were concentrated by
rotary evaporation.
For all fractions, interfering compounds were removed and analytes were fractionated with silica gel. The
hexane fraction contained all of the PCBs and the second fraction (40% dichloromethane, 60% hexane)
contained fra«s-nonachlor. PCBs and fra«s-nonachlor were determined by capillary gas chromatography
with electron-capture detection.
The dry deposition samples were prepared by extracting the Mylar® strips with dichloromethane and
hexane in an ultrasonic bath. The extracts were subjected to a cleanup procedure that employed
deactivated silica gel and deactivated alumina which removed most of the Apeizon® L grease from the
extracts. PCBs and fra«s-nonachlor were determined by capillary gas chromatography with electron-
capture detection. Because some grease remained in the concentrated extracts, a 1-m glass capillary pre-
column was attached to the front end of the GC column and the remaining grease was deposited on the
pre-column. Half of the 1-m pre-column was cut off after 3 to 5 sample injections and the pre-column
was replaced after 6 to 10 injections. Combined with frequent maintenance and replacement of the glass
injector liners in the GC, the pre-column prevented the residual grease from compromising the resolution
of the GC column.
To minimize the potential effects of sorption of PCBs and fra«s-nonachlor from the gas phase onto the
grease strips, the deposition of PCBs onto the field blank associated with each sample was used to correct
the sample results, based on the mass of PCBs normalized to the surface area of the sample.
2.6.4 Tributaries and Open Lake
The XAD-2® resins and glass fiber filters used to collect tributary samples were extracted separately with
a sequence of Soxhlet extractions, followed by volume reductions and rinses. Florisil and silica gel
column chromatography were used to cleanup tributary sample extracts prior to analysis, allowing
fractionation of the PCBs from fra«s-nonachlor and the majority of the other organochlorine pesticide
compounds in the samples. Following this fractionation, PCB congeners and fra«s-nonachlor were then
determined by the analyses of separate extracts using capillary gas chromatography with electron-capture
detection.
The XAD-2® resins and glass fiber filters used to collect open-lake samples were extracted separately
with a sequence of Soxhlet extractions, followed by volume reductions and rinses, in a fashion similar to
that for the tributary samples. For the open-lake samples, cleanup procedures included the use of
acidified silica gel and alumina to remove polar interferences from the sample extracts. Subsequent to the
2-24 April 2004
-------
PCB/trans-Nonachlor Study Overview
start of the study, the laboratory performing the analyses of the open-lake samples adjusted their extract
cleanup procedures and included a second acidified silica gel cleanup, in an attempt to overcome potential
interferences that were present on the XAD-2® resin used to collect water samples. As a result of those
changes, the laboratory did not obtain separate cleanup fractions containing the PCBs and trans-
nonachlor, but analyzed the sample extracts on two dissimilar GC columns (DB-5 and DB-1701). While
the DB-1701 column provided clear chromatographic separation of any fra«s-nonachlor in the sample,
this analyte coeluted with PCB 99 on the DB-5 column. As a result of the potential coelution, the
reported concentrations of PCB 99 in open-lake samples are probably biased by any fra«s-nonachlor
present in the samples.
2.6.5 Sediment
Approximately 15-30 g of thawed wet sediment was extracted with dichloromethane in an ultrasonic bath
held at 30°C. The extracts were dried over sodium sulfate and passed through a cleanup column. PCBs
were separated from pesticides and polyaromatic hydrocarbons by column chromatography fractionation.
PCBs and fra«s-nonachlor were determined by capillary gas chromatography with electron-capture
detection.
2.6.6 Lower Pelagic Food Web Organisms
Lower pelagic food web organism samples were thawed, rinsed with methanol, then extracted by a
Soxhlet procedure (four hours with methanol; then 16-24 hours with dichloromethane). The methanol
fraction was back-extracted with hexane. The hexane fraction containing the PCBs was combined with
the dichloromethane fraction and concentrated by rotary evaporation. Lipids were removed from the
extracts, and the extracts were cleaned up and eventually reduced to 200-300 [iL prior to analysis. PCBs
were determined by capillary gas chromatography with electron-capture detection. fra«s-Nonachlor was
analyzed by electron capture negative ionization (ECNI) gas chromatography/mass spectrometry
(GC/MS) with selected ion monitoring (SIM), using methane as the reagent gas.
2.6.7 Fish
Homogenized fish tissue samples were thawed and then extracted with a 90/10 mixture of petroleum
ether/ethyl acetate. Extracts were concentrated and the lipid content was determined from an aliquot of
the concentrated extract. The extracts were then prepared for analysis by separation and removal of lipids
using a gel permeation chromatography system. PCBs and fra«s-nonachlor were determined by electron
capture negative ionization (ECNI) gas chromatography/mass spectrometry (GC/MS) with selected ion
monitoring (SIM), using methane as the reagent gas.
2.7 Quality Implementation and Assessment
As described in Section 1.5.5, the LMMB quality assurance (QA) program prescribed minimum standards
to which all organizations collecting data were required to adhere. The goal of the QA program was to
ensure that all data gathered during the LMMB Study met defined standards of quality with specified
levels of confidence. Data quality was defined, controlled, and assessed through activities that included
development of study QAPPs, use of SOPs, and data verification. These activities are described in detail
in The Lake Michigan Mass Balance Study Quality Assurance Report (USEPA, 200 Ib).
April 2004 2-25
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Specific quality control elements implemented in the sampling and analysis of PCB and fra«s-nonachlor
included:
use of standard operating procedures and trained personnel for field sampling and laboratory analysis;
• determination of method sensitivity through calculation of method detection limits;
preparation and analysis of a variety of blanks to characterize contamination associated with specific
sample handling, storage, and analysis processes including field blanks, lab reagent blanks, bottle
blanks, trip blanks, and lab solvent blanks;
collection and analysis of field or laboratory duplicate samples;
• preparation and analysis of a variety of quality control samples including standard reference samples
and performance standards;
• use of internal and surrogate standards for all field samples;
use of a standardized data reporting format; and
• preparation and analysis of matrix spike samples to characterize the applicability of the analytical
method to the study sample matrices.
In addition, each researcher's laboratory was audited during an on-site visit at least once during the time
LMMB samples were being analyzed. The auditors reported positive assessments and did not identify
issues that adversely affected the quality of the data. Prior to data submission, each researcher submitted
electronic test files containing field and analytical data according to the LMMB data reporting standard.
GLNPO reviewed these test data sets for compliance with the data reporting standard and provided
technical assistance to the researchers.
An intercomparison study for the PCB analyses was conducted among all Pis and the EPA ORD
laboratory in Duluth, Minnesota. A detailed discussion of the intercomparison study is provided in The
Lake Michigan Mass Balance Study Quality Assurance Report (USEPA, 200 Ib).
Because all of the Pis were analyzing different types of samples, such as vapor, fish, and sediment, and
employing sample-specific preparation procedures, the intercomparison study focused on the instrumental
analysis portion of the study. Each of the Pis was provided with two solutions of PCB congeners
prepared by Ultra Scientific in consultation with Dr. Mike Mullin of EPA ORD. The solutions included
PCB congeners with retention times and response factors covering the ranges observed in environmental
samples as well as congeners that are known to be difficult to resolve.
The results submitted by each PI were evaluated by comparison to: 1) the gravimetric true value of the
solution as prepared and 2) the 95% confidence intervals for specific congeners based on the full set of
results from the study. A goal of 30 percent bias from the true gravimetric mean was established for the
intercomparison study.
This goal was not always met. Comparison of the results from each PI to the gravimetric true mean
indicated a high bias for some congeners and low bias for others. With few exceptions, results that
exhibited more than 30% bias from the true value were associated with coeluting congeners. The results
from all of the Pis were within the 95% confidence intervals calculated for the PCB congeners, indicating
that none of the results were likely to be outliers. However, the relatively low number of degrees of
freedom (i.e., eight laboratory participants) limit the statistical power of this evaluation.
The variability of the analytical results among Pis also was evaluated by calculating the relative standard
deviation (RSD) among results for each congener or group of coeluting congeners. Although the RSD
values for the majority of the congeners were less than 30%, the results for congeners 37, 77, 77+110, 81,
and 123+149 in one of the two solutions in the intercomparison study exceeded 30% RSD. Coelution is a
problem for all of these six congeners. However, these congeners represent only a small portion of the
2-26 April 2004
-------
PCB/trans-Nonachlor Study Overview
total mass of PCBs. Only the coeluting pair of 123+149 is present in the Mullin mix at greater than 1%,
and PCB-77 and PCB-81 represent only 0.15% and 0.08% of the Mullin mix, respectively. Therefore, the
RSD results greater than 30% for these congeners in the intercomparison study are not likely to have a
large influence on the study goal of developing a mass balance of PCBs for Lake Michigan.
Overall, the results of the intercomparison study suggest the Pis were successfully employing analytical
procedures consistent with what is expected for determination of PCB congeners in environmental
samples and did not indicate any major analytical difficulties associated with any of the labs in the study.
Prior to sample collection, Quality Assurance Project Plans (QAPPs) were developed by the Pis and
submitted to GLNPO for review. In the QAPPs, the Pis defined MQOs in terms of six attributes:
sensitivity, precision, accuracy, representativeness, completeness, and comparability. The MQOs were
designed to control various phases of the measurement process and to ensure that the total measurement
uncertainty was within the ranges prescribed by the DQOs. The MQOs for PCBs and trans-nonachlor are
listed in Section 7 of The Lake Michigan Mass Balance Study Quality Assurance Report (USEPA,
2001b).
The Pi-defined MQOs also were used in the data verification process. GLNPO conducted data
verification through the LMMB QA Workgroup. The workgroup was chaired by GLNPO's Quality
Assurance Manager and consisted of quality control coordinators that were responsible for verifying the
quality of specific data sets. Data verification was performed by comparing all field and QC sample
results produced by each PI with their MQOs and with overall LMMB Study objectives. If the results
failed to meet MQOs and corrective actions were not feasible, the results were flagged to inform data
users of the failure. These flags were not intended to suggest that data were not useable; rather they were
intended to caution the user about an aspect of the data that did not meet the predefined criteria. In
addition, a wide variety of flags were applied to the data to provide detailed information to data users.
For example, the flag LAC (laboratory accident, no result reported) was applied to sample results to
document that a sample was collected, but no result was reported due to a laboratory accident. The
frequencies of flags applied to PCB and fra«s-nonachlor study data are provided in the Quality
Implementation Sections of each of the following chapters. The flag summaries include the flags that
directly relate to evaluation of the MQOs to illustrate some aspects of data quality, but do not include all
flags applied to the data to document sampling and analytical information (such as LAC). In order to
provide detailed quality information to data users, the study data are maintained in the GLENDA database
with all applied flags. Detailed definitions of the flags can be found in the Allowable Codes Table on
GLNPO's web site atwww.epa.gov/glnpo under Result Remark, List of QC flags (lab_rmrk).
Comparability of the data from the various Pis was enhanced by the use of standardized reporting units
for samples from similar matrices. In addition, the analyses of PCBs incorporated the use of surrogate
compounds that were added to the sample before extraction. These surrogates provide an estimate of the
efficiency of the entire sample preparation, extraction, cleanup, and analysis process, as applied to each
sample. The recoveries of the surrogates were used by most of the investigators to correct the results of
the target PCBs for any apparent losses or gains of the surrogates. The dry deposition PCB data were not
corrected for surrogate recoveries, but were corrected for the potential contribution of PCBs absorbed into
the Apeizon® L grease directly from the vapor phase, using field blank results.
For fra«5-nonachlor analyses, the Pis had difficulty identifying a surrogate compound that worked well
for those matrices analyzed by gas chromatography with electron-capture detection (GC/ECD). The
analyses of lower pelagic food web organisms and fish were performed by GC/MS techniques, which
permitted the use of a stable 13C-labeled form of fra«s-chlordane as a surrogate. Therefore, the lower
pelagic food web and fish results were corrected for surrogate recoveries.
April 2004 2-27
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Of the GC/ECD analyses for fra«s-nonachlor, only the data from the atmospheric samples collected at
Sleeping Bear Dunes were corrected for surrogate recoveries. As with the PCB results, the dry deposition
fra«5-nonachlor data were not corrected for surrogate recoveries, but were corrected for the potential
contribution of fra«s-nonachlor absorbed into the Apeizon® L grease directly from the vapor phase, using
field blank results.
The Pis participating in the study also conducted real-time data verification. Pis applied best professional
judgement during sampling, analysis, and data generation, based on their experience monitoring PCB and
fra«5-nonachlor in the environment. In most cases, when sample results were questionable, the PI
reanalyzed the sample or clearly documented the data quality issues in the database through the
application of data quality flags or by including comments in the database field, "Exception to Method,
Analytical." Because the flags and comments are maintained in the database for each sample result, data
users are fully informed of data quality and can evaluate quality issues based on their intended use of the
data. The level of documentation that GLNPO is maintaining in the study database is unprecedented for a
database of this size and will serve as a model for future efforts.
GLNPO also conducted data quality assessments in terms of four of the six attributes used as the basis for
the MQOs, specifically sensitivity, precision, and bias. For example, system precision was estimated as
the mean relative percent difference (RPD) between results for field duplicate pairs. Similarly, analytical
precision was estimated as the mean RPD between results for laboratory duplicate pairs. Bias was
estimated using the mean recovery of spiked field samples or other samples of known concentration such
as laboratory performance standards. A summary of data quality assessments is provided for the PCB and
fra«5-nonachlor study data in the Quality Implementation Section of each of the following chapters.
2-28 April 2004
-------
Chapter 3
PCBs/trans-Nonachlor in Atmospheric Components
3.1 Results
Atmospheric samples were collected from November 1993 to October 1995 at nine sampling stations on
the shoreline of Lake Michigan, three stations outside of the Lake Michigan basin, 17 stations in the open
lake (over-water), and two stations in Green Bay, and analyzed for PCB congeners (Table 3-1) and trans-
nonachlor (Table 3-2). Results for "total PCBs" were determined by summing the results for the
individual PCB congeners. These analytes were measured in four separate atmospheric media or phases:
vapor (pg/m3), particulates (pg/m3), precipitation (pg/L), and dry deposition (pg/m2). In total, 306 vapor-
phase samples, 235 particulate samples, 209 precipitation samples, and 42 dry deposition samples were
collected and analyzed for PCBs. For fra«s-nonachlor, 303 vapor-phase samples, 220 particulate
samples, 206 precipitation samples, and 40 dry deposition samples were collected and analyzed. Eighteen
of the particulate samples analyzed for PCBs and fra«s-nonachlor were collected while the R/VLake
Guardian was steaming between stations. Because these samples were not collected at fixed locations,
they are considered spatial composites and are listed as such in Tables 3-1 and 3-2.
As noted in Chapter 2, there are 209 possible PCB congeners, and the investigators in this study reported
results for 65 to 110 of these congeners, depending on the capabilities of each laboratory. For the
purposes of this report, we are presenting summaries of the results for the following subset of analytes:
• PCB congener 33
• PCB congener 118
• PCB congener 180
• Total PCBs
• fra«5-nonachlor
Apr/72004 3-1
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 3-1. Number of Atmospheric Samples Collected and Analyzed for PCB Congeners and Total PCBs
Sampling Station
Shoreline
Atmospheric
Stations
Out-of- Basin
Atmospheric
Stations
Over-Water
Atmospheric
Stations
Beaver Island
Chicago SWFP
Crib Intake
Chiwaukee Prairie
III Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear
Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Spatial composites
Empire Michigan
GB17
GB24M
Harrison Crib
1
5
6
110
Sampling
Dates
03/1 5/94 to
09/26/95
07/0 1/94 to
10/03/95
03/1 5/94 to
10/02/95
12/06/93 to
10/02/95
03/1 5/94 to
10/20/95
03/1 2/94 to
10/08/95
03/1 5/94 to
10/13/95
11/23/93 to
10/07/95
11/23/93 to
10/08/95
03/1 5/94 to
10/20/95
04/0 1/94 to
10/08/95
04/0 1/94 to
07/31/94
04/30/94 to
10/11/95
04/0 1/94 to
07/29/94
04/1 2/95 to
04/12/95
10/1 7/94 to
09/20/95
08/1 1/94 to
08/11/94
05/1 0/94 to
10/11/95
05/1 1/94 to
10/10/95
08/25/94 to
10/12/95
04/08/95 to
09/23/95
Vapor
Samples
Analyzed
19
0
21
25
31
19
18
35
22
21
19
8
0
10
0
4
0
5
6
4
3
Part icu late
Samples
Analyzed
18
0
21
25
30
19
16
15
19
21
18
4
18
4
0
2
0
1
3
1
0
Dry Deposition
Samples
Analyzed
0
9
0
13
0
0
0
8
11
0
0
0
0
0
0
0
1
0
0
0
0
Precipitation
Samples
Analyzed
20
0
20
17
21
20
20
16
21
21
19
4
0
4
1
1
0
1
1
0
0
Total
Samples
Analyzed
57
9
62
80
82
58
54
74
73
63
56
16
18
18
1
7
1
7
10
5
3
3-2
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Sampling Station
Over-Water
Atmospheric
Stations
18M
23M
27M
280
310
380
40M
41
47M
MB19M
11M
Sampling
Dates
05/06/94 to
10/09/95
05/04/94 to
10/03/95
05/02/94 to
09/27/95
10/26/94 to
10/01/95
03/28/95 to
10/08/95
10/31/94 to
01/23/95
10/1 8/94 to
09/25/95
04/30/94 to
08/12/94
08/07/94 to
09/19/95
01/24/95 to
01/24/95
05/08/94 to
05/08/94
Total
Vapor
Samples
Analyzed
5
4
5
4
3
1
4
2
5
1
2
306
Part icu late
Samples
Analyzed
0
0
0
0
0
0
0
0
0
0
0
235
Dry Deposition
Samples
Analyzed
0
0
0
0
0
0
0
0
0
0
0
42
Precipitation
Samples
Analyzed
0
1
0
0
0
1
0
0
0
0
0
209
Total
Samples
Analyzed
5
5
5
4
3
2
4
2
5
1
2
792
April 2004
3-3
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 3-2. Number of Atmospheric Samples Collected and Analyzed for frans-Nonachlor
Sampling Station
Shoreline
Atmospheric
Stations
Out-of- Basin
Atmospheric
Stations
Over-Water
Atmospheric
Stations
Beaver Island
Chicago SWFP
Crib Intake
Chiwaukee Prairie
III Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear
Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Spatial composites
Empire Michigan
GB17
GB24M
1
5
6
110
18M
Sampling
Dates
03/1 5/94 to
09/26/95
07/0 1/94 to
10/03/95
03/1 5/94 to
10/02/95
12/06/93 to
10/02/95
03/1 5/94 to
10/20/95
03/1 2/94 to
10/08/95
03/1 5/94 to
10/13/95
11/23/93 to
10/07/95
11/23/93 to
10/08/95
03/1 5/94 to
10/20/95
04/0 1/94 to
10/08/95
04/0 1/94 to
07/31/94
04/30/94 to
10/11/95
04/0 1/94 to
07/29/94
04/1 2/95 to
04/12/95
10/1 7/94 to
09/20/95
05/1 0/94 to
10/11/95
05/1 1/94 to
10/10/95
08/25/94 to
10/12/95
04/08/95 to
09/23/95
05/06/94 to
10/09/95
Vapor
Samples
Analyzed
18
0
21
25
31
19
18
34
22
21
18
8
0
10
0
4
5
6
4
3
5
Part icu late
Samples
Analyzed
17
0
21
25
29
18
16
4
18
21
18
4
18
4
0
2
1
3
1
0
0
Dry Deposition
Samples
Analyzed
0
9
0
13
0
0
0
8
10
0
0
0
0
0
0
0
0
0
0
0
0
Precipitation
Samples
Analyzed
20
0
20
17
21
20
19
14
21
21
19
4
0
4
1
1
1
1
0
0
0
Total
Samples
Analyzed
55
9
62
80
81
57
53
60
71
63
55
16
18
18
1
7
7
10
5
3
5
3-4
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Sampling Station
Over-Water
Atmospheric
Stations
23M
27M
280
310
380
40M
41
47M
MB19M
11M
Sampling
Dates
05/04/94 to
10/03/95
05/02/94 to
09/27/95
10/26/94 to
10/01/95
03/28/95 to
10/08/95
10/31/94 to
01/23/95
10/1 8/94 to
09/25/95
04/30/94 to
08/12/94
08/07/94 to
09/19/95
01/24/95 to
01/24/95
05/08/94 to
05/08/94
Total
Vapor
Samples
Analyzed
4
5
4
3
1
4
2
5
1
2
303
Part icu late
Samples
Analyzed
0
0
0
0
0
0
0
0
0
0
220
Dry Deposition
Samples
Analyzed
0
0
0
0
0
0
0
0
0
0
40
Precipitation
Samples
Analyzed
1
0
0
0
1
0
0
0
0
0
206
Total
Samples
Analyzed
5
5
4
3
2
4
2
5
1
2
769
3.1.1 Vapor Fraction
Vapor-phase PCB congeners were detected in the vast majority of the samples collected from all LMMB
Study stations. Vapor-phase PCB 33 was detected in all but two samples (Table 3-3), vapor-phase PCB
118 was detected in all but one sample (Table 3-4), and vapor-phase PCB 180 was detected in all but five
samples (Table 3-5). Tables 3-3 to 3-6 present the results for the monthly composite vapor samples from
this study. As discussed in Chapter 2, the composite results represent either: 1) the physical compositing
of several individual samples collected during a calendar month to create one sample for analysis, or 2)
mathematical composites of the results from the analysis of the individual samples collected over a
calendar month. In some instances, both physical and mathematical composites were prepared within a
month. In these instances, the reported result is a mathematical composite based on both the physical
composite samples and the individual samples. The total number of composite results is shown for each
station as "N," along with the mean concentration, range, standard deviation, and relative standard
deviation (RSD). Tables 3-3 to 3-6 also indicate the percent of the individual sample results that were
below the sample-specific detection limit, (as opposed to the percent of the composite results). The mean
concentrations were calculated using the results reported by each laboratory (substitution of the detection
limit or other value was not used for results below the sample-specific detection limits).
Monthly composite vapor-phase PCB congener concentrations ranged from 0 pg/m3 for PCB 180 at the
Sleeping Bear Dunes and Brule Rivers stations and open-water station 27M, to 290 pg/m3 for PCB 33 at
the IIT Chicago sampling station (Tables 3-3 to 3-5). Monthly composite concentrations of vapor-phase
total PCBs ranged from 0 pg/m3 at Beaver Island and Brule River stations to 6300 pg/m3 at the IIT
April 2004
3-5
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Chicago station (Table 3-6). Mean monthly composite concentrations of vapor-phase PCBs ranged from
0.24 pg/m3 for PCB 180 at Sleeping Bear Dunes to 2600 pg/m3 for total PCBs at the IIT Chicago site.
The variability of the monthly composite concentrations differed among both stations and congeners, with
RSD values ranging from 32% to 150% for PCB 33, from 15% to 170% for PCB 118, and 45% to 160%
for PCB 180. Except for vapor-phase results for PCBs 118 and 180 at Indiana Dunes, and the PCB 33
results at Sleeping Bear Dunes, many of the highest RSD values are associated with sampling stations
with small numbers of total samples, particularly for the over-water stations and the remote shore-based
site at Eagle Harbor, suggesting that one of the monthly composite results may be driving the variability.
For stations with greater than 10 samples over the course of the study, the variability for vapor-phase
PCB 33 was greatest at Sleeping Bear Dunes (RSD = 120%) and greatest for PCB 118 and PCB 180 at
Indiana Dunes (RSDs of 140% and 150%, respectively).
Vapor-phase fra«s-nonachlor was detected much less frequently than PCB congeners (Table 3-7). Vapor-
phase fra«5-nonachlor was not detected in the samples from two over-water stations. Of the 28 sampling
stations, 13 stations had 13% to 50% of the individual samples below detection limits. Only one sample
was collected at Stations 380 and MB19M and each had a result of zero.
Concentrations of vapor-phase fra«s-nonachlor ranged from 0 pg/m3 at over-water stations 380 and
MB19M to 118 pg/m3 at Bondville. Non-zero mean monthly composite concentrations oftmns-
nonachlor for each sampling station ranged from 2.1 pg/m3 measured at Brule River to 43 pg/m3 measured
at Bondville.
3-6 April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Table 3-3. Monthly Composite Concentrations of Vapor-phase PCB 33 Measured in Samples Collected
Around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline
Atmospheric Stations
Out-of-Basin
Atmospheric Stations
Over-Water
Atmospheric Stations
Beaver Island
Chiwaukee Prairie
I IT Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB24M
1
5
6
110
18M
23M
27M
280
310
380
40M
47M
MB19M
11M
N
18
19
19
19
19
18
11
19
19
18
4
4
3
3
4
3
3
3
2
3
3
3
1
3
3
1
1
Mean
(pg/m3)
20
14
130
39
16
21
10
17
41
6.1
4.9
6.4
27
31
37
25
20
43
31
19
20
26
10
13
22
10
33
Range
(pg/m3)
2.1 to 54
2.0 to 35
24 to 290
4.6 to 120
2.4 to 48
2.6 to 71
2.0 to 44
3.4 to 37
4.0 to 130
0.62 to 21
0.89 to 12
4.3 to 8.7
3.5 to 72
8.0 to 51
20 to 55
16 to 30
2.1 to 54
7.8 to 100
23 to 40
6.0 to 34
7.0 to 45
10 to 53
NA
2.0 to 28
4.6 to 51
NA
NA
SD
(pg/m3)
17
9.3
87
30
13
17
12
11
33
5.7
5.0
2.2
39
22
14
7.8
29
53
12
14
22
23
NA
13
25
NA
NA
RSD (%)
83
66
69
76
80
83
120
64
80
94
100
34
150
70
38
32
140
120
37
76
110
91
NA
100
110
NA
NA
% Below DL*
0
0
0
0
0
0
0
0
10
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
* Value represents the percent of the individual samples collected and analyzed, not of the monthly composite samples
prepared from the individual samples.
NA = Not applicable
April 2004
3-7
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 3-4. Monthly Composite Concentrations of Vapor-Phase PCB 118 Measured in Samples Collected
Around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline
Atmospheric Stations
Out-of-Basin
Atmospheric Stations
Over-Water
Atmospheric Stations
Beaver Island
Chiwaukee Prairie
I IT Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear
Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB24M
1
5
6
110
18M
23M
27M
280
310
380
40M
47M
MB19M
N
18
19
19
19
19
18
11
19
19
18
4
4
3
3
4
3
3
3
2
3
3
3
1
3
3
1
Mean
(pg/m3)
19
2.2
29
6.8
2.2
6.6
1.2
1.8
1.6
0.83
2.9
1.2
5.2
39
6.5
42
5.7
6.2
6.5
4.1
4.2
3.9
1.0
2.4
3.2
1.0
Range
(pg/m3)
1.2 to 70
0.24 to 5.8
3.5 to 66
0.58 to 33
0.25 to 5.0
1.1 to 21
0.29 to 2.7
0.35 to 5.2
0.43 to 2.9
0.075 to 2.8
0.20 to 9.9
0.59 to 1.7
1.1 to 13
1.3to110
5.4 to 7.9
2.2 to 120
0.78 to 15
2.0 to 14
5.7 to 7.4
0.64 to 6.4
1.5 to 7.1
2.3 to 6.1
NA
0.48 to 4.9
1.2 to 6.3
NA
SD
(pg/m3)
21
1.5
22
9.2
1.5
6.5
0.75
1.3
0.78
0.85
4.7
0.50
6.4
61
1.0
67
8.4
7.1
1.2
3.1
2.8
2.0
NA
2.2
2.8
NA
RSD (%)
110
71
77
140
67
99
62
73
50
100
170
43
120
160
16
160
150
120
18
74
67
51
NA
95
87
NA
% Below DL*
0
0
0
0
0
0
3.8
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Value represents the percent of the individual samples collected and analyzed, not of the monthly composite samples
prepared from the individual samples.
NA = Not applicable
3-8
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Table 3-5. Monthly Composite Concentrations of Vapor-Phase PCB 180 Measured in Samples Collected
Around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline
Atmospheric Stations
Out-of-Basin
Atmospheric Stations
Over-Water
Atmospheric Stations
Beaver Island
Chiwaukee Prairie
I IT Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear
Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB24M
1
5
6
110
18M
23M
27M
280
310
380
40M
47M
MB19M
11M
N
18
19
19
19
19
18
11
19
19
18
4
4
3
3
4
3
3
3
2
3
3
3
1
3
3
1
1
Mean
(pg/m3)
11
0.54
4.3
1.6
0.47
1.0
0.24
0.47
0.52
0.27
0.33
0.36
0.73
16
2.4
15
0.93
1.8
1.6
1.0
1.4
1.1
0.47
0.67
0.75
0.46
55
Range
(pg/m3)
0.53 to 27
0.076 to 1.2
0.44 to 9.2
0.1 3 to 8.3
0.059 to 1.1
0.1 3 to 2.7
0.00 to 0.57
0.059 to 1.3
0.13 to 1.1
0.00 to 0.80
0.065 to 0.81
0.015 to 0.50
0.24 to 1.6
0.90 to 45
1.6 to 4.3
1.1 to 43
0.1 6 to 2.4
0.57 to 4.0
0.69 to 2.5
0.00 to 2.0
0.43 to 3.1
0.36 to 1.6
NA
0.087 to 1.4
0.17 to 1.6
NA
NA
SD
(pg/m3)
9.8
0.32
3.3
2.3
0.32
0.93
0.20
0.38
0.30
0.23
0.33
0.16
0.79
25
1.3
24
1.3
1.9
1.3
0.99
1.5
0.64
NA
0.69
0.77
NA
NA
RSD (%)
92
59
75
150
68
90
84
87
57
87
100
45
110
160
52
160
140
100
80
96
110
60
NA
100
100
NA
NA
% Below DL*
0
0
0
0
0
0
35
0
0
11
13
0
0
0
0
0
0
0
0
33
0
0
0
0
0
0
0
Value represents the percent of the individual samples collected and analyzed, not of the monthly composite samples
prepared from the individual samples.
NA = Not applicable
April 2004
3-9
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 3-6. Monthly Composite Concentrations of Vapor-Phase Total PCBs Measured in Samples Collected
Around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline Atmospheric
Stations
Out-of-Basin Atmospheric
Stations
Over-Water Atmospheric
Stations
Beaver Island
Chiwaukee Prairie
III Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB24M
1
5
6
110
18M
23M
27M
280
310
380
40M
41
47M
MB19M
11M
N
19
19
19
19
19
18
15
19
19
19
4
4
4
5
6
4
3
5
4
5
4
3
1
4
2
5
1
2
Mean
(pg/m3)
970
320
2600
680
350
490
380
400
250
110
260
170
940
990
670
1200
810
560
490
360
480
650
290
340
21
410
280
2200
Range (pg/m3)
0 to 2400
47 to 810
460 to 6300
88 to 2000
49 to 830
68 to 1300
54 to 2000
54 to 1400
44 to 590
0.00 to 390
29 to 800
87 to 260
7.8 to 3400
6.0 to 3600
9.8 to 1500
52 to 3900
84 to 2200
7.6 to 2200
21 to 1300
9.6 to 1000
4.9 to 1300
380 to 1200
NA
7.6 to 1000
16 to 25
8.5 to 1500
NA
8.9 to 4300
SD (pg/m3)
880
230
1900
580
260
410
550
360
150
110
370
77
1600
1500
590
1800
1200
930
600
410
570
430
NA
460
6.4
630
NA
3100
RSD
(%)
90
72
72
85
74
84
150
89
60
99
140
46
180
150
88
140
150
170
120
120
120
67
NA
140
31
150
NA
140
NA = Not applicable
3-10
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Table 3-7. Monthly Composite Concentrations of Vapor-Phase frans-Nonachlor Measured in Samples
Collected Around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline
Atmospheric
Stations
Out-of- Basin
Atmospheric
Stations
Over-Water
Atmospheric
Stations
Beaver Island
Chiwaukee Prairie
I IT Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear
Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB24M
1
5
6
110
18M
23M
27M
280
310
380
40M
41
47M
MB19M
Station 11M
N
18
19
19
19
19
18
15
19
19
18
4
4
4
5
6
4
3
5
4
5
4
3
1
4
2
5
1
2
Mean
(pg/m3)
3.9
9.3
29
19
7.5
14
5.3
10
43
2.1
3.7
8.8
2.9
13
14
12
5.2
8.3
14
5.6
5.8
6.8
0
6.6
3.2
7.2
0
4.3
Range (pg/m3)
0.10 to 20
0.30 to 25
1.5 to 80
1.0 to 61
Oto19
0.40 to 51
0.98 to 15
Oto33
1.9 to 120
Oto12
1.0 to 6.2
5.5 to 21
0.00 to 5.1
1.1 to 39
1.8 to 28
1.1 to 34
0.0 to 14
Oto31
3.8 to 25
Oto14
Oto19
1.9 to 16
NA
0.20 to 14
0 to 6.5
Oto14
NA
3.8 to 4.7
SD (pg/m3)
4.9
8.1
22
16
6.0
16
4.8
8.8
37
2.9
2.1
3.6
2.3
15
11
15
8.0
13
10
5.2
8.6
7.6
NA
6.2
4.6
5.8
NA
0.69
RSD (%)
130
86
77
84
81
110
86
86
87
140
58
41
79
110
81
130
150
150
74
93
150
110
NA
94
140
80
NA
16
% Below DL*
6
0
0
0
21
0
26
4.5
0
28
13
0
25
0
0
0
33
20
0
20
25
0
100
0
50
20
100
0
* Value represents the percent of the individual samples collected and analyzed, not of the monthly composite samples
prepared from the individual samples.
NA = Not applicable
April 2004
3-11
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
3.1.1.1 Temporal Variation
Vapor-phase PCB results exhibited a seasonal trend, with higher concentrations occurring in summer
months and lower concentrations occurring in winter months (Figure 3-1). This seasonal variation has
often been reported for semivolatile compounds (Hoff et al, 1992, Burgoyne and Kites, 1993, and Cortes
et al, 1998) and may be a result of the interaction of the vapor pressures of the PCBs and the increased
temperatures during summer months. The results for each monthly composite sample were plotted using
the mid-point of the compositing period as the date (e.g., the date midway between the start of the
collection of the first individual sample in the composite and the end of the collection of the last
individual sample in the composite).
All of the sites exhibited similar trends in vapor phase total PCB concentrations, with higher
concentrations generally occurring in the summer and lower concentrations in the winter, despite
differences between sites of an order of magnitude or more. For example, concentrations of vapor-phase
total PCBs measured at the urban-influenced site, Manitowoc, were 49 pg/m3 in February 1995 and were
17 times higher in August 1995 at 830 pg/m3. For the remote site, Beaver Island, vapor-phase total PCB
concentrations were 72 pg/m3 in February 1995 and were 31 times higher in August 1995, at 2300 pg/m3.
Vapor-phase fra«s-nonachlor results showed an even stronger seasonal variation than the vapor-phase
PCB results. All of the sites exhibited similar trends in vapor phase fra«s-nonachlor concentrations, with
higher concentrations generally occurring in the summer and lower concentrations in the winter, despite
differences between sites of an order of magnitude or more. For the urban site IIT Chicago,
concentrations of vapor-phase fra«s-nonachlor were 1.5 pg/m3 in February 1995 and were 50 times higher
in July 1995, at 80 pg/m3. For the rural Bondville station, vapor-phase fra«s-nonachlor was 2 pg/m3 in
February 1995 and was 60 times higher in July 1995 at 120 pg/m3.
Temporal variability could not be evaluated for the over-water stations because of the limited number of
composite samples (1 to 4 at any given over-water station) that do not represent either all four seasons in
any one year, or the entire time-span of the LMMB Study.
3-12 April 2004
-------
Figure 3-1. Temporal Variations of Total Vapor-Phase PCB (top) and frans-Nonachlor (bottom) Concentrations Measured at Lake Michigan Shoreline
and Out-of-Basin Stations from April 1994 to October 1995
"E
01
— 4000
0. 3000
~rS
o
Urban and Urban-Influenced Sampling Locations
-INDIANA DUNES
NT-CHICAGO
MUSKEGON
-MANITOWOC
CHIWAUKEE PRAIRIE
Apr-94 Jul-94 Oct-94 Jan-95 Apr-95 Jun-95 Sep-95
Urban and Urban-Influenced Sampling Locations
INDIANA DUNES
NT-CHICAGO
MUSKEGON
MANITOWOC
CHIWAUKEE PRAIRIE
o! 3000
T5
o
Remote and Rural Sampling Locations
—•—BEAVER ISLAND
—•—EAGLE HARBOR
SLEEPING BEAR DUNES
—1—BRULE RIVER
BONDVILLE
....... SOUTH HAVEN
Oct-94 Jan-95 Apr-95
Remote and Rural Sampling Locations
—•—BEAVER ISLAND
—•—EAGLE HARBOR
SLEEPING BEAR DUNES
—1—BRULE RIVER
—X—BONDVILLE
....... SOUTH HAVEN
XV
A
Sep-95
• r-94 Jul-94
Sep-95
Apr-94
3-13
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
PCB congener and total PCB concentrations were significantly different among the four seasons. Seasons
were defined as:
Spring (SP)
Summer (SU) =
Autumn (AU) =
Winter (WI) =
March 20 to June 20,
June 21 to September 22,
September 23 to December 21, and
December 22 to March 19
As illustrated in Figure 3-2, the concentrations of total PCBs measured at shoreline and out-of-basin
stations differed significantly between seasons (p<0.0001, two-way AN OVA with a Tukey pairwise
comparison, total PCB concentrations were log-transformed prior to conducting the test). Concentrations
of vapor-phase total PCBs are significantly higher in the summer and the seasons can be ranked in order
of decreasing monthly composite vapor-phase total PCB concentration as: Summer > Spring > Autumn
>Winter. Concentrations of PCB 118 and 180 also were significantly different for all seasons in the same
pattern as the total PCBs.
Figure 3-2. Seasonal Differences in Vapor-phase Total PCB Concentrations
Measured at Lake Michigan Shoreline and Out-of-basin Stations from April
1994 and October 1995
10000=,
CO
<
E
Ł=
O
Ł=
0)
O
c
O
O
1000=
100=
10=
1:SP(N=58) 2:SU(N=61) 3: AU (N=39) 4:WI(N=31)
Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile (top of box) results. Bars represent the
results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest edge of the box. Circles represent results
beyond 1.5*IQR from the box. The letters (A - D) above the boxes represent the results of the analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05). Concentration is plotted on a log scale.
For PCB 33, there was a statistically significant interaction between season and station, indicating that the
variations among seasons differed for the stations. The out-of-basin stations at Bondville and Eagle
Harbor and the shoreline stations at Muskegon and Sleeping Bear Dunes did not have a significant
difference for vapor-phase PCB 33 among seasons. The results for all other stations exhibited some
significant difference among seasons, with the same general trend as total PCBs, PCB 118 and PCB 180,
in which the mean concentrations were highest in summer and lowest in winter. At Eagle Harbor, the
lack of a seasonal difference likely was due to the lack of composite samples collected in autumn and
winter at this remote site.
3-14
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
For trans-nonachlor, concentrations differed significantly among seasons (p<0.0001, two-way ANOVA
with a Tukey pairwise comparison, trans -nonachlor concentrations were log-transformed prior to
conducting the test). Mean concentrations of vapor-phase fra«s-nonachlor are significantly higher in the
summer and the seasons can be ranked in order of decreasing monthly composite vapor-phase total PCB
concentration as: Summer > Spring > Autumn >Winter.
Figure 3-3. Seasonal Differences in frans-Nonachlor Concentrations
Measured at Lake Michigan Shoreline and Out-of-Basin Stations from April
1994 to October 1995
1000-1
ro
o
Ł=
0)
O
c
o
o
100-
10-
1:SP(N=57) 2:SU(N=61) 3: AU (N=38) 4:WI(N=31)
Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile (top of box) results. Bars represent the
results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest edge of the box. Letters above the boxes
represent the results of the analysis of variance and multiple comparisons test. Boxes with the same letter were not statistically different (at
alpha = 0.05). Concentration is plotted on a log scale.
3.1.1.2 Geographical Variation
Vapor-phase PCB congener and total PCB concentrations varied by sampling station (Table 3-9). Urban
and urban-influenced sites had higher mean monthly composite concentrations for the duration of the
study period than rural sites. For example, the mean vapor-phase total PCB concentrations were 2600,
460, and 400 pg/m3, respectively, for the urban, urban-influenced, and rural stations. A similar pattern
was observed for PCB 118 (Table 3-8). These data are consistent with what is expected, given that urban
and urban-influenced areas contain significant sources of vapor-phase PCBs. However, the mean
concentrations for total PCBs and PCB 118 at the remote sites are higher than those at the urban and
urban-influenced sites. The higher mean concentration for the remote sites is largely due to the high
concentration of PCBs at the Beaver Island station (Figure 3-4). The results for PCBs at Beaver Island
suggest the presence of an unknown source for PCBs at this station.
The mean concentration of vapor-phase total PCBs at over-water stations was 640 pg/m3 (Table 3-8),
which is higher than the mean concentrations for the urban-influenced, out-of-basin, and rural stations
(460, 190, and 400 pg/m3, respectively). The mean concentration of vapor-phase PCB 118 at over-water
stations was also higher than the other station types (Table 3-8). However, the mean concentrations of
April 2004
3-15
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
PCB at the over-water stations may be driven by the relatively high concentrations observed at Stations
1,5,6, and 11M (Tables 3-3 to 3-6), which are all fairly close to shore in the area near Chicago.
The mean concentration of vapor-phase trans-nonachlor at over-water stations was 8.1 pg/m3, which is
lower than the mean concentrations at all other types of stations except the remote stations.
Table 3-8. Mean Monthly Composite Concentrations of Vapor-Phase Total PCBs, PCB 118, and trans-
Nonachlor at LMMB Study Sampling Stations in and around Lake Michigan from April 1994 to October 1995
Vapor-phase
parameter
Total PCBs
PCB 118
frans-Nonachlor
Sampling Station Type
Shoreline
Over-water
Out-of-Basin
Shoreline
Over-water
Out-of-Basin
Shoreline
Over-water
Out-of-Basin
Urban
Urban-Influenced
Rural
Remote
Overall
Urban
Urban-Influenced
Rural
Remote
Overall
Urban
Urban-Influenced
Rural
Remote
Overall
N
19
75
19
34
147
62
42
19
75
19
29
142
43
41
19
75
19
33
146
62
41
Mean (pg/m3)
2600
460
400
710
790
640
190
29
4.4
1.8
12
8.9
13
1.4
29
12
10
4.6
13
8.1
21
Range (pg/m3)
460 to 6300
47 to 2000
54 to 1400
0.0 to 2400
0.0 to 6300
4.0 to 4300
0.0 to 800
3.5 to 66
0.24 to 33
0.35 to 5.2
0.29 to 70
0.24 to 70
0.48 to 150
0.075 to 9.9
1.5 to 80
0.0 to 61
0.0 to 33
0.10 to 20
0.0 to 80
0.0 to 39
0.0 to 120
RSD (%)
72
90
89
110
140
160
93
77
140
73
150
170
250
120
77
100
86
100
120
110
150
Vapor-phase PCB concentrations varied by station, with the highest mean concentrations generally at the
urban IIT Chicago station and the remote Beaver Island station. However, there is no clear trend of
concentrations and latitude (Figure 3-4).
3-16
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Figure 3-4. Vapor-phase PCB 118 Concentrations Measured at Shoreline and Out-of-basin Sampling
Stations around Lake Michigan from April 1994 to October 1995
35 -
_ 30
E
oo 25
m
a
15 -
5 -
n
n n
Vapor-phase fra«5-nonachlor concentrations also varied by sampling station. The mean monthly
composite sample results for fra«s-nonachlor in Table 3-8 exhibit a general trend of decreasing
concentrations from urban stations to remote stations. In addition, the fra«s-nonachlor data exhibit a
trend of increasing concentrations moving south across Lake Michigan (Figure 3-5).
April 2004
3-17
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 3-5. Vapor-phase frans-Nonachlor Concentrations Measured at Sampling Stations around
Lake Michigan from April 1994 to October 1995
40 -
^
&
Concentrations of vapor-phase PCB congeners in samples collected at over-water sampling locations also
differed among stations. Figure 3-6 illustrates a significant difference between sampling stations in
Southern Lake Michigan (sites south of 43° latitude) and stations in Northern Lake Michigan (sites north
of 43° latitude) for PCB 118. PCB 118 concentrations observed at southern sampling stations were
significantly higher than concentrations observed at northern sampling stations.
For vapor-phase fra«s-nonachlor and total PCBs, a significant difference between over-water sampling
stations in southern Lake Michigan (sites south of 43° latitude) compared to stations in northern Lake
Michigan did not occur.
3-18
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Figure 3-6. Concentrations of PCB118 in Vapor Measured in Over-water
Samples Collected in the Northern and Southern Areas of Lake Michigan
1000=,
100=
ro
o
in
"c
0)
o
o
O
North (N=10)
South (N=24)
Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile (top of box) results. Bars represent the
results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest edge of the box. Circles represent results
beyond 1.5*IQR from the box. Letters above the boxes represent the results of the analysis of variance and multiple comparisons test. Boxes
with the same letter were not statistically different (at alpha = 0.05). Northern sites are those north of 43° latitude.
3.1.2 Particulate Fraction
Particulate-phase PCB congeners were detected in the majority of the samples collected from all LMMB
Study stations. Tables 3-9 to 3-12 present the results for the monthly composite particulate-phase
samples from this study. As discussed in Chapter 2, the composite results represent either: 1) the
physical compositing of several individual samples collected during a calendar month to create one
sample for analysis, or 2) mathematical composites of the results from the analysis of the individual
samples collected over a calendar month. In some instances, both physical and mathematical composites
were prepared within a month. In these instances, the reported result is a mathematical composite based
on both the physical composite samples and the individual samples. The total number of composite
results is shown for each station as "N," along with the mean concentration, range, standard deviation,
and relative standard deviation (RSD). Tables 3-9 to 3-11 also indicate the percent of the individual
sample results that were below the sample-specific detection limit, not the percent of the composite
results. The mean concentrations were calculated using the results reported by each laboratory
(substitution of the detection limit or other value was not used for results below the sample-specific
detection limits).
Particulate-phase PCB congener concentrations ranged from 0 pg/m3 for PCB 180 at the Manitowoc and
Sleeping Bear Dunes sampling stations to 8.7 pg/m3 for PCB 118 at the IIT Chicago sampling station
(Tables 3-9 to 3-11). Concentrations of particulate-phase total PCBs ranged from 0 pg/m3 at the Beaver
Island station to 250 pg/m3 at the IIT Chicago station (Table 3-12). The mean particulate-phase PCB
concentrations ranged from 0.10 pg/m3 for PCB 180 at Sleeping Bear Dunes to 3.2 pg/m3 at the IIT
Chicago station. Mean particulate-phase total PCB concentrations ranged from 0.37 pg/m3 at over-water
Station 6 to 91 pg/m3 at the IIT Chicago station.
April 2004
3-19
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
The variability of the PCB concentrations differed among both stations and congeners, with RSD values
for PCB 33 ranging from 24% to 77%, from 26% to 190% for PCB 118, and 35% to 210% for PCB 180.
In contrast to the vapor-phase results, the greatest RSD values tend to be associated with the stations with
larger numbers of samples, while the lowest RSD values occur at the two stations with only four samples,
Eagle Harbor and Empire Michigan. However, the low RSD values at these two stations may be due to a
lack of seasonal variability since the results for these stations are only from one season. Over-water
stations 5 and GB24M have only one sample, therefore, no RSD can be calculated. For stations with
greater than 10 samples over the course of the study, the variability for particulate-phase PCB 33 was
greatest at IIT Chicago (RSD = 72%) and greatest for PCB 118, PCB 180, and total PCBs at Sleeping
Bear Dunes (RSDs of 190%, 210%, and 110%, respectively).
Particulate-phase fra«s-nonachlor was detected much less frequently than PCB congeners in the samples.
Except for the samples collected at the Empire Michigan station, fra«s-nonachlor was reported as being
below the sample-specific detection limit in 20 to 100% of the particulate-phase samples from the other
16 stations (Table 3-13). Concentrations of particulate-phase fra«s-nonachlor ranged from 0 pg/m3 at 12
stations to 2.6 pg/m3 at Bondville. Mean monthly composite concentrations of fra«s-nonachlor for each
sampling station ranged from 0.16 pg/m3 measured at GB24M to 1.2 pg/m3 measured at IIT Chicago, with
a concentration of 1.8 pg/m3 for the only sample collected at over-water Station 1.
Table 3-9. Monthly Composite Concentrations of Particulate-Phase PCB 33 Measured in Samples Collected
Around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline Atmospheric
Stations
Out-of-Basin
Atmospheric Stations
Over-water
Atmospheric Stations
Beaver Island
Chiwaukee Prairie
IIT Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB24M
5
N
17
19
19
19
19
16
14
17
19
18
4
4
1
1
Mean
(pg/m3)
0.52
0.52
1.7
0.63
0.43
0.38
0.59
0.58
0.66
0.46
0.32
0.30
0.34
1.3
Range
(pg/m3)
0.21 to 1.4
0.20 to 0.82
0.67 to 5.3
0.28 to 1.4
0.17 to 0.83
0.032 to 0.67
0.16 to 1.4
0.32 to 1.1
0.24 to 1.1
0.18to1.0
0.20 to 0.58
0.20 to 0.35
NA
NA
SD
(pg/m3)
0.26
0.16
1.2
0.29
0.17
0.19
0.40
0.20
0.22
0.21
0.18
0.071
NA
NA
RSD (%)
50
31
72
46
40
51
68
34
33
47
56
24
NA
NA
% Below DL*
0
0
0
3
5
25
7
11
0
0
0
0
100
0
Value represents the percent of the individual samples collected and analyzed, not of the monthly composite samples
prepared from the individual samples.
NA = Not applicable
3-20
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Table 3-10. Monthly Composite Concentrations of Particulate-phase PCB118 Measured in Samples
Collected Around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline Atmospheric
Stations
Out-of-Basin
Atmospheric Stations
Over-water
Atmospheric Stations
Beaver Island
Chiwaukee Prairie
III Chicago
Manitowoc
Muskegon
Sleeping Bear Dunes
South Haven
Indiana Dunes
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB24M
5
Spatial Composites
N
17
19
19
19
16
14
17
19
91
18
4
4
1
1
12
Mean
(pg/m3)
1.7
0.47
3.2
0.48
0.67
0.24
0.49
0.77
0.40
0.50
0.22
0.34
0.14
0.61
0.56
Range
(pg/m3)
0.51 to 4.0
0.19 to 0.83
1.5 to 8.7
0.20 to 1.3
0.17 to 1.6
0.023 to 1.8
0.20 to 1.9
0.35 to 1.7
0.1 9 to 0.89
0.14to1.4
0.1 4 to 0.28
0.25 to 0.44
NA
NA
0.065 to 2.0
SD
(pg/m3)
0.95
0.18
1.7
0.26
0.46
0.46
0.39
0.31
0.17
0.27
0.067
0.089
NA
NA
0.54
RSD (%)
56
40
54
54
69
190
80
41
44
53
30
26
NA
NA
96
% Below DL*
0
0
0
0
0
43
11
0
5
0
0
0
100
0
8
* Value represents the percent of the individual samples collected and analyzed, not of the monthly composite samples
prepared from the individual samples.
NA = Not applicable
Table 3-11. Monthly Composite Concentrations of Particulate-Phase PCB 180 Measured in Samples
Collected Around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline Atmospheric
Stations
Out-of-Basin
Atmospheric Stations
Over-Water
Atmospheric Stations
Beaver Island
Chiwaukee Prairie
III Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB24M
5
Spatial Composites
N
17
19
19
19
19
16
14
17
19
18
4
4
1
1
12
Mean
(pg/m3)
1.0
0.21
1.9
0.36
0.22
0.38
0.10
0.29
0.26
0.18
0.090
0.11
0.11
0.21
0.27
Range
(pg/ms)
0.26 to 2.6
0.067 to 0.34
0.97 to 4.6
0.10 to 0.71
0.0 to 0.74
0.1 4 to 0.83
0.0 to 0.82
0.046 to 1.5
0.078 to 1.3
0.00 to 0.69
0.046 to 0.1 3
0.065 to 0.1 5
NA
NA
0.047 to 0.71
SD
(pg/ms)
0.71
0.075
0.84
0.16
0.15
0.22
0.21
0.34
0.28
0.15
0.034
0.037
NA
NA
0.20
RSD (%)
69
36
45
45
69
59
210
120
110
82
37
35
NA
NA
74
% Below DL*
0
10
0
0
5
0
57
17
14
6
25
0
100
100
33
Value represents the percent of the individual samples collected and analyzed, not of the monthly composite samples
prepared from the individual samples.
NA = Not applicable
April 2004
3-21
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 3-12. Monthly Composite Concentrations of Particulate-phase Total PCBs Measured in Samples
Collected Around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline Atmospheric
Stations
Out-of-Basin
Atmospheric Stations
Over-Water
Atmospheric Stations
Beaver Island
Chiwaukee Prairie
I IT Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Spatial Composites
Empire Michigan
GB24M
1
5
6
N
18
19
19
19
19
16
15
18
19
18
4
18
4
2
1
3
1
Mean (pg/m3)
52
22
91
33
26
24
18
23
25
21
14
19
14
3.9
2.6
17
0.37
Range (pg/m3)
0.0 to 110
7.7 to 31
46 to 250
18 to 66
8.6 to 110
7.7 to 51
0.0 to 69
0.0 to 47
9.6 to 58
5.8 to 71
8.2 to 20
0.18 to 77
8.9 to 18
0.26 to 7.6
NA
0.25 to 48
NA
SD (pg/m3)
29
6.1
48
12
22
12
21
12
14
14
4.7
21
4.0
5.2
NA
27
NA
RSD (%)
56
28
53
36
84
49
110
52
54
66
33
110
29
130
NA
160
NA
NA = Not applicable
3-22
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Table 3-13. Monthly Composite Concentrations of Particulate-Phase frans-Nonachlor Measured in Samples
Collected Around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline Atmospheric
Stations
Out-of-Basin
Atmospheric Stations
Over-Water
Atmospheric Stations
Beaver Island
Chiwaukee Prairie
III Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB24M
1
5
6
Spatial Composites
N
17
19
19
18
18
16
4
17
19
18
4
4
2
1
3
1
18
Mean
(pg/m3)
0.35
0.48
1.2
0.64
0.58
0.47
0.30
0.46
0.81
0.43
0.24
0.34
0.16
1.8
0.36
0.74
0.41
Range
(pg/m3)
0.0 to 0.72
0.0 to 1.0
0.0 to 2.5
0.0 to 1.6
0.075 to 1.4
0.0 to 1.4
0.0 to 1.1
0.0 to 1.2
0.0 to 2.6
0.0 to 1.1
0.051 to 0.49
0.17 to 0.51
0.0 to 0.32
NA
0.0 to 1.0
NA
0.0 to 1.6
SD
(pg/m3)
0.22
0.25
0.68
0.45
0.42
0.39
0.55
0.32
0.89
0.31
0.20
0.14
0.23
NA
0.55
NA
0.46
RSD (%)
62
53
58
70
73
84
190
71
110
71
84
41
140
NA
150
NA
110
% Below DL*
29
24
20
34
22
31
75
33
48
39
50
0
100
0
67
100
61
Value represents the percent of the individual samples collected and analyzed, not of the monthly composite samples
prepared from the individual samples.
NA = Not applicable
3.1.2.1 Temporal Variation
Particulate-phase PCB congener and total PCB results exhibited different temporal trends among
congeners and also among stations. Season did not have a significant effect on particulate-phase PCB 33
concentrations, but did have a significant effect on particulate-phase total PCBs (p<0.0001, two-way
ANOVA, PCB concentrations were log-transformed prior to conducting the test). For particulate-phase
PCB 118 and PCB 180 concentrations, there was a significant interaction between station and season
(p<0.0361, two-way ANOVA, PCB concentrations were log-transformed prior to conducting the test).
The relationship between season and concentration differed for PCBs 118 and 180 at different stations.
Figure 3-7 illustrates the interaction between station and season for PCB 118.
The particulate-phase results for PCB 118 at the Beaver Island and IIT Chicago stations are highest in
April 1994 and generally decrease by September 1995, and variable in between. While some of the urban
and urban-influenced stations exhibit their highest concentrations of PCB 118 in May 1994 (e.g.,
Muskegon, Manitowoc, and Indiana Dunes), the overall decrease seen at Beaver Island and IIT Chicago is
not apparent at the other stations. The rural and remote stations other than Beaver Island do not show a
temporal trend in particulate-phase PCB 118 concentrations. The highest result at Sleeping Bear Dunes
occurred in January 1995 (1.8 pg/m3). However, this result may be due to contamination, as evidenced by
the field duplicate composite sample that had a concentration of only 0.19 pg/m3. The highest result at
IIT Chicago in April 1994 is especially striking, given that the other samples collected from this station
during April 1995 had lower results than the neighboring months.
April 2004
3-23
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 3-7. Temporal Variations in Particulate-phase PCB118 Concentrations Measured at Lake
Michigan Shoreline and Out-of-basin Stations from April 1994 to October 1995
Urban and Urban-Influenced Sites
I
-INDIANA DUNES
NT-CHICAGO
MUSKEGON
-MANITOWOC
CHIWAUKEE PRAIRIE
Apr-94
Apr-95
Sep-95
Remote and Rural Sites
-BEAVER ISLAND
-EAGLE HARBOR
SLEEPING BEAR DUNES
-BRULERIVER
BONDV1LLE
Apr-94
Apr-95
Sep-95
The particulate-phase total PCB concentrations were significantly higher in Spring compared to Autumn
and Summer (two-way ANOVA, with Tukey pairwise comparisons) and significantly higher in Winter
than Autumn (Figure 3-8).
3-24
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Figure 3-8. Seasonal Differences in Particulate-phase Total PCB
Concentrations Measured at Lake Michigan Shoreline and Out-of-basin
Stations from April 1994 to October 1995
1000-,
ro
Ł=
O
Ł=
0)
O
c
O
O
100-
BC
•
!
1: SP (N=58) 2: SU (N=59) 3: AU (N=38) 4: Wl (N=29)
Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile (top of box) results. Bars represent the
results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest edge of the box. Circles represent results
beyond 1.5*IQR from the box. The letters (A - D) above the boxes represent the results of the analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05). Concentration is plotted on a log scale.
The seasonal differences evident in Figure 3-8 were apparent even when the high April 1994 results were
removed from the data set. Given that the data were log-transformed before the comparisons, it is not
unexpected that removing the April 1994 results did not change the seasonal patterns.
For fra«s-nonachlor, concentrations measured in particulate-phase samples also were significantly
different among seasons, with mean concentrations highest in winter and lowest in summer (two-way
ANOVA, with Tukey pairwise comparisons, see Figure 3-9). However, the pattern of seasonal
differences was not the same as exhibited in the particulate-phase total PCB results.
April 2004
3-25
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 3-9. Seasonal Differences in Particulate-phase frans-Nonachlor
Concentrations Measured at Lake Michigan Shoreline and Out-of-basin
Stations from April 1994 to October 1995
3.0-,
ro
o
Ł=
0)
o
c
o
o
2.0-
1.0-
0.0-
B
•
BC
•
C
•
1: SP (N=53) 2: SU (N=57) 3: AU (N=33) 4: Wl (N=26)
Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile (top of box) results. Bars represent the
results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest edge of the box. Circles represent results
beyond 1.5*IQR from the box. The letters (A - D) above the boxes represent the results of the analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05).
Temporal variability could not be evaluated for the over-water stations because of the limited number of
composite samples (1 to 4 at any given over-water station) that do not represent either all four seasons in
any one year, or the entire time-span of the LMMB Study.
3.12.2 Geographical Variation
Particulate-phase PCB congener and total PCB concentrations varied by sampling station (Table 3-14).
As was noted for the vapor-phase results, urban and urban-influenced sites had higher mean monthly
composite concentrations for the duration of the study period than rural sites, consistent with the
hypothesis that urban and urban-influenced areas contain significant sources of particulate-phase PCBs.
As also noted for the vapor-phase results, the results for total PCBs and PCB 33 at the remote sites fall
between the urban and urban-influenced sites, suggesting the presence of an unknown source for PCBs at
the remote sites, likely near Beaver Island (see Figures 3-7 and 3-10).
3-26
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Table 3-14. Mean Monthly Composite Concentrations of Particulate-phase Total PCBs, PCB 33, and trans-
Nonachlor at LMMB Study Sampling Stations in and around Lake Michigan between April 1994 and October
1995
Particulate-phase
parameter
Total PCBs
PCB 33
frans-Nonachlor
Sampling Station Type
Shoreline
Over-Water
Out-of-Basin
Shoreline
Over-Water
Out-of-Basin
Shoreline
Over-Water
Out-of-Basin
Urban
Urban-Influenced
Rural
Remote
Overall
Urban
Urban-Influenced
Rural
Remote
Overall
Urban
Urban Influenced
Rural
Remote
Overall
N
19
73
18
33
143
29
41
19
73
17
31
140
18
41
19
71
17
21
128
29
41
Mean (pg/m3)
91
26
23
37
37
16
22
1.7
0.50
0.58
0.55
0.68
0.60
0.54
1.2
0.54
0.46
0.34
0.59
0.43
0.59
Range (pg/m3)
46 to 250
7.7 to 110
Oto47
0 to 110
0 to 250
0.18 to 77
5.8 to 71
0.67 to 5.3
0.032 to 1.4
0.32 to 1.1
0.16 to 1.4
0.032 to 5.3
0 to 1.6
0.18 to 1.1
0 to 2.5
0 to 1.6
0 to 1.2
0 to 1.1
0 to 2.5
0 to 1.8
0 to 2.6
RSD(%)
53
55
52
83
90
120
60
72
46
34
59
93
79
45
58
70
71
85
82
110
110
There was no apparent relationship between the particulate-phase concentration of PCB 118 and latitude
(see Figure 3-10).
There is an apparent relationship between the particulate-phase concentration of fra«s-nonachlor and
latitude (Figure 3-11). Except for the high mean particulate-phase concentration observed at the urban
IIT Chicago site, the mean concentration of fra«s-nonachlor generally decreases moving from south to
north across Lake Michigan.
April 2004
3-27
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 3-10. Particulate-phase PCB 118 Concentrations Measured at Lake Michigan Shoreline
and Out-of basin Stations from April 1994 to October 1995
00
00
o
Ol
o
o
o
c
ro
01
°
^
Figure 3-11. Particulate-phase frans-Nonachlor Concentrations Measured at Lake Michigan
Shoreline and Out-of-basin Stations from April 1994 to October 1995
<Ą
O
IS
| 1.2
O
O
o
c
ro
•S 0.4
C
ro
01
A
3-28
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
3.1.3 Precipitation Fraction
PCB congeners were detected in many of the precipitation samples collected from the LMMB Study
stations. However, the overall frequency of occurrence of PCBs in the precipitation samples was lower
than for the vapor-phase and particulate-phase samples. The frequency of precipitation samples with
results below the sample-specific detection limit was greatest for PCB 33 compared to either PCB 118 or
180.
The precipitation samples were collected as described in Chapter 2. The samples collected from the
shoreline and out-of-basin stations represent true 28-day composite samples. The precipitation samples
collected from the over-water stations represent single-day composite results. Tables 3-15 to 3-18 present
the results for the composite precipitation samples from this study. The total number of composite results
is shown for each station as "N," along with the mean concentration, range, standard deviation, relative
standard deviation (RSD), and the percent of the sample results that were below the sample-specific
detection limit. In calculating these summary statistics, the analytical results for each sample were
volume-weighted to account for differences in the total volume of precipitation that fell during each 28-
day period. The mean concentrations were calculated using the results reported by each laboratory
(substitution of the detection limit or other value was not used for results below the sample-specific
detection limits).
Precipitation PCB concentrations ranged from 0 pg/L for PCB 33 at 10 of the 12 shoreline and out-of-
basin sampling stations (see Table 3-15) to 5,500 pg/L for PCB 118 at Muskegon (Table 3-16). Mean
precipitation PCB concentrations ranged from 2.2 pg/L for PCB 33 at the Empire Michigan station to 470
pg/L for PCB 33 at the IIT Chicago station. For total PCBs, the mean concentrations in precipitation
ranged from 290 pg/L at the Eagle Harbor station to 16,000 pg/L at the IIT Chicago station. The
precipitation samples collected at the IIT Chicago site had the highest volume-weighted mean
concentrations of PCBs 33, 180, and total PCBs, while PCB 118 had highest volume-weighted mean
concentration at the Muskegon site (see Tables 3-15 to 3-18).
Many of the summary statistics for the precipitation samples collected from the over-water stations could
not be calculated because only one precipitation sample was collected at most of these stations.
fra«5-Nonachlor was detected even less frequently than the PCB congeners in the precipitation samples
(Table 3-19). Except for the samples collected at the IIT Chicago station, fra«s-nonachlor was reported
as being below the sample-specific detection limit in 75 to 100% of the precipitation samples from all
stations. The concentrations of fra«s-nonachlor in the precipitation samples ranged from 0 pg/L at every
site to a high of 630 pg/L at the Chiwaukee Prairie site. The mean concentrations of trans-nonachlor in
precipitation samples ranged from 0 pg/L at three over-water stations to 120 pg/L at the IIT Chicago
station. Similarly, the volume-weighted mean concentrations ranged from 0 pg/L at three over-water
stations to 100 pg/L at the IIT Chicago site.
As with the PCB results, many of the summary statistics for the precipitation samples collected from the
over-water station could not be calculated because only one precipitation sample was collected at most of
these stations.
April 2004 3-29
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 3-15. Monthly Composite Concentrations of PCB 33 Measured in Precipitation Samples Collected
around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline
Atmospheric
Stations
Out-of-basin
Atmospheric
Stations
Over-water
Atmospheric
Stations
Beaver Island
Chiwaukee Prairie
I IT Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB17
GB24M
1
5
23M
380
N
19
17
16
20
18
18
14
19
20
18
3
4
1
1
1
1
1
1
VWMean
(P9/L)
45
18
270
39
36
37
58
47
45
15
21
3.0
33
10
6.9
36
7.6
8.5
Mean
(pg/L)
42
19
470
38
51
67
83
74
71
32
16
2.2
33
10
6.9
36
7.6
8.5
Range (pg/L)
0.0 to 450
0.0 to 83
15 to 3200
0.0 to 160
0.0 to 350
0.0 to 470
13 to 300
0.0 to 620
0.0 to 310
0.0 to 310
0.0 to 47
0.0 to 8.8
NA
NA
NA
NA
NA
NA
SD
(pg/L)
100
26
800
38
85
120
72
150
84
78
27
4.4
NA
NA
NA
NA
NA
NA
RSD (%)
240
130
170
100
170
180
86
210
120
240
170
200
NA
NA
NA
NA
NA
NA
% Below
DL
68
59
6
50
50
50
7
47
30
72
67
100
100
100
100
100
100
100
NA = Not applicable
3-30
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Table 3-16. Monthly Composite Concentrations of PCB 118 Measured in Precipitation Samples Collected
around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline
Atmospheric
Stations
Out-of-basin
Atmospheric
Stations
Over-water
Atmospheric
Stations
Beaver Island
Chiwaukee Prairie
III Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB17
GB24M
1
5
23M
380
N
19
16
16
20
17
17
14
19
20
18
3
3
1
1
1
1
1
1
VWMean
(P9/L)
41
26
230
34
46
240
11
35
24
67
5.8
7.7
76
23
27
45
12
17
Mean
(pg/L)
53
38
440
41
81
360
16
110
34
61
4.9
5.7
76
23
27
45
12
17
Range (pg/L)
5.5 to 250
10 to 79
48 to 2300
8.2 to 230
6.3 to 880
13 to 5500
3.4 to 37
6.2 to 1600
11 to 93
0.0 to 770
0.0 to 11
0.0 to 17
NA
NA
NA
NA
NA
NA
SD
(pg/L)
67
23
710
48
210
1300
9.7
350
25
180
5.8
9.9
NA
NA
NA
NA
NA
NA
RSD (%)
130
61
160
120
250
360
59
320
74
290
120
170
NA
NA
NA
NA
NA
NA
% Below
DL
11
13
0
0
18
0
50
5
0
39
67
67
0
0
0
100
0
0
NA = Not applicable
April 2004
3-31
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 3-17. Monthly Composite Concentrations of PCB 180 Measured in Precipitation Samples Collected
around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline
Atmospheric
Stations
Out-of-basin
Atmospheric
Stations
Over-water
Atmospheric
Stations
Beaver Island
Chiwaukee Prairie
III Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB17
GB24M
1
5
23M
380
N
19
17
16
20
18
18
14
19
20
18
3
4
1
1
1
1
1
1
VWMean
(P9/L)
23
24
180
30
16
22
5.5
18
12
12
2.5
9.2
29
8.5
14
19
3.5
6.9
Mean
(pg/L)
33
41
320
38
35
25
7.7
37
18
14
3.2
11
29
8.5
14
19
3.5
6.9
Range (pg/L)
2.5 to 160
7.8 to 150
14 to 1900
4.6 to 210
0.0 to 210
0.0 to 50
0.0 to 17
5.6 to 370
0.0 to 80
0.0 to 45
0.0 to 7.4
2.2 to 20
NA
NA
NA
NA
NA
NA
SD
(pg/L)
42
39
520
47
53
12
6.3
81
18
13
3.8
8.1
NA
NA
NA
NA
NA
NA
RSD (%)
130
95
160
120
150
48
82
220
100
93
120
73
NA
NA
NA
NA
NA
NA
% Below
DL
21
18
0
10
33
17
93
32
40
50
100
75
100
100
0
100
100
100
NA = Not applicable
3-32
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Table 3-18. Monthly Composite Concentrations of total PCBs Measured in Precipitation Samples Collected
around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline
Atmospheric
Stations
Out-of-basin
Atmospheric
Stations
Over-water
Atmospheric
Stations
Beaver Island
Chiwaukee Prairie
I IT Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear
Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire Michigan
GB17
GB24M
1
5
23M
380
N
20
20
17
21
20
20
16
21
21
19
4
4
1
1
1
1
1
1
Mean (pg/L)
1900
1800
16000
1500
2600
2600
1300
3800
1700
1700
290
2000
2300
680
750
1500
360
510
Range (pg/L)
0.0 to 11 000
0.0 to 4700
0.0 to 11 0000
0.0 to 7200
0.0 to 18000
0.0 to 19000
0.0 to 2800
0.0 to 48000
0.0 to 4200
0.0 to 13000
0.0 to 700
520 to 4800
NA
NA
NA
NA
NA
NA
SD (pg/L)
2800
1200
28000
1500
4200
4000
880
10000
1100
2900
300
2000
NA
NA
NA
NA
NA
NA
RSD (%)
150
66
180
100
160
150
66
280
69
170
100
100
NA
NA
NA
NA
NA
NA
NA = Not applicable
April 2004
3-33
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 3-19. Monthly Composite Concentrations of frans-Nonachlor Measured in Precipitation Samples
Collected around Lake Michigan from April 1994 to October 1995
Sampling Station
Shoreline
Atmospheric
Stations
Out-of-basin
Atmospheric
Stations
Over-water
Atmospheric
Stations
Beaver Island
Chiwaukee
Prairie
III Chicago
Indiana Dunes
Manitowoc
Muskegon
Sleeping Bear
Dunes
South Haven
Bondville
Brule River
Eagle Harbor
Empire
Michigan
GB17
GB24M
1
5
23M
380
N
20
20
17
21
20
19
14
21
21
19
4
4
1
1
1
1
1
1
VWMean
(P9/L)
18
11
100
26
17
27
12
13
29
37
14
16
0.0
0.0
20
69
13
0.0
Mean
(pg/L)
20
43
120
37
17
33
13
26
27
53
11
39
0.0
0.0
20
69
13
0.0
Range (pg/L)
0.0 to 73
0.0 to 630
0.0 to 480
0.0 to 190
0.0 to 140
0.0 to 210
0.0 to 46
0.0 to 260
0.0 to 290
0.0 to 210
0.0 to 31
0.0 to 130
NA
NA
NA
NA
NA
NA
SD(pg/L)
27
140
140
58
42
59
13
59
67
82
15
62
NA
NA
NA
NA
NA
NA
RSD (%)
130
330
120
150
250
180
100
230
250
160
130
160
NA
NA
NA
NA
NA
NA
% Below
DL
80
90
41
76
90
79
100
76
76
79
75
100
100
100
100
100
100
100
NA = Not applicable
3.13.7 Temporal Variation
The PCB congeners and total PCBs exhibited no clear temporal trends in the precipitation samples
collected at any of the stations. Figure 3-12 presents the results for PCB 33 and fra«s-nonachlor over the
course of the LMMB Study, by station, differentiating the urban and urban-influenced stations from the
rural and remote stations. The results for PCB 33 at the IIT Chicago site in May 1994 are 5 - 10 times
higher than any other result for this congener throughout the study. Although there are several results for
fra«5-nonachlor that are higher than those from the IIT Chicago site in May 1994, the results for trans-
nonachlor exhibit the same sharp decline from May 1994 to June 1994 as was seen for the PCB 33 results
at the same site.
3-34
April 2004
-------
Figure 3-12. Temporal Variation in Precipitation PCB 33 (top) and frans-Nonachlor (bottom) Concentrations Measured at Lake Michigan Shoreline and
Out-of-basin Stations from March 1994 to October 1995
Urban and Urban-Influenced Sampling Locatioi
Remote and Rural Sampling Locations
_2500
„ l&UU
c
E 1000
-INDIANA DUNES
MUSKEGON
-MANITOWOC
CHIWAUKEE PRAIRIE
NT-CHICAGO
|
f 2100
d)
Ł 1400
; . •*
03/27/1994 06/25/1994 09/23/1994 12/22/1994 03/22/1995 06/20/1995
03/27/1994 06/25/1994
-*r- »"-»
09/23/1994 12/22/1994
*"-• •*
03/22/1995 06/20/1995 09/18/1995
Urban and Urban-Influenced Sampling Locatio
-INDIANA DUNES
MUSKEGON
MANITOWOC
CHIWAUKEE PRAIRIE
NT-CHICAGO
A
03/27/1994 06/25/1994 09/23/1994 12/22/1994 03/22/1995 06/20/1995 09/18/1995
03/27/1994 06/25/1994 09/23/1994 12/22/1994 03/22/1995 06/20/1995
3-35
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
There is a significant difference in the concentrations of PCB 33 in precipitation among seasons
(p=0.0248, ANOVA, with Tukey pairwise comparisons) (Figure 3-13). Concentrations are significantly
higher in winter than in summer and autumn. The variability during all four seasons is large enough that
there is considerable overlap in the box plots.
Figure 3-13. Seasonal Patterns of PCB 33 Concentrations in
Precipitation Measured at Lake Michigan Shoreline and Out-of-
basin Stations from March 1994 to October 1995
1
c
o
+3
1
-------
PCBs/trans-Nonachlor in Atmospheric Components
Figure 3-14. Percent of Precipitation PCB 33 Sample Results Reported as Zero,
by Season
The concentrations of trans-nonachlor also showed a significant difference among seasons (Figure 3-15).
The winter concentrations of fra«s-nonachlor are significantly higher than the summer concentrations.
Spring and Autumn concentrations are not significantly different from one another, or from either winter
or summer.
Figure 3-15. Seasonal Patterns of frans-Nonachlor Concentrations
in Precipitation Measured at Lake Michigan Shoreline and Out-of-
basin Stations from March 1994 to October 1995
1000:,
c
o
+=
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
The overall percentage of samples reported as zero for trans-nonachlor is much higher than for PCB 33.
Figure 3-16 shows the relationship with season, which is not as clear as for PCB 33. These differences
were significant, based on a Chi-square test at alpha=0.05 (p=0.0173).
Figure 3-16. Percent of Precipitation frans-Nonachlor Sample Results Reported
as Zero, by Season
I 40%
E
Ł
30%
3.13.2 Geographical Variation
Precipitation PCB congener and total PCB concentrations varied by sampling station. Table 3-20
provides the mean composite sample concentration, the range, and the RSD, for the precipitation samples
collected in the LMMB Study. The urban and urban-influenced sites had higher mean monthly composite
concentrations for the duration of the study period than remote sites, consistent with the hypothesis that
urban and urban-influenced areas contain significant sources of PCBs. However, the mean results for
total PCBs and PCB 33 in the precipitation samples from the rural sites are higher than those for both the
remote and the urban-influenced sites. The results for precipitation differ from those for the vapor-phase
and particulate-phase, where the remote sites showed higher than anticipated concentrations of PCBs,
suggesting that the unknown source near the Beaver Island site did not have an effect on PCB
concentrations in precipitation.
3-38
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Table 3-20. Mean Precipitation Concentrations of Total PCB, PCB 33 and frans-Nonachlor at LMMB Study
Sampling Stations in and around Lake Michigan between March 1994 and October 1995
Precipitation
Parameter
Total PCBs
PCB 33
trans-
Nonachlor
Sampling Station Type
Shoreline
Over-water
Out-of-basin
Shoreline
Over-water
Out-of-basin
Shoreline
Over-water
Out-of-basin
Urban
Urban-influenced
Rural
Remote
Overall
Urban
Urban-influenced
Rural
Remote
Overall
Urban
Urban-influenced
Rural
Remote
Overall
N
17
81
21
36
155
10
44
16
73
19
33
141
10
41
17
80
21
34
152
10
44
Mean (pg/L)
15800
2100
3800
1600
3700
1400
1500
470
44
74
60
100
11
50
120
33
26
17
38
26
37
Range (pg/L)
0.0 to 110,000
0.0 to 19,000
0.0 to 48,000
0.0 to 11, 000
0.0 to 110,000
361 to 4,800
0.0 to 13,000
15 to 3,200
0.0 to 4,700
0 to 620
0.0 to 450
0.0 to 3,200
0.0 to 36
0.0 to 310
0.0 to 480
0.0 to 630
0.0 to 260
0.0 to 73
0.0 to 630
0.0 to 130
0.0 to 290
RSD (%)
180
140
280
130
290
100
130
170
170
210
150
310
120
160
120
260
230
130
230
170
200
Although there is a general decrease in precipitation total PCB concentrations moving north from the high
concentrations observed at the IIT Chicago site, the variability in the data for the stations does not reveal
a clear trend (Figure 3-17).
April 2004
3-39
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 3-17. Precipitation-phase Total PCB Concentrations Measured at Lake Michigan Shoreline
and Out-of-basin Stations from March 1994 to October 1995
20000
— 15000
o
m
o
10000
5000
JlJ
The top portion of Figure 3-18 is a similar plot of trans-nonachlor concentrations in precipitation by
station. Again, aside from the general decrease in the concentrations north of the IIT Chicago site, there
is not a prominent trend by latitude. The bottom portion of Figure 3-18 presents the percentage of
precipitation samples in which fra«s-nonachlor was reported as zero. In general, the sites with lower
mean concentrations of fra«s-nonachlor in precipitation had higher percentages of sample reported as
zero. Conversely, the IIT Chicago site had the highest mean concentration in precipitation and the lowest
percentage of results reported as zero. The Sleeping Bear Dunes site also had a low percentage of results
reported as zero, although it had one of the lowest mean concentrations in this study.
3-40
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Figure 3-18. Precipitation-phase frans-Nonachlor Concentrations Measured at Lake Michigan
Shoreline and Out-of-basin Stations from March 1994 to October 1995 (top) and the Percent of
Sample Results Reported as Zero for each Sampling Location (bottom)
1
y
o
u
o
g
o
S
CO
I
0 -
_
T
T rln
, T
\ T 1 T
1 T
x x ,x ' x x x
I 40% -
April 2004
3-41
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
3.1.4 Dry Deposition
Dry deposition samples were collected from a limited number of sites during the LMMB Study using the
procedures described in Chapter 2. The frequencies of occurrence of the PCB congeners varied greatly.
Table 3-21 presents the results for the dry deposition samples from this study. The total number of
samples is shown for each station as "N," along with the mean concentration, range, standard deviation,
and relative standard deviation (RSD). The mean concentrations were calculated using the results
reported by each laboratory (substitution of the detection limit or other value was not used for results
below the sample-specific detection limits).
Note that the units for the dry deposition samples differ significantly from those for the vapor-phase and
particulate-phase samples. Based on the manner in which the samples are collected, dry deposition results
are reported here in terms of mass per unit area, in ng/m2, not the mass per unit volume units of pg/m3
used for the other phases. Also, PCB 33 was not measured in the dry deposition samples.
Dry-deposition PCB congener concentrations ranged from 140 ng/m2 for PCB 118 at the IIT Chicago
sampling station to over 3 million ng/m2 for PCB 180 at the South Haven sampling station. Total PCBs
in the dry deposition samples exhibited a similar range. Mean concentrations for PCBs in dry deposition
ranged from 450 ng/m2 at the IIT Chicago station to 1.5 million ng/m2 at the South Haven station. Mean
concentration for dry deposition total PCBs ranged from 1800 ng/m2 at the Chicago SWFP crib intake to
320,000 ng/m2 at the South Haven station.
The analysis of the dry deposition samples presented more difficulties than the other phases studied. As a
result, a larger proportion of data for the dry deposition samples were qualified or found invalid, thereby
complicating the summary statistics shown in Table 3-21. In some cases, only one sample had valid
results for a specific PCB congener, such that a range and standard deviation could not be determined. In
addition the very high concentrations of PCB 180 observed in a few samples at the South Haven station
significantly skew the mean, range, and standard deviations reported in Table 3-21 for this congener and
for the total PCBs.
The frequencies at which the PCB congeners were found above the sample-specific detection limits also
varied greatly by congener. PCB 118 was found below the detection limit in 33 to 100% of the dry
deposition samples, while there were fewer valid results for PCB 180 above the detection limit (Table 3-
21).
fra«5-Nonachlor was reported in many of the dry deposition samples, but generally at lower
concentrations than the PCBs (Table 3-21). The concentration of fra«s-nonachlor in dry deposition
samples ranged from 0 ng/m2 at South Haven to 209 ng/m2 at the Chicago SWFP Crib Intake. Most of
the results for trans-nonachlor were reported as being below the sample-specific detection limits (e.g., 60
to 100%).
3-42 April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Table 3-21. Monthly Composite Concentrations of PCBs and frans-Nonachlor Measured in Dry Deposition
Parameter
PCB118
PCB 180
Total PCBs
frans-
Nonachlor
Sampling Station
Chicago SWFP Crib Intake
III Chicago
Sleeping Bear Dunes
III Chicago
Sleeping Bear Dunes
South Haven
Chicago SWFP Crib Intake
Harrison Crib
III Chicago
Sleeping Bear Dunes
South Haven
Chicago SWFP Crib Intake
III Chicago
Sleeping Bear Dunes
South Haven
N
2
7
3
1
1
2
9
1
13
8
11
9
5
3
4
Mean (ng/m2)
604
453
745
3220
884
1,580,000
1830
5400
7060
6120
315,000
78.2
64.8
39.4
25.7
Range (ng/m2)
534 to 674
144 to 919
524 to 11 60
NA
NA
5240 to 3, 160,000
324 to 51 50
NA
1720 to 23,500
48.6 to 19,600
109 to 3,380,000
34.8 to 209
0.00 to 100
18.2 to 55.2
0.00 to 58.5
SD (ng/m2)
99.0
296
360
NA
NA
2,230,000
1710
NA
6480
7940
1,020,000
52.0
40.6
19.1
30.0
RSD (%)
16
65
48
NA
NA
141
94
NA
92
130
323
66
63
48
118
% Below DL
50
100
33
0
0
0
NC
NC
NC
NC
NC
89
60
100
75
NA = Not applicable. Only one result was reported.
NC = Not calculated. The total PCB concentration is the sum of the results for the individual PCB congeners, each of which
has a sample-specific detection limit. However, detection limits are not additive, so there is no meaningful way to specify
a detection limit for total PCBs.
3.2 Quality Implementation and Assessment
As described in Section 1.5.5, the LMMB QA program prescribed minimum standards to which all
organizations collecting data were required to adhere. The quality activities implemented for the PCBs
and fra«5-nonachlor monitoring portion of the study are further described in Section 2.7 and included use
of SOPs, training of laboratory and field personnel, and establishment of MQOs for study data. A
detailed description of the LMMB quality assurance program is provided in the Lake Michigan Mass
Balance Study Quality Assurance Report (USEPA, 200 Ib). A brief summary of data quality issues for
the atmospheric PCBs and fra«s-nonachlor data is provided below.
As discussed in Section 2.5, because data comparability was important to the successful development of
the mass balance model, the Pis used similar sample collection, extraction, and analysis methods for the
PCB and trans-nonachlor monitoring in this study. For a small portion of the study (June 15, 1994 to
September 22, 1994), a revision to the silica gel clean-up procedure for the determination of PCBs in
precipitation and particulate phases resulted in coelution of PCB 99 with fra«s-nonachlor. Affected
sample results are qualified with the high bias flag (HIB) in the database.
The Pis used surrogate spikes to monitor the bias of the analytical procedure. Analytical results for PCBs
in vapor, precipitation, and particulate phases were corrected for surrogate recoveries. Analytical results
for PCBs in dry deposition were not corrected for surrogate recoveries because the recoveries of the
surrogates (PCB 14 and PCB 65) were subject to chromatographic interferences. The PI noted that the
recoveries were too erratic to be used as reliable indicators of method accuracy. Sample results for PCBs
April 2004
3-43
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
associated with surrogate recoveries outside the MQO limits were qualified with the failed surrogate spike
flag (FSS).
Analytical results for fra«s-nonachlor were not corrected for surrogate recoveries, except for those
collected and analyzed by Indiana University (Sleeping Bear Dunes site August 1994 to October 1995.
Analytical results were considered invalid (INV) and qualified when surrogate recoveries were less than
10%. Some PCB data were found to be invalid for this reason, but the trans-nonachlor surrogate
recoveries were never below 10%.
Laboratory matrix spike samples also were used to monitor the bias of the analytical procedure. For the
dry deposition analyses, a spiked laboratory solvent blank also was prepared and analyzed. Analytical
results associated with matrix spike samples with recoveries below the MQO limits were qualified with
the failed matrix spike and low bias flags (FMS and LOB) and results associated with matrix spike
samples with recoveries higher than the MQO limits were qualified with the failed matrix spike and high
bias flags (FMS and HIB). Analytical results were considered invalid (INV) and qualified when the
analyte was undetected and recoveries for associated matrix spike samples were less than 10%. Some
PCB data were found to be invalid for this reason, but the fra«s-nonachlor surrogate recoveries were
never below 10%.
To characterize contamination associated with field and analytical activities, field blanks were obtained
for precipitation, particulate, and vapor samples at a subset of monitoring stations, and for dry deposition
strips. For precipitation, particulate, and vapor samples, filters and/or absorbent were installed in the
samplers for the normal sampling period but were not exposed to precipitation or air flow. The
precipitation field blank included a water rinse of the collector surfaces to check for contamination by
dry-deposited material that might have penetrated the cover and seal on the precipitation collector. Field
blanks were not collected at all stations and potential station-specific contamination associated with these
sites cannot be evaluated. However, contamination associated with sample collection, sampling
equipment, sample processing, shipping, storing, and analysis can be evaluated based on the field blanks
collected throughout the study.
PCB congeners were detected in all field blanks in all sample phases. This is not unexpected given the
ubiquitous nature of PCBs in the environment. For fra«s-nonachlor determinations, 80% of the field
blank results contained detectable concentrations of fra«s-nonachlor. When field blank concentrations
were within a factor of five of the concentration in an associated sample, the sample result was qualified
with the failed field blank flag (FFR) and also with the high bias flag (HIB). When sample concentrations
were indistinguishable from the associated field blank concentration, samples were determined to be
invalid and were qualified as such (INV). For PCBs in dry deposition, 30% of sample results were
qualified as invalid, based on field blank contamination. For trans-nonachlor, 23% of the sample results
for all phases were qualified as invalid, based on field blank contamination. Due to contaminated field
blanks and variable sample results for field duplicates, many of the PCB results for samples collected
from the bow of the R/V'Lake Guardian in 1994 were determined to be invalid and are qualified as such,
based on the potential for a shipboard source of PCB contamination (Miller, 1999). Samples collected
from the yardarm of the R/VLake Guardian and samples collected in 1995 were not affected.
For dry deposition samples, field blanks were collected with each sample and exposed to ambient
conditions only for the length of time (<30 minutes) required to set up the routine dry deposition samples.
Once the strips for the routine samples were set up, the field blank strips were returned to their sealed
containers for the duration of the routine sampling episode. Because each sample had an associated field
blank, the surface-area-normalized deposition of PCBs was subtracted from the corresponding field
sample to minimize artifacts associated with sorption of gas-phase organic chemicals to the grease coating
the deposition strips. When normalized to exposed surface area, the field blanks for land sites (n=35)
3-44 April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
averaged 5100 ng/m2 for total PCBs and the blanks for over-water sites (n=l 1) averaged 4700 ng/m2
(Franzee/'a/., 1998).
Laboratory blanks were prepared and analyzed for PCBs and fra«s-nonachlor in all phases. PCB
congeners were detected in all laboratory blanks in all sample phases. For example, all results reported
for PCB 1 for samples analyzed at Illinois Water Survey were determined to be invalid by the PI, based
on erratic laboratory background. Thus, the PCB 1 results were not included in the repotted total PCB
concentrations. When laboratory blank contamination was greater than the method detection limit, all of
the associated results were qualified with failed blank sample (FBS) and high bias (HIB). No laboratory
blanks contained fra«s-nonachlor above the method detection limit.
As discussed in Section 2.7, data verification was performed by comparing all field and QC sample
results produced by each PI with the MQOs and with overall LMMB Study objectives. Analytical results
were qualified when pertinent QC sample results did not meet the acceptance criteria defined by the
MQOs. These flags were not intended to suggest that data were not useable; rather they were intended to
caution the user about an aspect of the data that did not meet the predefined criteria. Tables 3-22 to 3-24
summarize the flags applied to the atmospheric PCB and fra«s-nonachlor data generated by each of the
Pis involved in the analysis of atmospheric samples. Table 3-23 addresses data generated at the Illinois
Water Survey. Table 3-23 addresses data generated at Indiana University. Table 3-24 addresses data
generated for dry deposition samples. Qualifier flags were not applied to the data for total PCBs because
they were the results of calculations rather than laboratory analyses.
The summary tables include the flags that directly relate to evaluation of the MQOs to illustrate some
aspects of data quality, but do not include all flags applied to the data to document sampling and
analytical information, as discussed in Section 2.7. As noted throughout this report, given the large
number of PCB congeners that were determined by the investigators in this study, it is not practical to
summarize the results for every congener.
As illustrated in Tables 3-23 and 3-24 and discussed in previous sections, PCB congeners and trans-
nonachlor were not detected in a substantial portion of precipitation samples. Particulate and vapor
samples more frequently contained PCB congeners and fra«s-nonachlor above detection limits than
precipitation samples. Fifty-one percent of precipitation samples analyzed at Illinois Water Survey (Table
3-22) contained PCB 33 below detection limits and were qualified with the MDL flag (less than method
detection limit) and 25% were qualified with the UND flag (analyte not detected). Sixty-four percent of
precipitation samples analyzed at Indiana University (Table 3-23) contained PCB 180 below detection
limits and were qualified with the MDL flag and PCB 180 was not detected in 21% of precipitation
samples.
Of the three PCB congeners, PCB 33 was most frequently below detection limits for precipitation samples
analyzed at Illinois Water Survey, whereas PCB 180 was most frequently below detection limits for
precipitation samples analyzed at Indiana University.
Apr/72004 3-45
-------
Table 3-22. Field Sample Flags Applied to Select PCB Congeners and frans-Nonachlor Results in Atmospheric Samples Analyzed at Illinois Water
Survey
Analyte
PCB 33
PCB 118
PCB 180
frans-nonachlor
Fraction
Particulate
Precipitation
Vapor
Particulate
Precipitation
Vapor
Particulate
Precipitation
Vapor
Particulate
Precipitation
Vapor
Flags
Sensitivity
MDL
5% (11)
51% (91)
0.8% (2)
2% (5)
11% (20)
0
8% (18)
28% (50)
1%(4)
35% (75)
78% (151)
8% (21)
UNO
0.5% (1)
25% (44)
0.8% (2)
0
0.6% (4)
0
0.9% (2)
5% (8)
1%(3)
16% (34)
55% (106)
7% (19)
Contamination
FFR
0.5% (1)
0
0
1%(3)
0
0
0.5% (1)
0
0.4% (1)
0
0
0.7% (2)
FBS
0
0
3% (8)
0
5% (8)
16% (42)
11% (23)
10% (18)
34% (91)
0
0
0
Holding Time
EHT
0.5% (1)
46% (82)
5% (13)
0.5% (1)
46% (81)
5% (13)
0.5% (1)
46% (82)
5% (13)
0.5% (1)
0.5% (1)
0.7% (2)
Precision
FFD
0
1%(2)
0.4% (1)
0
0.6% (1)
1%(4)
0
0.6% (1)
1%(3)
0
0
0.7% (2)
Bias
FPC
2% (5)
0
6% (16)
2% (5)
0
6% (16)
2% (5)
0
6% (16)
0
0
0
FSS
0.5% (1)
1%(2)
0.8% (2)
0.5% (1)
0.6% (1)
1%(3)
0.5% (1)
0.6% (1)
1%(3)
0
0
0
FMS
4% (8)
5% (9)
0
0
0
0
0
0
0
0
14% (27)
3% (8)
HIB
0
0
0
0
0
0
0
0
0.4% (1)
0
0
0
Invalid
INV
5% (11)
0
8% (19)
5% (11)
0
7% (19)
5% (11)
0
7% (19)
0
0
0
The number of routine field samples flagged is provided in parentheses. The summary provides only a subset of applied flags and does not represent the full suite of flags applied to the data.
MDL = Less than method detection limit (Analyte produced an instrument response but reported value is below the calculated method detection limit. Validity of reported value may be
compromised.)
UNO = Analyte not detected (Analyte produced no instrument response above noise.)
FFR = Failed field blank (A field blank sample, type unknown, associated with this analysis failed the acceptance criteria. It is unknown whether the blank that failed was a field blank or a lab
blank. Validity of reported value may be compromised.)
FBS = Failed blank sample (A blank sample associated with this analysis failed the acceptance criteria. It is unknown whether the blank that failed was a field blank or a lab blank. Validity of
reported value may be compromised.)
EHT = Exceeded holding time (Sample or extract was held longer than the approved amount of time before analysis. Validity of reported value may be compromised.)
FFD = Failed field duplicate (A field duplicate associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FPC = Failed performance check (A laboratory performance check sample associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FSS = Failed surrogate (Surrogate recoveries associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FMS = Failed matrix spike (A matrix spike associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
HIB = Likely biased high (Reported value is probably biased high as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery, blank contamination, or other
internal lab QC data. Reported value is not considered invalid.)
INV = Reported value is deemed invalid by the QC Coordinator.
3-46
-------
Table 3-23. Field Sample Flags A
Analyte
PCB33
PCB118
PCB 180
frans-nonachlor
Fraction
Particulate
Precipitation
Vapor
Particulate
Precipitation
Vapor
Particulate
Precipitation
Vapor
Particulate
Precipitation
Vapor
pplied to Select PCB Congener and frans-Nonachlor Results in Atmospheric Samples Analyzed at Indiana University
Flags
Sensitivity
MDL
7%(1)
7%(1)
0
43% (6)
57% (8)
4%(1)
21% (3)
64% (9)
4%(1)
25% (1)
79% (11)
26% (9)
UNO
0
0
0
0
0
0
36% (5)
21% (3)
31% (8)
50% (2)
21% (3)
0
Contamination
FFR
21% (3)
7%(1)
19% (5)
14% (2)
7%(1)
23% (6)
0
7%(1)
12% (3)
0
7%(1)
0
FBS
0
14% (2)
62% (16)
1
1
6
7%(1)
7%(1)
8% (2)
0
0
0
Holding Time
EHT
0
0
0
0
0
0
0
0
0
0
0
3%(1)
Precision
FFD
0
7%(1)
0
0
0
12% (3)
0
7%(1)
23% (6)
0
0
0
Bias
FPC
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
FSS
0
0
0
0
0
0
0
0
0
25% (1)
0
15% (5)
FMS
0
0
0
0
0
0
0
0
8% (2)
0
29% (4)
5
LOB
0
7%(1)
4%(1)
0
7%(1)
4%(1)
0
7%(1)
4%(1)
0
7%(1)
3%(1)
Invalid
HIB
0
0
19% (5)
0
0
12% (3)
0
0
0
0
0
6% (2)
The number of routine field samples flagged is provided in parentheses. The summary provides only a subset of applied flags and does not represent the full suite of flags applied to the data.
MDL = Less than method detection limit (Analyte produced an instrument response but reported value is below the calculated method detection limit. Validity of reported value may be
compromised.)
UNO = Analyte not detected (Analyte produced no instrument response above noise.)
FFR = Failed field blank (A field blank sample, type unknown, associated with this analysis failed the acceptance criteria. It is unknown whether the blank that failed was a field blank or a lab
blank. Validity of reported value may be compromised.)
FBS = Failed blank sample (A blank sample associated with this analysis failed the acceptance criteria. It is unknown whether the blank that failed was a field blank or a lab blank. Validity of
reported value may be compromised.)
EHT = Exceeded holding time (Sample or extract was held longer than the approved amount of time before analysis. Validity of reported value may be compromised.)
FFD = Failed field duplicate (A field duplicate associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FPC = Failed performance check (A laboratory performance check sample associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FSS = Failed surrogate (Surrogate recoveries associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FMS = Failed matrix spike (A matrix spike associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
LOB = Likely biased low (Reported value is probably biased low as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery or other internal lab QC data.
Reported value is not considered invalid.)
HIB = Likely biased high (Reported value is probably biased high as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery, blank contamination, or other
internal lab QC data. Reported value is not considered invalid.)
NA = This flag was not applied to this data set or this type of QC sample was not prepared and analyzed.
3-47
-------
Table 3-24. Field Sample Flags Applied to Select PCB Congener and frans-Nonachlor Results in Dry Deposition Atmospheric Samples
Analyte
PCB118
PCB180
frans-nonachlor
Flags
Sensitivity
MDL
37% (15)
0
70% (28)
UNO
0
0
8% (3)
Contamination
FFB
68% (28)
90% (38)
48% (19)
EHT
2%(1)
2%(1)
3%(1)
Holding Time
FFD
0
0
0
Precision
FMS
49% (20)
48% (20)
10% (4)
Bias
FSS
0
0
0
HIB
27% (11)
48% (20)
0
FPC
0
0
0
Invalid
INV
68% (28)
90% (38)
48% (19)
The number of routine field samples flagged is provided in parentheses. The summary provides only a subset of applied flags and does not represent the full suite of flags applied to the data.
MDL = Less than method detection limit (Analyte produced an instrument response but reported value is below the calculated method detection limit. Validity of reported value may be
compromised.)
UNO = Analyte not detected (Analyte produced no instrument response above noise.)
FFB = A field matrix blank associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.
EHT = Exceeded holding time (Sample or extract was held longer than the approved amount of time before analysis. Validity of reported value may be compromised.)
FFD = Failed field duplicate (A field duplicate associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FMS = Failed matrix spike (A matrix spike associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FSS = Failed surrogate (Surrogate recoveries associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
HIB = Likely biased high (Reported value is probably biased high as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery, blank contamination, or other
internal lab QC data. Reported value is not considered invalid.)
FPC = Failed performance check (A laboratory performance check sample associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
INV = Reported value is deemed invalid by the QC Coordinator.
Note: PCB 33 was not determined in the dry deposition samples.
3-48
-------
PCBs/trans-Nonachlor in Atmospheric Components
For dry deposition samples, 37% contained PCB 118 below detection limits and were qualified with the
MDL flag. However, PCB 118 was detected in all of the samples (i.e., none of the sample results were
qualified with the undetected flag). For fra«s-nonachlor, 78% of the precipitation samples analyzed at
Illinois Water Survey contained fra«s-nonachlor below detection limits and were qualified the MDL flag
and in 55% of precipitation samples fra«s-nonachlor was not detected. Seventy-nine percent of
precipitation samples analyzed at Indiana University contained fra«s-nonachlor below detection limits
and were qualified the MDL flag and in 21% of samples fra«s-nonachlor was not detected. Seventy
percent of dry deposition samples contained fra«s-nonachlor below detection limits and in 8% of dry
deposition samples, fra«s-nonachlor was not detected.
A substantial portion of precipitation samples were flagged for exceeding sample holding times for
determination of PCB congeners. For example, 46% of precipitation samples analyzed for PCB
congeners at Illinois Water Survey exceeded the established holding time. However, the holding times
for PCBs and many other environmental pollutants are not well-established and the effects on the sample
results generally are not known. PCBs are highly stable compounds and loss or degradation is considered
minimal even after 12 months if the samples are stored frozen. Loss or degradation rates may vary among
congeners, depending on the mechanism (e.g., biological transformation, evaporative loss, or photolysis).
To characterize contamination associated with field and analytical activities, field blanks were obtained
for precipitation, particulate, and vapor samples at a subset of monitoring stations, and for dry deposition
strips.
For the analyzed at Illinois Water Survey, 1% or less of sample results for PCBs and fra«s-nonachlor
were associated with field blanks that showed significant contamination and were qualified with the failed
field blank flag. A single result for PCB 180 in a vapor sample also was qualified with the high bias flag
due to this contamination. For the vapor samples analyzed at Indiana University, 23 % or less of sample
results were associated with field blanks that showed significant contamination and were qualified with
the failed field blank and high bias flags. Laboratory blanks also showed contamination for PCB
congeners in all phases. Vapor samples were most frequently flagged for laboratory blank contamination
with 62% of vapor sample results for PCB 33 generated at Indiana University being qualified with the
failed laboratory blank flag and 34% of sample results for PCB 180 generated at Illinois Water Survey
being qualified with the failed laboratory blank flag. As a result of contamination, 19% of the vapor
sample results for PCB 33 also were qualified the high bias flag.
Contamination was a significant issue for PCB congeners and fra«s-nonachlor in dry deposition samples.
The majority of dry deposition sample results are flagged for contamination and invalidated. Invalid
sample results will not be used in the LMMB model. Due to the ubiquitous nature of PCBs,
contamination can be an issue when analyzing samples with concentrations close to the method detection
limit. Overall, the large majority of sample results were not affected by contamination.
Field duplicates were collected and analyzed for all phases. Although field duplicates were not planned
for dry deposition, one field duplicate was collected. However, the dry deposition field duplicate was not
used to evaluate study data. For the samples analyzed at Illinois Water Survey, less than 1% of all the
field samples had associated duplicates with results outside the MQO limit and thus were qualified with
the failed field duplicate flag. For the samples analyzed at Indiana University, 23% of vapor samples
analyzed for PCB 180 had associated field duplicates with results outside the MQO limit and were
qualified with the failed field duplicate flag. Only one precipitation sample for PCB 180 and one for PCB
33 had an associated field duplicate outside the MQO limit.
Matrix spike samples results showed acceptable results for the large majority of study samples in all
phases, except for dry deposition. For samples analyzed at Illinois Water Survey, all of the spike
April 2004 3-49
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
recoveries for all sample results for PCB 118 and PCB 180 in all phases were within MQO limits. For
fra«s-nonachlor, 14% of the precipitation sample results were qualified with the failed matrix spike flag
for samples analyzed at Illinois Water Survey. For samples analyzed at Indiana University, all of the
spike recoveries for all sample results for PCB 33 and PCB 118 in all phases were within MQO limits.
For fra«s-nonachlor, 29% of precipitation sample results were qualified with the failed matrix spike flag
for samples analyzed at Indiana University. For analysis of PCBs and trans-nonachlor in dry deposition
samples, matrix effects presented analytical difficulties and 48% of the sample results were qualified with
the failed matrix spike flag for PCB 180 and 10% for fra«s-nonachlor.
Surrogate recoveries indicated acceptable results for the large majority of study samples. Of the results
for PCBs 33, 118, and 180 generated at the Illinois Water Survey, at most 1% of the vapor, precipitation,
and particulate samples were qualified with the failed surrogate spike flag (FSS). None of the trans-
nonachlor results generated at the Illinois Water Survey failed the surrogate recovery limits. None of the
results for PCBs 33, 118, or 180 analyzed at Indiana University failed the surrogate recovery limits, but
15% of the vapor samples and 25% of the particulate-phase samples analyzed there for fra«s-nonachlor
failed the surrogate recovery limits. As noted earlier in this section, the surrogate recoveries for the dry
deposition results were erratic and subject to interferences. As a result, no results for dry deposition
samples were qualified due to surrogate recovery problems and no surrogate recovery correction was
applied to the dry deposition results for any analyte.
As discussed in Section 1.5.5, MQOs were defined in terms of six attributes: sensitivity, precision,
accuracy, representativeness, completeness, and comparability. GLNPO derived its data quality
assessments based on a subset of these attributes. For example, system precision was estimated as the
mean relative percent difference (RPD) between the results for field duplicate pairs. Similarly, analytical
precision was estimated as the mean relative percent difference (RPD) between the results for laboratory
duplicate pairs. Tables 3-25 to 3-28 provide summaries of data quality assessments for several of these
attributes for the data for atmospheric PCB congener 33, 118, and 180 and frara-nonachlor.
Data quality assessments were conducted for two separate sets of results: those sample results that are
above 5 times the sample specific detection limits and those sample results that are below 5 times the
sample specific detection limits. Performing separate assessment illustrates the expected differences in
data quality for these sample groups that are due to increased variability in the analytical results when
sample concentrations are close to the detection limit of the analytical method. In addition, MQOs often
were set and applied differently for sample results that are greater than 5 times the detection limit versus
those results that are less than five times the method detection limit.
As discussed in this chapter, a significant number of sample results were below the calculated analytical
sample specific method detection limits. Seven percent of precipitation samples analyzed at Indiana
University were reported below the MDL and 51% of precipitation samples analyzed at Illinois Water
Survey were reported as below the MDL. For PCB 180, 86% of precipitation samples analyzed at
Indiana University were below the sample specific detection limits. For study samples that are below
detection limits, the variability of sample results is expected to be greater than for sample results that are
significantly above the detection limit and in the middle of the calibration range of the analytical method.
Data quality assessments based on field duplicates and laboratory duplicates also will reflect increased
variability because the sample concentrations are close to or below the detection limits.
System precision, estimated as the mean RPD between field duplicate results, varied by fraction. For
example, for PCB 118 particulate-phase samples analyzed at Illinois Water Survey, with results greater
than 5 times their associated sample specific MDL, had a mean RPD of 21% and vapor samples analyzed
at Illinois Water Survey, with results greater than 5 times their associated sample specific MDL, had a
mean RPD of 54%. System precision also varied by the relationship of sample concentrations to the
3-50 April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
sample specific detection limits. For example, for vapor-phase samples analyzed at Indiana University
for PCB 180, the mean RPD was 42% for sample results less than 5 times the sample specific detection
limit and was 22% for sample results greater than 5 times the sample specific detection limit. Similarly,
for vapor-phase samples analyzed at Indiana University for fra«s-nonachlor, the mean RPD was 53% for
sample results less than 5 times the sample specific detection limit and was 16% for sample results greater
than 5 times the sample specific detection limit.
In some cases, the system precision was unexpectedly greater for sample results at higher concentrations.
For example, for vapor samples analyzed at Illinois Water Survey, the mean RPD was 21% for sample
results less than 5 times the sample specific detection limit and was 40% for sample results greater than 5
times the sample specific detection limit. System precision also varied by congener, although a pattern
among congeners was not evident. For example, for precipitation samples analyzed at Indiana University,
the mean RPD was 34% for PCB 118 and was 53% for PCB 180, whereas for particulate samples, the
mean RPD was 55% for PCB 118 and was 23% for PCB 180. Duplicate pair samples with a reported
concentration of zero for either one of both samples could not be used in this assessment. Because of the
large number of results reported as zero, the system precision estimate is based on only a small number of
field duplicates and may not accurately reflect the system.
Analytical precision, estimated as the mean RPD between laboratory duplicates, only could be estimated
for a single particulate sample analyzed at Indiana University (a small number of laboratory duplicates
prepared and analyzed for these analytes in the study because of the large expense of these analyses). The
RPD between laboratory duplicates of particulate sample was 13% for PCB 33 and 34% for PCB 118.
This analytical precision for particulate samples was lower than the mean RPD for system precision,
where the mean RPD was 31% for PCB 33 and 55% for PCB 118.
Evaluation of matrix spike sample (LMS) recoveries shows a slight low bias overall for all PCB
congeners for phases of atmospheric samples, except dry deposition. For vapor samples, the mean LMS
recovery for PCB 33 was 93% for samples analyzed at Indiana University and at the Illinois Water
Survey. For PCB 180 in vapor samples, the mean LMS recovery was 94% for samples analyzed at
Illinois Water Survey and 96% for samples analyzed at Indiana University. For precipitation samples,
the mean LMS recovery for PCB 33 was 86% for samples analyzed at Illinois Water Survey and 91% for
samples analyzed at Indiana University. For fra«s-nonachlor, the low bias was more pronounced. For
example, for vapor samples, the mean LMS recovery was 80% for samples analyzed at Illinois Water
Survey and 70% for samples analyzed at Indiana University.
For dry deposition, a significant high bias was observed based on results of the LMS samples. For PCBs
118 and 180, the mean LMS recoveries were 626% and 10,798 % . For fra«s-nonachlor, the bias was
much less extreme, with a mean LMS recovery of 126%. Sample results associated with LMS results that
were outside the MQO were qualified with the failed matrix spike and high bias flags. As discussed
above, a significant portion of sample results for dry deposition were determined to be invalid based on
results of LMS and other QC samples. For the majority of PCB and fra«s-nonachlor results, the PI and
QC coordinator determined that the bias demonstrated by results for the LMS and other QC samples was
not strong enough to warrant flagging the data as either HIB or LOB. As discussed above, the sample
results that are flagged HIB are due, in part, to contamination in field and laboratory blanks.
April 2004 3-51
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 3-25. Data Quality Assessment for PCB 33 in Atmospheric Samples
Fraction (Lab)
Vapor (Illinois)
Precipitation (Illinois)
Particulate (Illinois)
Vapor (Indiana)
Precipitation (Indiana)
Particulate (Indiana)
Number of
Routine
Samples
Analyzed
251
179
208
26
14
14
System Precision Mean
Field Duplicate RPD (%)
<5*MDL
(0)
72% (3)
31% (10)
(0)
23% (11)
31% (8)
>5*MDL
34% (14)
(0)
(0)
16% (13)
(0)
(0)
Analytical
Precision
Mean Lab
Duplicate RPD
< 5* MDL
(0)
(0)
(0)
(0)
(0)
13% (1)
Bias
Mean LMS
Recovery (%)
93% (33)
86% (23)
82% (28)
93% (14)
91% (14)
90% (13)
Sensitivity
Samples
reported as
< MDL (%)
0.80%
51%
4.8%
(0)
7.1%
7.1%
The number of QC samples used in the assessment is provided in parentheses.
RPD= Relative percent difference
LMS= Laboratory matrix spike
Table 3-26. Data Quality Assessment for PCB 118 in Atmospheric Samples
Fraction (Lab)
Vapor (Illinois)
Precipitation (Illinois)
Particulate (Illinois)
Vapor (Indiana)
Precipitation (Indiana)
Particulate (Indiana)
Dry deposition
Number of
Routine
Samples
Analyzed
251
179
208
26
14
14
13
System Precision Mean
Field Duplicate RPD (%)
<5*MDL
(0)
43% (8)
28% (8)
45% (7)
34% (11)
55% (8)
(0)
>5*MDL
54% (14)
(0)
21% (2)
34% (6)
(0)
(0)
(0)
Analytical
Precision
Mean Lab
Duplicate RPD
< 5* MDL
(0)
(0)
(0)
(0)
(0)
34% (1)
(0)
Bias
Mean LMS
Recovery (%)
97% (34)
94% (23)
94% (28)
96% (14)
93% (14)
95% (13)
626% (5)
Sensitivity
Samples
reported as
< MDL (%)
0
11%
2.4%
3.9%
57%
43%
69%
The number of QC samples used in the assessment is provided in parentheses.
RPD= Relative percent difference
LMS= Laboratory matrix spike
3-52
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Table 3-27. Data Quality Assessment for PCB 180 in Atmospheric Samples
Fraction (Lab)
Vapor (Illinois)
Precipitation (Illinois)
Particulate (Illinois)
Vapor (Indiana)
Precipitation (Indiana)
Particulate (Indiana)
Dry deposition
Number of
Routine
Samples
Analyzed
251
179
208
26
14
14
13
System Precision Mean
Field Duplicate RPD (%)
<5*MDL
21%(3)
47% (9)
33% (9)
42% (3)
53% (7)
23%(5)
(0)
>5*MDL
40% (11)
(0)
10% (1)
22% (2)
(0)
(0)
(0)
Analytical
Precision
Mean Lab
Duplicate RPD
< 5* MDL
(0)
(0)
(0)
(0)
(0)
(0)
(0)
Bias
Mean LMS
Recovery (%)
94% (34)
93% (23)
97% (28)
96% (14)
92% (14)
95% (13)
10,798% (5)
Sensitivity
Samples
reported as
< MDL (%)
1.6%
29%
8.2%
35%
86%
57%
0
The number of QC samples used in the assessment is provided in parentheses.
RPD= Relative percent difference
LMS= Laboratory matrix spike
Table 3-28. Data Quality Assessment for frans-Nonachlor in Atmospheric Samples
Fraction (Lab)
Vapor (Illinois)
Precipitation (Illinois)
Particulate (Illinois)
Vapor (Indiana)
Precipitation (Indiana)
Particulate (Indiana)
Dry deposition
Number of
Routine
Samples
Analyzed
270
193
217
34
14
4
21
System Precision Mean
Field Duplicate RPD (%)
<5*MDL
all results =0
all results = 0
52% (10)
53% (10)
66% (8)
all results = 0
(0)
>5*MDL
24% (23)
(0)
(0)
16% (2)
(0)
(0)
(0)
Analytical
Precision
Mean Lab
Duplicate RPD
< 5* MDL
(0)
(0)
(0)
(0)
(0)
54% (1)
(0)
Bias
Mean LMS
Recovery (%)
80% (36)
73% (23)
75% (28)
70% (18)
75% (14)
70% (13)
126% (5)
Sensitivity
Samples
reported as
< MDL (%)
7.8%
78%
34%
27%
100%
75%
81%
The number of QC samples used in the assessment is provided in parentheses.
RPD= Relative percent difference
LMS= Laboratory matrix spike
April 2004
3-53
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
3.3 Data Interpretation
3.3.1 Atmospheric Sources
Atmospheric sources of PCBs and trans-nonachlor to Lake Michigan can include exchange of
contaminants from the vapor phase to the water, deposition of contaminants associated (e.g., bound or
sorbed) with particulates, and contaminants in precipitation. This report summarizes the concentrations
of PCB congeners, total PCBs, and fra«s-nonachlor reported during the LMMB Study in each of these
atmospheric phases: vapor, particulate, dry deposition, and precipitation. The data for these phases will
be used in the LMMB modeling efforts to evaluate the fluxes and loads of contaminants from each of
these sources.
PCBs and trans-nonachlor were detected in samples from all four atmospheric phases. PCBs were
detected above the sample-specific detection limits most frequently in the vapor-phase samples, followed
by the particulate-phase samples, and least often in the precipitation samples. This is consistent with the
findings of researchers that have estimated the atmospheric fluxes of PCB from the vapor and particulate
phase as much higher than from wet deposition (Franz et al., 1998).
While PCBs were detected in all four atmospheric phases, the frequency of occurrence and magnitude of
concentrations differed among the PCB congeners and among the four phases. Figure 3-19 shows the
mean percentage of individual PCB congeners that contributed to the total PCB concentration in vapor,
particulate, and precipitation phases. In the vapor phase, the lower molecular weight, less-chlorinated
congeners predominated, while particulate and precipitation phases contained a more diverse mixture of
PCB congeners, including higher molecular weight congeners. Baker and Eisenreich (1990) also found
that the more volatile tri- and tetrachlorobiphenyl congeners dominated the distribution of PCBs in the
atmospheric gas phase. The higher vapor pressures of these less-chlorinated PCB congeners favor
volatilization from the particulate and dissolved phases to the vapor phase. The differences in vapor
pressures among the PCB congeners may explain some of the differences in the frequencies of occurrence
of PCB congeners in the various phases. For example, the more volatile PCB 33 was less often detected
in the precipitation phase than the less volatile, more chlorinated PCB 180 (Figure 3-20).
To statistically evaluate the contribution of the various PCB congeners, the proportion of the total PCB
concentration attributed to the higher chlorinated (high molecular weight) congeners was calculated for
vapor-phase, particulate-phase, and precipitation samples. For the purposes of this evaluation, the higher
chlorinated congeners include the hexachloro-, heptachlor-, octachloro, and nonachlorobiphenyl
congeners, while the lower chlorinated congeners include the dichloro-, trichloro-, tetrachloro-, and
pentachlorobiphenyl congeners. Neither the three monochlorobiphenyl congeners nor the one
decachlorobiphenyl congener were reported by all laboratories in all phases, so they were not included in
this calculation.
The high molecular weight proportion was calculated as:
Vrri + n + n + TM
Z^\ 6 7 8 Q/
High molecular weight proportion =
C13 + C14 + C15 + C16 + C17 + C18 + C19)
3-54 April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Figure 3-19. Mean Percentage of Individual PCB Congener's Contribution to Total
PCB Concentrations
o>
O)
* DO
c O
d> Q.
O _
5 S
*t
re
o>
t 10
• 10
- 8
C O
0) Q.
O _
o> S
Q. O
re
o>
PCB Congener
8
6
P articulate
PCB Congener
Precipitation
PCB Congener
April 2004
3-55
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 3-20. Percentages of PCBs 33 and 180 in Precipitation Samples Reported as Zero, Below the
Detection Limits, and Above the Detection Limits
PCB 33 in precipitation
Reported as zero
23%
Reported above
detection limit —
52%
PCB 180 in precipitation
Reported as zero
6%
Reported above zero,
\_ but below detection
limit
25%
Reported above zero,
but below detection
limit
27%
There were three instances in which two or more congeners in the two categories coeluted and could not
be reported separately (PCBs 123 and 149; PCBs 111 and 131; and PCBs 105, 132, and 153), therefore,
none of the results for these congeners were used to calculate the proportions.
Based on a comparison of high molecular weight proportions in over 180 vapor, particulate, and
precipitation samples, there was a significant difference among the three phases (Figure 3-21).
Precipitation samples had the highest proportion of high molecular weight PCB congeners, followed by
particulate-phase samples. Vapor-phase samples contained the lowest proportion of high molecular
weight PCB congeners, consistent with the lower vapor pressures of the higher molecular weight
congeners.
Similarly to the low molecular weight PCB congeners, fra«s-nonachlor, which has a high vapor pressure
also was more prevalent in the vapor phase than in precipitation or the particulate phase. Mean vapor-
phase concentrations of fra«s-nonachlor were approximately 10 to 20 times higher than the particulate-
phase concentrations at the same sampling stations (see Tables 3-7 and 3-13). fra«s-Nonachlor was even
less common in precipitation samples. With the exception of the IIT Chicago site, 75 to 100% of
precipitation samples from the various LMMB sampling stations contained fra«s-nonachlor below the
sample-specific detection limit.
3.3.2 Atmospheric Concentrations
In this study, total PCB concentrations in the vapor phase ranged from 0 to 6300 pg/m3 at shoreline
sampling stations surrounding Lake Michigan. Average monthly composite concentrations of total PCBs
ranged from 320 to 2600 pg/m3 at these stations. These concentrations are comparable to vapor-phase
total PCB concentrations measured by other researchers over large water bodies near urban influences.
Brunciak et al. (2001) measured average vapor-phase total PCB concentrations of 1180 pg/m3 near
Baltimore, MD and 550 pg/m3 over the Northern Chesapeake Bay. Similarly, vapor-phase total PCB
concentrations ranged from 210 to 4780 pg/m3 over Galveston Bay, TX (Park et al, 2001). The lower
PCB concentrations measured at the more remote out-of-basin sampling stations in the LMMB Study
(averaging 110 to 260 pg/m3) were comparable to concentrations measured by other researchers at remote
sites across the northern hemisphere. Iwata et al. (1993) measured average vapor-phase total PCB
concentrations of 93, 130, 130, 320, and 290 pg/m3 over the Bering Sea, Gulf of Alaska, North Pacific
Ocean, Caribbean Sea, and North Atlantic Ocean, respectively.
3-56
April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Figure 3-21. Proportion of High Molecular Weight PCB Congeners in the
Vapor Phase, Particulate Phase, and Precipitation
CO
o
Q_
O
.Ł=
D)
I
8.
o
CL
1.00-,
-
0.75-
_
-
0.25-
-
.UIH
B
A
g
• 1
1
1
I 1
+
3 <
p- 8 -s
II
00
Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile (top of box) results. Bars represent the
results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest edge of the box. Circles represent results
beyond 1.5*IQR from the box. The As represent results beyond 3*IQR from the box. The letters (A - D) above the boxes represent the results
of the analysis of variance and multiple comparisons test. Boxes with the same letter were not statistically different (at alpha = 0.05).
Proportions were transformed by calculating the arcsine of the square root of the proportion prior to testing.
The range of vapor-phase total PCB concentrations measured in the LMMB Study during 1994 and 1995
also are relatively consistent with PCB concentrations historically measured in the Great Lakes region.
Baker and Eisenreich (1990) reported average atmospheric PCB concentrations over the Great Lakes of
300 to 3200 pg/m3 with no discernable trend from 1977 through 1986. Baker and Eisenreich (1990)
concluded that PCB concentrations have remained relatively constant in the atmosphere over this period.
In 1986, Baker and Eisenreich (1990) measured an average PCB concentration of 1200 pg/m3 over Lake
Superior. Hoffet al. (1992) measured monthly averages of 55 to 823 pg/m3 over Southern Ontario in
1988 to 1989. Also in 1989, Hornbuckle et al. (1993) measured average total PCB concentrations of 670
to 2200 pg/m3 over southern Green Bay, and 160 to 520 pg/m3 over northern Green Bay in 1989. From
1991 to 1993, Hornbuckle et al. (1995) measured 30 to 400 pg/m3 of total PCB in air samples collected at
the Sleeping Bear Dunes site. PCB concentrations measured in the LMMB Study during 1994 and 1995
are consistent with the range of previous measurements and do not clearly suggest trends of increasing or
decreasing PCB concentrations in the vapor phase, however, as HoSet al. (1992) noted, such long-term
trends may be difficult to detect due to the large amplitude of seasonal cycles and due to differences in
sampling locations and analytical methodologies in the various studies. Using longer-term monitoring
data and consistent IADN (Integrated Atmospheric Deposition Network) monitoring stations, Simcik et
al. (1999) were able to detect decreases in vapor-phase total PCB concentrations from 1991 to 1997 over
Lake Michigan and Lake Erie, with half-lives of 2.8 to 3.3 years.
Total PCB concentrations in precipitation averaged 1.3 to 16 ng/L at shoreline sampling stations and 0.29
to 1.7 ng/L at out-of-basin stations during the LMMB Study. These values also were relatively consistent
with total PCB concentrations measured in precipitation over other large water bodies across the U.S.
Leister and Baker (1994) measured a volume-weighted mean total PCB concentration of 1.6 ng/L in
April 2004
3-57
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
precipitation over the Chesapeake Bay, and Park et al. (2001) measured total PCB concentrations of 0.08
to 3.34 ng/L in precipitation over Galveston Bay, TX. Total PCB concentrations in snowfall at Lake
Tahoe ranged from 4.8 to 5.1 ng/L (Datta et al., 1998). In the Great Lakes region, Simcik et al. (2000)
reported lower PCB concentrations in precipitation over Lake Michigan and Lake Erie than over Lake
Huron and Lake Ontario. Simcik et al. (2000) also reported significant decreases in PCB concentrations
in precipitation over Lake Michigan from 1991 to 1997, with a half life of 6.9 years.
3.3.3 Seasonality
Concentrations of PCBs in the vapor phase were highly influenced by season (Figure 3-2). Total PCB
concentrations and individual congeners (PCB 118 and PCB 180) in the vapor phase peaked in mid
summer and reached minimum values in the winter. The mean total PCB concentration during the
summer at shoreline and out-of-basin stations was eight times higher than in the winter. fra«s-Nonachlor
followed the same trend, with vapor-phase mean concentrations in summer ten times higher than in the
winter. This finding is consistent with other researchers who have measured PCBs and chlordanes over
annual cycles (Hoff etal, 1992; Greened al., 2000). Hoff et al. (1992) concluded that this seasonal
pattern was due to volatilization and the effects of temperature and vapor pressures on the distribution of
these organic contaminants in the vapor phase. As temperature increases, distribution of these organic
contaminants within the atmosphere favors the vapor phase. Fluxes of PCBs from Lake Michigan to the
atmospheric vapor phase also may increase vapor-phase PCB concentrations during the summer
(Hornbucklee/'a/., 1993).
In contrast, the particulate-phase total PCB concentrations exhibited a different seasonal pattern than the
vapor-phase results (Figure 3-8). Particulate-phase total PCB concentrations were higher in spring and
winter than in summer and autumn. The particulate-phase fra«s-nonachlor concentrations exhibited a
different pattern than the corresponding vapor-phase results (Figure 3-9), and slightly different from the
pattern for the particulate-phase total PCBs, with the highest concentrations in the winter. These patterns
for total PCBs and fra«s-nonachlor also are consistent with the effects of temperature and vapor
pressures, in that at the lower temperatures during winter, the contaminants are less likely to volatilize off
of the surface of particulates into the vapor phase.
3.3.4 Regional Considerations
Atmospheric sampling stations in the LMMB Study were grouped into the following categories based on
their proximity to urban areas: urban, urban-influenced, rural, and remote (see Table 2-6 in Chapter 2).
Mean total PCB concentrations at urban sites were six times higher than at urban-influenced sites, seven
times higher than at rural sites, and four times higher than at remote sites. The highest total PCB
concentrations were at the IIT Chicago site (mean of 2600 pg/m3), which is expected since various
combustion sources and other sources are likely to be present in a large urban area. The second highest
total PCB concentrations were at the remote Beaver Island location (mean of 970 pg/m3), which was not
anticipated and suggests an unknown source near this otherwise relatively remote site in northern Lake
Michigan. The Beaver Island location had the highest concentrations of PCB 180 (mean of 11 pg/m3), a
highly chlorinated and relatively high molecular weight congener. Mean PCB 180 concentrations at
Beaver Island were 3 times higher than at the IIT Chicago site and 33 to 46 times higher than at any other
remote station. The high concentrations of relatively high molecular weight PCB congeners at the Beaver
Island site suggest a nearby source, rather than long-range transport of PCBs from distant sources.
However, a review of land use maps, literature sources, and other information was conducted during the
preparation of this report. That review did not uncover any obvious sources. Therefore, rather than
speculate about possible sources, the anomalous PCB results at Beaver Island are simply reported here.
3-58 April 2004
-------
PCBs/trans-Nonachlor in Atmospheric Components
Results were similar for participate phase PCB concentrations, with the highest concentrations at the IIT
Chicago and Beaver Island sites. Total PCB concentrations in precipitation also were highest at the IIT
Chicago site, but precipitation-phase PCBs at Beaver Island were comparable to other remote sites.
The results from this study demonstrate that the urban source of PCBs from the Chicago area significantly
influence PCB concentrations over Lake Michigan. A statistically significant difference in the vapor-
phase concentrations of PCB 118 between over-water stations north and south of 43° latitude (Figure 3-6)
was observed in this study. PCB concentrations over the southern portion of Lake Michigan, and
particularly near Chicago, were significantly higher than concentrations over northern Lake Michigan.
PCB 118 concentrations at over-water stations 1 and 6, near Chicago, were 6 to 39 times higher than at
any other over-water station. PCB 118 concentrations at these two stations also were higher than at the
IIT Chicago site.
Other researchers have noted a similar influence on atmospheric PCB concentrations over Lake Michigan
due to the urban Chicago area. As a part of the AEOLOS (Atmospheric Exchange Over Lakes and
Oceans) Project, Simcik et al. (1997) determined that gas-phase PCB concentrations over southern Lake
Michigan were highly influenced by the urban/industrial area from Evanston, IL to Gary, IN. Emissions
from these urban areas increased the average coastal atmospheric concentration above the continental
background by a factor of four. Gas-phase PCB concentrations ranged from 0.14 ng/m3 to 1.1 ng/m3 over
the lake and from 0.27 to 14 ng/m3 in the urban area. Total PCB concentrations in rain measured by
Offenberg and Baker (1997) over southern Lake Michigan ranged from 4.1 to 189 ng/L and were from 2
to 400 times higher than the measured regional background concentrations. Offenberg and Baker (1997)
concluded that the "urban plume" of Chicago increases PCB wet deposition loadings over southern Lake
Michigan by 50 to 400%.
In contrast to the PCB results, the vapor-phase trans-nonachlor concentrations suggest that there is a
significant source in the rural, agricultural area near Bondville, with generally decreasing concentrations
in more northern stations (Figure 3-5). fra«s-Nonachlor concentrations in the vapor phase were highest at
the Bondville station followed by the IIT Chicago station. Although trans-nonachlor is no longer
produced in the U.S., nor applied in agricultural practice, this apparent trend from south to north may
indicate the effects of historical agricultural applications, with lesser contributions from other historical
uses in urban areas such as Chicago.
April 2004 3-59
-------
Chapter 4
PCBs/trans-Nonachlor in Tributaries
4.1 Results
A total of 354 samples were collected from 11 tributaries that flow into Lake Michigan or Green Bay and
analyzed for PCBs and fra«s-nonachlor. The samples were collected as described in Section 2.5.2, by
pumping 80 to 160 L of river water through a cartridge packed with 250 g of XAD-2®, a macroreticular
resin that traps hydrophobic organic contaminants. A "pentaplate" filter was installed in the sampling
train front of the XAD-2® cartridge to collect the particulate matter suspended in the sample. Separate
analyses were performed on the XAD-2® resin and the filtered particulates from each sampling effort,
yielding results for operationally defined "dissolved" and "particulate" PCBs (Table 4-1) and trans-
nonachlor (Table 4-2). Interferences and laboratory accidents reduced the number of fra«s-nonachlor
results to 338 dissolved results and 350 particulate results.
As noted in Chapter 2, there are 209 possible PCB congeners, and the investigators in this study reported
results for 65 to 110 of these congeners, depending on the capabilities of each laboratory. From March
1994 through October 1994, the analyses performed at the University of Wisconsin, Wisconsin State Lab
of Hygiene determined results for 65 congeners or co-eluting congeners. In November 1994, the
laboratory instituted a change in their standard operating procedure that allowed them to report the results
for 78 congeners or coeluting congeners. For the purposes of this report, we are presenting summaries of
the results for the following subset of all of the analytes:
• PCB congener 33
• PCB congener 118
• PCB congener 180
• Total PCBs
• fra«5-nonachlor
Table 4-1. Number of Tributary Samples Analyzed for Dissolved and Particulate PCB Congeners and Total
PCBs
Tributary
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Sampling Dates
04/07/94 to 10/1 2/95
08/04/94 to 10/1 8/95
04/1 1/94 to 10/31/95
04/1 2/94 to 10/30/95
04/1 1/94 to 10/26/95
04/1 3/94 to 10/1 1/95
03/29/94 to 10/06/95
04/1 4/94 to 12/05/95
04/05/94 to 10/1 8/95
04/06/94 to 10/24/95
04/06/94 to 10/27/95
Total
Dissolved PCBs
39
15
47
38
28
24
38
28
28
36
33
354
Particulate PCBs
39
15
47
38
28
24
38
28
28
36
33
354
Total Samples
78
30
94
76
56
48
76
56
56
72
66
708
April 2004
4-1
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 4-2. Number of Tributary Samples Analyzed for Dissolved and Participate frans-Nonachlor
Tributary
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Sampling Dates
04/07/94 to 10/1 2/95
08/04/94 to 10/1 8/95
04/1 1/94 to 10/31/95
04/1 2/94 to 10/30/95
04/1 1/94 to 10/26/95
04/1 3/94 to 10/1 1/95
03/29/94 to 10/06/95
04/1 4/94 to 12/05/95
04/05/94 to 10/1 8/95
04/06/94 to 10/24/95
04/06/94 to 10/27/95
Total
Dissolved Trans-
Nonachlor
38
15
34
38
28
24
36
28
28
36
33
338
Part icu late Trans-
Nonachlor
38
15
47
37
27
24
38
28
27
36
33
350
Total Samples
76
30
81
75
55
48
74
56
55
72
66
688
The 11 tributaries were chosen for sampling by the Lake Michigan Tributary Coordinating Committee,
comprised of representatives from EPA, the Wisconsin Department of Natural Resources, the Michigan
Department of Natural Resources, and the U.S. Geological Survey offices in Wisconsin and Michigan.
The 11 sites represent the variety of types of river that drain into the Lake Michigan basin. Ten of the
eleven rivers were chosen because elevated concentrations of contaminants previously observed in fish
collected from these tributaries suggest that these rivers are contributing the highest contaminant loadings
to the lake. The exception was the Pere Marquette River in Michigan. This tributary was chosen as the
"background" site, with little anthropogenic input. The samples from the Pere Marquette River will be
used to estimate loads from the small portion of the Lake Michigan watershed that was not monitored in
this study. The 11 monitored tributaries represent greater than 90% of the total river flow into Lake
Michigan and an even higher percentage of the total tributary load of pollutants into Lake Michigan (see
Section 2.4.2).
The committee classified the tributaries into three categories, based on their "event responsiveness,"
meaning the degree to which their physical and hydrological characteristics respond to the flow changes
associated with precipitation events. The categories were: variable, stable, and super stable. The
classifications were used to establish the sampling frequency for each tributary (Table 4-3). All
tributaries were to be sampled monthly during the winter and during base (low) flow conditions, with
additional samples collected after precipitation events that increased the tributary flow by at least 20%.
The planned sampling frequencies were met for all but four of the tributaries (Grand Calumet,
Menominee, Milwaukee, and Sheboygan Rivers).
4-2
April 2004
-------
PCBs/trans-Nonachlor in Tributaries
Table 4-3. Tributary Classifications Relative to Responsiveness to Precipitation Events
Tributary
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Event Responsiveness
Stable
Super Stable
Stable
Stable
Stable
Stable
Variable
Stable
Super Stable
Variable
Stable
Number of Planned Sampling Events
High Flow
18
Low Flow
8
16
24
18
12
8
16
18
30
8
15
16
11
30
18
5
15
8
4.1.1 Temporal Variation
Many of the tributary samples were collected in response to precipitation events and these events may
not have occurred simultaneously across the entire Lake Michigan basin. As a result, the collection dates
of the samples sometimes vary greatly across the tributaries. Therefore, the tributary results were
examined by season, where the seasons were defined as:
Spring (SP) = March 20 to June 20,
Summer (SU) = June 21 to September 22,
Autumn (AU) = September 23 to December 21, and
Winter (WI) = December 22 to March 19
The concentrations of dissolved and particulate total PCBs exhibited a seasonal trend for many of the
tributaries, with higher mean concentrations occurring in summer months and lower mean concentrations
occurring in winter months. There were significant differences between seasons for the dissolved total
PCB concentrations in nine of the eleven tributaries, and significant differences between season for the
particulate total PCB concentration in six of the eleven tributaries. However, the trend was not consistent
across all of the tributaries. Based on F-tests of log-transformed concentration data, there were significant
interactions between tributary and season. The temporal variations in the dissolved and particulate
concentrations of individual PCB congeners did not exhibit trends that were consistent across all
tributaries, based on F-tests of log-transformed concentration data.
The mean seasonal concentrations of dissolved and particulate total PCBs across all 11 tributaries span at
least two orders of magnitude. The tributaries can be visually divided into two groups, based on PCB
concentration. Specifically, six of the eleven tributaries exhibit dissolved and particulate total PCB mean
concentrations that are less than 4 ng/L, and often less than 1 ng/L, across all four seasons. The results
are plotted separately for dissolved total PCBs (Figure 4-1) and particulate total PCBs (Figure 4-2).
April 2004
4-3
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 4-1. Temporal Variation in Total Dissolved PCB Concentrations Measured in Lake Michigan
Tributaries
Dissolved PCB Totals - All stations
FOX RIVER
GRAND CALUMET RIVER
GRAND RIVER
KALAMAZOO RIVER
MANISTIQUE RIVER
MENOMINEE RIVER
MILWAUKEE RIVER
MUSKEGON RIVER
PERE MARQUETTE
SHEBOYGAN RIVER
ST. JOSEPH RIVER
SU
AU
Wl
Dissolved PCB Totals - Lower-level Stations
—A—GRAND RIVER
—SK—MANISTIQUE RIVER
—•—MENOMINEE RIVER
—e— MUSKEGON RIVER
—«— PERE MARQUETTE
—B— ST. JOSEPH RIVER
Note: The letters (A - C) represent the results of the analysis of variance and multiple comparisons test. Points with the same letter were not
statistically different (at alpha = 0.05). Tributaries without letters are those where there were no significant differences between
seasons.
4-4
April 2004
-------
PCBs/trans-Nonachlor in Tributaries
Figure 4-2. Temporal Variation in Total Participate PCB Concentrations Measured in Lake Michigan
Tributaries
Participate PCB Totals - All Stations
FOX RIVER
GRAND CALUMET RIVER
GRAND RIVER
KALAMAZOO RIVER
MANISTIQUE RIVER
MENOMINEE RIVER
MILWAUKEE RIVER
MUSKEGON RIVER
PERE MARQUETTE
- -SHEBOYGAN RIVER
ST. JOSEPH RIVER
SU
AU
Wl
Participate PCB Totals - Lower-level Stations
GRAND RIVER
-MANISTIQUE RIVER
-MENOMINEE RIVER
MUSKEGON RIVER
-PERE MARQUETTE
-ST. JOSEPH RIVER
Note: The letters (A - B) represent the results of the analysis of variance and multiple comparisons test. Points with the same letter were not
statistically different (at alpha = 0.05). Tributaries without letters are those where there were no significant differences between
seasons.
The mean dissolved total PCB concentrations appear to peak in summer in three of the tributaries (Fox,
Menominee, and Sheboygan), while they appear to peak in the autumn for three other tributaries (Grand
River, Muskegon, and St. Joseph). However, for many of the tributaries, the results do not show
significant differences between seasons and those apparent peaks are not significantly different from the
mean concentrations of the adjacent seasons. For example, the seasonal mean dissolved total PCB
concentrations in the Grand Calumet River and the Grand River are not statistically significant different
April 2004
4-5
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
between any of the seasons. For Fox River and Kalamazoo River, the mean concentrations in spring and
summer are not significantly different, but the mean concentration in summer is significantly higher than
in autumn, which is significantly higher than in winter.
The summer mean concentration of dissolved total PCBs in the Menominee River is significantly
different from the autumn mean concentration, but neither the summer nor the autumn mean is
significantly different from the spring and winter means. The large increase in concentration that is
visible in Figure 4-1 is driven by one of the four summer results for this tributary, with a dissolved total
PCB result of 10.6 ng/L. There is no unambiguous evidence that indicates this high result for total
dissolved PCBs is due to contamination in the field or the laboratory, thus the result was not excluded
from the database. However, examination of the data qualifiers applied to the individual PCB congener
results by both the PI who produced the results and the data reviewers suggest some increased uncertainty
with this specific sample (e.g., PCB 33 was associated with a field blank that did not meet the acceptance
criteria and PCB 180 was reported with the suspected contamination flag), but these concerns did not
affect a large number of other congeners. Were this result excluded from the calculation of the mean
seasonal results, the mean summer result for dissolved total PCBs at the Menominee River would be on
the order of 1.1 ng/L, a value well in line with the other low-level stations.
The mean particulate total PCB concentrations appear to peak in either spring or summer in 9 of the 11
tributaries, with the lowest mean concentrations in the winter. However, the significance of the seasonal
differences varies by tributary. For example, in the Sheboygan River, the mean spring and summer
particulate total PCB concentrations are not significantly different from one another, but both are
significantly different from the winter mean concentration. In the Fox River, mean spring, summer, and
autumn particulate total PCB concentrations are not significantly different from one another, but all are
significantly different from the mean winter concentration. For the Kalamazoo, Milwaukee, Grand, and
St. Joseph Rivers, the spring mean particulate total PCB concentrations are never the lowest
concentrations of the four seasons and the winter concentrations are never the highest of the four seasons.
Despite the apparent increase in the mean total particulate PCB concentrations from summer to winter in
Figure 4-2 for the Grand Calumet River, there is no statistically significant difference across all four
seasons in this tributary. Among the six low-level tributaries, there are no significant differences among
the seasons for the mean particulate total PCB concentrations in the Manistique, Menominee, Muskegon,
and Pere Marquette Rivers.
The mean concentrations of dissolved and particulate fra«s-nonachlor show fewer significant differences
than the total PCB results (Figure 4-3). Eight of the eleven tributaries (Fox, Grand Calumet, Grand,
Menominee, Milwaukee, Muskegon, Pere Marquette, and St. Joseph) exhibit no statistically significant
differences in mean dissolved fra«s-nonachlor concentrations among the seasons. Of the other three
tributaries, the mean dissolved fra«s-nonachlor in the Kalamazoo River is never the lowest in spring or
summer, and never the highest in autumn, while in the Sheboygan River, mean dissolved fra«s-nonachlor
is never the lowest in the summer, or the highest in the winter. The dissolved trans-nonachlor results for
the Manistique River are characterized by a very high mean concentration in the winter which is
significantly different from the other three seasons, which in turn, are not significantly different from one
another. The very high winter mean concentration is repeated in the particulate fra«s-nonachlor results in
this tributary. Figure 4-4 illustrates the seasonal trends for dissolved and particulate fra«s-nonachlor in
the tributaries after removing the very high winter mean result for the Manistique River.
4-6 April 2004
-------
PCBs/trans-Nonachlor in Tributaries
Figure 4-3. Temporal Variation in Total Dissolved (top) and Participate (bottom) frans-Nonachlor
Concentrations Measured in Lake Michigan Tributaries
Dissolved transENonachlor
— 0.08
"a
_c_
c
.2
•s
•e 0.06
i—FOX RIVER
I—GRAND CALUMET RIVER
GRAND RIVER
i—KALAMAZOO RIVER
!—MANISTIQUE RIVER
1—MENOMINEE RIVER
— MILWAUKEE RIVER
MUSKEGON RIVER
i—PERE MARQUETTE
••--SHEBOYGAN RIVER
J—ST. JOSEPH RIVER
Participate rra/7s -Nonachlor
FOX RIVER
GRAND CALUMET RIVER
GRAND RIVER
KALAMAZOO RIVER
MANISTIQUE RIVER
MENOMINEE RIVER
MILWAUKEE RIVER
MUSKEGON RIVER
— PERE MARQUETTE
--SHEBOYGAN RIVER
—ST. JOSEPH RIVER
Note: The letters (A - B) represent the results of the analysis of variance and multiple comparisons test. Points with the same letter were not
statistically different (at alpha = 0.05). Tributaries without letters are those where there were no significant differences between
seasons.
April 2004
4-7
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 4-4. Temporal Variation in Total Dissolved (top) and Particulate (bottom) frans-Nonachlor
Concentrations Measured in Lake Michigan Tributaries without the Winter Mean for the Manistique River
Dissolved transENonachlor
1
c
.2
5
•Ł 0.02
-•—FOX RIVER
-•—GRAND CALUMET RIVER
-*—GRAND RIVER
-X—KALAMAZOO RIVER
-*—MANISTIQUE RIVER
-•—MENOMINEE RIVER
—I—MILWAUKEE RIVER
-e—MUSKEGON RIVER
•0--PEREMARQUETTE
-•••--SHEBOYGAN RIVER
-B—ST. JOSEPH RIVER
Particulate rra/7s -Nonachlor
c 0.06
>—FOX RIVER
I—GRAND CALUMET RIVER
i—GRAND RIVER
!—KALAMAZOO RIVER
i—MANISTIQUE RIVER
I—MENOMINEE RIVER
— MILWAUKEE RIVER
MUSKEGON RIVER
>—PERE MARQUETTE
••••SHEBOYGAN RIVER
i—ST. JOSEPH RIVER
Note: The letters (A - B) represent the results of the analysis of variance and multiple comparisons test. Points with the same letter were not
statistically different (at alpha = 0.05). Tributaries without letters are those where there were no significant differences between
seasons.
4-8
April 2004
-------
PCBs/trans-Nonachlor in Tributaries
Six of the tributaries (Grand Calumet, Menominee, Milwaukee, Muskegon, Pere Marquette, and
Sheboygan) exhibited no statistically significant differences in the particulate trans-nonachlor
concentrations across seasons. The Manistique River had one non-zero result in summer and one non-
zero result in winter, and all other results were reported as zero. As seen in the dissolved fra«s-nonachlor
result, the winter mean particulate concentration in the Manistique River was much higher than for any
other tributary. The bottom portion of Figure 4-4 shows the results for the tributaries without this high
mean particulate result.
The trends for the other four tributaries are such that the mean particulate fra«s-nonachlor concentrations
in the spring were never the lowest, and the winter mean concentrations were never the highest.
4.1.2 Geographical Variation
The concentrations of dissolved and particulate Figure 4-5. Mean Dissolved and Particulate Concentrations
PCBs and frara-nonachlor varied by tributary of PCB 33 in Lake Michigan Tributaries
over the course of the study (Tables 4-4 through
4-8). For example, the concentration of
dissolved PCB 33 ranged from 0 to 1.1 ng/L
and the concentration of particulate PCB 33
ranged from 0 to 4.2 ng/L. The mean dissolved
concentrations of PCB 33 ranged from 0.0067
ng/L in the Pere Marquette River to 0.76 ng/L
in the Grand Calumet River, while the mean
particulate concentration of PCB 33 ranged
from 0.00042 ng/L in the Pere Marquette River
to 1.5 ng/L in the Fox River (Figure 4-5). (The
particulate PCB 33 results were reported as zero
for all 27 samples from the St. Joseph River).
Lake
Michigan
16-
1.4 •
1.2
1 1-
a
i o.e •
Ł 0.6-
0.4 •
0.2
i
f
Hfceiiomi
JTv
L-" '
// Fox
/
i
•xi
Slieboyg.in
\
0 - Milwaukee
]
_
Other PCB congeners exhibited ranges and
mean concentrations similar to those observed
for PCB 33. The dissolved total PCB
concentrations ranged from 0 ng/L in four
tributaries to 48 ng/L in the Grand Calumet,
while particulate total PCB concentrations
ranged from 0 ng/L in four tributaries to 120
ng/L in the Sheboygan River. Mean dissolved
total PCB concentrations ranged from 0.43 ng/L
in the Pere Marquette River to 35 ng/L in the
Grand Calumet, while mean particulate
concentration ranged from 0.25 ng/L in the
Muskegon River to 55 ng/L in the Sheboygan River.
For PCB 33, the mean dissolved concentrations are higher than the mean particulate concentrations in
eight of the eleven tributaries (Figure 4-5), while the particulate concentrations are higher in the Fox,
Kalamazoo, and Sheboygan Rivers. The distribution between the dissolved and particulate fractions
appears to change for the higher molecular weight congeners (e.g., see Figure 4-6 for the mean
concentrations of PCB 180). For PCB 118, the mean concentrations of the particulate samples were
markedly higher than the dissolved concentrations in 9 of the 11 tributaries and essentially equal in the
April 2004
4-9
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
ii .••
Menominee and Muskegon Rivers (Table 4-5). For PCB 180, the mean concentrations of the particulate
samples were markedly higher than the dissolved concentrations in all 11 tributaries.
Concentrations of dissolved trans-nonachlor Figure 4-6. Mean Dissolved and Particulate Concentrations
ranged from 0 in seven tributaries to 0.19 ng/L of PCB 180 in Lake Michigan Tributaries
in the Manistique River, while particulate
trans-nonachlor ranged from 0 in five
tributaries to 0.38 ng/L in the Manistique River
(Table 4-9). (Note: The maximum dissolved
and particulate trans-nonachlor concentrations
occurred in the same sample from the
Manistique River, which is otherwise relatively
uncontaminated. These two values may be the
result of contamination of the sample during
collection).
Mean dissolved trans-nonachlor concentrations
ranged 0.0033 ng/L in the Menominee River to
0.026 ng/L in the St. Joseph River, while mean
particulate trans-nonachlor concentrations
ranged from 0.0028 ng/L in the Menominee
River to 0.074 ng/L in the St. Joseph River.
There are statistically significant differences
among the mean concentrations of dissolved
and particulate PCBs in the 11 tributaries. The
differences in the mean dissolved
concentrations of total PCBs are shown in
Figure 4-7. The mean dissolved total PCB
concentrations in the Grand Calumet and
Sheboygan Rivers were significantly higher than in all other tributaries. The mean dissolved total PCB
concentrations in the Pere Marquette River were significantly lower than in all other tributaries except the
Muskegon River. There is a statistically significant interaction between tributary and year for the particulate
total PCB results (p=0.0118, two-way ANOVA, total PCB concentrations were log-transformed prior to
conducting the test). Therefore, the mean particulate total PCB concentrations for all of the tributaries are
presented separately for 1994 and 1995 (Figure 4-8, top and bottom, respectively).
There was a statistically significant difference between the particulate total PCB concentrations from
1994 and 1995 at three of the 11 tributaries (Fox, Grand, and Muskegon Rivers), based on two-sample t-
tests of log-transformed PCB data. The mean particulate total PCB concentrations were significantly
higher in 1994 than in 1995 in the Fox and Grand Rivers, while in the Muskegon River, the 1995 mean
concentration was higher than in 1994. The differences between the results in 1994 and 1995 in these
three tributaries may be a function of the unequal distribution of samples across calendar years (e.g., there
are no 1994 data before April and no 1995 data after October), or the differences may be the result of
some other factors.
The distinction of the "lower-level" tributaries shown in Figure 4-2 also holds true when the same
dissolved and particulate total PCB results were used in Figures 4-7 and 4-8. The six tributaries with
relatively low concentrations of dissolved and particulate PCBs (Grand, Manistique, Menominee,
Muskegon, Pere Marquette, and St. Joseph) in Figure 4-2 can also be distinguished from the five
4-10
April 2004
-------
PCBs/trans-Nonachlor in Tributaries
tributaries with much higher mean concentrations (Fox, Grand Calumet, Kalamazoo, Milwaukee, and
Sheboygan) in Figures 4-7 and 4-8.
Figure 4-7. Mean Dissolved Total PCB Concentrations in Lake Michigan
Tributaries
(C
"o
I—
CD
O
CL
d
(Li
O
d
O
O
100=,
10:
B
T
-
1-
1
• • ETF1 ? 1
T 1 1 G
| | 1 ~r _^
*
p
o
I
m
o
d
m
Concentration is plotted on a log scale. Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile
(top of box) results. Bars represent the results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest
edge of the box. Circles represent results beyond 1.5*IQR from the box. The letters (A - G) above the boxes represent the results of the
analysis of variance and multiple comparisons test. Boxes with the same letter were not statistically different (at alpha = 0.05).
April 2004
4-11
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 4-8. Mean Particulate Total PCB Concentrations in Lake Michigan Tributaries in 1994 (top)
and 1995 (bottom)
1000=,
"L 100=
to
"o
m
o
o_
o
"ro
"d
o
d
o
O
10=
1 =
1994
EF
J_
EF
p
o
d
2]
m
m
m
m
o
o
D
d
3
m
m
a?
o
-=:
en
o
m
-Q
I
1000=,
^ 100=
O5
"5
m
o
CL
d
dl
o
d
o
O
10 =
1 =
1995
T
r
p
o
m
n
o
m M
[I] hj
8 I
— i — 10
m
o
o
ji m
Concentration is plotted on a log scale. Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile
(top of box) results. Bars represent the results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest
edge of the box. Circles represent results beyond 1.5*IQR from the box. The As represent results beyond 3*IQR from the box. The letters (A -
F) above the boxes represent the results of the analysis of variance and multiple comparisons test. Boxes with the same letter were not
statistically different (at alpha = 0.05).
4-12
April 2004
-------
PCBs/trans-Nonachlor in Tributaries
The results for trans-nonachlor exhibited less distinct geographical variations, compared to the PCB
results (Figure 4-9). For example, the concentrations of dissolved trans -nonachlor did not exhibit the
distinction of the "lower-level" tributaries shown in Figure 4-2 for the PCB results. Only two of those six
"lower-level" tributaries had statistically lower dissolved trans-nonachlor concentrations: the Manistique
and Menoninee Rivers. In marked contrast to the dissolved PCB results, the dissolved trans -nonachlor
results for the Fox River were the third lowest of the 11 tributaries.
Figure 4-9. Mean Dissolved frans-Nonachlor Concentrations in Lake
Michigan Tributaries
o
o
ro
c
o
2
w
c
TO
-I?
Concentration is plotted on a log scale. Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile
(top of box) results. Bars represent the results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest
edge of the box. Circles represent results beyond 1.5*IQR from the box. The As represent results beyond 3*IQR from the box. The letters (A -
E) above the boxes represent the results of the analysis of variance and multiple comparisons test. Boxes with the same letter were not
statistically different (at alpha = 0.05).
As with the particulate PCB results, there is a statistically significant interaction between tributary and
year for the particulate trans-nonachlor results (p=0.0001, two-way ANOVA, fra«s-nonachlor
concentrations were log-transformed prior to conducting the test). Therefore, the mean particulate trans-
nonachlor concentrations for all of the tributaries are presented separately for 1994 and 1995 (Figure 4-
10, top and bottom, respectively).
However, unlike the particulate PCB results, the distinctions between rivers with relatively low or
relatively high concentrations of particulate trans -nonachlor are not as clear, nor as consistent between
the two years. For example, the particulate trans-nonachlor results for the Fox River are not statistically
different from those in the Pere Marquette River in either 1994 or 1995. The concerns about the
anomalous particulate trans -nonachlor in the Manistique River are evident in the top and bottom portions
of Figure 4-10, where the 1994 mean result is among the lowest of all 11 tributaries, while the 1995 mean
result is the highest of all 11 tributaries.
April 2004
4-13
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 4-10. Mean Particulate frans-Nonachlor Concentrations in Lake Michigan Tributaries in 1994
(top) and 1995 (bottom)
10
3
u
1=
o
o
o
(C
.001
DE
*
M
-n
I
A
ABC T
S ^
*
3)
0 ^
1 1
1 1
m*4 IV
1
BC
^
•
FG*
Ł 1
S ^5
r3 5
§ m
Ti ?i
W -i
FG
S
s
m
NOMINEE:
1
fc.1^
CD
*
T G
Q
= c
1 i
5 o
S o
m z
7i ii
CT «
EF
E:
^
D
C
m
Ij
m
j*^
P
ABC
H
I
v>
m
(D
I
?
_L
AB
T
H
JOSEPH (i
Tl
_i
>J
CO
c
(J
c
o
o
u
05
10a
1 =
.1 =
.01 =
.001
D
*
•n
O
X
3
II
-2
BC AB
T T BC
0 > C
n S
S 1? §
m A S
A*
-MANISTIQU
m
D
?
m
z
O
S
z
m
m
BC
*
S D
5
-MUSKEGON
-MILWAUKE
D
T
-P. MARQUE
-I
-i
c
X
m
(D
o
z
AB
*
-ST. JOSEPH
i*-V
II
to
II
K)
K)
Concentration is plotted on a log scale. Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile
(top of box) results. Bars represent the results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest
edge of the box. Circles represent results beyond 1.5*IQR from the box. The *s represent instances where all but one observation was zero (0)
and the log of 0 is indeterminate. The letters (A - G) above the boxes represent the results of the analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05).
4-14
April 2004
-------
PCBs/trans-Nonachlor in Tributaries
Table 4-4. Concentrations of PCB Congener 33 Measured in Tributaries
Fraction
Dissolved
Participate
Tributary
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
N
39
15
35
36
24
15
37
24
27
28
28
39
15
25
34
27
24
34
28
28
16
27
Mean (ng/L)
0.43
0.76
0.018
0.095
0.0099
0.082
0.27
0.010
0.0067
0.20
0.020
1.5
0.46
0.0033
0.16
0.0016
0.00076
0.12
0.00088
0.00042
0.25
0.0
Range (ng/L)
0.066 to 0.93
0.39 to 1.1
0.0 to 0.043
0.035 to 0.1 7
0.0 to 0.028
0.015 to 0.56
0.094 to 0.94
0.0 to 0.067
0.0 to 0.025
0.12 to 0.31
0.0 to 0.052
0.067 to 4.2
0.18 to 1.1
0.0 to 0.026
0.023 to 0.38
0.0 to 0.0090
0.0 to 0.018
0.032 to 0.36
0.0 to 0.012
0.0 to 0.0083
0.048 to 0.46
0.0 to 0.0
SD (ng/L)
0.23
0.27
0.0099
0.032
0.0098
0.14
0.14
0.018
0.0087
0.058
0.015
0.99
0.26
0.0070
0.092
0.0029
0.0037
0.068
0.0027
0.0017
0.12
0.0
RSD (%)
55
36
55
34
99
170
52
170
129
29
74
64
57
210
56
180
490
56
310
400
47
-
% Below DL
0.0
0.0
23
0.0
50
20
0.0
71
93
0.0
50
0.0
0.0
84
0.0
100
100
0.0
100
100
0.0
100
April 2004
4-15
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 4-5. Concentrations of PCB Congener 118 Measured in Tributaries
Fraction
Dissolved
Participate
Tributary
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
N
39
15
45
38
27
22
37
28
27
36
32
39
15
47
36
27
23
38
28
28
36
32
Mean (ng/L)
0.044
0.14
0.013
0.058
0.0039
0.014
0.072
0.0068
0.0062
0.19
0.021
0.43
0.83
0.065
0.47
0.010
0.015
0.20
0.0068
0.017
1.9
0.083
Range (ng/L)
0.014 to 0.14
0.083 to 0.25
0.0 to 0.028
0.026 to 0.1 4
0 to 0.011
0.0058 to 0.040
0.037 to 0.1 9
0.0 to 0.019
0.0 to 0.028
0.084 to 0.31
0.0095 to 0.039
0.014 to 1.2
0.20 to 1.7
0.0045 to 0.1 4
0.088 to 0.98
0.0 to 0.032
0.0 to 0.030
0.030 to 0.60
0.0 to 0.014
0.0 to 0.046
0.22 to 4.2
0.0 to 0.2
SD (ng/L)
0.022
0.044
0.0052
0.023
0.0034
0.0074
0.026
0.0064
0.0065
0.058
0.0082
0.26
0.46
0.026
0.24
0.0080
0.0076
0.13
0.0035
0.013
1.1
0.045
RSD (%)
51
31
41
39
88
51
36
94
110
30
38
60
55
40
51
78
51
62
51
74
57
54
% Below DL
7.7
0.0
31
0.0
93
91
0.0
100
96
0.0
50
2.6
0.0
2.1
0.0
63
70
0.0
100
64
0.0
9.4
4-16
April 2004
-------
PCBs/trans-Nonachlor in Tributaries
Table 4-6. Concentrations of PCB Congener 180 Measured in Tributaries
Fraction
Dissolved
Participate
Tributary
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
N
38
14
42
37
28
21
38
28
28
36
32
39
15
44
36
25
24
38
27
26
36
31
Mean (ng/L)
0.013
0.022
0.0029
0.014
0.0015
0.0052
0.023
0.0076
0.0054
0.016
0.0065
0.20
0.66
0.049
0.19
0.0031
0.026
0.28
0.012
0.018
0.30
0.074
Range (ng/L)
0.0 to 0.044
0.0 to 0.081
0.0 to 0.010
0.0 to 0.056
0.0 to 0.014
0.0 to 0.024
0.011 to 0.11
0.0 to 0.023
0.0 to 0.020
0.0 to 0.055
0.0 to 0.023
0.011 to 0.50
0.32 to 1.5
0.0093 to 0.1 3
0.046 to 0.36
0.0 to 0.010
0.0 to 0.1 2
0.086 to 1.1
0.0 to 0.027
0.0 to 0.063
0.045 to 0.91
0.016 to 0.15
SD (ng/L)
0.0088
0.019
0.0026
0.010
0.0029
0.0060
0.016
0.0070
0.0059
0.011
0.0068
0.12
0.33
0.023
0.086
0.0031
0.025
0.17
0.0070
0.015
0.18
0.032
RSD (%)
69
86
90
66
190
110
70
93
110
67
110
58
50
46
45
100
95
61
61
84
59
43
% Below DL
68
36
95
57
96
100
21
89
96
50
91
5.1
0.0
0.0
0.0
92
42
0.0
85
58
0.0
6.5
April 2004
4-17
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 4-7. Concentrations of Total PCBs Measured in Tributaries
Fraction
Dissolved
Particulate
Tributary
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
N
39
15
47
38
28
24
38
28
28
36
33
39
15
47
38
28
24
38
28
28
36
33
Mean (ng/L)
14
35
0.76
6.9
0.76
1.4
13
0.58
0.43
26
1.0
39
41
1.6
16
0.41
0.52
11
0.25
0.47
55
1.9
Range (ng/L)
2.5 to 32
24 to 48
0.0 to 2.0
2.9 to 12
0.13 to 1.8
0.0 to 11
6.7 to 28
0.23 to 2.2
0.0 to 0.83
13 to 45
0.0 to 2.7
1.6 to 110
19 to 96
0.24 to 3.7
0.0 to 38
0.046 to 1.8
0.18 to 1.3
2.3 to 35
0.0 to 0.59
0.0 to 1.2
6.9 to 120
0.0 to 4.1
SD (ng/L)
7.6
6.5
0.35
2.1
0.39
2.1
4.0
0.40
0.19
8.3
0.53
25
22
0.63
9.6
0.37
0.27
6.2
0.14
0.32
31
0.98
RSD (%)
53
19
47
30
52
150
30
69
45
32
52
64
53
40
59
90
53
58
54
67
56
52
4-18
April 2004
-------
PCBs/trans-Nonachlor in Tributaries
Table 4-8. Concentrations of frans-Nonachlor Measured in Tributaries
Fraction
Dissolved
Participate
Tributary
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
N
38
15
34
38
28
24
36
28
28
36
33
38
15
47
37
27
24
38
28
27
36
33
Mean (ng/L)
0.0056
0.015
0.020
0.016
0.0083
0.0033
0.023
0.0094
0.0081
0.015
0.026
0.018
0.049
0.067
0.042
0.014
0.0028
0.037
0.0041
0.0098
0.040
0.074
Range (ng/L)
0.0 to 0.019
0.0090 to 0.033
0.0 to 0.046
0.0080 to 0.026
0.0 to 0.19
0.0 to 0.014
0.0 to 0.044
0.0 to 0.046
0.0 to 0.020
0.0060 to 0.035
0.010 to 0.045
0.0 to 0.1 7
0.021 to 0.12
0.014 to 0.18
0.017 to 0.076
0.0 to 0.38
0.0 to 0.012
0.011 to 0.22
0.0 to 0.021
0.0 to 0.027
0.0082 to 0.11
0.010 to 0.23
SD (ng/L)
0.0050
0.0067
0.0083
0.0040
0.036
0.0036
0.0093
0.0088
0.0054
0.0057
0.0071
0.026
0.028
0.030
0.016
0.073
0.0035
0.035
0.0047
0.0075
0.024
0.043
RSD (%)
89
44
40
26
430
110
41
94
67
39
28
145
56
44
39
520
130
96
110
76
58
57
% Below DL
97
67
2.9
74
96
100
25
96
96
81
15
74
0.0
0.0
2.7
96
100
18
96
85
19
3.0
4.2 Quality Implementation and Assessment
As described in Section 1.5.5, the LMMB QA program prescribed minimum standards to which all
organizations collecting data were required to adhere. The quality activities implemented for the PCBs
and fra«s-nonachlor monitoring portion of the study are further described in Section 2.7 and included use
of SOPs, training of laboratory and field personnel, and establishment of MQOs for study data. A
detailed description of the LMMB quality assurance program is provided in the Lake Michigan Mass
Balance Study Quality Assurance Report (USEPA, 200 Ib). A brief summary of data quality issues for
the tributary PCBs and fra«s-nonachlor data is provided below.
Quality Assurance Project Plans (QAPPs) were developed by the Pis and were reviewed and approved by
GLNPO. Each researcher trained field personnel in sample collection SOPs prior to the start of the field
season and analytical personnel in analytical SOPs prior to sample analysis. Each researcher submitted
test electronic data files containing field and analytical data according to the LMMB data reporting
standard prior to study data submittal. GLNPO reviewed these test data sets for compliance with the data
reporting standard and provided technical assistance to the researchers. In addition, each researcher's
laboratory was audited during an on-site visit at least once during the time LMMB samples were being
April 2004
4-19
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
analyzed. The auditors reported positive assessments and did not identify issues that adversely affected
the quality of the data.
As discussed in Section 2.5, because data comparability was important to the successful development of
the mass balance model, the Pis used similar sample collection, extraction, and analysis methods for the
PCB and trans-nonachlor monitoring in this study. However, as noted earlier in this section, after the
study began, changes were made to the procedures used for cleaning the XAD-2® resin and for the
analyses of the tributary samples. The first 35 field samples were analyzed on an older GC/ECD system
and were quantified against the Aroclor mixture prepared in 1985 by Dr. Mike Mullin at the EPA-Grosse
lie laboratory. These samples are identified in the data set with the text "Method 1293-11/11/94" in the
Exception to Method text field. After mid-November 1994, analyses were performed on a new GC/ECD
system that resulted in resolution of more PCB congeners (78 vs. 65) and lower method detection limits
(MDLs) than in the earlier analyses. In addition, samples analyzed after mid-November 1994 were
quantified against the 1994 version of the Mullin mix standard prepared exclusively for the LMMB
Study. These analytical changes were implemented after November 12, 1994, but affect all the tributary
samples collected from May 1994 to the end of the LMMB Study.
As discussed in Section 2.7, data verification was performed by comparing all field and QC sample
results produced by each PI with their MQOs and with overall LMMB Study objectives. Analytical
results were flagged when pertinent QC sample results did not meet acceptance criteria as defined by the
MQOs. These flags were not intended to suggest that data were not useable; rather they were intended to
caution the user about an aspect of the data that did not meet the predefined criteria. Table 4-9 provides a
summary of flags applied to the tributary PCB and trans-nonachlor data. The summary includes the flags
that directly relate to evaluation of the MQOs to illustrate some aspects of data quality, but does not
include all flags applied to the data to document sampling and analytical information, as discussed in
Section 2.7. Compared to other matrices, the percentage of results that were qualified for these criteria is
relatively small.
Pis used surrogate spikes to monitor the bias of the analytical procedure. The PCB results were corrected
for the recoveries of the surrogates. The trans-nonachlor results were not surrogate-corrected. Only
0.6% of the results of the tributary samples analyzed for dissolved PCB 33 (2 samples) were qualified for
surrogate recovery problems (Table 4-9).
Laboratory matrix spike samples also were used to monitor the bias of the analytical procedure. The
laboratory matrix spike samples were prepared from unexposed filters and XAD-2® cartridges that were
spiked with PCBs and trans-nonachlor. The results for the matrix spike samples were compared to the
MQO for spike recoveries (50 - 125%). Analytical results associated with matrix spike samples with
recoveries below the MQO limits were flagged with failed matrix spike and low bias and results
associated with matrix spike samples with recoveries higher than the MQO limits were flagged with failed
matrix spike and high bias. Analytical results were considered invalid and flagged as such when the
analyte was undetected and recoveries for associated matrix spike samples were less than 10%. No
tributary trans-nonachlor samples failed the matrix spike MQOs. Overall, only 1.4% of the samples were
associated with a matrix spike samples that failed the MQOs for a given PCB congener. None of the
results for PCBs 33, 118, or 180 were flagged as failing the matrix spike MQOs.
Field blanks were collected for PCBs and fra«s-nonachlor. When field blank contamination was greater
than 3.3 times the method detection limit, all of the associated results were flagged with the failed field
blank sample code (FRB). Field blanks were not collected at all stations, so potential station-specific
contamination associated with these sites cannot be evaluated. However, contamination associated with
sample collection and sampling equipment and sample processing, shipping, storing, and analyzing can be
evaluated based on the field blanks collected throughout the study. For dissolved PCB 33,3% of the field
4-20 April 2004
-------
PCBs/trans-Nonachlor in Tributaries
samples were associated with a field blank in which this congener was reported above the sample-specific
detection limit (Table 4-9). None of the field samples results for fra«s-nonachlor were qualified because
of field blank results.
Two types of laboratory blanks were prepared and analyzed for PCBs and fra«s-nonachlor. One type of
laboratory blank (LRB) consisted of an unexposed resin cartridge and filter that were extracted like a field
sample. Another type of laboratory blank (LDB) consisted of a volume of solvent processed through an
empty Soxhlet apparatus in the same fashion used to extract the field samples. After extraction, the
solvent was concentrated and analyzed like a field sample. The results for both types of laboratory blanks
were handled in the same fashion. When laboratory blank contamination was greater than the method
detection limit, all of the associated results were flagged. None of the field samples results for trans-
nonachlor were qualified because of laboratory blank results.
PCB congeners were reported detected in all of the laboratory blanks that were analyzed. The following
PCB congeners were detected in LDB (empty Soxhlet) blanks above the MDL: 15+17, 18, 87, 170+190,
180, and 206. The following PCB congeners were detected in LRB (unused resin cartridge and filter)
blanks above the MDL: 28+31,41+71+64,44,49,52, 87,95, 101, 170+190, 180, 194,208+195,201
and 206. The differences between the results for these two types of laboratory blanks provide an
indication of the congeners that are contributed by the resin and filter, as opposed to the laboratory
glassware. The resin and the filter appear to contribute congeners 28+31, 41+71+64, 44, 49, 52, 95, 101,
194, 208+195, and 201.
Trip blanks were prepared and analyzed for PCBs and fra«s-nonachlor. When trip blank contamination
was greater than 3.3 times the method detection limit, all of the associated results were flagged with failed
trip blank sample code (FFT). For dissolved PCB 33, 7% of the field samples were associated with a trip
blank in which this congener was reported above 3.3 times the method detection limit (Table 4-9). None
of the field samples were associated with atrip blank that contained fra«s-nonachlor above 3.3 times the
method detection limit.
Field duplicates were to be collected at a frequency of 5%. Duplicate samples collected within 5 minutes
of each other were considered field duplicates. However, an examination of the field collection records
indicated that some of the planned field duplicates were not collected within that 5-minute time frame as a
result of problems with equipment mobilization or the time required to pump the sample through the filter
and resin cartridge. Those "duplicates" that were collected more than 5 minutes apart were considered
"sequential field duplicates" and the data were labeled accordingly (e.g., SDF1 vs. FD1). Combining the
field duplicates and sequential field duplicates, the actual rate of collection of duplicates was 4.2%.
The results from the original field sample and the associated duplicate were compared on the basis of the
relative percent difference (RPD). The RPD value for each PCB congener and fra«s-nonachlor was
compared to the MQO for field duplicate precision. Only 0.3% of the field samples results for PCBs 33
and 180 were qualified because of the field duplicate precision (FFD) concerns (Table 4-9). None of the
fra«5-nonachlor results were qualified.
April 2004 4-21
-------
Table 4-9. Summary of Routine Field Sample Flags Applied to Select PCB Congeners and frans-Nonachlor in Tributary Samples
Analyte
PCB 33
PCB 118
PCB 180
frans-Nonachlor
Fraction
Dissolved
Particulate
Dissolved
Particulate
Dissolved
Particulate
Dissolved
Particulate
Flags
Sensitivity
MDL
9% (28)
5% (16)
30% (103)
21% (72)
47% (163)
18% (61)
52% (176)
26% (93)
UNO
17% (52)
47% (141)
9% (31)
4% (14)
25% (86)
5% (16)
14% (49)
15% (53)
Contamination
FFR
3% (10)
0
0
0
0
0
0
0
FFT
7% (23)
0
0
0
0
0
0
0
Precision
FFD
0.3% (1)
0
0
0
0.3% (1)
0
0
0
Bias
FSS
0.6% (2)
0
0
0
0
0
0
0
FMS
0
0
0
0
0
0
0
0
LOB
0
0
0
0
0
0
0
0
HIB
0
0
0
0
0
0
0
0
The number of routine field samples flagged is provided in parentheses. The summary provides only a subset of applied flags and does not represent the full suite of flags applied
to the data.
MDL = Less than method detection limit (Analyte produced an instrument response but reported value is below the calculated method detection limit. Validity of reported
value may be compromised.)
UNO = Analyte not detected (Analyte produced no instrument response above noise.)
FFR = Failed field blank (A field blank sample, type unknown, associated with this analysis failed the acceptance criteria. It is unknown whether the blank that failed was a
field blank or a lab blank. Validity of reported value may be compromised.)
FFT = A trip blank associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.
FFD = Failed field duplicate (A field duplicate associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FSS = Failed surrogate (Surrogate recoveries associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FMS = Failed matrix spike (A matrix spike associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
LOB = Likely biased low (Reported value is probably biased low as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery or other internal
lab QC data. Reported value is not considered invalid.)
HIB = Likely biased high (Reported value is probably biased high as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery, blank
contamination, or other internal lab QC data. Reported value is not considered invalid.)
4-22
-------
PCBs/trans-Nonachlor in Tributaries
As discussed in Section 1.5.5, MQOs were defined in terms of six attributes: sensitivity, precision,
accuracy, representativeness, completeness, and comparability. GLNPO derived data quality assessments
based on a subset of these attributes. For example, system precision was estimated as the mean relative
percent difference (RPD) between the results for field duplicate pairs. Similarly, analytical precision was
estimated as the mean relative percent difference (RPD) between the results for laboratory duplicate pairs.
Table 4-10 provides a summary of data quality assessments for several of these attributes for the tributary
PCB and trans-nonachlor data.
Because the relative variability of most measurement techniques increases as one approaches the
detection limit of the technique, the assessment of the field duplicate results were divided into two
concentration regimes. One measure of system precision was calculated for those field duplicate results
that were less than 5 times the method detection limit (MDL) of the analyte and a separate measure was
calculated for those field duplicate results that were greater than 5 times the MDL.
For PCBs 33 and 118, the dissolved measurements were much more precise for those samples above 5
times the MDL, compared to those samples below 5 times the MDL. The mean relative percent
difference (RPD) between the field duplicates decreased from 45% for dissolved PCB 33 field duplicates
below 5 times the MDL to 12% for the field duplicates above 5 times the MDL. For PCB 118, the mean
RPD dropped from 27% to 3.7% for the dissolved results. There were no field duplicate pairs with
dissolved concentrations of PCB 180 or trans-nonachlor that were above 5 times the MDL, so a similar
comparison is not possible for these two analytes. The precision of the particulate measurements varied
much less than that of the dissolved measurements. For particulate PCB 33 in field duplicates, the mean
RPD actually increased from 13% for samples below 5 times the MDL to 15% for those duplicates above
5 times the MDL. For PCB 118, the mean RPD decreased from 15% to 10%. For particulate PCB 180,
the mean RPDs were 14% and 13% , and they decreased from 13% to 8.6% for particulate trans-
nonachlor.
Analytical bias was assessed using the results from matrix spike samples. Because it is not practical to
prepare a sufficiently large volume (80 - 160 L) of water spiked with known amounts of both dissolved
and particulate analytes, the matrix spike samples were prepared in the laboratory by adding known
amounts of the PCBs and fra«s-nonachlor to a filter and XAD-2® resin cartridge that had never been
exposed in the field and then extracting the filter and the resin using the sample techniques employed for
the field samples. The mean recoveries of the analytes were excellent for the PCBs, ranging from 97% to
103% for the PCB congeners in Table 4-10, with no appreciable difference between the dissolved and
particulate fractions. The mean recoveries for fra«s-nonachlor were very good, at 86% and 87% for the
dissolved and particulate fractions respectively.
Thus, these results demonstrate that the analytical techniques applied to the field samples introduce little
or no bias into the PCB results, and a slight low bias into the trans-nonachlor results. However, it is not
possible to directly assess the capabilities of the sampling techniques to collect the dissolved and
particulate analytes from the field samples themselves, a problem that was discussed at length in the
quality assurance project plan for the LMMB Study (e.g., it is not practical to prepare large volumes (80 -
160 L) of water containing known concentrations the analytes of interest for routine use as reference
samples).
Analytical sensitivity was assessed on the basis of the percentage of study samples that were reported
with concentrations below the sample-specific detection limit (SSDL). The sensitivity varied by congener
for the PCBs, partly as a function of the analytical instrumentation and its response to the individual
congeners.
April 2004 4-23
-------
Table 4-10. Data Quality Assessment for Select PCB Congeners and frans-Nonachlor in Tributary Samples
Analyte/Number
Field Samples
PCB 33 -
309 Dissolved
297 Particulate
PCB 118-
346 Dissolved
349 Particulate
PCB 180-
342 Dissolved
341 Particulate
frans-Nonachlor -
338 Dissolved
350 Particulate
Parameter
System Precision - Mean Field Duplicate RPD (%), < 5 * SSDL
System Precision - Mean Field Duplicate RPD (%), > 5 * SSDL
Analytical Bias - Mean Laboratory Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < SSDL (%)
System Precision - Mean Field Duplicate RPD (%), < 5 * SSDL
System Precision - Mean Field Duplicate RPD (%), > 5 * SSDL
Analytical Bias - Mean Laboratory Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < SSDL (%)
System Precision - Mean Field Duplicate RPD (%), < 5 * SSDL
System Precision - Mean Field Duplicate RPD (%), > 5 * SSDL
Analytical Bias - Mean Laboratory Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < SSDL (%)
System Precision - Mean Field Duplicate RPD (%), < 5 * SSDL
System Precision - Mean Field Duplicate RPD (%), > 5 * SSDL
Analytical Bias - Mean Laboratory Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < SSDL (%)
Number of QC samples
Dissolved
9 field duplicate pairs
13 field duplicate pairs
64 Matrix Spikes
-
22 field duplicate pairs
5 field duplicate pairs
64 Matrix Spikes
-
24 field duplicate pairs
0 field duplicate pairs
64 Matrix Spikes
-
26 field duplicate pairs
0 field duplicate pairs
65 Matrix Spikes
-
Particulate
4 field duplicate pairs
10 field duplicate pairs
64 Matrix Spikes
-
10 field duplicate pairs
19 field duplicate pairs
64 Matrix Spikes
-
1 1 field duplicate pairs
18 field duplicate pairs
64 Matrix Spikes
-
22 field duplicate pairs
5 field duplicate pairs
65 Matrix Spikes
-
Assessment
Dissolved
45%
12%
98%
24%
27%
3.7%
101%
37%
27%
-
103%
69%
19%
-
86%
66%
Particulate
13%
15%
97%
53%
15%
10%
101%
24%
14%
13%
103%
22%
13%
8.6%
88%
41%
4-24
-------
PCBs/trans-Nonachlor in Tributaries
PCB congeners and fra«s-nonachlor were not detected in substantial portions of the dissolved and
particulate samples from the tributaries ("UND" flag in Table 4-9). These analytes were detected below
the sample-specific detection limits in substantial portions of the samples as well ("MDL" flag in Table 4-
9). For the three congeners listed in Table 4-9, the percentage of the samples with results reported below
the sample-specific detection limits increases with the congener number (e.g., with molecular weight),
suggesting that solubility may play a role in the distribution.
However, other factors affect this assessment of sensitivity, including both the extent of PCB
contamination in Lake Michigan and the expected partitioning of analytes between the dissolved and
particulate fractions. For example, only 24% of the dissolved PCB 33 results were below the SSDL,
while 69% of the dissolved PCB 180 results were below the SSDL. In contrast, 53% of the particulate
PCB 33 results were below the SSDL, while only 22% of the particulate PCB 180 results were below the
SSDL. These differences between PCB 33 and PCB 180 may reflect the physical properties of the two
congeners which indicate that PCB 33 is likely to be more soluble in water than PCB 180 and that PCB
180 is more likely to sorb to particulates. Conversely, the analytical sensitivities reported here may
reflect the fact that the mean concentrations of PCB 33 in the tributaries are generally higher than the
mean concentrations of either PCBs 118 or 180, thus fewer samples will contain PCB 33 below the
SSDL.
The sensitivity for tmns-nonachlor was similar to that for PCB 180, with 66% of the dissolved results
below the SSDL and 41% of the particulate results below the SSDL.
4.3 Data Interpretation
4.3.1 Comparison to Historical Studies
There appear to be relatively few historical data on PCBs and fra«s-nonachlor available for the tributaries
in the LMMB Study. Much of the published data focuses on the open lake, not the tributaries. Data for
the Fox River are available from the Wisconsin Department of Natural Resources (DNR) based on their
efforts to remediate PCB contamination in 39 miles of the lower Fox River emptying into Green Bay.
Those data are a combination of data collected in 1989 and 1990 and the data collected in the Fox River
as part of the LMMB Study in 1994 and 1995. The results are for total PCBs, without any fractionation
between dissolved and particulate phases. The individual results from each sample collected in 1989 and
1990 are not available in the DNR report, so no formal statistical comparisons could be made. The results
are presented in a graph in the DNR report, in which the total PCB concentrations appear to range from
near 0 to 120 ng/L in the 1989 - 1990 study, and from near 0 to 130 ng/L for the 1994 - 1996 data
(collected as part of the LMMB Study). Moreover, DNR concluded that "the Lower Fox River is the
source of 95% of the PCB load to Green Bay and is the single largest tributary load to Lake Michigan. "
The data from the LMMB Study presented in Section 4.2 show that the mean concentrations of PCB
congeners and total PCBs in the Fox River are among the highest of the 11 tributaries in this study.
When combined with the flow data from the LMMB Study, these concentration data can be transformed
into loads that can be compared to the loads from the other tributaries, and ultimately compared to the
conclusions of the DNR report.
PCB data collected from large volume samples similar to those in the LMMB Study were reported from
the Detroit River, which connects Lake Huron and Lake Erie (Froese et al. 1997). The samples were
collected on eight occasions between March and October 1995 and on one occasion in May 1996. The
investigators in that study reported that dissolved total PCB concentrations ranged from less than 5 ng/L
April 2004 4-25
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
to 13 ng/L, while participate total PCB concentrations ranged from less than 5 ng/L to 22 ng/L, with a
mean participate concentration of 10 ng/L. Those results fall within the same general ranges as the
dissolved and particulate results for the Fox, Grand Calumet, Kalamazoo, Milwaukee, and Sheboygan
Rivers in this study, and are higher than the results for many of the other Lake Michigan tributaries.
Another earlier study addressed PCB concentrations in 14 major tributaries that discharge into Lake
Michigan. Marti and Armstrong (1990) reported the results from between three and eight samples
collected from each of the 14 tributaries from 1980 to 1983. The 14 tributaries in that study included 10
of the 11 tributaries in the LMMB Study (only the Grand Calumet River was not included). The sample
volumes were approximately 16 liters, and were filtered and processed through a column of XAD-2®
resin, in a fashion similar to that used in the LMMB Study.
As with the WDNR data, the results were reported for "total PCBs," however, in addition to the total PCB
concentration (e.g., dissolved and particulate), Marti and Armstrong reported the percentage that represent
the particulate PCBs. Where possible, they also assigned the PCBs to one of three Aroclors or Aroclor
mixtures (1242 + 1248, 1254, and 1260).
For the 10 LMMB tributaries, Marti and Armstrong reported mean total PCB concentrations ranging from
9 to 103 ng/L, with the extreme values ranging from 4 to 262 ng/L in those 10 tributaries. Marti and
Armstrong attributed from 53% to 83% of the "total PCB" concentrations to particulate-phase PCBs. In
contrast, the sum of the mean dissolved and mean particulate total PCB concentrations in the LMMB
Study ranged from 0.8 to 76 ng/L, while the percentage of the total PCBs attributable to the particulates
ranges from 27% to 74%. Marti and Armstrong found the highest total PCB concentration in the Fox
River (262 ng/L). The three highest mean concentrations were reported for the Sheboygan, Fox, and
Milwaukee Rivers, at 103, 98, and 97 ng/L, respectively. The sums of the mean dissolved and particulate
total PCB results from the LMMB Study for these same three tributaries are 81, 53, and 24 ng/L,
suggesting that total PCB concentrations in these three tributaries decreased by 21% to 71% from 1980 to
1995.
Data from the study also suggest that PCB concentrations are influenced by river flows as well as
sediment PCB concentrations. Under low-flow conditions, total PCB concentrations in the Fox River
were relatively high, possibly the result of the release of PCBs from sediments into the river water. As
rivers flow increased, the total PCB concentrations decreased to the point of mean flow, and then
increased again at higher flows. The total PCB concentration increase with increased flow is believed to
be indicative of resuspension of PCB-contaminated sediments.
Marti and Armstrong also reported the results for 20 blanks processed through the filtration and extraction
procedures applied to the tributary samples. The mean PCB concentration in the XAD-2 resin blanks was
1.1 ng/L ± 1.4 ng/L. The mean total PCB concentration for the filter blanks was 0.81 ng/L ± 0.76 ng/L.
Therefore, although the Marti and Armstrong data from 1980-1983 suggest that there may have been
significant decreases in PCB concentrations by the time of the LMMB Study, the results for the blanks in
the Marti and Armstrong data are as large or larger than the LMMB Study total PCB results for at least
five of the LMMB tributaries, complicating the evaluation of any historical trends, especially for the less
polluted tributaries.
4-26 April 2004
-------
PCBs/trans-Nonachlor in Tributaries
4.3.2 Regional Considerations
The results from this study generally support Figure 4-11. Mean Dissolved and Particulate Total PCB
the assumptions used to design the study - Concentrations in Lake Michigan Tributaries
namely that there are several tributaries that
contribute large amounts of PCBs and trans-
nonachlor to Lake Michigan and that there are
other tributaries that have much lower
concentrations (Figure 4-11). The tributaries
that contribute the largest amounts of PCBs and
trans-nonachlor are those near the Chicago
metropolitan area and on the western shore of
Lake Michigan (e.g., Grand Calumet,
Milwaukee, Sheboygan, and Fox).
60 n
50
40 -
30
20 -
10 -
o J
Mennminee
Manistique
;
Lake
Michigan
Pare Maiquette
Milwaukee
D Dissolved
• Participate
Kalamazoo
As these results are converted into pollutant
loads to the lake, management decisions can
focus on those tributaries where reductions are
most practical and on those that will have the
greatest impact on the overall concentrations of
contaminants in Lake Michigan.
4.3.3 Other Interpretations and Perspectives
In the Wisconsin DNR study of the lower Fox
River described in Section 4.3.1, the
investigators noted a correlation between the
concentrations of chlorophyll a and particulate
PCBs in the Fox River. That relationship was
subsequently investigated in the Milwaukee and
Manitowoc rivers with similar results
(Fitzgerald and Steuer, 1996).
The possible relationship between particulate PCB concentrations and chlorophyll a was examined using
the LMMB Study data. The particulate PCB results for 39 samples from the Fox River demonstrate a
strong correlation with chlorophyll a for both individual PCB congeners and total PCBs, while the
correlations in the Milwaukee River are not as strong (Table 4-11).
Table 4-11. Correlation of Particulate PCB and frans-Nonachlor Concentrations with Chlorophyll a in the Fox
and Milwaukee Rivers
~^St. Joseph
1 b
K 0 IS
Particulate-Phase Analyte
PCB 33
PCB 118
PCB 180
Total PCBs
frans-Nonachlor
Correlation with Chlorophyll a (r)
Fox River (n=39)
0.845
0.901
0.884
0.873
0.642
Milwaukee River (n=37)
0.343
0.625
0.595
0.627
0.596
April 2004
4-27
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Fitzgerald and Steuer attribute the correlation to a combination of the low solubility of PCBs, and thus
their affinity for particle surfaces, and active uptake of PCBs by algal cells. As a result, they consider
PCBs in the particulate phase to be subdivided into the abiotic fraction, comprised of the PCBs associated
with suspended particles including resuspended river sediments, and the biotic fraction, comprised of
algae that have incorporated PCBs into their cells. This biotic fraction is the lowest link in the
incorporation of PCBs into the food web.
Given the differences observed between the Fox and Milwaukee Rivers in Table 4-11, the correlations
between particulate total PCBs and chlorophyll a were examined for all 11 tributaries in the LMMB
Study (Table 4-12). The r-values for the correlations range from -0.180 to 0.895, with the strongest
correlations in the Sheboygan, Kalamazoo, and Fox Rivers. The correlations do not appear to be related
to overall particulate PCB concentrations because some relatively clean rivers have high correlations
(e.g., St. Joseph), while some rivers with much higher PCB concentrations show very low or even
negative correlations (e.g., Grand Calumet and Muskegon).
Data were collected for "total solids" during the LMMB Study. Total solids include both the suspended
solids and the dissolved solids and therefore, the total solids results will overestimate the concentration of
solid particles in the sample. However, there are strongly positive correlations between particulate total
PCBs and total solids in many of the tributaries (Table 4-12). The correlations with total solids generally
are similar to the correlations with chlorophyll a in most of the tributaries. The exceptions are the Grand
Calumet and Muskegon Rivers. The correlation with total solids is very strong in the Grand Calumet
River, while the correlation with chlorophyll a is very low. This suggests that the particulate PCBs in this
tributary are almost exclusively "abiotic." In the Muskegon, both correlations are very low, and with
different signs.
Table 4-12. Correlation of Particulate Total PCB Concentrations with Chlorophyll a and Total Solids in Lake
Michigan Tributaries
Tributary
Fox River
Grand Calumet
Grand River
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Correlation with Particulate Total PCBs (r)
Chlorophyll a
0.873
0.094
0.619
0.877
0.443
0.274
0.627
-0.180
0.613
0.895
0.718
Total Solids
0.848
0.841
0.892
0.849
0.354
0.495
0.826
0.056
0.591
0.859
0.786
4-28
April 2004
-------
Chapter 5
PCBs/trans-Nonachlor in Open-lake Water
5.1 Results
Open-lake samples were collected from 38 sampling stations in Lake Michigan, 2 stations in Green Bay,
and 1 station in Lake Huron. A total of 350 samples were collected and analyzed for PCBs and trans-
nonachlor. Samples were collected as described in Section 2.5.3, by pumping 100 to 1000 L of lake
water through a column packed with 250 g of XAD-2®, a macroreticular resin that traps hydrophobic
organic contaminants. A "pentaplate" filter was installed in the sampling train in front of the XAD-2®
column to collect the particulate matter suspended in the sample. Separate analyses were performed on
the XAD-2® resin and the filtered particulates from each sampling effort, yielding results for operationally
defined "dissolved" and "particulate" PCBs (Table 5-1) and frara-nonachlor (Table 5-2).
The results from two samples collected at Station MB63 in September 1995 are not included in the
summary tables. The results for these two samples were several orders of magnitude higher than any
other samples collected in the LMMB Study and were removed from consideration based on a consensus
of the LMMB modeling team, leaving 348 samples from the rest of the study. Interferences and
laboratory accidents further reduced the number of dissolved PCB results to 347 and reduced the number
of fra«5-nonachlor results to 341 dissolved results and 347 particulate results.
Of the 38 sampling stations in Lake Michigan, 25 are
"permanent" monitoring stations used by GLNPO and
other investigators for a variety of studies. One
additional permanent station is located in Lake
Michigan near the mouth of Green Bay (GB100M)
and two permanent stations are located in Green Bay
itself (GB17 and GB24M). Twelve stations were
established for the purposes of the Lake Michigan
Mass Balance Study (the "MB" stations in Figure 5-1)
and the one station in Lake Huron (LH54M) serves as
a means to assess the flux of contaminants from Lake
Michigan into Lake Huron. The station locations are
shown in Figure 5-1.
As noted in Chapter 2, there are 209 possible PCB
congeners, and the investigators in this study reported
results for 65 to 110 of these congeners, depending on
the capabilities of each laboratory. Battelle Marine
Sciences Laboratory determined results for 105
congeners or co-eluting congeners.
For the purposes of this report, we are presenting
summaries of the results for the following subset of all
of the analytes:
• PCB congener 33
• PCB congener 118
• PCB congener 180
• Total PCBs
• fra«5-nonachlor
Figure 5-1. Open-lake Sampling Stations
April 2004
5-1
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 5-1. Numbers of Open-lake Samples Analyzed for Dissolved and Participate PCB Congeners and Total
PCBs
Sampling Station
1
3
5
6
6A
9
13
17
18M
19M
20
21
23M
24
25
26
27M
31
36
38
40M
41
43
45
47M
52
57
63
72M
110
140
180
240
280
310
340
380
GB17
GB24M
GB100M
LH54M
Sampling Dates
05/1 0/94 to 10/1 1/95
05/1 0/94 to 10/1 3/95
05/1 1/94 to 10/1 0/95
05/09/94 to 10/1 2/95
05/09/94 to 10/1 3/95
05/1 1/94 to 10/1 0/95
05/09/94 to 10/1 3/95
05/07/94 to 10/09/95
05/07/94 to 10/09/95
05/05/94 to 10/05/95
05/05/94 to 10/06/95
05/07/94 to 10/04/95
05/04/94 to 10/03/95
05/05/94 to 10/05/95
05/03/94 to 09/29/95
05/02/94 to 09/27/95
05/02/94 to 09/28/95
05/02/94 to 09/28/95
05/01/94 to 09/27/95
04/28/94 to 09/22/95
04/30/94 to 09/26/95
05/01/94 to 09/27/95
05/01/94 to 09/26/95
04/26/94 to 09/20/95
04/26/94 to 09/20/95
04/26/94 to 09/1 9/95
04/25/94 to 09/1 8/95
04/25/94 to 09/1 8/95
04/25/94 to 09/1 7/95
06/1 9/94 to 09/24/95
04/28/94 to 09/23/95
06/1 8/94 to 09/22/95
05/04/94 to 10/02/95
05/04/94 to 10/02/95
05/06/94 to 10/08/95
05/06/94 to 10/07/95
05/06/94 to 10/06/95
04/27/94 to 09/22/95
04/27/94 to 09/21/95
04/26/94 to 09/20/95
04/24/94 to 09/1 7/95
Total
Dissolved PCBs
5
5
11
6
5
6
5
8
17
11
6
9
17
5
6
3
16
5
4
6
11
9
5
5
13
8
5
5
8
6
10
11
10
14
7
11
13
8
10
9
13
347
Part icu late PCBs
5
5
11
6
5
6
5
8
17
11
6
9
17
5
5
4
16
5
4
6
11
9
5
5
13
8
5
5
8
6
10
11
11
14
7
11
13
8
9
10
13
348
Total Samples
10
10
22
12
10
12
10
16
34
22
12
18
34
10
11
7
32
10
8
12
22
18
10
10
26
16
10
10
16
12
20
22
21
28
14
22
26
16
19
19
26
695
5-2
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Table 5-2. Number of Open-lake Samples Analyzed for Dissolved and Participate frans-Nonachlor
Sampling Station
1
3
5
6
6A
9
13
17
18M
19M
20
21
23M
24
25
26
27M
31
36
38
40M
41
43
45
47M
52
57
63
72M
110
140
180
240
280
310
340
380
GB17
GB24M
GB100M
LH54M
Sampling Dates
05/1 0/94 to 10/1 1/95
05/1 0/94 to 10/1 3/95
05/1 1/94 to 10/1 0/95
05/09/94 to 10/1 2/95
05/09/94 to 10/1 3/95
05/1 1/94 to 10/1 0/95
05/09/94 to 10/1 3/95
05/07/94 to 10/09/95
05/07/94 to 10/09/95
05/05/94 to 10/05/95
05/05/94 to 10/06/95
05/07/94 to 10/04/95
05/04/94 to 10/03/95
05/05/94 to 10/05/95
05/03/94 to 09/29/95
05/02/94 to 09/27/95
05/02/94 to 09/28/95
05/02/94 to 09/28/95
05/0 1/94 to 09/27/95
04/28/94 to 09/22/95
04/30/94 to 09/26/95
05/0 1/94 to 09/27/95
05/0 1/94 to 09/26/95
04/26/94 to 09/20/95
04/26/94 to 09/20/95
04/26/94 to 09/1 9/95
04/25/94 to 09/1 8/95
04/25/94 to 09/1 8/95
04/25/94 to 09/1 7/95
06/1 9/94 to 09/24/95
04/28/94 to 09/23/95
06/1 8/94 to 09/22/95
05/04/94 to 10/02/95
05/04/94 to 10/02/95
05/06/94 to 10/08/95
05/06/94 to 10/07/95
05/06/94 to 10/06/95
04/27/94 to 09/22/95
04/27/94 to 09/2 1/95
04/26/94 to 09/20/95
04/24/94 to 09/1 7/95
Total
Dissolved frans-Nonachlor
5
5
10
6
5
6
5
8
16
11
5
9
17
5
5
3
16
5
4
6
11
9
5
5
13
8
5
5
8
6
10
11
10
12
7
11
13
8
10
9
13
341
Particulate frans-Nonachlor
5
5
11
6
5
6
5
8
17
11
6
9
17
5
5
4
16
5
4
6
11
9
5
5
13
8
5
5
7
6
10
11
11
14
7
11
13
8
9
10
13
347
Total Samples
10
10
21
12
10
12
10
16
33
22
11
18
34
10
10
7
32
10
8
12
22
18
10
10
26
16
10
10
15
12
20
22
21
26
14
22
26
16
19
19
26
688
April 2004
5-3
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
5.1.1 Temporal Variation
Temporal variation was assessed by examining the mean concentrations of dissolved and particulate total
PCBs across seven cruises of the R/VLake Guardian (Figures 5-2 and 5-3). The data from the January
1995 cruise were not included in this assessment because winter lake conditions only permitted the
collection of samples at four stations. In general, the mean dissolved and particulate concentrations of
total PCBs show little variation over time, with no discernable temporal or seasonal trends. An analysis
of variance found no differences between the concentrations by cruise. In the context of a mass balance,
the concentrations of PCBs and fra«s-nonachlor in the open lake reflect a large number of inputs, internal
processes, and outputs of a complex ecosystem. The observed total PCB concentrations in the open lake
over the course of this study may reflect competing temporal trends among those inputs, outputs, and
processes. However, the apparent lack of temporal trends shown in these figures is complicated by
concerns about contamination of the XAD-2® resin. Those concerns are discussed in detail in Section
5.2.3.
Figure 5-2. Dissolved Total PCB by Cruise
.01
Figure 5-3. Dissolved Total PCB by Cruise
CO
o
Q_
10.
.01
.001
T
T
Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile (top of box) results. Bars represent the
results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest edge of the box. Circles represent results
beyond 1.5*IQR from the box. The Xs represent results beyond 3*IQR from the box. Concentration is plotted on a log scale and the scales for
the two figures are different.
5.1.2 Geographical Variation
The concentrations of dissolved and particulate PCBs and frara-nonachlor varied by station over the
course of the study (Tables 5-3 through 5-12). EPA researchers at the Large Lakes Research Station in
Grosse lie, Michigan, used the data from the LMMB Study to prepare "contour plots" of the lake where
similar concentrations of PCBs and fra«s-nonachlor are indicated using a color scale. Examples of such
plots are shown in Figures 5-4 to 5-11 for the dissolved and particulate concentrations of PCBs 33, 118,
and 180, and fra«s-nonachlor. Similar plots for dissolved total PCBs and particulate total PCBs are
shown in Figures 5-12 and 5-13.
The data in these tables and figures include all of the valid data from open lake samples, except the two
samples at Station MB63 discussed in Section 5.1. The use and interpretation of the results in these tables
and figures is complicated by concerns about contamination of the XAD-2® resin. Those concerns are
discussed in detail in Section 5.2.3.
5-4
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Table 5-3. Concentrations of Dissolved PCB Congener 33 Measured in Open-lake Samples
Sampling Station
1
3
5
6
6A
9
13
17
18M
19M
20
21
23M
24
25
26
27M
31
36
38
40M
41
43
45
47M
52
57
63
72M
110
140
180
240
280
310
340
380
GB17
GB24M
GB100M
LH54M
N
5
5
11
6
5
6
5
8
17
11
5
9
17
5
6
3
16
5
4
6
10
9
5
5
12
7
5
5
8
6
10
11
10
14
7
11
13
8
10
8
13
Mean (ng/L)
0.00748
0.00622
0.0187
0.0101
0.00630
0.0684
0.00718
0.00639
0.00617
0.00624
0.00749
0.00660
0.0107
0.00727
0.00445
0.00584
0.00416
0.00562
0.0296
0.00535
0.00514
0.00306
0.00448
0.00528
0.00460
0.00510
0.00374
0.00338
0.00366
0.00894
0.00373
0.00717
0.00496
0.0307
0.00787
0.00585
0.00590
0.0127
0.00512
0.00421
0.00398
Range (ng/L)
0.00392 to 0.0122
0.00418 to 0.00742
0.00217 to 0.0975
0.00349 to 0.0235
0.00516 to 0.00909
0.00234 to 0.206
0.00567 to 0.00981
0.00358 to 0.00826
0.00240 to 0.00991
0.00487 to 0.00958
0.00571 to 0.0120
0.00331 to 0.0152
0.00400 to 0.0853
0.00585 to 0.0106
0.00302 to 0.00611
0.00451 to 0.00722
0.00191 to 0.00830
0.00349 to 0.01 12
0.00382 to 0.1 00
0.00398 to 0.00730
0.00332 to 0.00947
0.00237 to 0.00438
0.00247 to 0.00785
0.00290 to 0.00918
0.00123 to 0.00700
0.00344 to 0.00658
0.00278 to 0.00459
0.00252 to 0.00499
0.00270 to 0.00469
0.00277 to 0.0337
0.00 to 0.00530
0.00325 to 0.0335
0.00346 to 0.00763
0.00250 to 0.1 94
0.00590 to 0.01 18
0.00305 to 0.00899
0.00202 to 0.01 14
0.00287 to 0.0254
0.00335 to 0.00806
0.00228 to 0.00707
0.001 16 to 0.0122
SD (ng/L)
0.00304
0.00137
0.0303
0.00750
0.00166
0.0987
0.00164
0.00159
0.00231
0.00136
0.00268
0.00350
0.0193
0.00202
0.00127
0.00136
0.00202
0.00324
0.0472
0.00112
0.00178
0.000671
0.00216
0.00244
0.00184
0.00117
0.000742
0.00094
0.000785
0.0121
0.00158
0.00876
0.00114
0.0651
0.00212
0.00174
0.00219
0.00709
0.00154
0.00170
0.00270
RSD (%)
41
22
163
74
26
144
23
25
37
22
36
53
180
28
28
23
49
58
160
21
35
22
48
46
40
23
20
28
21
136
42
122
23
212
27
30
37
56
30
41
68
% Below DL
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
10
0
0
0
0
0
0
0
0
0
0
April 2004
5-5
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 5-4. Concentrations of Participate PCB Congener 33 Measured in Open-lake Samples
Sampling Station
1
3
5
6
6A
9
13
17
18M
19M
20
21
23M
24
25
26
27M
31
36
38
40M
41
43
45
47M
52
57
63
72M
110
140
180
240
280
310
340
380
GB17
GB24M
GB100M
LH54M
N
5
5
11
6
5
6
5
8
17
11
6
9
17
5
5
4
16
5
4
6
11
9
5
5
13
8
5
5
7
6
10
11
11
13
7
11
13
8
9
10
13
Mean (ng/L)
0.00148
0.00113
0.000634
0.000413
0.00205
0.000694
0.00208
0.000550
0.000573
0.000605
0.00101
0.000593
0.000615
0.000793
0.000676
0.000910
0.000497
0.000842
0.00476
0.000438
0.000207
0.000302
0.000740
0.000446
0.000360
0.000441
0.000732
0.00123
0.000112
0.000214
0.000566
0.000351
0.000466
0.000349
0.00331
0.00147
0.000825
0.0283
0.00108
0.000237
0.000245
Range (ng/L)
0.00 to 0.00530
0.00 to 0.00375
0.00 to 0.00361
0.00 to 0.00133
0.00 to 0.00770
0.000129 to 0.00126
0.00 to 0.00698
0.00 to 0.00194
0.00 to 0.00233
0.00 to 0.00244
0.000193 to 0.00195
0.00 to 0.00158
0.00 to 0.00382
0.000188 to 0.00225
0.000142 to 0.00157
0.0000872 to 0.00285
0.00 to 0.00166
0.00 to 0.00290
0.00 to 0.0180
0.00 to 0.00168
0.00 to 0.000649
0.00 to 0.00202
0.00 to 0.00305
0.0001 17 to 0.00145
0.00 to 0.00236
0.00 to 0.00238
0.00 to 0.00332
0.00 to 0.00516
0.00 to 0.000285
0.00 to 0.000648
0.00 to 0.00268
0.00 to 0.000833
0.00 to 0.00204
0.00 to 0.000890
0.000611 to 0.0100
0.00 to 0.00414
0.00 to 0.00266
0.0130 to 0.0578
0.000209 to 0.00340
0.00 to 0.000714
0.00 to 0.00125
SD (ng/L)
0.00218
0.00151
0.00100
0.000501
0.00320
0.000421
0.00286
0.000619
0.000788
0.000898
0.000694
0.000497
0.000942
0.000853
0.000627
0.00131
0.000519
0.00120
0.00884
0.000633
0.000258
0.000666
0.00130
0.000566
0.000681
0.000798
0.00145
0.00220
0.000135
0.000238
0.000789
0.000315
0.000615
0.000281
0.00330
0.00132
0.000692
0.0131
0.000977
0.000249
0.000377
RSD (%)
147
133
158
121
156
61
137
113
138
149
69
84
153
108
93
144
104
143
185
145
125
220
175
127
189
181
198
178
121
111
139
90
132
80
100
90
84
46
90
105
154
% Below DL
20
20
10
33
20
0
20
25
29
27
0
11
18
0
0
25
13
20
25
33
55
67
20
0
62
25
60
40
57
33
30
27
36
23
0
18
15
0
0
40
54
5-6
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Table 5-5. Concentrations of Dissolved PCB Congener 118 Measured in Open-lake Samples
Sampling Station
1
3
5
6
6A
9
13
17
18M
19M
20
21
23M
24
25
26
27M
31
36
38
40M
41
43
45
47M
52
57
63
72M
110
140
180
240
280
310
340
380
GB17
GB24M
GB100M
LH54M
N
5
5
11
6
5
6
5
8
17
11
5
9
17
5
6
3
16
5
4
6
10
9
5
5
12
7
5
5
8
6
10
11
10
14
7
11
13
8
10
8
13
Mean (ng/L)
0.00117
0.00129
0.00306
0.00227
0.000921
0.00356
0.00167
0.00318
0.00330
0.00363
0.00245
0.00328
0.00218
0.00275
0.00253
0.000825
0.00142
0.00260
0.00426
0.00363
0.00256
0.000788
0.00174
0.00171
0.00258
0.000790
0.00173
0.00088
0.000760
0.00230
0.00219
0.00265
0.00254
0.00420
0.00352
0.00286
0.00272
0.00406
0.00218
0.00218
0.00163
Range (ng/L)
0.00 to 0.00183
0.00 to 0.00252
0.00 to 0.00974
0.000869 to 0.00407
0.00 to 0.00250
0.00 to 0.00802
0.000797 to 0.00232
0.00 to 0.0128
0.00 to 0.0125
0.000223 to 0.01 19
0.00 to 0.00745
0.000858 to 0.0142
0.000394 to 0.00555
0.00 to 0.00896
0.000616 to 0.00948
0.000377 to 0.00120
0.000338 to 0.00522
0.00 to 0.0104
0.000273 to 0.0142
0.000216 to 0.0139
0.0000681 to 0.01 15
0.00 to 0.00208
0.000159 to 0.00396
0.000626 to 0.00353
0.000269 to 0.00994
0.00 to 0.00207
0.000202 to 0.00453
0.00039 to 0.00166
0.00 to 0.00208
0.000255 to 0.00846
0.0000899 to 0.00912
0.000228 to 0.00991
0.000738 to 0.00823
0.00 to 0.0183
0.00144 to 0.00552
0.00 to 0.00867
0.00 to 0.01 11
0.00195 to 0.00727
0.000302 to 0.00729
0.000422 to 0.00798
0.0000737 to 0.00947
SD (ng/L)
0.000747
0.00109
0.00302
0.00130
0.000978
0.00297
0.000636
0.00405
0.00331
0.00359
0.00305
0.00426
0.00149
0.00362
0.00346
0.000417
0.00109
0.00439
0.00663
0.00520
0.00338
0.000635
0.00146
0.00132
0.00342
0.000670
0.00192
0.00056
0.000676
0.00320
0.00294
0.00274
0.00259
0.00580
0.00167
0.00268
0.00328
0.00183
0.00206
0.00249
0.00254
RSD (%)
64
85
99
57
106
84
38
127
100
99
124
130
69
132
137
50
77
169
156
143
132
81
84
77
132
85
111
63
89
139
134
103
102
138
47
94
121
45
94
114
156
% Below DL
20
40
18
0
60
33
0
25
6
18
40
0
6
40
33
33
6
40
25
33
40
44
20
40
33
57
60
60
63
33
30
18
0
14
0
18
15
0
10
13
38
April 2004
5-7
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 5-6. Concentrations of Participate PCB Congener 118 Measured in Open-lake Samples
Sampling Station
1
3
5
6
6A
9
13
17
18M
19M
20
21
23M
24
25
26
27M
31
36
38
40M
41
43
45
47M
52
57
63
72M
110
140
180
240
280
310
340
380
GB17
GB24M
GB100M
LH54M
N
5
5
11
6
5
6
5
8
17
11
6
9
17
5
5
4
16
5
4
6
11
9
5
5
13
8
5
5
7
6
10
11
11
13
7
11
13
8
9
10
13
Mean (ng/L)
0.00252
0.00223
0.00195
0.00203
0.00358
0.00194
0.00378
0.00255
0.00215
0.00272
0.00307
0.00275
0.00284
0.00245
0.00172
0.00218
0.00224
0.00162
0.00254
0.00136
0.000581
0.000553
0.00130
0.00133
0.000618
0.00104
0.00114
0.00141
0.000637
0.00121
0.00140
0.00141
0.00176
0.00214
0.0101
0.00436
0.00371
0.0157
0.00233
0.00119
0.000839
Range (ng/L)
0.000610 to 0.00730
0.000269 to 0.00683
0.00031 5 to 0.00450
0.00 to 0.00402
0.00 to 0.0136
0.000354 to 0.00441
0.00 to 0.0131
0.0001 11 to 0.00420
0.00 to 0.0139
0.0000860 to 0.0124
0.001 11 to 0.00504
0.000704 to 0.0081 6
0.00 to 0.0129
0.000849 to 0.00510
0.000476 to 0.00288
0.00109 to 0.00352
0.000109 to 0.00702
0.000527 to 0.00258
0.000722 to 0.00457
0.000543 to 0.00267
0.00 to 0.00162
0.000229 to 0.000942
0.000407 to 0.00239
0.000450 to 0.00241
0.00 to 0.00166
0.00 to 0.00270
0.00 to 0.00232
0.00 to 0.002550
0.00 to 0.00136
0.000398 to 0.00204
0.000267 to 0.00280
0.0000435 to 0.00273
0.0000510 to 0.00368
0.0000664 to 0.00411
0.00240 to 0.0263
0.0000385 to 0.0122
0.0000735 to 0.00660
0.00 to 0.0326
0.000283 to 0.00410
0.00 to 0.00226
0.00 to 0.00230
SD (ng/L)
0.00282
0.00264
0.00134
0.00198
0.00567
0.00154
0.00537
0.00176
0.00376
0.00352
0.00156
0.00236
0.00325
0.00173
0.000914
0.00120
0.00224
0.000934
0.00177
0.000841
0.000494
0.000300
0.000957
0.000777
0.000556
0.000927
0.00106
0.00118
0.000495
0.000723
0.000988
0.00104
0.00149
0.00148
0.00852
0.00388
0.00228
0.0116
0.00121
0.000806
0.000821
RSD (%)
112
118
69
98
158
79
142
69
174
130
51
86
114
71
53
55
100
58
70
62
85
54
74
58
90
89
93
83
78
60
71
74
85
69
84
89
61
74
52
68
98
% Below DL
0
20
0
17
40
0
20
13
29
9
0
0
12
0
0
0
13
0
0
0
18
22
0
0
31
13
20
20
29
0
0
9
18
15
0
18
8
25
11
10
38
5-8
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Table 5-7. Concentrations of Dissolved PCB Congener 180 Measured in Open-lake Samples
Sampling Station
1
3
5
6
6A
9
13
17
18M
19M
20
21
23M
24
25
26
27M
31
36
38
40M
41
43
45
47M
52
57
63
72M
110
140
180
240
280
310
340
380
GB17
GB24M
GB100M
LH54M
N
5
5
11
6
5
6
5
8
17
11
5
9
17
5
6
3
16
5
4
6
10
9
5
5
12
7
5
5
8
6
10
11
10
14
7
11
13
8
10
8
13
Mean (ng/L)
0.000340
0.000143
0.00311
0.000400
0.000371
0.000882
0.000486
0.000383
0.000258
0.000314
0.000289
0.000263
0.000119
0.000757
0.000555
0.000518
0.000294
0.000535
0.00262
0.0000565
0.000107
0.000165
0.000228
0.000148
0.0000981
0.000151
0.000858
0.000183
0.0000796
0.000556
0.0000667
0.000116
0.000414
0.00145
0.000712
0.000170
0.000445
0.00109
0.000296
0.000137
0.000121
Range (ng/L)
0.00 to 0.00156
0.00 to 0.00228
0.00 to 0.000747
0.00 to 0.000345
0.00 to 0.00405
0.00 to 0.000347
0.00 to 0.0295
0.00 to 0.00160
0.00 to 0.00309
0.00 to 0.00211
0.00 to 0.000381
0.00 to 0.00211
0.00 to 0.000563
0.00 to 0.00258
0.00 to 0.000882
0.00 to 0.00139
0.00 to 0.000891
0.00 to 0.00223
0.00 to 0.0124
0.00 to 0.00261
0.00 to 0.00144
0.00 to 0.00982
0.00 to 0.000128
0.00 to 0.00346
0.00 to 0.00051 8
0.00 to 0.000630
0.00 to 0.000291
0.00 to 0.00060
0.00 to 0.0479
0.00 to 0.000458
0.00 to 0.00177
0.00 to 0.00476
0.00 to 0.00196
0.00 to 0.00334
0.00 to 0.001 14
0.00 to 0.00209
0.00 to 0.00404
0.00 to 0.00699
0.00 to 0.00186
0.00 to 0.000282
0.00 to 0.000527
SD (ng/L)
0.000684
0.000208
0.00880
0.000692
0.000694
0.00191
0.000921
0.000748
0.000642
0.000605
0.000411
0.000455
0.000260
0.00146
0.000855
0.000580
0.000516
0.00116
0.00481
0.0000637
0.000181
0.000252
0.000316
0.000145
0.000131
0.000246
0.00179
0.00024
0.000127
0.00124
0.000126
0.000183
0.000724
0.00343
0.00151
0.000441
0.000970
0.00240
0.000568
0.000117
0.000175
RSD (%)
201
175
153
133
208
159
283
187
223
190
189
195
158
249
142
173
218
154
236
217
259
183
113
218
170
138
98
133
172
146
173
217
192
193
112
175
212
220
192
85
144
% Below DL
80
70
89
83
80
100
64
80
83
80
90
75
91
82
60
78
88
50
71
80
82
50
100
69
80
60
100
100
57
80
67
67
73
60
33
69
71
63
80
100
85
April 2004
5-9
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 5-8. Concentrations of Participate PCB Congener 180 Measured in Open-lake Samples
Sampling Station
1
3
5
6
6A
9
13
17
18M
19M
20
21
23M
24
25
26
27M
31
36
38
40M
41
43
45
47M
52
57
63
72M
110
140
180
240
280
310
340
380
GB17
GB24M
GB100M
LH54M
N
5
5
11
6
5
6
5
8
17
11
6
9
17
5
5
4
16
5
4
6
11
9
5
5
13
8
5
5
7
6
10
11
11
13
7
11
13
8
9
10
13
Mean (ng/L)
0.000974
0.000943
0.000807
0.00115
0.00160
0.000654
0.00162
0.00111
0.000929
0.00123
0.00178
0.00110
0.00112
0.00143
0.000415
0.000919
0.000729
0.000553
0.00115
0.000329
0.000200
0.000114
0.000581
0.000552
0.000336
0.000532
0.000709
0.00093
0.000268
0.000221
0.000365
0.000277
0.000708
0.000751
0.00431
0.00219
0.00180
0.00644
0.000627
0.000593
0.000314
Range (ng/L)
0.00 to 0.00248
0.00 to 0.00364
0.00 to 0.00103
0.00 to 0.00168
0.00 to 0.00288
0.00 to 0.00125
0.00 to 0.00246
0.00 to 0.00568
0.00 to 0.000577
0.00 to 0.00531
0.00 to 0.00168
0.00 to 0.00278
0.00 to 0.000954
0.00 to 0.00547
0.0000299 to 0.00371
0.00 to 0.00241
0.00 to 0.00407
0.00 to 0.00183
0.00 to 0.00183
0.00 to 0.00161
0.00 to 0.00586
0.00 to 0.00332
0.00 to 0.00197
0.00 to 0.00477
0.00 to 0.00139
0.00 to 0.00232
0.00 to 0.00164
0.00 to 0.00207
0.00 to 0.00207
0.00 to 0.00276
0.00 to 0.00276
0.00 to 0.00142
0.00 to 0.00477
0.00 to 0.00516
0.00 to 0.00296
0.00 to 0.00307
0.000761 to 0.01 17
0.00415 to 0.0142
0.00 to 0.00238
0.00 to 0.00199
0.00 to 0.001 11
SD (ng/L)
0.00128
0.00121
0.000343
0.000488
0.00123
0.000485
0.000864
0.00238
0.000283
0.00219
0.000534
0.00104
0.000373
0.00151
0.00144
0.000886
0.00122
0.000800
0.000663
0.000664
0.00189
0.00157
0.000805
0.00139
0.000455
0.00100
0.000710
0.000913
0.000833
0.00119
0.00135
0.000511
0.00167
0.00218
0.00140
0.000902
0.00399
0.00332
0.000840
0.000730
0.000428
RSD (%)
132
170
300
145
173
181
107
149
128
135
146
93
135
162
81
80
109
193
88
120
86
137
245
77
227
173
129
98
115
126
118
78
135
152
153
124
93
52
134
123
136
% Below DL
40
55
89
46
40
71
36
40
50
40
50
38
55
47
17
22
29
60
31
40
18
50
83
15
82
60
40
60
29
40
50
17
45
20
50
38
0
0
44
50
62
5-10
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Table 5-9. Concentrations of Dissolved Total PCBs Measured in Open-lake Samples
Sampling Station
1
3
5
6
6A
9
13
17
18M
19M
20
21
23M
24
25
26
27M
31
36
38
40M
41
43
45
47M
52
57
63
72M
110
140
180
240
280
310
340
380
GB17
GB24M
GB100M
LH54M
N
5
5
11
6
5
6
5
8
17
11
6
9
17
5
6
3
16
5
4
6
11
9
5
5
13
8
5
5
8
6
10
11
10
14
7
11
13
8
10
9
13
Mean (ng/L)
0.172
0.187
0.303
0.168
0.169
0.364
0.199
0.180
0.166
0.169
0.181
0.189
0.174
0.207
0.140
0.121
0.131
0.144
0.299
0.185
0.145
0.104
0.142
0.232
0.133
0.115
0.121
0.107
0.118
0.257
0.178
0.164
0.159
0.245
0.373
0.179
0.182
0.653
0.166
0.289
0.129
Range (ng/L)
0.0796 to 0.227
0.107 to 0.137
0.0758 to 0.574
0.0864 to 0.1 65
0.1 38 to 0.200
0.0613 to 0.898
0.0713 to 0.865
0.1 03 to 0.788
0.0932 to 0.528
0.142 to 0.219
0.0791 to 0.248
0.0669 to 0.304
0.0996 to 0.232
0.00 to 0.319
0.101 to 0.352
0.170 to 0.312
0.0789 to 0.274
0.0722 to 0.253
0.1 85 to 0.949
0.1 23 to 0.243
0.1 13 to 0.384
0.128 to 0.270
0.00 to 0.269
0.0692 to 0.1 26
0.0806 to 0.216
0.00 to 0.272
0.00 to 0.1 67
0.0731 to 0.1 56
0.0682 to 0.204
0.1 08 to 0.243
0.100 to 0.209
0.1 57 to 0.280
0.0920 to 0.249
0.1 00 to 0.247
0.0673 to 0.944
0.0992 to 0.219
0.1 16 to 0.734
0.290 to 1.52
0.0844 to 0.247
0.00 to 0.634
0.0526 to 0.209
SD (ng/L)
0.0616
0.0566
0.282
0.0447
0.0271
0.360
0.0493
0.0285
0.0636
0.0450
0.106
0.0600
0.0652
0.0594
0.0557
0.0151
0.0474
0.0531
0.292
0.104
0.0720
0.0216
0.0550
0.197
0.0761
0.0503
0.0318
0.3058
0.0500
0.262
0.130
0.0641
0.0629
0.239
0.280
0.0421
0.0465
0.420
0.0602
0.256
0.0556
RSD (%)
36
13
85
26
16
99
93
102
73
16
39
38
27
59
38
29
40
36
75
23
56
26
50
21
39
57
44
31
42
30
27
25
32
40
97
37
98
64
36
89
43
April 2004
5-11
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 5-10. Concentrations of Participate Total PCBs Measured in Open-lake Samples
Sampling Station
1
3
5
6
6A
9
13
17
18M
19M
20
21
23M
24
25
26
27M
31
36
38
40M
41
43
45
47M
52
57
63
72M
110
140
180
240
280
310
340
380
GB17
GB24M
GB100M
LH54M
N
5
5
11
6
5
6
5
8
17
11
6
9
17
5
5
4
16
5
4
6
11
9
5
5
13
8
5
5
8
6
10
11
11
14
7
11
13
8
9
10
13
Mean (ng/L)
0.100
0.0830
0.0699
0.0563
0.140
0.0728
0.138
0.0707
0.0652
0.0757
0.103
0.0910
0.0754
0.0870
0.0522
0.0843
0.0630
0.0619
0.150
0.0575
0.0271
0.0307
0.0549
0.0527
0.0312
0.0416
0.0558
0.0654
0.0275
0.0424
0.0547
0.0500
0.0590
0.0581
0.297
0.134
0.103
1.02
0.0810
0.0459
0.0379
Range (ng/L)
0.0334 to 0.270
0.0344 to 0.1 44
0.0259 to 0.0877
0.0159 to 0.137
0.0217 to 0.449
0.0222 to 0.1 26
0.0196 to 0.170
0.0161 to 0.0774
0.0224 to 0.0996
0.0131 to 0.124
0.0141 to 0.0814
0.01 13 to 0.264
0.00836 to 0.269
0.0355 to 0.1 81
0.0129 to 0.245
0.0279 to 0.181
0.0104 to 0.115
0.0103 to 0.152
0.0776 to 0.744
0.0127 to 0.365
0.0211 to 0.122
0.0151 to 0.181
0.0144 to 0.0541
0.01 15 to 0.0734
0.0165 to 0.123
0.00637 to 0.0894
0.0157 to 0.112
0.0203 to 0.1 59
0.00 to 0.0663
0.0141 to 0.234
0.0129 to 0.109
0.01 10 to 0.420
0.0306 to 0.200
0.0199 to 0.0681
0.00 to 0.101
0.0232 to 0.1 08
0.0253 to 0.421
0.468 to 2.30
0.0402 to 0.1 37
0.0105 to 0.109
0.00976 to 0.1 03
SD (ng/L)
0.102
0.0867
0.0490
0.0445
0.174
0.0349
0.165
0.0416
0.0712
0.0783
0.0573
0.0479
0.0620
0.0653
0.0203
0.0548
0.0434
0.0354
0.183
0.0385
0.0132
0.0198
0.0443
0.0269
0.0239
0.0333
0.0512
0.055
0.0205
0.0236
0.0297
0.0261
0.0418
0.0358
0.237
0.106
0.0549
0.553
0.0341
0.0312
0.0268
RSD (%)
101
65
51
92
124
48
70
56
54
59
52
109
103
55
82
75
71
69
80
79
67
53
49
64
81
77
80
84
75
104
79
120
53
39
62
57
122
54
42
68
71
5-12
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Table 5-11. Concentrations of Dissolved frans-Nonachlor Measured in Open-lake Samples
Sampling Station
1
3
5
6
6A
9
13
17
18M
19M
20
21
23M
24
25
26
27M
31
36
38
40M
41
43
45
47M
52
57
63
72M
110
140
180
240
280
310
340
380
GB17
GB24M
GB100M
LH54M
N
5
5
10
6
5
6
5
8
16
11
5
9
17
5
5
3
16
5
4
6
11
9
5
5
13
8
5
5
8
6
10
11
10
12
7
11
13
8
10
9
13
Mean (ng/L)
0.00507
0.00545
0.0111
0.00870
0.00429
0.00430
0.00548
0.0236
0.00659
0.00773
0.00450
0.00366
0.00368
0.00579
0.00478
0.00440
0.00524
0.00402
0.00228
0.00432
0.00433
0.00475
0.00289
0.00397
0.00388
0.00375
0.00232
0.00296
0.00330
0.00728
0.00869
0.00481
0.00627
0.00637
0.00530
0.00551
0.00604
0.00560
0.00348
0.00531
0.00876
Range (ng/L)
0.00227 to 0.00863
0.00 to 0.01 19
0.000359 to 0.0124
0.00 to 0.0102
0.00 to 0.00594
0.00 to 0.00590
0.00160 to 0.0392
0.000807 to 0.00786
0.00124 to 0.0165
0.000917 to 0.00888
0.00 to 0.0463
0.00 to 0.1 45
0.00 to 0.0102
0.00 to 0.0131
0.00 to 0.00984
0.00 to 0.0150
0.00 to 0.0108
0.001 18 to 0.0101
0.00 to 0.0145
0.00 to 0.00962
0.00 to 0.0138
0.00 to 0.00913
0.00 to 0.0127
0.000786 to 0.0129
0.00 to 0.00998
0.000239 to 0.00650
0.00296 to 0.00542
0.00152 to 0.00494
0.00152 to 0.176
0.00108 to 0.00832
0.00 to 0.0193
0.00 to 0.00932
0.00 to 0.0169
0.00146 to 0.00789
0.00 to 0.00930
0.00 to 0.0139
0.00253 to 0.00749
0.00145 to 0.0159
0.00 to 0.00733
0.00 to 0.01 13
0.00 to 0.0774
SD (ng/L)
0.00303
0.00296
0.0118
0.00656
0.00302
0.00405
0.00338
0.0494
0.00393
0.00613
0.00423
0.00467
0.00367
0.00258
0.00361
0.00467
0.00432
0.00366
0.00457
0.00467
0.00376
0.00361
0.00251
0.000926
0.00300
0.00261
0.00266
0.00135
0.00207
0.00516
0.0136
0.00328
0.00368
0.00422
0.00184
0.00483
0.00383
0.00497
0.00213
0.00347
0.0210
RSD (%)
60
59
76
77
115
63
106
70
71
62
156
209
68
60
94
127
100
76
66
91
88
200
108
63
87
87
23
46
238
54
75
94
79
45
106
82
35
89
61
65
239
% Below DL
0
10
11
23
40
25
0
20
17
20
30
13
18
19
40
44
41
20
8
20
27
75
33
8
27
40
0
20
0
20
17
33
18
0
33
19
0
0
10
22
46
April 2004
5-13
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 5-12. Concentrations of Participate frans-Nonachlor Measured in Open-lake Samples
Sampling Station
1
3
5
6
6A
9
13
17
18M
19M
20
21
23M
24
25
26
27M
31
36
38
40M
41
43
45
47M
52
57
63
72M
110
140
180
240
280
310
340
380
GB17
GB24M
GB100M
LH54M
N
5
5
11
6
5
6
5
8
17
11
6
9
17
5
5
4
16
5
4
6
11
9
5
5
13
8
5
5
7
6
10
11
11
14
7
11
13
8
9
10
13
Mean (ng/L)
0.0026
0.0012
0.0027
0.0013
0.0040
0.0025
0.0014
0.0015
0.0018
0.0012
0.0021
0.0027
0.0021
0.0027
0.0014
0.0024
0.0022
0.0016
0.0031
0.0025
0.0011
0.0015
0.0023
0.0026
0.0014
0.0017
0.0025
0.00226
0.0016
0.0023
0.0029
0.0017
0.0023
0.0025
0.0030
0.0031
0.0025
0.0037
0.0020
0.0019
0.0011
Range (ng/L)
0.000509 to 0.00844
0.000365 to 0.00771
0.000109 to 0.00575
0.00 to 0.00644
0.000396 to 0.00687
0.000531 to 0.00298
0.00 to 0.0104
0.00121 to 0.00621
0.000571 to 0.00636
0.00 to 0.00245
0.000552 to 0.00652
0.000912 to 0.00342
0.00 to 0.00475
0.00 to 0.00844
0.000773 to 0.00403
0.00 to 0.00665
0.000391 to 0.00622
0.000468 to 0.00217
0.000180 to 0.00657
0.000423 to 0.00228
0.00 to 0.00889
0.00102 to 0.00639
0.000604 to 0.00683
0.000452 to 0.00679
0.00 to 0.00344
0.000887 to 0.00583
0.000264 to 0.00596
0.00 to 0.00542
0.00 to 0.00542
0.00 to 0.00261
0.00 to 0.00474
0.000508 to 0.00583
0.00 to 0.00409
0.000501 to 0.00570
0.000918 to 0.00643
0.00 to 0.00791
0.000887 to 0.00761
0.001 19 to 0.00893
0.000253 to 0.00355
0.00 to 0.00580
0.00 to 0.00307
SD (ng/L)
0.0033
0.0011
0.0029
0.0017
0.0023
0.0019
0.0010
0.0008
0.0024
0.0011
0.0013
0.0022
0.0016
0.0021
0.0007
0.0027
0.0021
0.0008
0.0024
0.0024
0.0010
0.0018
0.0020
0.0023
0.0018
0.0016
0.0026
0.0022
0.0009
0.0022
0.0023
0.0014
0.0022
0.0021
0.0022
0.0029
0.0021
0.0025
0.0011
0.0020
0.0009
RSD (%)
127
96
116
133
103
60
105
57
98
72
79
52
82
138
64
80
76
51
84
46
95
79
96
84
88
90
90
97
109
91
132
75
93
77
112
96
75
66
56
102
87
% Below DL
0
9
22
31
0
0
18
0
0
20
0
0
18
29
0
11
0
0
14
20
9
0
0
0
27
0
20
20
14
20
33
0
18
0
0
13
0
0
11
20
38
5-14
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Figures 5-4 to 5-13 illustrate the concentrations of PCBs 33, 118, 180, total PCBs, and fra«s-nonachlor in
the dissolved and particulate samples collected over the course of the LMMB Study.
Note: The color scales used in these contour plots vary with each plot. Therefore, although the red end
of the visible spectrum always represents higher concentrations than the violet end of the
spectrum, the absolute magnitude represented by each color differs with the contaminant and the
phase (dissolved versus particulate). Each plot includes a concentration scale and readers are
advised to consult those scales carefully when comparing the plots.
The plots of the dissolved PCB congeners (Figures 5-4 to 5-6) indicate that the concentrations of these
contaminants are generally lowest in the far northern areas of the lake that are removed from urban
influences. The highest dissolved concentrations generally are found in the southwest area of the lake,
centered around Station 9, which lies between the urban areas of Chicago and Milwaukee. The dissolved
PCB concentrations suggest that there may be a point source at Waukegan Harbor, Illinois. The
concentrations of these dissolved contaminants show some increase in Green Bay, with dissolved PCB
118 concentrations highest overall at Station GB 17, near the discharge of the Fox River. (The apparent
decrease in concentrations of these contaminants from Station GB 17 to the head of Green Bay likely is a
function of the lack of a sampling station further up Green Bay).
The plots of the particulate PCB congeners (Figures 5-7 to 5-9) illustrate the importance of contaminant
sources in Green Bay. The particulate PCB concentrations are highest in Green Bay, at Station GB 17,
with much lower particulate PCB concentrations in the remainder of the lake. The particulate
concentrations of PCBs 118 and 180 show a slight increase in the southeast portion of the lake, in the area
between the mouths of the St. Joseph and Kalamazoo Rivers. However, the concentrations of particulate
PCBs 118 and 180 in that area are still 2 to 5 times lower than in the upper reaches of Green Bay.
The patterns of dissolved concentrations of fra«s-nonachlor (Figure 5-10) are similar to those of the
dissolved PCB congeners, with an apparent increase in concentration in the southwest portion of the lake,
near Chicago. The increase in concentration on the western shore of the lake seen in Figure 5-10 suggests
that the Sturgeon Bay ship canal, which connects lower Green Bay with this portion of Lake Michigan,
results in the transfer of water containing dissolved trans-nonachlor from Green Bay into Lake Michigan.
The particulate concentrations of fra«s-nonachlor (Figure 5-11) are similar to those of the particulate PCB
congeners, with the highest concentrations in Green Bay, near Station GB 17. However, particulate
fra«5-nonachlor concentrations appear to be increased in areas of the lake adjacent to most of the major
urban areas around the lake. However, similar increases occur near the discharges of the Manistique and
Pere Marquette Rivers, which are not associated with urban areas, suggesting that the increases near the
urban area may be a function of river-borne sources of particulate trans-nonachlor, including
resuspension of contaminated sediments. The particulate fra«s-nonachlor results also suggest the transfer
of contaminants from lower Green Bay to Lake Michigan via the Sturgeon Bay ship canal.
The plots of dissolved and particulate total PCBs (Figures 5-12 and 5-13) illustrate the importance of the
sources of these contaminants in Green Bay. The dissolved total PCB concentrations are highest in Green
Bay. However, the dissolved total PCB concentrations in the lower portion of Lake Michigan are higher
than in the northern portion of the lake, with an apparent hot spot near Chicago. The particulate total
PCB concentrations are also highest in Green Bay and lower throughout the main portion of the lake, with
a slight increase in the southeast portion of the lake, similar to the particulate PCB 118 and 180 results.
The dissolved and particulate PCB results show less indication of the transfer of contaminants from lower
Green Bay to Lake Michigan via the Sturgeon Bay ship canal than the results for trans-nonachlor, with
increases for the congeners reported here evident only for dissolved PCB 118 and dissolved total PCBs.
April 2004 5-15
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 5-4. Concentrations of Dissolved PCB Congener 33 Measured in Open-lake Samples
•••
•
:
••
I I I
Lake Michigan Mass Balance Project
N
Sample Type: DISSOLVED
Sample Nome: PCB i3
Unll: (ng/L)
'
Survey Date:
4/24/94 - 10/13/95
Sample Depths:
OM - 300M
41 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
-89
-88
-87
-86
-85
-84
5-16
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Figure 5-5. Concentrations of Dissolved PCB Congener 118 Measured in Open-lake Samples
••
•
.: :
•
•••.
I I I I I I
Lake Michigan Mass Balance Project
N
'
Sample Type: DISSOLVED
Sample Nome: PCB 118
Unit: (ng/L)
0.00476
0.00436
0.00396
0.00356
0.00316
0.00276
0.00236
0.00196
0.00156
0.00116
0.00076
Survey Date:
4/24/9* - 10/13/95
Sample Depths:
OM - 300M
41 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
-89
-58
-87
-86
-85
-84
April 2004
5-17
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 5-6. Concentrations of Dissolved PCB Congener 180 Measured in Open-lake Samples
•
:
•
: -
•
•:
! I I
Lake Michigan Mass Balance Project
N
Sompl« Type: DISSOLVED
Sample Nome: PCB 180
Unit: (ng/L)
'
Survey Date:
4/24/94 - 10/13/95
Sample Depths:
OM - 300M
41 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
-89
-88
-87
-86
-85
-84
5-18
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Figure 5-7. Concentrations of Participate PCB Congener 33 Measured in Open-lake Samples
April 2004
5-19
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 5-8. Concentrations of Participate PCB Congener 118 Measured in Open-lake Samples
••
•
:
•:
I I l
Lake Michigan Mass Balance Project
N
'
Sample Type: PARTICULAR
Sample Name: PCB 118
Unit: (ng/L)
0.01912
0.01726
0.01541
0.01355
0.01169
0.00984
0.00798
0.00612
0.00427
0.00241
0.00055
Survey Date:
4/24/9* - 10/13/95
Sample Depths:
OM - 300M
41 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I i I I I
-89 -88 -87 -86 -85
-84
5-20
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Figure 5-9. Concentrations of Participate PCB Congener 180 Measured in Open-lake Samples
••
•
.; :
:
•:
Lake Michigan Mass Balance Project
N
'
Sample Type: PARTICULATf
Sample Nome: PCS 180
Unit: (ng/L)
0.00680
0.00613
0.00546
0.00479
0.00412
0.00346
0.00279
0.00212
0.00145
0.00078
0.00011
Survey Date:
4/24/94 - 10/13/95
Sample Depths:
OM - 300M
41 I I I I I I I I I I I I
-89 -88
I I I i I i I i I I i I I I i I I I i I I I I i I I i I i I
-87
-86
-85
-84
April 2004
5-21
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 5-10. Concentrations of Dissolved frans-Nonachlor Measured in Open-lake Samples
•••
•
.-, :
:
•:
I I l
Lake Michigan Mass Balance Project
N
'
Sample Type: DISSOLVED
Sample Name: frans-nonoch!or
Unit: (ng/L)
Survey Date:
4/24/94 - 10/13/95
Sample Depths:
OM - 300M
41 I I I I I I I I I I ! I I I I I I I I I I I I I I I 1 I I I I I I I I I I I I I I I
-89
-88
-87
-86
-85
-84
5-22
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Figure 5-11. Concentrations of Participate trans-Nonachlor Measured in Open-lake Samples
,:,
•
.-, :
:
•:
! I I
Lake Michigan Mass Balance Project
N
'
Sompl. Type: PARTICULATt
Sample Name: /rans-nonochlor
Unit: (ng/L)
Survey Date:
4/24/94 - 10/13/95
Sample Depths:
OM - 300M
41 I I I I I I I I I I I I I I I I I I I I I I I I I I I
-89
-88
-87
-86
-85
-84
April 2004
5-23
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 5-12. Concentrations of Dissolved Total PCBs Measured in Open-lake Samples
••
•
.; :
:
•:
I I l
Lake Michigan Mass Balance Project
N
Sample Type: DISSOLVED
Sample Nome: PCS lotpcb
Unit: (H9/L)
'
Survey Date:
4/24/9* - 10/13/95
Sample Depths:
OM - 300M
41 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
-89
-58
-87
-86
-85
-84
5-24
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Figure 5-13. Concentrations of Participate Total PCBs Measured in Open-lake Samples
:•
•
:
•:
I I l
Lake Michigan Mass Balance Project
Sompte Type: PARTICULATE
Sample Name: PCS totpcb
Unit: (ng/L)
IN
t
Survey Date:
4/24/94 - 10/13/95
Sample Depths:
OM - 300M
41 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
-89 -58 -87 -86 -85
-84
April 2004
5-25
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
5.2 Quality Implementation and Assessment
As described in Section 1.5.5, the LMMB QA program prescribed minimum standards to which all
organizations collecting data were required to adhere. The quality activities implemented for the PCBs
and fra«5-nonachlor monitoring portion of the study are further described in Section 2.7 and included use
of SOPs, training of laboratory and field personnel, and establishment of MQOs for study data. A
detailed description of the LMMB quality assurance program is provided in the Lake Michigan Mass
Balance Study Quality Assurance Report (USEPA, 200 Ib). A brief summary of data quality issues for
the open lake PCBs and fra«s-nonachlor data is provided below.
Quality Assurance Project Plans (QAPPs) were developed by the Pis and were reviewed and approved by
GLNPO. Each researcher trained field personnel in sample collection SOPs prior to the start of the field
season and analytical personnel in analytical SOPs prior to sample analysis. Each researcher submitted
test electronic data files containing field and analytical data according to the LMMB data reporting
standard prior to study data submittal. GLNPO reviewed these test data sets for compliance with the data
reporting standard and provided technical assistance to the researchers. In addition, each researcher's
laboratory was audited during an on-site visit at least once during the time LMMB samples were being
analyzed. The auditors reported positive assessments and did not identify issues that adversely affected
the quality of the data.
As discussed in Section 2.5, because data comparability was important to the successful development of
the mass balance model, the Pis used similar sample collection, extraction, and analysis methods for the
PCB and trans-nonachlor monitoring in this study.
5.2.1 Sample Collection
During examination of the field collection records for field duplicates, it was discovered that some field
duplicates were not actually collected at the same time as the field sample due to equipment mobilization.
Samples collected within five minutes of each other were considered field duplicates (FD1), and if more
than five minutes elapsed, the samples were considered sequential field duplicates (SFD1). Separate
labeling of these data points as FD1 and SFD1 was done in order to assess if precision differed based on
the elapsed time.
5.2.2 Data Assessments
As discussed in Section 2.7, data verification was performed by comparing all field and QC sample
results produced by each PI with their MQOs and with overall LMMB Study objectives. Analytical
results were flagged when pertinent QC sample results did not meet acceptance criteria as defined by the
MQOs. These flags were not intended to suggest that data were not useable; rather they were intended to
caution the user about an aspect of the data that did not meet the predefined criteria. Table 5-13 provides
a summary of flags applied to the open lake PCB and fra«s-nonachlor data. The summary includes the
flags that directly relate to evaluation of the MQOs to illustrate some aspects of data quality, but does not
include all flags applied to the data to document sampling and analytical information, as discussed in
Section 2.7.
Pis used surrogate spikes to monitor the bias of the analytical procedure. The PCB results were corrected
for the recoveries of the surrogates. The trans-nonachlor results were not surrogate-corrected. Only 1%
of each of the open-lake particulate results for PCBs 33, 118, and 180, and fra«s-nonachlor were qualified
because of surrogate recovery problems (Table 5-13). For the dissolved PCB samples from the open lake,
3 to 5% of the results for PCBs 33, 118, and 180 were qualified for surrogate recovery problems (Tables
5-13), while 19% of the dissolved fra«s-nonachlor results were qualified.
5-26 April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Laboratory matrix spike samples also were used to monitor the bias of the analytical procedure. The
results for the matrix spike samples were compared to the MQO for spike recoveries (50 - 125%).
Analytical results associated with matrix spike samples with recoveries below the MQO limits were
flagged with failed matrix spike and low bias flags, and results associated with matrix spike samples that
had recoveries higher than the MQO limits were flagged with failed matrix spike and high bias flags.
Analytical results were considered invalid and flagged as such when the analyte was undetected and
recoveries for associated matrix spike samples were less than 10%. None of the open-lake particulate
tmns-nonachlor results or PCB 33, 118, or 180 results failed the matrix spike MQOs. However, 8% of
the open-lake dissolved PCB 33 results, 14% of the open-lake dissolved PCB 118 results, and 71% of the
open-lake dissolved tmns-nonachlor results were flagged as failing the matrix spike MQOs. A maximum
of 1% of the samples were flagged as invalid.
Field blanks were collected for PCBs and trans-nonachlor. Field blanks were to be collected at a
frequency of 5%. Due to the limited availability of samplers and resin, the actual frequency was only
3.5% When field blank contamination was greater than 3.3 times the method detection limit, all of the
associated results were flagged with the failed field blank sample code (FFR). Field blanks were not
collected at all stations, so potential station-specific contamination associated with these sites cannot be
evaluated. However, contamination associated with sampling equipment, collection, processing,
shipping, storing, and analysis can be evaluated based on the field blanks collected throughout the study.
Large percentages of samples were associated with field blanks in which PCBs 33, 118, or 180, or trans-
nonachlor were reported above the sample-specific detection limit (Table 5-13). This issue is discussed in
greater detail in Section 5.2.3.
Trip blanks were collected for PCBs and fra«s-nonachlor. Trip blanks were to be collected at a frequency
of 5%. Due to the limited availability of samplers and resin, the actual frequency was only 2.2%. As
with the field blanks, large percentages of samples were associated with a trip blank in which PCBs 33,
118, or 180, or trans-nonachlor were reported above the sample-specific detection limit (Table 5-13).
This issue is discussed in greater detail in Section 5.2.3.
Field duplicates were to be collected at a frequency of 5%. Duplicate samples collected within 5 minutes
of each other were considered field duplicates. However, an examination of the field collection records
indicated that some of the planned field duplicates were not collected within that 5-minute time frame as a
result of problems with equipment mobilization or the time required to pump the sample through the filter
and resin cartridge. Those "duplicates" that were collected more than 5 minutes apart were considered
"sequential field duplicates" and the data were labeled accordingly (e.g., SDF1 vs. FD1). Combining the
field duplicates and sequential field duplicates, the actual rate of collection of duplicates was 7.6%.
The results from the original field sample and the associated duplicate were compared on the basis of the
relative percent difference (RPD). The RPD value for each PCB congener and trans-nonachlor was
compared to the MQO for field duplicate precision. None of the particulate PCB results were qualified
because of the field duplicate precision (FFD) concerns (Table 5-13). Only 2% of the particulate trans-
nonachlor results were so qualified. The percentage of dissolved PCB and fra«s-nonachlor results that
were qualified because of field duplicate precision concerns ranged from 0.3% to 2%.
As discussed in Section 1.5.5, MQOs were defined in terms of six attributes: sensitivity, precision,
accuracy, representativeness, completeness, and comparability. GLNPO derived data quality assessments
based on a subset of these attributes. For example, system precision was estimated as the mean relative
percent difference (RPD) between the results for field duplicate pairs. Similarly, analytical precision was
estimated as the mean relative percent difference (RPD) between the results for laboratory duplicate pairs.
Table 5-14 provides a summary of data quality assessments for several of these attributes for the open-
lake PCB and trans-nonachlor data.
April 2004 5-27
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Because the relative variability of most measurement techniques increases as one approaches the
detection limit of the technique, the assessment of the field duplicate results were divided into two
concentration regimes. One measure of system precision was calculated for those field duplicate results
that were less than 5 times the sample-specific detection limit (SSDL) of the analyte, and a separate
measure was calculated for those field duplicate results that were greater than 5 times the SSDL. None of
the open lake particulate sample field duplicate pairs contained fra«s-nonachlor concentrations above 5
times the SSDL.
The precision of the particulate field duplicate results above 5 times the SSDL ranged from approximately
5 to 11% for the PCB congeners (Table 5-14), while the precision of the particulate field duplicate results
below 5 times the SSDL ranged from approximately 23% to 55%.
None of the field duplicate pairs for dissolved samples contained PCB 180 above 5 times the SSDL, and
the precision of the dissolved field duplicate results for PCBs 33 and 118, and fra«s-nonachlor above the
SSDL ranged from approximately 17% to 36% (Table 5-14).
Analytical bias was assessed using the results from matrix spike samples. The mean recoveries were very
good for the particulate-phase PCBs and frara-nonachlor, ranging from 88.8% to 105% for the analytes in
Table 5-14. These results demonstrate that the analytical techniques applied to the field samples
introduced little or no bias into most of the results, and only a slight low bias was introduced into the
particulate PCB 118 results.
The matrix spike recoveries of the dissolved analytes were considerably more varied than the particulate
results. The recoveries of dissolved PCBs 33 and 180 were very good to excellent, at 80.9% and 109%,
respectively. However, the dissolved PCB 118 results indicate a significant high bias, with a mean
recovery of 157%, while the dissolved fra«s-nonachlor recoveries average only 33.7%, indicating a
significant low bias.
Analytical sensitivity was assessed on the basis of the percentage of study samples that were reported
with concentrations below the sample-specific detection limit (SSDL). The sensitivity varied by congener
for the PCBs, partly as a function of the analytical instrumentation and its response to the individual
congeners.
The three PCB congeners and fra«s-nonachlor were not detected in substantial portions (1 - 55%) of the
dissolved and particulate samples from the open lake ("UND" flag in Table 5-13). These analytes were
detected below the sample-specific detection limits in substantial portions (1 - 23%) of the samples as
well ("MDL" flag in Table 5-13). For the three congeners listed in Table 5-13, the percentage of the
dissolved samples with results reported below the sample-specific detection limits increases (i.e., 10, 18,
and 23%) with the congener number (e.g., with molecular weight), suggesting that solubility may play a
role in the distribution.
The percentages of fra«s-nonachlor results that were not detected or detected below the sample-specific
detection limits (Table 5-13) generally fell between the same percentages for the three PCB congeners,
and were most similar to the percentages for PCB 118 (e.g., 6 - 12% for trans-nonachlor and 5 - 18% for
PCB 118).
5-28 April 2004
-------
Table 5-13. Summary of Routine Field Sample Flags Applied to Select PCB Congeners and frans-Nonachlor in Open-lake Samples
Analyte
PCB 33 -
348 Dissolved
349 Particulate
PCB 118-
348 Dissolved
349 Particulate
PCB 180-
348 Dissolved
349 Particulate
frans-Nonachlor -
343 Dissolved
349 Particulate
Fraction
Dissolved
Particulate
Dissolved
Particulate
Dissolved
Particulate
Dissolved
Particulate
Flags
Sensitivity
MDL
0
1%(5)
18% (63)
8% (28)
23% (79)
3% (12)
8% (29)
6% (22)
UNO
1%(2)
24% (85)
5% (19)
5% (18)
55% (190)
39% (135)
12% (42)
6% (20)
Contamination
FFR
85% (297)
29% (101)
57% (200)
30% (104)
53% (183)
20% (71)
35% (11 9)
21% (73)
FFT
59% (206)
41% (142)
77% (269)
48% (169)
38% (131)
14% (48)
56% (193)
48% (169)
Precision
FFD
1%(2)
0
0.3% (1)
0
1%(3)
0
2% (6)
2% (6)
Bias
FSS
5% (16)
1%(4)
3% (10)
1%(4)
3% (10)
1%(4)
19% (66)
1%(2)
FMS
8% (29)
0
14% (48)
0
0
0
71% (244)
0
LOB
0
0
0
0
0
0
5% (16)
0
HIB
68% (238)
19% (65)
57% (197)
25% (86)
33% (11 5)
9% (33)
11% (37)
7% (26)
INV
1%(4)
0.3% (1)
1%(4)
0.3% (1)
1%(4)
0.3% (1)
0
0
The number of routine field samples flagged is provided in parentheses. The summary provides only a subset of applied flags and does not represent the full suite of flags applied
to the data.
MDL = Less than method detection limit (Analyte produced an instrument response but reported value is below the calculated method detection limit. Validity of reported value
may be compromised.)
UNO = Analyte not detected (Analyte produced no instrument response above noise.)
FFR = Failed field blank (A field blank sample, type unknown, associated with this analysis failed the acceptance criteria. It is unknown whether the blank that failed was a
field blank or a lab blank. Validity of reported value may be compromised.)
FFT = A trip blank associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.
FFD = Failed field duplicate (A field duplicate associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FSS = Failed surrogate (Surrogate recoveries associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FMS = Failed matrix spike (A matrix spike associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
LOB = Likely biased low (Reported value is probably biased low as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery or other internal
lab QC data. Reported value is not considered invalid.)
HIB = Likely biased high (Reported value is probably biased high as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery, blank
contamination, or other internal lab QC data. Reported value is not considered invalid.)
INV = Invalid
5-29
-------
Table 5-14. Data Quality Assessment for Select PCB Congeners and frans-Nonachlor in Open-lake Water Samples
Analyte/Number
Field Samples
PCB 33 -
344 Dissolved
348 Particulate
PCB 118-
344 Dissolved
348 Particulate
PCB 180-
344 Dissolved
348 Particulate
frans-Nonachlor -
343 Dissolved
349 Particulate
Parameter
System Precision - Mean Field Duplicate RPD (%), < 5 * SSDL
System Precision - Mean Field Duplicate RPD (%), > 5 * SSDL
Analytical Bias - Mean Laboratory Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < SSDL (%)
System Precision - Mean Field Duplicate RPD (%), < 5 * SSDL
System Precision - Mean Field Duplicate RPD (%), > 5 * SSDL
Analytical Bias - Mean Laboratory Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < SSDL (%)
System Precision - Mean Field Duplicate RPD (%), < 5 * SSDL
System Precision - Mean Field Duplicate RPD (%), > 5 * SSDL
Analytical Bias - Mean Laboratory Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < SSDL (%)
System Precision - Mean Field Duplicate RPD (%), < 5 * SSDL
System Precision - Mean Field Duplicate RPD (%), > 5 * SSDL
Analytical Bias - Mean Laboratory Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < SSDL (%)
Number of QC samples
Dissolved
8 field duplicate pairs
19 field duplicate pairs
24 matrix spikes
-
21 field duplicate pairs
4 field duplicate pairs
24 matrix spikes
-
10 field duplicate pairs
0 field duplicate pairs
24 matrix spikes
-
23 field duplicate pairs
1 field duplicate pair
24 matrix spikes
-
Particulate
7 field duplicate pairs
1 1 field duplicate pairs
22 matrix spikes
-
8 field duplicate pairs
13 field duplicate pairs
22 matrix spikes
-
6 field duplicate pairs
12 field duplicate pairs
22 matrix spikes
-
22 field duplicate pairs
0 field duplicate pairs
22 matrix spikes
-
Assessment
Dissolved
55.9%
17.1%
80.9%
0.3%
30.4%
35.6%
157%
23.0%
77.4%
-
109%
77.3%
41.0%
17.8%
33.7%
20.7%
Particulate
23.6%
11.3%
97.6%
25.9%
18.2%
4.96%
88.8%
13.2%
54.9%
8.98%
105%
42.2%
36.7%
-
95.1%
12.0%
5-30
-------
PCBs/trans-Nonachlor in Open-lake Water
As noted in Section 2.6.4, the laboratory did not obtain separate cleanup fractions containing the PCBs
and fra«s-nonachlor, but analyzed the sample extracts on two dissimilar GC columns (DB-5 and DB-
1701). While the DB-1701 column provided clear chromatographic separation of any fra«s-nonachlor in
the sample, this analyte coeluted with PCB 99 on the DB-5 column. As a result of the potential coelution,
the reported concentrations of PCB 99 in open-lake samples are probably biased by any fra«s-nonachlor
present in the samples.
5.2.3 Evaluation of Blanks
Because PCBs are a ubiquitous contaminant, both in the environment and in environmental testing
laboratories, the LMMB Study design included a wide range of types of blanks that were designed to
identify many of the potential sources of PCB contamination that might be encountered during the study.
Contamination of the samples from other sources was a particular concern because the study attempted to
investigate the very low concentrations present in the open lake. When the data were examined, a large
number of open-lake PCB sample results (20 - 90%) in the LMMB Study were flagged as being
associated with one or more blanks that exhibited signs of contamination.
The data presented in this report thus far include all of the sample results except those flagged as invalid.
Samples that were flagged with blank contamination were included in the analyses, and as a result, the
estimates of mean concentrations may be biased due to contributions from the various blanks. An
evaluation of the blank contaminants was conducted to examine the impacts of these contaminants on the
results and conclusions by comparing several alternative approach to flagging and treating sample results.
The reported concentrations of PCBs in the open lake were evaluated with regard to the results of the
three types of routine blanks that were prepared for the study. Blanks are important to consider when
estimating concentrations of PCB congeners in this study for several reasons including:
• Blank contamination is typical for PCB sampling and analysis, especially for low concentrations of
PCB congeners that are close to the detection limit of the analytical method,
• Blank contamination affected a significant number of field samples results collected in the LMMB
Study, and
• The analytical laboratory changed its resin cleaning procedures in the middle of the study to comply
with a revised criteria for "clean" resin set by GLNPO.
Mean concentrations of open lake PCB congeners were calculated in two ways: using all the LMMB
data, and using only those data that were not affected by contamination of the field reagent blank, the
laboratory dry blank, or the laboratory reagent blank. For the purposes of this evaluation, a field sample
result was considered unaffected by blank contamination if the results of all of the associated blank
samples were less than 1/3 of the concentration reported in the field sample.
The criteria for the evaluation were based on the measurement quality objectives (MQOs) established by
the principal investigators for open lake PCBs, when possible, because the Pis are most knowledgeable
about the performance of their sampling and analytical method.
Table 5-15 provides criteria used to evaluate whether to include a sample result in the estimation of the
mean concentrations in the open lake. The approach taken in this evaluation was conservative, in that it
was designed to leave as many samples as possible in the estimation of the mean. This evaluation did not
consider the effects of the field trip blanks on the sample results for several reasons. First, there were
fewer field trip blanks than the other types of blanks. Therefore, there is concern that the results for a
given field trip blank may not be as representative of the actual sample collection procedures as other
types of blanks. Secondly, the potential contamination illustrated by the field trip blank also could be
April 2004 5-31
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
evaluated using the field reagent blank results, because the field reagent blank should theoretically
capture most of the same sources of contamination.
In addition to the blank considerations listed in Table 5-15, data also were excluded if they were flagged
"Invalid" in the database. The invalid flag indicates that the PI and the QC coordinator deemed the data
to be unusable for any purpose.
All sample results were included as reported by the PI in the estimate of the mean. If the PI reported a
result as zero, then the zero was included in the estimate. A zero result should be interpreted as a
concentration that is below the sample-specific detection limit for that sample. In addition, the results that
were reported as a value below the sample-specific detection limit also were included in the estimate.
These results were flagged in the database with the "MDL" flag and should be interpreted as a
concentration that is below the sample-specific detection limit.
Table 5-15. Criteria Used to Evaluate Data to be Included in the Estimation of the Mean Concentrations of
PCB Congeners
Quality Control Consideration
Blanks, including:
• field reagent blanks,
• lab dry blanks1, and
• lab reagent blanks.1
Criteria
Exclude the sample result
when any of the three
associated blanks has a
result that is greater than
1/3 of the concentration in
the sample
Rationale
When a sample is associated with a blank that has greater
than 1/3 of the concentration in the sample result, the
result is likely to be biased high and contamination may be
a significant portion of the concentration reported in the
sample. The multiplier of 1/3 is based on the MQOs
established by the PI for several blank types.
1Lab dry blanks and lab reagent blanks were reported in mass units because there is no actual volume of "sample" pumped through the filter
and resin column. The sample results were compared to the results for these two types of blanks by converting the sample results to mass
units as well.
The estimation of the mean concentration of each PCB in the open lake was complicated by the use of
detection limits that are specific to each sample, rather than using one detection limit for each congener
across all lake samples. The sample-specific detection limits take into account the actual volume of lake
water pumped through the filter and resin column, which may differ between samples.
There are several approaches that may be used to estimate the mean concentration of each PCB congener
in the open-lake samples. One common approach is to substitute the sample-specific detection limit for
any result below that limit and use the result as reported for any result above the sample-specific detection
limit. However, that approach introduces a high bias into the mean concentration because no result used
in the mean will ever be less than the detection limit for that sample. Another common approach is to use
the concentrations as reported by the investigator. This approach recognizes that the actual
concentrations in the samples may range from zero to the sample-specific detection limit. The modelers
using the LMMB data are using the results as reported, including the values reported by the Pis as zero or
below the sample-specific detection limit. Therefore, this same approach was used to estimate the mean
concentrations for this evaluation.
The mean concentrations for the three PCB congeners are presented in Table 5-16 in four ways, using:
1. All data except those flagged invalid in the database,
2. Only data without associated blank failures as described in Table 5-15,
3. Only data without associated blank failures as described in Table 5-15 for 1994, and
4. Only data without associated blank failures as described in Table 5-15 for 1995.
5-32 April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
The differentiation between the 1994 and 1995 data was made because GLNPO lowered the acceptable
level of PCBs that could be in the XAD-2® resin and particulate collection filters at the beginning of 1995
and the Pis responded by changing their cleaning procedures. The standard deviation, the concentration
range, and the mean sample-specific detection limit are presented in Table 5-16 for the three congeners.
Table 5-16. Comparison of Summary Statistics for LMMB Open-lake PCB Congener Results after Removal of
Sample Results associated with Contaminated Blanks
Analyte
PCB 33
PCB 118
PCB 180
Fraction
Dissolved
Particulate
Dissolved
Particulate
Dissolved
Particulate
Data Included in Mean1
All data
Data Without Blank Failures
1994 Data Without Blank Failures
1995 Data Without Blank Failures
All data
Data Without Blank Failures
1994 Data Without Blank Failures
1995 Data Without Blank Failures
All data
Data Without Blank Failures
1994 Data Without Blank Failures
1995 Data Without Blank Failures
All data
Data Without Blank Failures
1994 Data Without Blank Failures
1995 Data Without Blank Failures
All data
Data Without Blank Failures
1994 Data Without Blank Failures
1995 Data Without Blank Failures
All data
Data Without Blank Failures
1994 Data Without Blank Failures
1995 Data Without Blank Failures
N
303
128
64
64
306
216
112
104
303
213
132
81
306
277
161
116
303
189
104
85
306
289
172
117
Mean (pg/L)
9.15
8.70
5.92
11.47
0.77
0.78
1.13
0.41
2.46
3.04
3.76
1.87
2.24
2.44
2.49
2.36
0.49
0.32
0.02
0.70
0.94
0.97
1.08
0.81
SD
(pg/L)
23.16
22.27
1.60
31.33
1.54
1.79
2.36
0.63
3.01
3.30
3.32
2.94
2.88
2.96
3.26
2.48
2.03
2.45
0.06
3.62
1.40
1.44
1.61
1.13
Range (pg/L)
0.00 - 205.62
0.00-193.50
0.00-11.78
2.37-193.50
0.00-18.01
0.00-18.01
0.00-18.01
0.00-3.89
0.00-18.33
0.00-18.33
0.00-15.53
0.00-18.33
0.00 - 26.26
0.00 - 26.26
0.00-26.26
0.00-15.23
0.00 - 29.53
0.00 - 29.53
0.00-0.31
0.00 - 29.53
0.00-11.72
0.00-11.72
0.00-11.72
0.00-6.15
Mean SSDL
(pg/L)
0.578
0.591
0.637
0.545
0.108
0.106
0.127
0.084
0.788
0.778
0.798
0.746
0.252
0.253
0.295
0.196
0.388
0.385
0.392
0.376
0.155
0.157
0.181
0.122
Figures 5-14 to 5-16 provide a graphical display of the means presented in Table 5-16, along with the
standard error of each mean (i.e., the standard deviation divided by the square root of the number of
observations in each mean), and the mean sample-specific detection limit (SSDL) for the congener.
The four bars in each graph above represent the mean concentrations of the analyte derived from: all
data, all data without the samples associated with contaminated blanks, all data from 1994 without
samples associated with contaminated blanks, and all data from 1995 without samples associated with
contaminated blanks. The y-axis units are the mean concentration of the PCB congener in ng/L. The
narrow vertical lines represent ± 1 standard error around the mean concentration. The horizontal line
across each graph is the approximate position of the mean sample-specific detection limit (SSDL) for the
results from all of the samples for that congener and phase. Note that the vertical scale use for the y-axis
differs from congener to congener, based on the observed range of mean concentrations.
April 2004
5-33
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
The results for dissolved PCB 33 and PCB 180 illustrate the potential for high bias in the mean
concentrations derived from all of the data. For both of these congeners, the mean dissolved
concentration decreases slightly when the data associated with the contaminated blanks are removed (e.g.,
the second bar in each graph labeled "data without QC failures"). In contrast, the results for dissolved
PCB 118 and the particulate fraction for all three congeners show an increase in the mean concentration
when the data associated with the contaminated blanks are removed.
The effect of the change in the acceptance criteria for the XAD-2® resin between 1994 and 1995 is not
consistent across the dissolved congeners. For PCB 118, the 1995 mean concentration of samples without
QC failures is about half of the 1994 mean concentration. However, for dissolved PCB 33, the trend is
exactly opposite, with the 1995 mean concentration approximately twice the 1994 mean concentration.
For dissolved PCB 180, the 1995 mean is actually 35 times higher than the 1994 mean.
For the particulate sample results, the mean concentrations were lower in 1995 for PCBs 33, 118, and
180. This suggests that the contribution of the blanks to the particulate results may have been less in the
1995 data then in the 1994 data.
The data for the mean SSDL for each congener and fraction also illustrate a significant aspect of the
situation. For dissolved PCB 33, the mean sample results, regardless of QC failures, are 7 to 10 times
higher than the mean SSDL value. For dissolved PCB 118, the ratio drops to 2 to 4 times the SSDL, and
the error bars for the mean dissolved PCB 180 encompass the mean SSDL for three of the four bar
graphs. For the particulate PCB results, the mean concentrations of all three congeners are at least five
times higher than the mean SSDL values. These results illustrate the congener-specific difficulties in
measuring open-lake concentrations that are near or below the capabilities of the analytical techniques.
5-34 April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Figure 5-14. Summary Statistics for LMMB Open-lake PCB 33 Results after Removal of Sample Results
associated with Contaminated Blanks
Dissolved PCB 33
17.00 -
16.00 -
15.00 -
14.00 -
13.00 -
12.00 -
11.00 -
10.00 -
9.00 -
8.00 -
7.00 -
6.00 -
5.00 -
4.00 -
3.00 -
2.00 -
1.00 -
n nn -
Mean SSDL
n =
n = 64
303 n = 1 28
n = 64
All Data
Data Without QC Failures 1994 Data Without QC Failures 1995 Data Without QC Failures
Participate PCB 33
1.40 -
1.20 -
1.00 -
0.80 -
0.60 -
0.40 -
0.20 -
0.00
.. Mean SSDL
n = 306
n = 112
n = 216
n = 104
All Data
Data Without QC Failures 1994 Data Without QC Failures 1995 Data Without QC Failures
April 2004
5-35
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 5-15. Summary Statistics for LMMB Open-lake PCB 118 Results after Removal of Sample Results
associated with Contaminated Blanks
Dissolved PCB 118
4.DU -
4.00 -
3.50 -
3.00 -
2.50 -
2.00 -
1.50 -
1.00 -
0.50 -
nnn -
Mean SSDL n = 132
n = 213
n = 303
T
1
T
1
n =
81
All Data Data Without QC Failures 1994 Data Without QC Failures 1995 Data Without QC Failures
Particu late PCB 118
3.UU -
2.50 -
2.00 -
|, 1.50 -
1.00 -
0.50 -
nnn -
Mean SSDL n = 277 n -
n = 306
T
1
1
1
161
n =
116
All Data Data Without QC Failures 1994 Data Without QC Failures 1995 Data Without QC Failures
5-36
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Figure 5-16. Summary Statistics for LMMB Open-lake PCB 180 Results after Removal of Sample Results
associated with Contaminated Blanks
Dissolved PCB 180
1.40 -
1.20 -
1.00 -
0.80 -
0.60 -
0.40 -
0.20 -
0.00
...... Mean SSDL
S03
189
n = 104
85
All Data
Data Without QC Failures 1994 Data Without QC Failures 1995 Data Without QC Failures
Participate PCB 180
1.40 -
1.20 -
1.00 -
0.80 -
0.60 -
0.40 -
0.20 -
nnn -
Mean SSDL
n =
n = 306 n = 289
T 1
1
1
172
n - 117
All Data
Data Without QC Failures 1994 Data Without QC Failures 1995 Data Without QC Failures
April 2004
5-37
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
5.3 Data Interpretation
The LMMB Study resulted in one of the largest collections of PCB and fra«s-nonachlor data ever
produced for any of the Great Lakes. The data for PCBs and fra«s-nonachlor from the LMMB Study
indicate that the concentrations of individual PCB congeners and fra«s-nonachlor vary among the stations
in the open lake, In addition, the concentrations of the congeners differ, both within a station, as well as
across the stations.
5.3.1 Comparison to Historical Studies
The magnitude of the LMMB data set makes it difficult to find comparable historical results that are
useful for comparisons. Even where data are available from other investigators for the same analytes, the
potential differences in spatial and temporal coverage present concerns that are further complicated by the
likely differences in the sampling and analytical procedures.
Three historical data sets have been identified by EPA's Large Lakes Research Station as relevant to such
comparisons. Swackhamer and Armstrong (1987) collected PCB data during September 1980 under a
project partially supported by USEPA Cooperative Agreement CR807836. Filkins et al. (1983), at the
Cranbrook Institute, collected PCB data during September 1981 under a project supported by USEPA
Cooperative Agreement CR810232. Pearson et al. (1996) collected PCB data during September 1991
under a project supported by GLNPO Grant No. GL995233.
These historical data represent the "total PCB" concentration in each sample, without regard for the
dissolved or particulate fraction, and without distinguishing among the PCB congeners. Therefore, the
historical results can only be compared to the sum of the dissolved and particulate PCB results from the
LMMB Study. Figures 5-17 to 5-20 present the results from the three historical studies and the LMMB
results for samples collected in September 1995. Figure 5-20 presents the summary plot for total PCBs
for the entire LMMB data set. As with the earlier contour plots, note that the concentrations scales differ
among the four plots. The data from the three historical studies represent samples collected in the top 30
meters of the lake, while the LMMB data include samples collected at greater depths as well. Additional
comparisons could be made using only the LMMB results for samples collected in the top 30 meters of
the lake. However, such plots were not available at the time of this report.
Swackhamer and Armstrong (1987) collected 45-L water samples from 19 stations in the open lake. The
water samples were filtered through glass-fiber filters on board the R/V Roger R. Simons. The filtrates
were passed through glass columns containing XAD-2® resin. The filter and the XAD-2® resin were
extracted separately, and the extracts were analyzed by GC/ECD. Total PCB concentrations were
determined by comparison to standards of Aroclor mixtures, as well as through the use of standards for
some individual PCB congeners.
The data from September 1980 (Swackhamer and Armstrong, 1987, Figure 5-17) ranged from 0.4 to 7.9
ng/L, with a mean concentration in the open lake of 1.8 ng/L. The total PCB concentration in central
portion of Lake Michigan is between 1.2 and 1.6 ng/L, with lower concentrations (< 0.8 ng/L) in the
extreme northern portion of the lake, and with concentrations as high as 6 to 8 ng/L near the shore in the
lower portion, close to major urban areas.
Filkins et al, (1983) collected approximately 120 L of water from a depth of 4 m at four stations in Lake
Michigan, as part of a larger study involving 21 stations in all five Great Lakes. The water samples were
collected as 3-L aliquots placed in multiple 1-gallon glass bottles. Methylene chloride was added to each
bottle, and the bottles were shaken for 3 minutes. After standing for two hours to allow the solvent and
water sample to separate, the methylene chloride was removed from each bottle and stored. The extracts
5-38 April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
from the individual bottles were concentrated and combined into a single final extract that was analyzed
by GC/ECD. Total PCB concentrations were determined by comparison to specific PCB congeners
identified in a mixed Aroclor standard. Filkins et al. (1983) provide data on the percentage of the sample
result represented by the method blanks associated with each sample. For the four Lake Michigan
samples, the associated blanks represent 14 to 122% of the associated sample result.
The data from September 1981 (Filkins, et al., 1983, Figure 5-18) represent the results from only four
samples in all of Lake Michigan. The total PCB concentrations in this study ranged from about 0.25 ng/L
in the extreme northern portion and the central portion of the lake, to about 0.31 ng/L at a station near
Chicago, with a similar concentration found at the mouth of Green Bay.
Pearson et al. (1996) collected approximately 100 L of water at each of 11 stations in Lake Michigan
abroad the R/VLake Guardian. Samples were filtered through glass fiber filters on board the ship. The
filtrate was collected in 70-L stainless steel tanks and returned to the laboratory were it was extracted
using a continuous flow liquid-liquid Goulden large-volume extraction device. Particulate and dissolved
extracts were analyzed separately, using GC/ECD. The total PCB concentrations were calculated by
summing the results determined for individual congeners using an internal standard method and surrogate
correction.
The data from September 1991 (Pearson, et al., 1996, Figure 5-19) includes significantly more stations
than in 1981. Total PCB concentrations range from 0.34 to 1.7 ng/L, with a lake-wide mean
concentration of 0.64 ng/L. Because the congener distribution patterns at two of the stations differed
from those at the other nine stations, Pearson et al. also calculated a mean concentration of 0.47 ng/L for
the nine stations alone. The northern portions of the lake contain approximately 0.5 ng/L of total PCBs,
with concentrations of about 1 ng/L in the southern portion of the lake. The hot spot apparent in the
southern portion of the lake has a maximum concentration of 1.7 ng/L. Overall, the total PCB
concentrations are lower than in 1980 throughout the lake.
The LMMB data from September 1995 (Figure 5-20) suggest further decreases in the total PCB
concentrations lake-wide since 1991. These data suggest that the concentrations in the northern portion
of the lake are less than 0.15 ng/L. A hot spot appears in the southern portion of the lake with a
maximum concentration of about 0.7 ng/L, roughly half the concentration is a similar hotspot found in
1991. Thus, the September 1995 total PCB data suggest a drop in concentration of about 50% from the
1991 results, and a drop of almost an order of magnitude from the 1980 results. In addition, the data in
Figure 5-21 for the entire LMMB Study suggest that the results from September 1995 (Figure 5-20) are
not unusual for the period of the LMMB Study itself.
Historical data for fra«s-nonachlor were not available, so no inferences can be made regarding changes in
the concentrations of this contaminant in Lake Michigan.
April 2004 5-39
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 5-17. Concentrations of Total PCBs Measured in Open-lake Samples in September 1980
••••
••:
: :
:
42
Lake Michigan Mass Balance Project
Sample Type; TOTAL
Sample Name: Total PC8
Unit: (ng/L)
IN
t
Survey Dote:
9/1/80 - 9/30/80
Sample Depths:
OM - 30M
41 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
-89 -87 -86 -84
5-40
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Figure 5-18. Concentrations of Total PCBs Measured in Open-lake Samples in September 1981
••••
•
: :
:
-
Lake Michigan Mass Balance Project
N
Sample Type: TOTAL
Sample Name: Total PCB
Unit: (f>9/L)
'
,
9/6/81 - 9/6/81
Sample Depths:
OM - 30M
41 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
-89 -88 -87 -86
April 2004
5-41
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 5-19. Concentrations of Total PCBs Measured in Open-lake Samples in September 1991
••••
:
: :
:
,:
Lake Michigan Mass Balance Project
N
Sample Type: TOTAL
Sample Name: Total PC8
Unit: (ng/L)
•
Survey Dote:
9/6/91 - 9/11/91
Sample Depths:
OM - 30M
41 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
-89 -88 -87 -86 -84
5-42
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Figure 5-20 Concentrations of Total PCBs Measured in LMMB Open-lake Samples in September
1995
:
46
;
:
-
Loke Michigan Mass Balance Project
IN
t
Sompl« Type: DISS+PART
Sample Name: Total PCS
Unit; (*g/L)
3.56505
1.36434
1.21275
1.06115
0.90956
0.75797
0.60637
0.45478
0.30319
0.15159
0.00000
Survey Dote:
9/16/95 - 10/13/95
Sample Depths:
OM - 3COM
41 I I i i i I I I I I I I I I I I I I ll I I I I I I I I ' I I I I I I
-89
-
-87
-86
-85
-84
April 2004
5-43
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 5-21. Concentrations of Total PCBs Measured in All LMMB Open-lake Samples
:
'
-
: :
•
42
Lake Michigan Mass Balance Project
N
'
Sample Type: OISS+PART
Sample Name: PCB lotpcb
Unit: (ng/L)
1.68411
1.52913
1.37414
1.21916
1.06417
0.90919
0.75420
0.59922
0.44423
0.28924
0.13426
Survey Dote:
4/24/94 - 10/13/95
Sample Depths:
OM - 300M
41 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 1 I I I I I
-89 -88 -86 -84
5-44
April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
5.3.2 Regional Considerations
Among the general trends evident in the contour plots for the particulate-phase results for PCBs 33, 118,
and 180 (Figures 5-7 to 5-9, shown earlier) are the presence of "hot spots" in the upper reaches of Green
Bay (on the western side of Lake Michigan). The apparent decrease in concentrations to the southwest of
the hot spot, e.g., towards the head of Green Bay, is an artifact of the lack of a sampling station further up
Green Bay (see Figure 5-1 for the locations of the two sampling stations in Green Bay). The lower
reaches of the Fox River are a known source of sediments containing high levels of PCBs (see Chapter 4),
and these hot spots in particulate PCB concentrations may be the result of resuspended sediments from
the Fox River that are carried into Green Bay.
The PCB and trans-nonachlor results were examined to determine if there were any statistically
significant differences between the northern and southern portions of Lake Michigan. For these
comparisons, the data from the LMMB Study were divided at approximately 44° north latitude (see
Figure 5-17 for the latitude). The dividing line at 44° N is not intended as a formal differentiation
between hydrographic basins in the lake, and other means of differentiating the results from north to south
could be considered. Rather, the line at 44° N yields approximately equal numbers of stations in each
portion of the lake. The results from the two stations in Green Bay and the one station in Lake Huron
were excluded from these comparisons. The stations in the lower portion of the lake include Stations 1
through 29, plus 310, 340, and 380. Stations 31 through 180 were in the northern portion of the lake.
The results of these comparisons for PCB congeners 33, 118, and 180, total PCBs and fra«s-nonachlor are
shown in Table 5-17 for both the dissolved and particulate samples. There are statistical interactions
between the effects of the cruise and the north/south division for some analytes, in which case, a
comparison between the northern and southern stations cannot be made.
Table 5-17. Results of North/South Comparisons of Open-lake Concentrations of PCBs and frans-Nonachlor
Fraction/Analyte
Significant Difference Between North and South?
Probability
Direction
Dissolved
PCB 33
PCB 118
PCB 180
Total PCBs
frans-Nonachlor
Yes
< 0.0001
South > North
Interaction with cruise
No
Yes
Yes
0.2326
< 0.0001
0.0043
NA
South > North
South > North
Particulate
PCB 33
PCB 118
PCB 180
Total PCBs
frans-Nonachlor
Interaction with cruise
Yes
Yes
< 0.0001
< 0.0001
South > North
South > North
Interaction with cruise
Interaction with cruise
NA = Not applicable
Of the 10 possible comparisons shown in Table 5-17, there were four interactions between the cruise and
location. In five of the other six possible comparisons, there was a statistically significant difference,
with concentrations in the southern portion of the lake greater than those in the northern portion of the
lake.
April 2004
5-45
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Samples were collected from more than one depth at many of the stations during the periods when the
lake was stratified. The choice of the depths of the samples was based on the position of the thermocline
and other factors, but not a clear cutoff at a specific depth. As a result, it was possible to compare the
concentrations of PCBs and fra«s-nonachlor between the samples collected near the surface and those
collected below the thermocline. The results of these comparisons for PCB congeners 33, 118, and 180,
total PCBs and fra«s-nonachlor are shown in Table 5-18 for both the dissolved and particulate samples,
and for the results from the lake overall, those in the northern portion, and those in the southern portion of
the lake. Generally speaking, most of the "shallow" samples were collected above 30 m, and all of the
deep samples were collected below 30 m.
For the dissolved PCB results, the samples from the greater of the two depths at a station (e.g., the deeper
samples) were significantly higher than the samples from the lesser of the two depths (e.g., the shallower
samples) for all of the analytes except dissolved PCB 33. The differences between depths were consistent
across the northern and southern stations, and for the lake overall.
For the particulate samples, the samples from the greater of the two depths at a station (e.g., the deeper
samples) were significantly higher than the samples from the lesser of the two depths (e.g., the shallower
samples) for all of the analytes except for particulate PCB 33 and fra«s-nonachlor in the northern portion
of the lake. In both of those cases, the differences apparent in the results from the samples in the southern
portion of the lake were sufficient to make the results different at depth in the lake overall.
These differences in concentrations with depth are consistent with the expected behavior of these
hydrophobic contaminants. PCBs and fra«s-nonachlor are likely to be introduced into the open lake in a
particulate form, either from atmospheric deposition or associated with particulate matter in tributary
flows, or become associated with particulate matter in the lake through biological processes. As that
particulate matter settles under the influence of gravity, the contaminants will settle too. In addition,
based on the bathymetry of the Mackinac Channel between Lake Michigan and Lake Huron, the sill
between the two lakes is at a depth of approximately 50m. Therefore, even during winter months when
the lakes are not thermally stratified, water below 50 m cannot flow out of Lake Michigan into Lake
Huron. Although mixing of deeper water and surface water does occur and the mixed water may flow out
of the lake, the deeper waters may retain their pollutant loads from historical sources long after the
surface waters of the lake.
5-46 April 2004
-------
PCBs/trans-Nonachlor in Open-lake Water
Table 5-18. Results of Depth Comparisons of Open-lake Concentrations of PCBs and frans-Nonachlor
Fraction/Analyte/Location
Significant Difference between Shallow and Deep
Results?
Probability
Direction
Dissolved
PCB 33 - Overall
PCB 33 - South
PCB 33 - North
PCB 118 -Overall
PCB 118 -South
PCB 118 -North
PCB 180 -Overall
PCB 180 -South
PCB 180 -North
Total PCB - Overall
Total PCB - South
Total PCB - North
frans-Nonachlor - Overall
frans-Nonachlor - South
frans-Nonachlor - North
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
0.3615
0.1570
0.5513
<0.0001
<0.0001
<0.0001
0.0006
0.0245
0.0093
< 0.0001
< 0.0001
0.0091
< 0.0001
< 0.0001
< 0.0001
NA
NA
NA
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Particulate
PCB 33 - Overall
PCB 33 - South
PCB 33 - North
PCB 118 -Overall
PCB 118 -South
PCB 118 -North
PCB 180 -Overall
PCB 180 -South
PCB 180 -North
Total PCB - Overall
Total PCB - South
Total PCB - North
frans-Nonachlor - Overall
frans-Nonachlor - South
frans-Nonachlor - North
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
<0.0001
<0.0001
0.3049
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
0.0013
< 0.0001
< 0.0001
< 0.0001
0.0003
< 0.0001
0.3507
Deep > Shallow
Deep > Shallow
NA
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
Deep > Shallow
NA
NA = Not applicable
April 2004
5-47
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
5.3.3 Other Interpretations and Perspectives
As noted in various earlier sections of this report, there are limitations to the interpretations of the LMMB
Study data presented here. Among the most basic considerations is the fact that this report has focused on
providing the results for only three of the PCB congeners. The rationale for the choice of congeners is
presented in Chapter 2. The interpretations suggested from the data for these three congeners may not
apply to all other PCB congeners studied in the LMMB, and it would be advisable to examine the actual
results for other congeners of interest before accepting the interpretations presented in this report.
The issues surrounding the evaluation of the blanks results in Section 5.2.3, and the impacts of the blanks
on the interpretation of the field sample results cannot be overemphasized. As implemented in the
LMMB Study, the sample collection and analysis procedures applied to PCBs and fra«s-nonachlor in
open-lake waters represent a carefully crafted balance among practicality, affordability, and the size of the
data set. Since the time that this study was conducted, more powerful analytical techniques such as high
resolution GC/MS have been routinely applied to PCB congener analyses. However, the cost of such
analyses would severely limit the number of samples that could be collected and analyzed the LMMB
Study. While high resolution GC/MS may be able to better resolve some of the congeners, it would not
necessarily better address the presence of PCBs in the blanks, except in instances where the contaminants
are not actually PCBs, but had similar GC retention times.
Assessing temporal variation within the LMMB Study is hampered by the fact that not all stations were
sampled on all cruises and that there are relatively few data from winter cruises. Comparisons between
shallow and deep water samples suffer from similar problems, in that not all stations had samples at more
than one depth. Therefore, the results for stations at different depths must be interpreted carefully.
Comparisons to historical data must consider not only the differences in the sizes of the various data sets
(with the LMMB data set generally far larger than any other), but must also take into account significant
differences in the sample collection and analysis procedures.
The concern about blanks for the LMMB data set is readily apparent in the data from Filkins et al. (1983)
and although the investigators in the other two studies do not present similar blank results, it would be
reasonable to assume that blanks would also have been a problem for those studies.
Finally, this report has used "contour" plots of PCB concentrations as a means of visually presenting parts
of a large and complex data set. However, as noted earlier, there are limitations to those plots,
particularly with regard to the identification of "hot spots" in Green Bay, as well as with the demarcation
of PCB concentrations in the open lake when relatively few samples were collected.
5-48 April 2004
-------
Chapter 6
PCBs/trans-Nonachlor in Sediment
6.1 Results
A total of 133 sediment samples were collected from 117 stations in Lake Michigan and 6 stations in
Green Bay and analyzed for PCBs and fra«s-nonachlor. The samples were collected as described in
Sections 2.2.4 and 2.5.4. Selection of sampling sites was based on prior information of the spatial
distribution of sediment grain size, composition (e.g., organic carbon), sediment layer thickness, mixed
layer depth, and 137Cs and 210Pb inventories. Emphasis was placed on sampling the more permanent
accumulations of recent sediment in the depositional basins of the lake. Sediment samples were collected
using three types of equipment: a box core, a Ponar dredge, and a gravity core.
The locations of the LMMB stations Figure 6-1. Sediment Sampling Stations
sampled by each of the three sampling
methodologies are shown in Figure 6-1,
along with the depositional classifications of
Cahill (1981). Cahill developed the class-
ifications of depositional, transitional and
nondepositional zones in Lake Michigan
from an examination of the physical
description of sediment from 286 stations,
grain size information, and echo-sounding
tracks.
Recently deposited sediments were obtained
from depositional basins through the use of
a specially designed box corer. A modified
Soutar box corer was used to retrieve cores
approximately 60 cm in length, with
extremely well preserved sediment-water
interfaces. The 25- x 25- x 60-cm core was
subsequently subcored with four 10-cm
diameter butyrate tubes. Care was taken to
preserve the interface and reduce distortion
of the core upon insertion of the tube and
during subsequent extrusion of the subcore.
Two of the subcores were used for analysis
of PCBs and fra«s-nonachlor, as well as to
determine the organic carbon (OC) content
of the sediment. A third subcore was used
for analysis of radionuclides, trace elements,
and biogenic components. The fourth
subcore was archived. The sections of two
cores from 0-1 cm were combined to allow for enough material for analysis of the PCBs and trans-nonachlor.
In transitional regions where recent deposits thinly overlie older glaciolacustrine clays, modern sediment
is difficult to collect with a box corer. Therefore, a Ponar dredge was used at most transitional and
nondepositional stations. However, this device does not collect modern sediment very efficiently, due to
the bow wave's impact on the uppermost flocculent material. A gravity corer was used instead of a box
corer at a few of the depositional stations, due to the lack of box coring capability of the vessel.
Gravity core
Ponar dredge
Transitional and
nondeposltional
zones
April 2004
6-1
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
In addition to the core samples, sediment Figure 6-2. Sediment Trap Locations
traps were deployed at eight locations |
(Figure 6-2), but the traps at two stations
could not be recovered. Sample retrieval was
successful at trap locations 1, 2, 5, 6, 7, and
As noted in Chapter 2, there are 209
possible PCB congeners and the invest-
igators in this study reported results for 65 to
110 of these congeners, depending on the
capabilities of each laboratory. The invest-
igators for the sediment portion of the study
determined results for 105 congeners or co-
eluting congeners.
For the purposes of this report, we are
presenting summaries of the results for the
following subset of the analytes:
• The coeluting pair of PCB congeners
28 + 31 (2,4,4'-trichlorobiphenyl and
2,4',5-trichlorobiphenyl)
• PCB congener 118
• PCB congener 180
• Total PCBs
• fra«5-nonachlor
The PI for the sediment sample analyses
focused on the coeluting pair 28 + 31 because of concerns that PCB congener 33 coeluted with the
pesticide heptachlor. Based on matrix spike results for heptachlor and given the possibility that
heptachlor was present in the sediments as a result of its historical use in the area around Lake Michigan,
the PI flagged all of the PCB 33 results and presented data for PCB 28 + 31.
6-2
April 2004
-------
PCBs/trans-Nonachlor in Sediment
6.1.1 Geographical Variation
In addition to Cahill's three depositional zones, Figure 6-3. Lake Michigan Basins
Lake Michigan can be divided into four
geographical regions, based on characterizations
of bottom morphology by the U.S. Department of
the Interior (1967) (Figure 6-3):
• The southern basin lies south of a line
between Milwaukee and Muskegon. In this
basin the bottom has a gentle relief and
maximum depth of 180 m. The southern
basin receives urban and industrial inputs of
PCBs from Waukegan Harbor, IL, and the
metropolitan areas of Chicago, IL, Gary, IN,
and Milwaukee, WI.
The divide area, located between the northern
and southern basins, is bounded by two
approximately east-west ridges. Two small
depositional basins reside within the divide,
the Grand Haven basin on the east side and
the Milwaukee basin on the west side. For
this study, the two basins' stations will be
lumped together and referred to as the
"Central basins."
The northern basin is bounded by a line
between Manitowoc and Manistee and
another line from Frankfort to Manistique.
Bottom topography of the northern basin is
much more complex than the southern basin.
A steep grade defines this basin resulting in
a maximum depth of 282 m. The northern
basin receives lower inputs from industrial
and urban centers relative to the southern
basin. However, the extent to which PCB-laden Green Bay sediments contribute to the northern basin's
inventory is unknown.
• The straits area, including Grand Traverse Bay, has a very irregular bottom with isolated depressions
where sediments accumulate. Its morphology is more similar to that of northern Lake Huron than
northern Lake Michigan.
Green Bay samples were not included in this analysis, because only six stations were visited, and these
stations were not the focus of the LMMB Study.
The summary statistics for the four basins are provided in Table 6-1.
April 2004
6-3
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 6-1. Summary Statistics for Surficial Sediments
Analyte
Organic Carbon*
PCB 28+31
PCB118
PCB 180
Total PCBs
frans-Nonachlor
Basin
Southern
Central
Northern
Straits region
Southern
Central
Northern
Straits region
Southern
Central
Northern
Straits region
Southern
Central
Northern
Straits region
Southern
Central
Northern
Straits region
Southern
Central
Northern
Straits region
N
50
22
32
16
50
22
31
16
50
22
31
16
50
22
31
16
50
22
31
16
49
21
32
16
Mean (ng/g)
22.634
22.755
21.479
12.172
4.593
2.563
1.879
0.405
3.271
2.668
2.176
0.404
1.562
1.223
1.070
0.175
69.724
50.541
41.599
7.271
0.599
0.638
0.560
0.168
Range (ng/g)
0.440 - 42.620
0.930 - 49.920
0.430-41.490
1.560-46.070
0.0490-16.486
0.110-11.558
0.0111-4.010
0.0625-1.640
0.0130-9.840
0.0404-7.190
0.00653 - 4.940
0.0655-1.660
0.00740 - 4.783
0.00570-3.512
0.00299 - 2.400
0.0264 - 0.745
0.545-219.288
1.240-149.390
0.138-91.220
1.056-28.600
0.00250 - 2.830
0.00576 - 2.243
0.00661-1.200
0.0296-0.681
Std. Dev. (ng/g)
14.346
17.280
15.999
16.042
3.841
2.855
1.466
0.493
2.608
2.478
1.709
0.493
1.318
1.179
0.873
0.238
53.114
49.406
33.102
9.187
0.522
0.644
0.456
0.217
* Units for organic carbon content are mg/g.
The mean concentrations of the PCB congeners and total PCBs in Table 6-1 exhibit a general trend of
decreasing concentrations from south to north, with the lowest concentrations in the Straits region. In
contrast, the mean concentrations of trans-nonachlor, while lowest in the Straits region, exhibit no south
to north trend. The organic carbon content data exhibit a pattern similar to that for fra«s-nonachlor.
6-4
April 2004
-------
PCBs/trans-Nonachlor in Sediment
Total PCBs (ng/gl
6.1.2 Frequency Distributions
Frequency distributions of total PCBs, trans- Figure 6-4. Frequency Distributions of Total PCBs, trans-
nonachlor and organic carbon (OC) content for Nonachlor, and Organic Carbon
surficial sediment samples from Lake Michigan
are shown in Figure 6-4. These data are not
normally distributed. It should also be noted that
these distributions are biased as a result of our
emphasis on sampling the depositional regions of
the lake, especially in the southern basin. ^
m
D
Total PCBs exhibit a wide range in con- Ł
centration. Although not a true bimodal
distribution, two distinct groups are evident
within the total PCB distribution. The first group
of samples, from nondepositional and transi-
tional stations, has very low concentrations that
exponentially decline in number with increasing
concentration (0-30 ng/g). The second group of
samples taken mainly from depositional sites (35
- 225 ng/g) is more normally distributed, though
tailing toward higher levels is evident.
I
While the distribution of trans-nonachlor also is Ł
skewed toward lower concentrations in non-
depositional and transitional regions (0 - 0.2
ng/g), this compound's distribution in the dep-
ositional region (0.25 - 2.25 ng/g) differs from
PCBs by being more evenly distributed. The
distribution of OC more closely resembles a bi-
modal distribution with the first peak
representing nondepositional and transitional
stations (0-11 mg/g), and the second peak
representing depositional stations (20 - 51 mg/g).
o
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
northern basin than straits region, while the southern and central basins have an even wider range due to a
few stations with very high concentrations. Only in the southern and central basins did total PCBs exceed
100 ng/g, and only in the southern basin did total PCBs exceed 150 ng/g. The distribution of PCB would
indicate that the accumulation of this contaminant is not evenly distributed throughout the lake but rather
is preferentially collecting in depositional areas.
Figure 6-5. Frequency Distributions of Total PCBs by Basin
Southern Basin
Northern Basin
N
Total PCBs (ng/g)
1-1 "f
Total PCBs (ng/g!
Central Basins
Straits Region
i..
Total PCBs (ng/g)
Total PC8s ing/gi
The distribution of fra«s-nonachlor for all basins is generally similar to that of total PCBs (Figure 6-6).
One notable exception is the relative similarity of ranges between basins for frara-nonachlor.
6-6
April 2004
-------
PCBs/trans-Nonachlor in Sediment
Figure 6-6. Frequency Distributions of frans-Nonachlor by Basin
Southern Basin
Northern Basin
trans nonachlor (ng/g)
trans nonachlor (ng/g)
Central Basins
Straits Region
trans nonachlor mg/gi
trans nonachlor (ng/gi
6.2 Quality Implementation and Assessment
As described in Section 1.5.5, the LMMB QA program prescribed minimum standards to which all
organizations collecting data were required to adhere. The quality activities implemented for the PCBs
and trans -nonachlor monitoring portion of the study are further described in Section 2.7 and included use
of SOPs, training of laboratory and field personnel, and establishment of MQOs for study data. A
detailed description of the LMMB quality assurance program is provided in the Lake Michigan Mass
Balance Study Quality Assurance Report (USEPA, 200 Ib). A brief summary of data quality issues for
the sediment PCBs and fra«s-nonachlor data is provided below.
Quality Assurance Project Plans (QAPPs) were developed by the Pis and were reviewed and approved by
GLNPO. Each researcher trained field personnel in sample collection SOPs prior to the start of the field
season and analytical personnel in analytical SOPs prior to sample analysis. Each researcher submitted
test electronic data files containing field and analytical data according to the LMMB data reporting
standard prior to study data submittal. GLNPO reviewed these test data sets for compliance with the data
reporting standard and provided technical assistance to the researchers. In addition, each researcher's
laboratory was audited during an on-site visit at least once during the time LMMB samples were being
analyzed. The auditors reported positive assessments and did not identify issues that adversely affected
the quality of the data.
April 2004
6-7
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
As discussed in Section 2.5, because data comparability was important to the successful development of
the mass balance model, the Pis used similar sample collection, extraction, and analysis methods for the
PCB and fra«s-nonachlor monitoring in this study.
6.2.1 Sample Collection
The LMMB work plan indicated that 131 sediment sampling sites were to be visited. Field records
indicated that only 117 stations were actually sampled. Of the eight sediment trap locations to be
sampled, sample retrieval occurred at trap locations 1, 2, 5, 6, 7, and 8. Numerous attempts were made to
retrieve samples from trap locations 3 and 4, but they were not successful.
The LMMB work plan specified that the sediment depth intervals to be sampled was from 0 to 1 cm.
Actual intervals sampled and analyzed ranged from 0-0.5 to 0-1.5 cm, due to sporadic slanting at the top
surface of the core after retrieval.
During examination of the field collection records for field duplicates, it was discovered that some field
duplicates were not actually collected at the same time as the field sample due to equipment mobilization.
Samples collected within five minutes of each other were considered field duplicates (FD1), and if more
than five minutes elapsed, the samples were considered sequential field duplicates (SFD1). Separate
labeling of these data points as FD1 and SFD1 was done in order to assess if precision differed based on
the elapsed time.
6.2.2 Data Assessments
As discussed in Section 2.7, data verification was performed by comparing all field and QC sample
results produced by each PI with their MQOs and with overall LMMB Study objectives. Analytical
results were flagged when pertinent QC sample results did not meet acceptance criteria as defined by the
MQOs. These flags were not intended to suggest that data were not useable; rather they were intended to
caution the user about an aspect of the data that did not meet the predefined criteria. Table 6-2 provides a
summary of flags applied to the sediment PCB and fra«s-nonachlor data. The summary includes the flags
that directly relate to evaluation of the MQOs to illustrate some aspects of data quality, but does not
include all flags applied to the data to document sampling and analytical information, as discussed in
Section 2.7. Compared to other matrices, the percentage of results that were qualified for these criteria is
relatively small.
Pis used surrogate spikes to monitor the bias of the analytical procedure. The PCB results were corrected
for the recoveries of the surrogates. The fra«s-nonachlor results were not surrogate-corrected. Only
0.5% of the PCB 28 +31 results, 2% of PCB 118 results, and 1% of the PCB 180 results were qualified
for surrogate recovery problems (Table 6-2).
Laboratory matrix spike samples also were used to monitor the bias of the analytical procedure. The
results for the matrix spike samples were compared to the MQO for spike recoveries (50 - 125%).
Analytical results associated with matrix spike samples with recoveries below the MQO limits were
flagged with failed matrix spike and low bias flags and results associated with matrix spike samples that
had recoveries higher than the MQO limits were flagged with failed matrix spike and high bias flags.
Analytical results were considered invalid and flagged as such when the analyte was undetected and
recoveries for associated matrix spike samples were less than 10%. No sediment fra«s-nonachlor samples
failed the matrix spike MQOs. None of the results for PCBs 28+31 or 180 were flagged as failing the
matrix spike MQOs. For PCB 118, 5% of the results were flagged as failing the matrix spike MQOs.
6-8 April 2004
-------
PCBs/trans-Nonachlor in Sediment
Field blanks were collected for PCBs and fra«s-nonachlor. When field blank contamination was greater
than 3.3 times the method detection limit, all of the associated results were flagged with the failed field
blank sample code (FRB). Field blanks were not collected at all stations, so potential station-specific
contamination associated with these sites cannot be evaluated. However, contamination associated with
sampling equipment, collection, processing, shipping, storing, and analysis can be evaluated based on the
field blanks collected throughout the study. For PCB 28+31, 11% of the field samples were associated
with a field blank in which this congener was reported above the sample-specific detection limit (Table 6-
2). None of the field sample results for fra«s-nonachlor were qualified because of field blank results.
Field duplicates were to be collected at a frequency of 5%. Duplicate samples collected within 5 minutes
of each other were considered field duplicates. However, an examination of the field collection records
indicated that some of the planned field duplicates were not collected within that 5-minute time frame as a
result of problems with equipment mobilization or the time required to pump the sample through the filter
and resin cartridge. Those "duplicates" that were collected more than 5 minutes apart were considered
"sequential field duplicates" and the data were labeled accordingly (e.g., SFD1 vs. FD1). Combining the
field duplicates and sequential field duplicates, the actual rate of collection of duplicates was 5.5%.
The results from the original field sample and the associated duplicate were compared on the basis of the
relative percent difference (RPD). The RPD value for each PCB congener and trans-nonachlor was
compared to the MQO for field duplicate precision. None of the field samples results for PCBs or trans-
nonachlor were qualified because of the field duplicate precision (FFD) concerns (Table 6-2).
As discussed in Section 1.5.5, MQOs were defined in terms of six attributes: sensitivity, precision,
accuracy, representativeness, completeness, and comparability. GLNPO derived data quality assessments
based on a subset of these attributes. For example, system precision was estimated as the mean RPD
between the results for field duplicate pairs. Similarly, analytical precision was estimated as the mean
RPD between the results for laboratory duplicate pairs. Table 6-3 provides a summary of data quality
assessments for several of these attributes for the sediment PCB and fra«s-nonachlor data.
Because the relative variability of most measurement techniques increases as one approaches the
detection limit of the technique, the assessment of the field duplicate results were divided into two
concentration regimes. One measure of system precision was calculated for those results that were less
than 5 times the sample-specific detection limit (SSDL) of the analyte and a separate measure was
calculated for those results that were greater than 5 times the SSDL. However, none of the sediment
sample results with corresponding field duplicates contained PCB 28+31, 118, or 180 at concentrations
below the SSDL. The precision of the field duplicate results above the SSDL ranged from 19 to 24% for
the PCB congeners in Table 6-3.
Three of the four field duplicate pairs contained fra«s-nonachlor below the SSDL and the mean precision
of those pairs was 37%, compared to one field duplicate pair containing fra«s-nonachlor above the SSDL,
with an RPD of 26%.
The variability of the laboratory measurement technique was assessed through the preparation and
analysis of laboratory duplicate pairs. Eleven laboratory duplicate pairs were analyzed for the PCB
congeners, and ten pairs were analyzed for fra«s-nonachlor. The laboratory duplicate precision ranged
from 9 to 13% for the PCB congeners and fra«s-nonachlor (Table 6-3). As expected, the laboratory
duplicate results are more precise than the field duplicate results.
April 2004 6-9
-------
Table 6-2. Summary of Routine Field Sample Flags Applied to Select PCB Congeners and frans-Nonachlor in Sediment Samples
Analyte
PCB 28 + 31
PCB 118
PCB 180
frans-Nonachlor
Flags
Sensitivity
MDL
5% (11)
3% (6)
6% (13)
28% (54)
UNO
0
0
0
0
Contamination
FFR
11% (23)
0
0
0
Precision
FFD
0
0
0
0
Bias
FSS
0.5% (1)
2% (3)
1%(3)
0
FMS
0
5% (10)
0
0
LOB
0
0
0
0
HIB
0
0
0
0
INV
0
0
0
0
FBK
0
0
0
0
FDL
0
0
0
0
The number of routine field samples flagged is provided in parentheses. The summary provides only a subset of applied flags and does not represent the full suite of flags applied
to the data.
MDL = Less than method detection limit (Analyte produced an instrument response but reported value is below the calculated method detection limit. Validity of reported
value may be compromised.)
UNO = Analyte not detected (Analyte produced no instrument response above noise.)
FFR = Failed field blank (A field blank sample, type unknown, associated with this analysis failed the acceptance criteria. It is unknown whether the blank that failed was a
field blank or a lab blank. Validity of reported value may be compromised.)
FFD = Failed field duplicate (A field duplicate associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FSS = Failed surrogate (Surrogate recoveries associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FMS = Failed matrix spike (A matrix spike associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
LOB = Likely biased low (Reported value is probably biased low as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery or other internal
lab QC data. Reported value is not considered invalid.)
HIB = Likely biased high (Reported value is probably biased high as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery, blank
contamination, or other internal lab QC data. Reported value is not considered invalid.)
INV = Invalid.
FBK = Failed laboratory blank (Laboratory reagent blank result greater than the SSDL).
FDL = Failed laboratory duplicate
6-10
-------
Table 6-3. Data Quality Assessment for Select PCB Congeners and frans-Nonachlor in Sediment Samples
Analyte/Number Field Samples
PCB 28 + 31
197 samples
PCB 118
189 samples
PCB 180
196 samples
frans-Nonachlor
185 samples
Parameter
System Precision - Mean Field Duplicate RPD (%), < 5 * SSDL
System Precision - Mean Field Duplicate RPD (%), > 5 * SSDL
Analytical Precision - Mean Laboratory Duplicate RPD (%), > 5 * SSDL
Analytical Bias - Mean Laboratory Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < SSDL (%)
System Precision - Mean Field Duplicate RPD (%), < 5 * SSDL
System Precision - Mean Field Duplicate RPD (%), > 5 * SSDL
Analytical Precision - Mean Laboratory Duplicate RPD (%), > 5 * SSDL
Analytical Bias - Mean Laboratory Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < SSDL (%)
System Precision - Mean Field Duplicate RPD (%), < 5 * SSDL
System Precision - Mean Field Duplicate RPD (%), > 5 * SSDL
Analytical Precision - Mean Laboratory Duplicate RPD (%), > 5 * SSDL
Analytical Bias - Mean Laboratory Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < SSDL (%)
System Precision - Mean Field Duplicate RPD (%), < 5 * SSDL
System Precision - Mean Field Duplicate RPD (%), > 5 * SSDL
Analytical Precision - Mean Laboratory Duplicate RPD (%), > 5 * SSDL
Analytical Bias - Mean Laboratory Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < SSDL (%)
Number of QC samples
0 field duplicate pairs
9 field duplicate pairs
1 1 lab duplicate pairs
24 matrix spikes
-
0 field duplicate pairs
9 field duplicate pairs
1 1 lab duplicate pairs
24 matrix spikes
—
0 field duplicate pairs
9 field duplicate pairs
1 1 lab duplicate pairs
24 matrix spikes
—
8 field duplicate pairs
0 field duplicate pairs
10 lab duplicate pairs
22 matrix spikes
—
Assessment
—
16%
10%
102%
6%
—
17%
9%
112%
3%
—
18%
13%
108%
7%
25%
—
12%
88%
28%
RPD = Relative percent difference
SSDL = Sample-specific detection limit
6-11
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Analytical bias was assessed using the results from matrix spike samples. The mean recoveries of the
analytes were excellent for the PCBs, ranging from 102% to 112% for the PCB congeners in Table 6-3.
The mean recovery for fra«s-nonachlor was very good, at 88%. Thus, these results demonstrate that the
analytical techniques applied to the field samples introduce little or no bias into the PCB results, and a
slight low bias into the trans-nonachlor results.
6.3 Data Interpretation
6.3.1 Comparison to Historical Studies
LMMB surficial sediment concentrations of total PCBs were compared to measurements made by
Swackhamer and Armstrong (1988), Hermanson et al. (1991), and Golden (1994). Swackhamer and
Armstrong (1988) collected 20 box cores from the southern and Grand Haven depositional basins in
1978-1980. Hermanson et al. (1991) collected five box cores from the northern, southern and Grand
Haven basins in 1984. More recently, Golden (1994) collected four box cores from the northern and
southern basins in 1991 and 1992. Only box cores taken from depositional regions were used for this
comparison because of the lack of permanent sediment accumulation in the transitional and
nondepositional regions.
Swackhamer and Armstrong (1988) and Hermanson et al. (1991) used Aroclor analysis to quantify total
PCBs. Aroclor analysis using COMSTAR, a multiple linear regression analysis program that provides the
best fit to Aroclor composition, has been shown to compare favorably (90-95%) with total PCBs
measurements using congener analysis (Burkhard and Weininger, 1987). Interpolated values from the
contour map of LMMB surficial PCBs (1994-1995) were compared to the measured historical values.
Plots of the interpolated versus measured total PCB values are shown in Figure 6-7.
Differences between LMMB values and historical measurements are compared to gains/loss lines (solid
line, 0% change; dashed lines, 50% gain or loss; dotted line, 100 % gain). In Figure 6-7A, the author,
basin, and station are designated for each data point (s= Swackhamer and Armstrong [1988];
h=Hermanson et al. [1991]; g=Golden [1994]). In Figure 6-7B, the stations are divided into basins
(n=north, c=central, s=south) followed by areas within basins (dh=deep hole, es=eastern side).
It appears that in the southern deep hole (sdh) and central deep hole (cdh) regions, total PCB levels have
either increased (35 -> 100%) or remained un-changed since they were measured in 1978-1984. In
contrast, 4 of 5 stations on the eastern side of the southern basin (ses) exhibited reductions in PCB
concentrations.
In the northern basin, surficial sediment total PCBs appear to have declined significantly (-40%) since
Hermanson's 1984 collection, and even over the 3-year period since Golden's 1992 collection. It is
possible that the northern basin's complex topography results in a very heterogenous distribution of total
PCB sediment concentrations. Thus, comparisons of historical data with recent interpolated values could
be misleading. There is other evidence, however, that a significant decline in total PCB levels is
occurring in the northern basin. A core profile of total PCBs at LMMB station 103 indicates that a 33%
decline in PCB concentrations has occurred since the early 1980s. Looking at Golden's sediment core
profiles, a loss of only 1-9% was noted for two northern cores between about 1984 and 1992. However, a
30% loss was noted between about 1980 and 1992.
6-12 April 2004
-------
PCBs/trans-Nonachlor in Sediment
Figure 6-7. Comparison of Measured Historical Total PCB Results in Surficial Sediment and
Interpolated Total PCB Results from Contours of the 1994-1995 LMMB Collections
100 150
Measured t-PCBs (ng/g)
150
-5
~
S 100
«
CD
O
D.
Q. 50
sdh
cdh
cdh
50 100 150
Measured t-PCBs (ng/g)
200
B
200
0% change 50% gain or loss 100% gain
See text for the key to the letter designations of each data point.
April 2004
6-13
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Since PCBs were banned from use in North America in 1979 and loadings to the Great Lakes have
decreased (Eisenreich etal, 1989; Rapaport and Eisenreich, 1988), reductions in surficial sediment PCB
content are expected. Decreasing sediment PCB concentrations have been observed in Lakes Ontario and
Superior (Golden et a/., 1993). The lack of a decline in sediment PCB content in the southern basin of
Lake Michigan may result from sediment focusing processes that redistribute the PCB inventory, and/or
continued external source inputs.
6.3.2 Other Interpretations and Perspectives
6.3.2.7 Relationship of PCB Congeners and trans-Nonachlor with Sediment Organic Carbon
Concentrations of PCBs and fra«s-nonachlor in surficial sediments increase with increasing organic
carbon (OC) content. The correlations of these two variables, however, were not very strong on a linear
scale (r2 = 0.25 - 0.67). The spread in the data suggest that the southern basin stations were significantly
higher in contamination. Dividing the stations into groups, based on basin morphology and regional
inputs, may explain some of the variability and improve OC correlation.
Generally, the concentrations of PCB congeners and tmns-nonachlor increased with sediment organic
carbon content according to a power equation,
[PCB] = a x [OC]b
This power equation best described the relationship of PCB congeners and frara-nonachlor with sediment
organic carbon in all four regions of Lake Michigan. Coefficients of the power relationship were
obtained by log transformation of the data and fitting the linear relationship,
log[PCB] = A + (B x log[0q)
Linear regressions for the four compounds in each of the four regions are shown in Figures 6-8 -6-11.
Coefficients for the relationships are listed in Table 6-4. The r2 values for these relationships, ranged
from 0.865 - 0.977, indicating a strong correlation between the two variables. This finding was expected
as hydrophobic compounds are known to partition strongly to organic matter. The southern basin PCB
relationships exhibited an additional feature not seen in other basins. A peak in PCB concentration was
observed at -25 mg/g OC for PCB 28+31 and at -35 mg/g for PCB 118 and PCB 180 (Figure 6-8).
fra«s-Nonachlor did not exhibit such a peak.
A comparison of coefficient results was used to discern any differences in compound behavior between
basins, as well as differences between compounds within a basin. Congener-specific behavior within a
basin was examined by comparing regression slopes (i.e., extent of association with OC). A slope greater
than 1 suggests that enrichment of the compound is occurring with increasing OC content. A slope of 1
also indicates a linear relationship of contaminant content with OC content, while a slope of < 1 indicates
a dilution of contaminant content with increasing OC content. Non-unity slopes may result from
differences in OC composition or PCB loadings between nearshore sandier stations and offshore siltier
stations.
Generally, slopes were >1 for all chlorinated organic contaminants in the southern, central and northern
basins. An exception occurred for PCB 28+31 in the southern and central basins where the slopes ranged
from 0.88 to 1. In the straits region, all compounds exhibited slopes <1.
Significant differences in slopes between compounds for a specific basin are designated by dissimilar
superscripts in Table 6-4. For the southern and central basins, a difference in PCB congener behavior was
6-14 April 2004
-------
PCBs/trans-Nonachlor in Sediment
noted. Within these basins, the slopes of the more highly chlorinated congeners, PCB 118 and PCB 180
were significantly greater than those of the trichlorobiphenyls, PCB 28+31. In the northern basin, only
congeners PCB 180 and PCB 28+31 were significantly different. For the straits region, no significant
difference in slopes was observed between congeners.
The relationship of fra«s-nonachlor with OC did not follow the same pattern as any particular PCB
congener. No significant differences between the slopes of trans-nonachlor and any of the PCB
congeners were noted for the central basins and straits region. However, in the southern basin, the slope
for trans-nonachlor was lower than that of PCB 180, while in the northern basin its slope was lower than
the slopes for PCB 118 and PCB 180.
Figure 6-8. Relationships of PCB/frans-Nonachlor
Concentrations and Organic Carbon Content for
the Southern Basin
Figure 6-9. Relationships of PCB/frans-Nonachlor
Concentrations and Organic Carbon Content for
the Central Basins
Southern Basin Surficial Sediments (0-1 cm)
Central Basins Surficial Sediments (0-1 cm)
Organic Carbon (mg/g)
*.• *
* PCBS115
PCB #180
pea *nfi
PCB #130
Organic Carbon (mg/g)
Organic Carbon [mg-rg)
April 2004
6-15
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 6-10. Relationships of PCBItrans-
Nonachlor Concentrations and Organic Carbon
Content for the Northern Basin
Northern Basin Sutficial Sediment (0-1 cm)
Figure 6-11. Relationships of PCBItrans-
Nonachlor Concentrations and Organic Carbon
Content for the Straits Region
Slrails Region Surficial Sediment (0-1 cm)
' 'I*
1r3n5 nonchlcr
Organic l.ait»n (mg-'gi
PCBS1 IS
PCBfflSO
PCBS113
PCBW1SQ
Organic Carbon (mg/g)
10
Organic Carbon [mg/g)
6-76
April 2004
-------
PCBs/trans-Nonachlor in Sediment
Table 6-4. Linear Regression Parameters of Log PCB and frans-Nonachlor versus Log OC Content in
Surficial Sediments in Lake Michigan
Basin
Southern
Central
Northern
Straits
Whole lake
Parameter
slope
intercept
i2
n
slope
intercept
f
n
slope
intercept
f
n
slope
intercept
i2
n
slope
intercept
i2
n
PCB 28+31
0.88093
-0.5799
0.7329
50
1.0033
-1.076
0.9160
21
1.2393ac
-1.4384
0.9611
31
0.85873
-1.3100
0.9138
16
1.091
-1.0893
0.7587
119
PCB 118
1.1461bc
-1.0717
0.9810
50
1.1884"
-1.2402
0.9562
22
1.3269bc
-1.4963
0.9718
31
0.81603
-1.2589
0.9189
16
1.2215
-1.2938
0.9024
120
PCB 180
1.2178"
-1.5113
0.9370
50
1.3326"
-1.7968
0.9457
22
1.3763"
-1.8870
0.9691
31
0.83073
-1.6658
0.8870
16
1.2945
-1.7365
0.9077
120
frans-Nonachlor
1.0761°
-1.7336
0.9048
48
1.1500ab
-1.8006
0.9278
21
1.1341"
-1.7966
0.9773
32
0.84963
-1.7002
0.9095
16
1.0878
-1.7645
0.9300
117
Dissimilar superscripts denote a significant difference (p<0.05) in slopes between compounds for a specific basin.
Several general observations can be made based on the relationships of organic carbon content and the
chlorinated organic contaminants for each of the basins (Table 6-5). For the high molecular weight
(HMW) congeners, PCB 118 and PCB 180, and fra«s-nonachlor, the straits region slopes were
significantly lower than the other basins, and the northern basin slopes were higher than those of the
southern basin (except for frara-nonachlor). For the low molecular weight (LMW) congeners, PCB
28+31, the northern basin slope was significantly higher than for all other basins. No significant
difference was detected between slopes for PCB 28+31 in the central, southern and straits regions.
Interestingly, the southern basin stations that had high OC content and lower than expected PCB
concentrations were located in the basin's deep hole at >100 m depth; fra«s-nonachlor did not exhibit
reduced concentrations at the deep-hole stations. If these stations are excluded from the regressions, then
the slopes of the southern basin (1.11 for PCB 28+31, 1.23 for PCB 118, 1.27 for PCB 180) are not
significantly different from those of the central and northern basins for all PCB congeners tested.
The intercept parameter is an indication of the level of contamination as a net result of loading and
removal rates. For fra«s-nonachlor, the intercept was similar for all four basins, suggesting that regional
inputs may be similar if removal rates are not significantly different between basins (Tables 6-7 and 6-8).
At the other end of the spectrum, PCB 28+31 had higher intercepts in the southern and central basins
relative to the northern basin and straits region. The HMW congeners, PCB 118 and PCB 180 had higher
intercepts in the southern basin than the northern basin, while the intercepts in the central basins and
straits region fell in between.
In summary, while it would appear at first glance that there is not much of a relationship of either PCB or
fra«5-nonachlor concentration to the OC content of surficial sediments in Lake Michigan, strong power
relationships resulted when the stations were divided into basin groups. The association between trans-
nonachlor concentration and sediment OC does not vary throughout the lake's basins, except in the straits
April 2004
6-17
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
region. For the PCB congeners, more variability in their associations with sediment OC was noted
between basins. The variability appears to result, in part, from differences in regional source input
functions. Other variables that control burial, such as sedimentation rate, residence time in the sediment
mixed layer, sediment focusing processes, and compound hydrophobicity, may also influence PCB
associations with particles and particulate OC in Lake Michigan.
Table 6-5. Significant Differences between Linear Regression Parameters among Lake Michigan Basins for
Log Analyte versus Log Organic Carbon
Analyte
PCB 28+31
PCB 118
PCB 180
frans-Nonachlor
Regression Parameter
slope
intercept
slope
intercept
slope
intercept
slope
intercept
Significant Differences at a = 0.05
Southern (So)
N
C, N, St
N, St
N, St
N, St
C, N
St
-
Central (C)
N
So, N, St
St
N
St
So
St
-
Northern (N)
C, St, So
C, So
So, St
C, So, St
So, St
So, St
St
-
Straits (St)
N
C, So
C, So, N
N, So
C, So, N
N
C, So, N
-
6.3.2.2 Contour Maps
The data for the 117 LMMB stations were used to Figure 6-12. Contour Plot of Organic Carbon Content
create contour maps of surficial sediment concen- in Lake Michigan Sediments
trations for organic carbon, total PCBs, PCB 28+31,
PCB 118, and PCB 180, and trans-nonachlor (Figures
6-12 - 6-17). Note that the contour maps for PCB
congeners are normalized to the highest concentration
measured in order to facilitate comparisons.
High OC concentrations (20-50 mg/g) are associated
with the accumulation of recent sediments in the
depositional and transitional regions of Lake Michigan
(Figure 6-12). In addition, the OC distribution in Lake
Michigan generally tracks with the depth contours of
the lake (Figure 6-18). Sediment OC concentrations >
30 mg/g are generally found at depths greater than 100
m. The other cluster of high OC content stations is
located in the eastern portion of the southern basin.
These sites are located at a depth range of 45 - 85 m
and have OC contents of 20 - 33 mg/g. The cluster of
low OC content stations (< 10 mg/g) includes those
generally located at depths less than 100 m. Stations
that do not fall within these three clusters are either
located in transition areas, or located in the straits area.
Stations in the straits region tend to have higher OC
content than other stations located at similar depths.
6-18
April 2004
-------
PCBs/trans-Nonachlor in Sediment
PCBs are also accumulating in the depositional and transitional regions of Lake Michigan (Figure 6-13).
In particular, PCBs accumulate at relatively high concentrations (> 100 ng/g) along the eastern side of the
southern basin, as well as at a few of the deeper stations in the southern and central basins. Enrichment of
most PCB congeners with increasing OC content in sediments was demonstrated by the slopes of greater
than 1 for the relationship of log PCB versus log OC(exceptthe straits region, Table 6-4). Localized
input sources may contribute to the depositional pattern and explain the enrichment of PCBs relative to
OC. Southern basin surficial sediment concentrations have a higher maximum, and greater areal coverage
for the high Total PCB concentration range (MOO ng/g) than the northern basin. Additionally, focusing
processes may be responsible for concentrating PCBs into select regions of the lake.
The contour maps of PCB 118 and PCB 180 exhibit similar distribution patterns to one another
throughout Lake Michigan (Figures 6-15 and 6-16). The accumulation of these PCB congeners occurs in
the depositional and transitional regions with the highest concentrations located in the deepest parts of the
southern and central basins, and in the upper half of the eastern side of the southern basin. The straits
region has some of the lowest concentrations of these two PCBs despite having some stations with
relatively high levels of OC content.
The contour map for PCB 28+31 provides evidence of accumulation in the surficial sediments of the
lake's depositional and transitional regions (Figure 6-14). Relative to PCB 118 and PCB 180, however,
elevated concentrations of PCB 28+31 are found on the eastern side of the southern basin and for Station
55, in the Grand Haven basin. A plot of the concentrations of PCB 28+31 versus PCB 180 for stations in
the southern and central basins illustrates the enhanced levels of this low molecular weight PCB congener
on the eastern side of the southern basin (Figure 6-19). In contrast, depositional/transitional stations
located in the rest of the lake, including the northern basin (not shown), are about a factor of four lower in
their concentrations of PCB 28+31, for a given concentration of PCB 180. PCBs 15+17 (not shown) also
demonstrated enrichment in this region, while PCB 101 (not shown) and PCB 118 had a one-to-one
correspondence with PCB 180 (Figure 6-19). Other stations with elevated levels of PCB 28+31 relative
to PCB 180 are located in nearshore nondepositional areas (< 25 km offshore) with low concentrations of
PCBs: along the northern coast, and western coast of the southern and central portion of the lake. Note
that Waukegan Harbor and Green Bay also have elevated levels of PCB 28+31 relative to PCB 180, due
to point-source contamination of Aroclors 1242 and 1248 (Swackhamer and Armstrong 1988, and
Manchester 1993).
April 2004 6-19
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 6-13. Contour Plot of Total PCBs in Lake
Michigan Sediments
Figure 6-14. Contour Plot of PCB 28+31 in Lake
Michigan Sediments
Figure 6-15. Contour Plot of PCB 118 in Lake
Michigan Sediments
Figure 6-16. Contour Plot of PCB 180 in Lake
Michigan Sediments
6-20
April 2004
-------
PCBs/trans-Nonachlor in Sediment
fra«s-Nonachlor is also accumulating in the depositional Figure 6-17. Contour Plot of frans-Nonachlor in
and transitional regions of Lake Michigan. Unlike PCBs, Lake Michigan Sediments
frara-nonachlor is not preferentially accumulating along
the eastern side of the southern basin. Nor does trans-
nonachlor have the elevated concentrations in the
southern basin relative to the northern basin that PCBs
exhibit. The log-log slope of 1.09 for fra«s-nonachlor and
OC suggests more of a linear relationship and less of a
power relationship with OC compared to PCBs.
Figure 6-18. Organic Carbon Content versus Depth in
Lake Michigan
o
° 20-
.* *
f S
100 150
Depth (m)
April 2004
6-21
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 6-19. Plot of PCB 28+31 Concentrations versus PCB 180 Concentrations in Lake
Michigan Sediments
ro
co
CM
CD
O
18
16
14
10
8 -
6 -
4
2
0
SES stations
nonSES stations
0123456
PCB#180(ng/g)
5 -
&>
ra
c
CO
5
CD
O
CL
3 -
2 -
1 -
All stations
4 6
PCB#180(ngyg)
10
12
6-22
April 2004
-------
Chapter 7
PCBs/trans-Nonachlor in the Lower Pelagic Food Web
7.1 Results
The lower pelagic food web was sampled from June 1994 through October 1995 for PCB and trans-
nonachlor analysis. Individual samples of the lower pelagic food web included mixed phytoplankton,
mixed zooplankton, Diporeia spp., and Mysis spp. Phytoplankton were collected by pumping water from
the water column at the optimum depth for maximum phytoplankton density, zooplankton were collected
in vertical tows, Diporeia spp. were collected in benthic tows, and Mysis spp. were collected in vertical
and benthic tows (see Section 2.5.5 for details of the sample collection procedures). Lower pelagic food
web samples were collected from 15 locations in Lake Michigan, including 9 stations within 4 designated
biological sampling areas (biota boxes) and 6 additional routine monitoring stations (Table 7-1).
*• Chicago biota box - a station in southern Lake Michigan basin near Chicago
*• Sturgeon Bay biota box - a combination of three stations on the western side of the northern Lake
Michigan basin near Sturgeon Bay, Wisconsin
»• Port Washington biota box - a combination of two stations in the central Lake Michigan basin near
Port Washington, Wisconsin
*• Saugatuck biota box - a series of three stations on the eastern side of the southern Lake Michigan
basin near Saugatuck, Michigan.
A total of 208 lower pelagic food web samples were collected and analyzed for fra«s-nonachlor, and 233
lower pelagic food web samples were collected and analyzed for PCBs (Table 7-1).
As noted in Chapter 2, there are 209 possible PCB congeners, and the investigators in this study reported
results for 65 to 110 of these congeners, depending on the capabilities of each laboratory. The University
of Minnesota determined results for 110 congeners or co-eluting congeners.
For the purposes of this report, we are presenting summaries of the results for the following subset of all
of the analytes:
• PCB congener 33
• PCB congener 118
• PCB congener 180
• Total PCBs
• fra«5-nonachlor
April 2004 7-1
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 7-1. Number of Lower Pelagic Food Web Samples Analyzed for PCB Congeners and frans-Nonachlor
Sample Type
Diporeia
Mysis
Phytoplankton
Sampling Locations
Chicago biota box
Sturgeon Bay
biota box
Port Washington
biota box
Saugatuck
biota box
Other
5
40
180
240
280
340
380
47M
Sampling Dates
06/26/94 to 10/1 0/95
06/1 8/94 to 09/23/95
08/1 0/94 to 09/22/95
06/2 1/94 to 10/02/95
06/2 1/94 to 10/0 1/95
06/25/94 to 10/06/95
06/25/94 to 10/06/95
06/1 7/94 to 06/1 7/94
Total
Chicago biota box
Sturgeon Bay
biota box
Port Washington
biota box
Saugatuck
biota box
Other
5
140
180
240
280
340
380
18M
27M
47M
06/26/94 to 10/1 0/95
06/1 8/94 to 09/23/95
08/1 0/94 to 09/22/95
06/2 1/94 to 10/02/95
06/2 1/94 to 10/0 1/95
06/25/94 to 10/06/95
06/24/94 to 10/06/95
06/22/94 to 10/09/95
08/09/95 to 08/09/95
06/1 7/94 to 09/1 9/95
Total
Chicago biota box
Sturgeon Bay
biota box
Port Washington
biota box
Saugatuck
biota box
Other
5
110
140
180
240
280
310
340
380
18M
23M
27M
41
47M
06/26/94 to 10/1 0/95
06/1 9/94 to 09/23/95
06/1 8/94 to 09/23/95
06/1 8/94 to 09/22/95
06/2 1/94 to 10/02/95
06/2 1/94 to 10/0 1/95
06/26/94 to 10/08/95
06/25/94 to 10/06/95
06/24/94 to 10/06/95
06/22/94 to 10/09/95
06/23/94 to 06/23/94
06/20/94 to 06/20/94
06/1 8/94 to 06/1 8/94
06/1 7/94 to 09/1 9/95
Total
Number of Samples
Analyzed for frans-
Nonachlor
6
5
4
5
6
6
6
1
39
6
6
5
6
6
6
6
5
1
5
52
7
6
6
6
6
5
6
6
7
6
1
1
1
6
70
Number of Samples
Analyzed for PCB
Congeners and Total
PCBs
6
5
4
5
6
6
6
1
39
6
6
5
6
6
6
6
6
1
5
53
7
6
6
6
6
6
6
6
7
6
1
1
1
6
71
7-2
April 2004
-------
PCBs and trans-Nonachlor in the Lower Pelagic Food Web
Sample Type
Zooplankton
Sampling Locations
Chicago biota box
Sturgeon Bay
biota box
Port Washington
biota box
Saugatuck
biota box
Other
5
110
140
180
240
280
310
340
380
18M
27M
47M
19M
Sampling Dates
06/26/94 to 10/1 0/95
06/1 9/94 to 09/23/95
06/1 8/94 to 09/23/95
06/1 8/94 to 09/22/95
06/2 1/94 to 10/02/95
06/2 1/94 to 10/0 1/95
06/26/94 to 10/08/95
06/25/94 to 10/06/95
06/25/94 to 10/06/95
06/22/94 to 10/09/95
06/20/94 to 06/20/94
06/1 7/94 to 09/1 9/95
01/24/95 to 01/24/95
Total
Total
Number of Samples
Analyzed for trans-
Nonachlor
7
5
5
5
5
2
6
5
2
1
1
2
1
47
208
Number of Samples
Analyzed for PCB
Congeners and Total
PCBs
7
6
6
6
6
6
6
6
7
6
1
6
1
70
233
7.1.1 Sample Type and Species Variation
PCB and trans-nonachlor concentrations measured in the lower pelagic food web differed significantly
among phytoplankton, zooplankton, Mysis spp., and Diporeia spp. samples (Figure 7-1). Concentrations
of PCB 33, PCB 118, PCB 180, total PCBs, and trans -nonachlor were highest in samples of Diporeia
spp., followed by Mysis spp., zooplankton, and phytoplankton, respectively (Table 7-2). Total PCB
concentrations were 9 times higher in Diporeia spp. than in phytoplankton, averaging 420, 250, 170, and
49 ng/g dry weight in Diporeia spp., Mysis spp., zooplankton, and phytoplankton samples, respectively.
Trans -Nonachlor concentrations were 19 times higher in Diporeia spp. than in phytoplankton, averaging
32, 25, 16, and 1.7 ng/g dry weight in Diporeia spp., Mysis spp., zooplankton, and phytoplankton
samples, respectively.
A portion of the difference in PCB and trans-nonachlor concentrations among lower pelagic food web
sample types is likely due to variations in the lipid content of the samples. Hydrophobic organic
contaminants such as PCBs and fra«s-nonachlor preferentially concentrate in the fatty tissues of
organisms, so those organisms with higher lipid content will likely concentrate more of these
contaminants. This is evidenced by the fact that lipid content was positively correlated with total PCB
and trans -nonachlor concentrations (r2 of 0.25 for total PCB and 0.40 for trans -nonachlor), and the lipid
content of phytoplankton was significantly lower than for the other sample types. The differences in lipid
content among the sample types, however, explained only a quarter to less than half of the variability in
total PCB and trans -nonachlor concentrations. Even when total PCB and fra«s-nonachlor concentrations
were normalized by lipid content, the trends in PCB and fra«s-nonachlor concentrations among the
sample types were almost always the same (Figure 7-2). Normalized total PCB and trans -nonachlor
concentrations in Diporeia spp. and Mysis spp. were significantly higher than in zooplankton and
phytoplankton, and normalized trans -nonachlor concentrations in zooplankton were significantly higher
than in phytoplankton. Normalized total PCB concentrations in zooplankton, however, were not
significantly different than in phytoplankton.
April 2004
7-3
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 7-1. Total PCB and frans-Nonachlor Concentrations in the Lower Pelagic
Food Web
O)
6
Ł=
O
O
CD
O
Cl-
io
"o
O)
6
Ł=
O
O
o
TO
1000 =
100,
10=
D
:
O)
CO
100g
10,
.1
Tl
~~Y
O
T3
Q)
^
O
N
O
0
•D^
^
O
3
^
^^,
CO
co'
S"
II
Oi
•$Ł
D
O
S
^
^
ii
fni
CD
N
O
o
•D^
03
3
Ł
O
CO
co'
II
01
K)
T3
O
CD
O
•D^
03
3
s:
o
00
Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile (top of box) results. Bars represent the
results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest edge of the box. Circles represent results
beyond 1.5*IQR from the box. The Xs represent results beyond 3*IQR from the box. The letters (A - D) above the boxes represent the results
of the analysis of variance and multiple comparisons test. Boxes with the same letter were not statistically different (at alpha = 0.05).
Concentration is plotted on a log scale.
7-4
April 2004
-------
PCBs and trans-Nonachlor in the Lower Pelagic Food Web
Table 7-2. Mean Concentrations of PCBs and frans-Nonachlor Measured in the Lower Pelagic Food Web
Analyte
PCB33
PCB118
PCB 180
Total PCBs
frans-Nonachlor
Sample Type
Diporeia
Mysis
Phytoplankton
Zooplankton
Diporeia
Mysis
Phytoplankton
Zooplankton
Diporeia
Mysis
Phytoplankton
Zooplankton
Diporeia
Mysis
Phytoplankton
Zooplankton
Diporeia
Mysis
Phytoplankton
Zooplankton
N
39
53
71
70
39
53
71
70
39
53
71
70
39
53
71
70
39
52
70
47
Mean
(ng/g)
0.99
0.64
0.27
0.53
15
13
1.6
5.5
17
9.4
1.4
6.3
420
250
49
170
32
25
1.7
16
Range (ng/g)
0 to 3.0
0 to 5.0
0 to 2.8
0 to 2.2
8.6 to 36
6.3 to 28
0.20 to 10
0.072 to 20
6.4 to 49
3.1 to 18
0.11 to 7.2
0.35 to 18
240 to 620
110 to 410
8.5 to 240
57 to 330
11 to 69
4.5 to 49
0 to 5.9
2.2 to 81
SD (ng/g)
0.76
0.85
0.37
0.59
5.2
3.6
1.5
3.4
7.3
3.4
1.2
4.0
100
61
38
74
12
9.3
1.0
16
RSD (%)
77
130
130
110
34
29
97
63
43
36
87
63
24
24
76
44
38
37
60
100
Below DL (%)
5.1
15
9.9
24
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2.9
0
April 2004
7-5
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 7-2. Normalized (by Lipid Content) Total PCB and frans-Nonachlor
Concentrations in the Lower Pelagic Food Web
-S3
^5
o
o
m
O
Q_
"ro
•5
;
-
-
;
1000-
-
_
-
-
mn
-
><
"C
a
7_
r
I — C
•
T _^_
I
•
i i i i
: N ^ o
r o >< .=•
I -g 2. o
> — CO ^5
j Q3 CD
r 3 3" n>'
S ii ^
r 0 Ui ^
r 3 " ji
O)
oo
o
Ł=
o
o
O
TO
Ł=
O
Ł=
TO
II
O5
00
CD
O)
10000g
1000=
_
-
100=
—
-
10=
-1
A A
T ~r
c • 1
1 i ' 1 ' 1 ' _L
i i
I
_L
Tl N ^ D
f | 1 f
-8 Ą w s
§ 1 T? ^
TT c? cn ^
w
CD
Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile (top of box) results. Bars represent the
results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest edge of the box. Circles represent results
beyond 1.5*IQR from the box. The letters (A - D) above the boxes represent the results of the analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05). Concentration is plotted on a log scale.
7-6
April 2004
-------
PCBs and trans-Nonachlor in the Lower Pelagic Food Web
7.1.2 Seasonal Variation
The lower pelagic food web was sampled in six separate cruises in June 1994, August 1994, October
1994, March 1995, August 1995, and September 1995. Two-way analysis of variance revealed that total
PCB concentrations inMysis spp., zooplankton, and phytoplankton differed significantly by station and
by sampling cruise. Seasonal and geographical variations, however, were small in comparison to
differences due to sample type (phytoplankton, zooplankton, Mysis spp., and Diporeia spp.). Figure 7-3
shows the seasonal variation in total PCB concentrations by sample type. While there were no absolute
seasonal trends in total PCB concentrations, average concentrations across stations were often higher in
the spring and early summer (June 1994 and March 1995) than in the late summer (August 1994, August
1995, and September 1995). Total PCB concentrations in Mysis spp. samples were significantly higher in
June 1994 and March 1995 than in August 1995. In zooplankton samples, total PCB concentrations were
significantly higher in March 1995 and October 1994 than in either of the August cruises (August 1994
and August 1995). In phytoplankton samples, total PCB concentrations were significantly higher in June
1994 than in August 1994. Total PCB concentrations in Diporeia spp. did not differ significantly among
cruises.
fra«5-Nonachlor concentrations in all lower pelagic food web sample types differed significantly by
station and by sampling cruise. Figure 7-4 shows the seasonal variation in trans-nonachlor concentrations
by sample type. Similarly to total PCB concentrations, trans-nonachlor concentrations for some sample
types were often higher in the spring and early summer than in the late summer. In phytoplankton
samples, trans-nonachlor concentrations were significantly higher in March 1995 than in September 1995
and significantly higher in June 1994 than in August 1994, October 1994, August 1995 or September
1995. In zooplankton samples, trans-nonachlor concentrations were significantly higher in March 1995
than in all other cruises. fra«s-Nonachlor concentrations in Mysis spp. samples were significantly higher
in June 1994, October 1994, and March 1995 than in September 1995. Diporeia spp. samples did not fit
the general trend of higher trans-nonachlor concentrations in spring and early summer than in late
summer. The only significant differences between trans-nonachlor concentrations in Diporeia spp.
samples were between the September 1995 and the August 1994 cruises.
Figure 7-3. Seasonal Variation of Total PCB Concentrations Measured in the Lower Pelagic Food
Web of Lake Michigan
u
c
o
O
00
o
0.
1
nJun-94
• Aug-94
rjOct-94
• Mar-95
rjAug-95
• Sep-95
Phytoplankton Zooplankton
Mysis
Diporeia
April 2004
7-7
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 7-4. Seasonal Variation in frans-Nonachlor Concentrations Measured in the Lower Pelagic
Food Web of Lake Michigan
u
c
o
O
u
re
tfc
I
Qjun-94
• Aug-94
DOct-94
• Mar-95
DAug-95
QSep-95
Phytoplankton Zooplankton Mysis
Diporeia
7.1.3 Geographical Variation
Sampling of the lower pelagic food web was focused in the following four biological sampling areas or
biota boxes:
*• Chicago biota box - a station in southern Lake Michigan basin near Chicago
*• Sturgeon Bay biota box - a combination of three stations on the western side of the northern
Lake Michigan basin near Sturgeon Bay, Wisconsin
»• Port Washington biota box - a combination of two stations in the central Lake Michigan basin
near Port Washington, Wisconsin
*• Saugatuck biota box - a series of three stations on the eastern side of the southern Lake
Michigan basin near Saugatuck, Michigan.
In addition to focused sampling in these areas, samples also were collected from six routine monitoring
sites throughout the lake (Table 7-1). Table 7-3 shows the concentrations of total PCBs measured in
lower pelagic food web samples collected from the various sampling locations.
7-8
April 2004
-------
PCBs and trans-Nonachlor in the Lower Pelagic Food Web
Table 7-3. Mean Concentrations of Total PCBs Measured in the Lower Pelagic Food Web at Various
Sampling Locations
Sample Type
Diporeia
Mysis
Phytoplankton
Zooplankton
Sampling Location
Chicago biota box
Sturgeon Bay biota box
Port Washington biota box
Saugatuck biota box
47M
Chicago biota box
Sturgeon Bay biota box
Port Washington biota box
Saugatuck biota box
18M
27M
47M
Chicago biota box
Sturgeon Bay biota box
Port Washington biota box
Saugatuck biota box
18M
23M
27M
41
47M
Chicago biota box
Sturgeon Bay biota box
Port Washington biota box
Saugatuck biota box
18M
27M
47M
19M
N
6
9
11
12
1
6
11
12
12
6
1
5
7
18
12
19
6
1
1
1
6
7
18
12
19
6
1
6
1
Mean
(ng/g)
450
350
440
440
340
200
250
230
300
250
180
250
55
40
41
72
34
44
49
52
35
180
110
180
200
220
140
170
280
Range (ng/g)
290 to 590
250 to 470
290 to 620
240 to 560
NA
110 to 330
180 to 320
190 to 270
190 to 410
190 to 320
NA
160 to 410
35 to 100
14 to 140
22 to 63
14 to 240
8.5 to 88
NA
NA
NA
27 to 46
110 to 270
57 to 270
83 to 300
87 to 310
140 to 270
NA
110 to 330
NA
SD (ng/g)
110
82
110
90
NA
82
52
24
54
45
NA
93
24
27
13
59
28
NA
NA
NA
7.8
62
59
59
66
56
NA
93
NA
RSD (%)
24
23
25
21
NA
40
21
10
18
18
NA
37
43
68
32
82
83
NA
NA
NA
22
35
56
33
33
25
NA
53
NA
Below DL
(%)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
NA = Not applicable. Summary statistics could not be calculated for a single data point.
Among the biota boxes total PCB concentrations were generally highest at the Saugatuck biota box and
lowest at the Sturgeon Bay biota box. Average total PCB concentrations in Mysis spp., phytoplankton,
and zooplankton were higher at the Saugatuck biota box than all other biota boxes, and average total PCB
concentrations in Diporeia spp., phytoplankton, and zooplankton were lower at the Sturgeon Bay biota
box than all other biota boxes. These differences were not statistically significant for all cases, but two-
April2004
7-9
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
way analysis of variance (accounting for sampling station and sampling cruise) revealed that total PCB
concentrations in two of the four lower pelagic food web sample types differed significantly. Total PCB
concentrations in My sis spp. were significantly higher at the Saugatuck biota box than at the Chicago
biota box or Port Washington biota box, and total PCB concentrations in zooplankton were significantly
lower at the Sturgeon Bay biota box than all other stations. This trend is consistent with the distribution
of PCBs in Lake Michigan sediments (see Chapter 6). PCBs accumulated in the eastern side of the
southern Lake Michigan basin, near Saugatuck, and were lower along the western shore and northern
basin, near Sturgeon Bay.
The trend of increased concentrations near the Saugatuck biota box and decreased concentrations near the
Sturgeon Bay biota box that was observed for total PCBs was not observed for trans-nonachlor
accumulation in the lower pelagic food web (Table 7-4). Average fra«s-nonachlor concentrations in
My sis spp. were highest at the Saugatuck biota box, but concentrations in phytoplankton and zooplankton
were highest at the Chicago biota box, and concentrations in Diporeia spp. were highest at the Port
Washington biota box. Average trans-nonachlor concentrations were lowest at the Sturgeon Bay,
Chicago, Port Washington, and Sturgeon Bay biota boxes for Diporeia spp., Mysis spp., phytoplankton,
and zooplankton, respectively. Two-way analysis of variance revealed that differences in fra«s-nonachlor
concentrations among sites were significant for Mysis spp. and zooplankton samples. fra«s-Nonachlor
concentrations in Mysis spp. were significantly higher in Saugatuck and Sturgeon Bay biota boxes than in
the Chicago biota box. fra«s-Nonachlor concentrations in zooplankton were significantly higher in the
Port Washington biota box than the Sturgeon Bay biota box.
The observed trend of trans-nonachlor accumulation in the lower pelagic food web was also consistent
with the geographical distribution of trans-nonachlor in sediments. fra«s-Nonachlor was not
preferentially accumulated in sediments along the eastern side of the southern basin (near the Saugatuck
biota box) as was the case for total PCBs (see Chapter 6). Rather, accumulation of fra«s-nonachlor in
Lake Michigan sediments was concentrated towards the center of the southern and central basins.
Consistent with these findings, fra«s-nonachlor was not generally higher at Saugatuck than at Sturgeon
Bay. In addition, trans-nonachlor in two of the sample types (Diporeia spp. and zooplankton) was higher
at the Port Washington biota box (which is in the center of the lake) than either the Saugatuck or Sturgeon
Bay biota boxes.
7-10 April 2004
-------
PCBs and trans-Nonachlor in the Lower Pelagic Food Web
Table 7-4. Mean Concentrations of frans-Nonachlor Measured in the Lower Pelagic Food Web at Various
Sampling Locations
Sample Type
Diporeia
Mysis
Phytoplankton
Zooplankton
Sampling Location
Chicago biota box
Sturgeon Bay biota box
Port Washington biota box
Saugatuck biota box
47M
Chicago biota box
Sturgeon Bay biota box
Port Washington biota box
Saugatuck biota box
18M
27M
47M
Chicago biota box
Sturgeon Bay biota box
Port Washington biota box
Saugatuck biota box
18M
23M
27M
41
47M
Chicago biota box
Sturgeon Bay biota box
Port Washington biota box
Saugatuck biota box
18M
27M
47M
19M
N
6
9
11
12
1
6
11
12
12
5
1
5
7
18
11
19
6
1
1
1
6
7
15
7
13
1
1
2
1
Mean (ng/g)
31
29
38
30
27
17
27
22
30
28
19
25
2.3
1.6
1.4
1.8
1.8
2.0
3.9
2.3
1.3
24
8.9
21
13
45
9.3
31
34
Range (ng/g)
19 to 49
19 to 51
19 to 69
11 to 46
NA
4.5 to 27
9.5 to 49
12 to 39
13 to 46
21 to 34
NA
20 to 33
1.0 to 3.6
0.60 to 3.7
0.0 to 2.5
0.71 to 3.4
0.46 to 5.9
NA
NA
NA
0.32 to 2.8
2.2 to 81
2.3 to 25
6.3 to 59
2.2 to 29
NA
NA
7.8 to 54
NA
SD (ng/g)
10
10
15
11
NA
7.9
11
7.4
10
5.2
NA
5.4
0.97
0.87
0.89
0.77
2.1
NA
NA
NA
0.86
27
6.7
19
9.8
NA
NA
32
NA
RSD (%)
32
35
40
38
NA
47
40
33
33
18
NA
22
42
54
63
44
120
NA
NA
NA
65
108
75
89
77
NA
NA
105
NA
Below DL (%)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
18
0
0
0
0
0
0
0
0
0
0
0
0
0
0
NA = Not applicable. Summary statistics could not be calculated for a single data point.
April 2004
7-11
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
7.1.4 Bioaccumulation
Persistent organic pollutants, such as PCBs and fra«s-nonachlor, typically accumulate in living organisms
above concentrations found in the water. This accumulation is due to the preferred partitioning of
hydrophobic organic contaminants in organic tissues (such as lipids) over water, uptake from food, and/or
reduced metabolism and elimination of persistent contaminants. The degree of accumulation is often
quantified by a bioaccumulation factor, which is the ratio of the concentration of pollutant in an organism
to the concentration of that pollutant in the water. When pollutants are increasingly accumulated with
each trophic level of a food chain (or biomagnified), a biomagnification factor can be used to quantify the
degree of accumulation from one trophic level to the next. A biomagnification factor is the ratio of the
concentration of pollutant in organisms at a particular trophic level to the concentration of that pollutant
in the next lowest trophic level.
To evaluate the degree of accumulation of PCBs and trans-nonachlor in the lower pelagic food web of
Lake Michigan, bioaccumulation factors were calculated for each sample type (Table 7-5).
Bioaccumulation factors were calculated as the mean concentration in a lower pelagic food web sample
type divided by the lake-wide mean concentration in Lake Michigan. Concentrations of total PCBs in the
lower pelagic food web were generally 105 to 106 times higher than dissolved concentrations of total
PCBs in Lake Michigan water, which averaged 0.18 ng/L (or 0.00018 ng/g assuming a water density of
Ig/mL). Bioaccumulation factors for total PCBs from water to the lower pelagic food web were 2.3 x
106, 1.4 x 106, 2.7 x 105, and 9.3 x 105 for Diporeia spp., Mysis spp., phytoplankton, and zooplankton,
respectively. On a congener-specific basis, bioaccumulation factors were generally lower for the less-
chlorinated PCB congeners and higher for the more-chlorinated congeners. Bioaccumulation factors for
PCB 33 ranged from 3.0 x 104to 1.1 x 105, while bioaccumulation factors for PCB 180 ranged from 2.9 x
106to3.5xl07.
Table 7-5. Bioaccumulation Factors for PCBs and trans-Nonachlor in the Lower Pelagic Food Web
Analyte
PCB 33
PCB 118
PCB 180
Total PCBs
frans-Nonachlor
Diporeia
1.1x105
6.2 x106
3.5 x107
2.3 x106
5.5 x106
Mysis
7.1 x104
5.1 x106
1.9x107
1.4x106
4.4 x106
Phytoplankton
3.0 x104
6.5 x105
2.9 x106
2.7 x105
3.0 x105
Zooplankton
5.8 x104
2.2 x106
1.3x107
9.3 x105
2.8 x106
The accumulation of trans -nonachlor was slightly greater than the accumulation of total PCBs in the
lower pelagic food web. Bioaccumulation factors for trans-nonachlor were 5.5 x 106, 4.4 x 106, 3.0 x 105,
and 2.8 x 106 for Diporeia spp., Mysis spp., phytoplankton, and zooplankton, respectively.
To evaluate the accumulation and transfer of PCBs and fra«s -nonachlor between trophic levels within the
lower pelagic food web, biomagnification factors also were calculated. Biomagnification factors between
primary producers and primary consumers were calculated as the concentration of contaminants in
Diporeia spp., Mysis spp., or zooplankton divided by the concentration in phytoplankton. Total PCB
biomagnification factors were 8.5, 5.1, and 3.4 for Diporeia spp., Mysis spp., and zooplankton,
respectively (Table 7-6). Higher bioaccumulation and biomagnification factors for Diporeia spp. could
be due to specific life history characteristics including this organism's close association with sediments,
which contained approximately 100,000 times the concentration of PCBs in water and slightly higher
concentrations of PCBs than phytoplankton. Bioaccumulation factors for Diporeia spp. compared to
sediments were 8.3 for total PCBs and 59 for trans-nonachlor.
7-12
April 2004
-------
PCBs and trans-Nonachlor in the Lower Pelagic Food Web
Similar to bioaccumulation factors, biomagnification factors increased with increasing chlorination of
PCB congeners. Biomagnification factors ranged from 1.9 to 3.6 for PCB 33, and from 4.5 to 12 for PCB
180 (Table 7-6). For trans-nonachlor, biomagnification factors were 18, 15, and 9.5 in Diporeia spp.,
Mysis spp., and zooplankton, respectively.
Table 7-6. Biomagnification Factors for PCBs and frans-Nonachlor between Primary Producers and Primary
Consumers
Analyte
PCB 33
PCB 118
PCB 180
Total PCBs
frans-Nonachlor
Diporeia 1 Phytoplankton
3.6
9.6
12
8.5
18
Mysis 1 Phytoplankton
2.4
7.9
6.7
5.1
15
Zooplankton / Phytoplankton
1.9
3.4
4.5
3.4
9.5
7.2 Quality Implementation and Assessment
As described in Section 1.5.5, the LMMB QA program prescribed minimum standards to which all
organizations collecting data were required to adhere. The quality activities implemented for the PCBs
and fra«5-nonachlor monitoring portion of the study are further described in Section 2.7 and included use
of SOPs, training of laboratory and field personnel, and establishment of MQOs for study data. A
detailed description of the LMMB quality assurance program is provided in The Lake Michigan Mass
Balance Study Quality Assurance Report (USEPA, 200 Ib). A brief summary of the quality of lower
pelagic food web PCB and fra«s-nonachlor data is provided below.
Quality Assurance Project Plans (QAPPs) were developed by the Pis and were reviewed and approved by
GLNPO. Each researcher trained field personnel in sample collection SOPs prior to the start of the field
season and analytical personnel in analytical SOPs prior to sample analysis. Each researcher submitted
test electronic data files containing field and analytical data according to the LMMB data reporting
standard prior to study data submittal. GLNPO reviewed these test data sets for compliance with the data
reporting standard and provided technical assistance to the researchers. In addition, each researcher's
laboratory was audited during an on-site visit at least once during the time LMMB samples were being
analyzed. The auditors reported positive assessments and did not identify issues that adversely affected
the quality of the data.
As discussed in Section 2.7, data verification was performed by comparing all field and QC sample
results produced by each PI with their MQOs and with overall LMMB Study objectives. Analytical
results were flagged when pertinent QC sample results did not meet acceptance criteria as defined by the
MQOs. These flags were not intended to suggest that data were not useable; rather they were intended to
caution the user about an aspect of the data that did not meet the predefined criteria. Table 7-7 provides a
summary of flags applied to the lower pelagic food web PCB and trans-nonachlor data. The summary
includes the flags that directly relate to evaluation of the MQOs to illustrate some aspects of data quality,
but does not include all flags applied to the data to document sampling and analytical information, as
discussed in Section 2.7. No results were qualified as invalid, thus all results are represented in the
analysis of lower pelagic food web PCB and trans-nonachlor concentrations presented in this report.
April 2004
7-13
-------
Table 7-7. Summary of Routine Field Sample Flags Applied to Select PCB Congeners and frans-Nonachlor in the Lower Pelagic Food Web
Analyte
PCB 33
PCB 118
PCB 180
frans-Nonachlor
Flags
Contamination
FBK
45% (104)
2% (4)
5% (12)
0
Precision
FFD
7% (16)
2% (4)
3% (7)
2% (5)
FDL
4% (9)
0
1%(2)
2% (5)
Bias
FMS
0
0
0
0
FSS
4% (9)
4% (9)
4% (9)
31% (65)
LOB
0
0
0
0
HIB
0
0
0
0
FPC
0
0
0
1%(3)
Invalid
INV
0
0
0
0
The number of routine field samples flagged is provided in parentheses. The summary provides only a subset of applied flags and does not represent the full suite of flags applied
to the data.
FBK = Failed blank (A related blank had a measurable value above the established QC limit when the blank was analyzed using the same equipment and analytical method.
Reported value may be suspect.)
FDL = Failed laboratory duplicate (A laboratory duplicate associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FFD = Failed field duplicate (A field duplicate associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FPC = Failed performance check (A laboratory performance check sample associated with this analysis failed the acceptance criteria. Validity of reported value may be
compromised.)
FSS = Failed surrogate (Surrogate recoveries associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FMS = Failed matrix spike (A matrix spike associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
LOB = Likely biased low (Reported value is probably biased low as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery or other internal
lab QC data. Reported value is not considered invalid.)
HIB = Likely biased high (Reported value is probably biased high as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery, blank
contamination, or other internal lab QC data. Reported value is not considered invalid.)
INV = Invalid (Reported value is deemed invalid by the QC Coordinator.)
7-1
-------
PCBs and trans-Nonachlor in the Lower Pelagic Food Web
Pis used surrogate spikes to monitor the bias of the analytical procedure. The PCB and fra«s-nonachlor
results were corrected for the recoveries of the surrogates. Only 4% of PCB results were qualified for
surrogate recovery problems (Table 7-7). For fra«s-nonachlor, 31% of results were qualified for
surrogate recovery problems (FSS). Surrogate recoveries were below the lower QC bound of 50% in 6%
of samples and above the upper QC bound of 125% in 25% of samples. The mean surrogate recovery for
fra«s-nonachlor, however, was 109%. Laboratory matrix spike samples also were used to monitor
analytical bias, and no results were qualified for failed matrix spikes. Based on an analysis of matrix
spikes, standard reference material recovery, blank contamination, and other internal QC data, the QC
coordinator did not qualify any samples as high or low biased. The QA report (USEPA, 200 Ib),
however, did mention that PCB 99 was prone to chromatographic interference in the plankton media,
which could lead to a potentially high bias for this congener and for total PCB values that contained a
significant proportion from PCB 99.
Field blanks, consisting of glass fiber filters, were collected for PCBs and fra«s-nonachlor analysis. It
was later determined that these glass fiber filter blanks were not representative of the plankton matrix, so
results were not flagged based on the results of field blanks. Laboratory blanks, consisting of a volume of
solvent processed through an empty Soxhlet apparatus in the same fashion used to extract the field
samples, also were prepared and analyzed for PCBs and fra«s-nonachlor. PCB congeners were detected
in all laboratory blanks analyzed. In accordance with the researcher's data qualifying rules, samples were
flagged for a failed blank (FBK) if the mass of the detected congener in the blank was greater than 10% of
the field sample mass or if the blank result was greater than the method detection limit. The level of
contamination varied by PCB congener, with only 2% of PCB 118 results flagged for failed blanks, and
with 45% of PCB 33 results flagged for failed blanks. Congeners 4+10, 31, 33, 44, 81, 87, 114+131,
123+149, 153, 158, and 170+190 were commonly detected in laboratory blanks. None of the field sample
results for fra«s-nonachlor were qualified because of laboratory blank results.
Field duplicates were collected at frequencies of 8%, 17%, 21%, and 11% for Diporeia spp., Mysis spp.,
phytoplankton, and zooplankton, respectively. Laboratory duplicates were prepared and analyzed at a
frequency of 8.1%. In accordance with the researcher's data qualifying rules for field and laboratory
duplicates, samples were flagged for a failed duplicate (FFD or FDL) if the relative percent difference
between duplicate results was greater than 30%. Only a small percentage of results (0 to 7% for trans-
nonachlor and the PCB congeners evaluated) were qualified for failed field or laboratory duplicates.
As discussed in Section 1.5.5, MQOs were defined in terms of six attributes: sensitivity, precision,
accuracy, representativeness, completeness, and comparability. GLNPO derived data quality assessments
based on a subset of these attributes. For example, system precision was estimated as the mean relative
percent difference (RPD) between the results for field duplicate pairs. Similarly, analytical precision was
estimated as the mean RPD between the results for laboratory duplicate pairs. Table 7-8 provides a
summary of data quality assessments for several of these attributes for the lower pelagic food web data.
Because the relative variability of most measurement techniques increases as one approaches the
detection limit of the technique, the assessments of the system and analytical precision were divided into
two concentration regimes. One measure of precision was calculated for those field and laboratory
duplicate results that were less than five times the method detection limit (MDL) of the analyte, and a
separate measure was calculated for those field and laboratory duplicate results that were greater than five
times the MDL.
April 2004 7-15
-------
Table 7-8. Data Quality Assessment for Select PCB Congeners and frans-Nonachlor in Lower Pelagic Food Web Samples
Analyte/Number Field
Samples
PCB 33
(233 samples)
PCB 118
(233 samples)
PCB 180
(233 samples)
frans-Nonachlor
(208 samples)
Parameter
System Precision - Mean Field Duplicate RPD (%), > 5 * MDL
System Precision - Mean Field Duplicate RPD (%), < 5 * MDL
Analytical Precision - Mean Lab Duplicate RPD (%), > 5 * MDL
Analytical Precision - Mean Lab Duplicate RPD (%), < 5 * MDL
Analytical Bias - Mean Lab Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < MDL (%)
System Precision - Mean Field Duplicate RPD (%), > 5 * MDL
Analytical Precision - Mean Lab Duplicate RPD (%), > 5 * MDL
Analytical Bias - Mean Lab Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < MDL (%)
System Precision - Mean Field Duplicate RPD (%), > 5 * MDL
Analytical Precision - Mean Lab Duplicate RPD (%), > 5 * MDL
Analytical Bias - Mean Lab Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < MDL (%)
System Precision - Mean Field Duplicate RPD (%), > 5 * MDL
Analytical Precision - Mean Lab Duplicate RPD (%), > 5 * MDL
Analytical Bias - Mean Lab Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < MDL (%)
Number of QC samples
24 field duplicate pairs
7 field duplicate pairs
12 lab duplicate pairs
5 lab duplicate pairs
20 lab matrix spike samples
-
34 field duplicate pairs
19 lab duplicate pairs
20 lab matrix spike samples
-
33 field duplicate pairs
19 lab duplicate pairs
20 lab matrix spike samples
-
29 field duplicate pairs
17 lab duplicate pairs
14 lab matrix spike samples
-
Assessment
28%
87%
39%
90%
87%
15%
16%
12%
68%
0%
23%
13%
93%
0%
15%
21%
77%
0.5%
RPD = Relative percent difference
MDL = Method detection limit
7-16
-------
PCBs and trans-Nonachlor in the Lower Pelagic Food Web
System precision was relatively consistent among the PCB congeners evaluated. In samples with
concentrations greater than five times the MDL, mean field duplicate RPDs were 28%, 16%, and 23% for
PCB congeners 33, 118, and 180, respectively, indicating good precision. Similarly, the mean field
duplicate RPD for tmns-nonachlor was 15%. For samples with concentrations less than five times the
MDL, precision was reduced, and the mean field duplicate RPD was 87% for PCB 33. For the remaining
congeners presented, all duplicate sample results were greater than five times the MDL. Analytical
precision was similar to system precision, and for two analytes, mean laboratory duplicate RPDs were
higher than mean field duplicate RPDs. This could suggest that the majority of the variability associated
with the measurement system for these analytes is due to the analytical component.
Analytical bias was evaluated by calculating the mean recovery of laboratory matrix spike samples
(LMS). Results indicated a slight low bias overall for all analytes. Mean LMS recoveries were 87%,
68%, 93%, and 77% for PCB 33, PCB 118, PCB 180, and fr«m?-nonachlor, respectively. The PI and QC
coordinator determined, however, that the bias was not strong enough to warrant flagging the data as low
biased (LOB).
Analytical sensitivity was evaluated by calculating the percentage of samples reported below the MDL.
No PCB 118 or PCB 180 results were below the MDL, and only 0.5% of trans-nonachlor results were
below the MDL. For PCB 33, 15% of sample results were reported below the MDL. Results from these
samples were not censored and were used as reported in the analysis of lower pelagic food web
contamination presented in this report.
7.3 Data Interpretation
7.3.1 Comparison to Historical Studies
In this study, total PCB concentrations in the lower pelagic food web averaged 420, 250, 170, and 49 ng/g
for Diporeia spp., Mysis spp., zooplankton, and phytoplankton, respectively. Jackson etal. (1998)
measured similar PCB concentrations in Lake Michigan biota in 1995 and found the same relative degree
of contamination (plankton
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
concentrations of 2.87 to 4.67 ng/L in Green Bay water, which is an order of magnitude higher than
average total PCB concentrations in Lake Michigan (see Chapter 5).
Total PCB concentrations in the lower pelagic food web of Lake Michigan were higher than measured by
other researchers in Swiss lakes (Berglund etal, 2000) and marine pelagic food webs (Harding etal.,
1997; Fisk etal, 2001b). In 19 Swiss lakes, Berglund etal. (2000) measured mean total PCB
concentrations of 28 and 33 ng/g in phytoplankton and zooplankton, respectively. Harding etal. (1997)
measured total PCB concentrations of 0.5 to 147 ng/g in plankton from the southern Gulf of St.
Lawrence, which is lower than the average concentration measured for zooplankton in the LMMB Study.
Fisk et al. (200 Ib) also measured lower total PCB concentrations in high Arctic marine zooplankton,
which averaged 30 ng/g total PCBs.
7.3.2 Seasonal Variation
In the LMMB Study, average PCB concentrations in the lower pelagic food web were often highest in the
spring and early summer (March - June) and lowest in the late summer and fall (August - September).
This finding agrees with the findings of Epplett et al. (2000), who found that concentrations of PCBs in
plankton in Lake Erie varied seasonally, with peaks in the spring or early summer (primarily June) and
decreasing concentrations throughout the summer. In the arctic marine environment, Hargrave et al.
(2000) also found seasonal variations in planktonic total PCB concentrations. Total PCBs were
maximized in the spring and early summer (May/June) and decreased in the late summer and fall
(August/September). Hargrave et al. (2000) concluded that equilibrium occurs rapidly between plankton
and water PCB concentrations. If finite amounts of dissolved PCBs are available for uptake, when
planktonic biomass levels change, there must be a rapid equilibrium reflected in increasing or decreasing
PCB concentrations. Hargrave et al. (2000) observed that the minimum PCB concentrations in plankton
that occurred in July and August corresponded with high particulate organic carbon concentrations
indicative of high production in the planktonic community.
Swackhamer and Skoglund (1993) and Stange and Swackhamer (1994) investigated uptake of PCBs by
several phytoplankton species that were exposed to these contaminants under controlled conditions. The
goals of these two studies included determining if kinetics or equilibrium partitioning of PCBs controlled
the bioaccumulation of hydrophobic organic contaminants such as PCBs. Cultures of phytoplankton were
exposed to a mixture of 40 PCB congeners that included representatives from all 10 levels of chlorination.
Exposures were carried out for 20 and 40 days, respectively, with samples of phytoplankton and water
collected at intervals throughout the study. PCB concentrations were measured in both the phytoplankton
and the water, in order to estimate bioaccumulation rates and bioaccumulation factors (BAFs).
Swackhamer and Skoglund (1993) held separate phytoplankton cultures at 11°C and 20°C to simulate
conditions that result in minimal algal growth (11°C) and average algal growth (20°C). The experiments
at 11 °C were carried out for 20 days, with duplicate samples of both algae and water collected at 0.2
days, 1 day, 3 days, and 20 days. They found that there was a relationship between the uptake of PCBs
and the growth rate of the phytoplankton. Under conditions of minimal growth (11°C), Swackhamer and
Skoglund found that the uptake of PCBs was consistent with equilibrium partitioning between the water
and lipids within the plankton cells. The logs of the calculated BAFs for the PCB congeners exhibited a
linear relationship with the logs of the octanol-water partitioning coefficients (Kow) for the contaminants,
with the more highly chlorinated congeners taking longer to reach equilibrium than the less chlorinated
congeners. However, even at 20 days, most of the congeners did not achieve equilibrium. Swackhamer
and Skoglund noted their results differed from many reports in the literature that suggest that equilibrium
is reached rapidly and that many modeling efforts assume that it is instantaneous.
7-18 April 2004
-------
PCBs and trans-Nonachlor in the Lower Pelagic Food Web
Under the average growth conditions (20°C), however, there was no correlation between log BAF and log
Kow for most of the PCB congeners studied. Under these average growth conditions, only congeners with
log Kow values less than 5.5 (e.g., mono- through trichlorinated congeners) exhibited a correlation
between log BAF and log Kow. Swackhamer and Skoglund offered two possible explanations for the
result under average growth conditions:
1. The kinetics of phytoplankton growth and contaminant uptake are of the same order. Thus, an
increase in biomass (organism growth) dilutes the concentration of PCBs in the organism, resulting in
a constant BAF over time.
2. Cellular metabolism increased during growth, leading to increased excretion of metabolic waste
products, comprised mostly of dissolved organic carbon. Hydrophobic contaminants such as PCBs
may associated with those metabolic waste products, and thus be excreted from the cells.
Stange and Swackhamer (1994) exposed phytoplankton cultures to the PCBs for 40 days at 11°C, in order
to investigate PCB uptake under minimal growth conditions that favored equilibrium partitioning. The
results of that study indicate that PCB uptake under these conditions is controlled by equilibrium
partitioning. Stange and Swackhamer identified an additional factor that may control PCB uptake of the
highly chlorinated congeners. The movement of those congeners through cell membranes (e.g., from the
water into the organism) may depend on their stereochemistry, with PCBs containing three or four
chlorine atoms in the ortho positions able to pass more easily through membranes into the cells. We note
that this additional factor is consistent with the "structure-activity relationship" theory underlying the
designation of 12 of the 209 PCB congeners as "toxic" (see Section 2.1.6).
The implications of the works of Swackhamer and Skoglund (1993) and Stange and Swackhamer (1994)
or the LMMB Study are that the seasonal variations in PCB concentrations in the lower pelagic food web
observed may be the result of the growth rates of phytoplankton species as well as patterns of dominance
of different species during the course of the year. The high PCB concentrations in spring may reflect
equilibrium partitioning processes that occur at colder water temperatures and low light conditions, while
the low PCB concentrations in later summer and fall may reflect the increased growth of organisms in
response to warmer water and increased daylight.
7.3.3 Bioaccumulation and Trophic Transfer
PCBs and trans-nonachlor significantly accumulated in the lower pelagic food web of Lake Michigan
above concentrations in the water column. Bioaccumulation factors from water to the lower pelagic food
web ranged from 105 to 106 for trans-nonachlor and from 104 to 107 for total PCBs depending on the PCB
congener and the compartment (e.g., Diporeia spp., Mysis spp., phytoplankton, or zooplankton). This is
similar to the bioaccumulation factors measured by other researchers. Willman etal. (1999) measured
bioaccumulation factors of 104 to 106 in Green Bay plankton. Oliver and Niimi (1988) also measured
bioaccumulation factors of 103 to 105 for plankton, 104 to 106 for Mysis, and 103 to 106 for amphipods in
Lake Ontario.
Within the lower pelagic food web, PCB and fra«s-nonachlor concentrations differed significantly among
the measured compartments. Concentrations were lowest in phytoplankton, at the base of the pelagic
food web. At the next trophic level (Figure 7-5), zooplankton contained significantly higher levels of
PCBs and fra«s-nonachlor. Total PCB concentrations increased by a factor of 3.4 in the trophic transfer
from phytoplankton to zooplankton, and trans-nonachlor concentrations increased by a factor of 9.5 in
this transfer. Other researchers have also measured significant increases in PCB concentrations from
phytoplankton to zooplankton. Willman et al. (1999) measured biomagnification factors of 1 to 10
between phytoplankton and zooplankton for tetra-, penta-, and hexachlorobiphenyl congeners. Willman
April 2004 7-19
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
et al. (1999) found that the tri-, hepta-, and octachlorobiphenyl congeners accumulated to a lesser degree.
Anderson et al. (1982) found approximately a 14 times increase in PCB concentrations from the
phytoplankton, Cladophora, to zooplankton (>153-|^m size range).
Other researchers have not found evidence of biomagnification of PCBs in the lower pelagic food web
and have suggested that differences in PCB concentrations are explained by factors such as lipid content,
age, size, or depuration rates. Berglund et al. (2000) did not find significant differences in total PCB
concentrations between phytoplankton and zooplankton, and Harding et al. (1997) did not find significant
differences in total PCB concentrations as plankton size varied (presumably accounting for differences
between phytoplankton and zooplankton). If fact, Berglund et al. (2000) noted that when concentrations
were normalized to lipid content, PCB concentrations in zooplankton were lower than in phytoplankton.
In the LMMB Study, the same was true when PCB concentrations were normalized based on lipid
content. Because the lipid content of zooplankton (19%) was substantially higher than for phytoplankton
(4.8%), lipid normalized total PCB concentrations were only 1100 ng/g lipid in zooplankton compared to
1500 ng/g lipid in phytoplankton. In the LMMB Study, however, lipid content did not explain all
observed differences in PCB and fra«s-nonachlor concentrations among the lower pelagic food web
compartments. Lipid-normalized total PCB concentrations were still significantly higher in My sis spp.
and Diporeia spp. than in phytoplankton, and lipid-normalized trans-nonachlor was significantly higher
in zooplankton, My sis spp., and Diporeia spp. than in phytoplankton (Figure 7-2).
In addition to zooplankton, Diporeia spp. and My sis spp. also occupy the second trophic level, however,
trophic transfer from phytoplankton is less direct for these species (Figure 7-5). While Diporeia spp.
feeds on phytoplankton, it is primarily a detrital feeder. My sis spp. is a non-selective filter feeder that
may feed on phytoplankton or zooplankton, such that this species may functionally occupy the second or
third trophic levels. PCB and trans-nonachlor concentrations were significantly higher in My sis spp. and
Diporeia spp. than in phytoplankton or zooplankton on either a dry-weight basis (Figure 7-1) or a lipid-
normalized dry weight basis (Figure 7-2). The higher concentrations of contaminants in Diporeia spp.
may be indicative of trophic transfer from phytoplankton, but more likely is due to this organism's close
association with the more heavily contaminated sediments. Even compared to sediments, however,
Diporeia spp. significantly accumulated organic pollutants. Biota-sediment accumulation factors for
Diporeia spp. were 8.3 for total PCBs and 59 for fra«s-nonachlor. Jackson et al. (1998) also suggested
that Diporeia spp. was more representative of sediment contamination, while plankton were more
representative of water column contamination.
My sis spp. are less associated with detrital sediments than Diporeia spp., and accumulation in this species
may be more directly linked to transfer through phytoplankton and zooplankton. Total PCB
concentrations in My sis spp. exceeded phytoplankton and zooplankton by factors of 5.1 and 1.5,
respectively. fra«s-Nonachlor concentrations in Mysis spp. exceeded phytoplankton and zooplankton by
factors of 15 and 1.6, respectively. Even when normalized to lipid content, Mysis spp. exceeded
phytoplankton and zooplankton PCB concentrations significantly, by factors of 1.8 and 2.4, and lipid-
normalized trans-nonachlor concentrations in Mysis spp. exceeded phytoplankton and zooplankton
significantly by factors of 5.4 and 2.7. These results are similar to the total PCB biomagnification factor
of 6.6 measured between plankton and Mysis by Oliver and Niimi (1988) in Lake Ontario. This factor
varied from approximately 1 to 12 depending upon the specific PCB congener. Fisk et al. (200 la) also
observed biomagnification from zooplankton to predatory invertebrates. Fisk et al. (200 la) calculated a
biomagnification factor of 7.8 for total PCB transfer from Calanus hyperboreus (an herbaceous copepod)
to Themisto libellula (a predatory amphipod), however, the authors noted that differences in
concentrations at this low level of the food chain may be due to differences in organism size and may be
more controlled by concentrations in water than in prey. The same factors could be controlling
bioaccumulation of PCBs in Mysis in this study, since Mysis were considerably larger than the
zooplankton species that were collected.
7-20 April 2004
-------
PCBs and trans-Nonachlor in the Lower Pelagic Food Web
Figure 7-5. Lower Pelagic Food Web Structure and Biomagnification Factors for Total PCBs (BMFp)
and frans-Nonachlor (BMFr)
Mysis
BMFp =1.5
BMFT=1.6
Diporeia
BMFp = 8.5
BMFT=18
BMFp = 5.1
BMFT=15
Zooplankton
BMFp = 3.4
BMFT = 9.5
Phytoplankton
7.3.4 Fractionation
Throughout the lower pelagic food web, PCB congeners were accumulated differentially, with more
highly chlorinated and more lipophilic congeners accumulated to a greater extent. For the three
congeners specifically highlighted in this study (PCB 33, PCB 118, PCB 180), bioaccumulation factors
generally increased by an order of magnitude from PCB 33 (a trichlorobiphenyl) to PCB 118 (a
pentachlorobiphenyl) and another order of magnitude from PCB 118 to PCB 180 (a heptachlorobiphenyl)
(Table 7-5). This differential accumulation of PCB congeners from water is expected based on the
increasing octanol-water partition coefficients with increasing PCB chlorination. Researchers have
described this relationship with linear regressions of log bioaccumulation factors versus log octanol-water
partition coefficients (Mackay, 1982; Oliver and Niimi, 1988).
Not only did bioaccumulation factors increase with increasing chlorination of PCB congeners, but
biomagnification factors also increased with increasing PCB chlorination. For example, biomagnification
factors from phytoplankton to zooplankton increased from 1.9 to 3.4 to 4.5, for PCB congeners 33, 118,
and 180, respectively (Table 7-6). While trends of increasing bioaccumulation factors and increasing
biomagnification factors were observed for the three congeners specifically highlighted in this study,
these trends were also generally true for all PCB congeners. Figure 7-6 shows the relative percentages of
individual PCB congeners in water, phytoplankton, zooplankton, and Mysis spp. In water, there is a
predominance of the less-chlorinated PCB congeners, and this predominance shifts progressively to more-
chlorinated congeners with increasing trophic levels.
April 2004
7-21
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 7-6. Mean Percentage of Individual PCB Congener Contribution to Total
PCB Concentrations
PCB Congener
The shift from less to more-chlorinated PCB congeners with increasing trophic level can be more easily
observed when PCB congeners are grouped by chlorination level homologs (e.g., di-, tri-, tetra-, penta-,
hexa-, hepta-, and octachlorobiphenyls). Figure 7-7 shows the percent change in the relative contribution
(to total PCB concentration) of PCB congener homologs between compartments. Positive percent
7-22
April 2004
-------
PCBs and trans-Nonachlor in the Lower Pelagic Food Web
changes indicate the relative enrichment of that homolog group between compartments, and negative
percent changes indicate the relative depletion of that homolog group between compartments. In transfers
from water to phytoplankton, the di- and trichlorobiphenyls are reduced in relative proportion to total
PCBs, while the penta-, hexa-, and heptachlorobiphenyls are enriched. The same general trend is seen in
the transfer from phytoplankton to zooplankton and the transfer from zooplankton to My sis.
Other researchers have observed this same trend. Jackson et al. (1998) found a relative shift from less-
chlorinated PCB congeners in the plankton to more-chlorinated PCB congeners in Mysis relicta and
Diporeia spp. Oliver and Niimi (1988) also observed differential PCB fractionation from water to
plankton to mysids. Less-chlorinated PCB congeners comprised a higher fraction of total PCBs in water
than in higher trophic levels. Similar to the LMMB Study, Willman et al. (1997) found that the penta-,
hexa-, and heptachlorobiphenyl congeners were enriched relative to other congeners as PCBs moved to
higher trophic levels from sediments to plankton to fish.
Figure 7-7. Relative Enrichment or Depletion of PCB Congener Homolog Groups between Compartments
15
• water to phyto
D phyto to zoo
D zoo to Mysis
4567
PCB Congener Homolog Group
8
9
April 2004
7-23
-------
Chapter 8
PCBs/trans-Nonachlor in Fish
8.1 Results
Lake Michigan fish were collected from April 1994 through November 1995 for PCB and trans-
nonachlor analysis. Forage fish species (alewife, bloater chub, deepwater sculpin, slimy sculpin, and
rainbow smelt) and piscivorous fish species (lake trout and coho salmon) were collected and analyzed.
Alewife and bloater chub were collected in two distinct size classes, and coho salmon were collected in
three distinct age classes. Overall, a total of 796 fish samples were collected for the analysis of PCBs and
fra«5-nonachlor (Table 8-1). Each sample was a composite of up to five fish of the same species and size
or age category. With the exception of coho salmon, fish were collected from the following three
biological sampling areas or biota boxes:
* Sturgeon Bay biota box - a series of three nearshore stations on the western side of the northern
Lake Michigan basin near Sturgeon Bay, Wisconsin
*• Port Washington biota box - a series of two mid-lake reef stations in the central Lake Michigan
basin near Port Washington, Wisconsin
*• Saugatuck biota box - a series of three nearshore stations on the eastern side of the southern
Lake Michigan basin near Saugatuck, Michigan.
Coho salmon were collected from various sites selected to follow the seasonal migration of coho, which
travel up Lake Michigan tributaries in the fall to spawn. During the summer, coho salmon were collected
from the east central and west central regions of the lake. During the fall, coho salmon were collected
from the northeastern side of the lake near the Platte River and on the western side of the lake near the
Kewaunee River.
As noted in Chapter 2, there are 209 possible PCB congeners, and the investigators in this study reported
results for 65 to 110 of these congeners, depending on the capabilities of each laboratory. The USGS
laboratory determined results for 80 congeners or co-eluting congeners.
For the purposes of this report, we are presenting summaries of the results for the following subset of all
of the analytes:
• PCB congener 33
• PCB congener 118
• PCB congener 180
• Total PCBs
• fra«5-nonachlor
Apr/72004 8-1
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 8-1. Number of Fish Samples Analyzed for PCB Congeners and frans-Nonachlor
Species/Size Category
Alewife< 120mm
Alewife> 120mm
Bloater< 160mm
Bloater> 160mm
Coho-Adult
Coho- Hatchery
Coho-Yearling
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
Sampling Dates
05/1 8/94 to 10/1 2/95
05/1 8/94 to 10/1 2/95
05/1 8/94 to 10/1 3/95
05/1 8/94 to 10/1 3/95
05/1 0/94 to 11/06/95
04/21/94 to 04/27/94
10/1 8/94 to 11/1 6/94
05/1 8/94 to 10/1 3/95
05/1 2/94 to 10/26/95
05/1 8/94 to 10/1 2/95
05/1 8/94 to 10/26/95
Total
Number of Samples Analyzed for PCBs and frans-Nonachlor
60
70
70
67
54
5
8
74
246
73
69
796
8.1.1 Species Variation
Tables 8-2 and 8-3 show the mean concentrations (on a wet-weight basis) of PCB 33, PCB 118, PCB 180,
total PCBs, and trans-nonachlor in various Lake Michigan fish species. PCB and fra«s-nonachlor
concentrations differed significantly among species (Figure 8-1). Significantly higher levels of total
PCBs and fra«s-nonachlor were observed in Lake trout, a top predator in the Lake Michigan pelagic food
web, than in any other fish species. Mean concentrations of PCB 33, PCB 118, PCB 180, total PCBs, and
tmns-nonachlor in lake trout were 1.4, 3.3, 3.4, 3.6, and 2.9 times higher than for any other species. This
trend was similar for dry-weight basis PCB and fra«s-nonachlor concentrations (Table 8-4). Mean dry-
weight basis total PCB concentrations in lake trout were from 1.2 to 16 times higher than in other species,
and mean dry-weight basis fra«s-nonachlor concentrations were 2.4 to 34 times higher in lake trout than
in other species.
When PCB and fra«s-nonachlor concentrations were compared among fish species on a lipid-normalized
basis, lake trout still contained higher levels of contamination than all other species with the exception of
adult coho salmon. Mean lipid-normalized total PCB and fra«s-nonachlor concentrations were highest in
adult coho salmon and second highest in lake trout. Lipid-normalized total PCB and fra«s-nonachlor
concentrations in these two top predator fish species were significantly higher than in any of the forage
fish species (Figure 8-2 and Table 8-5). The higher mean concentrations of lipid-normalized
contaminants in adult coho salmon were due to the relatively low lipid content in this species. Lipid
content in adult coho salmon averaged only 4% compared to 16% in lake trout. Of the species analyzed
in this study, only smelt contained lower lipid content (3.6%) than adult coho salmon.
The lowest total PCB and trans-nonachlor concentrations on a wet-weight, dry-weight, or lipid-weight
basis were consistently found in hatchery and yearling coho salmon. This species is raised in hatcheries
and annually stocked in Lake Michigan. Hatchery samples consisted of immature coho collected directly
from the Platte River hatchery, and yearling samples consisted of immature coho collected in Lake
Michigan. The reduced contamination in these sample types most likely reflects both the young age of
the fish and reduced contaminant exposure from hatchery food and water sources.
8-2
April 2004
-------
PCBs and trans-Nonachlor in Fish
Table 8-2. Mean Concentrations of Specific PCB Cone
Congener
PCB 33
PCB 118
PCB180
Species/Size Category
Alewife< 120mm
Alewife> 120mm
Bloater< 160mm
Bloater> 160mm
Coho-Adult
Coho- Hatchery
Coho-Yearling
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
Alewife< 120mm
Alewife> 120mm
Bloater< 160mm
Bloater> 160mm
Coho-Adult
Coho- Hatchery
Coho-Yearling
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
Alewife< 120mm
Alewife> 120mm
Bloater< 160mm
Bloater> 160mm
Coho-Adult
Coho- Hatchery
Coho-Yearling
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
N
60
70
69
67
54
5
8
74
246
72
68
60
70
70
67
54
5
8
74
246
73
69
59
70
70
67
54
5
8
74
246
73
69
Mean (ng/g)
8.3
23
3.9
4.0
19
0.0
0.0
0.16
33
1.7
2.0
8.3
23
28
39
36
6.5
7.7
34
130
16
22
5.6
15
25
29
25
2.4
7.7
29
100
9.0
19
eners in Lake Michigan Fish (Wet-weight Basis)
Range (ng/g)
0.0 to 49
0.0 to 83
0.0 to 36
0.0 to 36
0.0 to 65
0.0 to 0.0
0.0 to 0.0
0.0 to 7.4
0.0 to 230
0.0 to 18
0.0 to 19
1.4 to 23
12 to 34
8.2 to 56
16 to 76
5.7 to 82
5.4 to 8.7
4.5 to 19
4.1 to 110
4.2 to 790
7.9 to 28
0 to 42
1.1 to 12
7.7 to 25
7.8 to 45
15 to 63
4.2 to 50
2.0 to 3.5
5.4 to 16
5.9 to 83
8.2 to 490
4.3 to 14
7.2 to 52
SD (ng/g)
9.6
17
9.5
9.5
20
0.0
0.0
1.0
50
4.1
4.5
4.9
5.2
9.9
12
24
1.3
4.8
21
100
4.8
8.7
2.9
3.4
8.6
9.1
16
0.62
3.5
15
80
2.8
7.2
RSD (%)
120
75
240
240
110
-
-
620
150
240
230
59
22
35
32
68
21
62
61
77
31
40
53
23
35
32
65
26
45
51
78
31
39
Below DL (%)
45
13
84
82
37
100
100
97
50
83
82
0
0
0
0
0
0
0
0
0
0
1.4
0
0
0
0
0
0
0
0
0
0
0
April 2004
8-3
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 8-3. Mean Concentrations of Total PCBs and frans-Nonachlor in Lake Michigan Fish (Wet-weight
Basis)
Analyte
Total
PCBs
frans-
Nonachlor
Species/Size Category
Alewife< 120mm
Alewife> 120mm
Bloater< 160mm
Bloater> 160mm
Coho-Adult
Coho- Hatchery
Coho-Yearling
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
Alewife< 120mm
Alewife> 120mm
Bloater< 160mm
Bloater> 160mm
Coho-Adult
Coho- Hatchery
Coho-Yearling
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
N
60
70
70
67
54
5
8
74
246
73
69
60
70
70
67
54
5
8
74
246
73
69
Mean (ng/g)
250
580
650
830
810
120
200
420
3000
310
430
12
28
48
63
38
3.5
11
50
180
20
38
Range (ng/g)
58 to 750
320 to 910
230 to 1000
430 to 1400
130 to 1700
97 to 170
120 to 400
110 to 1200
180 to 15000
170 to 550
160 to 820
2.5 to 32
14 to 52
17 to 85
27 to 110
5.6 to 76
2.8 to 4.4
7.4 to 17
9.9 to 99
12 to 680
9.1 to 37
15 to 72
SD (ng/g)
150
140
180
210
520
27
90
200
2300
83
130
7.3
9.0
15
19
25
0.60
2.6
17
120
6.0
14
RSD (%)
60
24
28
25
64
22
46
48
76
27
30
59
32
31
30
65
17
23
34
65
30
37
Below DL (%)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8-4
April 2004
-------
PCBs and trans-Nonachlor in Fish
Figure 8-1. Total PCB and frans-Nonachlor Concentrations in Lake Michigan Fish
(Wet-weight Basis)
o
o
CQ
o
Q_
100000=,
10000E
1000=
100=
10
.0)
D)
o
O
o
ro
1000=,
100=
§
EF
DE
}>
i"
i?
A
—^
o
3
}>
i
i?
V
—^
o
3
DO
O
SI
5T
A
m
o
3
i
DO
O
91
5T
V
m
o
3
O
o
?
o
C
II
yi
o
o
~r
Q)
(D
-3
1
o
o
?,
(D
Q)
5'
(Q
D
OT
o
c
•g
3'
ii
i—
Q)
CD
H
O
II
OT
O)
o
c
•g
3
1
O>
3
(D_
^
II
~y
ii
23
II
en
ii
oo
EF
1
ST
S
(F
A
§
3
-^
(D"
s
t?
V
§
3
DO
O
SI
ST
A
8
3
3
i
DO
O
91
ff
V
8
3
3
O
o
o
i:
"5~
ii
01
O
o
o
SI
o>
^
O
o
o
3
(Q
D
O)
o
c
•g
^
ii
,-!-
Q)
(D
H
0
&
^
K)
OT
OT
o
c
-g
^
II
O5
rn
OT
3
-i
ii
II
o
II
en
ii
oo
Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile (top of box) results. Bars represent the
results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest edge of the box. Circles represent results
beyond 1.5*IQR from the box. The Xs represent results beyond 3*IQR from the box. The letters (A - G) above the boxes represent the results
of the analysis of variance and multiple comparisons test. Boxes with the same letter were not statistically different (at alpha = 0.05).
Concentration is plotted on a log scale.
April 2004
8-5
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 8-4. Mean Concentrations of Total PCBs and frans-Nonachlor in Lake Michigan Fish (Dry-weight Basis)
Analyte
Total
PCBs
frans-
Nonachlor
Species/Size Category
Alewife< 120mm
Alewife> 120mm
Bloater< 160mm
Bloater> 160mm
Coho-Adult
Coho- Hatchery
Coho-Yearling
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
Alewife< 120mm
Alewife> 120mm
Bloater< 160mm
Bloater> 160mm
Coho-Adult
Coho- Hatchery
Coho-Yearling
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
N
60
70
70
67
54
5
8
74
246
73
69
60
70
70
67
54
5
8
74
246
73
69
Mean (ng/g)
990
2100
2500
6700
2900
480
690
1700
7800
1400
1700
50
110
190
200
140
14
38
200
480
87
150
Range (ng/g)
230 to 3000
11 00 to 3500
930 to 4000
1300 to 5000
570 to 6000
370 to 710
400 to 1700
440 to 41 00
770 to 37000
660 to 2400
620 to 3600
9.8 to 130
52 to 230
69 to 310
110 to 390
24 to 300
11 to 19
29 to 67
40 to 380
48 to 1700
40 to 170
60 to 290
SD (ng/g)
580
580
750
760
1700
130
410
650
5400
390
530
31
38
59
61
82
2.9
12
69
280
28
55
RSD (%)
58
27
30
28
59
27
60
40
70
28
32
62
36
32
30
60
21
31
34
58
32
38
Below DL (%)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8-6
April 2004
-------
PCBs and trans-Nonachlor in Fish
Figure 8-2. Total PCB and frans-Nonachlor Concentrations in Lake Michigan Fish (Lipid-
weight Basis)
1000000|
^ 100000=
c
c 10000-
O
m
O
- 1000=
.2 -
"5 I
100-
FGH
T
1
i
^
0"
0s
A
i\j
o
3
•^
ii
O3
O
CD
T
1
^>
0"
0*
V
i\j
o
3
•^
ii
o
A
C CDE |
JT • |_
°
i i i
DO DO O
0 0 °
H. H. 0
00i
A V ^
O3 O3 §:
O O ^
3 3 ?
Zi Zi
=> => 01
"5~ "5~ "^
II II
^J CD
O -vl
B
f
H EFG JL, °.EF f
G
rfjg
o
0
0
~r
tu
j"
0
^
3~
ii
Ul
T g
1
O
j- — I t r^
? 1 TT
D C O) O)
o IB • q
? f? * g* 1
- — I c J-Z,
0" "5. o "5. ^
tu
^ '
3
(Q
ii
-25
3 C 3 "
T" ' — s 'T* tO
M ^ M ^
~~1 Ki CD
S ^ CD
^ -P^ -~^
O3
D)
Ł=
O
Ł=
O
O
O
TO
O
10000=,
1000=
100=
10
1
1
A
K>
O
3
^>
I
V
K>
O
3
DO
0
Q)
of
7
O3
O
3
3
DO
0
Q)
of
O3
O
3
3
O
0
o
_
3~
ii
01
O
0
o
i
Q)
O
0
„
I
O
0
o
0
03
3
(Q
D
O)
O
C
T3.
-*
II
1
03
5"
3
^
r-h
||
OT
OT
O
C
T3.
II
§
0_
^
II
II
Ol
Boxes represent the 25th percentile (bottom of box), 50th percentile (center line), and 75th percentile (top of box) results. Bars represent the
results nearest 1.5 times the inter-quartile range (IQR=75th-25th percentile) away from the nearest edge of the box. Circles represent results
beyond 1.5*IQR from the box. The letters (A - H) above the boxes represent the results of the analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05). Concentration is plotted on a log scale.
April 2004
8-7
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Table 8-5. Mean Concentrations of Total PCBs and frans-Nonachlor in Lake Michigan Fish (Lipid-weight
Basis)
Analyte
Total
PCBs
frans-
Nonachlor
Species/Size Category
Alewife< 120mm
Alewife> 120mm
Bloater< 160mm
Bloater> 160mm
Coho-Adult
Coho- Hatchery
Coho-Yearling
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
Alewife< 120mm
Alewife> 120mm
Bloater< 160mm
Bloater> 160mm
Coho-Adult
Coho- Hatchery
Coho-Yearling
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
N
60
70
70
67
54
5
8
74
246
73
69
60
70
70
67
54
5
8
74
246
73
69
Mean (ng/g)
6800
12000
10000
7900
27000
2500
3700
5700
17000
11000
6800
350
630
760
580
1200
73
180
710
1100
670
600
Range (ng/g)
880 to 19000
2300 to 59000
41 00 to 25000
31 00 to 25000
7300 to 130000
1700 to 4300
1400 to 13000
2800 to 13000
4000 to 77000
2600 to 32000
3000 to 17000
60 to 1100
130 to 3700
270 to 1800
280 to 1300
370 to 6500
59 to 110
110 to 530
370 to 2 100
340 to 3300
190 to 2000
190 to 1400
SD (ng/g)
5200
12000
4100
3700
22000
1000
4300
2300
9800
6700
2700
290
680
320
200
1000
23
150
290
490
440
240
RSD (%)
76
98
40
46
82
40
120
40
57
64
39
82
110
42
35
84
32
83
41
45
65
41
Below DL (%)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
8.1.2 Factors Affecting Contaminant Concentrations
In general, log-transformed total PCB and fra«s-nonachlor concentrations were highly correlated with
both lipid content and fish length (Table 8-6). For smelt and slimy sculpin, these correlations were not
significant (at the 95% confidence level) or were weak, and correlations with lipid content were weak in
alewife. For all other species, however, correlations were highly significant (p<0.0001) and r2 values
ranged from 0.27 to 0.89 for correlations between fish contaminant concentration and length and from
0.10 to 0.69 for correlations between fish contaminant concentration and lipid content. It should be noted
that analyzed fish samples were composites of up to five individual fish. Correlations with fish length
reflect the midpoint of the range offish lengths that were incorporated into the composite sample. It is
likely that correlations between contaminant concentrations and fish length would be stronger had
contaminant concentrations been measured in individual fish samples to allow for direct comparison of
length and contaminant concentration.
Multiple regression analysis was conducted to partition the effects of length and lipid content on fish
contamination (Table 8-7). Contaminant concentrations (log-transformed total PCB and log-transformed
8-8
April 2004
-------
PCBs and trans-Nonachlor in Fish
trans-nonachlor) in alewife and lake trout were significantly affected only by length and not by lipid
content. Lipid content remained correlated with contaminant concentrations (Table 8-6) through its
correlation with fish length, but in a multiple regression model, only length significantly accounted for
variability in alewife and lake trout contaminant concentrations. fra«s-Nonachlor concentrations in smelt
also were only affected by length, and not lipid content. In slimy sculpin, contaminant concentrations
(both PCB and fra«s-nonachlor) were significantly affected only by lipid content and not by fish length
(Table 8-7). For the remaining fish species, both length and lipid content or an interaction of the two
parameters significantly affected fish contaminant concentrations.
Table 8-6. Correlation Between Log-transformed Total PCBs and frans-Nonachlor Concentrations in Lake
Michigan Fish and Fish Length and Lipid Content
Analyte
Total
PCBs
frans-
Nonachlor
Species/Size
Category
Alewife
Bloater
Coho-Adult
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
Alewife
Bloater
Coho-Adult
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
Fish Length
Correlation
Coefficient
0.78
0.54
0.84
0.56
0.94
0.40
0.14
0.80
0.52
0.82
0.72
0.91
0.27
0.055
P
O.0001
O.0001
O.0001
0.0001
0.0001
0.0004
0.25
O.0001
O.0001
O.0001
O.0001
0.0001
0.021
0.65
A2
0.61
0.29
0.70
0.32
0.89
0.16
0.020
0.63
0.27
0.66
0.52
0.82
0.07
0.0031
Lipid Content
Correlation
Coefficient
0.25
0.31
0.73
0.62
0.83
-0.34
0.32
0.25
0.47
0.74
0.57
0.82
-0.20
0.311
P
0.0042
O.0001
O.0001
0.0001
0.0001
0.0035
0.017
0.0042
O.0001
O.0001
O.0001
0.0001
0.098
0.020
?
0.062
0.10
0.53
0.39
0.69
0.11
0.10
0.062
0.22
0.55
0.32
0.67
0.038
0.10
Contaminant concentrations generally increased with increasing lipid content and with increasing fish
length. Hydrophobic organic contaminants such as PCBs and fra«s-nonachlor preferentially concentrate
in the fatty tissues of organisms, so those organisms with higher lipid content are expected to contain
more of these contaminants. Older fish also are likely to accumulate higher levels of contaminants
because they have experienced longer exposure durations to environmental contaminants. As a surrogate
offish age, fish length is similarly correlated with fish contaminant concentrations. Contaminant
concentrations generally increased exponentially with increasing fish length, producing a linear
relationship between fish length and log concentration. Figure 8-3 shows the relationship between fish
length and contaminant concentrations in lake trout. The length of lake trout accounted for 89% (r2 =
0.89) of the variability in total PCB concentrations and 82% (r2 = 0.82) of the variability in trans-
nonachlor concentrations. Significant relationships between fish length and contaminant concentrations,
such as the one depicted in Figure 8-3 for lake trout, can be useful in setting and evaluating size-based
fish advisory levels.
April 2004
8-9
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 8-3. Relationship between Fish Length and Total PCB and frans-Nonachlor
Concentrations in Lake Michigan Lake Trout
log Y = 0.0020 X + 2.1
log Y = 0.0016 X+ 1.2
200 400 600 800
Fish Length (mm)
1000
Table 8-7. Results of Multiple Regression Significance Test for Effects of Fish Length and Lipid Content on
Concentrations of Total PCBs and frans-Nonachlor Concentrations in Lake Michigan Fish
Analyte
Total PCBsc
frans-Nonachlor0
Species/Size Category
Alewife
Bloater
Coho-Adult
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
Alewife
Bloater
Coho-Adult
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
p-value
Fish Length
<0.0001
<0.0001b
<0.0001b
0.0014
<0.0001
0.0019
0.080
O.0001
0.0004
<0.0001b
0.0001 b
0.0001
0.042
0.31
Lipid Content
0.43
0.014b
0.0002b
0.0001
0.33
0.016
0.011
0.35
0.036
O.0001b
0.0043b
0.45
0.21
0.014
Interaction3
NS
0.0058
0.0033
NS
NS
NS
NS
NS
NS
0.0006
0.015
NS
NS
NS
a NS indicates that the interaction term was not significant and was removed from the model.
b Due to the significant interaction term, interpretation of the effect of this variable on PCB and frans-nonachlor concentrations is
confounded by the remaining variable.
c Total PCB and frans-nonachlor concentrations were log transformed.
8-10
April 2004
-------
PCBs and trans-Nonachlor in Fish
8.1.3 Geographical Variation
With the exception of coho salmon, fish were collected from three biological sampling areas or biota
boxes (Saugatuck, Sturgeon Bay, and Port Washington) during the spring, summer, and autumn months.
Mean total PCB and fra«s-nonachlor concentrations in alewife and lake trout were highest at the
Saugatuck biota box, and concentrations in bloater chub, deepwater sculpin, and slimy sculpin were
highest at the Port Washington biota box (Figure 8-4). Total PCB concentrations in smelt were highest at
Sturgeon Bay, and fra«s-nonachlor concentrations in smelt were highest at Port Washington. For a given
species, mean contaminant concentrations differed by 6 to 46% among sampling stations.
Figure 8-4. Total PCB and frans-Nonachlor Concentrations in Fish from Three Biological
Sampling Stations in Lake Michigan
10000
D)
o
O
00
o
a.
1
1000
100
AAB
^>
y
«±A
D Sturgeon Bay
• Port Washington
D Saugatuck
1000
)
o
o
o
ro
m
c
us
100
D Sturgeon Bay
• Port Washington
D Saugatuck
//
Bars with the same letter were not statistically different (at alpha = 0.05). Bars with an asterisk indicate that there was significant interaction
between the effects of station and season.
April 2004
8-11
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Two-way analysis of variance (accounting for sampling station and season) revealed that for some
species, differences in contaminant concentrations among sampling stations were statistically significant.
Geographical trends in total PCB contamination, however, were species specific, with some species
containing higher contamination at Port Washington and other species containing higher contamination at
Saugatuck (Figure 8-4). Deepwater sculpin and small bloater chub from Port Washington contained
significantly higher levels of PCBs than the same species collected at Sturgeon Bay or Saugatuck, and
large bloater chub from Port Washington contained significantly higher levels of PCBs than the same
species collected at Sturgeon Bay. This trend was reversed, however, in lake trout. Lake trout from Port
Washington contained significantly lower levels of PCBs than lake trout from Sturgeon Bay or
Saugatuck. There were no significant differences among sites in slimy sculpin contamination. For the
remaining species, there was significant interaction between the effects of station and season, meaning
that significant differences between stations were only observed during given seasons. In the spring,
small alewife from Sturgeon Bay were significantly lower in PCBs than alewife from Port Washington or
Saugatuck. In spring, smelt from Sturgeon Bay contained significantly less PCBs than smelt from Port
Washington. In autumn, Port Washington smelt were significantly lower in PCBs than Sturgeon Bay or
Saugatuck smelt.
Total PCBs measured in the LMMB Study were highest in lake trout from Saugatuck, and lowest in lake
trout from Port Washington. This geographical pattern of PCB contamination in Lake Michigan lake
trout was previously reported by Madenjian et al. (1999a) and also observed by Miller et al. (1992).
Miller et al. (1992) found that lake trout at deep water reef locations, such as Port Washington, contained
lower PCB contamination than lake trout collected from nearshore locations, such as Saugatuck and
Sturgeon Bay. Using LMMB Study data, Madenjian et al. (1999a) identified significantly higher
concentrations of PCBs in lake trout from Saugatuck than from Sturgeon Bay or Port Washington and
significantly higher concentrations of PCBs in lake trout from Sturgeon Bay than from Port Washington.
Madenjian et al. (1999a) explained that these differences in PCB fish contamination levels among various
sites were likely due to differences in organism size or differences in lake trout diet at the sites. Lower
concentrations in lake trout from Port Washington were explained by the fact that the fish from this site
were smaller than fish from the other two sites. Lower concentrations at Sturgeon Bay than at Saugatuck
were explained by a lake trout diet that consisted of a higher proportion of more contaminated prey
species at Saugatuck than at Sturgeon Bay. Based on analyses of guts contents, Madenjian et al. (1999a)
determined that at Saugatuck, lake trout diet consisted of 55% alewife, 35% bloater, and 10% sculpins
and rainbow smelt. This diet contained a combined 0.64 mg/kg PCBs, compared to a 80% alewife and
20% rainbow smelt diet at Sturgeon Bay that contained a combined 0.53 mg/kg PCBs.
Similar to total PCB contamination, geographical trends in fra«s-nonachlor contamination were species
specific (Figure 8-4). Large alewife, bloater chub, and slimy sculpin from Port Washington contained
significantly higher fra«s-nonachlor concentrations than the same species at one or more other stations.
Large alewife and lake trout from Saugatuck contained significantly higher fra«s-nonachlor
concentrations than the same species at one or more other stations. Only slimy sculpin contained
significantly higher fra«s-nonachlor concentrations at Sturgeon Bay than other stations. For small
alewife, deepwater sculpin, and smelt, there was significant interaction between the effects of station and
season. fra«s-Nonachlor contamination in small alewife was significantly higher at Port Washington and
Saugatuck than at Sturgeon Bay during the spring, higher at Sturgeon Bay than Saugatuck during the
summer, and higher at Port Washington than Sturgeon Bay during autumn. In deepwater sculpin, trans-
nonachlor contamination was higher at Port Washington than Saugatuck during the summer. In smelt,
fra«5-nonachlor contamination was significantly higher at Sturgeon Bay than at Port Washington during
the spring and higher at Port Washington than Sturgeon Bay or Saugatuck during autumn.
In summary, differences in fish contamination levels among sites were relatively small (6% to 46%)
compared to species differences, which varied by more than a factor of 12, or differences attributed to fish
8-12 April 2004
-------
PCBs and trans-Nonachlor in Fish
size or lipid content, where contaminant levels varied over an order of magnitude (Figure 8-3).
Significant differences in fish contamination levels among sites were species specific. Some species
showed significantly higher contamination levels at Saugatuck, and other species showed significantly
higher contamination levels at Port Washington. In general, fish contamination levels at Sturgeon Bay
were lower than at the remaining sites. Differences in fish contamination levels among sites could be due
to increased water, sediment, and food contamination levels at specific sites. Contours of water and
sediment PCB levels did show increased concentrations (see Chapters 5 and 6) on the eastern shore of the
southern basin (near Saugatuck) and in the center of the southern basin (near Port Washington biota box).
PCB concentrations in the lower pelagic food web also were generally higher at Saugatuck (see Chapter
7). Differences in fish contamination levels also could be due to differences in fish size or lipid content at
the various sites, or as Madenjian et al. (1999a) suggested, differences in fish diets among the various
sites.
8.1.4 Seasonal Variation
Two-way analysis of variance (accounting for sampling station and season) revealed few significant
differences in contaminant concentrations among the three sampling seasons (spring, summer, and
autumn). No significant differences among season were observed for bloater chub, deepwater sculpin,
and lake trout total PCB concentrations. Total PCB concentrations in slimy sculpin were significantly
higher in summer than in spring. For alewife and smelt, there was significant interaction between the
effects of season and station. At Sturgeon Bay, small alewife contained significantly higher PCB
concentrations during the summer than during autumn, and smelt contained significantly higher PCB
concentrations during spring and summer than during autumn. At Port Washington, large alewife
contained significantly higher PCB concentrations during spring than during summer. At Saugatuck,
small alewife and smelt contained significantly higher PCB concentrations during spring than in summer
or autumn.
For fra«5-nonachlor, significant differences in contaminant concentrations among the three sampling
seasons were observed for some species but not for others. fra«s-Nonachlor concentrations in bloater
chub and lake trout did not differ significantly between seasons. Large alewife contained significantly
higher fra«s-nonachlor concentrations in spring than in summer or autumn, and slimy sculpin contained
significantly higher trans-nonachlor concentrations in summer than in spring. For small alewife,
deepwater sculpin, and smelt, there was significant interaction between the effects of season and station.
At Sturgeon Bay, small alewife, deepwater sculpin, and smelt contained significantly higher trans-
nonachlor concentrations during the spring and summer than during autumn. At Port Washington, small
alewife contained significantly higher trans-nonachlor concentrations during spring than during autumn,
and deepwater sculpin contained significantly higher fra«s-nonachlor concentrations during summer than
spring or autumn. At Saugatuck, small alewife contained significantly higher fra«s-nonachlor
concentrations during spring than in summer or autumn, and smelt contained significantly higher trans-
nonachlor concentrations during spring and summer than autumn.
In conclusion, most fish species did not show significant differences in fish contamination levels among
seasons. Those significant differences that were observed also were relatively small in comparison to
species differences and differences due to fish length or lipid content. As described for site variations,
differences in fish contamination levels among seasons also could be due to differences in fish size or
lipid content during the various seasons or differences in fish diet throughout the changing seasons.
April 2004 8-13
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 8-5. Total PCB and frans-Nonachlor Concentrations in Lake Michigan Fish during
Spring, Summer, and Autumn
10000
O)
o
o
00
o
0.
s
o
1000 -
100
/
1000
u>
6
c
O
o
o
ns
c
O
«F
(0
c
ns
100 -
10 --
AA
A A "
A A A
1
i- -7
^ ^
Bars with the same letter were not statistically different (at alpha = 0.05). Bars with an asterisk indicate that there was significant interaction
between the effects of station and season.
8.1.5 Bioaccumulation
Persistent organic pollutants, such as PCBs and fra«s-nonachlor, typically accumulate in living organisms
above concentrations found in the water. This accumulation is due to the preferred partitioning of
hydrophobic organic contaminants in organic tissues (such as lipids) over water, uptake from food, and/or
reduced metabolism and elimination of persistent contaminants. The degree of accumulation is often
quantified by a bioaccumulation factor, which is the ratio of the concentration of pollutant in an organism
to the concentration of that pollutant in the water. When pollutants are increasingly accumulated with
each trophic level of a food chain (or biomagnified), a biomagnification factor can be used to quantify the
degree of accumulation from one trophic level to the next. A biomagnification factor is the ratio of the
8-14
April 2004
-------
PCBs and trans-Nonachlor in Fish
concentration of pollutant in organisms at a particular trophic level to the concentration of that pollutant
in the next lowest trophic level.
To evaluate the degree of accumulation of PCBs and fra«s-nonachlor in fish species of Lake Michigan,
bioaccumulation factors were calculated for each species (Table 8-8). Bioaccumulation factors were
calculated as the mean dry-weight concentration in fish divided by the lake-wide mean concentration in
Lake Michigan. Concentrations of total PCBs in fish were generally 106 to 107 times higher than
dissolved concentrations of PCBs in Lake Michigan water, which averaged 0.18 ng/L (or 0.00018 ng/g).
Bioaccumulation factors for total PCBs from water to fish ranged from 5.5 x 106 for small alewife
(<120mm) to 4.3 x 107 for lake trout. Bioaccumulation factors were generally lower for the less-
chlorinated PCB congeners and higher for the more-chlorinated congeners. Bioaccumulation factors for
PCB 33 ranged from 6.8 x 104 to 9.4 x 106, while bioaccumulation factors for PCB 180 ranged from 4.7 x
107 to 5.6 x 10s. fra«5-Nonachlor accumulation was of the same magnitude as total PCB accumulation in
fish. Bioaccumulation factors for trans-nonachlor ranged from 8.7 x 106 to 8.3 x 107.
Table 8-8. Bioaccumulation Factors for PCBs and frans-Nonachlor in Lake Michigan Fish
Species/Size Category
Alewife< 120mm
Alewife> 120mm
Bloater< 160mm
Bloater> 160mm
Coho-Adult
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
Bioaccumulation Factor
PCB 33
3.8 x106
9.4 x106
1.8x106
1.5x106
7.0 x106
6.8 x104
9.2 x106
7.9 x105
8.8 x105
PCB 118
1.4x107
3.5 x107
4.4 x107
5.1 x107
5.2 x107
5.5 x107
1.4x108
2.8 x107
3.5 x107
PCB 180
4.7 x107
1.2 x108
2.0 x108
1.9 x108
1.8x108
2.4 x108
5.6 x108
8.2 x107
1.5x108
Total PCBs
5.5 x106
1.2 x107
1.4x107
1.5x107
1.6 x107
9.1 x106
4.3 x107
7.5 x106
9.1 x106
frans-Nonachlor
8.7 x106
1.8x107
3.2 x107
3.5 x107
2.4 x107
3.5 x107
8.3 x107
1.5x107
2.5 x107
To evaluate the accumulation and transfer of PCBs and trans-nonachlor between trophic levels within the
upper pelagic food web, biomagnification factors were calculated. Biomagnification factors were
calculated between forage fish species (alewife, bloater chub, sculpin, and smelt) and piscivorous fish
species (lake trout and coho salmon). Total PCB biomagnification factors from forage fish to piscivorous
fish were 1.6 and 4.2 for coho salmon and lake trout, respectively. fra«s-Nonachlor biomagnification
factors from forage fish to piscivorous fish were 0.96 and 3.4 for coho salmon and lake trout,
respectively. As evidenced by the biomagnification factor of less than one (<1) for coho salmon, trans-
nonachlor was not biomagnified in the trophic transfer from forage fish to coho salmon.
8.2 Quality Implementation and Assessment
As described in Section 1.5.5, the LMMB QA program prescribed minimum standards to which all
organizations collecting data were required to adhere. The quality activities implemented for the PCBs
and trans-nonachlor monitoring portion of the study are further described in Section 2.7 and included use
of SOPs, training of laboratory and field personnel, and establishment of MQOs for study data. A
detailed description of the LMMB quality assurance program is provided in The Lake Michigan Mass
Balance Study Quality Assurance Report (USEPA, 200 Ib). A brief summary of the quality offish PCB
and trans-nonachlor data is provided below.
April 2004
8-15
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Quality Assurance Project Plans (QAPPs) were developed by the Pis and were reviewed and approved by
GLNPO. Each researcher trained field personnel in sample collection SOPs prior to the start of the field
season and analytical personnel in analytical SOPs prior to sample analysis. Each researcher submitted
test electronic data files containing field and analytical data according to the LMMB data reporting
standard prior to study data submittal. GLNPO reviewed these test data sets for compliance with the data
reporting standard and provided technical assistance to the researchers. In addition, each researcher's
laboratory was audited during an on-site visit at least once during the time LMMB samples were being
analyzed. The auditors reported positive assessments and did not identify issues that adversely affected
the quality of the data.
As discussed in Section 2.7, data verification was performed by comparing all field and QC sample
results produced by each PI with their MQOs and with overall LMMB Study objectives. Analytical
results were flagged when pertinent QC sample results did not meet acceptance criteria as defined by the
MQOs. These flags were not intended to suggest that data were not useable; rather they were intended to
caution the user about an aspect of the data that did not meet the predefined criteria. Table 8-9 provides a
summary of flags applied to the fish PCB and fra«s-nonachlor data. The summary includes the flags that
directly relate to evaluation of the MQOs to illustrate some aspects of data quality, but does not include
all flags applied to the data to document sampling and analytical information, as discussed in Section 2.7.
Compared to other matrices, the percentage of results that were qualified for these criteria is relatively
small. No results were qualified as invalid, thus all results are represented in the analysis offish PCB and
fra«5-nonachlor concentrations presented in this report.
Pis used surrogate spikes to monitor the bias of the analytical procedure. The PCB and fra«s-nonachlor
results were corrected for the recoveries of the surrogates. Only 8 to 9% of PCB and fra«s-nonachlor
results were qualified for surrogate recovery problems (Table 8-9). Each of these samples was flagged
for surrogate recoveries that exceeded the upper MQO of 130% recovery. Laboratory matrix spike
samples also were used to monitor analytical bias. Only 2% of PCB 118, PCB 180, and fra«s-nonachlor
samples were flagged for associated failed laboratory matrix spikes that exceeded the upper MQO of
120% recovery. Performance check samples also were used to monitor analytical bias. For PCB 180, 5%
of samples were flagged for associated failed performance check samples, and 20% of fra«s-nonachlor
samples were flagged for associated failed performance check samples. Based on an analysis of matrix
spikes, standard reference material recovery, blank contamination, and other internal QC data, no samples
were qualified by the PI and QC coordinators as high or low biased.
Laboratory blanks of sodium sulfate were used to investigate the possibility of contamination. Corn oil
was added to the laboratory blanks after June 1996 to better represent the fish matrix. No total PCB
results for laboratory blanks were greater than 0.10 [ig, and all field sample results exceeded laboratory
blank concentrations by a factor of 50, therefore no sample results were qualified for failed blanks or
suspected contamination.
Duplicate samples were analyzed to evaluate the precision of analytical results. No field duplicates were
collected for the fish matrix, however, laboratory duplicates were analyzed at a frequency of one per
extraction batch. No results were flagged for duplicate results that exceeded the MQO of a 40% relative
percent difference.
8-16 April 2004
-------
Table 8-9. Summary of Routine Field Sample Flags Applied to Select PCB Congeners and frans-Nonachlor in Fish
Analyte
PCB 33
PCB 118
PCB 180
frans-Nonachlor
Flags
Contamination
FBK
0
0
0
0
Sensitivity
UNO
62% (490)
0.1%(1)
0
0
Precision
FDL
NA
NA
0
0
Bias
FMS
NA
2% (17)
2% (12)
2% (19)
FSS
9% (70)
8% (64)
8% (64)
8% (64)
LOB
0
0
0
0
HIB
0
0
0
0
FPC
NA
NA
5% (40)
20% (158)
Invalid
INV
0
0
0
0
The number of routine field samples flagged is provided in parentheses. The summary provides only a subset of applied flags and does not represent the full suite of flags applied
to the data.
FBK = Failed blank (A related blank had a measurable value above the established QC limit when the blank was analyzed using the same equipment and analytical method.
Reported value may be suspect.)
UNO = Analyte not detected (Analyte produced no instrument response above noise.)
FDL = Failed laboratory duplicate (A laboratory duplicate associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FMS = Failed matrix spike (A matrix spike associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
FSS = Failed surrogate (Surrogate recoveries associated with this analysis failed the acceptance criteria. Validity of reported value may be compromised.)
LOB = Likely biased low (Reported value is probably biased low as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery or other internal
lab QC data. Reported value is not considered invalid.)
HIB = Likely biased high (Reported value is probably biased high as evidenced by LMS (lab matrix spike) results, SRM (standard reference material) recovery, blank
contamination, or other internal lab QC data. Reported value is not considered invalid.)
FPC = Failed performance check (A laboratory performance check sample associated with this analysis failed the acceptance criteria. Validity of reported value may be
compromised.)
INV = Invalid (Reported value is deemed invalid by the QC Coordinator.)
NA = Not applicable. The relevant QC sample (e.g., duplicate, matrix spike, performance check) was not prepared or analyzed for this specific analyte.
8-17
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
As discussed in Section 1.5.5, MQOs were defined in terms of six attributes: sensitivity, precision,
accuracy, representativeness, completeness, and comparability. GLNPO derived data quality assessments
based on a subset of these attributes. For example, analytical precision was estimated as the mean relative
percent difference (RPD) between the results for laboratory duplicate pairs. Table 8-10 provides a
summary of data quality assessments for several of these attributes for fish data.
System precision could not be estimated for the analysis offish tissue data because field duplicates were
not collected for this matrix. While system precision was not estimated, analytical precision was
estimated from the results of laboratory duplicates. Analytical precision for fish PCB and trans-
nonachlor analysis was very good, with RPDs of only 5.3 and 6.9% for duplicate PCB 180 and trans-
nonachlor results, respectively. Laboratory duplicates were not analyzed for PCB 33 or PCB 118.
Analytical bias was evaluated by calculating the mean recovery of laboratory matrix spike samples
(LMS). Analytical bias was very low, with mean LMS recoveries of 99%, 98% and 90% for PCB 118,
PCB 180, and fra«s-nonachlor, respectively. Laboratory matrix spike samples were not analyzed for PCB
33.
Analytical sensitivity was evaluated by calculating the percentage of samples reported below the MDL.
No fra«5-nonachlor or PCB 180 results were below the MDL, and only 0.1% of PCB 118 results were
below the MDL. For less-chlorinated PCB congeners, such as PCB 33, a majority of sample results were
below the MDL. For PCB 33, 81% of 793 samples were reported below the MDL. Results from these
samples were not censored and were used as reported in the analysis offish contamination presented in
this report.
8-18 April 2004
-------
Table 8-10. Data Quality Assessment for Select PCB Congeners and frans-Nonachlor in Fish Sam
Analyte/Number Field Samples
PCB 33
(793 samples)
PCB 118
(796 samples)
PCB 180
(795 samples)
frans-Nonachlor
(796 samples)
Parameter
Analytical Precision - Mean Lab Duplicate RPD (%), > 5 * MDL
Analytical Bias - Mean Lab Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < MDL (%)
Analytical Precision - Mean Lab Duplicate RPD (%), > 5 * MDL
Analytical Bias - Mean Lab Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < MDL (%)
Analytical Precision - Mean Lab Duplicate RPD (%), > 5 * MDL
Analytical Bias - Mean Lab Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < MDL (%)
Analytical Precision - Mean Lab Duplicate RPD (%), > 5 * MDL
Analytical Bias - Mean Lab Matrix Spike Recovery (%)
Analytical Sensitivity - Samples Reported as < MDL (%)
pies
Number of QC samples
-
-
-
-
82 lab matrix spike samples
-
83 lab duplicate pairs
82 lab matrix spike samples
-
83 lab duplicate pairs
82 lab matrix spike samples
-
Assessment
NA
NA
81%
NA
99%
0.1%
5.3%
98%
0%
6.9%
90%
0%
RPD = Relative percent difference
MDL = Method detection limit
NA = Not applicable. Laboratory matrix spike samples were not analyzed for PCB 33 and laboratory duplicates were not analyzed for PCB 33 or PCB 118.
8-19
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
8.3 Data Interpretation
8.3.1 Comparison to Fish Advisory Levels
The Food and Drug Administration has set a tolerance level of 2000 ng/g (2 ppm) for PCBs in fish for
human consumption (21 CFR 109.30). Consistent with this tolerance level, the Great Lakes Fish
Consumption Advisory Task Force has set a fish advisory category of "no consumption" at PCB levels
above 2000 ng/g. Of the Lake Michigan fish analyzed in the LMMB Study, only lake trout contained
PCBs above the 2000 ng/g level. In fact, 56% of lake trout samples exceeded this tolerance level, and the
mean total PCB concentration for Lake Michigan lake trout was 3000 ng/g (or 3 ppm), which is 50%
above the 2000 ng/g tolerance level.
Figure 8-6 shows the percentages of lake trout and coho salmon samples falling into the various Great
Lakes fish advisory categories. No coho salmon or lake trout samples fell into the unrestricted
consumption category. Coho salmon primarily fell into the 1 meal/mo and 6 meals/yr categories. These
categories contained 46% and 44% of coho salmon samples, respectively, with only 9% of coho salmon
samples falling into the 1 meal/wk category. Lake trout primarily fell into the no consumption category
(56%), with only 0.4%, 17%, and 26% in the 1 meal/wk, 1 meal/mo, and 6 meals/yr categories,
respectively.
PCB contamination in fish is positively correlated with fish length, so fish advisories are tied to the size
of the fish collected. Based on the regression equation developed from LMMB Study data (Figure 8-3),
Lake Michigan lake trout above 575 mm were estimated to exceed the 2000 ng/g FDA tolerance level.
Lake trout between 425 and 575 mm would fall into the 6 meals/yr advisory category, and lake trout
below 425 mm would generally fall into the 1 meal/mo advisory category. Only one lake trout sample
contained less than 200 ng/g (0.2 ppm) total PCBs.
Figure 8-6. Percentage of Lake Michigan Coho Salmon and Lake Trout Samples within each
PCB Fish Advisory Category
to
0)
Q.
E
re
(O
II-
o
a)
O)
re
+j
a>
Ł
a>
a.
60
50 -
40 -
30 -
20 -
10 -
QCoho Salmon
• Lake Trout
o, Ł
"~~ -t-J
D) (/)
Ł= p
O
LO
V
C
3
if
o
o
CM
O
LO
8-20
April 2004
-------
PCBs and trans-Nonachlor in Fish
8.3.2 Comparison to Historical Studies
DeVault etal. (1996) and others (Miller et al., 1992; Huestis etal, 1996; Stow etal, 1995) have
observed dramatic declines in PCB concentrations in Great Lakes fish since the early 1970s. In Lake
Michigan lake trout, mean PCB concentrations declined from 23000 ng/g in 1974 to 2590 ng/g by 1986.
Between 1986 and 1992, there was little change in concentrations, with 3490 ng/g observed in 1992.
Total PCB concentrations in lake trout measured in the LMMB Study during 1994 and 1995 fit well with
this trend. Total PCBs in lake trout analyzed in the LMMB Study averaged 3000 ng/g, which is
consistent with the 2590 to 3490 ng/g range of mean total PCB concentrations observed from 1986 to
1992. DeVault et al. (1996) hypothesized that the leveling of lake trout PCB concentrations since 1986
(versus continuing decreases) could be due to significant changes in food web structure during that time
that has led to increased bioaccumulation. DeVault et al. (1996) suggested that the introduction of the
predacious cladoceran, Bythotrephes cederstroemi, in the early 1980s added an additional trophic level in
the pelagic food chain (phytoplankton > zooplankton > Bythotrephes > forage fish > lake trout), thus
increasing bioaccumulation at trophic levels above this insertion.
Hesselberg etal. (1990) measured similar decreases in PCB concentrations in Lake Michigan bloater
from 1976 to 1986. Total PCB concentrations in bloater decreased from 5700 ng/g in 1972 to 1640 ng/g
in 1986. By 1994 and 1995, total PCB concentrations in bloater were half of the 1986 levels. Total PCB
concentrations measured in the LMMB Study averaged 650 ng/g in small bloater chub and 830 ng/g in
large bloater chub.
Similar decreases in total PCB concentrations since the early 1970s have been observed for all Lake
Michigan fish species (Stow et al., 1995) and for fish species in other Great Lakes. DeVault et al. (1996)
reported significant decreases in PCB concentrations in lake trout from Lakes Superior, Huron, and
Ontario. In Lake Ontario, Huestis et al. (1996) observed declines in lake trout PCB levels of 80%
between 1977 and 1993, from 9060 ng/g in 1977 to 1720 ng/g in 1993. In contrast, however, Borgmann
and Whittle (1992) found no significant temporal trend from 1977 to 1988 in total PCB concentrations in
Lake Ontario smelt and sculpin.
fra«5-Nonachlor concentrations in Lake Michigan fish species were not measured routinely prior to 1986
(DeVault etal., 1996). From 1986 to 1992, trans-nonachlor concentrations in Lake Michigan lake trout
averaged 220 to 190 ng/g (DeVault et al, 1996). fra«s-Nonachlor concentrations measured in 1994 and
1995 during the LMMB Study were only slightly lower, at 180 ng/g, indicating that like total PCB
concentrations, trans-nonachlor concentrations in Lake Michigan lake trout have remained relatively
constant since the mid 1980s.
8.3.3 Regional Considerations
Among the Great Lakes, concentrations of PCBs in lake trout were highest in Lake Michigan (DeVault et
al., 1996). From 1986 to 1992, mean PCB concentrations in Lake Michigan lake trout ranged from 2590
to 3490 ng/g. Mean PCB concentrations ranged from 240 to 450 ng/g in Lake Superior trout, 1170 to
1570 ng/g in Lake Huron trout, 2180 to 2890 ng/g in Lake Ontario trout, and 1320 to 2200 ng/g in Lake
Erie walleye (DeVault etal., 1996).
In 1994, Kucklick and Baker (1998) measured PCBs and trans-nonachlor in the Lake Superior food web.
Total PCB concentrations ranged from 130 to 180 ng/g in bloater, from 42 to 52 ng/g in deepwater
sculpin, from 42 to 46 ng/g in slimy sculpin, and from 82 to 160 in lake trout. In Lake Michigan, bloater
chub, deepwater sculpin, slimy sculpin, and lake trout measured in the LMMB Study contained up to 5
times, 23 times, 17 times, and 93 times the total PCB concentrations measured in Lake Superior fish
species, respectively. fra«s-Nonachlor concentrations in Lake Superior ranged from 18 to 27 ng/g in
April 2004 8-21
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
bloater, from 7.6 to 12 ng/g in deepwater sculpin, from 6.9 to 7.6 ng/g in slimy sculpin, and from 9.7 to
21 ng/g in lake trout. In Lake Michigan, fra«s-nonachlor concentrations in these same species ranged
from 4 to 32 times the levels measured in Lake Superior fish.
Total PCB concentrations measured in Lake Michigan fish also were higher than reported levels in fish
from numerous other lakes and water bodies. Harding et al. (1997) measured total PCB concentrations of
155 ng/g wet-weight in the Gulf of St. Lawrence. In 19 Swedish lakes, Berglund et al. (2000) measured a
mean total PCB concentration of 55 ng/g dry-weight. In the relatively remote Lake Tahoe, Datta et al.
(1998) measured 9 to 14 ng/g total PCBs in lake trout. This is up to 1000 times lower than PCB
contamination of lake trout in Lake Michigan.
8.3.4 Factors Affecting Contaminant Concentrations
In the LMMB Study, fish contamination levels were primarily affected by species. Within species,
contaminant levels were significantly affected by fish length, lipid content, location, and season, but the
effect of these factors differed by species. In an evaluation of 20 years of data, Stow (1995) similarly
observed that the variability in fish PCB concentrations were explained by year, species, location, length,
and length/species effects.
For most species, fish length was strongly correlated with contaminant concentration (Table 8-6). This
effect has been commonly reported by other researchers (Stow, 1995; Harding et al., 1997; Huestis et al.,
1996), and is likely due to increased contaminant exposure with increasing fish age. Researchers,
however, have differed on the importance offish lipid content in controlling fish contaminant levels.
Harding et al. (1997) found that the best predictors of PCB contamination in fish were lipid content,
followed by size and age. Kucklick and Baker (1998) also determined that lipid content of organisms in
the Lake Superior food web explained 81% of the variability in wet-weight total PCB concentrations,
with trophic position exerting a smaller influence. The main influence of trophic position on total PCB
concentrations was shown to be due to the concurrent increase in lipid content with trophic position.
In contrast, Jackson and Schindler (1996) and Jackson et al. (2001) found little evidence of a relationship
between PCB concentration and lipid content within species. Stow (1995) also found that after
controlling for year, species, location, length, and length/species effects, lipid content showed no
relationship with PCB concentrations. In the LMMB Study, the effects of length and lipid content on fish
contamination levels varied by species. For some species (lake trout and alewife), lipid content did not
significantly affect fish contamination levels. For other species (slimy sculpin), lipid content did
significantly affect fish contamination levels and fish length did not. For most species, however, both
length and lipid content, or an interaction of the two parameters, significantly affected fish contamination
levels.
8.3.5 Bioaccumulation and Trophic Transfer
In the LMMB Study, five forage fish species and two piscivorous fish species were analyzed for
contaminant concentrations. Bioaccumulation and biomagnification of PCBs and fra«s-nonachlor within
these fish species are discussed here, and bioaccumulation and biomagnification within the entire Lake
Michigan food web are discussed in Chapter 9. PCBs and fra«s-nonachlor significantly accumulated in
Lake Michigan fish species above concentrations in the water column. Bioaccumulation factors from
water to fish ranged from 106 to 107, depending upon the species and the PCB congener. This is
comparable, but slightly higher than, bioaccumulation factors of 104 to 107 measured by Oliver and Niimi
(1988) in Lake Ontario. Bioaccumulation factors also were higher for fish than for plankton (see Chapter
7).
8-22 April 2004
-------
PCBs and trans-Nonachlor in Fish
Within the upper pelagic food web (fish), PCBs and fra«s-nonachlor were biomagnified in the trophic
transfer from forage fish species to some piscivorous fish. Among the fish species investigated, the
piscivorous lake trout accumulated significantly higher levels of PCBs and fra«s-nonachlor than any of
the forage fish species. This was true regardless of whether comparisons were made on a wet-weight,
dry-weight, or lipid-normalized basis. Biomagnification factors between forage fish species and lake
trout were 4.2 for total PCBs and 3.4 for fra«s-nonachlor. The piscivorous coho salmon, however,
accumulated significantly higher levels of PCBs and trans-nonachlor only when analyzed on a lipid-
normalized basis. Biomagnification factors between forage fish and coho salmon (on a dry-weight basis)
were below 1 for fra«s-nonachlor and near 1 for total PCBs (1.6), indicating that these contaminants were
not significantly biomagnified in the trophic transfer from forage fish to coho salmon.
These findings are consistent with those of other researchers who have calculated higher PCB transfer
efficiencies for lake trout than for coho salmon. Using data from 1975 to 1990, Jackson and Schindler
(1996) calculated PCB transfer efficiencies of 55% for lake trout and 50% for coho salmon. Using
LMMB Study data, Madenjian etal. (1998a) similarly estimated that coho salmon from Lake Michigan
retained 50% of the PCBs that are contained within their food and lake trout retained 80% of PCBs from
food (Madenjian et al, 1998b). Madenjian et al. (1998a) suggested that higher transfer efficiencies in
lake trout than coho salmon could be due to faster or more efficient gut uptake of PCBs in lake trout than
coho salmon. Jackson and Schindler (1996) also suggested that the higher PCB concentrations in lake
trout than other salmonids could be due to lower gross assimilation efficiencies for lake trout (-0.17) than
for other salmonids (-0.23). (Gross assimilation efficiency is a measure of the rate at which an animal
converts food into weight).
In conclusion, PCBs and fra«s-nonachlor were significantly accumulated in Lake Michigan fish.
Accumulation was significantly greater in the predacious lake trout than in forage fish species, indicating
classical biomagnification. Further evaluation of PCB and fra«s-nonachlor movement and accumulation
in the Lake Michigan ecosystem will be provided through the modeling efforts that are the focus of the
LMMB Study.
April 2004 8-23
-------
9.1
Chapter 9
Cross-Media Interpretations
Summary of PCB and frans-Nonachlor Concentrations in Lake Michigan
Compartments
PCB and trans-nonachlor levels were measured in Lake Michigan air, water, sediment, tributaries,
plankton, and fish. PCBs and trans -nonachlor were found throughout the Lake Michigan ecosystem.
PCBs were detected above sample-specific detection limits in all samples collected from all ecosystem
compartments except for tributaries (Table 9-1). Two percent of dissolved tributary samples and nine
percent of particulate tributary samples did not contain detectable levels of PCBs (i.e., not even one PCB
congener above its sample-specific detection limit). Within the various ecosystem compartments, an
average of 30 to 92 different PCB congeners were detected above sample-specific detection limits (Table
9-1). fra«5-Nonachlor was less frequently detected than PCBs. Detection frequency of trans-nonachlor
in environmental samples ranged from 29% in dry deposition to 100% in plankton and fish.
Table 9-1. Summary of Samples from each Ecosystem Compartment with Detectable Levels of PCBs and
frans-Nonachlor
Ecosystem
Compartment
Atmosphere
Tributary
Open Lake
Sediment
Plankton
Fish
Fraction
Vapor
Particulate
Precipitation
Dry Deposition
Dissolved
Particulate
Dissolved
Particulate
-
-
-
PCBs
% Samples with PCBs
Detected above
SSDLa
100
100
100
100
98
91
100
100
100
100
100
Average Number of PCB
Congeners Detected above
SSDLa
59
35
84
30
36
47
40
59
74
92
68
frans-Nonachlor
% Samples with frans-
Nonachlor Detected
above SSDLa
65
20
90
29
34
59
79
88
72
100
100
a Sample-specific detection limit
Figure 9-1 shows the distribution of PCBs throughout the atmosphere, tributaries, water column, and
sediments of Lake Michigan. Vapor-phase total PCB concentrations averaged from 21 to 2600 pg/m3 at
shoreline and over-water sampling stations. Higher atmospheric total PCB concentrations were generally
found above the southern Lake Michigan basin and southern shoreline, with the highest atmospheric
concentration observed at the IIT Chicago site (Figure 9-1). Lower atmospheric concentrations were
generally observed above the central and northern regions of the lake, however, some northern stations
such as Beaver Island maintained higher concentrations. Particulate-phase total PCB concentrations were
much lower than vapor-phase concentrations, averaging from 0.37 to 91 pg/m3 at shoreline and over-
water sampling stations. At individual stations, average particulate-phase total PCB concentrations were
April 2004
9-1
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
only 0.03% to 8.2% of vapor-phase concentrations. In precipitation, total PCB concentrations averaged
from 360 to 16000 pg/L at shoreline and over-water sampling stations, with the highest average
precipitation concentration at the IIT Chicago site. Average total PCB concentrations in precipitation at
all of the stations were greater than the average dissolved total PCB concentration in Lake Michigan, and
average precipitation concentrations at the IIT Chicago site were higher than dissolved total PCB
concentrations in all of the tributaries except for the Grand Calumet and Sheboygan Rivers.
In Lake Michigan tributaries, total PCB concentrations averaged from 0.43 to 35 ng/L in the dissolved
phase and from 0.25 to 55 ng/L in the particulate phase. Total PCB concentrations were highest in the
more urban and industrialized watersheds (Fox, Sheboygan, Milwaukee, Grand Calumet, and Kalamazoo
Rivers). In these tributaries, total PCB concentrations in the dissolved and particulate phases averaged
more than 23 ng/L. The remaining tributaries, which are comprised of more agricultural and forested
watersheds (Grand, Manistique, Menominee, Muskegon, Pere Marquette, and St. Joseph Rivers), all
contained less than 2.9 ng/L of total PCBs.
Within the Lake Michigan water column, total PCB concentrations averaged 0.18 ng/L in the dissolved
phase and 0.073 ng/L in the particulate phase. Dissolved total PCB concentrations were highest at the
southern end of Green Bay (Figure 9-1), presumably due to the significant PCB load from the Fox River
in combination with the reduced dilution available in Green Bay and limited mixing between Green Bay
and Lake Michigan. Total PCB concentrations in the dissolved phase averaged as high as 0.653 ng/L at
the lower end of Green Bay (Station GB17). Within Lake Michigan proper, total PCB concentrations in
the dissolved phase averaged from 0.104 to 0.373 ng/L at the various sampling stations. Contour plots of
dissolved phase total PCB concentrations indicate a general trend of higher concentrations in the southern
Lake Michigan basin than in the northern basin. This observation is consistent with atmospheric
concentrations over the lake, which also tend to be higher over the southern basin, particularly near
Chicago (Figure 9-1).
Total PCB concentrations in surficial sediments of Lake Michigan ranged from 0.138 to 219 ng/g.
Contour plots indicate that total PCBs are accumulating in the depositional and transitional regions of
Lake Michigan (Figure 9-1). In particular, PCBs accumulate at relatively high concentrations (> 100
ng/g) along the eastern side of the southern basin, as well as at a few of the deeper stations in the southern
and central basins. Among the lake's depositional and transitional zones, sediment PCB concentrations
generally increase from north to south. Total PCB concentrations averaged 69.7 ng/g in the southern
basin, 50.5 ng/g in the central basin, 41.6 ng/g in the northern basin, and 7.27 ng/g in the straits region.
Only in the southern and central basins did total PCBs exceed 100 ng/g, and only in the southern basin
did total PCBs exceed 150 ng/g. Increased PCB concentrations in the sediments of the southern basin are
consistent with the increased atmospheric PCB concentrations above the southern basin and the larger
number of urban and industrial sources surrounding the southern portions of the lake.
Figure 9-2 shows the distribution of fra«s-nonachlor throughout the atmosphere, tributaries, water
column, and sediments of Lake Michigan. Vapor-phase trans-nonachlor concentrations averaged from 0
to 29 pg/m3 at shoreline and over-water sampling stations. Like total PCB concentrations, higher
atmospheric fra«s-nonachlor concentrations were generally found above the southern Lake Michigan
basin and southern shoreline, with the highest atmospheric concentration observed at the IIT Chicago site
(Figure 9-2).
9-2 April 2004
-------
Figure 9-1. Concentrations of Total PCBs in the Atmosphere, Tributaries, Water Column, and Sediments of Lake Michigan
WATER
SEDIMENT
Vapor Phase
Total PCBs (pg/m3
Water Column
Diss. Total PCBs
(ng/L)
Sediment
Total PCBs (ng/g)
210
42
9-3
-------
Figure 9-2. Concentrations of frans-Nonachlor in the Atmosphere, Tributaries, Water Column, and Sediments of Lake Michigan
WATER
SEDIMENT
Vapor Phase
Frans-nonachlor
Water Column Diss
TYans-nonachlor
(ng/L)
Sediment
Trans-nonachlor
(ng/g)
2.4
0.4
0.0
9-4
-------
Cross-Media Interpretations
Participate -phase trans -nonachlor concentrations were much lower than vapor-phase concentrations,
averaging from 0.16 to 1.8 pg/m3 at shoreline and over-water sampling stations. At individual stations,
average particulate-phase fra«s-nonachlor concentrations were only 2.6% to 14% of vapor-phase
concentrations. In precipitation, fra«s-nonachlor concentrations averaged from 0.0 to 100 pg/L at
shoreline and over-water sampling stations, with the highest average precipitation concentration at the IIT
Chicago site. Average trans -nonachlor concentrations in precipitation at all except three of the stations
were greater than the average dissolved fra«s-nonachlor concentration in Lake Michigan, and average
precipitation concentrations at the IIT Chicago site were higher than dissolved trans -nonachlor
concentrations in all of the tributaries.
In Lake Michigan tributaries, fra«s-nonachlor concentrations averaged from 0.0033 to 0.026 ng/L in the
dissolved phase and from 0.0028 to 0.074 ng/L in the particulate phase. Unlike PCBs, which were
highest in urban and industrial influenced watersheds, fra«s-nonachlor concentrations were highest in the
heavily agricultural watersheds of the St. Joseph and Grand Rivers. Mid-range trans -nonachlor
concentrations were observed in the urban and industrial influenced watersheds of the Fox, Sheboygan,
Milwaukee, Grand Calumet, and Kalamazoo Rivers; and the lowest fra«s-nonachlor concentrations were
observed in the more forested watersheds of the Menominee, Manistique, Muskegon, and Pere Marquette
Rivers.
Within the Lake Michigan water column, fra«s-nonachlor concentrations averaged 0.0058 ng/L in the
dissolved fraction and 0.0021 ng/L in the particulate fraction. Among open-water sampling stations,
dissolved fra«s-nonachlor concentrations averaged from 0.00228 to 0.0236 ng/L, with the highest
concentration at Station 17 in the southern Lake Michigan basin (Figure 9-2). Contour plots of dissolved-
phase trans -nonachlor concentrations indicate a general trend of higher concentrations in the southern
Lake Michigan basin with isolated areas of high concentration in the northern basin. This observation is
consistent with atmospheric concentrations over the lake, which also tend to be higher over the southern
basin, particularly near Chicago (Figure 9-2). The apparent relationship between atmospheric and open-
lake concentrations may suggest atmospheric deposition drives open-water concentrations, or it may
suggest that fra«s-nonachlor may cycle between lake water and the atmosphere in manner similar to that
proposed by Mackay and Patterson (1986) for PCBs (see Section 2.1.4).
fra«5-Nonachlor concentrations in surficial sediments of Lake Michigan ranged from 0.00250 to 2.830
ng/g. Contour plots indicate that trans -nonachlor is accumulating in the depositional and transitional
regions of Lake Michigan (Figure 9-2). Unlike PCBs, trans -nonachlor does not exhibit elevated
concentrations in the southern basin relative to the central and northern basins, trans -Nonachlor
concentrations averaged 0.599 ng/g in the southern basin, 0.638 ng/g in the central basin, and 0.560 ng/g
in the northern basin. Also unlike PCBs, fra«s-nonachlor is not preferentially accumulating along the
eastern side of the southern basin. fra«s-Nonachlor is preferentially accumulated in the deeper regions of
the lake just to the east and west of the mid-lake reef identified as the Port Washington biota box in this
study (Figures 9-2 and 9-3).
Within living components of the Lake Michigan ecosystem, PCBs and fra«s-nonachlor were accumulated
at concentrations higher than in any abiotic ecosystem component. PCBs and trans -nonachlor exhibited
classical biomagnification, with concentrations increasing with increasing trophic level in the Lake
Michigan food web (see Section 9.2). Accumulation of PCBs in top predator fish has reached levels of
concern for human health and spawned fish consumption advisories throughout Lake Michigan and many
Lake Michigan tributaries.
April 2004 9-5
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 9-3. Lake Michigan Bathymetry
9-6
April 2004
-------
Cross-Media Interpretations
9.2 Bioaccumulation and Biomagnification
In the LMMB Study, classical bioaccumulation and biomagnification of PCBs and trans-nonachlor were
observed. These hydrophobic and lipophilic contaminants were bioaccumulated in living tissue at levels
well above water column or even sediment concentrations. Bioaccumulation factors for total PCBs
ranged from 2.7 x 105 to 4.3 x 107, and bioaccumulation factors for trans-nonachlor ranged from 3.0 x 105
to 8.3 x 107.
Not only were PCBs and trans-nonachlor- bioaccumulated in living tissue above water concentrations, but
these contaminants were biomagnified within the Lake Michigan food web. PCB and trans -nonachlor
concentrations increased with each successive trophic levels (Figure 9-4).
Figure 9-4. Total PCB and frans-Nonachlor Concentrations in Various Components of
the Lake Michigan Ecosystem
10000 -i
mnn
^_^_
"S mn
c5 m
o 1U
o
0 1
m '
o
Q. n 1
2
Ł nm
Onm
0.0001
1
<8
—
1 I
i 1
•o to
fl) "7^
" I
Q.
1 1
C
O
c
JS
o
a
•81
s
.2
o
Q.
Q'~
1
0)
O)
S
o
LL.
"S
o
i-
J3
1000 -,
S* mn
Q.
Q.
' — ' m
o
O 1
0 '
o n 1
o
m n m
o
C Q QQ1
< ^r
.E c c S o
T3 (8 (8 .2"
« "5. "5. Q
Woo
*Ji o
^ N
^^
Q.
t/1 ^
0, ^
2 «
o J3
LL
April 2004
9-7
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Total PCB concentrations increased from 49 ng/g in phytoplankton to 170 ng/g in zooplankton, to 280
and 420 ng/g in Mysis and Diporeict, to 1900 ng/g in forage fish, to 7800 ng/g in the top predator, lake
trout. fra«s-Nonachlor concentrations increased from 1.7 ng/g in phytoplankton, to 16 ng/g in
zooplankton, to 25 and 32 ng/g in Mysis and Diporeia, to 140 ng/g in forage fish, to 480 ng/g in lake
trout. From the bottom of the food web (phytoplankton) to the top of the food web (lake trout), total PCB
concentrations increased by a factor of 160, and fra«s-nonachlor concentrations increased by a factor of
280.
Figure 9-5 shows the biomagnification factors between the various components of the Lake Michigan
food web. The primary pelagic food web includes phytoplankton, zooplankton, forage fish, and lake
trout. Biomagnification factors between each of these trophic levels varied from 3.4 to 11 for total PCBs
and 3.4 to 9.5 for trans-nonachlor. For total PCBs, biomagnification was greatest from zooplankton to
forage fish (11), with lower biomagnification from phytoplankton to zooplankton (3.4) and from forage
fish to lake trout (4.2). For fra«s-nonachlor, biomagnification was high from phytoplankton to
zooplankton (9.5) and from zooplankton to forage fish (8.6) and lower for forage fish to lake trout (3.4).
Within this simplified pelagic food web, invertebrates such as Mysis may also play a role. Mysis may
feed on herbivorous zooplankton, adding an additional trophic level to the pelagic food web, or they may
be eaten directly by young top predators, effectively removing a trophic level.
The simplified benthic food web consists of benthic invertebrates, such as Diporeia, that may feed on
detritus or phytoplankton. The benthic invertebrates may then be preyed upon by bottom dwelling forage
fish, which in turn may be preyed upon by lake trout. Within the benthic food web, biomagnification
factors were greatest between phytoplankton and Diporeia (8.5 for PCBs and 18 for fra«s-nonachlor) and
between detritus and Diporeia (8.3 for PCBs and 59 for fra«s-nonachlor). Biomagnification factors were
lower between Diporeia and forage fish (4.5 for PCBs and 4.4 for fra«s-nonachlor) and between forage
fish and lake trout (4.2 for PCBs and 3.4 for fra«s-nonachlor).
It should be noted that bioaccumulation and biomagnification factors for total PCBs calculated from this
data are semi-quantitative. Total PCB concentrations were calculated from the sum of individual PCB
congeners analyzed, and the number and the specific congeners analyzed differed among the sample
matrices (water, plankton, and fish). This difference, however, should not affect conclusions concerning
the biomagnification of PCBs in Lake Michigan, because water and lower trophic levels were analyzed
for more PCB congeners than higher trophic levels. Total PCB concentrations in water, plankton, and
fish represent 123, 126, and 93 individual or coeluting congeners analyzed for the respective matrices.
Similar to the PCB and trans-nonachlor biomagnification observed in the LMMB Study, Oliver and
Niimi (1988) observed classical biomagnification of PCBs in the Lake Ontario food web. Oliver and
Niimi (1988) observed increases in total PCB levels at each trophic level with an overall increase of 86
times from plankton to top predators. This is slightly lower than observed in Lake Michigan in the
LMMB Study (factor of 160), however, Oliver and Niimi (1988) analyzed a mixture of phytoplankton
and zooplankton to characterize the plankton compartment and analyzed a mixture of coho salmon,
rainbow trout, and lake trout to characterize the top predator compartment. These approaches would tend
to decrease the overall biomagnification measured by Oliver and Niimi (1988) in comparison to the
LMMB Study data.
Koslowski et a/. (1994) observed similar biomagnification of PCBs in the Lake Erie food web, and
Kucklick and Baker (1998) observed similar biomagnification in Lake Superior. Both authors found
increasing PCB concentrations with increasing trophic level. Through multiple linear regression and path
analysis, Kucklick and Baker (1998) determined that the trophic level influenced PCB accumulation
levels both directly and indirectly through its effect on lipid content.
9-8 April 2004
-------
Figure 9-5. Biomagnification Factors for Total PCBs (BMFp) and frans-Nonachlor (BMFr) in a Simplified Lake Michigan Food Web
BMFP = 19
BMFT = 15
Herbivorous
Zooplankton
Phytoplankton
BMFp = 31
BMFT = 19
Concentrations of PCBs and frans-nonachlor in the sediment were used as a surrogate for detrital concentrations.
9-9
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
In the marine environment, authors have observed similar biomagnification of PCBs and other persistent
organic pollutants. Hop et al. (2002) and Fisk et al. (2001) investigated biomagnification through marine
food webs that have included water birds and mammals as top predators. In these studies,
biomagnification at the upper end of these food webs (in the homeothermic mammal and avian
populations) was even greater than at the lower end of the food web (in fish and invertebrates).
In contrast to the studies that have demonstrated biomagnification of PCBs in aquatic food webs,
Berglund et al. (2000) concluded that total PCB concentrations did not steadily increase with increasing
trophic level in 19 Swedish lakes. On a dry-weight basis, total PCB concentrations in fish were not
significantly different from concentrations in zooplankton, however, Berglund et al. (2000) only
investigated young-of-the-year fish and did not investigate piscivorous fish species. On a lipid-weight
basis, total PCB concentrations in fish were significantly higher than in zooplankton, but were not
significantly higher than in phytoplankton. As demonstrated in the LMMB Study, age (or length as a
surrogate for age) certainly affects the bioaccumulation of PCBs. In attempting to factor out this effect by
only using young-of-the-year fish, however, a true comparison of trophic levels cannot be adequately
made. Increases in the lifespan of organisms are a component of increasing trophic level.
10 i
9.3 Fractionation
While the discussion of bioaccumulation and biomagnification has focused primarily on total PCBs, each
of the 209 PCB congeners have differing physical and chemical properties and may be accumulated and
biomagnified differentially. In general, more-chlorinated PCB congeners are more hydrophobic and more
lipophilic. This chemical trend can be described by the octanol-water partition coefficient (KQW), which
is the ratio of the concentration of a substance preferentially dissolved in an octanol phase to the
concentration of that substance dissolved in the water phase. More-chlorinated PCB congeners have
higher octanol-water partition coefficients (Figure 9-6).
Because more-chlorinated PCB congeners are more Fjgure 9.5 Qctanol-water Partition Coefficients
hydrophobic and more lipophilic, these more-chlorinated (KOW) for pcB Congeners
congeners more readily bioaccumulate in living tissue.
Figure 9-7 shows the bioaccumulation factors for each of
the trophic levels versus Kow. The slope for each of the
trophic levels is positive, indicating that the more-
chlorinated PCB congeners with higher Kow values are
bioaccumulated to a greater degree. Slopes for these log-
log regressions range from 0.24 to 0.46. These regressions
also indicate increased bioaccumulation at each trophic
level. For a given Kow, PCBs are increasingly more
accumulated in phytoplankton, zooplankton, forage fish,
and lake trout.
Many other authors have documented this same trend of
increasing PCB bioaccumulation with increasing Kow or
increasing chlorination. Oliver and Niimi (1988) observed
that less-chlorinated PCB congeners comprised a higher
fraction of total PCBs in water than in higher trophic
levels in Lake Ontario. Koslowski et al. (1994) also
observed increased bioaccumulation with increased PCB congener chlorination in the Lake Erie food web.
Willman et al. (1997) similarly found that penta-, hexa-, and heptachloro congeners were enriched relative
to other congeners as PCBs moved to higher trophic levels from sediments to plankton to fish.
100 150
PCB Conyener
9-10
April 2004
-------
Cross-Media Interpretations
Figure 9-7. Bioaccumulation Factors in the Lake Michigan Food Web versus Log Octanol-Water
Partition Coefficients for Individual PCB Congeners
1.00E+10
1 .OOE+09
1 .OOE+08
1 .OOE+07
1 .ooE+oe
1 .OOE+05
1 .OOE+04
1 .OOE+03
•» _*•»**.» S *
A Lake Trout
• Forage Fish
Zooplankton
• Phytoplankton
10
LogK
ow
In the lower pelagic food web, it was observed that more-chlorinated PCB congeners not only
preferentially accumulated from water, but these more-chlorinated congeners also preferentially
accumulated in the transfer from prey to predator. For example, the slope of the log BAF versus log Kow
curve increases from phytoplankton to zooplankton. This indicates that relative to phytoplankton,
zooplankton preferentially accumulated more-chlorinated PCB congeners. This was not true of each
trophic transfer, however. The slopes of the curves did not increase with each increasing trophic level.
Figure 9-8 shows biomagnification factors plotted against log Kow values of the PCB congeners. Only in
the transfer from phytoplankton to zooplankton were more-chlorinated PCB congeners preferentially
accumulated. The slopes for zooplankton to forage fish and for forage fish to lake trout were negative
and close to zero. Kucklick and Baker (1998) also did not observe fractionation of PCB congeners
between predator and prey. More chlorinated PCB congeners with higher Kow values were not
selectively accumulated in predators.
April 2004
9-11
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 9-8. Biomagnification Factors in the Lake Michigan Food Web versus Log Octanol-Water
Partition Coefficients for Individual PCB Congeners
100
»Phytoplankton to Zooplankton
• Zooplankton to Forage Fish
A Forage Fish to Lake Trout
LogK
ow
9.4 Toxic PCB Congeners
It is important to independently consider the concentration and
accumulation of individual PCB congeners, because the
individual congeners bioaccumulate differentially and have
varying degrees of toxicity. The World Health Organization
has identified the PCB congeners listed in Table 9-2 as toxic
and "dioxin-like" based on structure-activity relationships (Van
den Berg et al., 1998). For each of these toxic PCB congeners,
the World Health Organization also has assigned toxicity
equivalency factors (TEFs), which relate the toxicity of each
congener to the toxicity of 2,3,7,8-tetra-chlorodibenzo-/?-
dioxin (TCDD). A compound with a TEF value of 1.0 is as
potent as TCDD, and a compound with a TEF value of 0.01 is
estimated to be 100 times less potent than TCDD. Based on the
assumption of additivity, the product of individual PCB
concentrations and TEFs can be summed to calculate the toxic
equivalent concentration (TEQ) of all dioxin-like compounds
in terms of TCDD.
Fish samples from the LMMB Study were analyzed for 80
individual PCB congeners including all of the toxic congeners
except for PCB 169. Sediment, tributary, water column, and
plankton samples were not analyzed for PCB 169 or PCB 126.
In the Lake Michigan food web, the toxic PCB congeners
bioaccumulated and biomagnified in the same general pattern
as described for total PCBs. Concentrations of these toxic
PCB concentrations increased in each successive trophic level
Table 9-2. Toxic PCB Congeners and
Toxicity Equivalency Factors (TEF)
PCB Congener
77
81
105
114
118
123
126
156
157
167
169
189
TEF
0.0001
0.0001
0.0001
0.0005
0.0001
0.0001
0.1
0.0005
0.0005
0.00001
0.01
0.0001
9-12
April 2004
-------
Cross-Media Interpretations
of the Lake Michigan food web (Figure 9-9). Of the toxic PCB congeners, concentrations of PCB 118 were
highest. This congener has aTEF of 0.0001, meaning that the compound is one ten-thousandth as toxic as
TCDD. PCB 126, which is the most toxic of the PCB congeners (TEF of 0.1), was found in fish tissue at
nearly the lowest level of toxic PCB congeners.
Figure 9-9. Concentration of Toxic PCB Congeners in the Lake Michigan Ecosystem
iuuu -
mn -
1 n
--> I V
"^ n 1
«; u- 1
Ł n n-i
o °-01
»n nm
U.UU I
a. n nnm
U.UUU I
Onnnrn
.UUUU I
Onnnnm -r
.UUUUU I
Onnnnnn-i
.UUUUUU I
n nnnnnnn-i
r
I
I
I
I
I
I
r
I
r
I
[
n water
nphyto
•zoo
a Forage Fish
• Lake Trout
77 81 105 114 118 123 126 156 157 167 169 189
PCB Congener
In open-lake water, the sum of the 12 toxic PCB congeners contributed 3.9% of the total PCB
concentration. In the lower pelagic food web (phytoplankton, zooplankton, Mysis, and Diporeid), the
toxic PCB congeners contributed 9.5 to 12% of the total PCB concentration. In fish, the toxic PCB
congeners contributed 11 to 19% of the total PCB concentration. The percentage of total PCBs attributed
to the toxic congeners did not increase by trophic level. In both phytoplankton (the base of the pelagic
food web) and lake trout (the top of the pelagic food web), the toxic PCB congeners contributed 11% of
the total PCB concentration. The highest percentage of toxic PCB congeners (19%) was observed in the
deepwater sculpin.
In relation to total PCBs, most of the toxic PCB congeners were bioaccumulated to a greater extent
(Figure 9-10). The toxicity of these congeners and their potential for bioaccumulating, in part, depends
upon their ability to penetrate cell membranes. In phytoplankton, bioaccumulation factors for all of the
toxic PCB congeners were higher than for total PCBs. In zooplankton, bioaccumulation factors for all of
the toxic PCB congeners except for PCB 114 were higher than for total PCBs. In forage fish and lake
trout, bioaccumulation factors for all of the toxic PCB congeners except for PCB 77, 114, and 123 were
higher than for total PCBs. This indicates that bioaccumulation based on total PCB values may
underestimate the bioaccumulation of the toxic PCB congeners.
The drop in BAFs for PCBs 77 and 123 that is apparent in Figure 9-10 may be an artifact of analytical
differences between the laboratories responsible for the analyses of lower pelagic food web organisms
and fish. In the lower pelagic food web analyses, PCBs 77 and 123 each coeluted with another PCB
congener that is not among the 12 toxic PCBs, while in the fish analyses, these two PCBs did not coelute
with any other congeners. Thus, concentrations of PCBs 77 and 123 may be biased high in the
phytoplankton and zooplankton samples, leading to higher BAFs for these two trophic levels, while the
BAFs for forage fish and trout do not include contributions from other congeners.
April 2004
9-13
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Figure 9-10. Bioaccumulation Factors (BAFs) for Toxic PCB Congeners and Total PCBs in Lake
Michigan Biota
1.00E-HD9
1.QOE-KB
U,
< 1.QOE4Q7
ffl
1.00E-KJ6
1.ODE-+05
-77
-81
-105
-114
-118
-123
-156
-157
167
189
-total
Phytoplankton Zooplankton Forage Fish Lake Trout
Table 9-3 shows toxic equivalent concentrations
(TEQs) calculated for the toxic PCB congeners in
terms of TCDD. TEQ values ranged from 0.0032
to 0.52 ng/g - TCDD equivalents. Even though
these TEQs include only dioxin-like PCBs and not
measured concentrations of dioxins and furans,
these values exceed EPA's recommended fish
consumption limits for dioxins/furans TEQs of
0.0012 ng/g. It should also be noted that these TEQ
values do not include the contribution of PCB 169,
because this congener was not analyzed in the
LMMB Study.
PCB 126, the most toxic PCB congener,
contributed 71 to 98% of the total TEQ values for
each species, with the exception of coho salmon
collected from the hatchery (in which PCB 126
was not detected). Surprisingly, the highest TEQ
value was not calculated for lake trout but was
calculated for large bloater chub. While total PCB
concentrations in lake trout were 3.6 times total
Table 9-3. Toxic Equivalent Concentrations (TEQs) for
Dioxin-like PCB Congeners in Lake Michigan Fish
Species/Size Category
Alewife< 120mm
Alewife> 120mm
Bloater< 160mm
Bloater> 160mm
Coho-Adult
Coho-Hatchery
Coho-Yearling
Deepwater Sculpin
Lake Trout
Smelt
Slimy Sculpin
TEQa (ng/g)
0.015
0.075
0.060
0.52
0.035
0.0032
0.010
0.048
0.24
0.039
0.029
a TEQ based on toxicity of dioxin-like PCB congeners relative to
2,3,7,8-tetrachlorodibenzo-p-dioxin(TCDD).
PCB concentrations in large bloater chub, concentrations of PCB 126 were higher in large bloater chub (5.1
ng/g) than in lake trout (2.0 ng/g). The higher concentration of this PCB congener in large bloater chub
caused the higher TEQ for large bloater chub. This finding also points out the importance of measuring and
evaluating individual PCB congener concentrations in addition to total PCB concentrations.
9-14
April 2004
-------
References
Anderson, M. L., C. P. Rice, and C. C. Carl. 1982. Residues of PCB in a Cladophora community along
the Lake Huron shoreline. Journal of Great Lakes Research. 8(1): 196-200.
Baker, J. E. and S. J. Eisenreich. 1990. Concentrations and Fluxes of Polycyclic Aromatic Hydrocarbons
and Polychlorinated Biphenyls across the Air-Water Interface of Lake Superior. Environmental
Science and Technology. 24: 342-352.
Ballschmiter, K. and M. Zell. 1980. Analysis of Polychlorinated Biphenyls (PCB) by Glass Capillary
Gas Chromatography. Fresenius Zeitschrift Fur Analytische Chemie. 302: 20-31.
Berglund, O., P. Larsson, G. Ewald, and L. Okla. 2000. Bioaccumulation and differential partitioning of
polychlorinated biphenyls in freshwater, planktonic food webs. Canadian Journal of Fisheries
and Aquatic Sciences. 57: 1160-1168.
Borgmann, U. and D. M. Whittle. 1992. DDE, PCB, and mercury concentration trends in Lake Ontario
rainbow smelt (Osmerus mordax) and slimy sculpin (Cottus cognatus): 1977 to 1988. Journal of
Great Lakes Research. 18(2): 298-308.
Brunciak, P. A., J. Dachs, T. P. Franz, C. L. Gigliotti, E. D. Nelson, B. J. Turpin, and S. J. Eisenreich.
2001. Polychlorinated biphenyls and particulate organic/elemental carbon in the atmosphere of
Chesapeake Bay, USA. Atmospheric Environment. 35(32): 5663-5677.
Burkhard, L. P. and D. Weininger. 1987. Analysis of polychlorinated biphenyls using multiple
regression with outlier detection and elimination. Analytical Chemistry. 59: 1187-1190.
Cahill, R. A. 1981. Geochemistry of Recent Lake Michigan Sediments. Circular 517: 94 pp. Illinois
Institute of Natural Resources, State Geological Survey Division.
Datta, S., L. L. McConnell, J. E. Baker, J. Lenoir, and J. N. Seiber. 1998. Evidence for atmospheric
transport and deposition of polychlorinated biphenyls to the Lake Tahoe basin, California-Nevada.
Environmental Science and Technology. 32: 1378-1385.
DeVault, D. S., R. Hesselberg, P. W. Rodgers, and T. J. Feist. 1996. Contaminant trends in lake trout
and walleye from the Laurentian Great Lakes. Journal of Great Lakes Research. 22(4): 884-895.
Dunnivant, F. M., and A. W. Elzerman. 1988. Aqueous solubility and Henry's law constant data for
PCB congeners for evaluation of quantitative structure-property relationships (QSPRs).
Chemosphere 17: 525-541.
Eisenreich, S. J., P. D. Capel, J. A. Robbins, and R.A. Bourbonniere. 1989. Accumulation and
diagenesis of chlorinated hydrocarbons in lacustrine sediments. Environmental Science and
Technology. 23: 1116-1123.
Epplett, T. D., S. Gewurtz, R. Lazar, and G. D. Haffner. 2000. Seasonal dynamics of PCBs in the
plankton of Lake Erie. Journal of Great Lakes Research. 26(1): 65-73.
Fiedler, H. 2001. Polychlorinated Biphenyls (PCB). (http://irptc.unep.ch/pops/stpeter/stpete2c.html).
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Fiedler H., H. Hoff, J. Tolls, C. Mertens, A. Gruber, and O. Hutzinger. 1994. Environmental Fate of
Organochlorines in the Aquatic Environment. Orgcmohalogen Compounds. 15: 199 pp.
ECO-INFORMA Press, Bayreuth. ISBN 3928379119.
Filkins, J. C., J. M. Townsend, and S. G. Rood. 1983. Organochlorines in Offshore Waters of the Great
Lakes, 1981. Report to the U.S. Environmental Protection Agency, Office of Research and
Development, ERL-Duluth, Large Lakes Research Station, Grosse lie, Michigan. 12pp.
Fisk, A. T., K. A. Hobson, and R. J. Norstrom. 200 la. Influence of chemical and biological factors on
trophic transfer of persistent organic pollutants in the Northwater Polynya marine food web.
Environmental Science and Technology. 35: 732-738.
Fisk, A. T., G. A. Stern, K. A. Hobson, W. J. Strachan, M. D. Loewen, and R. J. Norstrom. 200Ib.
Persistent organic pollutants (POPs) in a small, herbivorous, Arctic marine zooplankton (Calanus
hyperboreus): trends from April to July and the influence of lipids and trophic transfer. Marine
Pollution Bulletin. 43:93-101.
Franz, T. P., S. J. Eisenreich, and T. Holsen. 1998. Dry deposition of particulate polychlorinated
biphenyls and polycyclic aromatic hydrocarbons to Lake Michigan. Environmental Science and
Technology. 32(23): 3681-3688.
Golden, K. A.. 1994. Organochlorines in Lake Michigan Sediments. (Master of Science thesis,
University of Minnesota, Minneapolis, Minnesota, 1994).
Golden, K. A.., C. S. Wong, J. D. Jeremiason, S. J. Eisenreich, G. Sanders, J. Hallgren, D. L.
Swackhamer, D. R. Engstrom, and D. T. Long. 1993. Accumulation and preliminary inventory of
Organochlorines in Great Lakes sediments. Water Science and Technology. 28:19-31.
Green, M. L., J. V. DePinto, C. W. Sweet, and K. C. Hornbuckle. 2000. Regional Spatial and Temporal
Interpolation of Atmospheric PCBs: Interpretation of Lake Michigan Mass Balance Data.
Environmental Science and Technology. 34(9): 1833-1850.
Harding, G. C., R. J. LeBlanc, W. P. Vass, R. F. Addison, B. T. Hargrave, S. Pearre, A. Dupuis, and P. F.
Brodie. 1997. Bioaccumulation of polychlorinated biphenyls (PCBs) in the marine pelagic food
web, based on a seasonal study in the southern Gulf of St. Lawrence, 1976-1977. Marine
Chemistry. 56: 145-179.
Hargrave, B. T., G. A. Phillips, W. P. Vass, P. Bruecker, H. E. Welch, and T. D. Siferd. 2000.
Seasonality in bioaccumulation of Organochlorines in lower trophic level arctic marine biota.
Environmental Science and Technology. 34: 980-987.
Hermanson, M. H., E. R. Christensen, D. J. Buser, and L.-M. Chen. 1991. Polychlorinated biphenyls in
dated sediment cores form Green Bay and Lake Michigan. Journal of Great Lakes Research. 17:
94-108.
Hesselberg, R. J., J. P. Hickey, D. A. Nortrup, and W. A. Willford. 1990. Contaminant residues in the
bloater (Coregonus hoyi) of Lake Michigan, 1969-1986. Journal of Great Lakes Research. 16(1):
121-129.
R-2 April 2004
-------
References
Hoff, R. M., D. C. G. Muir, and N. P. Grift. 1992. Annual cycle of poly chlorinated biphenyls and
organohalogen pesticides in air in Southern Ontario. 1. Air concentration data. Environmental
Science and Technology. 26: 266-275.
Holson, T. M., G. J. Keeler, K. N. Noll, G. Fang, W. Lee, and J. Lin. 1993. Dry Deposition and Particle
Size Distributions Measured during the Lake Michigan Urban Air Toxics Study. Environmental
Science and Technology. 27(7): 1327-1333.
Hop, H., K. Borga, G. W. Gabrielsen, L. Kleivane, and J. U. Skaare. 2002. Food web magnification of
persistent organic pollutants in poikilotherms and homeotherms from the Barents Sea.
Environmental Science and Technology. 36: 2589-2597.
Hornbuckle, K. C., J. V. DePinto, M. L. Green, S. M. Miller, and J. J. Bogdan. 2001. Atmospheric
Deposition of Persistent Organic Pollutants: Results of the Lake Michigan Mass Balance Study.
Abstracts: June 10-14, 2001. International Association for Great Lakes Research, Green Bay,
Wisconsin.
Hornbuckle, K. C., D. R. Achman, and S. J. Eisenreich. 1993. Over-Water and Over-Land
Polychlorinated Biphenyls in Green Bay, Lake Michigan. Environmental Science and
Technology. 27(1): 87-98.
Hornbuckle, K. C., C. W. Sweet, R. F. Pearson, D. L. Swackhamer, and S. J. Eisenreich. 1995.
Assessing annual water-air fluxes of poly chlorinated biphenyls in Lake Michigan. Environmental
Science and Technology. 29: 869-877.
Huestis, S. Y., M. R. Servos, D. M. Whittle, and D. G. Dixon. 1996. Temporal and age-related trends in
levels of poly chlorinated biphenyl congeners and organochlorine contaminants in Lake Ontario
lake trout. Journal of Great Lakes Research. 22(2): 310-330.
Ivanov, V. and E. Sandell. 1992. Characterization of Polychlorinated Biphenyl Isomers in Sovol and
Trichlorodiphenyl Formulations by High-Resolution Gas Chromatography with Electron Capture
Detection and High-Resolution Gas Chromatography - Mass Spectrometry Techniques.
Environmental Science and Technology. 26: 2012-2017.
Iwata, H., S. Tanabe, N. Sakal, and R. Tatsukawa. 1993. Distribution of persistent organochlorines in the
oceanic air and surface seawater and the role of ocean on their global transport and fate.
Environmental Science and Technology. 27: 1080-1098.
Jackson, L. J. and D. E. Schindler. 1996. Field estimates of net trophic transfer of PCBs from prey fishes
to Lake Michigan salmonids. Environmental Science and Technology. 30: 1861-1865.
Jackson, L. J., S. R. Carpenter, J. Manchester-Neesvig, and C. A. Stow. 1998. Current concentrations of
PCBs in Lake Michigan invertebrates, a prediction test, and corroboration of hindcast
concentrations. Journal of Great Lakes Research. 24(4): 808-821.
Jackson, L. J., S. R. Carpenter, J. Manchester-Neesvig, and C. A. Stow. 2001. PCB congeners in Lake
Michigan coho (Oncorhynchus kisutch) and chinook (Oncorhynchus tshawytscha) salmon.
Environmental Science and Technology. 35: 856-862.
April 2004 R-3
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Koslowski, S. E., C. D. Metcalfe, R. Lazar, and G. D. Haffiier. 1994. The distribution of 42 PCBs,
including three coplanar congeners, in the food web of the western basin of Lake Erie. Journal of
Great Lakes Research. 20(1): 260-270.
Kucklick, J. R. and J. E. Baker. 1998. Organochlorines in Lake Superior's food web. Environmental
Science and Technology. 32:1192-1198.
Leister, D. L and Baker, J. E. 1994. Atmospheric deposition of organic contaminants to the Chesapeake
Bay. Atmospheric Environment. 28(8): 1499-1520.
Mackay, D. 1982. Correlation of bioconcentration factors. Environmental Science and Technology.
16(5): 274-278.
Mackay, D., S. Patterson, and W. H. Schroeder. 1986. Model Describing the Rates of Transfer Processes
of Organic Chemicals between Atmosphere and Water. Environmental Science and Technology.
20: 810-816.
Mackay D., W. Y. Shiu, and K. C. Ma. 1992. Illustrated Handbook of Physical-Chemical Properties
and Environmental Fate for Organic Chemicals. Vol. I+II. Boca Raton, FL. Lewis Publishers
Inc.
Madenjian, C. P., R. F. Elliot, L. J. Schmidt, T. J. DeSorcie, R. J. Hesselberg, R. T. Quintal, L. J.
Begnoche, P. M. Bouchard, and M. E. Holey. 1998a. Net Trophic Transfer Efficiency of PCBs to
Lake Michigan Coho Salmon from Their Prey. Environmental Science and Technology. 32(20):
3063-3067.
Madenjian, C. P., R. J. Hesselberg, T. J. DeSorcie, L. J. Schmidt, R. M. Stedman, R. T. Quintal, L. J.
Begnoche, and D. R. Passino-Reader. 1998b. Estimate of net trophic transfer efficiency of PCBs
to Lake Michigan lake trout from their prey. Environmental Science and Technology. 32: 886-
891.
Madenjian, C. P., T. J. DeSorcie, R. M. Stedman, E. H. Brown, Jr., G. W. Eck, L. J. Schmidt, R. J.
Hesselberg, S. M. Chernyak, and D. R. Passino-Reader. 1999a. Spatial Patterns in PCB
Concentrations of Lake Michigan Lake Trout. Journal of Great Lakes Research. 25(1): 149-159.
Madenjian, C. P., L. J. Schmidt, S. M. Chernyak, R. F. Elliott, T. J. DeSorcie, R. T. Quintal, L. J.
Begnoche, and R. J. Hesselberg. 1999b. Variation in Net Trophic Transfer Efficiencies among 21
PCB Congeners. Environmental Science and Technology. 33(21): 3768-3773.
Manchester, J. B. 1993. The Role ofPorewater in the Remobilization of Sediment-bound Poly chlorinated
Biphenyl Congeners. (Doctoral thesis, University of Wisconsin-Madison, 1993). 584pp.
Marti, E. A., and D. E. Armstrong. 1993. Polychlorinated Biphenyls in Lake Michigan Tributaries.
Journal of Great Lakes Research. 16(3): 396-405.
Miller, M. A., C. P. Madenjian, and R. G. Masnado. 1992. Patterns of organochlorine contamination in
lake trout from Wisconsin waters of the Great Lakes. Journal of Great Lakes Research. 18(4):
742-754.
R-4 April 2004
-------
References
Miller, S. M. 1999. Spatial and Temporal Variability of Organic and Nutrient Compounds in
Atmospheric Media Collected During the Lake Michigan Mass Balance Study. (Master of Science
thesis, State University of New York at Buffalo, 1999). 181 pp.
Offenberg, J. H. and J. E. Baker. 1997. Polychlorinated biphenyls in Chicago precipitation: enhanced
wet deposition to near-shore Lake Michigan. Environmental Science and Technology. 31: 1534-
1538.
Oliver, B. G. and A. J. Niimi. 1988. Trophodynamic analysis of poly chlorinated biphenyl congeners and
other chlorinated hydrocarbons in the Lake Ontario ecosystem. Environmental Science and
Technology. 22: 388-397.
Park, J. S., T. L. Wade, and S. Sweet. 2001. Atmospheric deposition of organochlorine contaminants to
Galveston Bay, Texas. Atmospheric Environment. 35(19): 3315-3324.
Pearson, R. F., Hornbuckle, K. C., Eisenreich, S. J., and Swackhamer, D. L. 1996. PCBs in Lake
Michigan Water Revisited. Environmental Science and Technology. 30: 1429-1436.
Rantanen, J. 1992. Industrial and Environmental Emergencies; Lessons Learned. Organohalogen
Compounds. 10,291-294.
Rappaport, R. A. and S. J. Eisenreich. 1988. Historical atmospheric inputs of high molecular weight
chlorinated hydrocarbons to eastern North America. Environmental Science and Technology. 22:
931-941.
Rochkind, M. L., J. W. Blackburn, and G. S. Sayler. 1986. Microbial Decomposition of Chlorinated
Aromatic Compounds. U.S. Environmental Protection Agency, Cincinnati, OH
EPA/600/2-86/090.
Shiu, W. Y. andD. Mackay. 1986. A Critical Review of Aqueous Solubilities, Vapor Pressures,
Henry's Law Constants, and Octanol-Water Partition Coefficients of the Polychlorinated
Biphenyls. Journal of Physical and Chemical Reference Data. 15 (2): 911 -929.
Simcik, M. F., H. Zhang, S. J. Eisenreich, and T. P. Franz. 1997. Urban contamination of the
Chicago/coastal Lake Michigan atmosphere by PCBs and PAHs during AEOLOS. Environmental
Science and Technology. 31: 2141-2147.
Simcik, M. F., I. Basu, C. W. Sweet, and R. A. Kites. 1999. Temperature dependence and temporal
trends of poly chlorinated biphenyl congeners in the Great Lakes atmosphere. Environmental
Science and Technology. 33: 1991-1995.
Simcik, M. F., R. M. Hoff, W. M. J. Strachan, C. W. Sweet, I. Basu, and R. A. Kites. 2000. Temporal
trends of semivolatile organic contaminants in Great Lakes precipitation. Environmental Science
and Technology. 34:361-367.
Stange, K., and D. L. Swackhamer. 1994. Factors Affecting Phytoplankton Species-specific
Differences in Accumulation of 40 Polychlorinated Biphenyls (PCBs). Environmental Toxicology
and Chemistry. 13(11): 1849-1860.
Stow, C. A. 1995. Factors associated with PCB concentrations in Lake Michigan salmonids.
Environmental Science and Technology. 29: 522-527.
April 2004 R-5
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Stow, C. A., S. R. Carpenter, and L. A. Eby. 1995. Evidence that PCBs are approaching stable
concentrations in Lake Michigan fishes. Ecological Applications. 5(1): 248-260.
Swackhamer, D. L., and D. E. Armstrong. 1987. Distribution and Characterization of PCBs in Lake
Michigan Water. Journal of Great Lakes Research. 13(1):24:36.
Swackhamer, D. L. and D. E. Armstrong. 1988. Horizontal and vertical distribution of PCBs in southern
Lake Michigan sediments and the effect of Waukegan Harbor as a point source. Journal of Great
Lakes Research. 14: 277-290.
Swackhamer, D. L., and R. S. Skoglund. 1993. Bioaccumulation of PCBs by Algae: Kinetics versus
Equilibrium. Environmental Toxicology and Chemistry. 12: 831-838.
Sweet, C. 2000. Sampling of Atmospheric PCBs in the Lake Michigan Mass Balance Study (LMMB),
1994-1995. [Internal document]. Great Lakes National Program Office.
Totten, L. A., C. L. Gigliotti, and S. J. Eisenreich. 2001. Re-evaluation of Air-water Exchange Fluxes of
PCBs in Green Bay and Southern Lake Michigan in AEOLOS. Abstracts: June 10-14, 2001.
International Association for Great Lakes Research, Green Bay, Wisconsin.
Trowbridge, A. G. and D. L. Swackhamer. 2001. An Analysis of Poly chlorinated Biphenyl
Concentrations in Lower Trophic Level Organisms of the Lake Michigan Foodweb. Abstracts:
June 10-14, 2001. International Association for Great Lakes Research, Green Bay, Wisconsin.
U.S. Department of the Interior. 1967. Water Quality Investigations, Lake Michigan Basin: Lake
Currents. Federal Water Pollution Control Administration Technical Report. Great Lakes Region,
Chicago, IL. 364 pp.
USEPA. 1975. Train Stops Manufacture of Heptachlor/Chlordane, Cites Imminent Cancer Risk. EPA
Press Release, July 30, 1975. (www.epa.gov/history/topics/legal/01.htm).
USEPA. 1984. Title 40 of the Code of Federal Regulations, part 136, Guidelines Establishing Test
Procedures for the Analysis of Pollutants, Appendix B, Definition and Procedure for the
Determination of the Method Detection Limit.
USEPA. 1988. Chlordane/Heptachlor termiticides; notification of cancellation and amendment of
existing stocks determination. Federal Register 53:11798-1 1805.
USEPA. 1995a. National Primary Drinking Water Regulations, Contaminant Specific Fact Sheets,
Inorganic Chemicals, Technical Version. U.S. Environmental Protection Agency, Office of
Water, Washington, DC. EPA 811/F-95/002-T.
USEPA. 1995b. National Primary Drinking Water Regulations, Contaminant Specific Fact Sheets,
Synthetic Organic Chemicals, Technical Version. U.S. Environmental Protection Agency, Office
of Water, Washington, DC. EPA 811/F-95/003-T.
USEPA. 1997a. Lake Michigan Mass Budget/Mass Balance Work Plan. U.S. Environmental Protection
Agency, Great Lakes National Program Office, Chicago, IL. EPA 905/R-97/018.
R-6 April 2004
-------
References
USEPA. 1997b. Mercury Study Report to Congress. U.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards and Office of Research and Development, Research Triangle
Park,NC. EPA 452/R-97/003.
USEPA. 1997c. The Enhanced Monitoring Program Quality Assurance Program Plan. U.S.
Environmental Protection Agency, Great Lakes National Program Office, Chicago, IL. EPA
905/R-97/017.
USEPA. 1997d. Lake Michigan Mass Balance Project (LMMB) Methods Compendium, Volume 1:
Sample Collection Techniques. U.S. Environmental Protection Agency, Great Lakes National
Program Office, Chicago, IL. EPA 905/R-97/012a.
USEPA. 1997e. Lake Michigan Mass Balance Project (LMMB) Methods Compendium, Volume 2:
Organic and Mercury Sample Analysis Techniques. U.S. Environmental Protection Agency, Great
Lakes National Program Office, Chicago, IL. EPA 905/R-97/012b.
USEPA. 1998. The Lake Michigan Mass Balance Project Quality Assurance Plan for Mathematical
Modeling, Version 3.0. U.S. Environmental Protection Agency, Great Lakes National Program
Office, Chicago, IL.
USEPA. 1999. National Recommended Water Quality Criteria-Correction. U.S. Environmental
Protection Agency, Office of Water, Washington, DC. EPA 822/Z-99/001.
USEPA. 2001a. Ambient Aquatic Life Water Quality Criteria for Atrazine. U.S. Environmental
Protection Agency, Office of Water, Washington, DC.
USEPA. 2001b. The Lake Michigan Mass Balance Study Quality Assurance Report. U.S.
Environmental Protection Agency, Great Lakes National Program Office, Chicago, IL. EPA
905/R-01/013.
USEPA. 2001c. Re suits of the Lake Michigan Mass Balance Study: Atrazine Data Report. EPA 905/R-
01/010. U.S. Environmental Protection Agency, Great Lakes National Program Office, Chicago,
IL.
USEPA. 2004. Results of the Lake Michigan Mass Balance Study: Mercury Data Report. EPA 905/R-
01/012. U.S. Environmental Protection Agency, Great Lakes National Program Office, Chicago,
IL.
Van den Berg, M, L. Birnbaum, A. T. C. Bosveld, B. Brunstrom, P. Cook, M. Feeley, J. P. Giesy, A.
Hanberg, R. Hasegawa, S. W. Kennedy, T. Kubiak, J. C. Larsen, F. X. R. van Leeuwen, A. K. D.
Liem, C. Nolt, R. E. Peterson, L. Poellinger, S. Safe, D. Schrenk, D. Tillitt, M. Tysklind, M.
Younes, F. Warn, and T. Zacharewski. 1998. Toxic Equivalency Factors (TEFs) for PCBs,
PCDDs, PCDFs for Humans and Wildlife. Environmental Health Perspectives. 106: 775-792.
Van Hoof, P. 2000. PCBs in Lake Michigan Surficial Sediments. [Internal document]. Great Lakes
National Program Office.
Willman, E. J., J. B. Manchester-Neesvig, and D. E. Armstrong. 1997. Influence of ortho-substitution on
patterns of PCB accumulation in sediment, plankton, and fish in a freshwater estuary.
Environmental Science and Technology. 31: 3712-3718.
April 2004 R-7
-------
Results of the LMMB Study: PCBs and trans-Nonachlor Data Report
Willman, E. J., J. B. Manchester-Neesvig, C. Agrell, and D. E. Armstrong. 1999. Influence of ortho-
substitution homolog group on polychlorobiphenyl bioaccumulation factors and fugacity ratios in
plankton and zebra mussels (Dreissenapolymorpho). Environmental Toxicology and Chemistry.
18(7): 1380-1389.
World Health Organization (WHO). 1988. IPCS International Programme on Chemical Safety. Health
and Safety Guide No. 13. Chlordane Health and Safety Guide.
(www.inchem.org/documents/hsg/hsg/hsg013.htm).
R-8 April 2004
------- |