&EPA
United States
Environmental Protection
Agency
Results of the Lake Michigan Mass
Balance Study: Mercury Data Report
February 2004
-------
U.S. Environmental Protection Agency
Great Lakes National Program Office (G-17J)
77 West Jackson Boulevard
Chicago, IL 60604
EPA 905 R-01-012
-------
Results of the Lake Michigan Mass Balance Study:
Mercury Data Report
Prepared for:
US EPA Great Lakes National Program Office
77 West Jackson Boulevard
Chicago, Illinois 60604
Prepared by:
Harry B. McCarty, Ph.D.,
Ken Miller,
Robert N. Brent, Ph.D., and
Judy Schofield
DynCorp (a CSC Company)
6101 Stevenson Avenue
Alexandria, Virginia 22304
and
Ronald Rossmann, Ph.D.
US EPA Office of Research and Development
Large Lakes Research Station
9311GrohRoad
Grosse He, Michigan 48138
February 2004
-------
Acknowledgments
This report was prepared under the direction of Glenn Warren, Project Officer, USEPA Great Lakes
National Program Office; and Louis Blume, Work Assignment Manager and Quality Assurance Officer,
USEPA Great Lakes National Program Office (GLNPO). The report was prepared by Harry B. McCarty,
Ken Miller, Robert N. Brent, and Judy Schofield, with DynCorp's Science and Engineering Programs,
and Ronald Rossmann, USEPA Large Lakes Research Station, with significant contributions from the
LMMB Principal Investigators for mercury and Molly Middlebrook, of DynCorp. GLNPO thanks these
investigators and their associates for their technical support in project development and implementation.
Ronald Rossmann wishes to thank Theresa Uscinowicz for assistance with collection, preparation, and
analysis of the samples; special thanks to staff of the NOAA Great Lakes Environmental Research
Laboratory, University of Wisconsin-Milwaukee Great Lakes Water Institute Center for Great Lakes
Studies, USEPA Great Lakes National Program, and USEPA Mid-Continent Ecology Division for
collection of the samples.
GLNPO also thanks the following reviewers of the draft report for their comments and observations:
Barbara Carney, U.S. Department of Energy; Dr. Malcolm Meaburn, NOAA/Great Lakes Environmental
Research Laboratory; Dr. Carl Watras, Environmental Research and Consulting; and Frank Anscombe
and Alexis Cain, EPA Region 5.
The information in this document has been funded wholly (or in part) by the U.S. Environmental
Protection Agency. Mention of trade names or commercial products constitute endorsement or
recommendation for use.
LMMB Principal Investigators for Mercury
Gerald Keeler, Ph.D. (atmosphere)
School of Public Health Environmental Health
Sciences
University of Michigan
Ann Arbor, Michigan
James Hurley, Ph.D. (tributary)
Bureau of Research
Wisconsin Department of Natural Resources
Monona Wisconsin, and
Water Science and Engineering Laboratory
University of Wisconsin
Madison, Wisconsin
Robert Mason, Ph.D. (open lake)
Chesapeake Biological Laboratory
University of Maryland
Center for Environmental Science
Solomons, Maryland
Ronald Rossmann, Ph.D. (sediment)
Large Lakes Research Station
USEPA
Grosse lie, Michigan
Edward Nater, Ph.D. (plankton)
Department of Soil, Water, and Climate
University of Minnesota
St. Paul, Minnesota
Jerome Nriagu, Ph.D. (fish)
School of Public Health
Department of Environmental Health Sciences
University of Michigan
Ann Arbor, Michigan
-------
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
1.3.1.2trans-Nonachlor
1.3.1.3Atrazine
1.3.1.4Mercury
1.3.2 Other Measured Parameters
1.3.3 Measured Compartments ..
1.4 Objectives
1.5 Design
1.5.1 Organization ....
1.5.2 Study Participants
-1
-1
-1
-2
-2
-2
-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
1.5.4.2Great Lakes Environmental Monitoring Database 1-11
1.5.4.3Public Access to LMMB Data 1-12
1.5.5 Quality Assurance Program 1-14
1.6 Project Documents and Products 1-15
Chapter 2 Mercury Study Overview 2-1
2.1 Mercury Introduction 2-1
2.1.1 Physical/Chemical Properties 2-1
2.1.2 Mercury Production, Uses, and Releases 2-1
2.1.3 Regulatory Background 2-3
2.1.4 Fate and Effects 2-4
2.1.5 Biological Transformations 2-5
2.1.6 Toxicity 2-5
2.2 Study Design 2-6
2.2.1 Description 2-6
2.2.2 Scope 2-7
2.2.3 Organization/Management 2-7
2.3 Sampling Locations 2-7
2.3.1 Atmospheric Components 2-7
2.3.2 Tributaries 2-8
2.3.3 Open Lake 2-11
2.3.4 Sediment 2-11
2.3.5 Lower Pelagic Food Web Organisms 2-14
2.3.6 Fish 2-14
2.4 Sampling Methods 2-15
2.4.1 Atmospheric Components 2-15
2.4.1.1 Vapor Fraction 2-15
2.4.1.2Particulate Fraction 2-15
2.4.1.3 Precipitation Fraction 2-16
-------
Results of the LMMB Study: Mercury Data Report
2.4.2 Tributaries 2-16
2.4.3 Open Lake 2-16
2.4.4 Sediment 2-16
2.4.5 Lower Pelagic Food Web Organisms 2-17
2.4.6 Fish 2-17
2.5 Analytical Methods 2-17
2.5.1 Atmospheric Components 2-18
2.5.1.1 Vapor Fraction 2-18
2.5.1.2Participate Fraction 2-18
2.5.1.3Precipitation Fraction 2-18
2.5.2 Tributaries 2-18
2.5.3 Open Lake Water 2-18
2.5.4 Sediment 2-18
2.5.5 Lower Pelagic Food Web Organisms 2-19
2.5.6 Fish 2-19
2.6 Quality Implementation and Assessment 2-19
Chapter 3 Mercury in Atmospheric Components 3-1
3.1 Results 3-1
3.1.1 Vapor Fraction 3-1
3.1.1.1 Geographical Variation 3-2
3.1.1.2Seasonal Variation 3-2
3.1.2 Particulate Fraction 3-4
3.1.2.1 Geographical Variation 3-4
3.1.2.2Seasonal Variation 3-5
3.1.3 Precipitation Fraction 3-6
3.1.3.1 Geographical Variation 3-6
3.1.3.2Seasonal Variation 3-8
3.2 Quality Implementation and Assessment 3-9
3.3 Data Interpretation 3-12
3.3.1 Atmospheric Sources 3-12
3.3.2 Seasonal Considerations 3-12
3.3.3 Regional Considerations 3-13
Chapter 4 Mercury in Tributaries 4-1
4.1 Results 4-1
4.1.1 Geographical Variation 4-2
4.1.1.1 Mercury 4-2
4.1.1.2Methylmercury 4-5
4.1.2 Seasonal Variation 4-9
4.1.3 Other Factors Affecting Tributary Mercury Concentrations 4-12
4.1.4 Mercury Forms 4-13
4.2 Quality Implementation and Assessment 4-15
4.3 Data Interpretation 4-19
4.3.1 Mercury Levels in Lake Michigan Tributaries 4-19
4.3.2 Comparison to Regulatory Limits 4-19
4.3.3 Seasonality 4-19
4.3.4 Regional Considerations 4-20
4.3.5 Mercury Fractions and Forms 4-20
IV
-------
Table of Contents
Chapter 5 Mercury in the Open-Lake Water Column 5-1
5.1 Results 5-1
5.1.1 Geographical Variation 5-2
5.1.2 Seasonal Variation 5-4
5.1.3 Vertical Variation 5-7
5.1.4 Mercury Forms 5-8
5.1.5 Other Factors Affecting Tributary Mercury Concentrations 5-9
5.2 Quality Implementation and Assessment 5-10
5.3 Data Interpretation 5-12
5.3.1 Mercury Levels in Lake Michigan 5-12
5.3.2 Comparison to Regulatory Limits 5-13
5.3.3 Lateral Variation 5-13
5.3.4 Temporal Variation 5-13
5.3.5 Vertical Variation 5-14
5.3.6 Mercury Fractions and Forms 5-15
Chapter 6 Mercury in Surficial Sediments 6-1
6.1 Introduction 6-1
6.1.1 Background 6-1
6.1.2 Study Objectives 6-1
6.2 Results 6-4
6.2.1 Mercury in Surficial Sediments 6-4
6.2.2 Mercury in Sediment Trap Samples 6-7
6.2.3 Moisture Content of Sediment Samples Collected by Ponar 6-10
6.2.4 Mercury Fluxes to Sediments 6-12
6.2.5 Horizontal Variation of Mercury and Mercury Fluxes 6-13
6.3 Quality Assurance 6-17
6.4 Data Interpretation 6-19
6.4.1 Comparison to Other Great Lakes Sediments 6-19
6.4.2 Comparison to Historical Lake Michigan Concentrations 6-20
6.4.3 Comparison to Historical Lake Michigan Horizontal Variations 6-25
6.4.4 Regional Lake Michigan Comparisons 6-28
6.4.5 Mercury Fluxes 6-28
6.4.6 Relative Importance of Regional Atmospheric Sources and Point Sources of
Mercury 6-30
6.5 Conclusions 6-30
Chapter 7 Mercury in Plankton 7-1
7.1 Results 7-1
7.1.1 Variation Among Sample Types 7-1
7.1.2 Temporal Variation 7-3
7.1.3 Geographical Variation 7-5
7.1.4 Bioaccumulation 7-5
7.2 Quality Implementation and Assessment 7-8
7.3 Data Interpretation 7-10
7.3.1 Mercury Levels in Lake Michigan Plankton 7-10
7.3.2 Seasonal Considerations 7-10
7.3.3 Bioaccumulation and Biomagnification 7-11
7.3.4 Other Interpretations and Perspectives 7-12
-------
Results of the LMMB Study: Mercury Data Report
Chapter 8 Mercury in Fish 8-1
8.1 Results 8-1
8.1.1 Variation Among Species 8-1
8.1.2 Factors Affecting Contaminant Concentrations 8-3
8.1.3 Geographical and Seasonal Variation 8-4
8.1.4 Bioaccumulation 8-5
8.2 Quality Implementation and Assessment 8-5
8.3 Data Interpretation 8-7
8.3.1 Comparison to Fish Advisory Levels 8-7
8.3.2 Regional Considerations 8-8
8.3.3 Factors Affecting Contaminant Concentrations 8-9
Chapter 9 Cross-Media Interpretations 9-1
9.1 Summary of Mercury Concentrations in Lake Michigan Compartments 9-1
9.2 Mercury Speciation 9-2
9.3 Bioaccumulation and Biomagnification 9-5
References R-l
-------
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. Components Sampled by Principal Investigators 2-7
Table 2-2. Watershed Characteristics for Tributaries Monitored in the LMMB Study 2-10
Table 2-3. Open-lake Cruise Dates 2-11
Table 2-4. Number of Fish Collected by Technique 2-14
Table 2-5. Number of Fish Collected by Species and Location 2-15
Table 3-1. Numbers of Atmospheric Samples Analyzed for Mercury 3-1
Table 3-2. Mean Mercury Concentrations Measured in the Vapor Phase 3-1
Table 3-3. Mean Mercury Concentrations Measured in the Particulate Phase 3-4
Table 3-4. Mean Mercury Concentrations by Station Measured in the Precipitation Phase 3-6
Table 3-5. Summary of Routine Field Sample Flags Applied to Mercury in Atmospheric Samples .3-10
Table 3-6. Data Quality Assessment for Mercury in Atmospheric Samples 3-11
Table 4-1. Number of Tributary Samples Analyzed for Mercury and Methylmercury 4-1
Table 4-2. Mean Mercury Concentrations Measured in Lake Michigan Tributaries 4-3
Table 4-3. Mean Methylmercury Concentrations Measured in Lake Michigan Tributaries 4-6
Table 4-4. Correlation of Tributary Mercury Levels with Tributary Flow 4-12
Table 4-5. Correlations of Total Mercury Levels in Lake Michigan Tributaries with Dissolved
Organic Matter (DOC), Particulate Organic Matter (POC), and Total Solids (TS) 4-13
Table 4-6. Percentages of Total Mercury Found in Various Forms 4-14
Table 4-7. Summary of Routine Field Sample Flags Applied to Mercury Data from
Lake Michigan Tributaries 4-15
Table 4-8. Summary of Routine Field Sample Flags Applied to Methylmercury Data from
Lake Michigan Tributaries 4-16
Table 4-9. Data Quality Assessment for Mercury Data from Lake Michigan Tributaries 4-18
Table 4-10. Data Quality Assessment for Methylmercury Data from Lake Michigan Tributaries .... 4-18
Table 5-1. Numbers of Open-Lake Samples Analyzed for Mercury 5-1
Table 5-2. Mean Particulate and Total Mercury Concentrations Measured in Open Lakes 5-3
Table 5-3. Mean Particulate and Total Mercury Concentrations by Cruise 5-4
Table 5-4. Mean Dissolved Mercury Concentrations by Cruise 5-9
Table 5-5. Summary of Routine Field Sample Flags Applied to Mercury in Open-lake Samples . . . 5-10
Table 5-6. Data Quality Assessment for Mercury in Open-lake Samples 5-12
Table 6-1. Concentrations of Mercury for each Lake Michigan Surficial Sediment Station 6-4
Table 6-2. Summary Statistics for Lake Michigan Surficial Sediment Mercury Concentrations 6-7
Table 6-3. Concentrations of Mercury in Sediment Trap Samples 6-8
Table 6-4. Mercury Summary Statistics for each Station at each Depth for Sediment Trap Samples . . 6-9
Table 6-5. Moisture Content of Samples Collected by Ponar 6-10
Table 6-6. Summary Statistics for Moisture Content Analyses of Samples Collected by Ponar . ... 6-12
Table 6-7. Net Mercury Flux to Lake Michigan Surface Sediments 6-12
Table 6-8. Summary Statistics for Net Mercury Fluxes to Lake Michigan Surface Sediments in
Depositional Basins 6-13
Table 6-9. Summary of Data Verification Flags Applied to Routine Field Sample Results for
Sediment Mercury 6-18
Table 6-10. Data Quality Assessment for Mercury in Sediment Samples 6-18
Table 6-11. Comparison of Lake Michigan Surficial Sediment Mercury Concentrations to
those at other Locations in the Great Lakes Basin 6-19
Table 6-12. Comparison of Current Lake Michigan Results to Historical Data 6-20
VII
-------
Results of the LMMB Study: Mercury Data Report
Table 6-13. Comparison of Lake Michigan Results at Station 15 to Historical Results for
the 0-3 cm Surficial Sediment Interval 6-24
Table 6-14. Comparison of Lake Michigan Results at Station 15 to Historical Results for
the 0-1 cm Surficial Sediment Interval 6-24
Table 6-15. Comparison of Mercury Concentrations in Various Basins of Lake Michigan for
Box Cores Only 6-28
Table 6-16. Comparison of Total Mercury Fluxes to Various Basins of Lake Michigan for
Box Cores Only 6-29
Table 6-17. Comparison of Total Mercury Fluxes for Lake Michigan Corrected for Cs-137
Focusing Factors to Fluxes for other Locations 6-30
Table 6-18. Comparison of Mercury Fluxes to Lake Michigan Surficial Sediments at Station 15
in 1981 and 1994 6-30
Table 7-1. Number of Plankton Samples Analyzed for Mercury in the LMMB Study 7-2
Table 7-2. Mercury Concentrations in Plankton Measured at Various Sampling Stations in
Lake Michigan 7-6
Table 7-3. Summary of Routine Field Sample Flags applied to Mercury in Plankton Samples 7-8
Table 7-4. Data Quality Assessment in Plankton Samples 7-10
Table 8-1. Number of Composite Fish Samples Analyzed for Mercury 8-1
Table 8-2. Mean Total Mercury Concentrations in Lake Michigan Fish (Wet-weight Basis) 8-2
Table 8-3. Mean Total Mercury Concentrations in Lake Michigan Fish (Dry-weight Basis) 8-3
Table 8-4. Summary of Routine Field Sample Flags for Fish Mercury 8-6
Table 8-5. Data Quality Assessment for Mercury in Fish Samples 8-7
Table 9-1. Summary of Samples from each Ecosystem Compartment with Detectable Levels
of Mercury 9-1
Table 9-2. Percent of Mercury Attributable to Methylmercury in Little Rock Lake 9-3
Table 9-3. Percent of Mercury Attributable to Methylmercury in 15 Lakes in Northern Wisconsin . . 9-4
VIII
-------
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 1-13
Figure 2-1. Global Mercury Cycle 2-5
Figure 2-2. Atmospheric Sampling Stations 2-8
Figure 2-3. Tributary Sampling Stations 2-9
Figure 2-4. Open-Lake Water Column Sampling Stations 2-11
Figure 2-5. Locations of Sediment Cores 2-12
Figure 2-6. Sediment Trap Locations 2-13
Figure 2-7. Sampling Stations for Lower Pelagic Food Web Organisms and Fish 2-14
Figure 2-8. Results from Intercomparison Study of Three LMMB Laboratories Analyzing
Mercury in Aqueous Samples 2-20
Figure 3-1. Mercury Concentrations in Atmospheric Vapor Measured at Four Lake Michigan
Shoreline Sites and One Out-of Basin Site (Bondville) 3-2
Figure 3-2. Arithmetic Monthly Means at each Station - Vapor Phase 3-3
Figure 3-3. Mercury Concentrations in Atmospheric Particles Measured at Five Lake Michigan
Shoreline Sites and One Out-of Basin Site (Bondville) 3-5
Figure 3-4. Arithmetic Monthly Means at each Station - Particulate Phase 3-6
Figure 3-5. Mercury Concentrations in Atmospheric Precipitation Measured at
Four Lake Michigan Shoreline Sites and One Out-of-basin Site (Bondville) 3-7
Figure 3-6. Arithmetic Monthly Means at each Station - Precipitation Phase 3-8
Figure 3-7. Volume-Weighted Monthly Means at each Station - Precipitation Phase 3-9
Figure 4-1. Total and Dissolved Mercury Concentrations in Lake Michigan Tributaries 4-4
Figure 4-2. Mean Total and Dissolved Mercury Concentrations Measured in
Lake Michigan Tributaries 4-5
Figure 4-3. Total and Dissolved Methylmercury Concentrations in Lake Michigan Tributaries .... 4-7
Figure 4-4. Mean Total and Dissolved Methylmercury Concentrations Measured in
Lake Michigan Tributaries 4-8
Figure 4-5. Seasonal Variation of Mercury Concentrations in Lake Michigan Tributaries 4-10
Figure 4-6. Seasonal Flow Patterns and Total Mercury Concentrations in Selected
Lake Michigan Tributaries 4-11
Figure 5-1. Mercury Concentrations Measured in Open-lake Water Column Samples 5-2
Figure 5-2. Particulate and Total Mercury Concentrations Measured in Open Lakes, by Cruise .... 5-6
Figure 5-3. Total Mercury Concentration versus Sample Depth During Stratified Conditions 5-7
Figure 5-4. Total Mercury Concentrations at Stations with Samples from Multiple Depths 5-8
Figure 6-1. Sampling Locations and Type of Sample Recovered between 1994 and 1996 6-2
Figure 6-2. Sediment Trap Locations 6-3
Figure 6-3. Mercury Concentrations (mg/kg) in Lake Michigan Surficial Sediments (1994-1996) .6-14
Figure 6-4. Lake Michigan Bathymetry with Depositional Basin Locations 6-15
Figure 6-5. Mercury Fluxes (ng/cm2/y) to Lake Michigan Surficial Sediments (1994-1996) 6-16
Figure 6-6. Station Locations for the 1969-1970 Kennedy et al. Mercury Results 6-21
Figure 6-7. Station Locations for the 1975 Cahill Mercury Results 6-22
Figure 6-8. Station Locations for 1981 Sediment Cores 6-23
Figure 6-9. Vertical Variation of Mercury in Core LM-81-HS 6-25
Figure 6-10. Mercury Concentrations (mg/kg) in 1969-1970 Lake Michigan Surficial Sediments . . 6-26
Figure 6-11. Mercury Concentrations (mg/kg) in 1975 Lake Michigan Surficial Sediments 6-27
IX
-------
Results of the LMMB Study: Mercury Data Report
Figure 7-1. Mercury Concentrations in Phytoplankton and Zooplankton Measured in
Lake Michigan 7-3
Figure 7-2. Mercury Concentrations in Phytoplankton (A) and Zooplankton (B) Measured in
Lake Michigan during Six Cruises 7-4
Figure 7-3. Mercury Concentrations in Phytoplankton (A) and Zooplankton (B) Measured at
Various Sampling Stations in Lake Michigan 7-7
Figure 8-1. Total Mercury Concentration (Wet-weight Basis) in Lake Michigan Fish 8-2
Figure 8-2. Relationship of Fish Length and Mercury Concentration 8-3
Figure 8-3. Total Mercury Concentrations in Lake Michigan Lake Trout of Various Sizes
from the Three Biological Sampling Stations 8-4
Figure 8-4. Percentage of Lake Michigan Coho Salmon and Lake Trout Samples within
each EPA- Recommended Fish Advisory Category 8-8
Figure 9-1. Mercury Concentrations in Various Components of the Lake Michigan Ecosystem .... 9-5
Figure 9-2. Biomagnification Factors for Mercury in Lake Trout (A) and Adult Coho (B) 9-6
-------
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 mercury data collected as part of the LMMB Study, and is one
in a series of data reports that documents the project.
Mercury is a naturally occurring transition metal, in Group II of the periodic table, with three possible
valences, or oxidation states, Hg°, Hg+1, and Hg+2. The principal mineral source of mercury in the
geosphere is cinnabar (HgS). Mercury also occurs as a trace element in other commercially significant
geologic deposits, including coal.
Elemental mercury is commonly used in barometers and thermometers. Its high reduction potential and
low resistivity make it ideal for use in battery cells, electrical switches, and fluorescent lamps. Elemental
mercury or inorganic mercury compounds are used as catalysts in the oxidation of organic compounds
and the production of chlorine and caustic soda. Elemental mercury is a principal component of the silver
amalgam used in dental fillings. Mercury may be used in gold mining operations because it forms an
amalgam with gold which then can be separated from the gold-bearing ore. Mercury compounds were
used for many years as antifungal agents in interior and exterior paints and at pulp and paper mills.
Global releases of mercury to the environment come from both natural and anthropogenic (caused by
human activity) sources. Many of these sources are the result of releasing geologically bound mercury to
the atmosphere. Once mercury enters the atmosphere, it becomes part of a global cycle of mercury
among land, water, and the atmosphere.
Study Design
In the LMMB Study, mercury was measured in atmospheric, tributary, open-lake water column, sediment,
lower pelagic food web organism, and fish samples. Methylmercury, a toxic organomercury compound
of environmental concern, also was measured in tributary samples. From March 1994 through October
1995, over 2300 samples were collected and analyzed by cold vapor atomic fluorescence spectrometry
(CVAFS) or cold vapor atomic absorption spectrometry (CVAA) (sediment samples only).
Atmospheric vapor, particulate, and precipitation samples were collected from five stations surrounding
Lake Michigan and one background station outside the Lake Michigan basin. Tributary samples were
collected from 11 rivers that flow into Lake Michigan. Open-lake water column samples were collected
from 15 sampling stations in Lake Michigan, 1 station 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
particulate matter were collected in sediment traps deployed at five stations in Lake Michigan. Samples
of phytoplankton and zooplankton were collected from 14 stations in Lake Michigan. Specimens of lake
ES-1
-------
Results of the LMMB Study: Mercury Data Report
trout and coho salmon were collected from eight stations in the lake and additional coho salmon were
collected from a hatchery used to stock Lake Michigan.
Mercury in Atmospheric Components
Vapor-phase mercury was detected in all of the samples collected from all LMMB Study stations.
Monthly composite concentrations of vapor-phase mercury ranged from 1.16 ng/m3 at the Chiwaukee
Prairie station to 2.2 ng/m3 at the IIT Chicago station. Vapor-phase mercury results exhibited a seasonal
trend, with higher concentrations occurring in summer months and lower concentrations occurring in
winter months. Vapor-phase mercury concentrations varied by sampling station. The urban station at IIT
Chicago had a higher mean monthly composite concentration for the duration of the study period than the
urban-influenced and rural sites.
Particulate-phase mercury was detected in all of the samples collected from all LMMB Study stations.
Concentrations of particulate-phase mercury in individual samples ranged from 1.05 pg/m3 at Sleeping
Bear Dunes to 494 pg/m3 at the IIT Chicago station. Particulate-phase mercury results exhibited a
seasonal trend at the Sleeping Bear Dunes station, with higher concentrations occurring in summer
months and lower concentrations occurring in winter months. However, there were no statistically
significant seasonal differences for the other five sampling stations. Particulate-phase mercury
concentrations varied by sampling station in a manner similar to that of the vapor-phase mercury
concentrations. The urban station at IIT Chicago had a higher mean monthly composite concentration for
the duration of the study period than the urban-influenced and rural sites.
Mercury was detected in all of the precipitation samples collected from the LMMB Study stations. The
mercury concentrations in individual samples of precipitation ranged from 2.09 ng/L at Sleeping Bear
Dunes to 137 ng/L at the rural Bondville station. The differences in precipitation mercury concentrations
between stations were much less significant than for the vapor-phase or particulate-phase samples. The
mean concentration at Sleeping Bear Dunes was significantly lower than those at IIT Chicago, Bondville,
and Chiwaukee Prairie, and the mean concentration at South Haven was significantly lower than that at
IIT Chicago. Seasonal differences in precipitation mercury concentrations were less evident than for the
other atmospheric phases, but summer concentrations tended to be higher than those in winter.
Mercury and Methylmercury in Tributaries
The dissolved mercury was detected in all of the samples from all of the tributaries. Dissolved mercury
concentrations in individual samples ranged from 0.202 ng/L in the Kalamazoo River to 40.8 ng/L in the
Fox River. The total mercury concentrations in individual samples ranged from 0.536 ng/L in the
Muskegon River to 191 ng/L in the Fox River. Particulate mercury concentrations were calculated as the
difference between the measured total and dissolved mercury concentrations. As a result of the low
concentrations of mercury present in many samples and the uncertainties in both the total and dissolved
measurement results, some of the calculated particulate mercury results were negative numbers. The
highest calculated particulate mercury concentration occurred in the Fox River at 153 ng/L.
The concentrations of dissolved and total mercury exhibited seasonal trends for many of the tributaries,
with higher mean concentrations occurring in the spring months and lower mean concentrations occurring
in winter months. However, the seasonal trends varied by tributary and many were tied to the seasonal
flow regimes in the rivers, which are dominated by high spring flows.
Methylmercury concentrations were often two orders of magnitude lower than the inorganic mercury
concentrations, with many samples having no detectable methylmercury in the dissolved phase. The
ES-2
-------
Executive Summary
seasonal trends in methylmercury concentrations varied by tributary and many were tied to the seasonal
flow regimes in the rivers, which are dominated by high spring flows.
Mercury in Open-lake Water
Total and participate mercury were detected in the majority of the samples collected from the open lake.
Except for the result of a single sample collected at Station 380, there was little difference in the mean
total or particulate mercury concentrations by station, nor were there any statistically significance
differences between the northern and southern portions of the lake. This relatively uniform distribution of
mercury within the lake is consistent with previous assessments that suggest that the primary source of
mercury is atmospheric rather than riverine.
Open-lake samples were collected at depths ranging from 1 to 150 m. There was only a weak correlation
between mercury concentrations and depth when the entire data set was examined. However, when only
the data for the summer and autumn were used, the correlations for total and particulate mercury became
stronger, as a result of the thermal stratification of the lake during these months. During periods of
stratification, samples collected at depths above 40 m generally had higher mercury concentrations.
Mercury in Sediments
Mercury was detected in all of the sediment samples and all of the sediment trap samples collected during
the study. Mercury concentrations in sediment samples ranged from 0.002 mg/kg to 0.260 mg/kg, while
concentrations in the sediment trap samples ranged from 0.021 mg/kg to 27 mg/kg.
Sediment mercury concentrations were higher along the eastern side of the lake and higher in the deeper
basins of the lake.
Mercury in Lower Pelagic Food Web Organisms
Except for one zooplankton sample, all plankton samples collected from Lake Michigan had detectable
concentrations of total mercury. Total mercury concentrations in phytoplankton ranged from 10.9 to 176
ng/g. Total mercury concentrations in zooplankton ranged from 11.0 to 376 ng/g. Total mercury
concentrations in zooplankton were statistically higher than those in phytoplankton.
Total mercury concentrations in zooplankton differed significantly by cruise, and were lowest in the
spring, peaked in late summer, and remained elevated throughout the fall. No statistically significant
differences in phytoplankton mercury concentrations were identified between cruises, although
phytoplankton mercury concentrations generally increased throughout the summer and were highest in
the fall.
Mercury bioaccumulation factors calculated in the LMMB Study were 1.07 x 105 for phytoplankton and
1.66 x 105 for zooplankton. These bioaccumulation factors are slightly higher than reported by other
researchers for other lakes in the region. LMMB Study results indicate the biomagnification of mercury
within the lower pelagic food web. Zooplankton mercury levels were significantly higher than
phytoplankton mercury levels. The biomagnification factor calculated between phytoplankton and
zooplankton in the LMMB Study was 1.55.
Mercury in Fish
Total mercury was detected in all of the fish samples collected for this study. Mercury concentrations in
adult lake trout ranged as high as 396 ng/g and averaged 139 ng/g. In coho salmon, mercury
ES-3
-------
Results of the LMMB Study: Mercury Data Report
concentrations ranged as high as 127 ng/g and averaged 79.9, 20.6, and 69.0 ng/g in hatchery, yearling,
and adult salmon, respectively. Mercury concentrations in lake trout were significantly higher than in
adult or yearling coho salmon. Adult coho salmon also were significantly higher in mercury
concentrations than yearling coho, which contained the lowest mean concentration of mercury.
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 mercury in Lake Michigan fish
were generally 105 to 106 times higher than total mercury concentrations in Lake Michigan water.
Bioaccumulation factors were 2.18 x 105 for yearling coho salmon, 7.58 x 105 for adult coho salmon, and
1.14 x 106 for adult lake trout.
Mercury concentrations in fish averaged 139 ng/g in lake trout and 69.0 ng/g in adult coho salmon. These
average values are approximately 10 times below the U.S. Food and Drug Administration's (FDA) action
level of 1000 ng/g (1 ppm) for fish tissue mercury content. Even the maximum mercury concentration
measured in the LMMB Study (396 ng/g) was well below the FDA action level. However, EPA guidance
for fish advisories is based on the methylmercury content of fish, and methylmercury was not measured in
fish in the LMMB Study. Therefore, the data from this study are not readily comparable to the EPA
guidance. However, based on the conservative assumption that 100% of total mercury was in the form of
methylmercury, 3% and 9% of lake trout and coho salmon, respectively, fell into the unrestricted
consumption category established in the EPA guidance for methylmercury. The most contaminated coho
salmon and lake trout specimens collected in the LMMB Study fell into the 4 meals/month and 2
meals/month restriction categories, respectively. For the average coho salmon sample, EPA guidance
would recommend restricting consumption to 12 meals per month; and for the average lake trout sample,
EPA guidance would recommend restricting consumption to 4 meals per month. This recommendation is
consistent with state-wide advisories for mercury that have been issued by several states. While Lake
Michigan fish mercury concentrations warrant some level of fish advisory, few fish advisories in Lake
Michigan have been based solely on mercury contamination, because Lake Michigan waters are generally
under more stringent fish advisories based on PCB contamination.
Mass Balance and Modeling Efforts
The data collection and quality assurance efforts described in this report were designed to support the
Lake Michigan Mass Balance 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, and will be described in later documents from GLNPO.
ES-4
-------
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 mercury 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).
1-1
-------
Results of the LMMB Study: Mercury Data Report
Figure 1-1. Simplified Mass Balance Approach
What is mass balance?
(rantftntcd
Svstem
Msiss
Massslol.f,i
= Massm,IKfol.n,cll + Massslorf(i + Mm,,,,,
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.
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.
1.3.1.1 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
1-2
-------
Project Overview
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).
1.3.1.2 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, fra«s-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.
In the LMMB Study, fra«s-nonachlor was selected as a model for the cyclodiene pesticides (USEPA,
1997a).
1-3
-------
Results of the LMMB Study: Mercury Data Report
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
-------
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).
1-5
-------
Results of the LMMB Study: Mercury 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
-------
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 — Phytoplankton and zooplankton were sampled and analyzed for species
diversity, taxonomy, and contaminant burden. 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
1-7
-------
Results of the LMMB Study: Mercury 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
Sheboygan
Milwaukee
Milwaukee River
Chiwaukee Prairie
NT Chicago
Chicago SWFP Intake
Grand Calumet Harbor
Muskegon River
Muskegon
Grand River
Kalamazoo River
^ South Haven
St. Joseph River
Benton Harbor
Indiana Dunes
A Atmospheric Station
Q Tributary Station
Biota Station
• Sediment Station
^ Water Column Station
1-8
-------
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 Environment 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.
1-9
-------
Results of the LMMB Study: Mercury 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
-------
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.
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
1-11
-------
Results of the LMMB Study: Mercury Data Report
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.
1.5.4.3 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).
1-12
-------
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
Store, Transmit, and
Upload Data to
GLENDA
GLNPO QA
Workgroup
Conduct Data
Verification (Merge Field
and Analytical Data using
RDMQ)
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
External Parties
Input Data to Study
Models
1-13
-------
Results of the LMMB Study: Mercury Data Report
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
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.
1-14
-------
Project Overview
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.
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 website 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
1-15
-------
Results of the LMMB Study: Mercury Data Report
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.
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 website 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 website
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 mercury, data reports are
being published for atrazine (USEPA, 200Ic) and PCBs and frara-nonachlor (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
-------
Chapter 2
Mercury Study Overview
2.1 Mercury Introduction
2.1.1 Physical/Chemical Properties
Mercury is a naturally occurring transition metal, in Group II of the periodic table, along with zinc and
cadmium. The atomic number for mercury is 80 and its atomic weight is 200.59 g/mole. Mercury is the
only metal that occurs in a liquid state at typical environmental temperatures. The melting point of
mercury is -39.87 °C, and its boiling point is 356.58 °C. Mercury has a density of 13.59 g/cm3 and a
vapor pressure of 0.00185 mm at 25 °C.
The solubility of mercury in water is approximately 0.28 ^moles/L (56.2 i-ig/L) at 25 °C. Its electrical
resistivity is 95.76 ^ohm-cm at 20 °C, making it an excellent electrical conductor. In fact, the value of
the ohm is formally defined on the basis of the resistance of a column of mercury of specific dimensions.
Mercury occurs naturally in the environment with three possible valences, or oxidation states, Hg°, Hg+1,
and Hg+2. The principal mineral source of mercury in the geosphere is cinnabar (HgS). Mercury is
extracted from this ore by roasting in an oxygen atmosphere to produce elemental mercury, which can be
further purified by distillation. Mercury also occurs as a trace element in other commercially significant
geologic deposits, including coal.
The reduction potential is 0.851 volts for the reaction:
Hg + 2 + 2e- +> Hg°
and 0.796 volts for the reaction:
Hg;2 + 2e~ <-x 2Hg°
placing mercury higher on the redox scale than most other metals.
2.1.2 Mercury Production, Uses, and Releases
Because it is a dense liquid at typical environmental temperatures and responds in a predictable fashion to
changes in temperature and pressure, elemental mercury is commonly used in barometers and
thermometers. Its high reduction potential and low resistivity make it ideal for use in battery cells,
electrical switches, and fluorescent lamps.
Elemental mercury is also used as a catalyst in the oxidation of organic compounds and the production of
chlorine and caustic soda. Elemental mercury is a principal component of the silver amalgam used in
dental fillings. Mercury may be used in gold mining operations because it forms an amalgam with gold
which then can be separated from the gold-bearing ore. It has been used in chlor-alkali plants around the
world. Historically, mercury compounds have been used in medicinal products, including topical
disinfectants such as Mercurochrome, and as a preservative in some vaccines and cosmetics. For many
years, mercuric chloride was used as a biocide to preserve water samples collected for analyses of other
environmental contaminants. Mercury compounds were used for many years as antifungal agents in
interior and exterior paints and at pulp and paper mills.
2-1
-------
Results of the LMMB Study: Mercury Data Report
According to the U.S. Geological Survey (USGS), there have been no domestic mines producing mercury
as a primary product since 1990 (USGS, 1999). Virtually all domestic mercury production involves
recovery or recycling of mercury from secondary sources such as spent batteries, mercury-containing
lamps, switches, dental amalgams, and wastes from laboratories and electrolytic processes.
Data from USGS for the period from 1995 to 1999 indicate that domestic production of mercury (from
secondary sources), as well as imports and exports of mercury, and industrial consumption of mercury
declined. In addition, world-wide mine production declined by approximately 40% over the same period,
from 3,190 to 1,970 metric tons. USGS estimated that domestic industrial consumption of mercury in
1997 was 346 metric tons (762,800 pounds). Data from EPA's Toxics Release Inventory (TRI) for 1997
indicate that 73,334 pounds of mercury (~ 10% of domestic production) were released to the environment
by facilities that were required to report releases to EPA. According to USGS, electrolytic production of
chlorine and caustic soda account for roughly half of the domestic use of mercury, with electrical
applications and products accounting for another 25%.
Global releases of mercury to the environment come from both natural and anthropogenic (caused by
human activity) sources. Many of these sources are the result of releasing geologically bound mercury to
the atmosphere. Once mercury enters the atmosphere, it becomes part of a global cycle of mercury
among land, water, and the atmosphere. In its 1997 Report to Congress on mercury, EPA estimated that
the global mercury cycle involved the release of 5,500 metric tons (12,130,000 pounds) of mercury to the
atmosphere from all natural and anthropogenic sources world-wide (USEPA, 1997b). Of that total, EPA
estimated that 158 metric tons (348,300 pounds) were contributed from anthropogenic sources in the U.S.
in 1994 - 1995, representing about 3% of the total global mercury input to the atmosphere. Of that 158
metric tons, approximately 87% came from combustion sources, and approximately 10% came from
manufacturing sources. A breakdown of these 1994 - 1995 anthropogenic emission estimates includes:
Combustion sources (87%)
- Coal-fired utility boilers (32.6%)
Municipal waste combustors (18.7%)
Commercial/industrial boilers (17.9%)
Medical waste incinerators (10.1%)
Hazardous waste combustors (4.4%)
All other combustion sources (3.3%)
• Manufacturing sources (10%)
- Chlor-alkali plants (4.5%)
Portland cement kilns - excludes those that burn hazardous waste (3.1%)
All other manufacturing sources (2/
Although it does not involve quantities of mercury similar to those used on an industrial scale, elemental
mercury is used in various cultural and religious practices of some Caribbean and Latin American
immigrants to the U.S., which may result in exposures that exceed current occupational standards (Riley,
et al, 2001). Frequently reported uses of mercury in such practices include those designed to bring luck
or ward off evil by:
Carrying a capsule, vial, or pouch containing elemental mercury on one's person
• Sprinkling it in a home or car
• Mixing it with perfume
• Burning a candle laced with mercury
Elemental mercury has also been used as a folk medicine treatment for gastroenteritis among some
Mexican Americans.
2-2
-------
Mercury Study Overview
In another study of such cultural and religious practices, Johnson (1999) reported that 64% of the mercury
users in that study in New York City dispose of mercury by throwing it in the trash, 27% flushed used
mercury down the toilet, and 9% disposed of mercury outdoors. Therefore, although the overall
quantities of mercury used in these cultural practices may pale in comparison to industrial uses, the
uncontrolled disposal practices could make such cultural uses significant sources of mercury to local
environments.
2.1.3 Regulatory Background
Efforts in the U.S. to regulate releases of mercury to the environment began shortly after the formation of
EPA in 1970. EPA regulates mercury under a wide range of environmental statutes. By 1976, the Office
of Water listed mercury as one of the 129 pollutants in the consent decree that resulted from NRDC v.
Train (8 ERC 2120, 1976). As a result, mercury is 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 0.012 i-ig/L of mercury. The freshwater acute WQC is 2.1
|ig/L. The WQC for human health is 0.05 |ig/L.
Under the Safe Drinking Water Act, EPA established a maximum contaminant level (MCL) of 2 i-ig/L in
1992. Under the auspices of the Resource Conservation and Recovery Act (RCRA), EPA placed mercury
on Appendix VIII (hazardous substances) and Appendix IX (groundwater monitoring), and established a
Universal Treatment Standard (UTS) of 25 |ig/L of mercury in non-wastewaters when subjected to the
toxicity characteristic leaching procedure (TCLP) and 150 i-ig/L in wastewaters. Mercury is included in
the Toxics Release Inventory (TRI) developed under the Emergency Planning and Community Right to
Know Act (EPCRA).
The use of mercury in paints was discontinued in 1991 under the Federal Insecticide, Fungicide, and
Rodenticide Act (FIFRA). Registrations of the last two mercury-based pesticides (Calochlor and
Calogran) were voluntarily cancelled by the manufacturer in 1993. In 1996, Congress enacted the
Mercury-Containing and Rechargeable Battery Management Act to phase out the use of mercury in
batteries. The Act limits the mercury content of "button" batteries to 25 mg per battery, prohibits the sale
of most other types of batteries containing mercury, and requires that manufacturers identify suitable
recycling facilities for any mercuric-oxide batteries it sells.
Mercury and mercury compounds are classified as hazardous air pollutants (HAPs) under the Clean Air
Act, and EPA has established national emission standards for mercury in five source categories: ore
processing facilities, mercury cell chlor-alkali plants, sewage sludge drying operations, municipal waste
combustors, and medical waste incinerators.
Discharges of mercury have been significantly limited under the Great Lakes Initiative (GLI), in
recognition of the impact of mercury on the Great Lakes ecosystem and the associated effects on human
health in the region. In 1995, EPA issued GLI guidance that recommends that a water quality criterion of
1.8 ng/L (0.0018 i-ig/L) for dissolved mercury for the protection of human health (FR Vol. 60 No. 56,
March 23, 1995, pp. 15366-15425).
Under the Federal Food, Drug, and Cosmetic Act, the Food and Drug Administration (FDA) banned most
uses of mercury in over the counter medications and limited the concentrations of mercury used as
preservatives in eye-area cosmetics. The FDA also regulates the use of mercury in dental amalgams,
classifying the silver-mercury alloy as a Class II medical device, thereby subjecting it to additional
controls and imposing safety regulations on its use and disposal.
2-3
-------
Results of the LMMB Study: Mercury Data Report
2.1.4 Fate and Effects
Unlike synthetic organic contaminants, mercury is a naturally occurring element, and therefore it cannot
be created or destroyed by chemical, biological, or physical processes. Rather, mercury can be
transformed by oxidation or reduction reactions, or it can combine with other elements to form inorganic
or organic mercury compounds. The organomercury compounds are characterized by a covalent bond
between the mercury atom and a carbon atom, making mercury unusual among metals (but not unique), in
that many metals form only ionic bonds with other elements.
The following are the mercury compounds most likely to be found under environmental conditions:
mercuric chloride (HgCl2), mercuric hydroxide (Hg [OH]2), mercuric sulfide (HgS), methylmercuric
chloride (CH3HgCl), methylmercuric hydroxide (CH3HgOH), and dimethyl mercury ([CH3]2Hg)
(USEPA, 1997b).
Due to the volatility of elemental mercury, the atmosphere is both an important reservoir and a major
component of the global mercury cycle. That global cycle encompasses the flux of mercury in its many
forms to and from the atmosphere, fresh and marine water bodies, and the land. The cycle includes a
natural component that is the result of mercury that originated in geologic deposits and that has been
released from those deposits by natural processes. The cycle has been significantly perturbed or modified
by human activities, and includes both regional and local sources and sinks of various forms of mercury.
Although a detailed discussion of the global mercury cycle is beyond the scope of this report, in general
terms, the cycle (Figure 2-1) is characterized by the following exchanges and transformations of mercury:
• Volatilization from land-based sources to the atmosphere
Volatilization from marine-based sources to the atmosphere
• Deposition from atmosphere to land, oceans, and other water bodies
Anthropogenic inputs of gaseous and particulate forms of mercury to the atmosphere from
combustion processes and municipal and industrial sources on land
Run-off of natural and anthropogenic mercury from land to freshwaters and oceans
• Exchanges between dissolved and particulate forms of mercury in oceans and lakes
Exchanges of mercury between inorganic and organic forms in the water and sediments of oceans and
lakes
Deposition of mercury in sediments of oceans and lakes
• Local and regional deposition of mercury from anthropogenic combustion sources and municipal and
industrial sources
2-4
-------
Mercury Study Overview
Figure 2-1. Global Mercury Cycle
Anthropogenic
Emissions
Re-emitted
Anthropogenic
& Natural
Re-emitted
Anthropogenic
& Natural
Local & Regional
Deposition
Global Terrestrial
Deposition
Global Marine
Deposition
Hg° -+— CH,Hg
Particulate
Removal
Adapted from Mason, Fitzgerald, and Morel (1994)
The residence time of elemental mercury in the atmosphere is estimated to be about one year (EPA,
1997b). As a result, mercury entering the atmosphere from any given source may be distributed globally,
making mercury a ubiquitous contaminant.
2.1.5 Biological Transformations
Mercury enters the food web primarily through aquatic systems, where it is associated with dissolved and
particulate forms of organic carbon (DOC and POC), and where it may undergo methylation by bacteria
in sediments or in the water column to form methylmercury (USEPA, 1997b). Methylmercury
accumulates in the tissues of aquatic organisms and methylmercury concentrations are magnified in
aquatic food webs, with highest concentrations often found in the top predators, including many game
fish. As a result, human exposure pathways related to terrestrial plants and grazing animals are much less
important than pathways related to consumption offish (USEPA, 1997b).
2.1.6 Toxicity
The effects of mercury exposure on organisms depend on the route of exposure and the form of mercury.
Many people are familiar with the "Mad Hatter" in Lewis Carroll's "Alice in Wonderland," whose
madness described the results of exposure of hatmakers to the mercuric nitrate used to shrink felt for hats.
While the etymology of the expression "mad as a hatter" is apparently subject to some debate, the effects
of exposure to elemental mercury vapors and/or soluble mercury salts were documented at the time. The
"Danbury shakes" was the name given to the neurological effects exhibited by hatmakers in Danbury,
Connecticut, in the 19th century.
Whether the route of exposure is through inhalation, dermal exposure, ingestion of food, or other means,
mercury and mercury compounds are readily transported throughout humans and animals by blood
circulation. Elemental mercury dissolved in the blood can cross the blood/brain barrier, where it can
accumulate in nerve tissue. Symptoms of chronic exposure to mercury vapors include: excitability,
2-5
-------
Results of the LMMB Study: Mercury Data Report
confusion and mental instability, personality changes, and fine tremors in the extremities. Mercury can
cause kidney damage, as the kidneys works to remove mercury from the bloodstream.
The effects of organomercury compounds, particularly methylmercury and dimethylmercury, are more
severe than for elemental mercury, given equivalent exposures or doses. Methylmercury is known to
have teratogenic effects in the children of mothers exposed to this organomercury compound. Mild
maternal exposures cause mainly neurological effects in the children, including developmental delays,
reduced intelligence, and altered muscle reflexes.
Much of the data on the direct effects of elemental and organomercury exposure are the result of studies
of long-term exposures of the people living around Minamata Bay, on the western coast of Kyushu, in
Japan. Beginning in 1956, a series of patients were identified as exhibiting symptoms of severe
convulsions, intermittent loss of consciousness, altered mental state, and ultimately permanent coma and
death. The common link among the patients was that they consumed large quantities offish from
Minamata Bay. A second outbreak of what became known as "Minamata disease" occurred in 1965 when
patients with the same symptoms were identified near Niigata City, far from Minamata. The affected
individuals were all fishermen living along the Agano River. In these cases, methylmercury was
identified in both the local fish that the patients consumed as well as in tissues from the patients' bodies.
Ultimately, the Japanese government publicly acknowledged that Minamata disease resulted from
environmental pollution. The source of the pollution in Minamata Bay was the untreated effluent from
the Nippon Chisso chemical manufacturing plant in Minamata City. Nippon Chisso produced
acetaldehyde and polyvinyl chloride (PVC) at the Minamata plant, and used large quantities of inorganic
mercury compounds as reaction catalysts. Although most of the mercury was recovered within the plant,
massive amounts were discharged in the wastewater over a period of decades, and much of it accumulated
in the sediments and biota of the bay. The methylmercury found in fish from the Agano River was
ultimately traced to the Showa Denko Company facility in Kase, on the upper reaches of the river (Ui,
1992).
The extreme toxicity of dimethylmercury came to the attention of the scientific community most recently
as the result of a tragic laboratory accident. In August 1996, Dr. Karen Wetterhahn, working at
Dartmouth College, was exposed to approximately 400 milligrams of dimethylmercury when a few drops
of a standard she was using to calibrate a nuclear magnetic resonance instrument accidentally spilled on
the back of her latex glove. The spill occurred in a hood and she cleaned up the spill and removed the
glove. Five months after the accident, she was admitted to the hospital exhibiting problems with her
speech, balance, and gait. Twenty-two days after the onset of these neurological symptoms, she did not
respond to visual or verbal stimuli, and lapsed into a coma. She died in June 1997, almost 300 days after
the accident (Nierenberg etal., 1998).
2.2 Study Design
2.2.1 Description
Mercury was chosen for inclusion in the LMMB Study as a representative of persistent, bioaccumulative
metals. Mercury was measured in vapor, precipitation, particulates, atmospheric dry deposition, water in
the open lake, tributaries, sediment, lower pelagic food web organisms, and top predator fish. The data
generated from this study were used to estimate an overall mass balance of mercury in Lake Michigan
(see Section 1.4). In addition, methylmercury was determined in tributary samples.
2-6
-------
Mercury Study Overview
2.2.2 Scope
To develop a mass balance for mercury in Lake Michigan, all significant sources and stores of mercury 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 2-1.
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.
2.2.3 Organization/Management
The responsibility for collecting and analyzing mercury samples from the various components was
divided among six principal investigators (Pis, see Table 2-1). Each principal investigator developed a
quality assurance project plan (QAPP) that was submitted to EPA's Great Lakes National Program Office
(GLNPO) for approval. 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 through 8.
Table 2-1. Components Sampled by Principal Investigators
Ecosystem Compartment
Atmosphere
Tributary
Open Lake
Sediment
Lower Pelagic
Food Web Organisms
Fish
Component
Vapor
Particulate
Precipitation
Dissolved Mercury and Methylmercury
Total Mercury and Methylmercury
Particulate matter
Total mercury
Surficial sediment
Resuspended sediment
Zooplankton
Phytoplankton
Lake Trout
Coho Salmon
Principal Investigator
Gerald Keeler, Ph.D., University of Michigan
School of Public Health Environmental Health
Sciences
James Hurley, Ph.D., University of Wisconsin
Water Science and Engineering Laboratory
Robert Mason, Ph.D., University of Maryland
Chesapeake Biological Laboratory
Ronald Rossmann, Ph.D., USEPA Large Lakes
Research Station
Edward Nater, Ph.D., University of Minnesota
Department of Soil, Water, and Climate
Jerome Nriagu, Ph.D., University of Michigan
Department of Environmental Health Sciences
School of Public Health
2.3 Sampling Locations
2.3.1 Atmospheric Components
Atmospheric samples were collected at five shoreline sampling stations and two open-lake sampling
stations within Lake Michigan (Figure 2-2). One of the shoreline sampling stations (George Washington
High School in Chicago) was used only once over the course of the study. In addition, one out-of-basin
land-based sampling station was established as a regional background site 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 field
2-7
-------
Results of the LMMB Study: Mercury Data Report
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 Figure 2-2. Atmospheric Sampling Stations
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 agricultural areas).
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 the
land-based IADN stations at Sleeping Bear Dunes 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 mercury sampling stations are shown in
Figure 2-2.
Atmospheric samples were collected from the R/V
Lake Guardian at two stations (Fig. 2-2) in the open
lake in July 1994 and January 1995. However, because
of the limited spatial and temporal coverage
represented by these open-lake atmospheric samples,
they were not included in the LMMB Study data set,
nor are they discussed in this report.
For vapor and particulate samples, one 24-h composite
sample was collected every 6 days using automated
sampling equipment. Precipitation samples were
collected by automated equipment that sensed the
presence of precipitation and collected samples from
each precipitation event during April through October.
Precipitation samples collected in November through March were collected on a weekly basis (e.g., each
sample represented the precipitation that fell during all of that week). 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.
2.3.2 Tributaries
Tributary samples were collected from 11 rivers that flow into Lake Michigan (Figure 2-3). 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.
2-8
-------
Mercury Study Overview
The 11 monitored tributaries represent greater Figure 2-3. Tributary Sampling Stations
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. 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.
Table 2-2 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 11 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.
2-9
-------
Table 2-2. 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
£ 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
aEPA'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.
2-10
-------
Mercury Study Overview
2.3.3 Open Lake
Open-lake water column samples were collected from Figure 2-4. Open-Lake Water Column Sampling
17 sampling locations on Lake Michigan, one Stations
sampling location in Green Bay, and one sampling
location on Lake Huron (Figure 2-4). Open-lake
samples were collected during six cruises of the R/V
Lake Guardian between June 1994 and September
1995. The dates of the six cruises are shown in Table
2-3.
Table 2-3. Open-lake Cruise Dates
Cruise Date
'*<
June 1994
August 1994
October/November 1994
March/April 1995
August 1995
September/October 1995
The first cruise during which mercury samples were
collected was in early summer (June 1994), after the
onset of stratification. The second and third surveys
were in late summer (August 1994) and fall (October
1994), during later stages of stratification. The fourth survey was conducted in March 1995, during non-
stratified conditions. The fifth and sixth surveys occurred in August and September 1995, during
stratification.
During stratification, samples were collected from two or three depths to represent the epilimnion and the
hypolimnion. When the water column was unstratified, samples at some stations were collected from
mid-depth, while at other stations, samples were collected from two depths.
2.3.4 Sediment
In 1994, 1995, and 1996, sediment samples were collected from Lake Michigan by box coring, Ponar
grabs, and gravity coring. The location of the sediment sampling stations and the sampling device used
are shown in Figure 2-5. The sediment sampling locations were selected to help define the three
depositional zones (depositional, transitional, and non-depositional).
In addition to grab samples of sediments, sediment traps were deployed at eight locations in Lake
Michigan (see Figure 2-6). The trap at Station 3, excluded from the figure but located in northern Lake
Michigan, was lost. Samples from the two traps at Station 6 had mercury chloride added as a preservative
to their collection bottles prior to deployment and therefore were not analyzed. The trap placed at a depth
of 245 m at Station 5 failed, and no sample was available from the trap at Station 4. Enough sample was
available for mercury analysis from Stations 1, 2, 5, 7, and 8. Samples from two depths were available
from Stations 7 and 8.
2-11
-------
Results of the LMMB Study: Mercury Data Report
Figure 2-5. Locations of Sediment Cores
A
Scale
ZZ I
50km 100km 150km
Lake Michigan Mass Balance
Project Sample Locations (1994-1996
»65 • /Pentwater
No Sample Collected or Available
Box Core Sample
Ponar Grab Samole
•* \ Grand Haven
2-12
-------
Mercury Study Overview
Figure 2-6. Sediment Trap Locations
Lake Michigan Mass Balanc
Sediment Trap Locations
2-13
-------
Results of the LMMB Study: Mercury Data Report
2.3.5 Lower Pelagic Food Web Organisms
Plankton samples were collected from 12 stations in Figure 2-7. Sampling Stations for Lower Pelagic Food
Lake Michigan selected by GLNPO and the Pis in Web Organisms and Fish
advance of sampling (Figure 2-7). The stations
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 2-7. Samples were
collected on several occasions, from June 1994 to
September 1995.
In addition, zooplankton samples were collected from
Station 10M in January 1995 and phytoplankton
samples were collected from Stations 23M and 41 in
June 1994. A total of 72 zooplankton and 71 phyto-
plankton samples were collected during the study.
2.3.6 Fish
Lake Michigan fish were collected from April 1994
through October 1995 for total mercury analysis.
Lake trout and coho salmon were collected using gill
nets, trawl nets, or other appropriate means (Table 2-
4). Up to five individual fish of the same species and
size or age category were combined to produce composite fish samples at each collection. In total, 693 adult
lake trout from 172 to 933 mm in length were collected from three of the four biological sampling areas or
biota boxes shown in Figure 2-7 (fish were not collected from the biota box at Station 5, near Chicago):
• Sturgeon Bay biota box — a series of three nearshore stations (110, 140, and 180) 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 (240 and 280) in the central
Lake Michigan basin near Port Washington, Wisconsin
• Saugatuck biota box — a series of three nearshore stations (310, 340, and 380) on the eastern side of
the southern Lake Michigan basin near Saugatuck, Michigan
Table 2-4. Number of Fish Collected by Technique
Species
Lake Trout
Coho salmon — adult
Coho salmon — yearling
Coho salmon — hatchery
Number of Fish Collected by Technique
Hook and Line
—
135
29
—
Gill Net
666
3
—
—
Bottom Trawl
27
—
—
—
Harvest Weir
—
—
9
—
Dip Net
—
—
—
25
These fish were used to prepare 156 trout composite samples that were analyzed for total mercury by cold
vapor atomic fluorescence spectroscopy (Table 2-5).
2-14
-------
Mercury Study Overview
A total of 201 coho salmon were collected in three distinct age classes (hatchery, yearlings, and adult).
Of the 201 fish, 138 were adult coho salmon collected from 54 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. These 138 adult coho salmon were used to prepare 32
composite samples for mercury analyses (Table 2-5). In addition, 38 yearling coho salmon were collected
from 22 locations to create 8 composite samples, and 25 young (hatchery) coho salmon were collected
directly from the Platte River hatchery, where the majority of Lake Michigan stocked salmon originate,
and were used to create 5 composite samples.
Table 2-5. Number of Fish Collected by Species and Location
Species
Lake Trout
Coho salmon — adult
Coho salmon — yearling
Coho salmon — hatchery
Total Number of Individual
Fish Collected
693
138
38
25
Number of Locations
3
54
22
1
Number of Composite
Samples Created
156
32
8
5
2.4 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 1997e). Field sampling for all media except sediment and fish adhered
to strict protocols for the sampling of trace metals using "clean" techniques. Sampling personnel were
outfitted with suits and gloves, "clean hands/dirty hands" techniques were employed, and pre-cleaned
polytetrafluoroethylene bottles and equipment were used. "Clean" techniques were not used for the
collection of sediments or fish, because these matrices were believed to contain significantly higher
mercury concentrations, so contamination from background sources would be less of a concern. Brief
summaries of the sampling procedures are provided below.
2.4.1 Atmospheric Components
2.4.17 Vapor Fraction
Vapor-phase mercury was quantitatively removed from air by amalgamation onto gold. Two gold-coated
borosilicate glass bead traps in quartz tubing (with glass fiber pre-filters) were used in series. The traps
were housed in a sampling box 3 m above the ground and maintained at 93 °C to prevent condensation.
Samples were collected for 12-24 hours at flow rates of 10 to 30 L/min.
2.4.12 Particulate Fraction
Particulate atmospheric components were collected using a filter pack assembly containing pre-treated
47-mm glass fiber filters housed in custom-made sampling boxes. The volume of air sampled was
measured with a calibrated dry test meter. The vacuum pumps attached to the sampling boxes were
specially designed for trace level mercury sampling. The apparatus was deployed 3 m above the ground,
and samples were collected for 12-24 hours at flow rates of 10 to 30 L/min.
2-15
-------
Results of the LMMB Study: Mercury Data Report
2.4.1.3 Precipitation Fraction
Precipitation samples were collected by automated equipment that sensed the presence of precipitation
and collected samples from each precipitation event during July 1994 to October 1994, and during each
precipitation event from April 1995 to October 1995. Precipitation samples collected from November
1994 through March 1995 were collected on a weekly basis (e.g., each sample represented the
precipitation that fell during all of that week). An automated sensor grid on the modified collector was
activated by precipitation, causing the lid of the sampler to open for wet-only collection of precipitation
samples. Samples were collected through a borosilicate funnel and in 1-L Teflon® bottles.
2.4.2 Tributaries
A small boat was anchored at the sampling site, above the centroid of the river. Water samples (500 mL)
were collected from two depths (0.2 x river depth and 0.8 x river depth). Water was pumped through a
Teflon® sampling tube (weighted with a Teflon® weight) and C-flex® pumphead tubing using a peristaltic
pump. Dissolved samples were collected using in-line filtration. Mercury samples were preserved in the
field with 10 mL of 50% HC1. Samples from the upper and lower depths were composited.
2.4.3 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. 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.
Teflon®-lined Go-Flo bottles were attached to Kevlar® lines with non-metallic weights. Two liters of
sample were collected for total mercury analysis. Samples were aliquotted and filtered in a clean room
onboard the ship. Particulate samples were collected onto 0.8-^m quartz fiber filters. Samples were
frozen on board and shipped overnight to the laboratory.
2.4.4 Sediment
Sediment samples were collected from 118 stations in Lake Michigan using two types of equipment
(Figure 2-5). Wherever sediments were sufficiently soft and fine grained to permit safe use of the box
corer, the box corer was preferred for sampling. After retrieval of the box core, four subcores were taken
from each box core. The subcore designated for radionuclide and mercury analyses was subsectioned at
1-cm intervals from top to bottom. The surficial 1 cm of each of these cores was analyzed for mercury.
Box cores were collected from 51 stations during the study.
The second, and less preferred, method of collection was grab sampling using a Ponar sampler. Many
sandy or stiff lake clay regions of sediment within the lake could not be box cored, so Ponar samples were
collected at these locations. When retrieved, the Ponar was carefully drained and opened. The surficial
1-cm sediment layer was removed from the grab sample. If the surficial sediment layer contained less
than 1 cm of recent sediment, then only the recent sediment was sampled. Recent sediment was visually
identifiable from older sediments by changes in cohesiveness, color, and grain size. Older sediments
were generally cohesive red-brown clays, whereas, recent sediments were brown to gray non-cohesive
silty and clayey sands. In most instances, there was at least 1 cm of recent sediment. This surficial 1-cm
layer was analyzed for mercury. Ponar samples were collected from 67 stations during the study.
2-16
-------
Mercury Study Overview
Sediment traps were deployed at eight locations (Figure 2-6). The trap at Station 3, located in northern
Lake Michigan (excluded from Fig. 2-6), was lost. Samples from the two traps at Station 6 had mercury
chloride added as a preservative to their collection bottles prior to deployment and therefore were not
analyzed. The trap at 245 m deep at Station 5 failed, and no sample was available from the trap at Station
4. Enough sample was available for mercury analysis from Stations 1, 2, 5, 7, and 8. Samples from two
depths were available from Stations 7 and 8. Details of trap sampling can be found in Eadie (1997a,
1997b). All samples and subsamples collected were placed in polyethylene bags or bottles, immediately
frozen on board the ship, and transported frozen to laboratory freezers (Edgington and Robbins 1997a).
2.4.5 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 lOO-^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
turbidity 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-^m and SOO-^m) 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. The required wet weight of material for mercury
analyses was usually obtained in one or two tows.
2.4.6 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.
Whole fish were then composited by age, location, species, and size range. Samples were homogenized
using a 40-quart vertical cutter mixer for large fish, a 12-quart vertical cutter for medium sized fish, or a
high-speed 2-quart cutter for small fish.
2.5 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 1997e). Brief summaries of the specifics of the analyses for each lake
component are provided in Sections 2.5.1 to 2.5.6. Except for the analyses of sediment samples, all of the
other media used cold vapor atomic fluorescence spectrometry (CVAFS) instrumentation and sample
preparation and analysis procedures that were similar to those described in EPA Method 1631 and Bloom
and Fitzgerald (1988). The sediment sample analyses were conducted using cold vapor atomic absorption
(CVAA) instrumentation.
2-17
-------
Results of the LMMB Study: Mercury Data Report
2.5.1 Atmospheric Components
2.5.17 Vapor Fraction
The mercury collected on gold-coated glass beads was thermally desorbed from the traps at 500 °C and
carried into a CVAFS analyzer.
2.5.12 Particulate Fraction
The glass fiber filters used to collect particulate atmospheric mercury were digested in 1.6 M nitric acid,
using a microwave digestion procedure to release the mercury from the particulate material. The mercury
in the digestate was then determined by oxidation with bromine monochloride, purge and trap, and
CVAFS.
2.5.13 Precipitation Fraction
The mercury in precipitation samples was determined by oxidation with bromine monochloride, purge
and trap, and CVAFS, without digestion.
2.5.2 Tributaries
Water samples from the tributaries were analyzed for mercury using the analytical techniques outlined in
EPA Method 1631. Briefly, the mercury in a 100-mL sample aliquot was oxidized to Hg+2 with bromine
monochloride. The sample was reduced with NH2OH-HC1 to destroy the free halogens, then reduced with
stannous chloride (SnCl2) to convert dissolved Hg+2 to volatile Hg°. The Hg° was separated from solution
by purging with an inert gas, collected onto a gold trap, and thermally desorbed from the trap into an inert
gas stream that carried the Hg° into the cell of a CVAFS analyzer for detection.
Water samples were analyzed for methylmercury using a combination of distillation, ethylation, gas
chromatography, and cold-vapor atomic fluorescence spectrometry. Briefly, methylmercury was distilled
from a water sample with heat and a flow of inert gas. The distillate was treated with sodium tetraethyl
borate, which converts the methylmercury to the more volatile methylethylmercury, which was separated
on a gas chromatographic column. The methylethylmercury was pyrolyzed and converted to Hg°, and
swept into the CVAFS analyzer for determination of mercury.
2.5.3 Open Lake Water
Water samples from the open lake were analyzed for mercury using the same techniques described above
for tributary samples.
2.5.4 Sediment
Sediment samples were freeze-dried in the laboratory in pre-weighed storage containers. The freeze-dried
samples were stored in these containers until subsamples were removed for analysis. Samples were
digested in one of two ways. Most surficial sediments were digested using a Leeman Labs, Inc.,
automated mercury system. The sediment trap samples and a few surficial sediment samples were
digested using a 1.6 M nitric acid solution and a microwave digestion system (Uscinowicz and Rossmann
1997). The Leeman automated digestion uses 50% aqua regia and potassium permanganate solutions and
provides a more vigorous digestion than the microwave procedure.
2-18
-------
Mercury Study Overview
All samples were analyzed using a Leeman Labs, Inc. automated mercury analysis system. The analysis
is based upon the cold vapor atomic absorption spectrophotometry (CVAAS) technique that reduces
divalent mercury in solution to elemental mercury vapor using stannous chloride. Argon is used to carry
the elemental mercury to the detector (Uscinowicz and Rossmann 1997).
2.5.5 Lower Pelagic Food Web Organisms
Freeze-dried plankton samples were placed in a PFA Teflon® digestion vessel with a 1:1 concentrated
sulfuric acid and nitric acid mixture, then placed in a 70 °C hot water bath overnight. Mercury was
determined by oxidation with bromine monochloride, purge and trap, and CVAFS.
2.5.6 Fish
Samples were digested in concentrated nitric acid by microwave digestion under high pressure and
temperature. Mercury analysis was performed using CVAFS.
2.6 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 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). Specific quality
control elements implemented in the sampling and analysis of mercury 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 procedural blanks;
collection and analysis of field or laboratory duplicate samples;
• analysis of standard reference materials;
preparation and analysis of a variety of quality control samples including performance standards;
• 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 September 1995, GLNPO conducted an intercomparison study involving the mercury Pis at the
Chesapeake Biological Laboratory (CBL), the University of Wisconsin Department of Natural Resources
(WDNR), and the University of Michigan Air Quality Laboratory (UMAQL). The performance of these
three laboratories could be more readily compared because they were analyzing similar sample matrices,
e.g., river water, lake water, and precipitation. The performance of the laboratories analyzing the
plankton, fish, and sediment samples could not be compared in a similar fashion, given the significant
differences in the sample preparation procedures used for each of these matrices. The study compared the
submersible pump collection technique performed by Gerald Keeler (University of Michigan) and the Go-
Flo bottle technique performed by Robert Mason (University of Maryland's Chesapeake Biological
Laboratory). Drs. Keeler and Mason collected samples from the same point aboard the R/VLake
Guardian. Dr. Hurley collected samples from an inflatable boat rowed several hundred yards from R/V
2-19
-------
Results of the LMMB Study: Mercury Data Report
Lake Guardian. Each of the Pis analyzed the samples in triplicate using the cold vapor atomic
fluorescence techniques described in Section 2.5.
The results are shown in Figure 2-8. The laboratory and sample fraction (total mercury vs. dissolved
mercury) are shown on the x-axis. The vertical bars represent the mean mercury concentration ± one
standard deviation for each laboratory/fraction combination. The Chesapeake Biological Laboratory only
provided data for total mercury. The mean total mercury concentrations from all three laboratories agree
within a factor of 1.4. The mean dissolved mercury concentrations from the two laboratories that
submitted dissolved mercury data agree within a factor of 1.8.
Figure 2-8. Results from Intercomparison Study of Three LMMB Laboratories Analyzing Mercury in
Aqueous Samples
0.7
c
Hg Concentration +/- 1 SD
o o o
GO j^ 01
| 0.2
0.1
0
i
i
i
i
i
i
CBL/Total WDNR/Total UMAQL/Total UMAQL/Dissolved WDNR/Dissolved
Lab/Fraction
In addition to the intercomparison study, 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.
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 measurement quality objectives (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 mercury are listed in Section 5 of The Lake Michigan Mass Balance Study Quality Assurance Report
(USEPA, 200Ib).
2-20
-------
Mercury Study Overview
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 mercury 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 website at:
www.epa. gov/glnpo under Result Remark, List of QC flags (lab_rmrk).
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 mercury
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 three 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
mercury study data in the Quality Implementation Section of each of the following chapters.
2-21
-------
Chapter 3
Mercury in Atmospheric Components
3.1 Results
From June 11, 1994 to October 30, 1995, atmospheric samples were collected from five shoreline
sampling station and one out-of-basin sampling station (Table 3-1 and Figure 2-2 in Chapter 2).
Atmospheric samples were collected from three separate sampling media or phases: vapor (ng/m3),
particulate (pg/m3) and precipitation (ng/L). A total of 387 vapor phase samples, 399 particulate phase
samples, and 407 precipitation phase samples were collected and analyzed for total mercury.
Table 3-1. Numbers of Atmospheric Samples Analyzed for Mercury
Sampling Station
Shoreline
Atmospheric
Sampling
Stations
Out-of-basin
Atmospheric
Sampling
Stations
Chiwaukee Prairie
George Washington
I IT Chicago
Sleeping Bear Dunes
South Haven
Bondville
Sampling Dates
7/1 9/94 to 10/30/95
7/1 9/94 to 7/25/94
6/1 1/94 to 10/30/95
6/23/94 to 10/30/95
6/1 9/94 to 10/30/95
6/24/94 to 10/30/95
Total
Number of
Vapor
Samples
Analyzed
73
1
80
801
79
74
387
Number of
Particulate
Samples
Analyzed
79
2
83
80
81
74
399
Number of
Precipitation
Samples
Analyzed
74
0
74
97
81
81
407
Samples
Analyzed
226
3
237
257
241
229
1193
10ne sample was invalid.
3.1.1 Vapor Fraction
Between 73 and 80 vapor-phase samples were collected from four shoreline atmospheric stations and one
out-of-basin station (Bondville, located in Illinois). In addition, one sample was collected at George
Washington High School. Because of the representativeness issues with using a single sample, this result
was not used in any of the analyses. The overall mean vapor-phase concentration was 2.44 ng/m3.
Table 3-2. Mean Mercury Concentrations Measured in the Vapor Phase
Sampling Station
Chiwaukee Prairie
George Washington H.S.
I IT Chicago
Sleeping Bear Dunes
South Haven
Bondville
N
73
1
80
79
79
74
Mean
(ng/m3)
2.20
2.31
3.62
2.12
2.16
2.06
Median
(ng/m3)
2.10
2.31
2.90
1.86
1.96
2.03
Range
(ng/m3)
1.16 to 5.68
NA
1.61 to 22.2
1.40 to 4.99
1.41 to 6.05
1.35 to 3.80
SD
(ng/m3)
0.740
NA
2.89
0.694
0.647
0.469
RSD
(%)
33.6
NA
80.0
32.8
29.9
22.7
Below DL
(%)
0
0
0
0
0
0
NA = Not applicable
3-1
-------
Results of the LMMB Study: Mercury Data Report
3.1.1.1 Geographical Variation
Mean vapor-phase mercury concentrations ranged from 2.06 ng/m3 at Bondville to 3.62 ng/m3 at IIT
Chicago (Table 3-2). The mean concentration at IIT Chicago was significantly greater than those of the
other stations, based on an analysis of variance (ANOVA) model with the Tukey method for pairwise
comparisons (results log-transformed prior to analysis). This was to be expected, because this station was
the only one classified as an urban sampling location. Among the remaining stations, only Chiwaukee
Prairie was located within 10 km of an urban area. The maximum concentration of 22.2 ng/m3 observed
at IIT Chicago was more than three times greater than the highest concentration observed at any of the
other stations (6.05 ng/m3 at South Haven). The differences in mercury concentrations at the five stations
are shown in Figure 3-1.
Figure 3-1. Mercury Concentrations in Atmospheric Vapor Measured at Four Lake
Michigan Shoreline Sites and One Out-of Basin Site (Bondville)
100-,
CO
<
E
o
"*i—«
03
I
O
O
CO
O
D
m
p
m
3~
M
^i
CO
x
X
o
I
o
>
o
o
oo
o
CO
D
O
C
I
1
Boxes represent the 25th (box bottom), 50th (center line), and 75th (box top) percentile 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. Xs represent results beyond 3*IQR from the box. Letters above the boxes represent
results of analysis of variance and multiple comparisons test. Boxes with the same letter were not statistically different (at alpha
= 0.05). The George Washington High School sampling site was not included in the analysis of variance due to the small number
of samples. C. Prairie = Chiwaukee Prairie, SBD=Sleeping Bear Dunes
3.1.1.2 Seasonal Variation
Beginning in July 1994, samples were collected approximately weekly at each station. Therefore, there
were multiple results from each station for each month in this interval, as well as one to two results during
June 1994 at three of the stations. A time plot of the monthly mean concentrations from each station is
presented in Figure 3-2.
3-2
-------
Mercury in Atmospheric Components
Figure 3-2. Arithmetic Monthly Means at each Station - Vapor Phase
6
§
u
£•
1
/""~\
/ \
/ \
- SLEEPING BEAR DUNES
NT-CHICAGO
CHIWAUKEE PRAIRIE
— • — BONDVILLE
SOUTH HAVEN
/ \
/ \
/ \
/ \ /\
1 \ / \
1 \
/ \
Jun-94
Sep-94
Dec-94
Mar-95
Jun-95
Sep-95
At IIT Chicago, there appears to be a difference in concentrations between the years 1994 and 1995.
With the exception of June 1994, for which only 2 samples were collected, the monthly means from 1994
are greater than any of the monthly means for 1995. Based on a two-sample t-test using Satterthwaite's
correction for differences in variability, this annual difference is significant (p<0.0001; using individual
log-transformed results). Annual differences are less noticeable for the other stations, however, the means
were significantly greater in 1994 for Bondville (p=0.0328) and Sleeping Bear Dunes (p=0.0058). These
differences may have been due to seasonality rather than annual shifts, as most samples collected in the
winter were collected in 1995.
Peaks occurred at IIT Chicago during July and August 1994, November 1994, and August 1995. Many of
the other stations also had peaks during summer months. For example, the maximum monthly means for
Sleeping Bear Dunes and Chiwaukee Prairie occurred during August 1994. At Bondville, the maximum
mean occurred during October 1994. At South Haven the maximum concentration occurred in March
1995, and in fact exceeded the mean at IIT Chicago during that month. After classifying individual
sample results according to season based on the collection date, significant differences between seasons
occurred at IIT Chicago (p=0.0014) and Chiwaukee Prairie (p=0.0228), but not the other stations, based
on a one-way ANOVA model, with results log-transformed prior to analysis. At IIT Chicago, the mean
concentration during summer was significantly greater than the means of sample concentrations collected
during spring and winter, based on the Tukey method for pairwise comparisons. At Chiwaukee Prairie,
the mean concentration of samples collected during summer was significantly greater than the mean
concentration during autumn.
3-3
-------
Results of the LMMB Study: Mercury Data Report
3.1.2 Particulate Fraction
Between 74 and 83 particulate-phase samples were collected from four shoreline atmospheric stations and
one out-of-basin station (Bondville). In addition, two samples were collected at George Washington High
School. Because of the representativeness issues with using only two samples, these results were not used
in any of the analyses. The overall mean particulate-phase concentration was 30.7 pg/m3.
Table 3-3. Mean Mercury Concentrations Measured in the Particulate Phase
Sampling Station
Chiwaukee Prairie
George Washington H.S.
I IT Chicago
Sleeping Bear Dunes
South Haven
Bondville
N
79
2
83
80
81
74
Mean
(pg/m3)
24.0
151
73.7
12.1
19.3
18.7
Median
(pg/m3)
19.9
151
50.4
10.9
18.5
17.4
Range (pg/m3)
3.03 to 108
58.6 to 244
8.25 to 494
1.05 to 41. 3
2. 10 to 69.0
4.04 to 62.5
SD (pg/m3)
18.2
131
77.2
8.28
12.2
11.0
RSD (%)
75.6
86.7
105
68.2
63.1
58.8
Below DL(%)
0
0
0
0
0
0
3.12.7 Geographical Variation
Mean particulate-phase mercury concentrations ranged from 12.1 pg/m3 at Sleeping Bear Dunes to 73.7
pg/m3 at IIT Chicago (Table 3-3). The mean concentration at IIT Chicago was greater than the maximum
concentrations at all stations other than Chiwaukee Prairie. Based on an ANOVA model with the Tukey
method for pairwise comparisons, the mean concentration at IIT Chicago was significantly greater than
those of the other stations and the mean concentration at Sleeping Bear Dunes was significantly lower
than those of the other stations (results log-transformed prior to analysis). These differences are not
unexpected, given the locations of the different stations. In addition to IIT Chicago being the only station
located in an urban area, Sleeping Bear Dunes is the only station located more than 50 km from an urban
area. The differences in mercury concentrations at the five stations are shown in Figure 3-3.
3-4
-------
Mercury in Atmospheric Components
Figure 3-3. Mercury Concentrations in Atmospheric Particles Measured at Five
Lake Michigan Shoreline Sites and One Out-of Basin Site (Bondville)
CO
<
E
c
.0
"•I— <
2
I
o
O
1000q
100=
1
D
-BONDVILLE(n=74)
1
i
"
O =
~° c
— f
2 3
m c
3- _5
.59 i
c
C
1
i
t_
= CO
H DO
•) D
5 1
3
"
)
0
•
1 1
1 1
1
OT
0
C
1
m
3~
M
Boxes represent the 25th (box bottom), 50th (center line), and 75th (box top) percentile 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 results of analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05). The George Washington High School sampling
site was not included in the analysis of variance due to the small number of samples.
C. Prairie = Chiwaukee Prairie, SBD = Sleeping Bear Dunes
3.12.2 Seasonal Variation
Beginning in July 1994, samples were collected approximately weekly at each station. Therefore, there
were multiple results from each station for each month in this interval, as well as one to two results during
June 1994 at three of the stations. A time plot of the monthly mean concentrations from each station is
presented in Figure 3-4.
Particulate sample concentrations from IIT Chicago seem to exhibit the same annual difference observed
in vapor samples, although to a lesser extent. Three of the four highest concentrations at IIT Chicago
occurred during 1994. However, it is worth noting that the June 1994 maximum was based on only two
samples and is therefore more variable than the other monthly means, which were based on at least four
samples. The difference between years was significant for IIT Chicago (p=0.0456), but not for the other
stations, based on a two-sample t-test run on the individual log-transformed results, with Satterthwaite's
correction for differences in variance.
Other than IIT Chicago, the stations did not exhibit much variability between months and there was little
evidence of any effects of seasonality. There was some consistency between these stations during May
1995, when all stations had relative minimum concentrations, and in September 1995, when all stations
had relative maximum concentrations. Mercury concentrations differed significantly between seasons
3-5
-------
Results of the LMMB Study: Mercury Data Report
only at Sleeping Bear Dunes (p=0.0311), based on a one-way ANOVA model with the Tukey method for
pairwise comparisons (results log-transformed prior to analysis). For this station, the mean concentration
of samples collected in summer was significantly greater than the mean concentration in winter.
Figure 3-4. Arithmetic Monthly Means at each Station - Participate Phase
E 150
2
SLEEPING BEAR DUNES
NT-CHICAGO
CHIWAUKEE PRAIRIE
—•—BONDVILLE
SOUTH HAVEN
Sep-94
Sep-95
3.1.3 Precipitation Fraction
Between 74 and 97 precipitation-phase samples were collected from four shoreline atmospheric stations
and one out-of-basin station (Bondville, located in Illinois). The overall mean precipitation-phase
concentration was 20.6 ng/L.
Table 3-4. Mean Mercury Concentrations by Station Measured in the Precipitation Phase
Sampling Station
Chiwaukee Prairie
I IT Chicago
Sleeping Bear Dunes
South Haven
Bondville
N
74
74
97
81
81
Mean
(ng/L)
23.1
26.1
15.2
18.1
22.1
Volume-
weighted
Mean (ng/L)
16.5
21.1
11.0
13.9
16.1
Median
(ng/L)
19.9
20.4
11.0
14.9
16.3
Range (ng/L)
4.47 to 134
5.45 to 74.6
2.09 to 63.7
3.21 to 110
5.32 to 137
SD
(ng/L)
18.3
15.5
12.0
14.8
18.3
RSD
(%)
79.1
59.5
78.9
81.9
82.5
Below
DL (%)
0
0
0
0
0
3.13.7 Geographical Variation
Mean precipitation-phase mercury concentrations ranged from 15.2 ng/L at Sleeping Bear Dunes to 26.1
ng/L at IIT Chicago (Table 3-4). In addition to the mean concentrations listed, means were also
calculated on a volume-weighted basis, which ranged from 11.0 ng/L at Sleeping Bear Dunes to 21.1
ng/L at IIT Chicago. Volume-weighting was done to minimize biases occurring due to small precipitation
events (low bias). The variability of the sample volumes collected at each station was high, with relative
standard deviations (RSDs) of approximately 100%. However, the volumes themselves did not differ
3-6
-------
Mercury in Atmospheric Components
greatly between stations, and the differences in volume-weighted means between stations were consistent
with the differences in arithmetic means. The formula for volume-weighted means is presented below:
7=1
7=1
where: c; = measured concentration in the rth sample,
v; = volume of the rth sample, and
n = number of samples.
Arithmetic means were compared using a one-way ANOVA model with the Tukey method for pairwise
comparisons. The mean concentration at Sleeping Bear Dunes was significantly lower than those at IIT
Chicago, Bondville, and Chiwaukee Prairie, and the mean concentration at South Haven was also
significantly lower than that at IIT Chicago. The difference between IIT Chicago and the other stations
for the precipitation phase is smaller than for the vapor and particulate phases. This is likely due to the
lack of an extremely high concentrations collected from this station. During a rain event, mercury is very
rapidly flushed out the atmosphere; hence, the first rain during an event has the highest mercury
concentrations. Therefore, short duration rain events have higher mercury concentrations than long
duration events because the lower mercury concentrations of rain later in an event tend to dilute the high
concentrations received early in an event. The differences in mercury concentrations at the five stations
are shown in Figure 3-5.
Figure 3-5. Mercury Concentrations in Atmospheric Precipitation Measured at
Four Lake Michigan Shoreline Sites and One Out-of-basin Site (Bondville)
1000=,
AB
AB
BC
D)
C
O
'•4-*
(0
-fc
0)
o
o
o
100=
10=
•
T
i
I
i
DO
O
1
r~
r~
m
•
T
i
1
J_
-C. PRAIRIE (n
A
T
i
i
-NT-CHICAGO
C
T
i
1
-SBD (n=97)
•
T
i
i
1
03
O
c
1
1
m
&
Boxes represent the 25th (box bottom), 50th (center line), and 75th (box top) percentile 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 results of analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05). C. Prairie = Chiwaukee Prairie, SBD = Sleeping
Bear Dunes.
3-7
-------
Results of the LMMB Study: Mercury Data Report
3.1.3.2 Seasonal Variation
Beginning in June 1994, samples were collected at least once during each month at each station except for
Chiwaukee Prairie, based on the occurrence of precipitation events. Sampling at Chiwaukee Prairie
began in July 1994. Monthly mean concentrations were calculated directly and through volume-
weighting at each station, and are presented as time plots in Figures 3-6 and 3-7, respectively.
Generally, a seasonal pattern can be seen when looking at the arithmetic means, with concentrations
greatest during the summer, and lowest during the winter. The only exception to this occurred in
December 1994 at IIT Chicago, which had a relatively high mean concentration of 31.2 ng/L. The
maximum monthly mean occurred in July 1995 for all stations except IIT Chicago, for which it occurred
in June 1995. Based on one-way AN OVA models using the Tukey method for pairwise comparisons,
there were significant differences in mean concentration between seasons at four of the five stations
(Bondville: p=0.0166, Chiwaukee Prairie: p=0.0045, IIT Chicago: p=0.0170, Sleeping Bear Dunes:
p=0.0008). At Chiwaukee Prairie, the mean concentration in summer was significantly greater than the
mean concentration in autumn, while the mean concentration in summer was greater than the mean in
winter at IIT Chicago. At Sleeping Bear Dunes, the mean concentration in summer was significantly
greater than those in both autumn and winter, and the mean concentration in spring was also greater than
the mean in autumn. No significant pairwise differences were found at Bondville. Unlike the vapor and
particulate phases, there were no significant differences between years for any of the stations.
Figure 3-6. Arithmetic Monthly Means at each Station - Precipitation Phase
B)
6
g
£-
2
| 20
\ / \ A
- SLEEPING BEAR DUNES
NT-CHICAGO
CHIWAUKEE PRAIRIE
• BONDVILLE
SOUTH HAVEN
/X
If *
Se(>94
Sep-95
The seasonal pattern is less distinct when examining the volume-weighted means. Maximum monthly
volume-weighted means occurred in different seasons for each station: in July 1994 at Chiwaukee Prairie,
in March 1995 at Bondville, in May 1995 at Sleeping Bear Dunes, in August 1995 at South Haven, and in
December 1994 at IIT Chicago. This last mean was the maximum at all stations, and contradicts the
expectations based on the seasonal patterns exhibited in Figure 3-6. This value was based on three
3-8
-------
Mercury in Atmospheric Components
samples, including one collected on December 4, 1994, with a volume of 170 mL and a concentration of
60.2 ng/L. All other precipitation samples with concentrations exceeding 60 ng/L had sample volumes
ranging from 22 to 83 mL. Therefore, this sample had a greater effect on the monthly volume-weighted
mean concentration than other high concentration, lower-volume samples. For example, a sample
collected at South Haven one week before had a concentration of 63.6 ng/L, but a volume of only 34 mL.
Figure 3-7. Volume-Weighted Monthly Means at each Station - Precipitation Phase
50 -,
40 -
30 -
9n
-
•
n
SLEEPING BEAR DUNES
NT-CHICAGO
CHIWAUKEE PRAIRIE
BONDVILLE
SOUTH HAVEN
A
/ \
x A / \
\ / \ M
A \ 1 \
A
\ /A / A
/ \ A \\ /
Jun-94
Sep-94
Dec-94
Mar-95
Jun-95
Sep-95
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 mercury
monitoring portion of the study are further described in Section 2.6 and included use of standard
operating procedures (SOPs), training of laboratory and field personnel, and establishment of method
quality objectives (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 atmospheric mercury 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.
3-9
-------
Results of the LMMB Study: Mercury Data Report
As discussed in Section 2.6, data verification was performed by comparing all field and quality control
(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 3-5 provides a summary of flags applied to the atmospheric mercury 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 2.6. One result for vapor mercury was qualified as invalid, and was not used in the analyses of
atmospheric mercury concentrations presented in this report.
Table 3-5. Summary of Routine Field Sample Flags Applied to Mercury in Atmospheric Samples
Flag
LOB, Low Biased Result
INV, Invalid Result
FFD, Failed Field Duplicate
FFT, Failed Trip Blank
FPC, Failed Lab Performance Check
MDL, Below Method Detection Limit
SDL, Below System Detection Limit
Number of QC Samples
Participate
—
—
—
43
219
NA
—
Precipitation
—
—
33
—
846
—
NA
Vapor
—
—
—
45
375
NA
—
Percentage of Samples Flagged (%)
Participate
1%(5)
0
—
1%(2)
1%(5)
NA
0
Precipitation
0
0
1%(2)
0
0
0
NA
Vapor
0
0.3% (1)
—
0.3% (1)
0
NA
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.
NA = Not Applicable
The analytical sensitivity of precipitation routine field samples was assessed through comparison to a
method detection limit (MDL) of 0.300 ng/L. For particulate and vapor field samples, analytical
sensitivity was assessed through comparison to system detection limits (SDL) equaling 1.00 pg/m3 and
0.200 ng/m3, respectively. If a sample result was below its appropriate limit, a "below MDL" or "below
SDL" flag was to be applied to that sample. However, because all sample concentrations were above the
corresponding limit, the MDL and SDL flags were not applied to any sample.
Field trip blanks were analyzed to assess the potential for contamination of routine field samples. A total
of 88 trip blanks were analyzed, 45 in the vapor phase, and 43 in the particulate phase. In accordance
with the researcher's data qualifying rules, samples were flagged for trip blank contamination (FTB) if the
associated blank concentration exceeded the SDL expressed as a mass (43.45 pg for particulate samples
and 0.084 ng for vapor samples). In the particulate phase, two samples were flagged for trip blank
contamination, based on associated blank masses 68.2 pg and 79.1 pg. The flagged particulate routine
field sample results, when expressed as masses, were approximately two and ten times the associated
blank masses. One additional sample in the vapor phase was flagged for blank contamination due to an
associated blank mass of 0.205 ng. The flagged vapor sample had a mass approximately 5 times greater
than the associated blank mass.
A total of 33 field duplicate samples were collected and analyzed to assess precision for the precipitation
phase. Field duplicates were collected at three of the five stations from which precipitation samples were
collected. In accordance with the researcher's data qualifying rules for field and laboratory duplicates,
samples were flagged for a failed duplicate (FFD) if the relative percent difference (RPD) between results
for a sample and its duplicate was greater than 25%. Two field duplicate pairs failed to meet this criteria,
3-10
-------
Mercury in Atmospheric Components
with RPDs of 25.7% and 62.5%. No field duplicate samples were collected for the participate or vapor
phases; therefore, the FFD flag was not applied to any samples from these phases.
Laboratory performance check samples were used to monitor analytical bias. Performance check samples
were run after every 6 samples, resulting in 1,440 total check samples. In accordance with the
researcher's data qualifying rules for performance checks, field samples were flagged for a failed
performance check (FPC) if the absolute percent difference for the associated performance check was
greater than 20%. The FPC flag was applied to five particulate field samples, due to performance check
percent differences of-28.8% and -29.1% (corresponding to percent recoveries of 71.2% and 70.9%,
respectively). These five samples were also qualified as being low biased by the QC Coordinator due to
the performance check recoveries. No other samples were qualified as being low biased or high biased
based on analyses of performance checks, blank contamination, or other internal QC data.
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 3-6 provides a summary of data quality
assessments for several of these attributes for atmospheric data.
Table 3-6. Data Quality Assessment for Mercury in Atmospheric Samples
Parameter
Number of Routine Samples Analyzed
System Precision, Mean Field Duplicate RPD (%), >SDL
Analytical Bias, Mean LPC RPD%
Analytical Sensitivity, Samples reported as
-------
Results of the LMMB Study: Mercury Data Report
phases ranged from -2.20% for particulate to 0.823% for precipitation. When expressed as percent
recoveries, these means correspond to 97.8% and 101%, respectively.
Analytical sensitivity was evaluated by calculating the percentage of samples reported below the SDL for
precipitation data and the percentage of samples reported below the MDL for the particulate and vapor
data. This percentage was 0% for all three phases.
3.3 Data Interpretation
3.3.1 Atmospheric Sources
Based on the results of this study, vapor, particulate and precipitation phases were all important sources
of mercury to Lake Michigan. All results from all three phases were above the associated method or
system detection limit. The mean vapor and particulate mercury concentrations of 2.44 ng/m3 and 30.7
pg/m3 (0.0307 ng/m3) were approximately 12 and 30 times greater than their associated SDLs. The mean
precipitation-phase mercury concentration of 20.6 ng/L was approximately 70 times greater than the
associated MDL.
3.3.2 Seasonal Considerations
Generally, the effect of season on mercury concentration depended on the phase and the station from
which the samples were collected. For vapor-phase mercury, significant differences between seasons
were observed only at IIT Chicago and Chiwaukee Prairie, with peak concentrations during the summer at
both stations. Both of these stations had greater levels in the summer of 1994 compared to 1995. For
particulate-phase mercury, significant seasonal differences were observed only at Sleeping Bear Dunes,
with peak concentrations occurring during the summer.
Seasonal patterns were most apparent in precipitation-phase mercury. Significant differences between
seasons occurred at four of the five stations. For each of these stations, the peak concentrations occurred
in summer and the lowest concentrations occurred either during autumn or winter. However, these
seasonal differences may have been partly due to the occurrence of smaller precipitation events during the
summer, compared to other seasons, which would result in smaller sample volumes, and hence, higher
mercury concentrations, during the initial wash out of mercury from the atmosphere.
When the data were examined using volume-weighted means, seasonal patterns became much less
distinct. However, for all stations other than Chicago IIT, the lowest volume-weighted means did occur
during the winter. This may have been due to differences in precipitation type, as the relationship
between mercury and precipitation may differ between warm-cloud processes and cold-cloud processes
(Landis et al., 2002). In a study of precipitation in mercury in the Lake Superior region, Glass et al.
(1986) found significantly greater mercury concentrations in rainfall than in snow. The seasonal pattern
was also similar to that observed at three sites in Wisconsin as part of the National Atmospheric
Deposition Program's (NADP) Mercury Deposition Network (WDNR, 1999). Volume-weighted mean
concentrations in that study were highest in the spring or summer for each site for all three years, other
than for one site in 1995, where the mean concentration was highest in the winter.
Significant differences between seasons were observed at only one LMMB station for particulate-phase
mercury. At the Sleeping Bear Dunes site, the mean concentration during summer was significantly
greater than the mean concentration during winter. This result is not consistent with results from past
studies. Particulate-phase mercury concentrations have previously been observed to be greater during the
winter compared to the summer in Maryland (Mason et al., 1997) and near Lake Michigan (Keeler et al.,
3-12
-------
Mercury in Atmospheric Components
1995). Concentrations at Sleeping Bear Dunes were similar during the two summers for which data were
collected.
3.3.3 Regional Considerations
For particulate and vapor-phase mercury, the mean concentration at IIT Chicago was significantly greater
than those at the other stations. For precipitation-phase mercury, the mean concentration was also
greatest at IIT Chicago, and was significantly higher than at two of the other stations. This was not
unexpected, as IIT Chicago was the only one of the five stations that could be classified as being located
in an urban area. It has been observed in the past that the Chicago area has significantly increased
mercury levels in dry deposition (Keeler, 1994) and precipitation around local urban/industrial areas
(Hoyer et al., 1995). The difference between IIT Chicago and the other stations was greater for
particulate-phase mercury than for the other phases. This may be due to the greater prevalence of the
mercuric form of mercury (Hg2+) in the particulate phase compared to the vapor phase. Mercuric mercury
is more soluble in water, and therefore more likely to be due to local sources (Lindberg and Stratton,
1998). Mason et al. (1997) found low levels of ionic mercury in precipitation, and hypothesized that this
was due to in-cloud oxidation processes being a significant source of mercury in precipitation, rather than
just the scavenging of particles or of gaseous ionic mercury.
The mean and median vapor-phase concentrations at IIT Chicago (mean: 3.62 ng/m3, median: 2.90 ng/m3)
were very close to those collected in Egbert, Ontario in 1990 (mean: 3.71 ng/m3, median: 2.90 ng/m3) by
Schroeder and Markes (1994). The station at IIT Chicago represents a major urban/industrial area and the
station in the Ontario study was located near Toronto, another major urban/industrial area. Thus, the
results from both studies may represent the influences of urban and industrial sources of mercury.
However, the samples from the Ontario study were all collected in the months of March and April, and
therefore cannot be interpreted as an annual estimate. The 49 mercury samples collected at IIT Chicago
in March and April 1995 had a mean of 2.26 ng/m3 and a median of 2.14 ng/m3, substantially lower than
the overall values. In addition to collecting samples in Egbert, Ontario, Schroeder and Markes (1994)
also measured mercury at Pt. Petre, Ontario. This site had lower mercury concentrations, with a mean of
2.21 ng/m3, comparable to the other stations in the LMMB data set. The Pt. Petre samples were collected
in the autumn only, however, and the LMMB stations had slightly lower results during these months.
While the difference in mean precipitation-phase mercury concentrations at IIT Chicago and the other
stations was not as large compared to the other phases in the study, the mean concentration at IIT Chicago
was still higher than for many sites in other studies. For example, samples of mercury in precipitation
have recently been collected as part of the National Atmospheric Deposition Program's Mercury
Deposition Network (MDN). The volume-weighted mean calculated from the MDN transition phase in
1995 was 10.25 ng/L, lower than the mean at all five LMMB stations (MDN, 1999). In addition, in an
assessment using data collected as part of the NADP, volume-weighted mean concentrations were
calculated for samples collected from seven sites in Wisconsin from 1995 to 1997 (WDNR, 1999). The
state-wide volume-weighted means for the three years ranged from 11.48 ng/L in 1997 to 15.75 ng/L in
1995. These means are similar to the volume-weighted mean concentrations from Chiwaukee Prairie
(16.5 ng/L), Bondville (16.1 ng/L), and South Haven (13.9 ng/L), but below the volume-weighted mean
of 21.1 ng/L from IIT Chicago. However, the maximum annual volume-weighted mean of 25.60 ng/L
from the seven Wisconsin sites, occurring at the rural Wildcat Mountain State site in western Wisconsin
in 1996, exceeded the volume-weighted mean at IIT Chicago. This mean was based on the results from
one of two sampling columns at that site, with the other column yielding in a mean of 13.81 ng/L. It is
worth noting that the mean concentration from this second column was greater than that of Sleeping Bear
Dunes (11.0 ng/L), the only atmospheric site from the LMMB located in a similarly rural area.
3-13
-------
Results of the LMMB Study: Mercury Data Report
Other recent studies have also shown spatial differences in mercury concentration in precipitation. Mason
et al. (2000) found higher levels of mercury flux at a site in Baltimore, compared to three other rural sites
in Maryland. Glass et al. (1986) measured mercury concentrations in snow pack collected from three
areas in Minnesota, one in Wisconsin, one in Upper Peninsula of Michigan, and one in Ontario within
watersheds that drain into Lake Superior. Samples of snow pack were collected at 10 to 17 specific
locations in each of these geographic areas. Measurements of mercury in snow from five of the six areas
were below those of IIT Chicago in this study. The means from these five areas ranged from 12 ng/L to
15 ng/L, with standard deviations ranging from 1 to 5 ng/L. The sixth sampling area was centered around
Grand Rapids, Minnesota, and had a mean concentration of 100 ng/L and a standard deviation of 173
ng/L. The mean concentration is substantially higher than the mean at IIT Chicago in this study, and may
be the result of contamination of samples from that area, or may represent a localized source of mercury.
In addition, the results may not be comparable to all of the LMMB data, because the samples were
collected in snow, rather than rain.
3-14
-------
Chapter 4
Mercury in Tributaries
4.1 Results
From March 29, 1994 to October 31, 1995, samples were collected from 11 tributaries that flow into Lake
Michigan (Figure 2-3 in Chapter 2). Samples were collected as described in Section 2.4.2 and analyzed
for total and dissolved mercury by cold-vapor atomic fluorescence spectrometry (see Section 2.5.2). A
total of 346 samples were collected and analyzed for dissolved mercury, and 353 samples were collected
and analyzed for total mercury (Table 4-1). In addition to the analysis of total and dissolved mercury, a
subset of samples was analyzed for methylmercury using a combination of distillation, ethylation, gas
chromatography, and cold-vapor atomic fluorescence spectrometry. A total of 203 samples were
analyzed for total methylmercury, and 204 samples were analyzed for dissolved methylmercury.
Table 4-1. Number of Tributary Samples Analyzed for Mercury and Methylmercury
Analyte
Mercury
Methylmercury
Tributary
Fox
Grand Calumet
Grand
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 10/1 7/95
04/05/94 to 10/1 8/95
04/06/94 to 09/1 9/95
04/06/94 to 10/27/95
Total
Fox
Grand Calumet
Grand
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
01/1 1/95 to 08/30/95
02/1 3/95 to 10/1 8/95
04/28/94 to 10/31/95
01/26/95 to 10/30/95
04/1 1/94 to 10/26/95
01/1 7/95 to 10/1 1/95
01/1 0/95 to 10/06/95
01/24/95 to 10/1 7/95
04/05/94 to 10/1 8/95
04/1 4/94 to 10/24/95
01/27/95 to 10/27/95
Total
Number of Samples Analyzed
Dissolved Fraction
38
15
46
38
27
23
36
27
28
35
33
346
17
7
31
16
20
12
21
11
22
32
15
204
Total Fraction
39
15
47
38
27
25
38
27
28
36
33
353
15
8
33
14
21
12
21
11
20
32
16
203
Total Number
of Samples
Analyzed
77
30
93
76
54
48
74
54
56
71
66
699
32
15
64
30
41
24
42
22
42
64
31
407
4-1
-------
Results of the LMMB Study: Mercury Data Report
4.1.1 Geographical Variation
4.1.1.1 Mercury
Total mercury concentrations measured in Lake Michigan tributaries ranged from 0.536 to 191 ng/L. In
the 11 tributaries monitored in the LMMB Study, mean total mercury concentrations ranged from 1.07
ng/L in the Muskegon River to 28.9 ng/L in the Fox River (Table 4-2). Analysis of variance (and
Tukey's pairwise comparison test) revealed that total mercury concentrations in the Fox River were
significantly higher than in any other Lake Michigan tributary (Figure 4-1). The mean total mercury
concentration in the Fox River was 2.7 to 27 times higher than in other Lake Michigan tributaries. The
Fox River watershed has long been highly industrialized and Hurley et al. (1998a) have suggested that the
main source of Fox River mercury loads is resuspension of contaminated sediments. Following the Fox
River, total mercury concentrations were highest in the Kalamazoo and Grand Calumet Rivers. Total
mercury concentrations in these tributaries were significantly higher (at the 95% confidence level) than in
any other tributary, except for the Fox River. These rivers are located to the south and southeast of Lake
Michigan (Figure 4-2), where urban and industrial land uses are predominant. The lowest total mercury
concentrations were observed in the Muskegon, Pere Marquette, Manistique, and Menominee Rivers
(Figure 4-2), which are the more northern tributaries that are primarily forested. Total mercury
concentrations in the Muskegon River were significantly lower than any other Lake Michigan tributary
(Figure 4-1). Hurley et al. (1998b) explained that the low mercury concentrations in this tributary may be
due to Lake Muskegon, which is located directly upstream of the sampling site and acts as a temporary
sink for contaminants.
Dissolved mercury concentrations were more consistent among tributaries than total mercury
concentrations. Mean dissolved mercury concentrations only ranged from 0.666 ng/L in the Grand
Calumet River to 3.71 in the Fox River. The remaining tributaries all contained mean dissolved mercury
levels between 1 and 2 ng/L (Table 4-2). Fewer significant differences in dissolved mercury
concentrations also were seen among tributaries (Figure 4-1 and Figure 4-2). Unlike total mercury
concentrations, dissolved mercury concentrations in the Fox River were not significantly higher than in
all other tributaries. Dissolved mercury concentrations in the Fox River were only significantly higher
than in three other tributaries (Grand Calumet, Muskegon, and Milwaukee Rivers). Following the Fox
River, mean dissolved mercury concentrations were highest in the Manistique and Menominee Rivers,
two tributaries that had among the lowest concentrations of total mercury. Dissolved mercury
concentrations in the Manistique River were significantly higher than in three other tributaries, and
dissolved mercury concentrations in the Menominee River was significantly higher than in two other
tributaries. The lowest mean dissolved mercury concentration was in the Grand Calumet River, which
was among the highest in total mercury concentrations. The mean dissolved mercury concentration at this
site was significantly lower than in seven other tributaries.
4-2
-------
Mercury in Tributaries
Table 4-2. Mean Mercury Concentrations Measured in Lake Michigan Tributaries
Fraction
Dissolved
Participate3
Total
Tributary
Fox
Grand Calumet
Grand
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Fox
Grand Calumet
Grand
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Fox
Grand Calumet
Grand
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
N
37
15
44
37
25
22
34
26
26
34
31
37
15
43
37
25
22
34
26
26
33
31
38
15
45
37
25
24
36
26
26
34
32
Mean
(ng/L)
3.71
0.666
1.68
1.62
1.99
1.87
1.15
1.08
1.79
1.64
1.46
25.8
9.26
4.29
9.00
1.08
1.92
2.93
-0.0058
1.09
3.02
4.04
28.9
9.93
6.02
10.6
3.07
3.63
4.08
1.07
2.88
4.52
5.40
Median
(ng/L)
1.44
0.628
1.39
1.22
2.06
1.71
0.963
0.730
1.12
1.59
0.912
22.1
8.00
3.23
8.81
0.447
1.75
2.45
0.215
0.758
3.12
4.18
23.5
8.63
4.87
10.3
2.71
3.33
3.62
0.984
2.46
4.72
5.29
Range
(ng/L)
0.786 to 40.8
0.261 to 1.37
0.400 to 8.29
0.202 to 7. 12
0.680 to 3.61
0.739 to 3.61
0.439 to 2.42
0.259 to 6.20
0.254 to 6.86
0.437 to 4.68
0.399 to 6.21
-11. 3 to 153
4.68 to 18.2
-3.54 to 46.6
0.786 to 23.7
-0.0865 to 13.3
-0.339 to 4.81
-0.320 to 18.6
-4.96 to 0.742
-5.40 to 7.67
-0.0094 to 7.42
-1.73 to 9.24
1.84 to 191
5.81 to 18.5
1.16 to 47.5
2.62 to 25.7
1.02 to 15.8
1.61 to 6.57
1.23 to 20.3
0.536 to 1.82
0.557 to 11. 5
0.712 to 9.25
1.38 to 14.5
SD
(ng/L)
8.75
0.341
1.32
1.41
0.815
0.861
0.594
1.13
1.56
0.928
1.42
26.2
4.34
7.16
5.56
2.61
1.57
3.06
1.08
2.49
1.59
2.33
30.5
4.29
6.91
5.77
2.89
1.57
3.19
0.354
2.59
2.00
2.70
RSD
(%)
236
51.2
78.9
87.3
40.9
46.1
51.7
105
87.0
56.5
97.2
101
46.9
167
61.8
242
81.7
104
—
228
52.9
57.6
106
43.2
115
54.3
94.2
43.3
78.1
33.1
90.1
44.1
50.1
Below DL
(%)
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
—
—
—
—
—
—
—
—
—
—
—
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
a Mercury concentrations in the particulate fraction were not directly measured. Particulate concentrations for each sample were calculated as
the difference between the measured total and dissolved concentrations. If measured dissolved concentrations were greater than measured
total concentrations, the calculated concentration in the particulate fraction was a negative number. Because particulate concentrations were
calculated from two measured values, these reported concentrations will contain more variability than measured values reported for dissolved
and total fractions. Also, the percent of samples below the detection limit could not be determined for the particulate fraction, because this
fraction was not directly measured and detection limits for this fraction were not developed.
4-3
-------
Results of the LMMB Study: Mercury Data Report
Figure 4-1. Total and Dissolved Mercury Concentrations in Lake Michigan Tributaries
o
c
o
O
LAMAZOO
1
^
-z.
CO
D
c
m
i
m
.NOMINEE I
1
' —
.WAUKEE (
i
a
ISKEGON (i
i
TJ
VIARQUETI
CO
I
EBOYGAN
CO
—i
JOSEPH (i
M
Ol
3
II
M
GO
CD
M
CD
m
M
CD
GO
M
ABC
ABC
AB
T
CD ABC ABCD
ABC
BCD
_L
1
p
o
>
l~
c
s
m
CD
70
>
CO
m
~z.
O
N
O
O
3~
M
GO
=! 5 c
D
C
m
m
m
m
m
3
M
GO
c
CO
7^
m
CD
O
M
CD
70
O
C
m
m
co
1
m
DO
O
3
II
GO
CO
c_
o
CO
m
TJ
3~
M
GO
Boxes represent the 25th (box bottom), 50th (center line), and 75th (box top) percentile 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.
Xs represent results beyond 3*IQR from the box. Letters above the boxes represent results of analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05).
4-4
-------
Mercury in Tributaries
Figure 4-2. Mean Total and Dissolved Mercury Concentrations Measured in Lake
Michigan Tributaries
Pere Marquette
Muskegon
Sheboygan ]
4.112 Methylmercury
The geographical pattern of methylmercury concentrations in Lake Michigan tributaries was very
different from that of total mercury. While total mercury concentrations were much higher in the Fox
River than in other tributaries, methylmercury concentrations in four other tributaries were higher than in
the Fox River (Table 4-3). Mean total methylmercury concentrations in Lake Michigan tributaries ranged
from 0.0424 ng/L in the Grand Calumet to 0.260 ng/L in the Sheboygan River (Table 4-3). Total
methylmercury concentrations in the Sheboygan River were significantly higher than in the St. Joseph,
Muskegon, Grand, and Grand Calumet Rivers (Figure 4-3). Total methylmercury concentrations in the
Grand Calumet were significantly lower than in the Sheboygan, Kalamazoo, and Menominee Rivers. No
other significant differences in total methylmercury were observed among Lake Michigan tributaries.
4-5
-------
Results of the LMMB Study: Mercury Data Report
Table 4-3. Mean Methylmercury Concentrations Measured in Lake Michigan Tributaries
Fraction
Dissolved
Participate3
Total
Tributary
Fox
Grand Calumet
Grand
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Fox
Grand Calumet
Grand
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Fox
Grand Calumet
Grand
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
N
17
7
31
16
20
12
21
11
22
32
15
15
7
29
14
19
12
20
11
20
29
15
15
8
33
14
21
12
21
11
20
32
16
Mean
(ng/L)
0.0419
0.0133
0.0479
0.0704
0.114
0.182
0.115
0.0363
0.0850
0.106
0.0915
0.118
0.0309
0.0492
0.0809
0.0073
0.0351
0.0552
0.148
0.0312
0.139
0.0081
0.162
0.0424
0.104
0.153
0.123
0.217
0.170
0.184
0.116
0.260
0.103
Median
(ng/L)
0.0420
0.0220
0.0240
0.0620
0.106
0.117
0.0774
0.0386
0.0733
0.0860
0.0393
0.134
0.0274
0.0500
0.0806
0.0080
0.0712
0.0370
0.0254
0.0270
0.0840
0.0474
0.170
0.0428
0.0993
0.147
0.128
0.196
0.117
0.0537
0.110
0.182
0.0846
Range
(ng/L)
0.00100 to 0.103
-0.0281 to 0.0527
-0.0212 to 0.404
-0.0137 to 0.240
0.0180 to 0.304
-0.00154 to 0.692
0.00977 to 0.487
0.0111 to 0.0508
-0.00700 to 0.428
-0.00868 to 0.371
0.000980 to 0.645
-0.0300 to 0.398
-0.0162 to 0.1 12
-0.225 to 0.1 72
-0.1 64 to 0.344
-0.247 to 0.203
-0.492 to 0.268
-0.281 to 0.568
-0.0023 to 1.08
-0.283 to 0.1 22
-0.226 to 0.767
-0.579 to 0.236
0.0150 to 0.413
-0.00804 to 0.0883
-0.00600 to 0.232
0.0647 to 0.33
0.0210 to 0.340
0.0971 to 0.331
0.0220 to 0.651
0.00881 to 1.1 3
-0.00796 to 0.202
0.038 to 0.822
0.0252 to 0.286
SD
(ng/L)
0.0254
0.0300
0.0779
0.0649
0.0624
0.198
0.126
0.0128
0.0839
0.0848
0.161
0.115
0.0494
0.0762
0.115
0.0879
0.214
0.209
0.319
0.0834
0.193
0.184
0.106
0.0297
0.0593
0.0773
0.0699
0.0762
0.170
0.323
0.0514
0.206
0.0639
RSD
(%)
60.7
226
163
92.3
54.6
109
110
35.2
98.6
79.7
175
97.8
160
155
142
1210
609
379
216
268
139
2280
65.3
70.0
57.0
50.6
56.7
35.1
100
176
44.4
79.1
61.8
Below DL
(%)
23.5
42.9
41.9
18.8
5.00
8.33
4.76
9.09
9.09
3.13
6.67
—
—
—
—
—
—
—
—
—
—
—
6.67
12.5
6.06
0.00
0.00
0.00
0.00
9.09
5.00
0.00
0.00
' Mercury concentrations in the particulate fraction were not directly measured. Particulate concentrations for each sample were calculated as
the difference between the measured total and dissolved concentrations. If measured dissolved concentrations were greater than measured
total concentrations, the calculated concentration in the particulate fraction was a negative number. Because particulate concentrations were
calculated from two measured values, these reported concentrations will contain more variability than measured values reported for dissolved
and total fractions. Also, the percent of samples below the detection limit could not be determined for the particulate fraction, because this
fraction was not directly measured and detection limits for this fraction were not developed.
4-6
-------
Mercury in Tributaries
Figure 4-3. Tote
Tributaries
? 1.2CH
co :
~ 1.00=
o -
0 0.80J
3 0.60J
cu :
1 0.40;
§ 0.20J
£2 0.00=
o -
\\ and Dissolved Methylmercury Concentrations in Lake Michigan
A
A
ABC
ABC
ABC AD T
-r AB AB BC
B<~ -r * "T -i- A.BC r^~l
-- I I f— 1 _L |— j [ | i— i '-r-' QH
^
""OO^SSSS^COCO
! 1 | | 1 | 1 1 | | |
II
N)
fel
s
o
C
o
o
Z5
O
cu
E
t~
4— '
CU
"^
^
cu
>
o
en
Q
0.80-1
-
0.60-
-
-
0.40;
-
-
0.20-
-
0.00=
-
.n 9n~
o
A ABCD
*
ABC
ABCD
n •*•
T * AB
AB * •
ABCD • ' •
| _
CD "
BCD T r^ TT BCD -,- .-L.
—I- ~T" _1 . [ | | | M 1 , 1 1 ^
^^^^T 1 ^^TT^
"^OOXSSSS^COC^
O' ^o"^"t»m — c ' 7T *^
! | | | | | I | 1 | i
Boxes represent the 25th (box bottom), 50th (center line), and 75th (box top) percentile 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.
Xs represent results beyond 3*IQR from the box. Letters above the boxes represent results of analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05).
The geographical pattern of dissolved methylmercury concentrations in Lake Michigan tributaries also
were different from that of dissolved mercury. Mean dissolved methylmercury concentrations ranged
from 0.0133 ng/L in the Grand Calumet to 0.182 ng/L in the Menominee River. Dissolved
4-7
-------
Results of the LMMB Study: Mercury Data Report
methylmercury concentrations in the Menominee were significantly higher than in the Muskegon, Fox,
Grand, and Grand Calumet Rivers. Dissolved methylmercury concentrations in the Sheboygan and
Manistique Rivers were significantly higher than in the Fox and Grand Rivers, and dissolved
methylmercury concentrations in the Milwaukee River were significantly higher than in the Grand River.
While the more northern and forested watersheds had lower total mercury concentrations, these tributaries
did not have corresponding lower concentrations of methylmercury (Figure 4-4). Methylmercury
concentrations in the Manistique, Menominee, Pere Marquette, and Muskegon Rivers were not
significantly lower than in any other sites, with the exception of the Muskegon River being significantly
lower than the Sheboygan River in total methylmercury. Similarly, those industrialized sites that had the
highest total mercury levels (Fox, Kalamazoo, and Grand Calumet Rivers), did not have corresponding
high methylmercury concentrations. Total methylmercury concentrations in these tributaries were not
significantly higher than in any other site.
Figure 4-4. Mean Total and Dissolved Methylmercury Concentrations Measured in
Lake Michigan Tributaries
4-8
-------
Mercury in Tributaries
4.1.2 Seasonal Variation
Tributary samples were collected for mercury analysis throughout seven consecutive seasons (Spring
1994 through Autumn 1995). Analysis of variance (with Tukey's pairwise comparison test) revealed that
total mercury concentrations differed significantly among season in six of the eleven tributaries (Figure 4-
5). In the Fox River, winter total mercury concentrations were significantly lower than in any other
season. In the Kalamazoo River, winter and autumn concentrations of total mercury were significantly
lower than spring or summer concentrations. In the Manistique River, spring concentrations of total
mercury were significantly higher than in other seasons. In the Muskegon River, spring total mercury
concentrations were significantly higher than summer concentrations. In the Pere Marquette and
Sheboygan Rivers, spring total mercury concentrations were significantly higher than concentrations
during autumn.
While seasonal patterns varied among tributaries, total mercury concentrations were generally higher in
the spring and lower in the winter. Spring concentrations of total mercury were higher than winter values
in ten of the eleven tributaries, and these differences were statistically significant in three of the
tributaries. In all six tributaries that showed significant seasonal differences, total mercury concentrations
were significantly higher in the spring than in other seasons.
Methylmercury concentrations differed significantly among seasons in four tributaries (Figure 4-5). In
the Fox and Manistique Rivers, total methylmercury concentrations during the winter were significantly
lower than in all other seasons. In the Pere Marquette and Sheboygan Rivers, total methylmercury
concentrations during the winter were significantly lower than in the spring. Similar to total mercury
concentrations, total methylmercury concentrations were generally higher in the spring and lower in the
winter. Spring concentrations of total methylmercury were higher than winter values in eight of eleven
tributaries and these differences were statistically significant in four of these tributaries.
In most of the tributaries with significant seasonal differences in total mercury and methylmercury
concentrations, difference were tied to the seasonal flow regimes of the tributaries. The flow regimes of
many of these tributaries were dominated by high spring flows, which coincided with higher mercury
concentrations. Low mercury concentrations in the winter also coincided with lower tributary flows.
Figure 4-6 demonstrates this effect in the Manistique, Sheboygan, and Fox Rivers. Ice cover in the winter
in many of these tributaries may also lead to reduced mixing and resuspension of contaminated sediments,
which would result in lower total mercury concentrations during the winter. The hydrograph for the Fox
River also demonstrates that high mercury concentrations are often associated with peak flow events
throughout the year. Many of the highest total mercury concentrations measured in the Fox River
coincided with high storm event flows. Indeed, tributary mercury concentrations were correlated with
flow in many of the tributaries (see Section 4.1.3).
4-9
-------
Results of the LMMB Study: Mercury Data Report
Figure 4-5. Seasonal Variation of Mercury Concentrations in Lake Michigan Tributaries
0)
c
o
is 1 n
c iu
0
O
3
& 1
fl) '
§
"J5
0
n 1
A
A
,-
A
B
-
-
A
A
B
-
A
BB
B
i-i
-
-
i-i
-i
D Spring
• Summer
DAutumn
QWinter
A AB
i— i
B AB
-
A
"KB
A
B
A
r-iAB
L_B.
A
B
t
-i
("^ s^? tw1 Cr
&
-------
Mercury in Tributaries
Figure 4-6. Seasonal Flow Patterns and Total Mercury Concentrations in Selected Lake
Michigan Tributaries
8000
100
Jan-95 May-95 Aug-95
r-94
1
Jul-94 Oct-94 Jan-95 May-95 Aug-95 Nov-95
4-11
-------
Results of the LMMB Study: Mercury Data Report
4.1.3 Other Factors Affecting Tributary Mercury Concentrations
As previously mentioned (see Section 4.1.2), peaks in mercury concentrations in some tributaries
coincided with either spring high flow conditions or high flows related to storm events. Significant
positive correlations existed between flow and total mercury concentrations (both log transformed) in six
tributaries (the Fox, Grand, Sheboygan, Milwaukee, Menominee, and Manistique Rivers). In these six
tributaries, r2 values indicated that flow accounted for 17 to 65% of the variability in total mercury
concentrations (Table 4-4). For methylmercury, only two tributaries (the Fox and Menominee Rivers)
exhibited significant positive correlations with flow.
Table 4-4. Correlation of Tributary Mercury Levels with Tributary Flow
Fraction
Total Mercury
Total
Methylmercury
Tributary
Fox
Grand
Grand Calumet
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Fox
Grand
Grand Calumet
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
N
38
45
13
37
25
24
36
26
26
34
32
15
31
7
14
21
12
21
11
19
32
16
Correlation Coefficient
0.417
0.431
0.311
-0.0729
0.806
0.662
0.656
0.258
0.271
0.595
0.136
0.579
-0.00648
-0.0705
-0.313
0.344
0.589
0.357
-0.273
-0.227
0.310
0.285
i2
0.174
0.185
0.0965
0.00532
0.649
0.438
0.430
0.0666
0.0732
0.354
0.0185
0.335
0.0000410
0.00497
0.0980
0.118
0.347
0.128
0.0743
0.0517
0.0963
0.0811
p-value
0.0091
0.0032
0.302
0.668
<0.0001
0.0004
<0.0001
0.203
0.181
0.0002
0.458
0.0238
0.972
0.881
0.276
0.127
0.0440
0.112
0.418
0.349
0.0838
0.285
Because most of the mercury in the water column is bound to dissolved or suspended organic matter
(USEPA, 1997c), mercury concentrations are expected to correlate with measures of solids and organic
carbon. In coordination with tributary sampling of mercury, samples also were analyzed for dissolved
organic carbon (DOC), particulate organic carbon (POC), and total solids (TS). Four of the eleven
tributaries showed significant positive correlations between total mercury and DOC concentrations (Table
4-5). Seven tributaries showed significant positive correlations between total mercury and POC
concentrations. In these seven tributaries, POC accounted for 23 to 62% of the variability in total
mercury concentrations. The strongest correlations, however, were between TS and total mercury
concentrations. All but the Muskegon River exhibited significant positive correlations between TS and
total mercury. Total solids accounted for up to 82% of the variability in total mercury concentrations. It
is possible that the POC and DOC correlations were auto-correlations, due to the attachment of not only
mercury, but also POC and DOC, to the total solids.
4-12
-------
Mercury in Tributaries
Table 4-5. Correlations of Total Mercury
(DOC), Participate Organic Matter (POC),
Levels in Lake Michigan Tributaries with Dissolved Organic Matter
and Total Solids (TS)
Analyte
DOC
POC
TS
Tributary
Fox
Grand
Grand Calumet
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Fox
Grand
Grand Calumet
Kalamazoo
Manistique
Menoninee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Fox
Grand
Grand Calumet
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
N
38
42
15
34
24
22
34
26
26
33
31
37
42
13
33
25
23
34
25
26
29
30
38
45
14
36
25
24
35
25
26
32
31
Correlation Coefficient
-0.221
0.341
0.463
0.221
0.531
0.281
0.511
-0.192
0.259
0.676
-0.116
0.625
0.220
0.776
0.0805
0.347
0.500
0.638
-0.0868
0.651
0.790
0.476
0.786
0.606
0.855
0.817
0.663
0.832
0.881
-0.215
0.823
0.908
0.775
i2
0.0488
0.116
0.215
0.0488
0.282
0.0791
0.261
0.0368
0.0670
0.457
0.0134
0.391
0.0483
0.602
0.00648
0.120
0.250
0.407
0.00753
0.424
0.624
0.227
0.618
0.367
0.731
0.668
0.439
0.693
0.777
0.0464
0.677
0.824
0.600
p-value
0.182
0.0273
0.0820
0.209
0.0076
0.205
0.0020
0.348
0.202
<0.0001
0.536
<0.0001
0.162
0.0018
0.656
0.0896
0.0151
<0.0001
0.680
0.0003
<0.0001
0.0078
<0.0001
<0.0001
<0.0001
<0.0001
0.0003
<0.0001
<0.0001
0.301
<0.0001
<0.0001
<0.0001
4.1.4 Mercury Forms
Total and dissolved fractions of mercury were directly measured in the LMMB Study, and mercury in the
particulate fraction was calculated by subtraction. Tributaries varied greatly in the contribution of
mercury from the dissolved and particulate fractions. Tributaries ranged from the Muskegon River, with
the impact of Lake Muskegon, where virtually all of the total mercury (99%) was attributable to the
4-13
-------
Results of the LMMB Study: Mercury Data Report
dissolved fraction, to the Grand Calumet River, where virtually all of the total mercury (92%) was
attributable to the particulate fraction (Table 4-6). There was a distinct separation of tributaries that were
dominated by the dissolved mercury fraction and tributaries that were dominated by the particulate
mercury fraction. The Menominee, Manistique, Pere Marquette, and Muskegon Rivers were dominated
by the dissolved mercury fraction. Each of these tributaries contained greater than 50% of total mercury
in the dissolved fraction, and the Manistique, Pere Marquette, and Muskegon Rivers contained greater
than 75% of total mercury in the dissolved fraction. These tributaries are the more northern tributaries
with more forested watersheds.
The Fox, Grand Calumet, and Kalamazoo Rivers were dominated by mercury in the particulate fraction.
Each of these tributaries contained more than 75% of total mercury in the particulate fraction. These
three tributaries are among the most urbanized and industrialized watersheds evaluated in the study.
In addition to measurement of total and dissolved mercury, methylmercury was measured in the total and
dissolved fractions. In most of the tributaries, methylmercury comprised less than 6% of the total
mercury (Table 4-6). This is consistent with USEPA (1997c) reports that less than 10% of total mercury
in a water column typically exists as a methylmercury complex. The one exception was the Muskegon
River, where methylmercury accounted for an average of 21% of total mercury. As Hurley et al. (1998b)
explained, Lake Muskegon is located directly upstream of the Muskegon River sampling site. This lake
traps particulates and particulate-bound contaminants, which reduces the load of particulate mercury in
the Muskegon River. As evidence of this, the Muskegon River had the lowest particulate mercury
concentration (virtually zero), the lowest particulate organic carbon concentration (0.537 mg/L), and the
lowest total solids concentration (3.04 mg/L). In addition to reducing the particulate load of mercury,
Lake Muskegon could provide favorable conditions for the methylation of mercury. This could explain
the much higher percentage of methylmercury in the Muskegon River than other tributaries.
Methylmercury is the bioavailable form of mercury that is readily accumulated and biomagnified in
aquatic food webs. While methylmercury accounts for less than 10% of the total mercury in surface
waters, methylmercury typically accounts for more than 90% of total mercury in fish tissue (Watras and
Bloom, 1992).
Table 4-6. Percentages of Total Mercury Found in Various Forms
Tributary
Fox
Grand Calumet
Grand
Kalamazoo
Manistique
Menominee
Milwaukee
Muskegon
Pere Marquette
Sheboygan
St. Joseph
Mean Percent of Total Mercury as a
Dissolved
15
8
43
19
78
54
34
99
80
38
29
Particulate
85
92
57
81
22
46
66
0.64
20
62
71
Methylmercury
0.97
0.48
2.6
2.0
4.7
5.3
5.2
21
5.6
5.9
2.1
a The dissolved and particulate fractions are mutually exclusive and add to 100% of the total mercury. The percent of total mercury in the form
of methylmercury is presented separately, however, this portion may exist in either dissolved or particulate fractions as well and is already
accounted for in those fractions.
4-14
-------
Mercury in Tributaries
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 mercury
monitoring portion of the study are further described in Section 2.6 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, 2001b). A brief summary of the quality of tributary mercury and
methylmercury 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 2.6, 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. Tables 4-7 and 4-8
provide a summary of flags applied to the tributary mercury and methylmercury data, respectively. The
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, as discussed in Section 2.6. A total of 15 dissolved mercury and 15 total mercury samples
were flagged as invalid by the PI. These samples were invalidated because they were prepared and
analyzed without a Tenax TA® pretrap (see section 3.19 of USEPA 1997b) and data quality was
significantly reduced. These samples were not used in any of the statistical analyses described in this
report. For methylmercury, no samples were flagged invalid, and therefore, all results were used in the
statistical analyses described in this chapter.
Table 4-7. Summary of Routine Field Sample Flags Applied to Mercury Data from Lake Michigan Tributaries
Flag
INV, Invalid Result
EHT, Exceeded Holding Time
FDL, Failed Lab Duplicate
FFD, Failed Field Duplicate
FSL, Failed Lab Fortified Spike
Number of QC samples
Dissolved
—
—
340 lab duplicate
groups
49 field duplicate pairs
65 lab fortified spike
samples
Total
—
—
347 lab duplicate
groups
49 field duplicate pairs
53 lab fortified spike
samples
Percentage of Samples Flagged
Dissolved
4% (15)
0
4% (15)
3% (9)
1%(3)
Total
4% (15)
0
2% (6)
3% (11)
1%(2)
The most frequently applied data validation flag for methylmercury data was for exceeding sample
holding times. More than half of the samples analyzed for methylmercury (55% of dissolved
methylmercury, and 57% of total methylmercury samples) were analyzed beyond the 2-year established
4-15
-------
Results of the LMMB Study: Mercury Data Report
holding time. The median holding time for methylmercury samples was 1,358 days, and samples were
held as long as 1,897 days prior to methylmercury analysis. The MQOs for holding times were based on
educated, conservative assessments by the Pis, however, the appropriateness of these holding times have
not been rigorously determined and the effects of extended holding times have not been investigated in
the tributary matrix. All total and dissolved mercury samples were analyzed within the 2-year holding
time, and therefore, no total or dissolved mercury results were flagged for exceeding the holding time.
Table 4-8. Summary of Routine Field Sample Flags Applied to Methylmercury Data from Lake Michigan
Tributaries
Flag
INV, Invalid Result
EHT, Exceeded Holding Time
FDL, Failed Lab Duplicate
FFD, Failed Field Duplicate
FSL, Failed Lab Fortified Spike
Number of QC samples
Dissolved
—
—
14 lab duplicate pairs
28 field duplicate groups
19 lab fortified spike
samples
Total
—
—
1 1 lab duplicate pairs
30 field duplicate groups
25 lab fortified spike
samples
Percentage of Samples Flagged
Dissolved
0
55% (11 3)
3% (6)
10% (21)
16% (33)
Total
0
57% (11 7)
0.5% (1)
9% (18)
19% (38)
Field blanks were analyzed to assess the potential for contamination of routine field samples. For total
and dissolved mercury, a total of 36 blanks were analyzed, including 12 field reagent blanks, 12 field
tubing blanks and 12 field filter blanks. Two field tubing blanks and one field reagent blank contained
greater than 1 ng/L mercury and were flagged as contaminated according to the established MQOs. The
maximum mercury concentration in these blanks was 1.2 ng/L. In addition, one other field reagent blank
and associated field filter blank were flagged because the difference between these two blank
concentrations and their associated field tubing blank was greater than 0.50 ng/L. In total, 14% of the
blanks were flagged for contamination. However, because the blanks could not be associated with
individual field samples, no field samples were flagged for blank failures. For methylmercury, no blank
contamination flags were applied to the field samples. One field trip blank sample was analyzed, with a
concentration of-0.0050 ng/L. Negative values are possible for methylmercury due to the analytical
methodology, which involves the subtraction of results from two analytical steps.
Field and laboratory duplicate samples were analyzed to assess the precision of the measurement system.
A total of 88 and 60 valid field duplicate samples were analyzed for mercury and methylmercury,
respectively, including 2 cases where a methylmercury field sample had multiple duplicates. All field
duplicate samples were classified as "sequential" because the duplicates were not collected within five
minutes of the original sample due to equipment mobilization and sample pumping time. At least three
sequential field duplicates were collected from each tributary for total and dissolved mercury analysis.
For methylmercury analysis, at least one sequential field duplicate was collected from every tributary
except for the Fox River. In accordance with the researcher's data qualifying rules for field duplicates,
total and dissolved mercury samples were flagged for a failed field duplicate (FFD) based on a maximum
relative percent difference (RPD) of 30% for samples greater than 5 times the method detection limit
(MDL) and 50% for samples less than 5 times the MDL. A total of 9 dissolved mercury samples and 11
total mercury samples exceeded these maximum RPD limits. For methylmercury, a maximum RPD limit
of 30% was used if all results were above 0.10 ng/L (approximately 5 times the MDL), and an absolute
difference of 0.030 ng/L was used if at least one result was below 0.10 ng/L. These criteria were
exceeded for 39 field duplicate pairs, however, only 8 of these pairs failed using the RPD criterion. The
4-16
-------
Mercury in Tributaries
remaining 31 pairs failed based on the absolute difference criterion, with the maximum absolute
difference between duplicates equaling 1.1 ng/L.
For total and dissolved mercury analysis, at least one laboratory duplicate was prepared for all but 19
field samples. For some samples, multiple laboratory duplicates (up to 4) were prepared. Laboratory
duplicates also were prepared for several field duplicate samples. For methylmercury analysis, laboratory
duplicates were prepared for only 25 field samples, with no more than one laboratory duplicate prepared
for a given sample. In accordance with the researcher's data qualifying rules for lab duplicates, total and
dissolved mercury samples were flagged for a failed duplicate (FDL) based on a maximum RPD level (or
RSD if more than one lab duplicate was analyzed for a given sample) of 20% for samples greater than 5
times the MDL and 50% for samples less than 5 times the MDL. A total of 15 dissolved and 6 total
mercury sample pairs exceeded these maximum RPD/RSD criteria, with a maximum RPD/RSD of 80%
calculated. For methylmercury, the rules for determining lab duplicate failure were the same as those
used for determining field duplicate failure. These criteria were exceeded for 7 laboratory duplicate pairs.
Three of these pairs failed using the RPD criterion and 4 pairs failed based on the absolute difference
criterion. The maximum RPD measured for methylmercury samples was 107%, and the maximum
absolute difference (between field sample and duplicate) was 0.34 ng/L.
To monitor the potential bias of analytical results, the laboratory prepared and analyzed a total of 162
laboratory fortified spike samples (LSFs). Samples were flagged for a failed lab fortified spiked sample
(FSL) if the associated spike recovery was below 70% or above 130%. The FSL flag was applied to 1%
of the total and dissolved mercury samples, due to two recoveries below the lower limit, with a minimum
of 66%, and three recoveries above the upper limit, with a maximum of 159%. The FSL flag was applied
to 16% of dissolved methylmercury and 19% of total methylmercury samples, due to one recovery below
the lower limit (69%) and four above the upper limit, with a maximum of 153%. Based on analysis of
laboratory spikes, blank contamination, and other internal QC data, the QC coordinator did not qualify
any samples as high or low biased.
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 field sample and duplicate result for laboratory duplicate pairs. Tables 4-9 and 4-10
provide summaries of data quality assessments for several of these attributes for tributary mercury and
methylmercury data, respectively. The results of laboratory and field duplicate samples revealed good
system and analytical precision for total and dissolved mercury data when the results were above 5 times
the given MDL. System precision was described by mean RPDs of 17% and 20% for dissolved and total
field duplicate samples, respectively. Analytical precision was even greater, with RPDs as low as 7.5%
and 5.1% for dissolved and total mercury samples, respectively. When results were less than 5 times the
MDL, mean RPDs were much higher. For field duplicates, the mean RPD was 45% for the 7 dissolved
duplicate pairs and 182% for the one total duplicate pair. For laboratory duplicates, the mean RPDs were
14% for dissolved mercury samples and 54% for total mercury samples.
Methylmercury results were less precise than total and dissolved mercury results. For results that were
greater than 5 times the MDL, mean field duplicate RPDs were 47% for dissolved methylmercury and
27% for total methylmercury. Mean laboratory duplicate RPDs were 47% and 13% for dissolved and
total methylmercury, respectively, when all results were above 5 times the MDL. When results were less
than 5 times the MDL, mean field duplicate RPDs were 99% and 51% for dissolved and total
methylmercury, respectively. Mean laboratory duplicate RPDs were 62% and 26% for dissolved and
total methylmercury, respectively.
4-17
-------
Results of the LMMB Study: Mercury Data Report
Analytical bias was evaluated by calculating the mean recovery of LSF samples. Results indicated very
little overall bias for analytical results. The mean LSF recovery for total and dissolved mercury was
103%. For methylmercury, the mean LSF recovery for dissolved samples was 99%, and the mean LSF
recovery for total methylmercury was 110%.
Analytical sensitivity was evaluated by calculating the percentage of samples reported below the
corresponding MDL (0.10 ng/L for total and dissolved mercury, and 0.019 ng/L fortotal and dissolved
methylmercury). Only one dissolved mercury sample, or 0.3% of the data, and no total mercury samples,
were below the detection limit for total mercury. For methylmercury, 31 dissolved samples (15% of the
data) and 6 total samples (3% of the data) were below the MDL. Results from these samples were not
censored and were used as reported in the analysis of tributary mercury data presented in this report.
Table 4-9. Data Quality Assessment for Mercury Data from Lake Michigan Tributaries
Parameter
Number of Routine Samples Analyzed
Number of Sequential Field Duplicates Analyzed
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 LFS (%)
Analytical Sensitivity, Samples reported as < MDL (%)
Assessment3
Dissolved
346
49
45% (7)
17% (34)
14%(29)b
7.5% (338)b
103% (65)
0%
Total
353
49
182%(1)
20% (46)
54%(1)b
5.1%(381)b
103% (53)
0%
a Number of QC samples used in the assessment is provided in parentheses
b Includes laboratory duplicates of field duplicate samples
LFS = Laboratory Fortified Spike
MDL = Method Detection Limit
Table 4-10. Data Quality Assessment for Methylmercury Data from Lake Michigan Tributaries
Parameter
Number of Routine Samples Analyzed
Number of Sequential Field Duplicate Groups Analyzed
System Precision, Mean Field Duplicate RPD (%), < MDL
System Precision, Mean Field Duplicate RPD (%), > MDL
Analytical Precision, Mean Lab Duplicate RPD (%), < MDL
Analytical Precision, Mean Lab Duplicate RPD (%), > MDL
Analytical Bias, Mean LFS (%)
Analytical Sensitivity, Samples reported as < MDL (%)
Assessment3
Dissolved
204
28
99% (22)
47% (3)
62% (9)
47% (4)
99% (19)
15%
Total
203
30
51% (17)
27% (12)
26% (3)
13% (8)
110% (25)
3%
a Number of QC samples used in the assessment is provided in parentheses
LFS = Laboratory Fortified Spike
MDL = Method Detection Limit
4-18
-------
Mercury in Tributaries
4.3 Data Interpretation
4.3.1 Mercury Levels in Lake Michigan Tributaries
Total mercury concentrations in Lake Michigan tributaries averaged from 1.07 ng/L in the Muskegon
River to 28.9 ng/L in the Fox River. Following the Fox River, the Kalamazoo and Grand Calumet Rivers
averaged approximately 10 ng/L in total mercury. The remaining tributaries averaged from 1 to 6 ng/L in
total mercury. These mercury levels are comparable to mercury concentrations measured in other
Midwestern tributaries. In a survey of 39 Wisconsin rivers, Hurley et al. (1995) measured a mean total
mercury concentration of 7.94 ng/L during the spring and 3.45 ng/L during the fall. This is consistent
with LMMB Study data, where a majority of tributaries averaged between 3 and 7 ng/L total mercury.
Similarly, Thompson-Roberts et al. (1999), measured average total mercury concentrations of 3 to 19
ng/L in 23 wetlands of the St. Lawrence River. Balogh et al. (1998) reported total mercury
concentrations below 4 ng/L in the St. Croix River, below 10 ng/L in the headwaters of the Mississippi
River, and routinely above 10 ng/L in the Minnesota River. In a summary of surface water mercury
levels nationwide, USEPA (1997c) reported that total mercury levels in lakes and streams are typically
well under 20 ng/L, however, elevated levels may be found in lakes and streams thought to be impacted
by anthropogenic mercury sources. This is consistent with the results of this study, where all tributaries
except for the Fox River were below 20 ng/L, and the Fox River is suspected of being impacted by
resuspension of contaminated sediments from legacy sources (Hurley etal, 1998a).
4.3.2 Comparison to Regulatory Limits
The average concentrations of mercury in Lake Michigan tributaries were all below EPA's nationwide
freshwater water quality criterion for human health protection of 50 ng/L, and only the Fox River
exceeded the chronic water quality criterion for protection of aquatic life (12 ng/L). When compared to
the more stringent water quality criteria recommended for Great Lakes states, three tributaries exceed the
Great Lakes water quality criterion for human health (1.8 ng/L dissolved mercury) and eight tributaries
exceed the Great Lakes water quality criterion for wildlife (1.3 ng/L dissolved mercury). The Fox,
Manistique, and Menominee Rivers exceed the human health criterion, and all tributaries except for the
Muskegon, Milwaukee, and Grand Calumet Rivers exceed the wildlife criterion.
4.3.3 Seasonality
While tributaries differed in their seasonal patterns of flow and mercury concentrations, many of the Lake
Michigan tributaries exhibited significantly lower mercury concentrations during the winter and higher
mercury concentrations in conjunction with spring high-flow conditions or event flows during the
summer and fall. Balogh et al. (1998) similarly found that total mercury concentrations in the Minnesota,
St. Croix, and Mississippi Rivers varied seasonally with lowest levels during the winter, increasing
concentrations during spring runoff, and fluctuating concentrations throughout the spring, summer, and
fall in response to precipitation runoff events. In the Minnesota River, Balogh et al. (1997) reported total
mercury concentrations from less than 1.0 ng/L during the winter months to greater than 35 ng/L
following spring runoff. When comparing just spring and fall concentrations, Hurley etal. (1995) found
strong seasonal variability in 39 Wisconsin Rivers, with total mercury concentrations approximately two
times higher in the spring than in the fall.
In tributaries that are dominated by particulate mercury, lower total mercury concentrations during the
winter are tied to lower suspended solids concentrations during the winter. The low-flow conditions that
occur during the winter in conjunction with the ice cover that forms over many Lake Michigan tributaries
contribute to reduced turbulence and reduced sediment resuspension. This reduced suspended sediment
load during the winter decreases particulate, and therefore total, mercury concentrations in the water
4-19
-------
Results of the LMMB Study: Mercury Data Report
column (Hurley et al., 1998a). This conclusion is consistent with correlations of total mercury with
particulate organic carbon concentrations, total solids concentrations, and suspended particulate matter
identified in this and other studies (Hurley et al., 1998a; Balogh et al., 1998; Balogh et al., 1997).
Seasonal differences in the fluxes of mercury from Lake Michigan tributaries were even more apparent
than seasonal differences in mercury concentrations alone. Hurley et al. (1998b) investigated the fluxes
of mercury from Lake Michigan tributaries during three flow regimes: spring, base flow, and event. For
all tributaries except the Grand Calumet, base flow fluxes were considerably lower than fluxes during
either spring or event conditions. In comparing spring and event fluxes, Hurley et al. (1998b) found that
the patterns of mercury flux and flow regimes differed among the tributaries. In the Fox, St. Joseph, and
Manistique Rivers, fluxes associated with the spring flows were much greater than those associated with
summer and fall events. In contrast, mercury fluxes in the Grand and Kalamazoo Rivers were greater
during summer and fall events than during spring flows. These differences were explained in part by
differences in watershed land use patterns (Hurley et al., 1998b). The Grand and Kalamazoo River
watersheds contain significant agricultural land cover with increased particulate erosion susceptibility
during precipitation events.
4.3.4 Regional Considerations
Of the 11 Lake Michigan tributaries evaluated in the LMMB Study, total mercury concentrations were
highest in the Fox River. Average total mercury concentrations in the Fox River were 2.7 times higher
than in any other tributary. The maximum total mercury concentration of 191 ng/L measured in the Fox
River was more than four times higher than the maximum concentration measured in any other tributary.
Following the Fox River, total mercury concentrations were highest in the Grand Calumet and Kalamazoo
Rivers. Total mercury concentrations in these two rivers were significantly higher than in any other
tributary, except for the Fox River. Each of these rivers (the Fox, Grand Calumet, and Kalamazoo) have
significantly urbanized and industrialized watersheds, which suggests anthropogenic sources. In more
intensive surveys of the lower Fox River that included longitudinal transect sampling and analysis of
sediment cores, Hurley et al. (1998a) concluded that mercury enrichment in the Fox River was due to
resuspension of historically contaminated sediments. Mercury concentrations of up to 5.69 |ig/g in
deeper sediment cores (18-cm composites) in conjunction with scouring from high flow events were
sufficient to produce the water column mercury levels measured at the mouth of the Fox River. Hurley et
al. (1998b) also measured mercury levels in the suspended particulate matter on a ng/g basis and
concluded that the Fox and Grand Calumet Rivers contained particles that were highly enriched with
mercury compared to the other tributaries. Levels of mercury in particles from the remaining tributaries
were generally 50 to 200 ng/g and in the range reported for Midwestern soils.
While the highest total mercury concentrations were observed in urban and industrial watersheds, the
lowest total mercury concentrations were observed in predominantly forested and wetland watersheds.
The more-northern Muskegon, Manistique, Pere Marquette, and Menominee Rivers contained the lowest
total mercury concentrations, averaging only 1.07 to 3.63 ng/L. Hurley et al. (1995) also found that
mercury yields varied by watershed land use patterns in 39 Wisconsin rivers. Mean spring concentrations
and yields of mercury were highest in urban watersheds, followed by wetland and forest watershed, with
lowest values in agricultural watersheds.
4.3.5 Mercury Fractions and Forms
Tributaries also differed in the fractions and forms of mercury present. In each of the three most
mercury-contaminated tributaries (Fox, Grand Calumet, and Kalamazoo Rivers), mercury was
predominantly in the particulate fraction. Particulate mercury accounted for 85%, 92%, and 81% of total
mercury in the Fox, Grand Calumet, and Kalamazoo Rivers, respectively. In the least contaminated
4-20
-------
Mercury in Tributaries
tributaries (the Muskegon, Manistique, Pere Marquette, and Menominee Rivers), total mercury
concentrations were dominated by the dissolved fraction. The dissolved fraction accounted for 54% to
99% of total mercury in these tributaries. In fact, the Manistique, Menominee, and Pere Marquette Rivers
contained the second, third, and fourth highest average dissolved mercury concentrations. Hurley et al.
(1998b), however, notes that on a flux basis, inputs of dissolved mercury from the Fox, Kalamazoo,
Grand, and St. Joseph Rivers are of the same magnitude as those from the dissolved mercury-dominated
tributaries.
Balogh et al. (1998) found similar results when investigating mercury in diverse Minnesota river basins.
In the more forested and wetland-dominated watershed of the St. Croix River, the dissolved fraction
dominated mercury mobility, while the particulate fraction dominated mercury mobility in the agricultural
Minnesota River watershed. Dissolved mercury accounted for over 62% of the total mercury in the St.
Croix River and less than 10% of the total mercury in the Minnesota River. Likewise, wetland/forest
watersheds in Wisconsin were dominated by mercury fluxes in the filtered fraction, while agricultural
watersheds were dominated by mercury fluxes in the particulate fraction (Hurley et al., 1995).
With the exception of the Muskegon River (where methylmercury accounted for 21% of total mercury),
methylmercury accounted for only 0.48% to 5.9% of total mercury in Lake Michigan tributaries. In a
study of 39 Wisconsin rivers, Hurley et al. (1995) similarly found that methylmercury accounted for an
average of less than 2.2% to 6.4% of total mercury. Lake Michigan tributaries such as the Fox, Grand
Calumet, and Kalamazoo Rivers that had the highest total mercury concentrations did not have
correspondingly high methylmercury concentrations. These tributaries ranked fifth, sixth, and tenth in
total methylmercury concentrations among the tributaries. Hurley et al. (1998b) cautioned, however, that
just because those sites with high total mercury levels contained only a small portion of mercury in more
bioavailable dissolved and methyl forms, these loads should not be discounted as inert. These particulate-
bound contaminants can be deposited in Lake Michigan sediments and undergo methylation,
reintroducing biologically available mercury to the Lake Michigan system.
4-21
-------
Chapter 5
Mercury in the Open-Lake Water Column
Open-lake water column samples were collected during six cruises of the R/VLake Guardian conducted
from April 1, 1994 to October 22, 1995. Samples were collected at 17 sampling locations, including 15
stations in Lake Michigan, 1 location in Green Bay and 1 location in Lake Huron (see Figure 2-4).
Samples were collected at depths ranging from 1 m to 150 m. Samples were collected as described in
Section 2.4.3 and analyzed for total and particulate mercury by cold-vapor atomic fluorescence
spectrometry (see Section 2.5.3). In addition, dissolved mercury results were calculated by subtracting
the particulate mercury result from the total mercury result, when results from both fractions were
reported.
5.1 Results
A total of 121 samples were analyzed for particulate mercury, and a total of 125 samples were analyzed
for total mercury (Table 5-1). Particulate mercury results ranged from 0.027 ng/L to 0.30 ng/L, with
approximately 8% of the samples below the associated daily detection limit. Total mercury results ranged
from 0.037 ng/L to 0.78 ng/L, with approximately 4% of the samples below the associated daily detection
limit. Combining data from all depths and all cruises, the lake-wide mean mercury concentrations
measured in this study were 0.33 ng/L for total mercury and 0.11 ng/L for particulate mercury.
Table 5-1. Numbers of Open-Lake Samples Analyzed for Mercury
Sampling Station
GB24M
LH54M
05
140
180
18M
23M
240
27M
280
340
380
40M
41
47M
19M
72M
Sampling Dates
08/08/94 to 09/20/95
08/03/94 to 09/1 6/95
08/24/94 to 10/1 0/95
06/1 8/94 to 09/23/95
04/07/95 to 04/07/95
06/22/94 to 10/09/95
06/23/94 to 10/03/95
06/2 1/94 to 10/02/95
06/20/94 to 09/27/95
04/01/95 to 04/01/95
08/2 1/94 to 10/06/95
03/26/95 to 03/26/95
10/1 8/94 to 09/25/95
06/1 8/94 to 10/22/94
06/1 7/94 to 09/1 9/95
08/1 9/94 to 10/05/95
08/04/94 to 09/1 7/95
Total
Particulate Samples
7
10
8
8
1
12.
12
7
10
1
7
1
7
4
12
8
6
121
Total Mercury
Samples
7
10
8
8
1
12
12
9
10
1
7
1
8
5
12
8
6
125
Total Number of
Samples
14
20
16
16
2
24
24
16
20
2
14
2
15
9
24
16
12
246
' One sample was invalid.
GB = Green Bay station
LH = Lake Huron station
5-1
-------
Results of the LMMB Study: Mercury Data Report
5.1.1 Geographical Variation
From 1 to 12 samples were collected at each of 17 different stations in Lake Michigan, Green Bay, and
Lake Huron. The mean concentrations are shown in Figure 5-1, and descriptive statistics of the
particulate and total mercury concentrations reported at each station are presented in Table 5-2. Mean
particulate mercury concentrations ranged from 0.029 ng/L at Station 380 to 0.17 ng/L at Station GB24M
in Green Bay. The maximum mean particulate mercury concentration in Lake Michigan was 0.13 ng/L,
and occurred at five different stations. Mean total mercury concentrations ranged from 0.25 ng/L at
Station 41 to 0.78 ng/L at Station 380. While the mean particulate and total mercury concentrations
collected at Station 380 were extremely low and high, respectively, compared to the other stations, these
means only represent a single sample result at this station. Therefore, it is unlikely that these means are
representative of the mercury concentrations at that station.
The highest mean particulate mercury value was in Green Bay (GB24M). This finding is not unexpected,
due to the large inputs of mercury, particularly in the particulate phase, from the Fox River (see Chapter
4). While particulate mercury concentrations were slightly higher in Green Bay than other sampling sites,
there were no significant differences among site in particulate mercury concentrations, based on a one-
way Analysis of Variance (ANOVA) model using log-transformed results (p=0.1685). Mean total
mercury concentrations were relatively consistent throughout Lake Michigan. No statistical differences
were observed among sampling sites, based on a one-way ANOVA model using log-transformed results
(p=0.2309).
Figure 5-1. Mercury Concentrations Measured in Open-lake Water Column Samples
0.8
0.7 -
0.6 -
•=• 0.5 -
C
o
I 0.4
o
5
O
I 0.3 H
0.2 -
0.1 -
i
J
1
.N!^" .\V'
Station
Stations are from Lake Michigan except for GB24M (Green Bay) and LH54M (Lake Huron). Bars show the mean mercury
concentration of samples collected at each station for the duration of the study. Error bars are standard error.
5-2
-------
Mercury in the Open-Lake Water Column
Table 5-2. Mean Particulate and Total Mercury Concentrations Measured in Open Lakes
Fraction
Particulate
Total
Sampling Station
140
180
18M
23M
240
27M
280
340
380
40M
41
47M
5
GB24M
LH54M
19M
72M
140
180
18M
23M
240
27M
280
340
380
40M
41
47M
5
GB24M
LH54M
19M
72M
N
8
1
11
12
7
10
1
7
1
7
4
12
8
7
10
8
6
8
1
12
12
9
10
1
7
1
8
5
12
8
7
10
8
6
Mean
(ng/L)
0.12
0.13
0.095
0.13
0.087
0.12
0.063
0.13
0.029
0.073
0.11
0.13
0.10
0.17
0.13
0.12
0.12
0.40
0.32
0.28
0.30
0.30
0.33
0.49
0.30
0.78
0.35
0.25
0.29
0.37
0.33
0.28
0.33
0.39
Median
(ng/L)
0.12
0.13
0.094
0.11
0.063
0.12
0.063
0.13
0.029
0.073
0.10
0.13
0.10
0.19
0.12
0.13
0.13
0.42
0.32
0.27
0.30
0.27
0.28
0.49
0.30
0.78
0.30
0.25
0.28
0.33
0.29
0.34
0.30
0.34
Range
(ng/L)
0.049 to 0.1 9
NA
0.030 to 0.1 5
0.031 to 0.24
0.038 to 0.1 6
0.030 to 0.30
NA
0.05 to 0.19
NA
0.038 to 0.11
0.097 to 0.14
0.035 to 0.28
0.032 to 0.1 5
0.076 to 0.30
0.079 to 0.27
0.027 to 0.20
0.057 to 0.1 7
0.21 to 0.61
NA
0.14 to 0.46
0.21 to 0.48
0.19 to 0.48
0.22 to 0.57
NA
0.22 to 0.39
NA
0.19 to 0.57
0.19 to 0.30
0.075 to 0.48
0.19 to 0.55
0.1 6 to 0.56
0.037 to 0.49
0.20 to 0.54
0.30 to 0.59
SD
(ng/L)
0.053
NA
0.033
0.065
0.046
0.077
NA
0.047
NA
0.029
0.020
0.070
0.040
0.076
0.054
0.066
0.045
0.14
NA
0.11
0.086
0.10
0.12
NA
0.062
NA
0.15
0.040
0.12
0.14
0.14
0.15
0.11
0.12
RSD
(%)
44
NA
35
51
53
63
NA
37
NA
40
18
53
40
45
41
53
36
35
NA
38
29
34
36
NA
21
NA
44
16
42
37
44
52
33
30
Below DL
(%)
13
0.0
0.0
0.0
14
10
0.0
0.0
0.0
0.0
0.0
17
25
14
20
0.0
0.0
0.0
0.0
8.3
8.3
0.0
0.0
0.0
0.0
0.0
13
0.0
8.3
0.0
0.0
10
0.0
0.0
NA = Not applicable
GB = Green Bay station
LH = Lake Huron station
5-3
-------
Results of the LMMB Study: Mercury Data Report
Statistical comparisons also were performed after combining the 15 stations in Lake Michigan into two
different basins. For these comparisons, the data from the LMMB Study were divided at approximately
44° north 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. The latitude limit was instead chosen to remain consistent with analyses performed on
PCB and atrazine data. The results from the stations in Green Bay and Lake Huron were excluded from
these comparisons. Based on the 44° N dividing line, six of the 15 Lake Michigan stations were
categorized as being in the northern basin (40M, 41, 47M, 72M, 140 and 180).
The results of the basin comparisons were similar to those of the comparisons of individual stations.
For both particulate and total mercury, there were no significant differences in mercury concentration
between basins (particulate: p = 0.1046; total: p = 0.2523) or between stations nested within basin
(particulate: p = 0.3869; total: p = 0.0805).
The lack of spatial differences is consistent with previous assessments that suggest that the primary
source of mercury is atmospheric rather than riverine (Mason and Sullivan, 1997). The effect of the
variability in mercury concentration among the tributaries, as discussed in Chapter 4, is only seen in the
slightly greater particulate mercury concentration in Green Bay at station GB24M. However, the total
mercury concentration at this station did not exhibit any effect of the Fox River, as the mean
concentration of 0.30 ng/L was below the overall mean total mercury concentration. Therefore, it is
likely that most of the mercury from the Fox River is removed to the sediment rather than staying in the
water column (Sullivan and Mason, 1998).
5.1.2 Seasonal Variation
Samples were collected during six cruises: June 1994, August 1994, October/November 1994,
March/April 1995, August 1995 and September/October 1995. During each cruise, up to 2 samples were
collected at each station. Descriptive statistics for particulate and total mercury for each cruise are
presented in Table 5-3.
Table 5-3. Mean Particulate and Total Mercury Concentrations by Cruise
Fraction
Particulate
Total
Sampling Cruise
June 1994
August 1994
Oct/Nov. 1994
March/April 1995
August 1995
Sept/Oct. 1995
June 1994
August 1994
Oct/Nov. 1994
March/April 1995
August 1995
Sept/Oct. 1995
N
12
23
18
23
23
21
14
23
20
24
23
21
Mean
(ng/L)
0.16
0.16
0.12
0.11
0.12
0.052
0.34
0.29
0.33
0.38
0.36
0.24
Median
(ng/L)
0.13
0.15
0.12
0.11
0.10
0.043
0.29
0.27
0.31
0.36
0.35
0.24
Range
(ng/L)
0.097 to 0.28
0.10 to 0.30
0.053 to 0.20
0.029 to 0.21
0.052 to 0.30
0.027 to 0.1 2
0.19 to 0.61
0.075 to 0.54
0.16 to 0.59
0.037 to 0.78
0.23 to 0.56
0.14 to 0.37
SD
(ng/L)
0.060
0.047
0.039
0.040
0.062
0.024
0.12
0.12
0.12
0.16
0.10
0.062
RSD
(%)
37
30
32
37
53
46
35
41
36
41
29
26
Below DL
(%)
0.0
43
0.0
0.0
0.0
0.0
0.0
13
0.0
8.3
0.0
0.0
5-4
-------
Mercury in the Open-Lake Water Column
Mean participate mercury concentrations generally decreased over the course of the study, ranging from
0.16 ng/L in the June and August 1994 cruises to 0.052 ng/L in the autumn 1995 cruise. Based on a one-
way ANOVA model, the difference between cruises was significant (p<0.0001). Subsequent Tukey
pairwise comparisons showed that the means for the first two cruises were significantly greater than the
means for the last three cruises, and that the mean of the last cruise was significantly lower than the
means for all other cruises (Figure 5-2A). Unlike particulate mercury, mean total mercury concentrations
did not appear to follow a trend. The maximum mean total mercury concentration occurred in
March/April 1995, rather than in summer 1994. However, similar to particulate mercury, the minimum
concentration occurred in September/October 1995. A one-way ANOVA model comparing mean total
mercury concentrations between cruises was statistically significant (p = 0.0015). Tukey pairwise
comparisons showed that the means of the March/April and August 1995 cruises were significantly
greater than the mean for the September/October 1995 cruise and that the mean of the March/April cruise
was significantly greater than the mean of the August 1994 cruise (Figure 5-2B).
Because the timing of the cruises differed between the two years of collection, it is difficult to interpret
the concentration differences between cruises as seasonal or annual differences. Cruises 2 and 5 occurred
during August, however, and differences could be interpreted as due to differences between 1994 and
1995. Based on profiles of temperature and pH, Sullivan and Mason (1998) concluded that productivity
in the lake was lower in the summer of 1994 compared to the summer of 1995. They hypothesize that the
increase in pH from August 1995 to September/October 1995 is evidence of the pH-induced precipitation
of calcite, a mineral form of calcium carbonate, and they conclude that the seasonal dynamics of the lake
differed between the two years of the LMMB Study. These differences in dynamics may have an effect
on the concentrations and partitioning of mercury in the lake.
5-5
-------
Results of the LMMB Study: Mercury Data Report
Figure 5-2. Particulate and Total Mercury Concentrations Measured in Open Lakes, by Cruise
Particulate Mercury
0.35n
0.281! -p
B
0.21-
AB
B
ro
^ 0 "1 4-
Q)
o
o
0 0.07-
0 00=
1 c
1 — 1 — 1 1 1 u
o o o o o o
E E E E E E
u)' (/)' (/)' (/)' (/)' w"
CD CD CD CD CD CD
-^ ro co ^. cn o
Total Mercury
0.80n
-
§ 0.60-
1 0.40^
w :
c
o o 20-
O
0 00-
ABC
T '
J_
o
c
8
ABC
3C
[ f
1
Lnr— '
-1-
O O
c c
8 8
A
AB
T
i
C
— — i —
T JL
-L- | 1
J_
o o o
c c c
WWW
000
II
N)
OO
Cruise 1 = June 1994, Cruise 2 = August 1994, Cruise 3 = September/October 1994, Cruise 4 = March/April 1995, Cruise 5 =
August 1995, and Cruise 6 = September/October 1995
Boxes represent the 25th (box bottom), 50th (center line), and 75th (box top) percentile 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. Xs represent results beyond 3*IQR from the box. Letters above the boxes represent
results of analysis of variance and multiple comparisons test. Boxes with the same letter were not statistically different (at alpha
= 0.05).
5-6
-------
Mercury in the Open-Lake Water Column
5.1.3 Vertical Variation
Open-lake samples were collected at depths ranging from 1 to 150 m. The correlation between sampling
depth and mercury concentration (both log-transformed) was weak for particulate (r2 = 0.057) mercury,
and did not differ significantly from 0 (p = 0.539). The correlation between depth and concentration for
total mercury (r2 = -0.203) was also somewhat weak, but differed significantly from 0 (p = 0.0235). The
overall weak correlation between depth and concentration may be due to station variability and variability
among cruises conducted during completely mixed or thermally stratified conditions. If correlations are
calculated based on only the samples collected during stratified conditions, (i.e., cruises during late
summer and autumn months), the negative correlation for total mercury strengthens (r2 = -0.393, p =
0.0002), while the particulate mercury correlation remains weak (r2 = 0.047, p = 0.666). The correlation
is presented graphically in Figure 5-3. While the relationship does not appear strong, concentrations
collected at depths above (shallower than) 40 meters were significantly greater than those collected below
40 meters, based on a two-sample t-test (p = 0.0008).
Figure 5-3. Total Mercury Concentration versus Sample Depth During Stratified Conditions
S2
I »
.n
'S. so
01
Q
ro
V)
• «.«»«»-
« • » »**
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Hg Concentration (ng/L)
To further account for station and cruise variability, a paired t-test was used to compare the mercury
concentration at the deeper depth (hypolimnion) to the concentration at the shallower depth (epilimnion)
where samples were collected at two depths for a given cruise and station. Two sample pairs for which
both depths were either above 20 meters or below 20 meters were not included in the analyses, leaving 27
pairs for particulate mercury and 28 pairs for total mercury. These pairs were collected either during a
late summer cruise (August 1994, August 1995) or an autumn cruise (October/November 1994,
September/October 1995). When tests were conducted separately by these seasonal categories, there was
a significant difference between the two depths for total mercury during the late summer (p = 0.0141),
where the concentration was greater in the hypolimnion, but not for the autumn (p = 0.7337). These
comparisons are shown in Figure 5-4. There were no significant differences between the two depths for
either season for particulate mercury (Summer: p = 0.1230, Autumn: p = 0.7867). The lack of a
difference between depths during the autumn cruises may be due to a decomposing thermocline late in the
fall season (i.e., the end of stratification).
5-7
-------
Results of the LMMB Study: Mercury Data Report
Figure 5-4. Total Mercury Concentrations at Stations with Samples from Multiple Depths
0.6
c 0.4
o
'c
E
"o
Q.
>.
I
c
o
c
o
O
E1 0.2
3
o
01
•o
o
0.2 0.4
Mercury Cone, in Epilimnion (ng/L)
0.6
Statistical comparisons were also conducted to compare mercury concentrations for the two seasonal
categories defined above separately for the epilimnion and hypolimnion samples. Based on two-sample
Wests with the Satterthwaite correction for differences in variability, there was a significant difference in
total mercury concentration between the two seasons for the shallower, epilimnion samples (Summer >
Autumn, p<0.004), but not for the deeper, hypolimnion samples (p = 0.766). Therefore, it would appear
that the cruise differences discussed in the previous section, (i.e., the low concentrations in the autumn
1995 cruise) were mainly driven by concentration differences in the epilimnion rather than in the
hypolimnion. For particulate mercury there was a significant difference for both the epilimnion
(Summer>Autumn, p = 0.010) and hypolimnion (Summer>Autumn, p = 0.002) samples. Therefore, it
would appear that the cruise differences in particulate mercury were driven by differences in both
stratification levels of the lake.
5.1.4 Mercury Forms
Total and particulate phases of mercury were measured in Lake Michigan during the LMMB Study, and
mercury in the dissolved phase was calculated by subtraction. Calculated dissolved mercury
concentrations ranged from -0.12 ng/L to 0.75 ng/L. The calculated dissolved mercury concentrations for
six samples were negative, including three samples collected at the station in Lake Huron, and three
others from different stations collected during the August 1994 cruise. These negative values generally
reflect the low concentrations of total mercury in the samples overall, and reflect the analytical
uncertainties in both the total and particulate mercury concentrations for the samples. Dissolved mercury
concentrations differed significantly by cruise (p = 0.0077), but not by station (p = 0.1730), based on
ANOVA models (results log-transformed when possible prior to analysis). Tukey pairwise comparisons
between cruises revealed that the dissolved mercury concentration during March/April 1995 was
significantly greater than the concentration during August 1994. Descriptive statistics of calculated
dissolved mercury concentrations are presented in Table 5-4 below. The relative standard deviations
5-8
-------
Mercury in the Open-Lake Water Column
(RSDs) for dissolved mercury during each cruise are greater than the RSDs for particulate or total
mercury. This is because the dissolved mercury results were calculated, rather than measured, which
increases the variability of the results.
Table 5-4. Mean Dissolved Mercury Concentrations by Cruise
Sampling Cruise
June 1994
August 1994
Oct/Nov. 1994
March/April 1995
August 1995
Sept/Oct. 1995
N
12
23
18
23
23
21
Mean (ng/L)
0.19
0.13
0.21
0.28
0.24
0.19
Median (ng/L)
0.16
0.11
0.18
0.27
0.24
0.19
Range (ng/L)
0.078 to 0.42
-0.1 2 to 0.36
0.055 to 0.51
-0.076 to 0.75
-0.026 to 0.43
0.033 to 0.31
SD (ng/L)
0.097
0.13
0.14
0.17
0.13
0.070
RSD (%)
50
100
66
62
52
37
In addition, the ratio of particulate to total mercury was calculated for each sample. For five of the six
cruises, the mean ratios were below 0.50 (i.e., total mercury concentration more than double the
particulate mercury concentration), ranging from 0.24 to 0.46. The only cruise for which this was not true
was the August 1994 cruise, which had a mean ratio of 0.68. These differences between the August 1994
cruise and the rest of the data do not appear to be due to seasonality, as seen by the much lower ratios for
the August 1995 cruise (mean=0.36).
5.1.5 Other Factors Affecting Tributary Mercury Concentrations
In previous studies, it has been observed that mercury concentration is correlated positively with DOC
and negatively with pH (Watras et a/., 1995). Samples were analyzed for both DOC and pH during the
LMMB Study. However, the samples collected for DOC and pH were not the same samples in which
mercury was analyzed. While pH and DOC samples were collected at the same stations during the same
day that mercury samples were collected, the sample depths were generally not the same. Therefore,
correlations between mercury and DOC and pH could not be calculated. However, if mercury was
associated with either pH or DOC, then any spatial or temporal differences observed in mercury may also
be observed in the other parameters, either in the same direction (DOC) or opposite direction (pH).
To assess this possible relationship, ANOVA models for the effect of station and cruise were conducted
for both pH and DOC. While pH and DOC samples were collected at more stations and cruises than
those for which mercury samples were collected, these added samples were not included in the analyses.
Based on the ANOVA models, pH did not differ significantly among the 15 Lake Michigan stations for
which mercury samples were collected (p=0.941), but DOC concentrations did differ significantly among
stations (p=0.0017; results were log-transformed prior to analysis). Subsequent Tukey pairwise
comparisons showed that the DOC levels at Station 72M were significantly lower than for three other
stations (180, 280 and 340). However, this was not consistent with the mercury results, as the mean
mercury concentration for this station was slightly greater than the overall mean for both the particulate
and total fractions. ANOVA comparisons of pH and DOC among cruises showed that mean pH differed
significantly among cruises (p<0.0001), but mean DOC did not differ significantly (p = 0.0531; results
were log-transformed prior to analysis). Subsequent Tukey pairwise comparisons showed that pH during
the two August cruises was significantly greater than during the spring 1995 and two autumn cruises, and
that mean pH during the June 1994 cruise was significantly greater than during the autumn 1994 and
spring 1995 cruises. This shows some evidence of an inverse relationship, as total mercury peaked in the
spring, while pH was lowest.
5-9
-------
Results of the LMMB Study: Mercury 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 mercury
monitoring portion of the study are further described in Section 2.6 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, 2001b). A brief summary of the quality of the open-lake mercury 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 2.6, 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 5-5 provides a
summary of flags applied to the open-lake mercury 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 2.6. One
particulate mercury result was qualified as invalid due to a suspected leak in the sample, and was not used
in the analyses of open-lake mercury concentrations presented in this report.
Table 5-5. Summary of Routine Field Sample Flags Applied to Mercury in Open-lake Samples
Flag
INV, Invalid Result
DDL, Below Daily Detection Limit
EHT, Exceeded Holding Time
FDL, Failed Lab Duplicate
FFD, Failed Field Duplicate
FFR, Failed Field Blank
FPC, Failed Lab Performance Check
Number of QC samples
Particulate
—
—
—
45 lab duplicate
groups
18
13
Total
—
—
—
63 lab duplicate
groups
18
17
114
Percentage of Samples Flagged (%)
Particulate
0.8% (1)
8% (10)
0
8% (10)
7% (8)
0
19% (23)
Total
0
4% (5)
0
18% (22)
6% (7)
0
26% (33)
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.
Holding time flags were applied based on a criterion of 120 days between sampling and analysis. All data
met this criterion, with a maximum lag between sampling and analysis of 115 days.
5-10
-------
Mercury in the Open-Lake Water Column
The analytical sensitivity of field samples was assessed through analysis of daily detection limits. A
different limit was calculated for each day of analysis, with a maximum of 12 field samples associated
with a given daily detection limit. A "below daily detection limit" flag (DDL) was applied if a given field
sample concentration fell below its associated daily detection limit. The DDL flag was applied to 8% of
particulate mercury sample results and to 4% of total mercury sample results.
Field reagent blanks were analyzed to assess the potential for contamination of routine field samples. A
total of 24 valid field reagent blanks were analyzed, with concentrations ranging from -0.33 ng/L to 0.099
ng/L. In accordance with the researcher's data qualifying rules for field blanks, these blank results were
compared to a maximum of 0.10 ng/L. Because this level was never exceeded, no blanks or associated
samples were flagged with associated blank failure.
A total of 31 field duplicate samples and 133 laboratory duplicate samples were analyzed to assess
precision. The laboratory duplicate samples include both replicate analyses of field samples and field
duplicates, with up to 3 duplicates associated with a given field sample. From each cruise (except the
January 1995 cruise that visited only two sites), duplicate samples were collected at one to three stations.
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 (RPD) (or relative standard
deviation, RSD, where more than one laboratory duplicate was prepared for a given field sample)
between results for a sample and its duplicate was greater than 20%. This criterion was not met for 15
field duplicate pairs and for 32 laboratory duplicate groups. The maximum field duplicate RPD was 96%,
and the maximum laboratory duplicate RPD/RSD was 109%. While these RPDs were high, they were
based on low concentrations which were either below the daily detection limit or only slightly above.
Laboratory performance check samples were used to monitor analytical bias. In accordance with the
researcher's data qualifying rules for laboratory performance checks, samples were flagged for a failed
performance check (FPC) if the associated concentration was outside the concentration range of 0.80 to
1.2 ng (corresponding to 80% to 120% recovery). Based on application of this criterion, 23% of the field
samples were associated with a failed performance check. These flags were applied based on 28
performance check results exceeding 1.2 ng, with a maximum of 1.7 ng. Based on an analysis of
laboratory spikes, blank contamination, and other internal QC data, the QC coordinator did not qualify
any samples as high or low biased.
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 or RSD between the results for laboratory duplicate groups. Table 5-6 provides a summary of data
quality assessments for several of these attributes for open-lake data. The mean RPD for field duplicate
sample results was 28% for particulate mercury and 21% for total mercury, where both the sample and
duplicate results were greater than the daily detection limit. The mean RPD/RSDs for laboratory
duplicate samples were 15% and 17% for particulate and total mercury, respectively, where all results
were above the daily detection limit.
Analytical bias was evaluated by calculating the mean recovery of laboratory performance check samples
(LPC). Results indicated a slight positive bias, with a mean recovery of 110%. This bias applies to both
particulate and total mercury, as the LPC samples were not associated with a specific fraction.
Analytical sensitivity was evaluated by calculating the percentage of samples reported below the daily
detection limit. The mean daily detection limit was 0.063 ng/L, and ranged from 0.010 ng/L to 0.26 ng/L.
The majority of field samples were above the corresponding daily detection limit, with only 8% of
5-11
-------
Results of the LMMB Study: Mercury Data Report
participate mercury sample results and 4% of total mercury sample results falling below the given limit.
Results from these samples were not censored and were used as reported in the analysis of open-lake
mercury data presented in this report.
Table 5-6. Data Quality Assessment for Mercury in Open-lake Samples
Parameter
Number of Routine Samples Analyzed
Number of Field Duplicates Analyzed
System Precision, Mean Field Duplicate RPD (%
System Precision, Mean Field Duplicate RPD (%
Analytical Precision, Mean Lab Duplicate RPD (%
Analytical Precision, Mean Lab Duplicate RPD (%
), DDLa
), DDLa
Analytical Bias, Mean LPC (percent recovery)
Analytical Sensitivity, Samples reported as
-------
Mercury in the Open-Lake Water Column
total mercury concentration from these lakes was 1.48 ng/L for total mercury and 0.37 ng/L for
particulate mercury. Watras and Bloom (1992) also measured total mercury in the lower trophic levels of
an acidified basin and a reference basin in Little Rock Lake in Wisconsin in 1990. The mean total
mercury concentration in the reference basin was 0.0011 ng/g, or 1.1 ng/L. Mercury concentrations
similar to those measured in Lake Michigan were measured in three drainage lakes in Manitoba, with total
mercury concentrations ranging from 0.2 to 1.1 ng/L (Bloom and Effler, 1990, based on their personal
communication with J.W.M. Rudd).
The differences in mercury concentration between Lake Michigan and the lakes measured in previous
studies are not surprising, given the inherent differences between the lakes. In addition to the greater area
and depth of Lake Michigan, there are also differences in the chemistry of the lakes. For example, the
mean DOC and pH for the LMMB Study were 1.57 mg/L and 8.20, respectively. In contrast, mean DOC
concentrations and pH measured in 23 Wisconsin lakes were 6.62 mg/L and 6.17, respectively (Watras et
a/., 1995). Monson and Brezonik (1998) also reported DOC concentrations in 12 lakes in northeastern
Minnesota that were similar to those in the Wisconsin lakes, ranging from 4.5 to 10.2 mg/L, and similar
pH levels, ranging from 6.2 to 6.8. In addition, correlations between total mercury and various chemical
parameters were reported by Watras et al. (1995), with mercury having a strong positive correlation with
DOC (r2 = 0.93) and a strong negative correlation with pH (r2 = -0.51). However, these correlations do
not necessarily explain the mercury differences between Lake Michigan and the other two studies, as
correlations do not necessarily imply a causal relationship.
5.3.2 Comparison to Regulatory Limits
The freshwater water quality criterion established by EPA for human health protection is 50 ng/L for
mercury. This is more than an order of magnitude above the mean concentration measured in the lakes in
this study (0.33 ng/L). The mean concentration in this study is also less than the criteria for human health
(1.8 ng/L) and wildlife (1.3 ng/L) for the Great Lakes states.
5.3.3 Lateral Variation
Neither total mercury nor particulate mercury differed significantly between the 15 stations in Lake
Michigan at which samples were collected. This lends support to the theory that the primary source of
mercury to Lake Michigan is atmospheric (Sullivan and Mason, 1998), rather than riverine. A larger
level of riverine input would have been suggested if stations located closer to tributaries, especially
GB24M, had higher levels of mercury. The lack of spatial variability in concentrations in Lake Michigan
was also supported by the generally homogeneous levels of pH and DOC in Lake Michigan samples.
Only DOC exhibited significant differences between stations, as one northern Lake Michigan station had
a lower DOC concentration than three of the other stations.
5.3.4 Temporal Variation
Seasonal patterns in the total and particulate results were not clear, due to differences in the timing of the
cruises in the two years of the study. For total mercury, the mean concentration was greatest during the
fourth cruise (March/April 1995), and was significantly greater than for two other cruises. This cruise
was the only one that occurred during the spring, which suggests that the difference may be due to a
seasonal effect. Peak mercury concentrations in lakes during the spring were also observed by Monson
and Brezonik (1998) in 12 lakes in Minnesota and by Bloom and Effler (1990) in the Onondaga Lake in
New York. However, seasonal patterns during summer and autumn seemed to differ between the two
years in the LMMB Study. The September/October 1995 cruise had the lowest mean total mercury
concentration and was significantly lower than the other two cruises run in 1995. The October/November
1994 cruise did not exhibit a similar drop in concentration, but in fact, had a mean mercury concentration
5-13
-------
Results of the LMMB Study: Mercury Data Report
slightly greater than the other two 1994 cruises. These differences in patterns may have been partially
due to a calcite precipitation event occurring in 1995 (Sullivan and Mason, 1998). A drop in mercury
concentration has not generally been observed in prior studies. Monson and Brezonik (1998), in fact,
observed an increase in concentration in autumn compared to summer. The lower concentration in
autumn 1995 is also unexpected, based on the lower productivity level in 1994 (Sullivan and Mason,
1998).
Unlike total mercury, particulate mercury concentrations did not peak during the spring, instead they were
greatest in the June 1994 and August 1994 cruises. Similar to total mercury, the lowest concentrations
were observed during the September/October cruise. These results were not consistent with the
productivity level differences in the two years. It is worth noting that the August 1994 cruise had greater
daily detection limit values than the other cruises. Based on the low levels and variability of
concentrations measured in this study, any differences could have been strongly affected by slight levels
of contamination.
Seasonal differences were also affected by lake stratification. The four late summer and autumn cruises
included samples from multiple depths at most stations, representing the epilimnion and hypolimnion
levels of the lake. For total mercury, the concentrations in the epilimnion were significantly greater in the
summer compared to the autumn.
While mercury concentrations differed by cruise, DOC concentrations did not. This was unexpected,
given the strong positive correlations observed between DOC and total mercury in past studies. Mean pH
did differ significantly between cruises, with peak levels occurring during summer, and lower levels
occurring during the spring and autumn.
5.3.5 Vertical Variation
Total and particulate mercury concentrations were generally higher at depths closer to the surface, though
the effect of depth on concentration was not strong. Higher concentrations were expected near the
surface, because atmospheric deposition is considered to be the primary source of mercury input (Sullivan
and Mason, 1998). This effect of depth was greater during the late summer cruises, i.e., during peak
stratification conditions. The timing of the two autumn cruises differed, as the autumn 1994 cruise began
in mid-October, whereas the autumn 1995 cruise began in mid-September. However, there were not
enough pairs collected during these two cruises to assess the effect the timing difference had on
stratification of mercury.
Samples analyzed for total mercury were also collected at different depths and seasons from Lake
Onondaga in New York (Bloom and Effler, 1990). Similar to the current study, differences in
concentration between surface and hypolimnion depths (measured at 18 m) were greatest during the
summer. However, the direction of the difference was not the same, as the total mercury concentration
was greater in the hypolimnion. Similar to the current study, the difference between depths was minimal
in autumn. Total mercury concentrations in the hypolimnion also exceeded concentrations in the
epilimnion in Devil's Lake in Wisconsin (Herrin, etal., 1998). While epilimnion concentrations were
similar to those observed in this study (ranging from 0.10 to 1.0 ng/L), hypolimnion concentrations were
as high as 2.0 ng/L.
A possible difference between the relationship between depth and concentration in Lake Michigan and in
the other lakes is the greater depth of Lake Michigan. The maximum depths of Lake Onondaga and
Devil's Lake are 20.5 m and 14 m, respectively. The depths of the Lake Michigan stations from which
mercury samples were collected ranged from 27 m to 259 m. This difference could have played a role in
the relationship between depth and mercury concentration. For smaller, shallower lakes, the role of
5-14
-------
Mercury in the Open-Lake Water Column
sediment resuspension, compared to atmospheric input, will likely be greater than for larger, deeper lakes.
This increased role of sediment resuspension would result in a greater level of mercury in the
hypolimnion compared to the surface of the lake.
5.3.6 Mercury Fractions and Forms
For five of the six cruises, the majority of the total mercury was in the dissolved, rather than particulate,
phase. This result is similar to that observed by Watras et al. (1995) in 23 Wisconsin lakes. However,
Bloom and Effler (1990) found that the majority of total mercury was in the particulate fraction in Lake
Onondaga. In addition, Bloom and Effler observed that the percentage of total mercury in the particulate
fraction was greatest in the autumn. The authors hypothesized that this was due to the coagulation of
suspended matter after lake turnover.
5-15
-------
Chapter 6
Mercury in Surficial Sediments
6.1 Introduction
6.1.1 Background
Sediments can be a reservoir of trace contaminants. This is true for mercury, which is strongly attached
to particles in the water column and settles to the lake bottom along with the particles to become the
building blocks of the sediment. Surficial sediment particles enriched in mercury may be resuspended by
currents and waves and transported by currents to new locations. Eventually a particle with its associated
mercury is buried by particles deposited at a later date. Once the particle ceases to physically interact
with the water column, it becomes a part of the permanent sediment record.
From the standpoint of mass balance modeling, those particles that can be resuspended and transported
elsewhere within the lake are of interest. Contaminants associated with these particles are subject to
transport, creating a flux of materials from one location in the lake to another location in the lake. These
surficial sediment particles are also subject to interactions with the food chain, resulting in contaminant
exposures to organisms.
6.1.2 Study Objectives
With respect to mercury in sediments, the LMMB Study was designed to describe the horizontal
variability of mercury in the surficial sediments of Lake Michigan (Figure 6-1). By agreement among
principal investigators of the sediment, surficial sediments were defined as the surficial 1 cm of sediment.
Based upon experience, it was decided that this was the depth of sediment most likely available for
resuspension. To ascertain the character of resuspended sediments, sediment trap samples were also
collected at a number of locations in the lake (Figure 6-2). The locations were to be representative of
depositional and non-depositional locations. The specific objectives of the sediment mercury study were
to:
Document concentrations of mercury in surficial sediments,
• Describe the horizontal variation of mercury in surficial sediments,
• Estimate the flux of mercury to the surficial sediments,
• Describe the horizontal variation in mercury fluxes to the surficial sediments, and
Define the concentration of mercury and its time variation in resuspended sediments.
For Lake Michigan Mass Balance modeling and project objectives, the reader is referred to the modeling
and project plans (USEPA, 1995c and 1995d).
6-1
-------
Results of the LMMB Study: Mercury Data Report
Figure 6-1. Sampling Locations and Type of Sample Recovered between 1994 and 1996
A Naubinway
' Manistique
N
Scale
Okm 50km 100km 150km
Menorninee
Escanaba
Charievoix
Green Bay
Manitowoc
Sheboygan
Milwaukee
Racine
Waukegan
Mackinaw City
Lake Michigan Mass Balance
Project Sample Locations (1994-1996)
Legend
Muskegon
Grand Haven
Saugatuck
South Haven
Benton Harbor
Chicago
'Michigan City
No Sample Collected or Available
Box Core Sample
Ponar Grab Sample
Gary
6-2
-------
Mercury in Surficial Sediments
Figure 6-2. Sediment Trap Locations
A
N
Naubinway
Okm 50km 100km 150km
Mackinaw City
Lake Michigan Mass Balance
Sediment Trap Locations
Milwaukee
Waukegan
Grand Haven
Saugatuck
South Haven
Benton Harbor
Chicago
Michigan City
Gary
6-3
-------
Results of the LMMB Study: Mercury Data Report
6.2 Results
6.2.1 Mercury in Surficial Sediments
Surficial sediments were collected using the box corer and Ponar grab sampling techniques (Section
2.4.4). From July 18, 1994 to May 22, 1996, at least one surficial sediment sample was collected at each
of 118 stations, for a total of 126 samples (Table 6-1). (Note: The station numbers used for the sediment
sample collection effort do not correspond to the station identifiers used for the open-lake water samples
described in Chapter 5).
At six stations, both a Ponar grab sample and box core sample were collected; and at another single
station, one Ponar grab sample and two box core samples were collected. When more than one sample
was collected at a given station using both the Ponar grab and box core devices, the result from the box
core sample was used in the data analysis provided in this report because box-coring was the preferred
sampling method (see Section 2.4.4). The mean mercury concentration in Lake Michigan surficial
sediments was 0.078 mg/kg and the median value was similar (0.079 mg/kg) (Table 6-2).
Table 6-1. Concentrations of Mercury for each Lake Michigan Surficial
Sediment Station
Station Number*
1
2
4
6
7
8
9
10
11
13
14
15
16
17
18
19
20
21
22
24
25
26
27
28
29
30
Hg Concentration (mg/kg)
0.018
0.0074
0.033
0.006
0.072
0.074
0.092
0.021
0.040
0.10
0.012
0.10
0.096
0.14
0.12
0.088
0.020
0.24
0.13
0.15
0.13
0.11
0.15
0.032
0.17
0.036
LMMB Sample Number
sdlp
sd2p
sd4p
sd6p
942356
sd8p
942321
sd10p
sd11p
940532
sd14p
940608
sd16p
284
940659
940152
sd20p
940285
940590
940011
943564
951475
940143
sd28p
940102
sdSOp
6-4
-------
Mercury in Surficial Sediments
Table 6-1. Concentrations of Mercury for each Lake Michigan Surficial
Sediment Station
Station Number*
31
33
34
35
36
37
38
39
40
41
43
44
46
47
48
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
Hg Concentration (mg/kg)
0.11
0.12
0.19
0.011
0.14
0.10
0.11
0.22
0.003
0.15
0.011
0.050
0.14
0.032
0.14
0.016
0.018
0.019
0.15
0.12
0.15
0.012
0.084
0.13
0.036
0.006
0.17
0.11
0.14
0.10
0.14
0.17
0.049
0.006
0.013
0.16
0.008
0.012
0.022
0.016
0.086
LMMB Sample Number
950164
943117
951426
sd35p
941861
941802
sd38p
942545
sd40p
943106
sd43p
sd44p
942472
sd47p
942412
sd50p
sd51p
sd52p
942190
942959
941965
sd56p
sd57p
950648
sd59p
sd60p
951452
941233
951795
sd64p
942019
950451
sd67p
sd68p
sd69p
950461
sd71p
sd72p
sd73p
sd74p
sd75p
6-5
-------
Results of the LMMB Study: Mercury Data Report
Table 6-1. Concentrations of Mercury for each Lake Michigan Surficial
Sediment Station
Station Number*
76
77
78
79
80
81
82
83
84
85
86
87
88
89
91
92
95
96
97
98
99
100
101
102
103
104
106
107
108
109
110
111
112
113
114
115
116
117
118
120
121
Hg Concentration (mg/kg)
0.13
0.012
0.14
0.15
0.12
0.020
0.14
0.18
0.034
0.15
0.14
0.14
0.034
0.028
0.14
0.006
0.26
0.0045
0.15
0.012
0.13
0.007
0.13
0.029
0.12
0.008
0.011
0.12
0.16
0.004
0.10
0.005
0.14
0.15
0.034
0.023
0.006
0.026
0.007
0.11
0.011
LMMB Sample Number
sd76p
sd77p
952700
941148
940829
sd81p
940402
951686
sd84p
951283
951271
940399
sd88p
sd89p
sd91p
sd92p
sd95p
sd96p
951877
sd98p
950097
sdlOOp
951858
sd102p
943042
sd104p
sd106p
951823
950698
sd109p
952550
sd111p
952533
sd113p
sd114p
sd115p
sd116p
sd117p
sd118p
952569
sd121p
6-6
-------
Mercury in Surficial Sediments
Table 6-1. Concentrations of Mercury for each Lake Michigan Surficial
Sediment Station
Station Number*
122
123
124
125
126
127
128
129
130
131
Hg Concentration (mg/kg)
0.012
0.006
0.004
0.002
0.15
0.006
0.012
0.002
0.016
0.041
LMMB Sample Number
sd122p
sd123p
sd124p
sd125p
950880
sd127p
sd128p
sd129p
sd130p
sd131p
The station numbers used for the sediment sample collection effort do not correspond to the station
identifiers used for the open-lake water samples described in Chapter 5.
Table 6-2. Summary Statistics for Lake Michigan Surficial Sediment Mercury Concentrations
Descriptive Statistic
Mean Concentration (mg/kg)
Standard Deviation of Mean (mg/kg)
Median Concentration (mg/kg)
Minimum Concentration (mg/kg)
Maximum Concentration (mg/kg)
Number of Observations
Result
0.078
0.065
0.079
0.002
0.260
118
To visually display the results, all data were contoured using a linear variogram with no drift kriging. In
order to contour mercury concentrations in surficial sediments, it was necessary to assign a concentration
to the boundary of the lake as well as to locations from which sediment samples could not be recovered.
This boundary concentration was set at 0.0035 mg/kg, the average mercury concentration measured in
sand that was relatively free of silt- and clay-sized particles. For contouring fluxes, the net mercury flux
chosen for the boundary was 1.2 ng/cm2/y (Rossmann 1999, Rossmann and Edgington 2000), the
estimated regional atmospheric flux of mercury. Without other processes operative, this would be the
flux to locations along the shoreline. While these assumptions are oversimplifications (especially in areas
impacted by local high fluxes of mercury), the selected boundary conditions represent the most
reasonable values that can be obtained without additional data.
6.2.2 Mercury in Sediment Trap Samples
Resuspended sediments were collected using sediment traps (Section 2.3.4). A total of 65 samples from 7
different traps, representing 5 stations, were analyzed for total mercury. Sixteen trap samples from one
station having two traps could not be analyzed due to the use of mercury chloride as a preservative.
Results for each trap are contained in Table 6-3. (Note: The station numbers used for the sediment trap
sample collection effort do not correspond to the station identifiers used for either the sediment core
samples in Table 6-1 or the open-lake water samples described in Chapter 5). Approximately 50% of all
results had a mercury concentration <0.5 mg/kg. Mercury concentrations at 30 m water depth were
highest at Station 7 and were lowest at Station 5 (Table 6-4).
6-7
-------
Results of the LMMB Study: Mercury Data Report
Table 6-3. Concentrations of Mercury in Sediment Trap Samples
Station
Number*
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
8
8
8
8
8
8
8
8
8
8
8
8
Trap
Number
5
5
5
5
5
5
5
5
5
5
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
7
7
7
7
7
7
7
7
7
7
7
7
Sequence
Number
2
9
10
11
12
15
16
17
18
23
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
23
1
2
3
4
5
9
10
11
12
13
14
15
Trap Water
Depth (m)
30
30
30
30
30
30
30
30
30
30
155
155
155
155
155
155
155
155
155
155
155
155
155
155
155
155
155
155
155
30
30
30
30
30
30
30
30
30
30
30
30
Sample
Number
ST314
ST321
ST322
ST323
ST324
ST327
ST328
ST329
ST330
ST335
ST337
ST338
ST339
ST340
ST341
ST342
ST343
ST344
ST345
ST346
ST347
ST348
ST349
ST350
ST351
ST352
ST353
ST354
ST358
ST359
ST360
ST361
ST362
ST363
ST367
ST368
ST369
ST370
ST371
ST372
ST373
Mercury Concentration
(mg/kg)
27
6.1
4.2
4.8
2.9
3.9
2.2
5.8
6.9
11.
1.6
3.0
2.4
2.0
0.95
3.0
1.8
1.1
0.47
0.47
0.38
0.42
0.43
0.40
0.33
0.44
0.66
0.51
1.4
1.1
0.30
0.37
0.36
0.21
0.66
0.91
0.55
0.32
0.43
1.2
2.5
6-8
-------
Mercury in Surficial Sediments
Table 6-3. Concentrations of Mercury in Sediment Trap Samples
Station
Number*
8
8
8
8
8
2
2
8
8
8
8
1
1
5
5
5
5
5
5
5
5
5
5
5
Trap
Number
7
7
7
7
7
9
9
6
6
6
6
8
8
3
3
3
3
3
3
3
3
3
3
3
Sequence
Number
16
18
20
21
23
1
7
1
2
3
4
1
2
3
4
10
11
12
13
14
15
16
17
18
Trap Water
Depth (m)
30
30
30
30
30
77
77
51
51
51
51
45
45
30
30
30
30
30
30
30
30
30
30
30
Sample
Number
ST374
ST376
ST378
ST379
ST381
ST382
ST385
ST388
ST389
ST390
ST391
ST395
ST396
ST469
ST470
ST476
ST477
ST478
ST479
ST480
ST481
ST482
ST483
ST484
Mercury Concentration
(mg/kg)
0.44
1.2
0.64
2.5
1.2
0.30
0.71
0.39
0.61
0.42
0.27
0.31
0.44
0.79
0.42
0.53
0.24
0.27
0.22
0.22
0.25
0.55
0.84
0.43
* The station numbers used for the sediment trap sample collection effort do not correspond to the station identifiers
used for either the sediment core samples in Table 6-1 or the open-lake water samples described in Chapter 5.
Table 6-4. Mercury Summary Statistics for each Station at each Depth for Sediment Trap Samples
Station
1
2
5
7
7
8
8
Depth (m)
45
77
30
30
155
30
51
Number of
Samples
2
2
11
10
19
17
4
Mean
(mg/kg)
0.37
0.51
0.43
7.5
1.1
0.87
0.42
Standard
Deviation (mg/kg)
NA
NA
0.23
7.4
0.90
0.69
0.14
Median
(mg/kg)
0.37
0.51
0.42
5.3
0.66
0.64
0.40
Minimum
(mg/kg)
0.31
0.30
0.22
2.2
0.33
0.21
0.27
Maximum
(mg/kg)
0.44
0.71
0.84
27
3.0
2.5
0.61
It should be noted that the mass of sediment collected from the traps was often too small to complete all
analyses targeted in the LMMB Study. Samples were collected and analyzed for mercury only in those
cases where the amount of material available in trap samples was sufficient, and this generally
6-9
-------
Results of the LMMB Study: Mercury Data Report
corresponded with periods of high sediment fluxes. Because of these shortcomings in the data set at this
time, the authors do not wish to interpret the sediment trap any further.
6.2.3 Moisture Content of Sediment Samples Collected by Ponar
The moisture content ([wet weight - dry weight]/[dry weight] x 100) was measured on each Ponar sample
received (Table 6-5). Ponar samples were collected not only from those regions with bottoms too hard to
core, but also from areas inaccessible to the ship. Most of these areas were very sandy. A few were very
silty sands.
Table 6-5. Moisture Content of Samples Collected by Ponar
Station Number*
1
2
4
6
8
9
10
11
13
14
16
19
20
28
30
35
36
38
40
43
44
47
50
50
51
52
56
57
59
60
64
67
68
69
LMMB Sample Number
sdlp
sd2p
sd4p
sd6p
sd8p
sd9p
sd10p
sd11p
sd13p
sd14p
sd16p
sd19p
sd20p
sd28p
sdSOp
sd35p
sd36p
sd38p
sd40p
sd43p
sd44p
sd47p
sd50p
sd50p
sd51p
sd52p
sd56p
sd57p
sd59p
sd60p
sd64p
sd67p
sd68p
sd69p
Moisture Content (%)
25
22
24
22
70
67
31
53
76
32
43
88
30
44
53
26
75
65
22
24
52
40
28
23
42
44
32
52
52
23
74
52
27
31
6-10
-------
Mercury in Surficial Sediments
Table 6-5. Moisture Content of Samples Collected by Ponar
Station Number*
71
72
72
73
74
75
76
77
81
84
88
89
91
92
95
96
98
98
100
102
104
106
109
111
111
113
114
115
116
117
118
121
122
123
124
125
127
128
129
130
131
LMMB Sample Number
sd71p
sd72p
sd72p
sd73p
sd74p
sd75p
sd76p
sd77p
sd81p
sd84p
sd88p
sd89p
sd91p
sd92p
sd95p
sd96p
sd98p
sd98p
sdlOOp
sd102p
sd104p
sd106p
sd109p
sd111p
sd111p
sd113p
sd114p
sd115p
sd116p
sd117p
sd118p
sd121p
sd122p
sd123p
sd124p
sd125p
sd127p
sd128p
sd129p
sd130p
sd131p
Moisture Content (%)
26
23
22
47
42
64
74
28
23
54
27
32
78
20
63
23
24
27
22
36
25
27
22
25
23
90
26
37
30
69
22
20
27
24
21
24
38
34
18
33
68
The station numbers used for the Ponar sample collection effort do not correspond to the station identifiers
used for the open-lake water samples described in Chapter 5.
6-11
-------
Results of the LMMB Study: Mercury Data Report
Because these data are only for Ponar samples collected from primarily sandy areas, the mean and median
moisture contents are relatively low compared to silt- and clay-rich sediments (Table 6-6). The minimum
of 18% moisture represents a fairly pure sand, while the maximum of 90% moisture represents a very silt-
or clay-rich sand that might even be classified as a silt or clay.
Table 6-6. Summary Statistics for Moisture Content Analyses
of Samples Collected by Ponar
Descriptive Statistic
Mean Moisture Content (%)
Standard Deviation of Moisture Content (%)
Median Moisture Content (%)
Minimum Moisture Content (%)
Maximum Moisture Content (%)
Number of Samples
Result
39
19
31
18
90
75
6.2.4 Mercury Fluxes to Sediments
Because sedimentation rates have been measured at all box-core stations (Edgington and Robbins 1997b,
Robbins and Edgington 1997), mercury fluxes were calculated for each site (Table 6-7). The flux is equal
to the Pb-210 sedimentation rate times the mercury concentration. At locations where box cores could
not be collected, the net sedimentation rate is essentially zero; hence, the net flux is also zero. With a
mean of 7.2 ng/cm2/y, mercury fluxes ranged between 0.85 and 32 ng/cm2/y (Table 6-8).
Table 6-7. Net Mercury Flux to Lake Michigan Surface Sediments
Station Number*
7
9
9
13
15
17
18
19
21
21
22
24
25
26
27
29
31
33
33
34
36
Total Hg Flux (ng/cm2/y)
6.4
19
2.3
3.8
23
1.4
2.2
8.4
2.9
3.9
17
10
4.5
2.2
4.7
6.0
3.2
2.0
3.8
5.0
8.1
Station Number*
54
55
58
61
62
63
65
66
70
78
79
80
82
83
85
86
87
97
99
101
103
Total Hg Flux (ng/cm2/y)
0.94
31
4.8
18
1.1
4.0
1.6
6.2
14
2.8
5.8
2.8
13
6.4
3.8
4.4
16
3.2
5.2
7.5
3.3
6-12
-------
Mercury in Surficial Sediments
Table 6-7. Net Mercury Flux to Lake Michigan Surface Sediments
Station Number*
37
39
41
46
48
53
Total Hg Flux (ng/cm2/y)
2.7
8.5
32
6.5
0.85
1.9
Station Number*
107
108
110
112
120
126
Total Hg Flux (ng/cm2/y)
5.4
8.0
2.6
9.4
8.0
7.1
* The station numbers used for the sediment sample collection effort do not correspond to the station identifiers used
for the open-lake water samples described in Chapter 5.
Table 6-8. Summary Statistics for Net Mercury Fluxes to Lake Michigan
Surface Sediments in Depositional Basins
Descriptive Statistic
Mean Net Mercury Flux (ng/cm2/y)
Standard Deviation of Mean Net Flux (ng/cm2/y)
Median Net Mercury Flux (ng/cm2/y)
Minimum Net Mercury Flux (ng/cm2/y)
Maximum Net Mercury Flux (ng/cm2/y)
Number of Samples
Result
7.2
6.9
4.9
0.85
32
54
6.2.5 Horizontal Variation of Mercury and Mercury Fluxes
Mercury concentrations and their resulting fluxes varied with location in the lake. Mercury
concentrations were higher along the eastern side of the lake than its western side (Figure 6-3). Mercury
concentration contours were coincident with those for the bathymetry of the lake (Figure 6-4). Unlike the
concentration contours, those for mercury flux were not coincident with the lake bathymetry (Figure 6-5).
Regions of highest flux were compressed along the eastern side of the lake.
6-13
-------
Results of the LMMB Study: Mercury Data Report
Figure 6-3. Mercury Concentrations (mg/kg) in Lake Michigan Surficial Sediments (1994-1996)
Naubinway
f
N
Manistique
Escanaba
Scale
: u
Okm 50km 100km 150km
Menominee
Charlevoix
Mackinaw City
Milwaukee
Waukegan
Chicago
Michigan City
Gary
Lake Michigan Mass Balance
Manistee Pr°JeCt 1994-1 "6 Surficial
Sediment (0-1 cm) Mercury
Concentrations (ng/g)
Legend
220 ng/g
200 ng/g
180 ng/g
160 ng/g
140 ng/g
120 ng/g
100 ng/g
80 ng/g
60 ng/g
40 ng/g
20 ng/g
Ong/g
-20 ng/g
Muskegon
Grand Haven
Saugatuck
South Haven
Benton Harbor
6-14
-------
Mercury in Surficial Sediments
Figure 6-4. Lake Michigan Bathymetry with Depositional Basin Locations
Naubinway
f
N
Manistique
Escanaba
Scale
Okm 50km 100km 150km
Menominee
Charievoix
Green Bay
Manitowoc
Sheboygan
Milwaukee
Racine
Waukegan f
Chicago
Mackinaw City
Lake Michigan Basins and
Water Depths (m) Data from
Cahill (1975) and This Study
Muskegon
Grand Haven
Saugatuck
South Haven
Benton Harbor
Michigan City
Legend
1280m
260m
240m
220m
200m
!l80m
160m
140m
120m
100m
80 m
60 m
40 m
20m
0 m
Gary
6-15
-------
Results of the LMMB Study: Mercury Data Report
Figure 6-5. Mercury Fluxes (ng/cm2/y) to Lake Michigan Surficial Sediments (1994-1996)
Naubinway
Mackinaw City
Charievoix
Scale
i
Okm 50km 100km 150km
Lake Michigan Mass Balance
Project 1994-1996 Surficial
Sediment (0-1 cm) Mercury
Fluxes (ng/cm2/y)
Legend
Milwaukee
Racine
Waukegan
p
Chicago
Michigan City
Gary
36 ng/cm2/y
33 ng/cm2/y
30 ng/cm2/y
27 ng/cm2/y
24 ng/cm2/y
21 ng/cm2/y
1 8 ng/cm2/y
1 5 ng/cm2/y
12 ng/cm2/y
9 ng/cm2/y
6 ng/cm2/y
3 ng/cm2/y
0 ng/cm2/y
-3 ng/cm2/y
6-16
-------
Mercury in Surficial Sediments
6.3 Quality Assurance
As described in Section 1.5.5, the LMMB quality assurance program prescribed minimum standards to
which all organizations collecting data were required to adhere. The quality activities implemented for
the mercury monitoring portion of the study are further described in Section 2.6 and included use of
SOPs, training of laboratory and field personnel, and establishment of measurement quality objectives
(MQO) 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 mercury data is provided below.
As discussed in Section 6.1.2, the mass of resuspended sediment collected from the sediment traps was
often too small to complete all analyses targeted in the LMMB Study. Because trap samples were
collected and analyzed for mercury only during relatively high sediment flux periods, the mercury
concentrations measured in sediment traps reflect those in resuspended sediments during these higher flux
periods.
For some field and quality control (QC) samples, multiple analyses were conducted either on the field
sample or the sample extract. Sample results were reported as average values of replicate results when
available and are identified as average values in the Great Lakes Environmental Database (GLENDA)
database. A standard reference material (SRM) from the National Institute of Standards and Technology
(NIST) was included with sample batches to monitor performance of the analytical system. The Buffalo
River Sediment, SRM 2704 (no longer available), has a certified value of 1.47 mg/kg. SRM samples
were prepared and analyzed using the same extraction procedure as the field samples and were included
with every group of samples extracted. Laboratory reagent blanks also were included with sample
batches and were prepared and analyzed using the same extraction procedure as the samples. The mean
mercury concentration measured in the blanks in a given batch was used to assess blank contamination in
each sample. More than 80% of blanks were below the method detection limit (MDL).
Sediment samples were extracted using two different procedures. Most surficial sediments were extracted
using a Leeman Labs, Inc. automated mercury system (Leeman Labs, Inc., 1993). All sediment trap
samples and a few surficial sediment samples were extracted using a microwave digestion system
(Uscinowicz and Rossmann, 1997). The Leeman automated extraction uses 50% aqua regia and
potassium permanganate solutions and is more vigorous than the microwave extraction, which uses a 10%
nitric acid solution. Mean recoveries of mercury in the NIST standard reference material samples were
97% for the automated digestion and 90% for the microwave digestion. This may be due, in part, to the
smaller sediment sample mass that is used in the microwave digestion procedure compared to the
automated digestion procedure, which requires as much as ten times the sample mass used in the
microwave procedure. Also, the concentration of acid used in the extraction is greater for the automated
extraction. Regardless of the extraction method, mercury concentrations measured in the SRM samples
were within acceptance criteria for 100% of the sample analyses.
As discussed in Section 2.6, all data were verified by comparing the field and QC sample results produced
by each principal investigator (PI) with their MQOs and with overall LMMB Study objectives. Field
sample results were flagged when pertinent QC sample results did not meet 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. Table
6-9 provides a summary of flags applied to the sediment mercury data. The summary provided below
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 2.6.
6-17
-------
Results of the LMMB Study: Mercury Data Report
Table 6-9. Summary of Data Verification Flags Applied to Routine Field Sample Results for Sediment
Mercury
Flag
EHT, Exceeded Holding Time
FBS, Failed Blank Sample
FDL, Failed Lab Duplicate
FFD, Failed Field Duplicate
FSR, Failed Standard Reference Material
GTL, Greater than Operating Range
SCX, Suspected Contamination
UDL, Below Sample Specific Detection Limit
Number of QC Samples
—
41 lab reagent blank samples
34 lab duplicate pairs
4 field duplicate pairs
40 SRM samples
—
—
—
Percentage of Samples Flagged
0
0
2% (4)
1%(2)
0
3% (5)
2% (3)
<1%(1)
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.
All of the sediment samples were analyzed for mercury within the required holding time. Of the 41
laboratory reagent blank samples (LRBs) prepared and analyzed, none of the sample results exceeded the
MQO and therefore, none of the routine field samples were flagged for a failed blank sample (FBS).
Only 2% of the field sample results had associated laboratory duplicates with results above the maximum
RPD/RSD of 15%, the acceptance criteria. The maximum RPD/RSD for these sample groups was 48%.
Three percent of samples contained mercury concentrations that were greater than the operating range of
the analytical system. These results are flagged in the database and should be considered estimated
values. Two percent of the field samples were flagged for suspected contamination, based on laboratory
notation that the samples were potentially contaminated during sample preparation and analysis in the
laboratory. The laboratory notation is included in the database for these sample results in a comment field
(exception to method text).
MQOs were defined in terms of six attributes: sensitivity, precision, accuracy, representativeness,
completeness, and comparability. GLNPO derived data quality assessments based on three 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 6-10 provides a summary of data quality
assessments for several of these attributes for the sediment mercury study data.
Table 6-10. Data Quality Assessment for Mercury in Sediment Samples
Parameter
Number of Routine Samples Analyzed
System Precision
Mean Field Duplicate RPD (%), Samples >MDLS
Analytical Precision
Mean Lab Duplicate RPD (%), Samples >MDLS
Analytical Bias, Mean SRM3(%)
Analytical Sensitivity, Samples reported as
-------
Mercury in Surficial Sediments
System precision, estimated as the mean relative percent difference for field duplicates, was 38%.
However, because only four field duplicates were collected and analyzed, this estimate may not
accurately reflect the variability associated with sampling and analytical activities. Analytical precision,
estimated as the mean relative percent deviation for laboratory duplicates, was much lower, at 8.5%,
suggesting that either the small number of field duplicates did not accurately reflect the variability
associated with sampling and analytical activities, or the variability associated with sampling is much
greater than that associated with the analytical activities. This latter possibility is not unexpected for
sediment sampling. Analytical bias, estimated as the mean recovery of standard reference materials, was
92%, which indicates a slight low bias in the analytical results. More than 99% of samples contained
mercury concentrations above the detection limit.
6.4 Data Interpretation
Lake Michigan surficial sediments have elevated mercury concentrations compared to pre-settlement
concentrations. Fluxes of mercury to the lake from atmospheric, tributary, and shoreline sources are
redistributed within the lake by wave action and current transport. This leads to a definitive distribution
pattern of mercury concentrations in Lake Michigan sediments and fluxes to those sediments. Only the
mercury results for surficial sediment are discussed in this chapter. A discussion of moisture content is
not included in this chapter.
6.4.1 Comparison to Other Great Lakes Sediments
Excluding Green Bay, Lake Michigan surficial sediments have relatively low mercury concentrations
(Table 6-11).
Table 6-11. Comparison of Lake Michigan Surficial Sediment Mercury Concentrations to those at other
Locations in the Great Lakes Basin
Location and
Years
Green Bay
1987-1990
Superior 1983
North Channel
1973
Georgian Bay
1973
Huron 1969
St. Clair1970
Erie 1971
Ontario 1968
Michigan
1994-1996
N
74
31
55
117
163
55
243
248
118
Mean
(ng/g)
360
180
150
260
220
630
610
650
78
Standard
Deviation (ng/g)
270
180
230
1000
160
630
700
510
65
Median
(ng/g)
280
140
NA
NA
NA
NA
NA
NA
73
Minimum
(ng/g)
6
27
8
12
54
70
13
32
2
Maximum
(ng/g)
1100
960
1100
9500
800
2600
7500
2100
260
Reference and Surficial
Interval Sampled
Rossmann and Edgington
(2000) 0-1 cm
Rossmann (1999)
0-2 cm
Thomas (1974)
0-3 cm
Thomas (1974)
0-3 cm
Thomas (1974)
0-3 cm
Thomas (1974)
0-3 cm
Thomas (1974)
0-3 cm
Thomas (1974)
0-3 cm
Rossmann (this study)
0-1 cm
NA = Not applicable
6-19
-------
Results of the LMMB Study: Mercury Data Report
In the main basin of Lake Michigan, mean mercury concentration is nearly one-half of those found in all
the other Great Lakes, making the lake relatively uncontaminated with mercury. However, it should be
noted that Lake Michigan sediments are being compared to much earlier results for other locations.
Contamination of sediments was historically higher than at present. It should also be noted that the
surficial sediment intervals compared are different in thickness and represent different periods of time that
are integrated to produce the mercury concentration reported for the homogenized layer of sediment. Due
to the difference in sampling year and sediment thickness, cautions should be used when comparing
LMMB data to these other studies. Recent data having similar time intervals represented by the top
interval are insufficient to be representative of Lakes Superior, Huron, Erie, and St. Clair sediments.
Note should be made of the fact that Green Bay, a bay of Lake Michigan, has sediments that are
contaminated with mercury relative to other Great Lakes locations. The contamination of these sediments
has been attributed to historical industrial practices in the Fox River drainage basin (Rossmann and
Edgington, 2000).
6.4.2 Comparison to Historical Lake Michigan Concentrations
For Lake Michigan, several historic data sets exist for mercury in surficial sediments (Table 6-12).
Kennedy et al. (1971) reported on mercury concentrations in the 0-1 and 0-5 cm intervals of surficial
sediments collected during 1969 and perhaps 1970. Samples were collected from the southern basin of
Lake Michigan from 31 sites (Figure 6-6). A much more comprehensive collection was made in 1975
and reported in Cahill (1981). The surficial 3 cm of sediment were collected from 254 locations from all
basins of the lake (Figure 6-7). Mercury results are available for one of three sediment cores collected
from southern Lake Michigan in 1981 (Figure 6-8). Results for the LM-81-HS core are reported by
Pirrone et al. (1998). Additional details for that core are reported here.
Table 6-12. Comparison of Current Lake Michigan Results to Historical Data
Years
Collected
1969-1970?
1975
1981
1994-1996
N
31
254
1
118
Mean
(ng/g)
150
110
200
78
Standard
Deviation
(ng/g)
100
110
—
65
Median
(ng/g)
120
60
—
73
Minimum
(ng/g)
30
20
—
2
Maximum
(ng/g)
380
670
—
260
Reference and
Surficial Interval Sampled
Kennedy et al. (1971)
surficial 0-1 through 0-5 cm
Cahill (1981)
0-3 cm
Pirrone et al. (1998) 0.5 cm
Rossmann (this study)
0-1 cm
6-20
-------
Mercury in Surficial Sediments
Figure 6-6. Station Locations for the 1969-1970 Kennedy et a/. Mercury Results
Naubinway
/-^~v-^_
Manistique
A
N
Escanaba
Scale
^HZI
Okm 50km 100km 150km
Menominee
Chartevoix
Green Bay
Manitowoc
Sheboygan
Milwaukee
Racine
Waukegan
Mackinaw City
1969-1970 Lake Michigan
Sediment Stations Reported
by Kennedy etal. (1971)
South Haven
Benton Harbor
Chicago
Michigan City
Gary
6-21
-------
Results of the LMMB Study: Mercury Data Report
Figure 6-7. Station Locations for the 1975 Cahill Mercury Results
Naubinway
Manistique / U49V49
• •
S48T48 U48V48W48X48
N
Escanaba.
N47 P47
• •
M46 046
^
V
U46 W46
Scale
Okm 50km 100km 150km
Menominee'
'Mackinaw City
t45U45V45 '
'• • •
t) L44M44 O44 Q44R44§WT44U44V44
0 K43L43 N43 P43 ^43343743 U43V43V
K42 M42 O42 CW2R42S42T42U42
J41 L41 N41 P41Q41R41S41T41 / C/7a/~/eVO/X
rt sl&r 140 K40 M40 O40 Q40
^ /NJ..CVL m * m m — m
T40
M32
H39 J39 L39 N39O39^ ( Q39
138 • M38 <23l3P38Q38|'
H37 J37 L37 N37
• • • •
G36 136 M36 03f
/F35 H35 J35 K|§5
G34 134 K34 M34
• • • •
E32 H33 J33 L33
Green Bay I 032 132 K32
F31 H31 J31 L31
• • • •
E30 G30 130 K30 M30
• * • •
0^9 F29 H29 J29 L29
Manltowoc^ E|8 ^ "^ K28
027 F27 H27 J27
fC26 E26 G26 126 K26
• • • •
025 F25 H25 J2S L2Sj
)C24 * * G24 ' 124 " K24 */penfwafer
1975 Lake Michigan
Sediment Stations
Reported by Cahill (1981)
Sheboygan
023
H23 J23
C22 E22 G22 I22 K22
• * • • I
021 F21 H21 J21 L2\
1 • • • • 4
C20 E20 G20 I20 K20
• 019* • • •
B19 • F19 H19 J19 L19
* c^s E^S * G«J * iis * M*B M^Muskegon
B17 017 F17 H17 ju L17
Milwaukee ( C16 E^6 Gi6 'Jf K16 M1S \Grand Haven
B15 015 F15 H15 U15 N1
C14 E14 G14 114 K14 M14
Racine ( * Di3 * Fi3 " Hi3 * Ji3 * Li3 " Ni
U2 E^2 0^2 112 Kiz M12 Saugatuck
011 F11 H11 J11 L11
C10 E10 G10 110 K10 M10
• • • • •
I 89 D9 F9 H9 J9 L9
Waukeganl c& EB G& ie KS MS/South Haven
07 F7 H7 J7 L7
Chicago
C6 E6 G6 16 K6
• • • •
D5 F5 H5 J5
• • • •
E4 G4 14 K4/
03 F3 H3 J3
• • •
E2 G2
/Benton Harbor
FI ^^Michigan City
Gary
6-22
-------
Mercury in Surficial Sediments
Figure 6-8. Station Locations for 1981 Sediment Cores
Manistique
Naubinway
N
Escanaba
Scale
^^Zl
Okm 50km 100km 150km
Charlevoix
Menominee i / <£"
Green Bay
Manitowoc
Sheboygan
Milwaukee
Racine
Waukegan
Chicago
1981 Lake Michigan
Sediment Station Locations
Mackinaw City
Muskegon
Grand Haven
Saugatuck
South Haven
Benton Harbor
Michigan City
Gary
6-23
-------
Results of the LMMB Study: Mercury Data Report
A general comparison of these results can be potentially misleading. It appears that mercury may be
decreasing in surficial sediments between 1969 and 1996. The problem is derived from using grab
samples that penetrated a variety of sediment depths and sampled a variety of sediment types. Samples
reported by Cahill (1981) represented a homogenate of the surficial 3 cm. It is possible that collection to
a depth of 3 cm penetrated to older sediments generally known to be more contaminated with mercury,
skewing the results toward a higher concentration (Rossmann and Edgington, 2000). Also, samples
collected in 1975 and 1994-1996 were representative of a variety of sediment types. Samples collected
from sandy areas will skew the results toward lower mercury concentrations. Mercury is associated with
fine-grained sediments. Depending upon the station distribution for a data set, results may be biased to
various regions of the lake basin. Thus, a direct comparison of data set that have different station
distributions and different depths of surficial sediment can lead to incorrect conclusions. To avoid these
problems, it is best to compare only those sediments collected at the same sampling interval from the
same station in a depositional basin.
There is only one location for which a direct comparison with historical data may be made. LMMB
Station 15 (Figure 6-1) is coincident with Station K8 reported by Cahill (1975). It is within 5 km of
Station 105 reported by Kennedy etal. (1971) and Station LM-81-HS reported by Pirrone etal. (1998).
Two comparisons can be made. The first is a comparison for the surficial sample interval of 0 - 3 cm
sediment depth, and the second is a comparison for the interval of 0 - 1 cm sediment depth. For the 0 - 3
cm surficial sediment depth interval, there is a distinct decrease in mercury concentration between 1969
and 1975 which continues through 1981 (Table 6-13).
Table 6-13. Comparison of Lake Michigan Results at Station 15 to Historical Results for the 0 - 3 cm
Surficial Sediment Interval
Year Collected
1969
1975
1981
Mercury Concentration (ng/g)
300
240
180
Reference and Surficial Interval Sampled
Kennedy ef a/. (1971)
Cahill (1975)
Pirrone ef a/. (1998)
This decrease is also evident for the 0 - 1 cm surficial sediment depth interval comparison (Table 6-14).
Thus, there has been a decrease in mercury concentrations in surficial sediments between 1969 and 1994.
The decrease between 1969 and 1975 was at the rate of 4.3 ng/cm2/y and that between 1975 and 1981 was
10 ng/cm2/y. The resolution for these 0-3 cm of surficial sediments was roughly 5 years. A more
realistic recent rate of mercury decline is derived from the 0 - 1 cm results, where the resolution is less
than one year. The most recent rate of decrease between 1981 and 1994 was 3.8 ng/cm2/y.
Table 6-14. Comparison of Lake Michigan Results at Station 15 to Historical Results for the 0 - 1 cm
Surficial Sediment Interval
Year Collected
1981
1994
Mercury Concentration (ng/g)
150
100
Reference and Surficial Interval Sampled
Pirroneefa/. (1998)
This study
The recent decrease in mercury concentrations in surficial sediments is corroborated by results for the
1981 core that are reported by Pirrone et al. (1998) and presented in a slightly different manner here
(Figure 6-9). At this station, pre-1800 background mercury concentrations ranged between 8 and 14 ng/g.
Between 1930 and 1950, peak mercury concentrations as high as 460 ng/g were reached. By the late
1950s, mercury reached its maximum concentration ranging between 300 and 450 ng/g. After 1970,
6-24
-------
Mercury in Surficial Sediments
mercury concentrations began to decrease. The decrease that was noted in the 1981 core has continued
through 1994 at this location in the lake.
Figure 6-9. Vertical Variation of Mercury in Core LM-81-HS
LM-81-HS
Total Mercury
500
400
300
•5?
en
200
100
1982 1980 1976 1971 1966 1961 1954 1945 1932 1919 1902 1873 1839 1804 1763
Mid-Interval Pb-210 Date
Each bar represents one interval of the core.
6.4.3 Comparison to Historical Lake Michigan Horizontal Variations
Horizontal variations of mercury in surficial sediments can be compared to two previously published data
sets. As discussed in the previous section, absolute concentrations cannot be compared; however,
patterns of variation can be discussed. The Kennedy et al. (1971) data set for the period of 1969-1970
covers southern Lake Michigan (Figure 6-6). The distribution pattern (Figure 6-10) is similar to that for
1994-1996 (Figure 6-3) and mimics the bottom topography (Figure 6-4). The same is true for the Cahill
(1981) data set collected in 1975 (Figure 6-11). The 1975 data set is very extensive and includes stations
from the entire lake (Figure 6-7). These two data sets and the current one produce mercury distribution
patterns that are similar and conform to the lake's bathymetry, sediment and mercury sources, and water
circulation pattern.
6-25
-------
Results of the LMMB Study: Mercury Data Report
Figure 6-10. Mercury Concentrations (mg/kg) in 1969-1970 Lake Michigan Surficial Sediments
Naubjnway
Manistique
Scale
Okm 50km 100km 150km
Chartevoix
Green Bay
Mackinaw City
Sheboygan
Milwaukee
Racine
Waukegant V «>
Chicago
1969-1970 Lake Michigan
Surficial Sediment (0-1 to
0-5 cm) Mercury Concentrations
(ng/g) as Reported by
Kennedy et al. (1971)
Legend
Muskegon
Grand Haven
Q^USaugatuck
South Haven
Benton Harbor
Michigan City
Gary
400 ng/g
360 ng/g
320 ng/g
280 ng/g
240 ng/g
200 ng/g
160 ng/g
120 ng/g
80 ng/g
40 ng/g
0 ng/g
6-26
-------
Mercury in Surficial Sediments
Figure 6-11. Mercury Concentrations (mg/kg) in 1975 Lake Michigan Surficial Sediments
Naubinway
f
N
Manistique
Escanaba
Scale
Okm 50km 100km 150km
Menominee
Chartevoix
Green Bay
Manitowoc
Milwaukee (,
Racine
Waukegan
Mackinaw City
Sheboygan) {//%%
^Muskagon
Grand Haven
'! Saugatuck
South Haven
'//Benton Harbor
Chicago
Michigan City
Gary
1975 Lake Michigan Surficial
Sediment (0-1 cm) Mercury
Concentrations (ng/g) as
Reported by Cahill (1981)
Legend
1460 ng/g
220 ng/g
200 ng/g
180 ng/g
160 ng/g
140 ng/g
120 ng/g
100 ng/g
80 ng/g
60 ng/g
40 ng/g
20 ng/g
Ong/g
6-27
-------
Results of the LMMB Study: Mercury Data Report
6.4.4 Regional Lake Michigan Comparisons
Because bathymetry and currents control observed mercury distributions in surficial sediments, it is
important to compare regional mercury concentrations for only the depositional basins of the lake (Table
6-15). When this is done, it becomes apparent that the mean mercury concentration varies very little
between basins. Mean mercury concentrations range between 120 and 160 ng/g, within the observed
standard deviations (Table 6-15). Two anomalies are noteworthy. First, the minimum concentration in
the Southern Basin is substantially lower than those for the other basins. The reason for this is unknown.
Second, the maximum concentration for the Waukegan Basin is considerably higher than those for the
other basins, suggesting a historic or current source, containing high mercury concentrations, to that
basin. Other than these noted differences, the sediments in the lake's depositional basins are amazingly
similar, suggesting either a similar regional source of mercury to the lake most likely delivered through
atmospheric pathways, or a well-mixed lake that redistributes inputs extremely well prior to
sedimentation to the lake bottom.
Table 6-15. Comparison of Mercury Concentrations in Various Basins of Lake Michigan for Box Cores
Only
Basin
Southern
Waukegan
Grand Haven
Milwaukee
Sarian
Algoma South
Algoma Central
Algoma North
Traverse
N
15
11
6
3
1
8
7
2
1
Mean
(ng/g)
120
160
150
130
160
140
130
130
160
Standard
Deviation (ng/g)
31
65
19
15
—
17
15
—
—
Median (ng/g)
130
130
150
130
—
140
130
—
—
Minimum (ng/g)
72
100
120
110
—
120
100
110
—
Maximum
(ng/g)
180
320
170
140
—
180
150
150
—
6.4.5 Mercury Fluxes
The amount of material available in trap samples limited the number of samples available for mercury
analyses. This limitation translates to a data bias because trap samples having enough material available
for mercury analysis represent relatively high sediment flux periods. As a result, mercury fluxes to traps
(0.049 to 3.7 ng/cm2/d) are always higher than fluxes to the sediment (0.0055 to 0.063 ng/cm2/d) at the
trap locations. Therefore, further discussion of mercury concentrations in and fluxes to sediment traps is
not warranted due to the bias.
As with mercury concentrations, mean mercury fluxes did not significantly vary from basin to basin of
the lake (Table 6-16). All fluxes were within one standard deviation of one another. Of interest are the
considerably higher minimum fluxes to the Algoma Basin relative to the other basins. In general, basins
that are towards the west side of the lake have lower mean and median fluxes than those on the east side
of the lake. Of significant note are the relatively high maximum mercury fluxes to the Southern and
Grand Haven Basins. Both of these basins are on the east side of the lake. These high fluxes could be
related to the transport of materials from the southwestern and southern shore of the lake to the eastern
shore, especially in the spring. This event occurs annually and the resulting plume has suspended
particulate matter concentrations 4 to 10 times that of the lake (Eadie et a/., 1996). A large amount of
particulate matter, with its associated contaminants, is transported along the eastern shore, where it settles
to the lake floor and accumulates in the Southern and Grand Haven deposition basins (Figure 6-5).
6-28
-------
Mercury in Surficial Sediments
Table 6-16. Comparison of Total Mercury Fluxes to Various Basins of Lake Michigan for Box Cores Only
Basin
Southern
Waukegan
Grand Haven
Milwaukee
Sarian
Algoma South
Algoma Central
Algoma North
Traverse
N
15
11
6
3
1
8
7
2
1
Mean
(ng/cm2/y)
10
3.4
10
3.3
14
6.9
5.2
7.6
8.0
Standard Deviation
(ng/cm2/y)
8.7
1.9
12
1.9
—
4.9
2.5
—
—
Median
(ng/cm2/y)
6.5
2.9
4.0
4.0
—
5.1
5.2
—
—
Minimum
(ng/cm2/y)
0.85
1.4
0.94
1.1
—
2.8
2.6
7.1
—
Maximum
(ng/cm2/y)
32
8.5
31
4.8
—
16
9.5
8.0
—
Fluxes to Lake Michigan in the vicinity of Station 15 (Figure 6-1) have decreased since 1981. In order to
compare fluxes between the two years, it is necessary to correct fluxes for sediment focusing. Sediment
focusing is the process by which fine-grained particles and their associated contaminants are winnowed
from the coarser fraction of sediments by wave and current action. Winnowing and resuspension occurs
in regions that are shallow enough to have wave and current velocities high enough to initiate sediment
grain movement. The resuspended materials are transported until they settle from the water column. For
each particle and associated contaminants, the process is repeated until the particle settles in a region
where winnowing and resuspension no longer occur. These regions are the depositional basins. For the
contaminants that are preferentially associated with fine-grained sediment particles, the
re suspension/transport process can result in a depletion or enhancement of a contaminant's net flux to any
one location. For sedimentary basins, the result is an enhancement of contaminate concentrations and
fluxes called sediment focusing. Sediment focusing can be estimated using parameters whose fluxes to a
lake's surface are equal at all locations. This is true for historically bomb-generated Cs-137 whose fluxes
to the region are well documented, and naturally derived Pb-210, whose flux is well known. Both of
these are mixed well in the atmosphere and were deposited to the lake's basin as a uniform flux from the
atmosphere. Because they, like contaminants, are associated with the fine-grained components of
sediment, they also are subject to sediment focusing. Because their fluxes are known, the degree of
sediment focusing can be calculated for them and then applied to observed contaminant fluxes. When
there is an excess of either of these radionuclides, the focusing factor is greater than one (depositional
basins). In regions of active winnowing and resuspension, the focusing factor may be less than one,
indicating depletion.
The Pb-210 and Cs-137 focusing factors are not always equal. The reason for this is unknown, but it is
reasoned to be related to each radionuclide associating with a different particle type. Because we do not
always know which focusing factor to apply to a particular contaminant, an average of the two focusing
factors can be used. For this study, however, the Cs-137 focusing factor will be used for the purpose of
comparison of Lake Michigan mercury fluxes to those of Green Bay and Lake Superior. For both those
locations, the Cs-137 focusing factor was used because only the Cs-137 factor was available for Lake
Superior. Mercury fluxes to Lake Michigan are very similar to those for the open waters of Lake
Superior, but are considerably lower than those to Green Bay (Table 6-17).
6-29
-------
Results of the LMMB Study: Mercury Data Report
Table 6-17. Comparison of Total Mercury Fluxes for Lake Michigan Corrected for Cs-137 Focusing
Factors to Fluxes for other Locations
Location
Lake Michigan
Lake Superior
Green Bay
Mean
(ng/cm2/y)
3.4
3.2
19
Standard Deviation
(ng/cm2/y)
1.8
1.1
30
Median
(ng/cm2/y)
3.2
2.8
14
Reference
this study
Rossmann (1999)
Rossmann and Edgington (2000)
A good illustration of the use of a sediment focusing factor is the region of the lake around Station 15.
Total uncorrected mercury fluxes to the surficial 1 cm of sediment are very similar in magnitude (Table 6-
18). When corrected for sediment focusing, it becomes apparent that the flux of mercury to this region of
the lake has decreased from 13 ng/cm2/y in surficial sediments collected in 1981, to 4.1 ng/cm2/y for
surficial sediments collected in 1994. This is consistent with the observed trend of decreasing mercury
concentrations in surficial sediments.
Table 6-18. Comparison of Mercury Fluxes to Lake Michigan Surficial Sediments at Station 15 in 1981 and
1994
Year
1981
1994
Total Mercury Flux (ng/cm2/y)
22
23
Total Mercury Flux Corrected for Focusing Factor (ng/cm2/y)
13
4.1
6.4.6 Relative Importance of Regional Atmospheric Sources and Point Sources of Mercury
To estimate the relative contribution of regional atmospheric and local point-source mercury fluxes to
measured total mercury fluxes, the total mercury fluxes were corrected with the Cs-137 focusing factor.
For Lake Michigan, atmospheric mercury fluxes account for 50% of the total mercury flux. This is higher
than that for Lake Superior (38%) and Green Bay (15%). Fluxes of mercury to Green Bay are dominated
by point sources derived from historic industrial use of mercury within the region (Rossmann and
Edgington, 2000).
6.5 Conclusions
Lake Michigan surficial sediments have low mercury concentrations relative to Green Bay. The mean
concentration was 0.078 mg/kg. The mean net total mercury flux to the depositional basins was 7.2
ng/cm2/y. Mercury fluxes to Lake Michigan sediments were similar to those for Lake Superior open-
water sediments and considerably lower than those to Green Bay sediments. There was little variation in
mercury concentration or fluxes from basin to basin of the lake. Mercury concentration distribution
patterns in surficial sediments are similar to historic patterns and conform to the bathymetry. Fluxes do
not conform to the bathymetry and are elevated along the eastern shore of the lake. Regional atmospheric
fluxes of mercury account for 50% of the total mercury flux to recent surficial sediments. Both mercury
concentrations in, and fluxes to, surficial sediments have decreased since the 1970s.
6-30
-------
Chapter 7
Mercury in Plankton
7.1 Results
Phytoplankton and zooplankton were collected in Lake Michigan from June 1994 through October 1995
for total mercury analysis. Phytoplankton samples were collected by pumping water from the optimum
depth in the water column for maximum phytoplankton density through 10-^m phytovibe nets.
Zooplankton samples were collected in vertical tows using nested 102-^m and 500-^m plankton nets (see
Section 2.4.5 for details of the sample collection procedures). Plankton samples were collected from 15
locations, including 9 stations within 4 designated biological sampling areas (biota boxes) and 6
additional routine monitoring stations (see Figure 2-7 in Chapter 2). A total of 157 samples were
collected and analyzed for total mercury by cold vapor atomic fluorescence spectroscopy (Table 7-1).
7.1.1 Variation Among Sample Types
All plankton samples collected from Lake Michigan, except one zooplankton sample, contained total
mercury levels above sample-specific detection limits, which averaged 8.65 ng/g for phytoplankton and
7.82 ng/g for zooplankton. Total mercury concentrations in phytoplankton ranged from 10.9 to 176 ng/g
and averaged 35.0 ng/g. Total mercury concentrations in zooplankton ranged from 11.0 to 376 ng/g and
averaged 54.3 ng/g. Based on a paired t-test using log-transformed mercury data, Lake Michigan
zooplankton contained significantly higher (at the 95% confidence level) levels of mercury than
phytoplankton (Figure 7-1).
The significantly higher levels of mercury found in zooplankton compared to phytoplankton suggest the
bioaccumulation and biomagnification of mercury in the lower pelagic food web of Lake Michigan.
PCBs and fra«s-nonachlor also were found to bioaccumulate and biomagnify in the Lake Michigan food
web (USEPA, 2004). For PCBs and fra«s-nonachlor, a portion of the difference between zooplankton
and phytoplankton concentrations was due to the lipid content in the two groups. This was not true for
mercury accumulation. Mercury concentrations in zooplankton and phytoplankton were not correlated
with lipid content (r2 of 5% and 0.9% for phytoplankton and zooplankton, respectively), and generalized
linear model results showed that lipid content did not explain a significant portion of variability in
mercury data either directly, or through interaction with trophic level (phytoplankton/zooplankton).
While organic contaminants such as PCBs and fra«s-nonachlor are preferentially accumulated in fatty
tissues, mercury does not appear to be preferentially accumulated in such tissues. Mercury has been
shown to preferentially bind to sulfhydryl groups in proteins, and in fish, accumulate in muscle tissue
(USEPA, 1999b).
7-1
-------
Results of the LMMB Study: Mercury Data Report
Table 7-1. Number of Plankton Samples Analyzed for Mercury in the LMMB Study
Sample Type
Phytoplankton
Zooplankton
Sampling Location
Biota Box
Chicago biota box
Sturgeon Bay biota box
Port Washington biota box
Saugatuck biota box
Other
Station
05
110
140
180
240
280
310
340
380
18M
23M
27M
40M
47M
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
06/20/94 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 10/08/95
06/23/94 to 10/03/95
06/20/94 to 08/1 0/95
08/1 2/94 to 04/1 2/95
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
05
110
140
180
240
280
310
340
380
18M
23M
27M
40M
47M
19M
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
06/20/94 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 10/08/95
08/1 9/94 to 10/03/95
06/20/94 to 08/1 0/95
10/1 8/94 to 04/1 2/95
06/1 7/94 to 09/1 9/95
01/24/95 to 01/24/95
Total
Total
Number of Samples
Analyzed
7
6
6
6
5
6
6
6
7
6
6
3
3
5
78
7
6
6
5
6
6
6
6
7
6
5
4
2
6
1
79
157
7-2
-------
Mercury in Plankton
Figure 7-1. Mercury Concentrations in Phytoplankton and Zooplankton Measured in
Lake Michigan
1000-,
o
U-J
CD
-------
Results of the LMMB Study: Mercury Data Report
Figure 7-2. Mercury Concentrations in Phytoplankton (A) and Zooplankton (B)
Measured in Lake Michigan during Six Cruises
A
1000-,
O)
o
"-i—'
£
I
o
O
100;
10-
o
5
(D
ii
CO
o
5
(D
II
-N
O
5
(D
CO
II
CO
O
5
en'
(D
II
K>
A
o
5
en'
(D
en
II
K>
A
L^
j
i_
_L
i
-L
I
i
— i —
o
5
(D
ii
K>
1000q
O)
A
100-
o>
o
c
o
o
10
1
1
_L
o c
5 c
(D n
AB
_^ T
i
5 Q
5
(D
B r^
• 1 ' 1
1 1 1
1 1 _l_ ^1
X
99-
5 5 c
' '
II
CO
II
CO
II
S
II
i\j
II
K>
(Cruise 1 = June 1994, Cruise 2 = August 1994, Cruise 3 = September/October 1994, Cruise 4 = March/April 1995, Cruise 5 = August 1995,
and Cruise 6 = September/October 1995)
Boxes represent the 25th (box bottom), 50th (center line), and 75th (box top) percentile 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.
Xs represent results beyond 3*IQR from the box. Letters above the boxes represent results of analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05).
7-4
-------
Mercury in Plankton
7.1.3 Geographical Variation
Plankton samples were collected from 15 sampling stations in Lake Michigan (see Figure 2-7 in Chapter
2). Nine of these sampling stations were focused in the following four biological sampling areas or biota
boxes:
»• Chicago biota box — around Station 5 in the southern Lake Michigan basin near Chicago
»• Sturgeon Bay biota box — a combination of three stations (110, 140, and 180) on the western side
of the northern Lake Michigan basin near Sturgeon Bay, Wisconsin
*• Port Washington biota box — a combination of two stations (240 and 280) in the central Lake
Michigan basin near Port Washington, Wisconsin
»• Saugatuck biota box — a series of three stations (310, 340, and 380) 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 LMMB monitoring
sites throughout the lake (Table 7-1). Table 7-2 shows the concentrations of total mercury measured in
plankton collected from the various sampling locations.
Considering all 15 individual sampling stations, two-way analysis of variance (accounting for cruise and
sampling station) revealed no significant differences among sampling stations in phytoplankton or
zooplankton mercury concentrations (Figure 7-3). When combining data within biota boxes,
phytoplankton mercury concentrations still did not vary significantly among the biota box stations. The
highest individual (176 ng/g) and mean (46.9 ng/g) phytoplankton mercury concentrations were observed
at the Saugatuck biota box, but this site also contained the greatest variability, and differences between
this site and other sites were not statistically significant (at the 95% confidence level).
Zooplankton mercury concentrations did vary significantly among biota boxes, however, no distinct trend
was observed. A significant interaction occurred between the biota box and cruise variables, such that
significant differences between stations were cruise-dependent. During Cruise 1, zooplankton mercury
concentrations at the Saugatuck biota box were significantly higher than at the Sturgeon Bay biota box.
During Cruise 3, zooplankton mercury concentrations at the Port Washington biota box were significantly
higher than at the Saugatuck biota box. During Cruise 6, zooplankton mercury concentrations at the
Chicago biota box were significantly higher than at the Saugatuck biota box.
7.1.4 Bioaccumulation
Mercury is known to accumulate in living organisms at levels far above concentrations in the water
column. The degree of this 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.
In the LMMB Study, bioaccumulation factors for mercury were calculated as the mean concentration of
mercury in phytoplankton or zooplankton divided by the lake-wide mean concentration of total mercury
in Lake Michigan. Concentrations of total mercury in Lake Michigan plankton were generally 105 times
higher than total mercury concentrations in Lake Michigan water, which averaged 0.328 ng/L (or
0.000328 ng/g, assuming the density of water is 1 g/mL). Bioaccumulation factors from water to
phytoplankton were 1.07 x 105 and from water to zooplankton were 1.66 x 105.
7-5
-------
Results of the LMMB Study: Mercury Data Report
To evaluate the accumulation and transfer of mercury 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 zooplankton
divided by the concentration in phytoplankton. The biomagnification factor for mercury between
phytoplankton and zooplankton was 1.55.
Table 7-2. Mercury Concentrations in Plankton Measured at Various Sampling Stations in Lake Michigan
Sample Type
Phytoplankton
Zooplankton
Sampling Station
Biota Box
Chicago biota box
Sturgeon Bay biota
box
Port Washington
biota box
Saugatuck biota
box
Other
Chicago biota box
Sturgeon Bay biota
box
Port Washington
biota box
Saugatuck biota
box
Other
Station
05
110
140
180
combined
240
280
combined
310
340
380
combined
18M
23M
27M
40M
47M
05
110
140
180
combined
240
280
combined
310
340
380
combined
18M
23M
27M
40M
47M
19M
N
7
6
6
6
18
5
6
11
6
6
7
19
6
6
3
3
5
7
6
6
5
17
6
6
12
6
6
7
19
6
5
4
2
6
1
Mean
(ng/g)
35.3
31.6
20.4
19.4
23.8
29.6
28.8
29.2
78.7
27.0
36.8
46.9
29.4
44.2
40.2
30.4
38.2
75.0
45.6
47.8
52.9
48.5
49.3
60.5
54.9
112
44.7
40.0
64.2
36.0
46.7
63.1
41.5
44.5
11.0
Range (ng/g)
21. 5 to 56.3
11.6 to 64.1
10.9 to 37.3
11. 2 to 30.5
10.9 to 64.1
14.0 to 58.8
14.7 to 48.7
14.0 to 58.8
16.8 to 176
15.1 to 66.6
12.3 to 96.4
12.3 to 176
12.9 to 69.2
13.4 to 111
15.0 to 77.7
24.0 to 38.7
13.3 to 89.9
45.3 to 177
15.3 to 72.4
23.5 to 65.1
23.7 to 97.5
15.3 to 97.5
29.6 to 86.5
29.6 to 94.8
29.6 to 94.8
31. 9 to 376
30.5 to 67.2
32.8 to 50.0
30.5 to 376
29.2 to 55.4
37.1 to 62.4
18.5 to 152
38.0 to 45.0
30.0 to 71. 3
NA
SD
(ng/g)
12.7
21.5
9.31
7.26
14.5
17.4
13.3
14.5
58.9
19.7
28.1
42.9
21.3
37.7
33.1
7.53
31.0
45.9
22.5
16.7
30.3
22.0
19.4
26.9
23.1
131
13.5
5.96
76.9
9.65
11.4
60.2
4.95
17.2
NA
RSD
(%)
36.2
67.8
45.7
37.4
60.8
58.6
46.3
49.6
74.9
72.9
76.4
91.5
72.6
85.2
82.2
24.8
81.2
61.2
49.4
34.9
57.2
45.4
39.4
44.5
42.1
117
30.1
14.9
120
26.9
24.5
95.4
11.9
38.6
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
0
0
0
0
0
100
NA = Not applicable
7-6
-------
Mercury in Plankton
Figure 7-3. Mercury Concentrations in Phytoplankton (A) and Zooplankton (B) Measured
at Various Sampling Stations in Lake Michigan
A
1000q
o
"ro
§
o
O
A
T
100-
10
A
y
A
JL
—
T A
Ss
T y y
A
1
—
^
1
A
r^
B
A
~
A
JQ
i
T
A
A T
Q?B
A
T
—
^ -L 1 T
g
-* -N 00 00
o o o ^
-N
O
B
II II
O3 O3
II
O3
II II
O3 O3
II
Ol
II
CO
II II
O3 O3
ii
O3
II II
CO Ol
1000z,
I
O
'-4-»
ro
(U
o
c
o
O
100-
10-
o
Ol
->• -N 00
O O O
S~ S~ S~
ii ii ii
O3 O3 Ol
-» (O (O
00 CO -N
=S ^ O
(O (O
^i oo
CO CO
00
o o o o
S~ S~ S~ S~
-N
O
Boxes represent the 25th (box bottom), 50th (center line), and 75th (box top) percentile 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.
Xs represent results beyond 3*IQR from the box. Letters above the boxes represent results of analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05).
7-7
-------
Results of the LMMB Study: Mercury Data Report
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 mercury
monitoring portion of the study are further described in Section 2.6 and included use of SOPs, training of
laboratory and field personnel, and establishment of method quality objectives (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, 2001b). A brief summary of the quality of plankton
mercury 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 2.6, 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 7-3 provides a
summary of flags applied to the plankton mercury 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 2.6. No
results were qualified as invalid, thus all results are represented in the analysis of plankton mercury
concentrations presented in this report.
Table 7-3. Summary of Routine Field Sample Flags applied to Mercury in Plankton Samples
Flag
EHT, Exceeded Holding Time
FBS, Failed Blank Sample
FDL, Failed Lab Duplicate
FFD, Failed Field Duplicate
FLS, Failed Lab Spike
SCF, Suspected Field Contamination
UDL, Below Sample-Specific Detection Limit
Number of QC Samples
—
18 lab reagent blank samples
31 lab duplicate samples
38 field duplicate samples
1 1 lab fortified spiked samples
—
—
Percentage of Samples Flagged
75% (11 8)
44% (69)
0
4% (6)
0
1%(2)
1%(1)
The most frequently applied data validation flag was for exceeding sample holding times. Seventy-five
percent of samples were analyzed beyond the 420-day established holding time. The median holding
time for frozen plankton samples was 614 days, and frozen samples were held as long as 896 days prior to
mercury analysis. The MQOs for holding times were based on educated, conservative assessments by the
Pis, however, the appropriateness of these holding times has not been rigorously determined and the
effects of extended holding times have not been investigated in the plankton matrix. Because
phytoplankton samples were analyzed for total mercury, as opposed to the determination of mercury
7-8
-------
Mercury in Plankton
species, possible conversion of mercury among individual species during the extended holding times
would not likely affect total mercury measurements and loss of mercury would likely be negligible.
Laboratory reagent blanks were analyzed to assess the potential for contamination of routine field
samples. A total of 18 laboratory reagent blanks were analyzed, and 11 of these 18 blanks contained
detectable mercury. Forty-four percent of routine field samples were associated with (e.g., analyzed in
the same batch) one of these 11 blanks that contained detectable mercury and were flagged for a failed
blank (FBS). While 44% of routine field samples were flagged for associated blank failure, the maximum
level of mercury detected in laboratory reagent blank samples was 0.1 ng/g, which is 100 times less than
the lowest measured mercury concentration in plankton samples (10.9 ng/g). For this reason,
contamination is not believed to significantly affect the reported plankton mercury results.
In addition to laboratory reagent blanks, laboratory dry blanks were analyzed at a frequency of 1 per 12
routine field samples. These blank results were not used to flag data, because they were not linked to
specific routine field samples. Like laboratory reagent blanks, measured concentrations in laboratory dry
blanks were 0.1 ng/g or below, further indicating that contamination did not significantly affect reported
plankton mercury results. While blank sample analysis indicates no pervasive contamination, two
samples were flagged for suspected field contamination based on a hydraulic fluid spill on the deck of the
sampling vessel during the June 1994 sampling at Station 310.
A total of 38 field duplicate samples and 31 laboratory duplicate samples were analyzed to assess
precision. From each cruise (except the January 1995 cruise that visited only two sites), duplicate
samples were collected at one to six stations. Laboratory duplicates were prepared at a frequency of at
least 2 per set of 24 routine field samples. 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 results for a sample and its duplicate was greater than 30%. No laboratory
duplicates failed to meet this criteria, and only 6 of the 38 field duplicates were flagged.
Laboratory fortified spike samples were used to monitor analytical bias, and no results were qualified for
failed laboratory spikes. Based on an analysis of laboratory 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.
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 7-4 provides a
summary of data quality assessments for several of these attributes for plankton data. The results of
laboratory and field duplicate samples revealed good system and analytical precision for plankton data.
The mean RPD for field duplicate samples was 19.8% and the mean RPD for laboratory duplicate
samples was 11.2%.
Analytical bias was evaluated by calculating the mean recovery of a standard reference material (SRM)
from the National Institute of Standards and Technology and the mean recovery of laboratory fortified
spike samples (LFS). Results indicated very little overall bias for analytical results. Mean recoveries for
SRM 1515, an apple leaf sample with a certified value of 0.044 mg/kg, were 98%, and mean LFS
recoveries were 103%, just slightly above and below the ideal recovery of 100%.
Analytical sensitivity was evaluated by calculating the percentage of samples reported below the sample-
specific detection limit. Only one sample, or 0.6% of the data, was below the detection limit. Results
7-9
-------
Results of the LMMB Study: Mercury Data Report
from this sample were not censored and were used as reported in the analysis of plankton mercury data
presented in this report.
Table 7-4. Data Quality Assessment in Plankton Samples
Parameter
Number of Routine Samples Analyzed
System Precision, Mean Field Duplicate RPD (%), >MDL
Analytical Precision, Mean Lab Duplicate RPD (%), >MDL
Analytical Bias, Mean SRM (%)
Analytical Bias, Mean LFS (%)
Analytical Sensitivity, Samples reported as
-------
Mercury in Plankton
(1998) observed seasonal variations in plankton mercury concentrations with the lowest values occurring
in spring and increasing throughout the summer. Similarly, Kirkwood et al. (1999) observed increases in
phytoplankton mercury concentrations in the hypolimnion throughout the summer season in two
Canadian lakes. In Devil's Lake, Wisconsin, Herrin et al. (1998) noted that mercury concentrations in the
water of the hypolimnion increased during stratification, and that mercury concentrations in Daphnia
peaked near the time of lake turnover in the fall (Herrin et al., 1998). Concentrations of methylmercury
in phytoplankton and zooplankton increased two to four-fold between peak stratification and complete
mixing. Herrin et al. (1998) concluded that mercury (particularly methylmercury) stored in the anoxic
hypolimnion during summer stratification is an important source of mercury to the food chain during
turnover. While plankton mercury levels measured in the LMMB Study increased in the late summer and
fall as described by Herrin et al. (1998) in Devil's Lake, water column concentrations in Lake Michigan
did not follow the same trend. No seasonal differences in epilimnetic or hypolimnetic mercury levels
were observed in the LMMB Study (see Chapter 5). The Lake Michigan main lake hypolimnion is
always oxic.
7.3.3 Bioaccumulation and Biomagnification
Mercury bioaccumulation factors calculated in the LMMB Study were 1.07 x 105 for phytoplankton and
1.66 x 105 for zooplankton. These bioaccumulation factors are slightly higher than reported by other
researchers for other lakes in the region. Bioconcentration factors in phytoplankton and zooplankton
from a north-central Wisconsin lake were approximately 3 x 104 and 5 x 104, respectively (Watras and
Bloom, 1992). Similarly, bioaccumulation factors for plankton in 12 Minnesota lakes were
approximately 3 x 104 (Monson and Brezonik, 1998).
In addition to bioaccumulation of mercury in the lower pelagic food web, LMMB Study results indicate
the biomagnification of mercury within the lower pelagic food web. Zooplankton mercury levels were
significantly higher than phytoplankton mercury levels. The biomagnification factor calculated between
phytoplankton and zooplankton in the LMMB Study was 1.55. Other studies have also documented the
biomagnification of mercury within the lower pelagic food web. Watras and Bloom (1992) measured
higher mercury and methylmercury levels in zooplankton than phytoplankton in both reference and
acidified lakes.
Tremblay et al. (1998) concluded biomagnification in the planktonic food web of Canadian reservoirs
based on observed increases in methylmercury with increasing plankton size. Tremblay et al. (1998)
measured biomagnification factors of 2.5 to 3 between adjacent trophic levels within the planktonic food
web. These biomagnification factors are likely higher than those calculated for Lake Michigan because
they are calculated based on methylmercury levels rather than total mercury levels.
While methylmercury concentrations were not measured in plankton and water during the LMMB Study,
Watras and Bloom (1992) concluded that it is the methylmercury species that is most efficiently
bioaccumulated and transferred up aquatic food chains. Methylmercury bioaccumulation factors were
considerably higher (3 x 105 and 1 x 106 for phytoplankton and zooplankton, respectively) than
bioaccumulation factors calculated based on total mercury concentrations. To further emphasize the
importance of methylmercury in bioaccumulation and biomagnification, Back and Watras (1995)
observed biomagnification of methylmercury from seston (which included phytoplankton and other
organic suspended matter) to herbivorous zooplankton, but reported that total mercury levels did not
increase between these trophic levels. Watras and Bloom (1992) also found that methylmercury becomes
a progressively greater fraction of total mercury as trophic levels increase. For instance, 5% of total
mercury in water was methylmercury; 13% of phytoplankton total mercury was methylmercury; 29% of
zooplankton mercury was methylmercury; and >90% offish mercury was methylmercury.
7-11
-------
Results of the LMMB Study: Mercury Data Report
7.3.4 Other Interpretations and Perspectives
Researchers have identified various physical and chemical properties within studied lakes that have
correlated with plankton mercury levels in the lakes. In general, mercury accumulation in plankton has
been observed to increase with increasing water concentrations, and decreasing pH, however, researchers
have not all agreed on the importance of these factors or additional factors in affecting bioaccumulation.
Sorensen et al. (1990) found that concentrations of mercury in zooplankton from 80 northern Minnesota
lakes correlated with mercury in water, mercury in fish, zooplankton density (negative correlation), pH
(negative correlation), and total organic carbon. Westcott and Kalff (1996) found that water color and pH
together were the best predictors of methylmercury levels in plankton from 24 Ontario lakes.
Methylmercury concentrations also were positively correlated with drainage ratio and percent wetlands in
the catchment (Westcott and Kalff, 1996). In contrast, Tremblay et al. (1995) found that zooplankton
mercury concentrations in 73 Canadian lakes were poorly correlated with catchment area, primary
production, total organic carbon, and sediment mercury levels. Monson and Brezonik (1998) found no
correlations of plankton mercury levels with acid-neutralizing capacity, pH, dissolved organic carbon,
sulfate, chlorophyll, or phosphorus in 12 northern Minnesota lakes. Back and Watras (1995) also found
no relationship between total mercury in zooplankton and pH in 12 northern Wisconsin lakes.
In a direct comparison between the acidified and reference basins of Little Rock Lake, Watras and Bloom
(1992) found that pH greatly influenced mercury accumulation, particularly in the methylmercury form.
Mean concentrations of total mercury in phytoplankton and zooplankton were 20-30% higher in the
acidified lake (pH 4.7) than in the reference lake (pH 6.1), and mean concentrations of methylmercury
were 2-4 times higher in the acidified lake. The acidified conditions also appeared to greatly affect the
fraction of mercury that is in the form of methylmercury. In the acidified lake, methylmercury comprised
>90% of the total mercury in Cladocera, whereas <30% of total mercury in Cladocera from the non-
acidified lake was methylmercury. Watras and Bloom (1992) concluded that it is the methylmercury
form of mercury that is preferentially bioaccumulated and transferred up aquatic food chains, so greater
proportions of methylmercury at lower trophic levels in the food chain will likely lead to greater
biomagnification of mercury at higher levels of the food chain.
Later work by Watras et al. (1998) demonstrates that the bioaccumulation of mercury depends not only
on the form of mercury under consideration (e.g., methylmercury versus inorganic mercury), but also on
the particular chemical species within each form (e.g., "neutral" species such as CH3HgCl° and
CH3HgOH° behave differently than ionized forms such as CH3Hg+). Some of the differences in
bioaccumulation are a function of interactions and correlations with other water quality characteristics
such as pH and dissolved organic carbon (DOC). The LMMB Study did not measure methylmercury in
the water or all of the trophic levels of biota, nor were particular mercury species measured within any of
the media. Therefore, it is unlikely that the results from this study can be used to delineate specific
bioaccumulation mechanisms or pathways. Rather, the bioaccumulation factors reported in this chapter
are relatively simple approximations of the transfer of mercury from the water column to the various
trophic levels that are indicative of general trends in mercury concentrations.
Finally, the zooplankton data from Watras and Bloom (1992) represent results for organisms that were
fractionated by size and sorted by species prior to analyses. Watras and Bloom (1992) contrast their
results with bioaccumulation factors calculated from mixed assemblages of zooplankton, in which
"obscure small but important difference in bioaccumulation. " The plankton results from the LMMB
Study are based on aggregate samples without regard for species. Thus, although the LMMB results
demonstrate that there is bioaccumulation of mercury within the lower pelagic food web, the calculated
bioaccumulation factors may not represent the accumulation that occurs between particular species within
the lake ecosystem.
7-12
-------
Chapter 8
Mercury in Fish
8.1 Results
Lake Michigan fish were collected from April 1994 through October 1995 for total mercury analysis (see
Section 2.4.6 for details of the sample collection procedures and Section 2.5.5 for the details of the
analysis procedures). Lake trout and coho salmon were collected using gill nets, trawl nets, or other
appropriate means. Up to five individual whole fish of the same species and size or age category were
combined to produce composite fish samples at each collection. Adult lake trout from 172 to 933 mm in
length were collected from three biological sampling areas or biota boxes (see Figure 2-7 in Chapter 2):
*• Sturgeon Bay biota box — a combination of three stations (110, 140, and 180) on the western side
of the northern Lake Michigan basin near Sturgeon Bay, Wisconsin
»• Port Washington biota box — a combination of two stations (240 and 280) in the central Lake
Michigan basin near Port Washington, Wisconsin
*• Saugatuck biota box — a series of three stations (310, 340, and 380) on the eastern side of the
southern Lake Michigan basin near Saugatuck, Michigan
Coho salmon were collected in three distinct age classes (hatchery, yearlings, and adult). 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 (see Figure 2-7 in Chapter 2). In addition, young coho salmon (hatchery) were collected directly
from the Platte River hatchery, where the majority of Lake Michigan stocked salmon originate. Overall, a
total of 201 composite samples of lake trout and coho salmon were collected and analyzed for total
mercury by cold vapor atomic fluorescence spectroscopy (Table 8-1).
Table 8-1. Number of Composite Fish Samples Analyzed for Mercury
Species-Size Category
Coho-Hatchery
Coho-Yearling
Coho-Adult
Lake Trout
Sampling Dates
04/21/94 to 04/27/94
10/1 8/94 to 11/1 6/94
05/1 0/94 to 10/25/94
05/1 2/94 to 10/26/95
Total
Number of Composite Samples
5
8
32
156
201
8.1.1 Variation Among Species
Table 8-2 shows the mean concentration of total mercury (on a wet-weight basis) in Lake Michigan coho
salmon and lake trout. Mercury concentrations in adult lake trout ranged as high as 396 ng/g and
averaged 139 ng/g. In coho salmon, mercury concentrations ranged as high as 127 ng/g and averaged
79.9, 20.6, and 69.0 ng/g in hatchery, yearling, and adult salmon, respectively. Analysis of variance
revealed that mercury concentrations in lake trout were significantly higher than in adult or yearling coho
salmon (Figure 8-1). Adult coho salmon also were significantly higher in mercury concentrations than
yearling coho, which contained the lowest mean concentration of mercury (20.6 ng/g). Coho salmon
collected directly from the hatchery surprisingly contained higher mercury levels (average of 79.9 ng/g)
than yearling or adult coho salmon and were not significantly different from lake trout mercury levels.
8-1
-------
Results of the LMMB Study: Mercury Data Report
This is surprising because smaller, younger fish generally contain lower levels of bioaccumulative
contaminants than older, larger fish. Among adult coho salmon and lake trout, fish length was highly
correlated with total mercury concentrations (see Section 8. 1 .2). Higher mercury concentrations in
hatchery samples than in adult coho may be due to differences in exposures between the hatchery and
Lake Michigan or differences in uptake and elimination rates between hatchery and adult fish. Also,
given the smaller number of composites of hatchery and yearling salmon, the mean values calculated for
these groups may be less representative of their respective populations than mean values calculated for
adult salmon and lake trout.
Table 8-2. Mean Total Mercury Concentrations in Lake Michigan Fish (Wet-weight Basis)
Species/Size
Category
Coho- Hatchery
Coho-Yearling
Coho-Adult
Lake Trout
N
5
8
32
156
Mean (ng/g)
79.9
20.6
69.0
139
Median
(ng/g)
81.2
18.1
69.8
130
Range (ng/g)
70.0 to 88.0
13.7 to 38.6
23.3 to 127
19.5 to 396
SD (ng/g)
6.77
7.85
35.9
83.8
RSD (%)
8.48
38.0
52.0
60.1
Below DL (%)
0
0
0
0
Figure 8-1. Total Mercury Concentration (Wet-weight Basis) in Lake Michigan
Fish
1000=,
I
o
O
=3
O
100:
10
AB
0
-
3
M
GO
0
-
-<
CD
Q)
5'
CO
O
C
rt-
Ol
CD
Boxes represent the 25th (box bottom), 50th (center line), and 75th (box top) percentile 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 results of analysis of variance and multiple comparisons
test. Boxes with the same letter were not statistically different (at alpha = 0.05).
The trends observed in fish mercury concentrations were the same on a dry-weight basis (Table 8-3).
Lake trout contained the highest mercury levels, followed by hatchery, adult, and yearling coho salmon.
As with wet-weight basis results, dry-weight mercury concentrations in lake trout were significantly
higher than in adult or yearling coho salmon, and mercury concentrations in adult coho salmon were
significantly higher than in yearling coho salmon.
8-2
-------
Mercury in Fish
Table 8-3. Mean Total Mercury Concentrations in Lake Michigan Fish (Dry-weight Basis)
Species/Size
Category
Coho-Hatchery
Coho-Yearling
Coho-Adult
Lake Trout
N
5
8
32
156
Mean (ng/g)
317
71.3
248
373
Median
(ng/g)
331
57.4
255
341
Range (ng/g)
269 to 344
43.1 to 156
98.8 to 504
83.5 to 929
SD (ng/g)
30.1
36.2
119
200
RSD (%)
9.52
50.7
47.9
53.6
Below DL (%)
0
0
0
0
8.1.2 Factors Affecting Contaminant Concentrations
Log-transformed total mercury concentrations in Lake Michigan fish were highly correlated (p<0.0001)
with fish length and lipid content. Fish length was positively correlated with adult lake trout and adult
coho salmon mercury levels with r2 values of 0.856 and 0.824, respectively (i.e., 85.6% and 82.4% of the
variability observed in lake trout and adult coho salmon mercury concentrations are attributable to the fish
length). It should be noted that the fish samples analyzed 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, therefore
allowing for direct comparison of length and contaminant concentration. Figure 8-2 shows the
relationship between fish length and mercury concentrations in Lake Michigan lake trout and coho
salmon. Mercury concentrations generally increased exponentially with increasing fish length, producing
a linear relationship between fish length and log concentration. Because fish length is often used as a
surrogate measure for fish age, this trend indicates either the increased accumulation of pollutants in older
fish that have experienced longer duration exposures to mercury, or exposures to higher mercury
concentrations.
Figure 8-2. Relationship of Fish Length and Mercury Concentration
1000
O)
c
£ 100
2
•4-1
0)
o
c
o
O
o
0)
• Lake Trout
• Coho Salmon
log Y = 0.00156X + 1.12
log Y= 0.00277 X + 0.301
200 400 600 800
Fish Length (mm)
1000
Mercury concentrations in Lake Michigan fish also were strongly correlated with fish lipid content
(p<0.0001). Lipid content was positively correlated with adult lake trout and adult coho salmon mercury
levels with r2 values of 0.684 and 0.531, respectively. This correlation, however, was likely due to the
8-3
-------
Results of the LMMB Study: Mercury Data Report
intercorrelation between fish length and lipid content. Lipid content was significantly correlated with fish
length (r2 = 0.798 for lake trout; r2 = 0.486 for adult coho salmon), which was in turn correlated with
mercury concentration. In general, mercury accumulation in fish is associated with proteins and storage
in muscle tissue rather than storage in fatty tissues, where organic contaminants are accumulated, so lipid
content is not considered a controlling variable in fish mercury concentrations. In the case of lake trout,
multiple regression analysis supported the assumption that lipid content correlation with mercury
concentration was a result of the intercorrelation between lipid content and fish length. Multiple
regression analysis revealed that mercury concentrations in lake trout were not significantly affected by
fish lipid content, when controlling for fish length. For adult salmon, however, multiple regression
analysis revealed that mercury concentration was significantly affected by fish length, lipid content, and
the interaction of these two factors.
8.1.3 Geographical and Seasonal Variation
Lake trout were collected from three biological sampling areas or biota boxes (Sturgeon Bay, Port
Washington, and Saugatuck) during the spring, summer, and autumn months. Two-way analysis of
variance (accounting for sampling station and season) revealed that mercury concentrations in lake trout
did not differ significantly (at the 95% confidence level) among seasons but did differ significantly
among biota boxes. This analysis was not conducted for coho salmon mercury data because coho were
collected from various locations throughout the lake, rather than from the designated biota boxes, and
coho composite samples occasionally consisted offish from different sampling sites.
Mercury concentrations in lake trout from the three biota boxes averaged 165 ng/g at Sturgeon Bay, 114
ng/g at Port Washington, and 127 ng/g at Saugatuck. Tukey's multiple comparison test revealed that the
mercury concentration in lake trout from Port Washington was significantly lower than in lake trout from
Sturgeon Bay. This difference, however, is primarily due to differences in the size offish collected from
the sites. The length of lake trout from Port Washington averaged 536 mm, compared to an average of
629 mm for lake trout from Sturgeon Bay. Because fish mercury concentrations are so strongly
correlated with fish length, decreased fish mercury concentrations at Port Washington could be due to the
smaller size offish from this site. Multiple regression analysis was used to evaluate differences between
biota boxes while considering fish length. Figure 8-3 compares the mercury versus fish length
regressions for fish collected at each of the biota boxes.
Figure 8-3. Total Mercury Concentrations in Lake Michigan Lake Trout of Various Sizes
from the Three Biological Sampling Stations
1000
-5*
"3)
i. 100
u
o
o
£<
I
0)
10 -
200
400 600
Fish Length (mm)
800
1000
8-4
-------
Mercury in Fish
While differences among biota boxes are small, multiple regression analysis determined that the
regression intercept for Saugatuck is significantly lower than for the other two sampling locations. When
comparing similarly sized fish from the three biota boxes, lake trout from Saugatuck contained
significantly lower mercury concentrations than lake trout from Sturgeon Bay or Port Washington.
8.1.4 Bioaccumulation
Mercury is known to accumulate in living organisms at levels far above concentrations in the water
column. The degree of this 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.
In the LMMB Study, 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 mercury in
Lake Michigan fish were generally 105 to 106 times higher than total mercury concentrations in Lake
Michigan water, which averaged 0.328 ng/L (or 0.000328 ng/g assuming a water density of 1 g/mL).
Bioaccumulation factors were 2.18 x 105 for yearling coho salmon, 7.58 x 105 for adult coho salmon, and
1.14 x 106 for adult lake trout. Bioaccumulation factors were not calculated for hatchery coho salmon,
because these samples were not collected from Lake Michigan.
The fish species analyzed for mercury content in the LMMB Study (coho salmon and lake trout)
represented only top predator fish species. While forage fish species were collected and analyzed for
PCBs and fra«s-nonachlor, these species were not analyzed for mercury. For this reason,
biomagnification of mercury in the upper pelagic food web could not be assessed. Biomagnification from
the lower pelagic food web (plankton) to the upper pelagic food web (fish) is discussed in Chapter 9 of
this report.
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 mercury
monitoring portion of the study are further described in Section 2.6 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, 2001b). A brief summary of the quality offish mercury 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.
8-5
-------
Results of the LMMB Study: Mercury Data Report
As discussed in Section 2.6, 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 8-4 provides a
summary of flags applied to the fish mercury 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 2.6. No results were
qualified as invalid, thus all results are represented in the analysis offish mercury concentrations
presented in this report.
Table 8-4. Summary of Routine Field Sample Flags for Fish Mercury
Flag
EHT, Exceeded Holding Time
FBS, Failed Blank Sample
FDL, Failed Lab Duplicate
FMS, Failed Matrix Spike
FRS, Failed Lab Reference Sample
FSR, Failed Standard Reference Material
Number of QC Samples
—
44 lab reagent blank samples
153 lab duplicate groups
9 lab matrix spike samples
24 lab reference samples
24 standard reference material
samples
Percentage of Samples Flagged (%)
0.5% (1)
0
5% (10)
0
0
1%(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.
Few data quality flags were applied to fish mercury data. Of the 201 routine field samples analyzed for
mercury, only 1 sample was flagged for exceeding sample holding time, 10 samples were flagged for
failed laboratory duplicates, and 2 samples were flagged for a failed standard reference material. The one
sample that was flagged for sample holding time exceeded the 1095-day criterion by 3 days. The average
holding time for analyzed samples was 680 days.
Field duplicate samples could not be collected for the fish matrix, because individual fish are not expected
to contain identical mercury concentrations. Laboratory duplicate samples, however, were prepared by
subsampling collected fish samples. Of the 153 laboratory duplicate groups that were analyzed, only 10
exceeded the MQO of 25% relative percent difference (RPD). RPDs for these failed duplicate samples
ranged from 25.2% to 31.3%.
A total of 44 laboratory reagent blanks were analyzed to assess the potential for contamination of routine
field samples. All of these samples contained less than 1 ng mercury, so no samples were flagged for
failed laboratory reagent blanks. Blank sample results ranged from 0 to 0.92 ng, which is more than 4
times below the lowest sample result of 4.1 ng. This indicates no significant contamination of routine
field samples.
To evaluate the bias of analytical results, the laboratory analyzed matrix spike samples, laboratory
reference samples that consisted of previously analyzed Lake Erie fish, and standard reference materials
(SRM) from the National Institute of Standards and Technology. Two SRMs were used for this study:
SRM 1566a, an oyster tissue sample with a certified value of 0.0642 mg/kg (no longer available) and
SRM 1515, apple leaves, with a certified value 0.044 mg/kg.
No samples were flagged for failed matrix spikes or laboratory reference samples. Recoveries for matrix
spike samples ranged from 82% to 109%. Recoveries for laboratory reference samples ranged from 90%
8-6
-------
Mercury in Fish
to 115%. Only one standard reference material sample, which was associated with two routine field
samples, was flagged for recovery beyond the MQO of 80-120%. This sample achieved a recovery of
133%. Based on the analysis of laboratory matrix spike samples, laboratory reference samples, standard
reference materials, laboratory reagent blank samples, and other internal QC data, the QC coordinator did
not qualify any samples as high or low biased.
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 groups. Table 8-5 provides a
summary of data quality assessments for several of these attributes. The results of laboratory duplicate
samples revealed good analytical precision for fish data. The mean RPD for laboratory duplicate samples
was 11.7%.
Table 8-5. Data Quality Assessment for Mercury in Fish Samples
Parameter
Number of Routine Samples Analyzed
Analytical Precision, Mean Lab Duplicate RPD (%), >MDL
Analytical Bias, Mean SRM (%)
Analytical Bias, Mean IMS (%)
Analytical Bias, Mean LRS (%)
Analytical Sensitivity, Samples reported as
-------
Results of the LMMB Study: Mercury Data Report
Figure 8-4 shows the percentages of coho salmon and lake trout from the LMMB Study that fall into each
of the advisory categories recommended by EPA for methylmercury contamination (USEPA, 2000).
Since methylmercury was not measured in fish during the LMMB Study, samples were assigned to each
category based on the conservative assumption that 100% of total mercury was in the form of
methylmercury. Only 3% and 9% of lake trout and coho salmon, respectively, fell into the unrestricted
consumption category. The most contaminated coho salmon and lake trout specimens collected in the
LMMB Study fell into the 4 meals/month and 2 meals/month restriction categories, respectively. For the
average coho salmon sample, EPA guidance would recommend restricting consumption to 12 meals per
month; and for the average lake trout sample, EPA guidance would recommend restricting consumption
to 4 meals per month. This recommendation is consistent with state-wide advisories for mercury that
have been issued by several states. For instance, Illinois has placed a state-wide methylmercury advisory
of one meal per week of predator fish to protect sensitive populations (women of childbearing age and
children). While Lake Michigan fish mercury concentration warrants some level offish advisory, few
fish advisories in Lake Michigan have been based solely on mercury contamination, because Lake
Michigan waters are generally under more stringent fish advisories based on PCB contamination.
Figure 8-4. Percentage of Lake Michigan Coho Salmon and Lake Trout Samples within each EPA-
Recommended Fish Advisory Category
40 -,
35 -
(A
-? 30 -J
o
0)
O)
TO
4-*
0)
o
0)
D.
20 -
15 -
10 -
5 -
0 J
rjCoho Salmon
• Lake Trout
6
CO
6 £
CD ^-
D) -t
c J2
o ro
-------
Mercury in Fish
reported an average mercury concentration of 120 ng/g in lake trout collected in 1988. Borgmann and
Whittle (1991) also reported that mercury concentrations in Lake Ontario lake trout had decreased
steadily to this level from an average of 240 ng/g in 1977. Cappon (1984) measured similar total mercury
levels in Lake Ontario lake trout fillets, but much higher concentrations in coho salmon fillets. Mercury
concentrations in lake trout fillets ranged from 160 to 290 ng/g and averaged 230 ng/g. Mercury
concentrations in coho salmon fillets ranged from 220 to 800 ng/g and averaged 420 and 460 ng/g in two
separate fillet cross-sections.
Mercury concentrations of top predators from Lake Michigan were generally lower than those from
smaller inland lakes. In a 1999 EPA report on fish mercury data from 1990 to 1995, the weighted mean
concentration of mercury in walleye from lakes across Michigan was 375 ng/g (USEPA, 1999b). In a
survey of 80 northern Minnesota lakes, Sorensen et al. (1990) measured an average mercury
concentration of 450 ng/g (range 140 to 1500 ng/g) in a standard 550 mm northern pike. Rose et al.
(1999) measured an average mercury concentration of 390 ng/g in largemouth bass from 24 lakes in
Massachusetts. In a study of 219 Wisconsin lakes, average concentrations of mercury in 450 to 500 mm
walleye ranged from 390 to 830 ng/g, depending upon the acid neutralizing capacity of the lakes (Lathrop
etal., 1991).
Mercury concentrations in forage fish species were not analyzed in the LMMB Study, so mercury
biomagnification within the upper pelagic food web could not be documented. Mercury concentrations
measured in top predator species during the LMMB Study, however, were higher than for forage fish
species measured by other researchers. Brazner and DeVita (1998) measured mercury concentrations of
9.4 to 31 ng/g in young-of-the-year yellow perch from Green Bay. Mercury concentrations in young-of-
the-year spottail shiners from Green Bay ranged from 10.5 to 33.5 ng/g. These concentrations are from 2
to 15 times lower than average mercury concentrations measured in top predators. Similarly, Borgmann
and Whittle (1992) measured mercury levels of 37 ng/g and 32 ng/g in 1988 from Lake Ontario smelt and
slimy sculpin, respectively.
8.3.3 Factors Affecting Contaminant Concentrations
In the LMMB Study, fish mercury concentrations varied primarily by species and by fish length. Lake
trout contained significantly more mercury than coho salmon, and for both species, mercury content
increased with fish length. Regression equations to describe mercury content based on the length of Lake
Michigan lake trout and coho salmon were calculated, with r2 values of 0.856 and 0.824, respectively.
This correlation with fish length has been well documented and is the basis for size-specific fish
advisories. Higher mercury concentrations are accumulated in larger fish because these fish are generally
older and have experienced longer exposure durations to environmental concentrations, giving them more
time to accumulate pollutants that are not easily degraded or eliminated.
In investigating fish mercury levels in a wide variety of lakes, researchers have identified other lake-
specific factors that influence mercury concentrations in fish. Sorensen et al. (1990) found that mercury
levels in northern pike from Minnesota lakes were correlated with mercury in water, mercury in
zooplankton, total organic carbon, iron, and pH (negative correlation). In a study of 219 Wisconsin lakes,
concentrations of mercury in walleye increased with increasing fish length and with decreasing acid
neutralizing capacity (Lathrop etal., 1991). Mean mercury concentrations ranged from 180 ng/g in the
smallest walleye (250 to 349 mm) from high acid neutralizing capacity lakes (>1500 |ieq/L) to 1470 ng/g
in the largest walleye (>650 mm) from low acid neutralizing capacity lakes (<100 |^eq/L). Rose et al.
(1999) measured fish mercury levels in 24 Massachusetts lakes. Mercury concentrations in top predators
(largemouth bass) were positively associated with fish weight, lake size, and watershed characteristics.
Lake pH was not correlated with mercury concentrations in largemouth bass, but was correlated with
mercury concentrations in brown bullhead and yellow perch.
8-9
-------
Chapter 9
Cross-Media Interpretations
9.1 Summary of Mercury Concentrations in Lake Michigan Compartments
Mercury was found throughout the Lake Michigan ecosystem, with concentrations measured in air, water,
sediment, tributaries, plankton, and fish samples collected from in and around the lake. Mercury was
found in the majority of samples at levels above the corresponding detection limit for each ecosystem
compartment (Table 9-1). Other than one sediment sample and one plankton sample, total mercury was
detected in every sample in all media other than the open-lake water column. A total of 8 particulate
mercury samples and 4 total mercury samples in the open-lake water column did not contain detectable
levels of mercury. Comparisons of these frequencies should be done with care, due to the different types
of detection limits used in the different mercury data sets. The type of detection limit used for the
atmospheric phase or analytical fraction was described by the PI responsible for the analyses. Samples
were only analyzed for methylmercury in the tributary compartment, and methylmercury was detected at
levels above the MDL for the majority of the tributary samples. Approximately 15% and 3% of the
dissolved and total samples, respectively, did not have detectable levels of methylmercury.
Table 9-1. Summary of Samples from each Ecosystem Compartment with Detectable Levels of Mercury
Ecosystem
Compartment
Atmosphere
Tributary
Tributary
Methylmercury
Open Lake
Sediment
Plankton
Fish
Atmospheric Phase or
Analytical Fraction
Vapor
Particulate
Precipitation
Dissolved
Total
Dissolved
Total
Particulate
Total
Total
Total
Total
Detection Limit Type
System Detection Limit
System Detection Limit
Method Detection Limit
Method Detection Limit
Method Detection Limit
Method Detection Limit
Method Detection Limit
Daily Detection Limit
Daily Detection Limit
Sample-Specific Detection Limit
Sample-Specific Detection Limit
Method Detection Limit
% Samples with Mercury Above
Detection Limit
100%
100%
100%
100%
100%
85%
97%
92%
96%
99.5%
99%
100%
Vapor-phase mercury concentrations averaged from 2.06 to 3.62 ng/m3 at five different shoreline and out-
of-basin stations. The highest concentrations of vapor-phase mercury were detected at the IIT Chicago
station, at the southern end of Lake Michigan. Particulate-phase mercury concentrations were lower than
vapor-phase concentrations, with means ranging from 12.1 pg/m3 to 73.7 pg/m3. At individual stations,
the mean vapor-phase concentration was 49 to 175% greater than the mean particulate-phase
concentration. As with the vapor phase, the higher particulate-phase mercury concentrations were found
at the IIT-Chicago site. Mean precipitation-phase mercury concentrations ranged from 15.2 to 26.1 ng/L.
When calculated by weighting the concentrations according to the sample volume, mean precipitation-
phase mercury concentrations ranged from 11.0 to 21.1 ng/L. As with the other two atmospheric phases,
the highest mean concentration in the precipitation phase was measured at the IIT Chicago station.
Total mercury concentrations in Lake Michigan tributaries averaged from 1.07 to 28.9 ng/L, and
dissolved mercury concentrations averaged from 0.666 to 3.71 ng/L. When calculated using the
9-1
-------
Results of the LMMB Study: Mercury Data Report
differences between total mercury and dissolved mercury concentrations in individual samples, mean
particulate mercury concentrations averaged from -0.0058 to 25.8 ng/L. In all cases, the highest mean
concentrations were measured in the Fox River, a tributary that empties into Green Bay. The Fox River
watershed is highly industrialized, and is suspected of being impacted by resuspension of contaminated
sediments from legacy sources (Hurley etal, 1998a; Rossmann and Edgington, 2000). Generally, among
the other tributaries, mercury levels were higher in more urban/industrialized areas, and lower in
primarily agricultural/forested areas.
Within the open-lake water column, total mercury concentrations averaged from 0.25 to 0.78 ng/L, and
particulate mercury concentrations averaged from 0.029 to 0.17 ng/L. Generally, mercury was well
mixed in the water column, as there was little variability in concentration among stations. While there
was a slightly greater concentration of particulate mercury in Green Bay, there was no corresponding
increase of total mercury.
The mean mercury concentration measured in precipitation samples was approximately 2.6 times greater
than the mean total mercury concentration measured in the tributaries. With the exception of the Fox
River tributary, all mean precipitation-phase mercury concentrations were greater than the mean total
mercury concentration at any tributary. The mean mercury concentration in the Fox River was greater
than the mean concentration in precipitation at any of the atmospheric stations. The overall mean
precipitation-phase and tributary concentrations were 64 and 24 times greater than the mean total mercury
concentration in the water column, respectively.
Total mercury concentrations measured in surficial sediments ranged from 0.002 to 0.26 mg/kg. Higher
levels of mercury tended to accumulate in the sediments in deeper locations of the lake. Net fluxes of
mercury ranged from 0.85 to 21 ng/cm2/y and were highest along the eastern shore in response to the
dominant water currents in the lake. Additional samples were collected from five different sediment trap
stations, including two which were set at two different depths. Mean mercury concentrations for the
different trap stations and depths ranged from 0.21 to 28 mg/kg. The highest mercury concentrations
were found in samples collected from traps located in the southern basin of Lake Michigan. Both
mercury concentrations and fluxes to surficial sediments have decreased since the 1970s.
9.2 Mercury Speciation
As discussed in Section 2.1.6 of this report, the organic compounds methylmercury and dimethylmercury
have a greater toxicity than inorganic mercury, given equivalent doses. Methylmercury is generally the
dominant form of mercury in higher levels of the aquatic food web. Methylmercury usually forms
through methylation of inorganic mercury by bacteria in sediments or in the water column. Therefore,
although atmospheric deposition and tributary flows are major sources of inorganic mercury to the lake,
they may not be major sources of methylmercury.
Among the ecosystem components in Lake Michigan from which mercury samples were collected,
methylmercury samples were collected from only the tributary component. While total mercury levels
were greatest from the Fox River, and other tributaries located near urban/industrial sources, this was not
the case for methylmercury. Tributaries located in mostly agricultural and forested areas such as the
Menominee and Muskegon rivers had among the highest methylmercury concentrations. The Grand
Calumet River, which had one of the highest mean total mercury concentrations and is located near the
Chicago/Gary urban area, had the lowest mean methylmercury concentration.
The relative contribution of methylmercury to the total mercury concentrations measured in the tributaries
was evaluated by calculating the percentage of the mean methylmercury concentration to the mean total
9-2
-------
Cross-Media Interpretations
mercury concentration in each tributary. The percentage of mean methylmercury concentration to mean
total mercury concentration ranged from 0.48% to 21% in the 11 tributaries. The 21% figure was for the
Muskegon River, and the percentage contributions for the other 10 tributaries were all less than 6%.
These lower percentages are consistent with other estimates of the contribution of methylmercury to total
mercury in the water column (USEPA, 1997b), which indicate that methylmercury constitutes less than
10% of the total mercury concentration in water samples.
The percentage of methylmercury is greater in plankton and fish than in water samples. For example,
Watras and Bloom (1992) measured both methylmercury and total mercury in various trophic levels in a
basin of the Little Rock Lake, Wisconsin. Little Rock Lake is in north-central Wisconsin, in a relatively
remote area with no industrial activity, and with restricted public-access. The lake is fed by groundwater
and is used as an experimental lake. The lake has been artificially divided into two basins, one of which
is acidified relative to the rest of the lake.
Watras and Bloom (1992) found that the percentage of methylmercury to total mercury in the water
column and biota varied with pH as shown in Table 9-2, below.
Table 9-2. Percent of Mercury Attributable to Methylmercury in Little Rock Lake
Ecosystem Component
Water column
Phytoplankton
Zooplankton
Fish
% Methylmercury of the Total Mercury
Reference Basin (pH = 6.1)
5
13
29
>90
Acidified Basin (pH = 4.7)
12
31
91
>90
Both basins of the experimental lake were acidic, with a pH of 4.7 in the acidified basin, and 6.1 in the
reference basin. In contrast, the mean pH measured in Lake Michigan for the LMMB Study was 8.2.
Mason and Sullivan (1997) and Sullivan and Mason (1998) reported methylmercury concentrations in
Lake Michigan that ranged from the detection limit of 5 pg/L to 42 pg/L, with an epilimnetic mean of 6
pg/L for August 1994 and 8.2 for October/November of 1994. (When calculating the mean, the detection
limit of 5 pg/L was substituted for any sample result below the detection limit.) The hypolimnetic mean
for August 1994 was 8 pg/L; whereas, the hypolimnetic concentrations for two samples in October/
November 1994 were 17 and 42 pg/L. These concentrations represent 2-3% of the mean total mercury
concentration (Sullivan and Mason, 1998), which are lower than those reported by Watras and Bloom
(1992).
A subsequent study of small lakes in northern Wisconsin by Watras et al. (1998) included the two basins
of Little Rock Lake and 13 others lakes. The majority of these lakes are precipitation-domination seepage
lakes in which the flows are dominated by precipitation, rather than riverine flow. Watras et al. (1998)
measured the concentrations of total mercury, dissolved mercury, total methylmercury, and dissolved
methylmercury in samples of dissolved organic carbon (DOC), microseston, zooplankton, and small fish.
The microseston in the lakes in that study primarily consists of phytoplankton, bacterioplankton, and
cellular debris. The zooplankton were collected in 153-^m mesh nets (a slightly larger mesh than used in
the LMMB Study). It total, 727 yellow perch (Perca flavescens) and 139 golden shiners (Notemigonus
crysoleucas) were collected during the spring and summer of 1994, ranging from one to seven years in
age. Total mercury and total methylmercury were also measured in surficial sediments collected from
these lakes.
Watras et al. (1998) reported the percentage of methylmercury relative to the total mercury concentration
in the DOC, microseston, zooplankton, and small fish, as shown in Table 9-3. The mean pH of the lakes
9-3
-------
Results of the LMMB Study: Mercury Data Report
on the study was 6.25, slightly higher than the reference basin in Little Rock Lake, and still well below
the mean pH of 8.2 in the LMMB Study.
Table 9-3. Percent of Mercury Attributable to Methylmercury in 15 Lakes in Northern Wisconsin
Ecosystem Component
Dissolved organic carbon
Microseston (includes phytoplankton)
Zooplankton
Fish
% Methylmercury of the Total Mercury
11%
18%
57%
95%
Except for the fish, the percentages of methylmercury in Table 9-3 are intermediate to the results for the
two basins on Little Rock Lake shown in Table 9-2. The results for the fish are comparable to those in
Table 9-2, where the fish are listed as ">90%."
There has been one report of methylmercury in Lake Michigan sediments. Rossmann et al. (2001)
reported methylmercury results for surficial sediments samples from Lake Michigan that were originally
collected in 1994 - 1996 as part of the LMMB, but not analyzed for methylmercury as part of this study.
The methylmercury concentrations ranged between 0.16 and 1.7 ng/g, with a mean and median of 0.57
and 0.45 ng/g, respectively. The methylmercury concentration varied between 0.11 and 1.4% of the total
mercury concentration. The mean and median fraction of methylmercury were 0.42 and 0.35%,
respectively.
The results from Watras et al. (1998) for methylmercury in surficial sediments of the lakes in northen
Wisconsin are slightly higher than those from Rossmann et al. (2001), with a range of 0.5 to 7.4 ng/g,
with a mean of 2.6 ng/g. The methylmercury concentration varied between 0.5 and 3.9% of the total
mercury concentration, with a mean fraction of methylmercury of 1.5%.
Studies comparing methylmercury and total mercury levels in fish have consistently shown that the
majority of the measured total mercury consists of methylmercury. Herrin et al. (1998) measured
mercury in bluegill and shiners in Devil's Lake, Wisconsin, and found that methylmercury accounted for
nearly all of the total mercury in both species. However, they also found that methylmercury accounted
for 26% to 58% of total mercury in open water, higher than most estimates. Rossmann et al. (2003)
reported mean total and methylmercury concentrations in forage fish to be 0.051 and 0.34 mg/kg,
respectively with methylmercury concentrations accounting for 60 and 91% of the total mercury for
various species. Francis et al. (1998) also measured methylmercury and total mercury in various fish
species in an estuary of Lake Erie. While mercury concentrations were frequently below detection limits,
the percentages attributable to methylmercury were usually greater than 90% in common carp and
channel catfish.
Unlike the two studies described above, Cappon (1984) measured methylmercury and total mercury in
lake trout and coho salmon in Lake Ontario, allowing greater comparability with the LMMB Study. On
average, methylmercury accounted for 71% of total mercury in both lake trout and coho salmon, a much
lower percentage than those observed in the other studies. The levels of total mercury in lake trout in
Lake Ontario were slightly higher, with a mean of approximately 165 ng/g on a wet-weight basis,
compared to 139 ng/g in the LMMB Study. The levels of total mercury in Lake Ontario, however, were
much higher, with a mean of approximately 240 ng/g, compared to 69 ng/g in the LMMB Study.
Therefore, it is unclear whether the percentages of methylmercury from the study in Lake Ontario were
unusually low due to taxonomic differences, or due to unusually high total mercury results.
9-4
-------
Cross-Media Interpretations
9.3 Bioaccumulation and Biomagnification
Mean mercury concentrations in the biota and mean concentrations in the water column and surficial
sediments are presented in Figure 9-1. Within living components of the Lake Michigan ecosystem,
mercury accumulated at concentrations higher than in any abiotic ecosystem component, with the
exception of surficial sediments. Bioaccumulation factors for mercury ranged from 1.1 x 105 in
phytoplankton to 1.1 x 106 in lake trout.
Figure 9-1. Mercury Concentrations in Various Components of the Lake Michigan
Ecosystem
W 1 \J\J\J
®~ 1 00
Q.
c m
.2
"'S .
CO 1
2 i
+j
Si n 1
o
5 n m
o
> n nni
3
P n nnm
| 1
0)
*r
In addition to accumulating in living tissue at concentrations above those in the water, mercury also was
magnified within the Lake Michigan food web (Figure 9-2). Total mercury concentrations increased from
35 ng/g in phytoplankton to 55 ng/g in zooplankton, a factor of 1.55. While samples of forage fish were
not initially analyzed for mercury in the LMMB, an approximate two-step biomagnification factor can be
calculated between zooplankton and the predator fish. The mean dry-weight mercury concentration in
adult coho was 248 ng/g, and the mean mercury concentration in lake trout in adult coho was 373 ng/g.
These concentrations correspond to biomagnification factors of 4.57 and 6.88, compared to zooplankton,
respectively. From the bottom of the food web (phytoplankton) to the top of the food web (lake trout),
mercury concentrations increase by a factor of 10.7.
9-5
-------
Results of the LMMB Study: Mercury Data Report
Figure 9-2. Biomagnification Factors for Mercury in Lake Trout (A) and Adult Coho (B)
(A)
500
O)
PHYTOPLANKTON
ZOOPLANKTON
TROUT
(B)
500
PHYTOPLANKTON
ZOOPLANKTON
ADULT COHO
Because forage fish samples were not analyzed for mercury, biomagnification of mercury in the upper
pelagic food web could not be estimated and compared to that calculated in total PCBs and trans-
nonachlor (USEPA, 2003). However, biomagnification factors between zooplankton and predator fish
species were much lower than those calculated for PCBs and fra«s-nonachlor. For total PCBs the
biomagnification factor between zooplankton and My sis was 1.5, and the factor between My sis and lake
trout was 31, yielding an estimated factor of 46.5 between zooplankton and lake trout. Similarly, the
biomagnification factor between zooplankton and My sis for fra«s-nonachlor was 1.6 and the factor
between My sis and lake trout was 19, yielding an estimated factor of 30.4 between zooplankton and lake
trout. These factors were much larger than the corresponding factor of 6.88 for mercury. The
biomagnification factor between phytoplankton and zooplankton was also smaller for mercury, at 1.55,
compared to 3.4 for total PCBs and 9.5 for fra«s-Nonachlor.
9-6
-------
Cross-Media Interpretations
The biomagnification of mercury in Lake Michigan occurred at a higher rate compared to previous
studies in other lakes. For example, Herrin et al. (1998) measured a biomagnification factor for mercury
of 2.7 between Daphnia and bluegills in Devil's Lake, Wisconsin in 1994 and 1995. The mean dry-
weight total mercury concentrations measured in bluegills in Devil's Lake were 575 ng/g in 1994 and 324
ng/g in 1995. These levels are comparable with the mean of 373 ng/g measured in Lake Michigan in this
study. The smaller biomagnification factor from Devil's Lake was likely due to greater mercury
concentrations in the Daphnia compared to the zooplankton in Lake Michigan and the fact that bluegills
are a step lower in the food chain than lake trout.
While mean total mercury concentrations were not reported for Daphnia in Devil's Lake, the mean
methylmercury concentrations of 186 and 100 ng/g in 1994 and 1995 were 3.38 and 1.82 times greater
than the mean total mercury concentration in zooplankton in Lake Michigan. The open-water total
mercury concentrations were also greater in Devil's Lake, with a mean of 3.0 ng/L total mercury, almost
an order of magnitude greater than the mean concentration observed in Lake Michigan. In addition, the
biomagnification factors calculated in Devil's Lake were based on methylmercury, not total mercury.
While the total mercury and methylmercury levels were comparable in that study, total mercury levels
may have been considerably greater in Daphnia than the measured methylmercury concentrations, which
would yield a biomagnification factor which would be greater than one calculated based on total mercury.
For example, Watras and Bloom (1992) found that 29% of zooplankton mercury in Little Rock Lake, WI
was methylmercury; whereas >90% of total mercury measured in fish was methylmercury. The
comparability of the two studies may also be affected by the taxonomic differences of the sampled fish.
Bluegills tend to be smaller than trout and will likely be lower on the food web than lake trout. In an
EPA survey of mercury concentrations offish (USEPA, 1999b), bluegills were found to have lower
concentrations than most other fish species from which samples were collected, including largemouth
bass, walleye and northern pike. Bluegill caught in Wisconsin for this survey had comparable mercury
concentrations to those measured in Devil's Lake.
Francis et al. (1998) also found evidence of mercury biomagnification in Old Woman Creek, an estuary
of Lake Erie. However, bioaccumulation and biomagnification factors could not be calculated, due to the
prevalence of open-water, plankton, and fish samples without detectable levels of mercury. The detection
limits reported in that study were higher than those for the LMMB Study, by up to two orders of
magnitude in water and fish tissue samples. While mercury was also not detected in zooplankton
samples, the detection limits for plankton samples in the two studies were comparable. However, the
authors did conclude that bioaccumulation was occurring, based on higher levels of mercury, and greater
rates of detection, in predatory catfish and bowfin.
9-7
-------
References
Back, R. C. and C. J. Watras. 1995. Mercury in zooplankton of northern Wisconsin lakes: taxonomic
and site-specific trends. Water, Air, and Soil Pollution. 80: 931-938.
Balogh, S.J., M.L. Meyer, and D. K. Johnson. 1997. Mercury and suspended sediment loadings in the
lower Minnesota River. Environmental Science and Technology. 31: 198-202.
Balogh, S.J., M.L. Meyer, and D. K. Johnson. 1998. Transport of mercury in three contrasting river
basins. Environmental Science and Technology. 32:456-462.
Bloom, N.S. and W.F. Fitzgerald. 1988. Determination of Volatile Mercury Species at the Picogram
Level by Low-Temperature Gas Chromatography with Cold-Vapor Atomic Fluorescence
Detection. Anal. Chim. Acta. 208: 151.
Bloom, N. and S.W. Effler. 1990. Seasonal variability in the mercury speciation of Onondaga Lake
(New York). Water, Air and Soil Pollution. 53:215-265.
Borgmann, U. and D. M. Whittle. 1991. Contaminant concentration trends in Lake Ontario lake trout
(Salvelinus namaycush): 1977 to 1988. Journal of Great Lakes Research. 17(3): 368-381.
Borgmann, U. and D. M. Whittle. 1992. DDE, PCB, and mercury concentration trends in Lake Ontario
rainbow smelt (Osmerus mordax) and slimy sculpin (Coitus cognatus}: 1977 to 1988. Journal of
Great Lakes Research. 18(2): 298-308.
Brazner, J. and W. DeVita. 1998. PCBs, DDE, and mercury in young-of-the-year littoral fishes from
Green Bay, Lake Michigan. Journal of Great Lakes Research. 21(1): 83-92.
Cahill, R. A. 1981. Geochemistry of Recent Lake Michigan Sediments. Illinois Institute of Natural
Resources, State Geological Survey Division, Circular 517, 94 pp.
Cappon, C. J. 1984. Content and chemical form of mercury and selenium in Lake Ontario salmon and
trout. Journal of Great Lakes Research. 10(4): 429-434.
Eadie, B. J. June 1997a. Quality assurance plan for the use of sediment traps, pp. 319-324. In USEPA
Lake Michigan Mass Balance Study (LMMB) Methods Compendium, Volume 3, Metals,
Conventionals, Radiochemistry, and Biomonitoring Sample Analysis Techniques, EPA-905-R-97-
012c, Great Lakes National Program Office and Office of Water.
Eadie, B. J. June 1997b. Trap sample splitting (wet): use of sediment traps for the measurement of
particle and associated contaminant fluxes, pp. 245-249. Lake Michigan Mass Balance Study
(LMMB) Methods Compendium, Volume 1, Sample Collection Techniques, EPA-905-R-97-
012a, Great Lakes National Program Office and Office of Water.
Eadie, B. J., D. J. Schwab, R. A. Assel, N. Hawley, M. B. Lansing, G. S. Miller, N. R. Morehead, and J.
A. Robbins. 1996. Development of recurrent coastal plume in Lake Michigan observed for first
time. EOS, Transaction, American Geophysical Union. 77:337-338.
R-1
-------
Results of the LMMB Study: Mercury Data Report
Edgington, D. N. and J. A. Robbins. June 1997a. Standard operating procedures for collection of
sediment samples, pp. 239-244. Lake Michigan Mass Balance Study (LMMB) Methods
Compendium, Volume 1, Sample Collection Techniques, EPA-905-R-97-012a, Great Lakes
National Program Office and Office of Water.
Edgington, D. N. and J. A. Robbins. June 1997b. Determination of the activity of lead-210 in sediments
and soils, pp. 341-345. Lake Michigan Mass Balance Study (LMMB) Methods Compendium,
Volume 3, Metals, Conventional, Radiochemistry, and Biomonitoring Sample Analysis
Techniques, EPA-905-R-97-012c, Great Lakes National Program Office and Office of Water.
Francis, D. R., D. J. Jude, and J. A. Barres. 1998. Mercury distribution in the biota of a Great Lakes
estuary: Old Woman Creek, Ohio. Journal of Great Lakes Research. 24(3): 595-607.
Glass, G. E., E. N. Leonard, W. H. Chan, and D. B. Orr. 1986. Airborne mercury in precipitation in the
Lake Superior region. Journal of Great Lakes Research. 12(1): 37-51.
Herrin, R. T., R. C. Lathrop, P. R. Gorski, and A. W. Andren. 1998. Hypolimnetic methylmercury and
its uptake by plankton during fall destratification: a key entry point of mercury into lake food
chains? Limnology and Oceanography. 43(7): 1476-1486.
Hoyer, M., J. Burke and G. Keeler. 1995. Atmospheric sources, transport and deposition of mercury in
Michigan: two years of event precipitation. Water, Air, and Soil Pollution. 80, 199-208.
Hurley, J.P., J. M. Benoit, C.L. Babiarz, M.M. Shafer, A.W. Andren, J. R. Sullivan, R. Hammond, and
D.A. Webb. 1995. Influences of watershed characteristics on mercury levels in Wisconsin
rivers. Environmental Science and Technology. 29: 1867-1875.
Hurley, J.P., S. E. Cowell, M.M. Shafer, and P.E. Hughes. 1998a. Partitioning and transport of total and
methyl mercury in the lower Fox River, Wisconsin. Environmental Science and Technology. 32:
1424-1432.
Hurley, J.P., S. E. Cowell, M.M. Shafer, and P.E. Hughes. 1998b. Tributary loading of mercury to Lake
Michigan: importance of seasonal events and phase partitioning. The Science of the Total
Environment. 213: 129-137.
Johnson, C. 1999. Elemental mercury use in religious and ethnic practices in New York City.
Population and Environment. 20(5): 443-453.
Keeler, G. 1994. Lake Michigan Urban Air Toxics Study. Funded by USEPA Atmospheric Research
and Exposure Assessment Laboratory. EPA/600/SR-94/191.
Keeler, G., G. Glinsorn, and N. Pirrone. 1995. Particulate mercury in the atmosphere: its significance,
transport, transformation and sources. Water, Air, and Soil Pollution. 80: 159-168.
Kennedy, J., R. R. Ruch, andN. F. Shimp. 1971. Distribution of mercury in unconsolidated sediments
from southern Lake Michigan. Illinois State Geological Survey, Environmental Geology Notes
Number 44, Studies of Lake Michigan Bottom Sediments Number Seven, 18 pp.
Kirkwood, A. E., P. Chow-Fraser, and G. Mierle. 1999. Seasonal mercury levels in phytoplankton and
their relationship with algal biomass in two dystrophic shield lakes. Environmental Toxicology
andChemistry. 18(3): 523-532.
R-2
-------
References
Landis, M.S., A.F. Vette, and G. J. Keeler. 2002. Atmospheric mercury in the Lake Michigan basin:
influence of the Chicago/Gary urban area. Environmental Science and Technology. 36(21),
4508-4517.
Lathrop, R. C., P. W. Rasmussen, and D. R. Knauer. 1991. Mercury concentrations in walleyes from
Wisconsin (USA) lakes. Water, Air, and Soil Pollution. 56: 295-307.
Leeman Labs, Inc. 1991. PS200 Automated mercury analyzer set-up and operating manual version of
November 1991.
Leeman Labs, Inc. 1993. AP'200 automated mercury preparation system manual. Revision C (11/20/93).
Mason, R. P., W. F. Fitzgerald, and M. M. Morel. 1994. The Biogeochemical Cycling of Elemental
Mercury: Anthropogenic Influences. Geochimica Cosmochimica Acta. 58(15): 3191-3198.
Mason, R. P., N.M. Lawson and K. A. Sullivan. 1997. The concentration, speciation and sources of
mercury in Chesapeake Bay precipitation. Atmospheric Environment. 31(21): 3541-3550.
Mason, R. P. and K. A. Sullivan. 1997. Mercury in Lake Michigan. Environmental Science and
Technology. 31: 942-947.
Mason, R. P., N.M. Lawson and G. R. Sheu. 2000. Annual and seasonal trends in mercury deposition in
Maryland. Atmospheric Environment. 34: 1691-1701.
MDN, 1999. Mercury deposition network National Atmospheric Deposition Program,
http://nadp.sws.uiuc.edu/mdn.
Monson, B. A. and P. L. Brezonik. 1998. Seasonal patterns of mercury species in water and plankton
from softwater lakes in northeastern Minnesota. Biogeochemistry. 40: 147-162.
Nierenberg, D. W. , R. E. Nordgren, M. B. Chang, R. W. Siegler, M. B. Blayney, F. Hochberg, T. Y.
Toribara, E. Cernichiari, and T. Clarkson. 1998. Delayed Cerebellar disease and death after
accidental exposure to dimethylmercury. New England Journal of Medicine. 338: 1672-1676.
NIST, 1990. Certificate of Analysis for Standard Reference Material 2704 Buffalo River Sediment.
National Institute of Standards and Technology, Gaithersburg, MD. 6 pp.
Pirrone, N., I. Allegrini, G. J. Keeler, J. O. Nriagu, R. Rossmann, and J. A. Robbins. 1998. Historical
records of mercury pollution in North America. Atmospheric Environment. 32(5): 929-940.
Riley, D. M., A. Newby, T. O. Leal-Almeraz, and V. M. Thomas. 2001. Assessing Elemental Mercury
Vapor Exposure from Cultural and Religious Practices. Environmental Health Perspectives. 109
(8): 779-784.
Robbins, J. A. and D. N. Edgington. June 1997. Protocol for standard analysis for cesium-137, pp. 337-
340. Lake Michigan Mass Balance Study (LMMB) Methods Compendium, Volume 3, Metals,
Conventionals, Radiochemistry, and Biomonitoring Sample Analysis Techniques, EPA-905-R-97-
012c, Great Lakes National Program Office and Office of Water.
R-3
-------
Results of the LMMB Study: Mercury Data Report
Rohrer, T. K., J. C. Forney, and J. H. Hartig. 1982. Organochlorine and heavy metal residues in standard
fillets of coho and chinook salmon of the Great Lakes - 1980. Journal of Great Lakes Research.
8(4): 623-634.
Rose, J., M. S. Hutcheson, C. R. West, O. Pancorbo, K. Hulme, A. Cooperman, G. DeCesare, R. Isaac,
and A. Screpetis. 1999. Fish mercury distribution in Massachusetts, USA lakes. Environmental
Toxicology and Chemistry. 18(7): 1370-1379.
Rossmann, R. September 1996. Quality Assurance Plan, Lake Michigan Mass Balance Project -
Mercury in Sediments. USEPA Community-Based Science Support Staff, Grosse lie, MI. 14 pp.
Rossmann, R. 1999. Horizontal and vertical distributions of mercury in 1983 Lake Superior sediments
with estimates of storage and mass flux. Journal of Great Lakes Research. 25(4): 683-696.
Rossmann, R. and J. Barres. 1988. Trace element concentrations in near-surface waters of the Great
Lakes and methods of collection, storage and analysis. Journal of Great Lakes Research. 14(2):
188-204.
Rossmann, R. and D. N. Edgington. 2000. Mercury in 1987-1990 Green Bay, Lake Michigan surficial
sediments. Journal of Great Lakes Research. 26(3): 323-339.
Rossmann, R., J. C. Filkins, and M. Levine. 2003. Total and methyl mercury concentrations in 1994-5
Lake Michigan forage fish. 46th Conference on Great Lakes Research, International Association
for Great Lakes Research, Chicago, Illinois, June 23-26, 2003. pp. 14-15 of abstracts.
Rossmann, R., K. R. Rygwelski, and J. C. Filkins. 2001. Methyl mercury in Lake Michigan surficial
sediments. Presented at the 44th Conference on Great Lakes Research, International Association
for Great Lakes Research, Green Bay, Wisconsin, June 10-14, 2001. p. 113 of abstracts.
Schroeder, W. H. and J. Markes. 1994. Measurements of atmospheric mercury concentrations in the
Canadian environment near Lake Ontario. Journal of Great Lakes Research. 20(1): 240-259.
Sorensen, J. A., G. E. Glass, K. W. Schmidt, J. K. Huber, and G. R. Rapp, Jr. 1990. Airborne mercury
deposition and watershed characteristics in relation to mercury concentrations in water,
sediments, plankton, and fish of eighty northern Minnesota lakes. Environmental Science and
Technology. 24(11): 1716-1727.
Sullivan, K. A. and R. P. Mason. 1998. The concentration and distribution of mercury in Lake Michigan.
The Science of the Total Environment. 213: 213-228.
Thomas, R. L.. 1974. The distribution and transport of mercury in the sediments of the Laurentian Great
Lakes system. Proceedings of the International Conference on Transport of Persistent
Chemicals in Aquatic Ecosystems, pp. I-1 -1-16.
Thompson-Roberts, E. S., F.R. Pick, and G. E. M. Hall. 1999. Total Hg in water, sediment, and four
species of aquatic macrophytes in the St. Lawrence River, near Cornwall, Ontario. Journal of
Great Lakes Research. 25:294-304.
Tremblay, A., M. Lucotte, and D. Rowan. 1995. Different factors related to mercury concentration in
sediments and zooplankton of 73 Canadian lakes. Water, Air, and Soil Pollution. 80: 961-970.
R-4
-------
References
Tremblay, A., M. Lucotte, and R. Schetagne. 1998. Total mercury and methylmercury accumulation in
zooplankton of hydroelectric reservoirs in northern Quebec (Canada). The Science of the Total
Environment. 213: 307-315.
Ui, J. 1992. Industrial Pollution in Japan, edited by Jun Ui, United Nations University Press.
Uscinowicz, T. and R. Rossmann. June 1997. Standard operating procedures for digestion and analysis
of sediment for total mercury using the cold vapor technique with the Leeman Labs, Inc.
automated mercury system, pp. 473-503. Lake Michigan Mass Balance Study (LMMB) Methods
Compendium, Volume 2, Organic and Mercury Sample Analysis Techniques, EPA-905-R-97-
012b, Great Lakes National Program Office and Office of Water.
USEPA. National Toxics Rule. 58 FR 60848.
USEPA. Stay of Federal Water Criteria for Metals. 40 CFR 131.36.
USEPA. Stay of Federal Water Criteria for Metals. 60 FR 22208.
USEPA. Water Quality Guidance for the Great Lakes System. 40 CFR 132.
USEPA. 1995a. National Primary Drinking Water Regulations, Contaminant Specific Fact Sheets,
Inorganic Chemicals, Technical Version. U.S. Environmental Protection Agency, Office of
Water, Washington, D.C. 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, D.C. EPA 811/F-95/003-T.
USEPA. 1995c. Lake Michigan Mass Balance Project: Modeling Work Plan. Office of Research and
Development, National Health and Environmental Effects Research Laboratory, Mid-Continent
Ecology Division, Community-Based Science Support Staff, Large Lakes Research Station,
Grosse lie, Michigan, 37 pp.
USEPA. 1995d. Lake Michigan Mass Budget/Balance Work Plan. Great Lakes National Program
Office, Chicago, Illinois, 155 pp.
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.
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. EPA452/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.
R-5
-------
Results of the LMMB Study: Mercury Data Report
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. 1999a. National Recommended Water Quality Criteria-Correction. U.S. Environmental
Protection Agency, Office of Water, Washington, DC. EPA 822/Z-99/001.
USEPA. 1999b. The National Survey of Mercury Concentrations in Fish: Data Base Summary 1990-
1995. U.S. Environmental Protection Agency, Office of Water, Washington, DC. EPA 823/R-
99/014.
USEPA. 2000. Guidance for Assessing Chemical Contaminant Data for Use in Fish Advisories. EPA
823/B-00/008. U.S. Environmental Protection Agency, Office of Water, Washington, D.C.
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. Results 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: Poly chlorinated Biphenyls and trans-
Nonachlor Data Report. EPA 905/R-01/011. U.S. Environmental Protection Agency, Great
Lakes National Program Office, Chicago, IL.
USGS. 1999. U.S. Geological Survey, Mineral Commodity Summaries, January 1999.
Watras, C. J., R. C. Back, S. Halvorsen, R. J. M. Hudson, K. A. Morrison, and S. P. Wente. 1998.
Bioaccumulation of mercury in pelagic freshwater food webs. The Science of the Total
Environment. 219: 183-208.
Watras, C. J. and N. S. Bloom. 1992. Mercury and methylmercury in individual zooplankton:
implications for bioaccumulation. Limnology and Oceanography. 37(6): 1313-1318.
Watras, C. J., K. A. Morrison, J. S. Host, and N. S. Bloom. 1995. Concentrations of mercury species in
relationship to other site-specific factors in the surface waters of northern Wisconsin lakes.
Limnology and Oceanography. 40(3): 556-565.
Westcott, K. and J. Kalff 1996. Environmental factors affecting methyl mercury accumulation in
zooplankton. Canadian Journal of Fisheries and Aquatic Sciences. 53: 2221-2228.
WDNR. 1999. Wisconsin Mercury Deposition Network Summary Report, Volume 2. PUB-AM-302-99.
Wisconsin Department of Natural Resources, Madison, WI.
R-6
------- |