SOURCES AND PATHWAYS FOR MERCURY TO MICHIGAN’S INLAND LAKES
By
Matthew James Parsons

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILSOPHY
Geological Sciences
2011

ABSTRACT
SOURCES AND PATHWAYS FOR MERCURY TO MICHIGAN’S INLAND LAKES
By
Matthew James Parsons
Mercury has a unique geochemical cycle due to its ability to exist as a gas at Earth
surface conditions. As a result Hg can be transported across a region (e.g., the Midwest) or
across the globe. The importance of local watershed scale versus regional/global sources is still
debated. The over-arching hypothesis driving this research is that local scale sources and
watershed scale processes have more impacts on the rate of Hg accumulation in the aquatic
environment than regional to global scale sources. To evaluate this hypothesis,

210

Pb dated

mercury sediment chronologies were compared among inland lakes within the State of Michigan
and investigated using the following approaches: 1) examine the spatial and temporal trends of
Hg loading to the environment to discriminate regional and local scale processes, 2) use modern
rates of recovery to identify those watershed attributes that influence the transport of Hg through
the watershed and 3) use diagnostic ratios of polycyclic aromatic hydrocarbons (PAHs) to infer
sources of Hg contamination.
Results of temporal and spatial analysis show that 1) mercury accumulation rates for
many lakes are still increasing; 2) focusing corrected anthropogenic Hg inventories are similar to
those determined from Great Lakes sediments suggesting a regional source and 3) anthropogenic
accumulation rate profiles are similar by sub-region and share episodic accumulation events
indicative of local scale sources.

Contemporary recovery rates were calculated for mercury based on concentration and
accumulation rate profiles. These recovery rates were then compared to watershed attributes.
Agricultural, forested and wetland attributes were all found to be significantly correlated.
However, those cover types found along highly dynamic watershed flowpaths were found to be
most significant.
Mercury and PAHs can share common sources and pathways resulting in similar
environmental fates. Diagnostic ratios of PAHs are often used to differentiate source of these
contaminants to the environment and may be useful to identify sources for Hg. Ratios of PAHs
indicate that a variety of sources exist including biomass (e.g., grass or coal) and petroleum (e.g.,
vehicle or natural gas) combustion. Sources are unique on regional to sub-regional scales with
biomass combustion a dominant source for lakes in the southeastern portion of the Lower
Peninsula.
Deer Lake, an US EPA Area of Concern, in Michigan’s Upper Peninsula was
contaminated with Hg as a result of a mining related laboratory activities. Sediment cores
collected from the lake revealed that the natural recovery remediation strategy employed was
resulting in deposition of “clean” sediments and Hg concentrations should reflect those of nearby
lakes in ca. 10y. Most recently a terrestrial pathway for Hg was identified using Al:Hg ratios
which could hamper the process of delisting Deer Lake as an Area of Concern.
The results of this research support the hypothesis that local scale processes are more
significant than regional to global sources for Hg to Michigan’s lakes. This has important
implications for the development of new environmental legislation limiting Hg output from coalfired power plants and other energy generating facilities.

DEDICATION
This dissertation is dedicated to my wife Lindsey, son William and daughter Avery…the loves of
my life

iv

ACKNOWLEDGEMENTS

I would like to thank my advisor, David Long and my committee members Lina Patino, Tom
Voice and Mantha Phanikumar for their advice and guidance. I would especially like to thank
Sharon Yohn for preparing me for the toils of the grant and Ryan Vannier for taking over. The
office staff of the Department of Geological whom without I’d still be figuring out where to
park. Thanks to the Michigan Department of Environmental Quality who supported and funded
this project with special thanks to Glen Schmidt our noble captain.
Finally, thanks to my family and friends who never doubted that I would finish…someday.

v

TABLE OF CONTENTS

LIST OF TABLES

viii

LIST OF FIGURES

ix

CHAPTER 1. INTRODUCTION
Hypothesis and Approach
References

1
5
8

CHAPTER 2. SPATIAL AND TEMPORAL TRENDS OF MERCURY LOADINGS TO
MICHIGAN INLAND LAKES
Abstract
Introduction
Methods
Sediment Focusing
Calculations
Quantification of Mercury
Results and Discussion
Anthropogenic Inventories
Temporal Trends
Recent Trends
References

12
12
12
14
17
17
20
20
20
23
33
36

CHAPTER 3. A REGIONAL ANALYSIS ON THE EFFECTS OF WATERSHED
ATTRIBUTES ON THE RECOVERY OF INLAND LAKES FROM MERCURY
ENRICHMENT
Abstract
Introduction
Methods
Core collection and recovery estimates
Determination of watershed attributes
Results and Discussion
Using sediment chronologies to define recovery
Patterns of watershed attributes
Patterns of concentration recovery rates
Patterns of flux recovery rates
Factors influencing concentration recovery rates
Factors influencing flux recovery rates
References

40
40
40
45
45
46
56
56
57
58
59
59
64
69

vi

CHAPTER 4. INFERRING SOURCES FOR MERCURY TO INLAND LAKES USING
SEDIMENT CHRONOLOGIES OF POLYCYCLIC AROMATIC HYDROCARBONS
Abstract
Introduction
Materials and Methods
Results and Discussion
Patterns of PAH and Hg
Diagnostic PAH Ratios
References

74
74
74
77
79
79
91
110

CHAPTER 5. ASSESSING THE NATURAL RECOVERY OF A LAKE CONTAMINATED
WITH HG USING ESTIMATED RECOVERY RATES DETERMINED BY SEDIMENT
CHRONOLOGIES
116
Abstract
116
Introduction
117
History and Site Description
120
Sampling and Analytical Methods
122
Results and Discussion
125
210
Effects of Lake Level Drawdown on
Pb Profiles
125
Impact of Mining Activities on Mercury Concentrations in Deer Lake
134
The Influence of watershed processes on mercury concentrations in Deer Lake
137
Estimating Recovery Using Hg Accumulation Rates and Sediment Concentrations
140
Conclusions
147
References
150
CHAPTER 6. SUMMARY AND RECOMMENDATIONS
Summary
Recommendations

154
154
158

APPENDIX A. SEDIMENT CHEMICAL DATA

163

APPENDIX B. SUPPORTING INFORMATION FOR CHAPTER 2
References

184
190

APPENDIX C. METHODS FOR ACQUISITION OF WATERSHED CHARACTERISTIC
DATA
193
APPENDIX D. ERROR ANALYSIS
References

212
221

vii

LIST OF TABLES

Table 3.1 Watershed attributes of study lakes: watershed area (WA), lake area (LA)
watershed to lake area ratio (WA:LA), urban (URB), agriculture (AG),
forest (FOR), upland (UP), barren (BAR), and wetland (WET)

49

Table 4.1 PAH parameters and abbreviations

78

Table 4.2 ΣPAH (ng/g dry wt) and Hg (mg/kg dry wt) concentration in the surfical
sediments and at maximum. The first column shows the Pearson’s
Correlation coefficient between Hg and ΣPAH.

80

Table A1
Table A2

210
210

Pb activities for Deer Lake South sediment cores

165

Pb activities for Deer Lake North sediment cores

168

Table A3 Metal concentrations for Deer Lake South core (mg/kg)

171

Table A4 Metal concentrations for Deer Lake North core (mg/kg)

177

210

Pb data for 26 Michigan lakes

188

Table C1 IFMAP Reclassification Scheme

193

Table C2 Landuse reclassification scheme

208

Table C3 Land use classifications for the 1978 MIRIS dataset

210

Table D1 SRM 1515 Population Statistics

217

Table D2 SRM 1633b Population Statistics

217

Table D3 Descriptive Sample Statistics and Bias

218

Table D4 SRM 1515 LOD

219

Table D5 SRM 1633b LOD

219

Table S1

viii

LIST OF FIGURES

Figure 2.1. Inland lake trends monitoring program: sediments study lakes, trends of
Hg loading over the last decade, and coal burning power plants. An up-arrow
indicates accumulation rates had increased toward the surface, a down-arrow
indicates accumulation rates had decreased. Solid triangles mean that accumulation
rates were elevated but are not increasing or decreasing. Solid circles indicate that
accumulation rates had returned to pre-industrial rates. Open stars represent coal
burning power plants greater than 25 MW generating capacity as of 2005.

16

Figure 2.2 Focusing-corrected anthropogenic Hg inventories from 18 inland
Michigan lakes. Inventories are plotted along a south to north gradient. The boxes
represent the range of anthropogenic inventories for Lake Michigan and Lake
Superior (Long et al. 1995).

22

210

Figure 2.3.
Pb dated focusing-corrected anthropogenic accumulation rates for
four southwestern Michigan lakes, dates older than 1850 were truncated to highlight
more recent loading. Note the change in scale of the x-axis.

24

210

Figure 2.4.
Pb dated focusing-corrected anthropogenic accumulation rates for 4
southeastern Michigan lakes. Note the change in scale of the x-axis.

25

210

Figure 2.5.
Pb dated focusing-corrected anthropogenic accumulation rates for five
Upper Peninsula of Michigan lakes, dates older than 1850 were truncated to
highlight more recent loading. Note the change in scale of the x-axis.

26

210

Figure 2.6.
Pb dated focusing-corrected anthropogenic Hg accumulation rates for
two north-central Michigan lakes. Note the change in scale of the x-axis.

27

210

Figure 2.7.
Pb dated focusing-corrected anthropogenic accumulation rates for
four northwestern Michigan lakes, dates older than 1850 were truncated to highlight
more recent loading. Note the change in scale of the x-axis.

28

Figure 3.1. Location of study lakes.

44

Figure 3.2: Mercury concentration profiles (solid markers), results of profile
smoothing (open markers) and recovery rate (solid line) for study lakes.

54

Figure 3.3. Mercury accumulation rate (Acc Rate) profiles (solid markers), results of
profile smoothing (open markers) and recovery rate (solid line) for study lakes.

61

ix

Figure 4.1. Study lakes

82

Figure 4.2a. Concentration profiles of Hg and PAH isomer groups for Avalon Lake.

83

Figure 4.2b. Concentration profiles of Hg and PAH isomer groups for Birch Lake.

84

Figure 4.2c. Concentration profiles of Hg and PAH isomer groups for Crystal Lake.

85

Figure 4.2d. Concentration profiles of Hg and PAH isomer groups for Muskegon
Lake.

86

Figure 4.2e. Concentration profiles of Hg and PAH isomer groups for Otter Lake.

87

Figure 4.2f. Concentration profiles of Hg and PAH isomer groups for Round Lake.

88

Figure 4.2g. Concentration profiles of Hg and PAH isomer groups for Sand Lake.

89

Figure 4.3a. Selected ratios of PAH isomers in sediment of Avalon Lake. Vertical
lines represent specific PAH diagnostic ratios used to distinguish source, see text for
specific examples.

94

Figure 4.3b. Selected ratios of PAH isomers in sediment of Birch Lake. Vertical lines
represent specific PAH diagnostic ratios used to distinguish source, see text for
specific examples.

95

Figure 4.3c. Selected ratios of PAH isomers in sediment of Crystal Lake. Vertical
lines represent specific PAH diagnostic ratios used to distinguish source, see text for
specific examples.

96

Figure 4.3d. Selected ratios of PAH isomers in sediment of Muskegon Lake. Vertical
lines represent specific PAH diagnostic ratios used to distinguish source, see text for
specific examples.

97

Figure 4.3e. Selected ratios of PAH isomers in sediment of Otter Lake. Vertical lines
represent specific PAH diagnostic ratios used to distinguish source, see text for
specific examples.

98

Figure 4.3f. Selected ratios of PAH isomers in sediment of Round Lake. Vertical
lines represent specific PAH diagnostic ratios used to distinguish source, see text for
specific examples.

99

Figure 4.3g. Selected ratios of PAH isomers in sediment of Sand Lake. Vertical lines
represent specific PAH diagnostic ratios used to distinguish source, see text for
specific examples.

100

x

Figure 4.4. Industrial and utility energy emitters according to the 2002 Mercury
Emissions Inventory for the State of Michigan.

104

Figure 4.5. Current cement kilns in the State of Michigan. The historic cement kiln
located in Cement City, MI is not shown but would have been located approximately
20 km due west of Sand Lake.

107

Figure 5.1. Deer Lake study area. Expanded view shows an areal photograph of the
study area clipped to the Deer Lake shoreline during the major drawdown. Yellow
circles indicate sampling locations. Also shown are the approximate locations of the
Rope’s Gold Mine, Carp River Dam and the City of Ishpeming WWTP Outfall.

119

Figure. 5.2. Excess
the

137

210

Pb profiles (solid circles) from the South and North Basins and

Cs profile (open circles) from the North Basin of Deer Lake. Double arrows
210

Pb profile; CRS model dates were
indicate the large deviation in the excess
assigned to the beginning and end of the large deviation for their associated mass
depths.

126

Figure 5.3. Profiles of mercury (open markers) and lead (solid markers) normalized to
their respective maximums to highlight similarities between the profiles.

131

Figure 5.4. Dry mass profiles from the South and North basins of Deer Lake. Double
210
arrows indicate the large deviation in the excess
Pb profile; dates are assigned to
the beginning and end of the large deviation for their associated mass depths.

133

Figure 5.5. Hg concentration (mg/kg) profiles from the South and North basins of
Deer Lake (open circles) and production of usable iron ore from Michigan (closed
circles, metric ton(x1000)).

135

Figure 5.6. Profiles of Al (open circles) and Hg:Al ratio (solid circles) for the South
and North basins of Deer Lake. Double arrows indicate the beginning and end of the
210
large deviation in excess
Pb activity.

139

Figure 5.7. Focusing corrected Hg accumulation (µg/m2/y) rates for the South and
North basins of Deer Lake.

144

Figure 5.8. Hg FCAR profiles for the South and North basins of Deer Lake, 1980 to
present. The dashed line shows the linear model used to determine the time to
recovery, equations used are also shown. The dash-dot-dash line shows the model
applied to predict the year of recovery.

145

Figure 5.9. Hg concentration profiles for the South and North basins of Deer Lake,
1980 to present. The dashed line shows the linear model used to determine the time
to recovery, equations used are also shown. The dash-dot-dash line shows the model
applied to predict the year of recovery.

146

xi

210

Figure S1.
Pb dated focusing-corrected anthropogenic accumulation rates for four
Michigan lakes, dates older than 1850 were truncated to highlight more recent
loading. Note the change in scale on the x-axis.

184

210

Figure S2.
Pb dated focusing-corrected anthropogenic accumulation rates for three
Michigan lakes, dates older than 1850 were truncated to highlight more recent
loading. Note the change in scale on the x-axis.

xii

185

CHAPTER 1
INTRODUCTION

Mercury contamination and its subsequent methylation and bioaccumulation in aquatic
food-webs has resulted in elevated levels of mercury in fish and resulted in statewide fish
consumption advisories for the Great Lakes region, including inland lakes throughout the state of
Michigan (Keeler et al., 1994, Watras et al., 1994). Fish consumption has been shown to be a
primary pathway of mercury, and its more toxic form methylmercury, exposure to humans
(Hightower and Moore, 2003). Toxicity issues related to methylmercury contaminated fish
consumption was highlighted in the 1950’s and 1960’s with deaths in Minamata Bay and
Niigata, Japan (Tsubaki and Irukayama, 1977). Although reported emissions of mercury have
declined in the past decade (Engstrom and Swain, 1997, USEPA, 2008), the amount of river
miles and lake acreage under fish consumption advisories have continued to increase (US EPA,
2004). The cause of which is two-fold:1) due to its geochemical cycle mercury can be
transported long distances creating a contaminated water body in an otherwise pristine setting
(Tseng et al., 2004), and 2) waterbodies that were previously untested are being confirmed to
contain fish that have Hg levels above regulatory limits (US EPA, 2004).
Mercury is ubiquitous in the environment, and is characterized by an atmospheric
residence time of one year (Fitzgerald, 1989). As a result, Arctic region lakes (Hermanson,
1998) and Antarctic ice cores (Vandal et al., 1993) are contaminated with mercury though they
are far from anthropogenic sources. Anthropogenic sources of mercury to the environment
include coal fired utilities, smelting activities, medical and municipal waste incineration,
agricultural and paint fungicides, gold and silver mining activities, chlor-alkali facilities, and
municipal waste streams (Jasinski, 1994). Many industrial applications for mercury have been
1

disallowed (e.g., agricultural fungicides) in the United States and substitutes are encouraged to
be used (e.g., alcohol in thermometers) consequently consumption has decreased over the last
three decades (Brooks, 2003); but mercury manometers, electronic switches and lamps, batteries,
and dental amalgams continue to add to U.S. consumption. Even with consumption declines
realized, all environmental records do not reflect a concomitant decline in mercury accumulation
(Engstrom and Swain, 1997, Engstrom et al., 1994, Lorey and Driscoll, 1999, Swain et al.,
1992b).
It is known that atmospheric deposition is the principle pathway for mercury
contamination of inland lakes (Fitzgerald, 1989, Fitzgerald et al., 1998, Mason et al., 1994a,
Watras et al., 2000). However, the impact of global (e.g., China) versus regional (e.g., Midwest
power plants) versus more local scale processes (e.g., increased runoff from agricultural fields) is
still debated (Evers et al., 2011). Determining the source for Hg to inland lakes has important
implications for future environmental legislation (US EPA, 2011)which would seek to cap Hg
emissions from power generating facilities in an attempt to further reduce Hg exposure. The
environmental legacy of historic Hg emissions also plays an important role in any future
legislative action since it is that Hg that now resides in soils that will continue to pose an
exposure threat even when Hg emissions from all anthropogenic sources are reduced to zero
(Mason et al., 1994b). It will be important then to more fully understand the processes that result
in Hg transport at the watershed scale.
A variety of approaches have been used to determine current and historical rates of
contaminant loading to the environment including: atmospheric deposition studies, ombotrophic
bogs, and lake sediment cores. Atmospheric deposition pathways have been studied to
determine rates of mercury deposition (Glass and Sorensen, 1999, Landis and Keeler, 2002,

2

Landis et al., 2002). Several Midwest states have been included in the National Atmospheric
Deposition Program (NADP) providing a continuously updated data set for calibration of
atmospheric cycling and deposition models (Illinois State Water Survey, 2008). The
International Association of Deposition Networks (IADN) (Hillery et al., 1998) has been
collecting similar data for the last decade throughout the Great Lakes Region. Ombotrophic
bogs throughout the world have also been used to determine current and historical rates of
contaminant deposition (Benoit et al., 1998, Bindler, 2003).
Lake sediment cores have been used to measure contaminant-loadings of atmospheric
and terrestrial inputs (Callender and Rice, 2000, Engstrom and Swain, 1997, Swain et al., 1992b,
Yohn et al., 2002b). Studies of lake sediment cores from the Midwest and Adirondacks have
shown decreasing mercury accumulation rates attributed to a reduction in atmospheric deposition
(Engstrom and Swain, 1997, Lorey and Driscoll, 1999, Swain et al., 1992a). The Great Lakes
region has been of particular interest in contaminant studies due to several industrialized areas
within the basin and the revenue dollars it provides to Great Lakes states through recreation,
tourism, and sport fishing. Sediment cores collected from Lake Superior had maximum
concentrations of mercury in the surficial sediments for the majority of the samples collected
(Rossmann, 1999). Approximately 70% of the mercury inventory was attributed to
anthropogenic activity (Rossmann, 1999). Surficial sediment concentrations from Lake
Michigan show a decline from historical data sets and estimate contributions from local sources
to be approximately 50% of total mercury loadings (Rossmann, 2002).
Most regulatory actions have been intended to reduce loadings of mercury from
atmospheric sources, such as the burning of fossil fuels and waste incineration facilities that
create a regional signal which would be expected to consist of a uniform deposition or
3

accumulation rate across a broad spatial scale (Engstrom and Swain, 1997). However, local
sources and processes acting on the watershed scale will not display uniform loading over large
spatial scales. These sources and processes may become more important over time, as emissions
from industrial activities are reduced, and therefore it is becoming increasingly important to
understand local scale influences on mercury loadings to aquatic ecosystems.
Investigation of sediment cores of inland lakes can aid in determining regional and local
signals of contaminant loadings. Previous work investigated the anthropogenic accumulation of
lead and cadmium using multivariate statistics and metal ratios to determine current and legacy
land use effects on anthropogenic accumulation rates of metals to inland lakes (Yohn et al.,
2003). Results showed that regional and local signals can be determined from temporal and
spatial trends of anthropogenic metals. And the efficacy of environmental legislation, such as
the Clean Air Act, on reducing regional signals of anthropogenic accumulation of lead can be
tested (Yohn, 2003). For example, it has been demonstrated that the anthropogenic accumulation
of lead and copper have not declined to background levels, as one would expect if a regional
source was eliminated (Yohn et al., 2002a). Therefore a different, possibly local, source may be
contributing to the continued accumulation of anthropogenic chemicals.
Various sources can contribute to sediment deposition within lakes including terrestrial,
atmospheric, and in-lake processes which mix with anthropogenic inputs (Yohn et al., 2001)
creating difficulty in elucidating source for a metal of interest. Heyvaert (2000) used titanium
ratios to normalize mercury concentrations to correct for terrestrial inputs, a watershed correction
factor. However, it may be more appropriate to look at background correction for mercury due
its geochemical cycle (Strunk, 1991). Using background correction assumes that the majority of
anthropogenic mercury deposited to the landscape occurs during atmospheric deposition and is
4

retained by soils and does not readily mobilize after deposition except during processes such as
fire, clear cutting, urbanization, landscape disturbance (i.e., agricultural practices), and flooding.
Hypothesis and Approach
This study attempts to elucidate sources and pathways for Hg to inland lakes, and
importance of local and regional scale sources. The over-arching hypothesis is that: local and
watershed scale processes have more impact on the rate of Hg accumulation in the aquatic
environment than regional to global scale sources.
Three approaches will be used to test this hypothesis, and they are:
1. examine the temporal changes of Hg accumulation over spatial scales. Local scale
sources and watershed scale processes would be reflected in the sediment record as
discontinuities in anthropogenic accumulation rates at multiple time periods and
spatial scales and inventories across spatial scales,
2. use modern recovery rates of recovery to identify those watershed attributes that
influence the movement of Hg through the watershed. Differences in processes
amongst watersheds (e.g., agriculture v. forest) impact the rates of modern recovery
from mercury contamination, and
3. use diagnostic ratios of polycyclic aromatic hydrocarbons (PAHs) as tracers of Hg
contamination. Sources of PAHs can be differentiated using diagnostic ratios and Hg
and PAHs share common sources and pathways thus PAH ratios can be used to infer
sources of Hg.

5

Sediment was collected from inland lakes throughout the State of Michigan, dated with

210

Pb

and analyzed for metals and organic contaminants to investigate this hypothesis. Results of
analysis from individual lakes can be found in several year-end reports (Parsons et al., 2004,
Parsons et al., 2006, Yohn et al., 2001, Yohn et al., 2002c, Yohn et al., 2003) prepared for the
Michigan Department of Environmental Quality as part of the Trends Monitoring Program:
Sediments (Michigan, 2011). This dissertation is composed of five chapters beyond the
introduction. Chapter 2 explores the use of anthropogenic accumulation rates and inventories
examined at temporal and spatial scales, respectively, focusing on the identification of source
regions. Chapter 3 investigates the use of geographic information system derived watershed
attributes on the rate of recovery of inland lakes from Hg contamination. The fourth chapter uses
diagnostic ratios of PAHs to infer sources of Hg; while the fifth, investigates the historic and
modern sources of Hg to Deer Lake near Ishpeming, MI, a U.S. EPA Area of Concern (USEPA,
2007)using Hg:metal ratios and changes in metal profiles that are consistent with the known
history of the lake. The final chapter offers a summary and suggestions for future work.

6

REFERENCES

7

REFERENCES
Benoit, J.M., Fitzgerald, W.F., Damman, A.W.H., 1998. The biogeochemistry of an
ombrotrophic bog: Evaluation of use as an archive of atmospheric mercury deposition.
Environmental Research 78, 118-133.
Bindler, R., 2003. Estimating the natural background atmospheric deposition rate of mercury
utilizing ombrotrophic bogs in southern Sweden. Environmental Science and Technology
37, 40-46.
Brooks, W.E. United States Geological Survey. Minerals Yearbook - Mercury; Reston, VA:
Government Printing Office, 2003.
Callender, E., Rice, K.C., 2000. The urban environmental gradient: Anthropogenic influences on
the spatial and temporal distributions of lead and zinc in the environment. Environmental
Science and Technology 24, 232-238.
Engstrom, D.R., Swain, E.B., 1997. Recent declines in atmospheric mercury deposition in the
upper Midwest. Environmental Science and Technology 31, 960-967.
Engstrom, D.R., Swain, E.B., Henning, T.A., Brigham, M.E., Brezonik, P.L., 1994. Atmospheric
Mercury Deposition to Lakes and Watersheds - a Quantitative Reconstruction from
Multiple Sediment Cores. Environmental Chemistry of Lakes and Reservoirs 237, 33-66.
Evers, D.C., Wiener, J.G., Driscoll, C.T., Gay, D.A., Basu, N., Monson, B.A., Lambert, K.F.,
Morrison, H.A., Morgan, J.T., Williams, K.A., Soehl, A.G., 2011. Great Lakes Mercury
Connections: The Extent and Effects of Mercury Pollution in the Great Lakes Region.
Biodiversity Research Institute. Gorham, MA. Report BRI 2011-18, 44 pages. Available
from: http://www.briloon.org/mercuryconnections/GreatLakes.
Fitzgerald, W.F., 1989. Atmospheric and oceanic cycling of mercury. Academic Press.
Fitzgerald, W.F., Engstrom, D.R., Mason, R.P., Nater, E.A., 1998. The case for atmospheric
mercury contamination in remote areas. Environmental Science and Technology 32, 1-7.
Glass, G.E., Sorensen, J.A., 1999. Six-year trend (1990-1995) of wet mercury deposition in the
Upper Midwest. USA. Environmental Science & Technology 33, 3303-3312.
Hermanson, M.H., 1998. Anthropogenic mercury deposition to arctic lake sediments. Water, Air,
and Soil Pollution 101, 309-321.
Hightower, J., Moore, D., 2003. Mercury Levels in High-End Consumers of Fish. Environmental
Health Perspectives 111, 604-608.
Hillery, B.R., Simcik, M.F., Basu, I., Hoff, R.M., Strachan, W.M.J., Burniston, D., Chan, C.H.,
Brice, K.A., Sweet, C.W., Hites, R.A., 1998. Atmospheric deposition of toxic pollutants
8

to the Great Lakes as measured by the integrated atmospheric deposition network.
Environmental Science & Technology 32, 2216-2221.
National Atmospheric Deposition Program (NRSP-3), Illinois State Water Survey, NADP
Program Office, 2204 Griffith Drive, Champaign, IL 61820, Available online at
http://nadp.sws.uiuc.edu/mdn/. (Accessed: February 5, 1998).
Jasinski, S.M. The Materials Flow of Mercury in the United States; Government Printing
Office, 1994.
Keeler, G.J., Hoyer, M.E., Lamborg, C.H., 1994. Mercury Pollution: Integration and Synthesis.
In: Measurements of Atmospheric Mercury in the Great Lakes Basin.
Landis, M.S., Keeler, G.J., 2002. Atmospheric mercury deposition to Lake Michigan during the
Lake Michigan Mass Balance Study. Environmental Science and Technology 36, 45184524.
Landis, M.S., Vette, A.F., Keeler, G.J., 2002. Atmospheric mercury in the Lake Michigan basin:
Influence of the Chicago/Gary urban area. Environmental Science and Technology 36,
4508-4517.
Lorey, P., Driscoll, C.T., 1999. Historical trends in mercury deposition in Adirondack Lakes.
Environmental Science and Technology 33, 718-722.
Mason, R.P., Fitzgerald, W.F., Morel, F.M., 1994a. The biogeochemical cycling of elemental
mercury: anthropogenic influences. Geochimica at Cosmochimica Acta 58, 3191-3198.
Mason, R.P., Fitzgerald, W.F., Morel, F.M.M., 1994b. The Biogeochemical Cycling of
Elemental Mercury - Anthropogenic Influences. Geochimica et Cosmochimica Acta 58,
3191-3198.
State of Michigan, 2011. Sediment Chemistry, Available online at
http://www.michigan.gov/deq/0,4561,7-135-3313_3686_3728-32365--,00.html.
(Accessed: September 5, 2011).
Parsons, M.J., Yohn, S.S., Long, D.T., Giesy, J.P., Bradley, P.W., 2004. Inland Lakes Sediment
Trends: Sediment Analysis Results for Five Michigan Lakes. Lansing, MI. pages.
Available from:
Parsons, M.J., Yohn, S.S., Long, D.T., Patino, L.C., Stone, K.A., 2006. Inland Lakes Sediment
Trends: Mercury Sediment Analysis Results for 27 Michigan Lakes. Lansing, MI. 47
pages. Available from:
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,
683-696.

9

Rossmann, R., 2002. Lake Michigan 1994-1996 surficial sediment mercury. Journal of Great
Lakes Research 28, 65-76.
Strunk, J.L. 1991. The extraction of mercury from sediment and the geochemical partitioning of
mercury in sediments from Lake Superior. Thesis, Michigan State University. Dept. of
Geology, 1991.,
Swain, E.B., Engstrom, D.R., Brigham, M.E., Henning, T.A., Brezonik, P.L., 1992a. Increasing
Rates of Atmospheric Mercury Deposition in Midcontinental North-America. Science
257, 784-787.
Swain, E.B., Engstrom, D.R., Brigham, M.E., Henning, T.A., Brezonik, P.L., 1992b. increasing
rates of atmospheric mercury deposition in Midcontinental North America. Science 257,
784-787.
Tseng, C.M., Lamborg, C., Fitzgerald, W.F., Engstrom, D.R., 2004. Cycling of dissolved
elemental mercury in Arctic Alaskan lakes. Geochimica et Cosmochimica Acta 68, 11731184.
Tsubaki, T., Irukayama, K., 1977. Minamata Disease: Methylmercuy Poisoning in Minamata and
Niigata, Japan. Elsevier.
U.S. EPA, 2004. National Listing of Fish Advisories, Available online at
http://www.epa.gov/waterscience/fish/. (Accessed: September 25, 2011).
US EPA, 2011. Proposed Rule: National Emission Standards for Hazardous Air Pollutants From
Coal- and Oil-Fired Electric Utility Steam Generating Units and Standards of
Performance for Fossil-Fuel-Fired Electric Utility, Industrial-Commercial-Institutional,
and Small Industrial-Commercial-Institutional Steam Generating Units. US
EPA.Government Printing Office, Washington, D.C. 40 CFR Parts 60 and 63.
Deer Lake River Area of Concern, Available online at
http://www.epa.gov/glnpo/aoc/drlake.html. (Accessed: January 30, 2008).
Toxic Release Inventory (TRI) Program, Available online at http://www.epa.gov/tri. (Accessed:
January 30, 2008).
Vandal, G.M., Fitzgerald, W.F., Boutron, C.F., Candelone, J.P., 1993. Variations in Mercury
Deposition to Antarctica over the Past 34,000 Years. Nature 362, 621-623.
Watras, C.J., Bloom, N.S., Hudson, J.M., Gherini, S., Munson, R., Claas, S.A., Marrison, K.A.,
Hurely, J., Wiener, J.G., Fitzgerald, W.F., Mason, R., Vandal, G., Powell, D., Rada, R.,
Rislov, L., Winfrey, M., Elder, J., Krabbenhoft, D., Andren, A.W., Babiarz, C., Porcella,
D.B., Huckabee, J.W., 1994. Sources and fates of mercury and methylmercury in
Wisconsin lakes. In: C.J. Watras, J.W. Huckabee (Eds.), Mercury Pollution: Intergration
and Synthesis. CRC Press, Ann Arbor, pp. 153-179.

10

Watras, C.J., Morrison, K.A., Hudson, R.J., Frost, T.M., Kratz, T.K., 2000. Decreasing mercury
in northern Wisconsin: temporal trends in bulk precipitation and a precipitationdominated lake. Environmental Science and Technology 34, 4051-4057.
Yohn, S.S., 2003. Regional versus local influences on lead and cadmium loading in the Great
Lakes region. Applied Geochemistry In press.
Yohn, S.S., Long, D.T., Fett, J.D., Patino, L., Giesy, J.P., Kannan, K., 2002a. Assessing
environmental change through chemical-sediment chronologies from inland lakes. Lakes
Reserv Res Manage 7, 217-230.
Yohn, S.S., Long, D.T., Fett, J.D., Patino, L.C., Giesy, J.P., Kannan, K., 2002b. Assessing
environmental change through chemical-sediment chronologies from inland lakes. Lakes
& Reservoirs: Research and Management 7, 217-230.
Yohn, S.S., Long, D.T., Giesy, J.P., Fett, J.D., Kannan, K., 2001. Inland lakes sediment trends:
sediment analysis results for two Michigan Lakes. Lansing, MI. 49 pages. Available
from:
Yohn, S.S., Long, D.T., Giesy, J.P., Scholle, L.K., Patino, L.C., Fett, J.D., Kannan, K., 2002c.
Inland lakes sediment trends: sediment analysis results for five Michigan Lakes. East
Lansing. 64 pages. Available from:
Yohn, S.S., Long, D.T., Giesy, J.P., Scholle, L.K., Patino, L.C., Parsons, M., Kannan, K., 2003.
Inland Lakes Sediment Trends: sediment analysis results for six Michigan Lakes. East
Lansing. 66 pages. Available from: http://www.deq.state.mi.us/documents/deq-wb-swassedimenttrend0203finalreport.pdf.

11

CHAPTER 2
SPATIAL AND TEMPORAL TRENDS OF MERCURY LOADINGS TO MICHIGAN
INLAND LAKES
Abstract
Several studies of chronologies of mercury (Hg) in inland lake sediments have
demonstrated that Hg accumulation had decreased in recent decades. However, episodic
mercury accumulation events were recorded in some of these lakes, but not investigated
in detail. Recent decreases had been attributed to the reduction of regional Hg
consumption and secondary removal during process waste treatment. In addition to
regional sources, local sources, including watershed disturbance, might significantly
contribute to Hg loading. Here, mercury chronologies of Hg loadings based on dated
sediment cores are presented for 26 inland, Michigan lakes. Although spatial trends of
anthropogenic inventories suggest a regional pattern dominated by human activities, subregional to local scale sources are also found to be significant. Temporal trends show
episodic Hg accumulation events superimposed on a more general, long-term trend.
Episodic increases common to lakes suggest a common source or processes common to
lakes. Episodic increases unique to a lake indicate a more local scale source. Similar Hg
profiles from lakes that are geographically proximal provide evidence for sub-regional to
regional scale sources. Local sources and pathways for mercury to inland lakes need to
be more fully understood to effectively reduce Hg loading to the environment.

Introduction
Mercury (Hg) concentrations greater than historical levels, as recorded in
sediment cores, have been observed in the high northern latitudes (Bindler et al., 2001,
Gobeil et al., 1999, Hermanson, 1998), New England (Kamman and Engstrom, 2002,
12

Lorey and Driscoll, 1999), the Midwestern United States (Engstrom and Swain, 1997,
Swain et al., 1992), Western United States (Gray et al., 2005, Menounou and Presley,
2003), and Europe (Wihlborg and Danielsson, 2006, Yang and Rose, 2003). Although
regulatory controls have attempted to limit mercury emissions (1997), the number of
river miles and fresh water lakes in which Hg concentrations exceed regulatory
guidelines continue to increase. The reasons for this increase have been attributed to
monitoring efforts extending to water bodies that were previously untested (2004), or
increased releases from human activities and subsequent deposition in remote areas
(Fitzgerald et al., 1998). Consumption of contaminated fish has been shown to be a
principle exposure route of methylmercury to humans (Hightower and Moore, 2003,
1997). As a result of elevated health risks, several Midwestern states have advised
consumers to reduce consumption of fish from all of the Great Lakes and all Michigan
inland lakes due to potential Hg exposure (2004).

Freshwater sediments retain mercury transported to the lake system (Engstrom et
al., 1994) and when coupled with dating techniques, sediments can record the history of
mercury loadings over time. Dated sediment cores have been used to investigate the
historic accumulation of Hg in sediments of the North American Great Lakes (Long et
al., 1995) and inland lakes (Engstrom and Swain, 1997, Kamman and Engstrom, 2002,
Lorey and Driscoll, 1999, Swain et al., 1992). While analysis of Great Lakes sediment
provides significant information about deposition of Hg on larger spatial scales, it is
difficult to determine local watershed scale contribution to Hg loadings. Investigating
inland lakes across a region allows examination of both regional- and, local, including

13

watershed scale, contribution to Hg accumulation. Earlier determinations of sediment
chronologies have demonstrated that Hg loadings in some areas of the United States
(USA) such as the Western United States (Gray et al., 2005, Menounou and Presley,
2003), Midwestern United States (Engstrom and Swain, 1997, Swain et al., 1992), New
England (Kamman and Engstrom, 2002, Perry et al., 2005), the Adirondack region of
New York (Lorey and Driscoll, 1999), and Europe (Wihlborg and Danielsson, 2006,
Yang and Rose, 2003) have decreased in recent years. However, several of the lakes
included in previous studies had recent increases in Hg accumulation.

The results reported in this paper are part of a larger study, “Inland Lakes Trends
Monitoring Program: Sediments,” to determine trends of organic and inorganic
contaminant accumulation in Michigan’s inland lakes (2006). Here the chronologies of
Hg loading into lakes were determined from 26

210

Pb-dated, confirmed by

137

Cs and

stable Pb, sediment cores from inland lakes throughout Michigan. Anthropogenic
inventories of Hg loadings to lakes were investigated at several spatial scales, and Hg
accumulation rates, examined temporally to test the hypothesis that local-scale sources of
Hg significantly contribute to Hg accumulation in inland lakes. Understanding local
scale sources for Hg will become increasingly important in assessing the efficacy of
control measures for regional sources.

Methods
During the summers of 1999 through 2004, 26 lake sediment cores were collected
aboard the U.S. Environmental Protection Agency (EPA) R/V Mudpuppy or Michigan
14

Department of Environmental Quality (MDEQ) M/V Nibi (Figure 2.1). Lakes were
selected to cover all regions of the state and varied in their watershed (area) and
hydrological characteristics, such as whether they were headwater or seepage lakes. An
MC-400 Lake/Shelf multi-corer was used to extract four simultaneous cores for

210

Pb-

dating and analysis of Hg and other metals. The core was sectioned on shore shortly after
retrieval at 0.5-1.0 cm intervals for the first 5-8 cm and 1 cm intervals for the remainder
of the core; core lengths varied from 50-60 cm. Sectioning of the

15

Gratiot

Imp

Round

Witch
Round D

Mullett
Avalon
Torch
Elk
Houghton

Crystal B

Cadillac

Hackert

Shupac
George

Higgins

Littlefield
Otter

Crystal M

Muskegon

Cass
Gull
Paw Paw
Birch

Whitmore
Sand

Figure 2.1. Inland lake trends monitoring program: sediments study lakes, trends of Hg
loading over the last decade, and coal burning power plants. An up-arrow indicates
accumulation rates had increased toward the surface, a down-arrow indicates
accumulation rates had decreased. Solid triangles mean that accumulation rates were
elevated but are not increasing or decreasing. Solid circles indicate that accumulation
rates had returned to pre-industrial rates. Open stars represent coal burning power plants
greater than 25 MW generating capacity as of 2005.

16

cores used for determining Hg concentrations and

210

Pb was conducted at the same

stratigraphic resolution. More detailed descriptions of the sectioning protocols and
results of

210

Pb are presented elsewhere (Yohn et al., 2002a, Yohn et al., 2001, Yohn et

al., 2002b, Yohn et al., 2003). Specific dating related information and methodology can
be found in Appendix B.

Sediment Focusing. Sediment focusing is the movement of fine grained
materials from erosional to depositional zones of a lake basin, and can be quantified by
calculating a focusing factor, which is the ratio of observed to expected
The observed

210

210

Pb inventory.

Pb inventory is the product of sediment dry mass and excess

summed over all core sections. The expected inventory of

210

210

Pb

Pb for this study region is

2

0.574 Bq/cm (Golden et al., 1993). Accounting for sediment focusing allows
comparison of inventories and accumulation rates among lakes (Golden et al., 1993).

Calculations. Anthropogenic Hg concentration was calculated by subtracting the
long-term geochemical background Hg concentration from the total concentration. The
background concentration was defined as that average concentration from the deepest
core section to the point where concentrations began to increase that could be associated
with human activities, primarily after colonization of the area by Europeans. This point
was determined graphically by determining where the concentration increased
significantly. Anthropogenic inventories, the mass of chemical in a sediment core due to

17

human inputs, were calculated as the product of anthropogenic Hg concentration and
sediment dry mass for each individual core section. Results for each core section were
then summed over the entire core to determine the inventory. The anthropogenic
accumulation rate was calculated as the product of the sedimentation rate and the
anthropogenic Hg concentration. Inventories and accumulation rates were corrected for
sediment focusing.

Sedimentation rates in lakes vary considerably, and when sedimentation rates are
great, bottom sections of the sediment core will not contain background concentrations of
Hg. Therefore, it was necessary to estimate the focusing-corrected background
accumulation rate (FCBGAR). These estimates were based on results of a backward
stepwise multivariate regression model taking into account physical variables and
surficial soil properties of the watershed. This model assumes that soil properties have
not significantly changed since the pre-industrial period and soil properties that control
the mobility of Hg do not vary among watersheds. Physical variables entered into the
model included: lake area, watershed area, and watershed:lake area ratio (WS:LA). Soil
properties were determined via dasymetric mapping of STATSGO (USDA) soil
characteristics using ArcInfo (ESRI, 1998) and included: cation-exchange capacity,
%organic matter, %clay, carbonate as CaCO3, kfact and kffact, which is a slope adjusted
kfact value. Parameters for variables to enter the model included tolerance greater than
2

0.6 and probability to enter/leave less than 0.15. The final model (n=14, Adjusted r =
0.66) is presented in equation 1; variables were significant at p<0.05 (one-tail t-statistic),
tolerances were greater than 0.75, and residuals were randomly distributed.
18

FCBGAR = 5 .016 + 0 .225 % ORG + 0 .015 WA − 0 .091 LA ,

(1)

where: %ORG is the percent organic matter in soil, WA is the watershed area and LA is
2

the area of the lake. The average (± SD) modeled FCBGAR was 7.9 (3.7) µg Hg/m /yr,
2

whereas the average of the observed FCBGAR was 6.0 (3.5) µg Hg/m /yr. The intercept,
which would represent the FCBGAR in the hypothetical absence of the variables above,
2

was reasonably close to the median FCBGAR of the observed values, 5.8 µg Hg/m /yr.
Thus, the predicted FCBGAR values were used to approximate anthropogenic
accumulation rates for lakes that did not reach geochemical background Hg
concentrations. Background estimations were not used to calculate anthropogenic
inventories due to the temporal variation of mercury accumulation; however, these
estimates would not change the accumulation rate profile.

Previous models have used WA and LA to predict Hg accumulation and organic
matter has been demonstrated to play a significant role in the cycling of mercury in the
watershed (Lorey and Driscoll, 1999, Swain et al., 1992) . Although the positive
relationship between WA and FCBGAR was expected, the negative relationship between
LA and FCBGAR was not anticipated. Removal of LA from the model consistently
2

resulted in poorer estimates, lower adjusted r , of FCBGAR. Greater inputs of Hg would
be expected for large lakes and thus background values would reflect this relationship.
We suspect that the LA variable may in part explain differences in sedimentation rates
19

observed in the larger lakes. Observed background Hg concentrations and sedimentation
rates, both of which would result in a low FCBGAR, were both least for larger lakes.
Organic matter is important in the biogeochemical cycling of Hg in soils (Grigal, 2003,
Meili, 1991). Greater concentrations of organic matter in soils effectively retain Hg and
other metals. More effective retention in soils might lead to greater concentrations during
runoff events resulting in the transport of Hg from the watershed (Scherbatskoy et al.,
1998).

Quantification of Mercury. Total Hg concentrations were determined in
lyphophilized sediment by use of a Lumex Thermal Decomposition Atomic Absorption
Spectrometer in accordance with EPA Method 7473 (USEPA, 1998). Calibration of the
instrument and in-run validation were performed using standard reference material
(SRM) 1515 Apple Leaves or SRM1633b Coal Fly Ash; detection limits were 0.69 ng
Hg/g, dw and 0.58 ng Hg/g, dw, respectively. In-run SRM validation measurements were
consistently within the range specified.

Results and Discussion
Anthropogenic Inventories. Anthropogenic inventories are an accounting of the
long-term loading of a contaminant to a lake. Analysis of the spatial patterns of
inventories among lakes can reveal patterns of source in a region. For the purposes of
this discussion a region has been defined as the entire state of Michigan whereas subregions are smaller areas within the region, such as southeast Michigan. A regional
pattern of deposition is indicated if inventories decrease along some gradient, such as

20

population density or industrial development. The source for the pattern of deposition
may be common to all lakes, a common source, or similar watershed processes occurring
at different rates, a source common to all lakes. For example, Pb inventories in these
study lakes decrease from the South to the North, implying a regional source (Yohn et al.,
2004). Lead from the use of leaded gasoline is the most commonly accepted cause for
this trend; greater lead inventories in the southern, more populated and industrialized
portion of the state trend toward lesser inventories in the northern, less populated,
portions of the state. A regional source may also be present if inventories are similar
among lakes within a sub-region. Focusing-corrected copper inventories were found to
be similar among depositional basins of Lake Michigan. (Kolak et al., 1998). This
suggests a regional source for copper entering Lake Michigan.
Recent measurements of atmospheric deposition of Hg for the Great Lakes region
have demonstrated a South to the North gradient (Keeler and Dvonch, 2005). If this
general South to North trend has been persistent over time then inventories plotted along
this spatial gradient should follow a similar pattern, evidence of a regional source.
Results of focusing-corrected anthropogenic Hg inventories for 18 Michigan lakes
(Figure 2.2) were plotted along a South to North gradient to investigate the possibility of
a regional source. Mercury inventories varied among lakes, ranging from 0.3 µg Hg/cm
2

in Round Lake to 1.6 µg Hg/cm in Houghton Lake (Figure 2.2). Anthropogenic Hg
2

inventories were found to vary slightly in Lakes Michigan (0.08 – 0.17 µg Hg/cm ) and
2

Superior (0.02 – 0.15 µg Hg/cm ), with anthropogenic inventories decreasing from the
South to the North (Long et al., 1995). If the range of inventories determined in the
21

2

Great Lakes represents a regional estimate of Hg loading for this region (both the Upper
and Lower Peninsulas of Michigan), then deviations from these patterns suggest the
influence of sub-regional to local scale sources. Most anthropogenic inventories from
inland lakes are consistent with a regional source; however, notable exceptions include,

2.0
1.9
1.8
1.7
1.6 South
1.5
1.4
1.3
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

North

Lk Michigan
Lk Superior

Birch
Gull
Crystal M
Littlefield
Cadillac
Houghton
Higgins
Crystal B
Shupac
Elk
Torch
Avalon
Mullet
Round D
Imp
Witch
Round
Gratiot

Anthropogenic Hg Inventory (µ g/cm 2 )

Higgins, Houghton and Gull lakes.

Figure 2.2. Focusing-corrected anthropogenic Hg inventories from 18 inland Michigan
lakes. Inventories are plotted along a south to north gradient. The boxes represent the
range of anthropogenic inventories for Lake Michigan and Lake Superior (Long et al.,
1995).

22

The relative importance of local sources can be determined by comparing the
anthropogenic inventories of lakes from the same geographic region because these lakes
would be expected to have similar loadings due to long range atmospheric loading so that
differences could be attributed to more local sources. Lakes Higgins and Houghton, in
the central Lower Peninsula, and Elk and Torch lakes, in the NW Lower Peninsula, are
examples of geographically proximal lakes. Anthropogenic inventories differ among
these pairs of lakes indicating that local, likely watershed-scale, sources were
contributing to Hg loadings. Crystal M and Littlefield lakes lie within the central portion
of the state and have similar anthropogenic inventories. Inventories for these lakes are
within the range of those from Lake Michigan suggesting that Hg loading in this area is
most likely due to regional sources of Hg. The differences among the Upper Peninsula
lakes, Gratiot, Imp, Round D, and Round, imply that local sources contribute
significantly to Hg loadings; however, these lakes are within the range of anthropogenic
inventories found in Lake Superior. Portions of the Upper Peninsula are more
mineralized than other areas of the state and have undergone intense mining, which may
have lead to increased export of Hg from the watershed (Kerfoot et al., 1999, Kerfoot et
al., 2004).
Temporal Trends. In general, anthropogenic accumulation began in the late
1800s and reached nominal peak values during the 1950s to 1980s depending on the lake
(Figures 1.3-1.7). Episodic events have been superimposed on this general trend. These
events may be unique to a lake or may be common among many lakes in a sub-region.
Episodic increases common among lakes indicate that they resulted from increased
accumulation and not post-depositional mobilization or spatial variability within the

23

sediments. Episodic increases that are common among lakes may be due to a common
source or watershed processes that are common to all lakes. Profiles are grouped by subregion to highlight accumulation rate patterns. Profiles for all of the other lakes that were
studied can be found in Appendix B.
Birch

Muskegon

Paw Paw

Gull

2000

1950

1900

1850

0

20

40 0

500 0
200
-2 -1
FCAAR (µg m y )

210

400 0

100

200

Figure 2.3.
Pb dated focusing-corrected anthropogenic accumulation rates for four
southwestern Michigan lakes, dates older than 1850 were truncated to highlight more
recent loading. Note the change in scale of the x-axis.

24

Cass

Otter

Sand

Whitmore

2000

1950

1900

1850

0

200

400 20

40
60 0
20
-2 -1
FCAAR (µg m y )

210

40 0

50

100

Figure 2.4.
Pb dated focusing-corrected anthropogenic accumulation rates for 4
southeastern Michigan lakes. Note the change in scale of the x-axis.

25

Gratiot

Imp

Round

Round D

10

20 0
10 20 0
-2 -1
FCAAR (µg m y )

Witch

2000

1950

1900

1850

0

50

100 0

210

10

20 0

50

Figure 2.5.
Pb dated focusing-corrected anthropogenic accumulation rates for five
Upper Peninsula of Michigan lakes, dates older than 1850 were truncated to highlight
more recent loading. Note the change in scale of the x-axis.

26

Higgins

Houghton

2000

1950

1900

1850

1800

1750

1700

0

50

100

150
0
-2 -1
FCAAR (µg m y )

210

500

1000

Figure 2.6.
Pb dated focusing-corrected anthropogenic Hg accumulation rates for two
north-central Michigan lakes. Note the change in scale of the x-axis.

27

Crystal B

Elk

Mullett

Torch

2000

1950

1900

1850

0

20

40 0

100
200 0
10
-2 -1
FCAAR (µg m y )

210

20 0

10

Figure 2.7.
Pb dated focusing-corrected anthropogenic accumulation rates for four
northwestern Michigan lakes, dates older than 1850 were truncated to highlight more
recent loading. Note the change in scale of the x-axis.

28

20

Anthropogenic Hg profiles from lakes in southwestern Michigan reflect the
proximity to industrialized areas and also include several episodic accumulation events
(Figure 2.3). The proximity of this sub-region to the Chicago, IL/Gary, IN area suggests
that these lakes should reflect the impact of Hg from this highly industrialized region. It
has been suggested previously that the Chicago/Gary area is a source of Hg to southern
Lake Michigan (Landis and Keeler, 2002, Landis et al., 2002) and it has been proposed
that the Chicago/Gary industrial complex was a major source for lead, cadmium and zinc
to these study lakes (Yohn, 2004). The similarity, excluding episodic accumulation
events, of the loading profiles observed for Gull and Birch lakes implies a source of Hg
common to both lakes was present until the early 1990s, although the onset of the
anthropogenic Hg accumulation occurred earlier in Birch Lake. The recent recovery of
Gull Lake to a rate of anthropogenic loading that is near zero is notable, since this was
observed in only 2 of the 26 study lakes studied (see Littlefield Lake, Figure S1 in
Appendix B ). Accumulation rates in Birch and Muskegon lakes decreased toward the
surface whereas the rates in Lake Paw Paw have increased. The uniqueness of the
accumulation rates observed near the surface of the core from Lake Paw Paw and the
unique episodic events (ca 1960 and 1984) suggests a local source. Paw Paw Lake was
dated with a linear dating model using a single sedimentation rate (see Appendix B,
Table S1). Meaning, in Paw Paw Lake, episodic events were due to increased Hg
concentration on the sediment and not increased sediment delivery. Episodic events that
were common among all southwestern lakes include a peak in the early 1990s and early

29

to mid 1970s. Great anthropogenic Hg accumulation rates in Paw Paw and Muskegon
lakes suggest that local-scale sources are significant contributors of Hg to these lakes.
Southeastern Michigan is a highly populated and industrialized area of the state.
This area also has the highest number of coal-fired utilities in the State; although all of
these facilities lie to the east of the study lakes (Figure 2.1). Winds are generally
westerly or southerly in this region. Mercury accumulation rates are greater in lakes from
this sub-region which may be a consequence of greater atmospheric deposition in this
region compared to less urban regions of Michigan (2005). Mercury in the cores from
Whitmore, Cass, and Otter lakes exhibit large anthropogenic accumulation rates with
peak rates that were amongst the greatest measured for all of the lakes studied (Figure
2.4). Lakes Otter and Whitmore, which are situated in the northern part of this subregion, exhibit rapid increases in anthropogenic Hg accumulation during the latter part of
the 1940s into the early 1950s. This implies a common source or a source that is
common to both lakes of Hg loading to both lakes during that period. Furthermore,
several of the episodic increases in Hg loading occurred simultaneously among these
lakes, suggesting a source of Hg loading common to both lakes. Episodic events in the
early and mid 1990s are common among all of the southeastern lakes. During the early
2000s, episodic events occurred in Sand and Otter Lakes. When accumulation profiles in
these lakes were compared to those observed in the surficial portion of the core from
Whitmore Lake it is possible that Whitmore Lake is currently experiencing a spike in Hg
accumulation. If Whitmore Lake were revisited, the most recent sediments should show a
decline in accumulation rate. Dating-model error, measurement error and field
methodologies all contribute to possible errors in sediment core chronologies. However,

30

the observation that so many episodic events “line-up” among lakes suggests that the
events are real. Similar to Paw Paw and Muskegon lakes, the accumulation rates in Cass
Lake were anomalously great; these rates suggest a unique source or a sub-regional
source in close proximity to the lake. Recently, decreases in Hg loading have been
observed among all of the lakes in the southeastern sub-region (Figure 2.1), generally
from maximal loading rates, in anthropogenic Hg accumulation. Sediment cores from
lakes in urban areas of Minneapolis, MN have demonstrated that loading rates of Hg have
decreased in recent years (Engstrom and Swain, 1997). The hypothesis stated for the
decreased Hg loading to urban Minneapolis lakes is that limited growth in already builtup areas has slowed the potential for new sources of Hg to the lakes. This also
adequately explains the recent declines observed for southeastern Michigan (suburban
Detroit and Flint) lakes in our study; although the reasons for episodic increases are
unclear at this time.
Rates of anthropogenic Hg loading to five Upper Peninsula (UP) lakes, Gratiot,
Imp, Round D, Round, and Witch, appears to be sub-regional in nature (Figure 2.5).
Other than the episodic accumulation events, this conclusion is further supported by the
similarity of profiles in the eastern portion of the UP (Round D and Round) compared to
those in the western UP (Imp and Witch). Both Round D and Round lakes appear to have
reached a steady state accumulation rate shortly after reaching maxima in the early 1900s
and 1950s respectively. Then, during the late 1970s, Hg accumulation rates in both lakes
increased and followed similar trajectories. Witch and Imp lakes both reached maximal
Hg accumulation rates in the 1960s followed by a steady decline until the 1990s. After
1990, both Imp and Witch lakes exhibited greater accumulation rates until 2000, followed

31

by a more recent decrease. Except for a unique accumulation rate profile in more recent
times, certain episodic increases in Hg accumulation during the 1910s, 1940s, and 1980s
in Gratiot Lake were common to other western UP lakes, suggesting that Gratiot Lake
would be included in this sub-region. The episodic events after 1975 in Gratiot were
superimposed on a generally increasing trend. The extent of these events suggests that
the source was local, most likely watershed, in scale.
Anthropogenic Hg accumulation rate profiles from Lakes Higgins and Houghton
in the north-central portion of the Lower Peninsula of Michigan exhibit simultaneous
occurrence of episodic increases (Figure 2.6). A notable deviation was observed during
th

the mid to late 18 century in Houghton Lake. Reasons for not observing this
phenomenon in other Michigan lakes that contain sediment of adequate age to record the
disturbance are not yet clear, but may be due to the unique geography of the Houghton
Lake area. Higgins Lake lies 14 km due north of Houghton Lake, and their watersheds
share a common boundary. Houghton Lake lies at or near the headwaters of three major
river systems, two of which flow into Lake Michigan (Muskegon and Manistee rivers)
and one into Lake Huron (Au Sable River). Native Americans commonly used rivers as
efficient transportation corridors from West to East (Quimby, 1962) and the advantage of
efficient travel may have led to early settlement. Native American tribes during this
period were known to travel inland over 100 km from the Lake Michigan shoreline
(Quimby, 1962). Other metals, such as, aluminum (Al), zinc and cadmium, do show an
accumulation rate increase during this period suggesting a terrestrial input (Yohn et al.,
2003). The disturbance that led to the increase of these metals is unclear, but may be due
to the early trading and land-clearing activities in the region.
32

Prior to 1930, the Hg accumulation rate profiles of 4 northwestern Michigan
lakes, Elk, Torch, Crystal B and Mullett were similar (Figure 2.7), which suggests a
regional to sub-regional source of Hg. After the 1930s, lakes Mullett and Torch appear to
have been influenced by a more sub-regional source, whereas Elk and Crystal lakes show
the effects of more local sources. Loading rates of Al to Elk Lake decreased during the
th

mid-20 century episodic Hg increase. These two observations, taken together, suggest
that the source was a point source or local atmospheric deposition rather than terrestrial
(Long et al., 2000). Mercury accumulation rates in Crystal B Lake were found to be
increasing towards the sediment water interface, which is unique for this sub-region,
suggesting a local source.
While it is clear that some episodic accumulation events are due to local-scale
sources it is unclear at this time the reasons for these events. Possible sources being
considered include: waste incinerators (Liu et al., 2007), fuel combustion (Lynam and
Keeler, 2006), dental office waste (Drummond et al., 2003), forest fire (Turetsky et al.,
2006), clear-cutting of forest (Porvari et al., 2003), pulp and paper mills (Hynynen et al.,
2004), and cement manufacturing (Pacyna et al., 2006).

Recent Trends. Hg concentration profiles in sediments from several lakes have
recorded recent increases in Hg loading (Engstrom and Swain, 1997, Lorey and Driscoll,
1999, Perry et al., 2005, Swain et al., 1992). Over the last decade in Michigan, 8 of the
26 lake profiles showed anthropogenic Hg accumulation rates that increased toward the
surface, 11 showed trends that indicate anthropogenic loading had declined, 5 showed
elevated accumulation rates but no clear increasing or decreasing trend, and 2 had
33

anthropogenic accumulation rates at background levels (Figure 2.1). Considering the
lakes experiencing increased loading, the lack of a spatial trend suggests that the source
was not regional in scale. Lakes that are increasing are not located in areas of the state
that have a high density of coal-fired power plants (Figure 2.1), further evidence of the
importance of local-scale sources of Hg to these lakes. In addition, these results are
inconsistent with a decrease in Hg emissions (Engstrom and Swain, 1997). Possible local
sources include: mobilization from watershed soils, industry, and agriculture. The
increase observed in sediment cores from Adirondack lakes have been attributed to a
decrease in the retention of Hg by watershed soils (Lorey and Driscoll, 1999). Some
researchers have attributed the increase recorded in western Minnesota lakes to Hg-laden
soils delivered from agricultural fields (Engstrom and Swain, 1997). At this time it is
unclear as to why we observe increased accumulation rates in more recent periods but
one or all of the above may have contributed.

34

REFERENCES

35

REFERENCES
Michigan Department of Environmental Quality, 2007. Sediment Chemistry, Available
online at http://www.michigan.gov/deq/0,1607,7-135-3313_3686_3728-32365-,00.html. (Accessed: January 15, 2007).
Bindler, R., Renberg, I., Appleby, P.G., Anderson, N.J., Rose, N.L., 2001. Mercury
accumulation rates and spatial patterns in lake sediments from west Greenland: A
coast to ice margin transect. Environmental Science and Technology 35, 17361741.
Drummond, J.L., Cailas, M.D., Croke, K., 2003. Mercury generation potential from
dental waste amalgam. Journal of Dentistry 31, 493-501.
Engstrom, D.R., Swain, E.B., 1997. Recent declines in atmospheric mercury deposition
in the upper Midwest. Environmental Science and Technology 31, 960-967.
Engstrom, D.R., Swain, E.B., Henning, T.A., Brigham, M.E., Brezonik, P.L., 1994.
Atmospheric Mercury Deposition to Lakes and Watersheds - a Quantitative
Reconstruction from Multiple Sediment Cores. In: Environmental Chemistry of
Lakes and Reservoirs. pp. 33-66.
ARC/INFO, Environmental Systems Research Institute, 1998. ARC/INFO
Fitzgerald, W.F., Engstrom, D.R., Mason, R.P., Nater, E.A., 1998. The case for
atmospheric mercury contamination in remote areas. Environmental Science and
Technology 32, 1-7.
Gobeil, C., MacDonald, R.W., Smith, J.M., 1999. Mercury profiles in sediments of the
Arctic Ocean basin. Environmental Science and Technology 33, 4194-4198.
Golden, K.A., Wong, C.S., Jeremiason, J.D., Eisenreich, S.J., Sanders, M.G., Hallgren, J.,
Swackhammer, D.L., Engstrom, D.R., Long, D.T., 1993. Accumulation and
preliminary inventory of organochlorines in Great Lakes sediments. Water
Science and Technology 28, 19-31.
Gray, J.E., Fey, D.L., Holmes, C.W., Lasorsa, B.K., 2005. Historical deposition and
fluxes of mercury in Narraguinnep Reservoir, southwestern Colorado, USA.
Applied Geochemistry 20, 207-220.
Grigal, D.F., 2003. Mercury sequestration in forests and peatlands: A review. Journal of
Environmental Quality 32, 393-405.
Hermanson, M.H., 1998. Anthropogenic mercury deposition to arctic lake sediments.
Water, Air, and Soil Pollution 101, 309-321.

36

Hightower, J., Moore, D., 2003. Mercury Levels in High-End Consumers of Fish.
Environmental Health Perspectives 111, 604-608.
Hynynen, J., Palomaki, A., Merilainen, J.J., Witick, A., Mantykoski, K., 2004. Pollution
history and recovery of a boreal lake exposed to a heavy bleached pulping
effluent load. Journal of Paleolimnology 32, 351-374.
Kamman, N.C., Engstrom, D.R., 2002. Historical and present fluxes of mercury to
Vermont and New Hampshire lakes inferred from Pb-210 dated sediment cores.
Atmospheric Environment 36, 1599-1609.
Keeler, G.J., Dvonch, J.T., 2005. Atmospheric mercury: A decade of observations in the
Great Lakes. In: N. Pirrone, K. Mahaffey (Eds.), Dynamics of Mercury Pollution
on Regional and Global Scales: Atmospheric Processes and Human Exposures
around the World. Kluwer Ltd., Norwell, MA.
Kerfoot, W.C., Harting, S., Rossmann, R., Robbins, J.A., 1999. Anthropogenic copper
inventories and mercury profiles from Lake Superior: Evidence for mining
impacts. Journal of Great Lakes Research 25, 663-682.
Kerfoot, W.C., Harting, S.L., Jeong, J., Robbins, J.A., Rossmann, R., 2004. Local,
regional, and global implications of elemental mercury in metal (copper, silver,
gold, and zinc) ores: Insights from Lake Superior sediments. Journal of Great
Lakes Research 30, 162-184.
Kolak, J.J., Long, D.T., Beals, T.M., Eisenreich, S.J., Swackhamer, D.L., 1998.
Anthropogenic inventories and historical and present accumulation rates of
copper in Great Lakes sediments. Applied Geochemistry 13, 59-75.
Landis, M.S., Keeler, G.J., 2002. Atmospheric mercury deposition to Lake Michigan
during the Lake Michigan Mass Balance Study. Environmental Science and
Technology 36, 4518-4524.
Landis, M.S., Vette, A.F., Keeler, G.J., 2002. Atmospheric mercury in the Lake Michigan
basin: Influence of the Chicago/Gary urban area. Environmental Science and
Technology 36, 4508-4517.
Liu, B., Keeler, G.J., Dvonch, J.T., Barres, J.A., Lynam, M.M., Marsik, F.J., Morgan,
J.T., 2007. Temporal variability of mercury speciation in urban air. Atmospheric
Environment 41, 1911-1923.
Long, D.T., Eisenreich, S.J., Swackhammer, D.L. Volume III - Metal Concentrations in
Sediments of the Great Lakes; Government Printing Office, 1995.
Long, D.T., Simpson, S.S., Geisey, J.P., Fett, J.D., 2000. Inland lakes sediment trends:
sediment analysis results for five Michigan Lakes. Lansing, MI. 17 pages.
37

Available from: http://www.deq.state.mi.us/documents/deq-swq-gleas9900sedtrendreport.pdf.
Lorey, P., Driscoll, C.T., 1999. Historical trends of mercury deposition in Adirondack
lakes. Environmental Science and Technology 33, 718-722.
Lynam, M.M., Keeler, G.J., 2005. Automated speciated mercury measurements in
Michigan. Environmental Science and Technology 39, 9253-9262.
Lynam, M.M., Keeler, G.J., 2006. Source-receptor relationships for atmospheric mercury
in urban Detroit, Michigan. Atmospheric Environment 40, 3144-3155.
Meili, M., 1991. The Coupling of Mercury and Organic-Matter in the Biogeochemical
Cycle - Towards a Mechanistic Model for the Boreal Forest Zone. Water, Air, and
Soil Pollution 56, 333-347.
Menounou, N., Presley, B.J., 2003. Mercury and other trace elements in sediment cores
from central texas lakes. Archives of Environmental Contamination and
Toxicology 45, 11-29.
Pacyna, E.G., Pacyna, J.M., Fudala, J., Strzelecka-Jastrzab, E., Hlawiczka, S., Panasiuk,
D., 2006. Mercury emissions to the atmosphere from anthropogenic sources in
Europe in 2000 and their scenarios until 2020. Science of the Total Environment
370, 147-156.
Perry, E., Norton, S.A., Kamman, N.C., Lorey, P.M., Driscoll, C.T., 2005.
Deconstruction of Historic Mercury Accumulation in Lake Sediments,
Northeastern United States. Ecotoxicology 14, 85-99.
Porvari, P., Verta, M., Munthe, J., Haapanen, M., 2003. Forestry practices increase
mercury and methyl mercury output from boreal forest catchments.
Environmental Science and Technology 37, 2389-2393.
Quimby, G.I., 1962. A Year with a Chippewa Family, 1763-1764. Ethnohistory 9, 217239.
Scherbatskoy, T., Shanley, J.B., Keeler, G.J., 1998. Factors controlling mercury transport
in an upland forested catchment. Water Air and Soil Pollution 105, 427-438.
Swain, E.B., Engstrom, D.R., Brigham, M.E., Henning, T.A., Brezonik, P.L., 1992.
Increasing Rates of Atmospheric Mercury Deposition in Midcontinental NorthAmerica. Science 257, 784-787.
Turetsky, M.R., Harden, J.W., Friedli, H.R., Flannigan, M., Payne, N., Crock, J., Radke,
L., 2006. Wildfires threaten mercury stocks in northern soils. Geophysical
Research Letters 33.
38

U.S. EPA, 2004. National Listing of Fish Advisories, Available online at
http://www.epa.gov/waterscience/fish/. (Accessed: September 25, 2011).
State Soil Geographic (STATSGO) Database for Michigan, Available online at
http://www.ncgc.nrcs.usda.gov/products/datasets/statsgo/index.html. (Accessed:
September 25, 2011).
USEPA. Office of Solid Wastes. Mercury in solids and solutions by thermal
decomposition, amalgamation, and atomic absorption spectrophotometry;
Government Printing Office, 1998.
USEPA. Office of Research and Development Office of Air Quality Planning and
Standards. Mercury Study Report to Congress; Washington, DC: Government
Printing Office, 1997.
Wihlborg, P., Danielsson, A., 2006. Half a century of mercury contamination in Lake
Vanern (Sweden). Water, Air, and Soil Pollution 170, 285-300.
Yang, H.D., Rose, N.L., 2003. Distribution of mercury in six lake sediment cores across
the UK. Science of the Total Environment 304, 391-404.
Yohn, S., Long, D., Fett, J., Patino, L., 2004. Regional versus local influences on lead
and cadmium loading to the Great Lakes region. Applied Geochemistry 19, 11571175.
Yohn, S.S. 2004. Natural and Anthropogenic Factors Influencing Spatial and Temporal
Patterns of Metal Accumulation in Inland Lakes. Thesis, Michigan State
University, East Lansing, MI.
Yohn, S.S., Long, D.T., Fett, J.D., Patino, L., Giesy, J.P., Kannan, K., 2002a. Assessing
environmental change through chemical-sediment chronologies from inland lakes.
Lakes Reserv Res Manage 7, 217-230.
Yohn, S.S., Long, D.T., Giesy, J.P., Fett, J.D., Kannan, K., 2001. Inland lakes sediment
trends: sediment analysis results for two Michigan Lakes. Lansing, MI. 49 pages.
Available from:
Yohn, S.S., Long, D.T., Giesy, J.P., Scholle, L.K., Patino, L.C., Fett, J.D., Kannan, K.,
2002b. Inland lakes sediment trends: sediment analysis results for five Michigan
Lakes. East Lansing. 64 pages. Available from:
Yohn, S.S., Long, D.T., Giesy, J.P., Scholle, L.K., Patino, L.C., Parsons, M., Kannan, K.,
2003. Inland Lakes Sediment Trends: sediment analysis results for six Michigan
Lakes. East Lansing. 66 pages. Available from:
http://www.deq.state.mi.us/documents/deq-wb-swassedimenttrend0203finalreport.pdf.

39

CHAPTER 3
A REGIONAL ANALYSIS ON THE EFFECTS OF WATERSHED ATTRIBUTES ON THE
RECOVERY OF INLAND LAKES FROM MERCURY ENRICHMENT.
Abstract
Sediments from inland lakes record recent declines in mercury (Hg) loading to the
environment. The rate of decline or recovery rate varies among lakes which we hypothesize to
be due to differences in how watershed attributes influence the transport of Hg through the
watershed. Sediment cores were collected, dated using lead-210, and analyzed for Hg from 13
inland lakes throughout the State of Michigan to test this hypothesis. Slopes of linear recovery
calculated from the most recent declines in concentration and accumulation rate were compared
to watershed attributes using Pearson’s correlation coefficient. Agricultural, forested and
wetland attributes were all important determinants of recovery rate. However, the strongest
correlations were found for land cover types along highly conductive watershed flow paths or
within close proximity to the lake. These results highlight the importance of watershed attributes
and the transport of Hg from the catchment to the lake. However, this work also emphasizes the
importance of location within the watershed as an important determinant of contemporary
recovery.

Introduction
How aquatic systems are recovering from man-caused stresses continues to be a critical
question for both researchers and natural resource managers. Of particular concern is recovery
from mercury (Hg) contamination because organic forms bioaccumulate in aquatic food webs
and consumption of tainted fishes can impact human health (Hightower and Moore, 2003,
USEPA, 1997). Environmental archives (e.g., sediment chronologies) have been shown to be
reliable recorders of anthropogenic activities and have demonstrated the efficacy of
40

environmental legislation (Bindler et al., 2001, Engstrom and Swain, 1997, Kamman and
Engstrom, 2002, Long et al., 2010, Lorey and Driscoll, 1999, Swain et al., 1992). However for
Hg, sediment concentrations and fluxes for most lakes are still elevated above geochemical
background values (Bindler et al., 2001, Engstrom et al., 2007, Kamman and Engstrom, 2002,
Parsons et al., 2007, Swain et al., 1992) This condition is in part due to continued Hg emissions
and the transport of previously deposited Hg, so called “legacy Hg”, from the watershed to the
aquatic ecosystem due to natural processes that might be exacerbated by human disturbances.
However, as anthropogenic Hg emissions continue to decrease, which some have suggested for
the Great Lakes region (Evers et al., 2011, Keeler and Dvonch, 2005), one might hypothesize
that transport through the watershed might play an ever more important role in the rates of
recovery of ecosystems from Hg contamination. The activities that result in differences in Hg
transport through the watershed will need to better defined, especially if those activities result in
a slower recovery
Sediment profiles of heavy metals (e.g., Cd, Cu, Pb, Zn and Hg) when influenced by
human activities consist of three primary temporal phases, the first being a time which can be
considered geochemical background, or reference condition, which in the State of Michigan
occurs prior to European settlement, ca. 1850. Throughout inland lakes of Michigan the next
phase recognized in sediment profiles is the onset of deforestation. This is followed by the
increased production and consumption of heavy metals during the industrial revolution (Long et
al., 2010) which is evidenced by rapidly increasing sediment concentrations and fluxes starting
th

in the late 19 century. The final phase occurs after the passage of environmental legislation,
i.e., Clean Air Act (US EPA, 2008), when the loading of metals to the environment significantly
decreased. This final, or recovery, phase is evidenced by decreasing sediment concentrations
41

and fluxes. The rate of recovery, defined here as the slope of decreasing concentrations or
fluxes, can differ among lakes which may be due in part to proximity to local sources, but also
due to differences in watershed attributes or differing rates of landscape disturbance.

Export rates of Hg from watersheds have been shown to be influenced by land use
(Babiarz et al., 1998, Hurley et al., 1995b). Although the transport of Hg through the watershed
is not fully understood some generalizations can be made. All catchments are sinks for
atmospherically deposited Hg and rates of retention can be greater than 90% (Aastrup et al.,
1991, Harris et al., 2007). Those landscapes that experience higher rates of erosion (e.g.,
agricultural) would be expected to have a higher probability to export rather than retain Hg
(Munthe et al., 2007a). Urban areas, in the absence of local emitters, have been suggested to
export less Hg then most other land use types due to a lack of activities that disturb the landscape
(Engstrom and Swain, 1997). Forests have been shown to be sinks for Hg; but can release Hg
during storm events (Munthe et al., 2007b), forest fires (Turetsky et al., 2006), and logging
(Munthe and Hultberg, 2004, Porvari et al., 2003). There does not appear to be a consensus
about the effects of wetlands on Hg export. Some have shown that wetland type or presence
does not influence Hg export (St. Louis et al., 1996), while others have found that Hg is less
likely to be retained when wetlands are present (Hurley et al., 1995a, Mierle and Ingram, 1991).

Both sediment concentration and accumulation rate chronologies are commonly used to
represent system state due to the unique information they provide about environmental condition.
Concentration profiles provide information about environmental exposure that may be useful to
natural resource managers and differences with respect to watershed concentrations that aid in

42

determining the state of recovery of the system from disturbance (see Long et al., 2010).
Knowledge of the accumulation rates can be important since concentration can be diluted or
amplified depending upon the changes in flux of allochthonous and autochthonous materials.
For example, if the flux of Hg to sediment were fixed, consequences of an increase in lake
productivity (e.g., increased calcium carbonate deposition) would dilute the Hg concentration
recorded in the sediment. However, the Hg accumulation rate would remain unchanged since the
increase in sedimentation rate would offset the decrease in Hg concentration. Likewise if the
source of the terrestrial input would for some reason change, e.g., logging activities strip native
top soil, see (Long et al., 2010), and the atmospheric flux of Hg is again fixed then the
concentration might be expected to decrease. The Hg accumulation rate would increase;
however, since the flux of soil material would be expected to be much greater than under
previous conditions. Under this set of circumstances the definition of recovery is confounded
since concentrations are decreasing, an indication of recovery, but accumulation rate is
increasing. A study of system recovery then must compare concentrations and accumulation
rates since each can provide alternative interpretations or additional information about movement
of Hg through the watershed.

Sediment cores used for this study were collected from 14 inland lakes throughout the
State of Michigan (Figure 3.1) as part of a larger Inland Lakes Trends Monitoring Program
funded by the Michigan Department of Natural Resources and Environment. Lakes included in
this study vary in watershed land use from highly urban landscapes in southeast Michigan to the
remote, highly forested region, of Michigan’s Upper Peninsula (Table 3.1). Slopes of the
concentrations and flux declines were compared against watershed attributes in order to

43

investigate the relationship between the recovery of inland lakes from anthropogenic Hg
contamination and watershed land use.

¯
!Round D
(
!
(

Charlevoix !
Mullett
(
! Elk
(
!
(
Torch
Houghton !
(

George
!
(

!Hackert
(
!Nichols
(
! Muskegon
(

! Otter
(
!
(

! Gull
(

100
Kilometers

! Birch
(

Figure 3.1. Location of study lakes.
44

Cass

Methods
Core collection and recovery estimates. Sediment cores were collected using an Ocean
Instruments MC-400 from the deepest portion of the lake from the MDNRE R/V Nibi or US EPA
R/V Mudpuppy between 1999 and 2007. Sediment cores were sectioned on shore at 0.5 to 1.0
cm increments and stored on ice in acid-cleaned polypropylene jars during. In the laboratory,
samples for mercury analysis were transferred to Whirl-Pak bags and lyophilized. Mercury
analysis was performed using a Lumex Mercury Analyzer following US EPA Method 7473
(USEPA, 1998). Samples for radiometric dating were submitted to the Freshwater Institute at
the University of Manitoba. The radionuclide

210

Pb was measured in every slice to determine
137

sedimentation rates and assign dates and cesium-137 (

Cs) was measured in select slices to

determine peak activities. Cesium-137 analyses were performed by direct counting of 5-10 g of
sediment on a γ-spectrometer (Ge-Li) semiconductor detector. Samples of 1-3 g were analyzed
for

210

Pb and radium-226 by leaching in 6N hydrochloric acid in the presence of a polonium-209

tracer, autoplating on a silver disk, and counting the disk on an α-spectrometer, determining
210

Pb via its polonium-210 daughter product. Sediment slice ages were estimated using the

constant rate of supply (CRS) (Appleby and Oldfield, 1978), constant flux:constant
sedimentation (CF:CS) (Golden et al., 1993), or segmented CF:CS model (Heyvaert et al.,
2000). Assigned dates were verified using peak activities of
Pb (Alfaro-De la Torre and Tessier, 2002).

45

137

Cs (Appleby, 2001) and stable

Recovery rates for concentration and accumulation rates were determined by calculating
the slope of the most recent decreasing trend of Hg, measured from the most recent peak or
change in slope trajectory. To avoid inclusion of episodic or short term increases (Figure 3.2) in
the calculation of slope the Hg profile was first smoothed using smooth command in Matlab
(Mathworks, 1994 - 2008). Slopes were compared to watershed attributes using Pearson’s
product-moment correlation coefficient (r) which assumes a normal distribution and where
necessary data were transformed to meet this assumption. Statistical analyses were performed in
the statistical environment R (The R Foundation for Statistical Computing, 2008). Outliers, as
indicated by statistical boxplots, were removed so that these data did not overly influence the
results. Only statistically significant results (p ≤ 0.10) are reported. In the case of zeros in the
dataset and the requirement to transform the data a constant was added to each original value
equivalent to the value of the first quartile squared divided by the value of the third quartile.

Determination of watershed attributes. Land use data were obtained from the 2001
IFMAP/GAP Lower and Upper Peninsula datasets (Michigan Department of Natural Resources,
2003) for each watershed. Land use was classified using the upper-most level Anderson
classification and resulted in 7 land use types: urban (URB), agriculture (AG), forest (FOR),
barren (BAR), upland (UPL), water (WAT) and wetland (WET). Lowland forests originally
included as wetland were reclassified as forested land use. Flow inverted (FLI) and flow
inverted squared (FLI2) land use values, for the watershed and within 100 m of the lake
shoreline, were calculated by determining the flow length (i.e., the pathway a drop of water
would take to the lake when dropped on the watershed) using the FLOWLENGTH command in
2

ArcGIS (ESRI, 1998). The inverse (1/flowlength) and inverse squared (1/flowlength ) of the

46

2

flowlength were then multiplied by the land use with the 30 m grid cell of individual land use
coverages assigned a 1, for presence, or 0 for absence. This method has the potential to provide
information about land use in close proximity or along highly conductive flow paths that would
not be represented by other measures. All land use variables were also calculated for a 100 m
buffer around the lake shoreline and converted to proportions for this analysis. By convention
land use type will precede FLI and FLI2 to indicate these transformations (e.g., AG.FLI is read
flow length inverted agriculture). Likewise land use calculated within the buffer is given the
suffix M (e.g., AG.FLI.M is read flow length inverted agriculture in the 100 m buffer).
Sulfate deposition (SD), used here as a proxy for coal combustion (Yohn, 2004), was
estimated for the region by using inverse distance weighting of the 1990-2009 average sulfate
deposition values for 19 National Atmospheric Deposition Program stations around the Great
Lakes region (Illinois State Water Survey, 2008). Soil erodibility, expressed here as the k-factor
(K.F) and slope corrected k-factor (K.FF), was estimated using dasymetric mapping techniques
using the State Soil Geographic Database (STATSGO) (USDA, 1994).
Four physical variables were investigated along with land cover attributes: latitude,
watershed area (WA), calculated with the lake removed, lake area (LA), and WA:LA.
Watersheds were delineated from National Elevation Dataset (NED) 1 arc second digital
elevation models (DEM), available from the USGS seamless website (USGS, 2008), using the
watershed command in ESRI ArcGIS (ESRI, 1998). Others have shown that WA:LA is an
important determinant for sediment Hg fluxes (Swain et al., 1992). Latitude was investigated as
a proxy for atmospheric deposition rates which have been shown to decrease from the South to
the North in this region (Keeler and Dvonch, 2005). Since latitude only captures a regional trend
Hg emissions estimates were included to detect possible correlations to local sources that may be
47

unique to a sub-region or an individual lake. Mercury emissions were calculated using the 2002
U.S. EPA Hg emissions estimates for the State of Michigan (USEPA, 2006) for 50 and 100 km
buffers around the lake. A complete list of the watershed characteristics for lakes in this study
can be found in Table 1.

48

Table 3.1. Watershed attributes of study lakes: watershed area (WA), lake area (LA), watershed
to lake area ratio (WA:LA), urban (URB), agriculture (AG), forest (FOR), upland (UP), barren
(BAR), and wetland (WET).
Flux
Concentration
Recovery
Recovery
Rate
Rate
2
(µg/m /y/y)
(µg/kg/y)
Lake
Birch
-2.7
-0.22
Cass
-1.9
-1.14
-1
-0.57
Charlevoix
Elk
-1.5
-0.25
George
-5.6
-1.45
Gull
-1.3
-0.47
Hackert
-4.3
-0.57
Houghton
-4.5
-0.55
Mullett
-0.3
-0.09
Muskegon
-4.5
-2.74
Nichols
-6.5
-0.41
Otter
-1.8
-0.52
Round D
-7.8
-1.13
Torch
-1.6
-0.27

Lake
Birch
Cass
Charlevoix
Elk
George
Gull
Hackert
Houghton
Mullett
Muskegon
Nichols
Otter
Round D
Torch

URB
6.40%
24.80%
3.40%
4.20%
5.00%
5.60%
5.90%
3.70%
2.80%
7.00%
0.90%
8.30%
0.20%
4.20%

AG
48.30%
0.50%
17.30%
20.40%
6.30%
59.80%
33.80%
0.80%
7.20%
17.20%
0.00%
4.80%
0.00%
18.60%

Latitude
41.8669
42.6036
45.2632
44.8690
44.3997
42.3963
43.9827
44.3203
45.4850
43.2333
43.7292
43.2164
46.1462
44.9326

FOR
27.80%
37.50%
61.00%
53.60%
59.90%
21.00%
42.30%
62.20%
68.10%
47.40%
86.40%
49.60%
78.60%
54.30%

49

WA
LA
2
2
(km ) (km ) WA:LA
16.9
1.2
14.2
38.7
5.2
7.5
765.1 69.8
11
85.8 31.3
2.7
5.1
0.8
6.3
55.3
8.2
6.8
1.6
0.5
3.2
382.7 81.2
4.7
1239.5 70.3
17.6
884.7 16.8
52.7
0.9
0.7
1.3
0.5
0.3
1.5
6.8
1.8
3.8
120.8 76
1.6

UP
7.80%
18.90%
14.80%
18.60%
17.40%
6.10%
10.20%
14.60%
17.70%
18.00%
2.00%
16.70%
3.40%
19.40%

BAR
0.10%
0.10%
0.30%
0.20%
0.20%
0.10%
0.20%
0.20%
0.20%
0.40%
0.20%
0.00%
0.10%
0.20%

WET
4.40%
11.00%
2.30%
1.80%
9.40%
4.90%
4.30%
16.40%
3.20%
7.00%
5.40%
17.90%
15.40%
1.50%

Table 3.1 (cont’d).
SD
Lake
(kg/ha/y) K.F K.FF
Birch
17.7
0.17 0.09
Cass
14.3
0.22 0.11
Charlevoix
11.8
0.17 0.19
Elk
13
0.19 0.27
George
12.2
0.22 0.04
Gull
16.5
0.26 0.12
Hackert
15.1
0.26 0.03
Houghton
13.5
0.13 0.06
Mullett
11.5
0.16 0.14
Muskegon
16.7
0.21 0.1
Nichols
15.2
0.17 0.09
Otter
12
0.21 0.1
Round D
8
0.15 0.14
Torch
12.6
0.19 0.29

Lake
Birch
Cass
Charlevoix
Elk
George
Gull
Hackert
Houghton
Mullett
Muskegon
Nichols
Otter
Round D
Torch

FOR.M
40.20%
42.60%
47.00%
46.20%
41.50%
45.00%
42.70%
19.70%
43.70%
5.30%
87.00%
45.60%
71.90%
44.80%

URB.M
23.10%
17.30%
1.60%
6.50%
12.80%
20.30%
11.20%
40.60%
13.30%
31.20%
3.00%
9.60%
0.00%
7.90%

UP.M BAR.M
10.00% 0.00%
9.80% 0.00%
15.30% 7.60%
10.20% 0.20%
14.10% 0.30%
12.20% 0.00%
10.00% 0.20%
8.60% 0.40%
8.00% 0.00%
16.50% 8.40%
3.70% 0.00%
14.90% 0.00%
4.10% 0.10%
12.20% 0.40%

50

AG.M
3.70%
0.00%
3.00%
7.10%
0.00%
1.00%
14.60%
0.00%
0.80%
0.20%
0.00%
0.00%
0.00%
2.90%

WET.M
21.20%
9.00%
11.90%
9.00%
8.10%
11.00%
6.80%
9.00%
16.20%
11.10%
2.20%
23.70%
3.20%
7.00%

FOR.M
40.20%
42.60%
47.00%
46.20%
41.50%
45.00%
42.70%
19.70%
43.70%
5.30%
87.00%
45.60%
71.90%
44.80%

Table 3.1 (cont’d)
Lake
URB.FLI
Birch
14.20%
Cass
19.30%
Charlevoix 11.60%
Elk
6.70%
George
11.40%
Gull
12.80%
Hackert
9.00%
Houghton
23.60%
Mullett
8.20%
Muskegon
26.80%
Nichols
0.30%
Otter
9.50%
Round D
0.20%
Torch
7.90%

AG.FLI FOR.FLI
28.40% 38.20%
0.10% 41.60%
11.60% 51.70%
17.10% 49.00%
2.20% 57.20%
23.90% 39.20%
21.40% 46.60%
0.80% 42.60%
5.60% 60.80%
3.30% 25.60%
0.00% 89.80%
2.20% 48.60%
0.00% 80.90%
7.90% 54.90%

UP.FLI BAR.FLI
9.70%
0.00%
14.00%
0.00%
15.00%
2.20%
14.40%
0.20%
16.70%
0.10%
11.20%
0.00%
10.40%
0.10%
11.70%
0.60%
13.90%
0.10%
17.80%
4.50%
2.60%
0.00%
18.10%
0.00%
2.00%
0.10%
16.30%
0.30%

Lake
WET.FLI URB.FLI2 AG.FLI2 FOR.FLI2 UP.FLI2
Birch
8.20%
20.00%
6.50%
42.20%
11.70%
Cass
10.90%
15.50%
0.00%
40.50%
9.10%
Charlevoix
5.00%
15.60%
4.00%
43.50%
13.60%
Elk
5.30%
4.50%
9.30%
50.60%
9.20%
George
7.20%
11.30%
0.20%
51.40%
14.40%
Gull
6.70%
14.30%
3.10%
43.30%
16.20%
Hackert
4.90%
10.20%
15.10%
46.10%
8.90%
Houghton
15.10%
36.40%
0.10%
27.20%
10.20%
Mullett
6.80%
9.90%
1.80%
52.90%
9.50%
Muskegon
9.30%
22.20%
0.20%
8.50%
17.10%
Nichols
3.60%
0.00%
0.00%
91.40%
3.20%
Otter
18.90%
7.10%
0.60%
45.70%
20.90%
Round D
5.90%
0.00%
0.00%
69.40%
1.60%
Torch
3.70%
6.30%
3.30%
49.80%
13.50%

51

Table 3.1 (cont’d).
Lake
BAR.FLI2 WET.FLI2 URB.FLI.M AG.FLI.M FOR.FLI.M
Birch
0.00%
19.00%
23.90%
3.30%
43.80%
Cass
0.00%
11.50%
18.00%
0.00%
40.30%
6.00%
8.60%
17.90%
3.10%
41.60%
Charlevoix
Elk
0.20%
8.90%
6.20%
7.80%
50.90%
George
0.00%
8.60%
15.10%
0.00%
47.40%
Gull
0.00%
10.10%
18.60%
1.30%
45.40%
Hackert
0.10%
5.50%
11.80%
14.80%
45.50%
Houghton
0.50%
10.80%
43.40%
0.00%
23.30%
Mullett
0.00%
12.20%
13.70%
1.10%
49.90%
Muskegon
7.30%
12.40%
28.10%
0.20%
5.30%
Nichols
0.00%
1.80%
0.50%
0.00%
90.70%
Otter
0.00%
21.20%
9.00%
0.00%
47.80%
Round D
0.10%
4.30%
0.00%
0.00%
69.40%
Torch
0.40%
6.80%
8.20%
2.90%
48.00%

Lake
UP.FLI.M BAR.FLI.M WET.FLI.M URB.FLI2.M AG.FLI2.M
Birch
11.50%
0.00%
17.10%
23.20%
2.20%
Cass
9.20%
0.00%
10.20%
17.30%
0.00%
Charlevoix
13.80%
6.10%
8.90%
18.50%
2.80%
Elk
9.30%
0.10%
9.20%
5.30%
7.50%
George
15.50%
0.00%
8.90%
13.40%
0.00%
Gull
14.80%
0.00%
9.60%
16.00%
1.20%
Hackert
9.50%
0.20%
5.40%
11.50%
13.40%
Houghton
9.70%
0.40%
8.70%
40.90%
0.00%
Mullett
8.90%
0.00%
12.90%
11.70%
1.10%
Muskegon
17.80%
7.20%
12.40%
26.10%
0.20%
Nichols
3.50%
0.00%
2.10%
0.00%
0.00%
Otter
18.30%
0.00%
21.20%
7.60%
0.00%
Round D
2.40%
0.20%
4.20%
0.00%
0.00%
Torch
12.80%
0.30%
7.00%
7.40%
2.60%

52

Table 3.1 (cont’d).
Hg.50 Hg.100
Lake
FOR.FLI2.M UP.FLI2.M BAR.FLI2.M WET.FLI2.M (Lbs/y) (Lbs/y)
Birch
40.60%
11.70%
0.00%
21.70%
67
178.6
Cass
38.40%
7.90%
0.00%
11.40%
1025.9 3127.7
Charlevoix
38.60%
13.30%
7.00%
9.40%
77.8
118.5
123.2
Elk
49.60%
7.70%
0.20%
9.80%
78.9
George
47.50%
13.30%
0.00%
9.20%
1.2
864
Gull
41.10%
16.60%
0.00%
10.90%
77.8
404
Hackert
45.30%
8.40%
0.10%
5.70%
56.5
222.4
129.1
Houghton
23.70%
9.40%
0.40%
9.20%
22.1
Mullett
50.00%
8.40%
0.00%
13.20%
21
110
Muskegon
4.60%
16.70%
6.90%
13.10%
170.2
255.5
317.7
Nichols
91.60%
3.40%
0.00%
1.40%
30.8
Otter
44.90%
20.90%
0.00%
21.80%
353.7 1306.2
Round D
64.50%
1.70%
0.10%
4.50%
30.6
74.7
Torch
46.30%
12.20%
0.40%
7.70%
78.9
119.6

53

Birch

Cass
2010

Charlevoix
2000

2000
2005

1980

1950
2000

1960

1900
1940

1995
1850

1920

1990

1900

1800
1985

1880

1750

1980
1860

1820

1700

1975

1840
0

0.1
0.2
Concentration (mg/kg)

1970

0

George
2010

0.5
1
Concentration (mg/kg)

1650

0

Gull
2000

0.2
0.4
Concentration (mg/kg)
Hackert

2000

2000
1950

1980

1990
1960

1900
1980

1940
1850

1970

1920

1960

1800

1950

1900
1880

1750

1940

1860

1700
1930

0

0.1
0.2
Concentration (mg/kg)

0

0.05
0.1
Concentration (mg/kg)

1840

0

0.1
0.2
Concentration (mg/kg)

Figure 3.2: Mercury concentration profiles (solid markers), results of profile smoothing (open
markers) and recovery rate (solid line) for study lakes.

54

Figure 3.2 (cont’d)
Houghton

Mullett

Muskegon
2010

2000

2000

1950

1950

1900

1900

1850

1850

1980

1800

1800

1970

1750

1750

1960

1700

0

2
4
Concentration (mg/kg)

1700

Nichols

2000

1990

0

0.05
0.1
Concentration (mg/kg)

1950

Otter

0

0.5
1
Concentration (mg/kg)
Round

2010
2000

2000
2000

1990
1950

1950
1980
1970

1900

1900

1960

1950

1850

0

0.2
0.4
Concentration (mg/kg)

1940
1850
0.1
0.15
0.2
0.25
0
0.1
0.2
Concentration (mg/kg)
Concentration (mg/kg)

55

Figure 3.2 (cont’d)
Torch
2000

1950

1900

1850

1800

1750

0

0.05
0.1
Concentration (mg/kg)

Results and Discussion
Using sediment chronologies to define recovery. The geochemistry of lake sediment is
determined by three primary sources of material 1) what is deposited, wet or dry, from the
atmosphere, 2) autocthonous production e.g., calcium carbonate deposition and 3) erosional
inputs from the catchment. Groundwater input could also be considered a source, but for this
discussion it is ignored since inputs of Hg would be small compared to the other three. This
makes the sediment archive a composite of these sources and a recorder of differences in the
rates at which material is delivered to the system. The recovery portion of the sediment archive
is then a reflection of decreases in deposition and watershed export, autocthonous production is
not considered since Hg would not be supplied by this process but rather its geochemical form
altered. Although it is sometimes difficult to differentiate between atmospheric deposition and
catchment export, it could be assumed that atmospheric deposition has not changed in the last
56

decade, or longer, as has been shown for this region (Prestbo and Gay, 2009) or that atmospheric
deposition across the region would not vary enough to be reflected in sediment archives.
Differences in the recovery slope are then mostly influenced by variation in watershed processes.
A recovery slope closer to zero (plotting Hg on the abscissa and age on the ordinate) would be
indicative of efficient removal of Hg from the landscape leading to the slow decline observed in
the Hg profile. This could be from watershed activities that denude the landscape e.g.,
agriculture or watershed processes that lead to the rapid release of Hg. Conversely, a steep
recovery slope suggests that Hg is being retained within the watershed quite possibly a result of a
lack of activities that disturb the landscape or watershed attributes that tend to retain Hg.

Patterns of watershed attributes. The areal extent of the watersheds in this study ranged from
2

2

0.95 to more than 1300 km with the majority falling below 200 km (Table 1). Ratios of
watershed area to lake area varied between 1.1 and 53 with most less than 10. Watersheds with a
higher percentage of URB are located in the southeastern Lower Peninsula; those high in AG
along the western half of the Lower Peninsula. Highly forested catchments are found in the
northern Lower and Upper Peninsula and those in the southwest Lower Peninsula are the least
forested. Soil erodibility values, K.F and K.FF, were generally low, less than 0.2 for most
watersheds, indicative of clay soils found throughout the state. Sulfate deposition decreased
from the South to the North which follows the general population and industrialization trend for
the State, with the exception that sulfate deposition was higher in the southwest than the
southeast Lower Peninsula(Yohn, 2004). This is likely an indication of the importance of the
large Chicago, IL/Gary, IN industrial complex located near southwest Michigan.

57

Patterns of concentration recovery rates. Concentration recovery rates varied from -0.30 to -1 -1

7.8 µg Hg kg y with the more positive values indicating a slower recovery rate, i.e., max and
reference concentrations being equal a return to reference conditions would take more time. In
lakes Gull, Houghton and Elk recovery rates were calculated from the mid-1960s until present.
In other lakes the recovery rate had not changed for the last 10 y prior to sampling. One
exception was Round D lake which exhibited a sharp decline in concentration around 2000, in
this case the most recent slope, which contained 8 sediment slices, was used (Figure 3.2). This
suggests that the choice of using the 2001 land use data set rather than an earlier data set was
justified. Round D and Nichols lakes have the fastest rate of recovery while Charlevoix and
Mullett lakes are recovering much more slowly (8-26 times slower, respectively). Lakes with
very rapid concentration recovery rates have small watersheds that are primarily forested with
very little urban or agricultural land uses. Differences in land use distribution between the
slowest and fastest recovery rates appear at first glance to be a greater percentage of agricultural
land use in watersheds of lakes with slower rates of recovery (Table 1). Similar rates of recovery
were found for some lakes within close proximity of each other. This may be due to comparable
rates of atmospheric deposition for these systems or could be due to a similar distribution of land
use and rates of disturbances within their watersheds. For example, Elk and Torch lakes, located
in northwest Lower Peninsula of Michigan have similar land use distributions within their,
primarily forested, watersheds (Table 1) and similar concentration recovery rates. The rate of
recovery for Gull Lake, located in the southwestern portion of the Lower Peninsula, is similar to
the northwestern Lower Peninsula lakes. However, it has a very different land use distribution
within it’s, primarily agricultural, watershed and different rates of atmospheric deposition of Hg
would be expected between these two areas of the State (Keeler and Dvonch, 2005). Others have
58

shown agricultural land-use to export Hg more rapidly than forested landscapes (Babiarz et al.,
1998, Hurley et al., 1995a), but this does not help to explain why Gull Lake was similar to the
northwestern Lower Peninsula lakes.

Patterns of flux recovery rates. Rates of flux recovery ranged from -0.09 to -2.7 µg Hg m

-2 -1

y

-1

y where more positive values again represent a slower rate of recovery. Much like the slopes
of concentration recovery, flux recovery slopes had not changed 10 y prior to sampling the lake.
Unlike concentration recovery; however, flux recovery rates were most rapid for Muskegon and
George whereas Birch and Mullett lakes were the slowest. Common among the concentration
and flux recovery rates was that Mullett had the slowest rate of recovery and that George had a
high rate of recovery. Lake Charlevoix which did have a slow concentration recovery was
amongst the more rapid flux recovery lakes. Lakes with more rapid flux recovery rates tended to
have a higher percentage of URB while those with slower rates had higher AG land cover within
a 100 m buffer of the lake. As with concentration recovery, lakes that are in close proximity to
one another tended to have similar flux recovery rates, e.g., Torch and Elk.

Factors influencing concentration recovery rates. Watershed area, WA:LA, SD, LAT, soil
erosion factors, and Hg emissions were not found to be significantly correlated to the rate of
concentration recovery. This might have been expected if the rate of recovery is dependent upon
land use and rate of disturbance of watershed soils rather than atmospheric deposition and
proximity to local emission sources.
The strongest correlation of the concentration recovery rate to land use was found for
agricultural land, but only along watershed flowpaths (LN AG.FLI2: r = -0.75, p ≤ 0.004;
59

Hackert removed as outlier)) (Figure 3.3), the inverse relationship indicates that as AG.FLI2
increases the rate of concentration recovery slows. Other measures of agriculture within the
watershed were also found to be significant but more poorly correlated, (AG.FLI: r = -0.62, p =
0.025; Birch, Gull and Hackert removed as outliers) and (LN AG: r = -0.54, p = 0.087; Birch and
Gull removed as outliers). The association of slowed concentration recovery to AG found here is
consistent with others that have found agricultural land use to be an important determinant for
Hg accumulation (Balogh et al., 1998, Engstrom et al., 2007). However, these results suggest
that where these activities are located within the landscape to be more explanatory than total
areal extent.
Wetlands within the watershed were found to be moderately positively correlated (WET:
r = 0.65, p = 0.02) to concentration recovery. The positive relationship means that lakes with
greater areal extent of wetlands showed faster rates of recovery. Conversely, wetlands within the
100 m buffer of the shoreline were found to be negatively correlated (WET.M: r = -0.61, p =
0.005, Birch and Otter removed as outliers). What this demonstrates is that the proximity of the
wetland to the lake shoreline plays an important role in the cycling of Hg in the watershed.
Wetlands have been reported to be sources of methyl-Hg, but there is disagreement in the
literature if they act as sinks (St. Louis et al., 1996) or sources (Hurley et al., 1995b, Mierle and
Ingram, 1991) for total Hg.

We interpret the contradiction between WET and WET.M to mean

that wetlands on the landscape can act to sequester Hg, but when they are located close to the
lake they can slowly release Hg, thus slowing the rate of recovery.

60

Birch

Cass

Charlevoix

2010

2000

2000
2005

1980

1950
2000

1960

1900
1940

1995
1850

1920

1990

1900

1800
1985

1880

1750

1980
1860

1820

1700

1975

1840
0

10

20

Acc Rate (µg m

30

1970

-2 -1

y )

0

200

400

Acc Rate (µg m

George

1650

-2 -1

y )

50

100

Acc Rate (µg m

Gull

2010

0

-2 -1

y )

Hackert

2000

2000

2000
1950

1980

1990
1960

1900
1980

1940
1850

1970

1920

1960

1800

1950

1900
1880

1750

1940

1860

1700
1930
10

20

Acc Rate (µg m

30
-2 -1

y )

0

10

20

Acc Rate (µg m

30

-2 -1

y )

1840

0

20
Acc Rate (µg m

40
-2 -1

y )

Figure 3.3. Mercury accumulation rate (Acc Rate) profiles (solid markers), results of profile
smoothing (open markers) and recovery rate (solid line) for study lakes.
61

Figure 3.3 (cont’d)
Houghton

Mullett

Muskegon
2010

2000

2000

1950

1950

1900

1900

1850

1850

1980

1800

1800

1970

1750

1750

1960

1700

0

500
Acc Rate (µg m

1000

1700

-2 -1

y )

2000

1990

0

10

20

Acc Rate (µg m

Nichols

30

1950

0

-2 -1

y )

200

400

Acc Rate (µg m

Otter

-2 -1

y )

Round

2010
2000

2000
2000

1990
1950

1950
1980
1970

1900

1900

1960

1950

1850

0

10
Acc Rate (µg m

20
-2 -1

y )

1940
100

150

200

Acc Rate (µg m

62

250

-2 -1

y )

1850

0

10

20

Acc Rate (µg m

30

-2 -1

y )

Figure 3.3 (cont’d)
Torch
2000

1950

1900

1850

1800

1750

0

10
Acc Rate (µg m

20
-2 -1

y )

Forested land cover types along hydrodynamic flowpaths were also found to be
significantly correlated to concentration recovery (FOR.FLI: r = -0.67, p = 0.02, Nichols and
Round D removed as outliers; FOR.FLI2: r = -0.64, p = 0.06, Nichols, Round D, Houghton, and
Muskegon removed as outliers). This result was not expected since forests have been found to
be net sinks of Hg (Aastrup et al., 1991, Engstrom and Swain, 1997, Harris et al., 2007); if Hg is
efficiently being sequestered then a more rapid recovery would be expected for highly forested
landscapes. These results suggest that Hg is being released from forests along hydronamic
flowpaths; however, thus slowing concentration recovery. This may be due to the ease of which
run-off containing Hg laden particles can reach the lake using the high conductivity flowpaths
and the fact that mercury concentrations are generally elevated in the litter and organic soils of
forests due to the accumulation of Hg over decadal or longer time spans (Obrist et al., 2011).
63

Factors influencing flux recovery rates. As with concentration recovery, none of the physical
attributes of the watersheds, e.g., WA, were correlated to flux recovery rates and neither were
soil erodibility, Hg emissions or sulfate deposition.

The relationships found to be significant for flux recovery were quite similar to those of
concentration recovery. Percentage wetlands within the watershed was found to have the
greatest correlation (Ln WET: r = 0.68, p = 0.01). As with concentration recovery, a proximity
measure for wetlands was also found to be significantly correlated to flux recovery, WET.FLI
(Ln WET.FLI: r = -0.69, p = -0.02; Houghton and Otter removed as outliers). The reasons for
the relationships found for wetland variables correlated to flux recovery are similar to those for
concentration recovery. Wetlands within the watershed act as traps for Hg leading to a faster
recovery as WET increases; whereas, wetlands along the highly conductive flowpaths of the
watershed tend to supply Hg to the lake, resulting in slower recovery. Proximity measures of
forest cover types were also found to be significant (FOR.FLI: r = -0.60, p = 0.05; Nichols and
Round D removed as outliers, FOR.FLI2: r = -0.60, p = 0.09; Round D, Nichols, Houghton, and
Muskegon removed as outliers). Similar to the results found for concentration recovery, flux
recovery became slower as the percentage of forest along hydronamic flowpaths increased. The
arguments made for the relationship between concentration recovery and forest land cover are
the same as those that can be made for flux recovery.

Noticeably absent from the analysis of flux recovery was the agricultural land cover
types. This was not expected since agricultural land cover is generally associated with erosion of

64

the top soil layers leading to the export of contaminant laden particles (Balogh et al., 1998,
Engstrom et al., 2007). It would have been anticipated then that increases in agricultural land
cover would have led to a decreased rate of flux recovery. The reasons for the absence are
unclear, but may be related to how drastically sedimentation rates are changed within highly
agricultural watersheds compared to those lakes in highly forested landscapes and the changes in
soil matrix that occurs as a result of erosional processes (Engstrom et al., 2007).

Surprisingly, watershed attributes such as WA:LA, urban land cover, and Hg emissions
were not identified as significant determinants of flux or concentration recovery. The ratio of
watershed to lake area has well been established as a predictor of Hg accumulation rates
(Kamman and Engstrom, 2002, Lorey and Driscoll, 1999, Swain et al., 1992). The model
suggests that as WA:LA increased Hg accumulation rate is predicted to rise due to the greater
surface area for atmospheric deposition and potential for export to the lake. It was expected then
that lakes with a larger WA:LA would have slower rates of recovery since the soil pool available
for export is greater. However, this relationship appears to break down when lakes are not in
close proximity to each other or the group of lakes is diverse (Drevnick et al., 2011, Muir et al.,
2009) as would be the case for lakes in this study (Figure 3.1, Table 3.1).

Urban landscapes have been suggested to export less Hg than other land cover types in
recent decades due to decreases in atmospheric deposition and the lack of watershed disturbance
(Engstrom and Swain, 1997). It would then be anticipated that URB would be correlated to the
rate of recovery with those lakes high in urban land cover having faster recovery rates. Our data,
although representing a broad range of watershed characteristics, had a maximum urban land

65

cover of 25% (Cass Lake) with most lakes less than 10%. Interestingly, Cass Lake had one of
the fastest flux recovery rates which may be a reflection of greater urban land cover. However,
the lack of distribution and representation of true “urban” lakes (URB > 50%) may be
contributing to the lack of significant correlation.

It was hypothesized that recovery rates would be slower in lakes that had high emission
rates, but emissions were not found to be significantly correlated regardless of the radius chosen.
Other variables used to represent atmospheric deposition, e.g., SD and LAT, were also found
unimportant. The lack of relationship could be related to the atmospheric transport of Hg and the
possibility that Hg is being deposited far from the source; although, many studies have shown
that the majority of Hg deposited derives from local to regional scale sources (Donahue et al.,
2006, Keeler et al., 2006, Menounou and Presley, 2003). It has been shown that only a portion
of the Hg deposited to the catchment is immediately transported to lake with most of the
watershed load being made up of “legacy” Hg (Harris et al., 2007). This means that it may have
been more appropriate to use estimates from earlier decades rather than current emissions.
However, if the pattern of Hg emissions has not changed greatly, a likely scenario, then this
analysis would have resulted in similar outcomes. If the majority of Hg is retained in the
watershed, mostly sorbed to soil organic matter (Grigal, 2003, Meili, 1991, Obrist et al., 2011,
Schuster, 1991, Skyllberg et al., 2000), it is then those activities within the catchment that disturb
the soil Hg pool that control export and therefore recovery. That is not to say that atmospheric
emission and deposition of Hg are unimportant but rather the relative impact of these activities
compared to watershed processes is small when comparing lakes within a region.

66

Although this analysis singles out certain watershed attributes that influence recovery
from Hg enrichment it is likely that it is a combination of factors that in coordination determine
the shape of Hg profiles and the trajectory of recovery. That being said, this work has important
implications for natural resource and watershed managers in that the locations of activities and
processes within the watershed were found to be significant which may help investigators
identify areas of interest in future catchment studies of Hg cycling.

67

REFERENCES

68

REFERENCES
Aastrup, M., Johnson, J., Bringmark, E., Bringmark, L., Iverfeldt, A., 1991. Occurrence and
Transport of Mercury within a Small Catchment-Area. Water Air and Soil Pollution 56,
155-167.
Alfaro-De la Torre, M.C., Tessier, A., 2002. Cadmium deposition and mobility in the sediments
of an acidic oligotrophic lake. Geochimica et Cosmochimica Acta 66, 3549-3562.
Appleby, P.G., 2001. Chronostratigraphic Techniques in Recent Sediments. In: W. M. Last, J. P.
Smol (Eds.), Tracking Environmental Change Using Lake Sediments. Kluwer Academic
Publishers, Boston, pp. 171-205.
Appleby, P.G., Oldfield, F., 1978. The calculation of 210Pb dates assuming a constant rate of
supply of unsupported 210Pb to the sediment. Catena 5, 1-8.
Babiarz, C.L., Hurley, J.P., Benoit, J.M., Shafer, M.M., Andren, A.W., Webb, D.A., 1998.
Seasonal influences on partitioning and transport of total and methylmercury in rivers
from contrasting watersheds. Biogeochemistry 41, 237-257.
Balogh, S.J., Meyer, M.L., Johnson, D.K., 1998. Transport of mercury in three contrasting river
basins. Environmental Science & Technology 32, 456-462.
Bindler, R., Renberg, I., Appleby, P.G., Anderson, N.J., Rose, N.L., 2001. Mercury
accumulation rates and spatial patterns in lake sediments from west Greenland: A coast to
ice margin transect. Environmental Science and Technology 35, 1736-1741.
Donahue, W.F., Allen, E.W., Schindler, D.W., 2006. Impacts of coal-fired power plants on trace
metals and polycyclic aromatic hydrocarbons (PAHs) in lake sediments in central
Alberta, Canada. Journal of Paleolimnology 35, 111-128.
Drevnick, P.E., Engstrom, D.R., Driscoll, C.T., Swain, E.B., Balogh, S.J., Kamman, N.C., Long,
D.T., Muir, D.C.G., Parsons, M.J., Rolfhus, K.R., Rossmann, R., 2011. Spatial and
Temporal Patterns of Mercury Accumulation in Lacustrine Sediments across the
Laurentian Great Lakes Region. Environmental Pollution In Press.
Engstrom, D.R., Balogh, S.J., Swain, E.B., 2007. History of mercury inputs to Minnesota lakes:
Influences of watershed disturbance and localized atmospheric deposition. Limnology
and Oceanography 52, 2467-2483.
Engstrom, D.R., Swain, E.B., 1997. Recent declines in atmospheric mercury deposition in the
upper Midwest. Environmental Science and Technology 31, 960-967.
ARC/INFO, Environmental Systems Research Institute, 1998. ARC/INFO
Evers, D.C., Wiener, J.G., Driscoll, C.T., Gay, D.A., Basu, N., Monson, B.A., Lambert, K.F.,
Morrison, H.A., Morgan, J.T., Williams, K.A., Soehl, A.G., 2011. Great Lakes Mercury
69

Connections: The Extent and Effects of Mercury Pollution in the Great Lakes Region.
Biodiversity Research Institute. Gorham, MA. Report BRI 2011-18, 44 pages. Available
from: http://www.briloon.org/mercuryconnections/GreatLakes.
Golden, K.A., Wong, C.S., Jeremiason, J.D., Eisenreich, S.J., Sanders, M.G., Hallgren, J.,
Swackhammer, D.L., Engstrom, D.R., Long, D.T., 1993. Accumulation and preliminary
inventory of organochlorines in Great Lakes sediments. Water Science and Technology
28, 19-31.
Grigal, D.F., 2003. Mercury sequestration in forests and peatlands: A review. Journal of
Environmental Quality 32, 393-405.
Harris, R.C., Rudd, J.W.M., Amyot, M., Babiarz, C.L., Beaty, K.G., Blanchfield, P.J., Bodaly,
R.A., Branfireun, B.A., Gilmour, C.C., Graydon, J.A., Heyes, A., Hintelmann, H.,
Hurley, J.P., Kelly, C.A., Krabbenhoft, D.P., Lindberg, S.E., Mason, R.P., Paterson, M.J.,
Podemski, C.L., Robinson, A., Sandilands, K.A., Southworth, G.R., Louis, V.L.S., Tate,
M.T., 2007. Whole-ecosystem study shows rapid fish-mercury response to changes in
mercury deposition. Proceedings of the National Academy of Sciences of the United
States of America 104, 16586-16591.
Heyvaert, A.C., Reuter, J.E., Slotton, D.G., Goldman, C.R., 2000. Paleolimnological
reconstruction of historical atmospheric lead and mercury deposition at Lake Tahoe,
California-Nevada. Environmental Science and Technology 34, 3588-3597.
Hightower, J., Moore, D., 2003. Mercury Levels in High-End Consumers of Fish. Environmental
Health Perspectives 111, 604-608.
Hurley, J.P., Benoit, J.M., Babiarz, C.L., Shafer, M.M., Andren, A.W., Sullivan, J.R., Hammond,
R., Webb, D.A., 1995a. Influences of Watershed Characteristics on Mercury Levels in
Wisconsin Rivers. Environmental Science & Technology 29, 1867-1875.
Hurley, P.J., Benoit, J.M., Babiarz, C.L., Shafer, M.M., Andren, A.W., Sullivan, J.R., Hammond,
R., Webb, D.A., 1995b. Influences of watershed characteristics on mercury levels in
Wisconsin rivers. Environmental Science and Technology 29, 1867-1875.
National Atmospheric Deposition Program (NRSP-3), Illinois State Water Survey, NADP
Program Office, 2204 Griffith Drive, Champaign, IL 61820, Available online at
http://nadp.sws.uiuc.edu/mdn/. (Accessed: February 5, 1998).
Kamman, N.C., Engstrom, D.R., 2002. Historical and present fluxes of mercury to Vermont and
New Hampshire lakes inferred from Pb-210 dated sediment cores. Atmospheric
Environment 36, 1599-1609.
Keeler, G.J., Dvonch, J.T., 2005. Atmospheric mercury: A decade of observations in the Great
Lakes. In: N. Pirrone, K. Mahaffey (Eds.), Dynamics of Mercury Pollution on Regional

70

and Global Scales: Atmospheric Processes and Human Exposures around the World.
Kluwer Ltd., Norwell, MA.
Keeler, G.J., Landis, M.S., Norris, G.A., Christianson, E.M., Dvonch, J.T., 2006. Sources of
mercury wet deposition in Eastern Ohio, USA. Environmental Science & Technology 40,
5874-5881.
Long, D.T., Parsons, M.J., Yansa, C.H., Yohn, S.S., McLean, C.E., Vannier, R.G., 2010.
Assessing the response of watersheds to catastrophic (logging) and possible secular
(global temperature change) perturbations using sediment-chemical chronologies.
Applied Geochemistry 25, 143-158.
Lorey, P., Driscoll, C.T., 1999. Historical trends in mercury deposition in Adirondack Lakes.
Environmental Science and Technology 33, 718-722.
Matlab, The Mathworks, 1994 - 2008. Matlab Student 7.0.1.
Meili, M., 1991. The Coupling of Mercury and Organic-Matter in the Biogeochemical Cycle Towards a Mechanistic Model for the Boreal Forest Zone. Water, Air, and Soil Pollution
56, 333-347.
Menounou, N., Presley, B.J., 2003. Mercury and other trace elements in sediment cores from
central texas lakes. Archives of Environmental Contamination and Toxicology 45, 11-29.
Michigan Department of Natural Resources, 2003. IFMAP/GAP land cover. In: MDNR (Ed.),
Lansing, MI,
Mierle, G., Ingram, R., 1991. The Role of Humic Substances in the Mobilization of Mercury
from Watersheds. Water Air and Soil Pollution 56, 349-357.
Muir, D.C.G., Wang, X., Yang, F., Nguyen, N., Jackson, T.A., Evans, M.S., Douglas, M., Kock,
G., Lamoureux, S., Pienitz, R., Smol, J.P., Vincent, W.F., Dastoor, A., 2009. Spatial
Trends and Historical Deposition of Mercury in Eastern and Northern Canada Inferred
from Lake Sediment Cores. Environmental Science & Technology 43, 4802-4809.
Munthe, J., Bodaly, R.A., Branfireun, B.A., Driscoll, C.T., Gilmour, C.C., Harris, R., Horvat,
M., Lucotte, M., Malm, O., 2007a. Recovery of mercury-contaminated fisheries. Ambio
36, 33-44.
Munthe, J., Hellsten, S., Zetterberg, T., 2007b. Mobilization of mercury and methylmercury from
forest soils after a severe storm-fell event. Ambio 36, 111-113.
Munthe, J., Hultberg, H., 2004. Mercury and methylmercury in runoff from a forested
catchment: concentrations, fluxes, and their response to manipultions. Water Air and Soil
Pollution: Focus 4, 607-618.

71

Obrist, D., Johnson, D.W., Lindberg, S.E., Luo, Y., Hararuk, O., Bracho, R., Battles, J.J., Dail,
D.B., Edmonds, R.L., Monson, R.K., Ollinger, S.V., Pallardy, S.G., Pregitzer, K.S.,
Todd, D.E., 2011. Mercury Distribution Across 14 US Forests. Part I: Spatial Patterns of
Concentrations in Biomass, Litter, and Soils. Environmental Science & Technology 45,
3974-3981.
Parsons, M.J., Long, D.T., Yohn, S.S., Giesy, J.P., 2007. Spatial and Temporal Trends of
Mercury Loadings to Michigan Inland Lakes. Environmental Science and Technology 41,
5634-5640.
Porvari, P., Verta, M., Munthe, J., Haapanen, M., 2003. Forestry practices increase mercury and
methyl mercury output from boreal forest catchments. Environmental Science &
Technology 37, 2389-2393.
Prestbo, E.M., Gay, D.A., 2009. Wet deposition of mercury in the US and Canada, 1996-2005:
Results and analysis of the NADP mercury deposition network (MDN). Atmospheric
Environment 43, 4223-4233.
Schuster, E., 1991. The Behavior of Mercury in the Soil with Special Emphasis on Complexation
and Adsorption Processes - A Review of the Literature. Water Air and Soil Pollution 56,
667-680.
Skyllberg, U., Xia, K., Bloom, P.R., Nater, E.A., Bleam, W.F., 2000. Binding of mercury(II) to
reduced sulfur in soil organic matter along upland-peat soil transects. Journal of
Environmental Quality 29, 855-865.
St. Louis, V.L., Rudd, J.W.M., Kelly, C.A., Beaty, K.G., Flett, R.J., Roulet, N.T., 1996.
Production and loss of methylmercury and loss of total mercury from boreal forest
catchments containing different types of wetlands. Environmental Science & Technology
30, 2719-2729.
Swain, E.B., Engstrom, D.R., Brigham, M.E., Henning, T.A., Brezonik, P.L., 1992. Increasing
Rates of Atmospheric Mercury Deposition in Midcontinental North-America. Science
257, 784-787.
R version 2.6.2, 2008. R version 2.6.2
Turetsky, M.R., Harden, J.W., Friedli, H.R., Flannigan, M., Payne, N., Crock, J., Radke, L.,
2006. Wildfires threaten mercury stocks in northern soils. Geophysical Research Letters
33.
US EPA, 2008. Air Pollution Prevention and Control. US EPA.Government Printing Office,
Washington, D.C.
State Soil Geographic (STATSGO) Database for Michigan, Available online at
http://www.ncgc.nrcs.usda.gov/products/datasets/statsgo/index.html. (Accessed:
September 25, 2011).

72

USEPA. Office of Solid Wastes. Mercury in solids and solutions by thermal decomposition,
amalgamation, and atomic absorption spectrophotometry; Government Printing Office,
1998.
USEPA. Office of Research and Development Office of Air Quality Planning and Standards.
Mercury Study Report to Congress; Washington, DC: Government Printing Office, 1997.
US EPA, 2010. National Emissions Inventory, Available online at
http://www.epa.gov/ttn/chief/index.html. (Accessed: June 1, 2010).
United States Geological Survey, The National Map Seamless Server, Available online at
http://seamless.usgs.gov/index.php. (Accessed: September 25, 2011).
Yohn, S.S. 2004. Natural and Anthropogenic Factors Influencing Spatial and Temporal Patterns
of Metal Accumulation in Inland Lakes. Thesis, Michigan State University, East Lansing,
MI.

73

CHAPTER 4
INFERRING SOURCES FOR MERCURY TO INLAND LAKES USING SEDIMENT
CHRONOLOGIES OF POLYCYCLIC AROMATIC HYDROCARBONS

Abstract

Sediment chronologies from inland lakes suggest the influence of local to sub-regional scale
sources for Hg. However, source apportionment using sediment chronologies is difficult due to
the mixing of sources and pathways. Mercury and polycyclic aromatic hydrocarbons (PAH)
concentration profiles often share common sources and pathway into the environment.
Therefore, sediment cores from seven inland lakes of Michigan were collected for the
measurement of PAHs and Hg and dated using

210

Pb. To infer sources of PAHs and Hg,

assuming common source, diagnostic ratios of kinetic and thermodynamic PAH isomers were
used. Ratios indicate the existence of modern combustion sources to each lake and historic
combustion sources to lakes near cement kilns and an iron foundry. Coal combustion sources
were identified for two lakes near urban centers. Whereas a petroleum combustion source was
identified for a lake that has a coal fired power plant along its shoreline. These results therefore
have implications for the cycling of Hg on local to regional scales.
Introduction
Several studies have shown that much of the anthropogenic mercury (Hg) emitted to the
atmosphere is a result of the combustion of coal and petroleum products (Bindler, 2003,
Fitzgerald, 1989, Kamman and Engstrom, 2002, Keeler et al., 2006, Landis et al., 2002,
Lindberg et al., 2007, Long et al., 1995, Menounou and Presley, 2003, Pacyna et al., 2006,
Watras et al., 1994). As a result of its physical properties Hg can reside in the atmosphere for up
to a year resulting in global transport (Fitzgerald et al., 1998). For example, several studies have
74

shown that Hg emitted in Asia has reached the North American continent (Boutron et al., 1998,
Durnford et al., 2010, Jaffe et al., 2005, Jaffe and Strode, 2008, Obrist et al., 2008, Reidmiller et
al., 2009, Valente et al., 2007). Still others have demonstrated that much of the Hg emitted to
the atmosphere can be attributed to local and regional sources (Donahue et al., 2006, Keeler et
al., 2006). As current environmental legislation (US EPA, 2011) attempts to reduce Hg release
determining the source, especially the distinction between local, regional and global, becomes of
impending importance. Environmental archives of Hg (e.g., sediment and ice cores) have been
used by others to elucidate the likely source of Hg by comparing the record to national
production and consumption records (Engstrom and Swain, 1997) and event-based markers (e.g.,
major volcanic eruptions) (Schuster et al., 2002). However, it is difficult to recognize the
potential of local sources using these methods since only national trends are used or, in the case
of ice cores, no local sources exist. The study of source apportionment has been approached
using atmospheric deposition collectors and multivariate statistical techniques providing
evidence of local to global scale sources (Dvonch et al., 1999, Dvonch et al., 1998, Keeler et al.,
2006). Their use is limited; however, due to the lack of atmospheric collectors and the uneven
distribution of collectors throughout a region, nation or the world (Evers et al., 2011).

Inland lakes are widely distributed and when many lakes can be sampled across a broad
region have the potential to discriminate between local and regional sources. However, since
sediments incorporate both direct atmospheric sources and watershed processes the interpretation
of source can become difficult. This may be overcome if an independent marker could be used
to infer source. Polycyclic aromatic hydrocarbons (PAHs) are primarily formed during the
incomplete combustion of petroleum products and organic material (e.g., vegetation) (Hites et

75

al., 1977). Due to their volatility PAHs can remain in the atmosphere after emission which can
result in long distance transport (Killin et al., 2004, Sofowote et al., 2011, Usenko et al., 2010),
but can also be deposited nearer the source as a result of their particle reactivity (Barrick and
Prahl, 1987). Because they are particle reactive sediment chronologies can be used to infer the
record of their release to the environment (Hites et al., 1977). There are a variety of activities
that result in a release of PAHs into the environment (e.g., coal combustion and vehicular
exhaust) and the sedimentary record is a mixture of all sources which can make source
apportionment difficult. Sources of PAHs can be natural (e.g., oil seeps, bitumens, and forest
fires) or anthropogenic (fossil fuels and combustion) (Yunker and Macdonald, 2003, Yunker et
al., 2002). Unique diagnostic ratios of PAHs have been shown to be useful in the identification
of types of sources (Howe et al., 2004, Simcik et al., 1999, Sofowote et al., 2010, Yunker and
Macdonald, 2003, Zheng et al., 2011).

There have been many studies that measured metals and PAHs in the environment (Di
Leonardo et al., 2009, Donahue et al., 2006, Gnandi et al., 2011, Mzoughi and Chouba, 2011).
However, few of these studies investigated the correlations in environmental loading between
them even though there is strong evidence that they might share common sources and pathways.
For example, the sum of PAHs measured in aerosols from the North Sea was found to be
significantly correlated to lead and other heavy metals (Chester et al., 1993). In a study designed
to investigate the impact of vehicular emissions on roadways a multivariate statistical analysis
showed that Hg tended to cluster with the PAHs rather than the other trace metals, such as lead
and zinc (Zechmeister et al., 2006) which could be an indication of the shared combustion source
and pathway for Hg and PAHs in vehicular exhaust. In New Orleans, LA, USA it was found that

76

PAH ratios indicated a pyrolytic source and that the PAHs were moderately correlated to trace
metals (Wang et al., 2004).

A significant problem with the use of diagnostic ratios of PAHs is the differential
degradation of individual PAHs which are driven by differences in vapor pressure and resistance
to photolytic decay (Yunker et al., 1999, Zhang et al., 2005). Differences in the rate of
degradation are greatest amongst the lower molecular weight PAHs, e.g., anthracene, making the
use of diagnostic ratios that compare lower molecular weight PAHs troublesome. Corrections
for differential degradation can be made when meteorological and source data are available
(Zhang et al., 2005); however, in most cases, and is the case in this study, these data are
unavailable. To overcome this issue most studies of PAHs in the environment use multiple
diagnostic ratios of PAHs (Yunker and Macdonald, 2003).

This work is a portion of the larger Michigan Department of Natural Resources and
Environment Sediment Trends Monitoring Program (Michigan, 2011b). Lakes for this study
were chosen to investigate the hypothesis that Hg is derived mainly from local to regional scales
sources. If true then diagnostic ratios should identify local to regional scale sources or show a
lack of impact from global scale combustion sources.
Materials and Methods
A detailed description of core collection (Parsons et al., 2007),

210

Pb dating (Parsons et al.,

2007), Hg (Parsons et al., 2007) and PAH analysis (Kannan et al., 2005) can be found elsewhere.
Briefly, cores were collected aboard the MDNRE R/V Nibi or US EPA R/V Mudpuppy using an
Ocean Instruments MC-400 and sectioned on-shore immediately after collection. Separate cores
77

were used for

210

Pb dating, metals and organic analyses. Sectioning of the Hg and

210

Pb core

was performed at equal intervals, 0.5 cm increments for the first 8 cm and 1 cm intervals
thereafter. The entire organics core was sectioned at 1 cm intervals. Due to the limited quantity
of sediment collected at 1-cm intervals, equal amounts of sediment were combined from
successive sediment samples starting from the top of the core.

Table 4.1. PAH parameters and abbreviations.
Mass Compound
Abbreviation
152
Acenaphthylene
Ayl
154
Acenaphthene
Aen
166
Fluorene
F
178
Phenanthrene
Ph
Anthracence
An
202
Pyrene
Py
Fluoranthene
Fl
228
Chrysene
Ch
Benz[a]anthracene
BaA
252
Benzo[b]fluoranthene
BF
Benzo[j]fluoranthene
BF
Benzo[k]fluoranthene
BF
276
Benzo[ghi]perylene
BghiP
Indeno[1,2,3-cd]pyrene
IP
278
Dibenz[a,h]anthracene
DhA

Mercury analysis was performed using an Ohio Lumex Hg Analyzer according to EPA
Method 7473. Measurements of
Winnipeg’s Freshwater Institute;

210

Pb were made by Paul Wilkinson at the University of

137

Cs was measured to verify age estimates. The US EPA’s 16

priority PAHs were analyzed using gas chromatography-mass spectrometry. Sediments for
organics analysis were Soxhlet extracted then treated with activated copper. Extracts were
passed through 10 g of activated Florisil (60-100 mesh; Sigma, St. Louis, MO) and the fraction
eluted with 100 mL of 20% dichloromethane in hexane and then concentrated. Results of PAHs
78

are reported here as ng/g dry weight. Total PAHs (ΣPAH) is defined here as the sum of all 16
US EPA priority PAHs, Table 1.

Since the organics core was not sectioned at the same intervals as the

210

Pb core dates for

the organic core were estimated by calculating the local polynomial regression of depth versus
age using the loess command in the statistical package R (The R Foundation for Statistical
Computing, 2008). Mercury concentrations were then estimated by interpolation using the
estimated dates from the regression analysis.

Results and Discussion
Patterns of PAH and Hg. Lakes chosen for this study (Figure 4.1) had at least a moderately
high correlation (Pearson’s r > 0.7, Table 1) between ΣPAH and Hg possibly indicating a
common source for these contaminants. The pattern of peak and surficial ΣPAH loosely follows
the South to North industrialization and population trend, a pattern which was also found for
peak lead (Pb) concentrations for this region (Yohn et al., 2004). The higher surficial and peak
ΣPAH concentrations in Muskegon and Otter suggest a more local than regional source which
was found for Pb. Peak and surficial ΣPAHs were greatest in Otter and lowest in Avalon Lake.
The year of peak ΣPAH concentration occurred earliest in Avalon, 1915, and Round lakes, 1929,
whereas peak concentration was observed in the surficial sediments of Crystal Lake. The timing
of peak ΣPAHs in Avalon and Round are not consistent with the peak in atmospheric deposition
described by others (Barrick and Prahl, 1987, Fernandez et al., 2000, Lima et al., 2003,
Wakeham et al., 1980, Yunker and Macdonald, 2003); peaks in ΣPAH near the turn of the last
century have been attributed to the burning of wood, e.g., forest fires and home heating
79

(Fernandez et al., 2000). Crystal Lake showed a peak in the middle of the last century (Figure
4.2c) consistent with an anthropogenic combustion source, but had higher concentrations in the
surficial sediment which may be evidence of newer sources. The year of peak ΣPAH
concentrations corresponded to the year of peak Hg concentration in Round Lake. Muskegon,
Otter and Sand lakes had peak ΣPAH years similar to that of Hg; whereas, Avalon, Birch and
Crystal lakes peak ΣPAH and Hg concentrations did not occur in the same year. Surficial Hg
concentrations were lowest in Avalon and greatest in Round and there is a positive correlation
between surficial Hg and ΣPAHs concentrations. Lakes with greatest PAH and Hg
concentrations tend to be located near urban centers of the State; Muskegon near the City of
Muskegon, Otter near the City of Flint, and Sand near the City of Detroit. The exception being
Round, located in the Michigan’s Upper Peninsula, which could be described as a remote lake
lacking significant urban development other than sparse shoreline housing.
Table 4.2. ΣPAHs (ng/g dry wt) and Hg (mg/kg dry wt) concentration in the surficial sediments
and at maximum. The first column shows the Pearson’s correlation coefficient between Hg and
ΣPAH.
Lake
r
Hg
Hg Peak Year
Hg
ΣPAHs
ΣPAHs
ΣPAH
(surficial) (peak)
(surficial)
(peak)
Peak Year
Avalon
Birch
Crystal
Muskegon
Otter
Round
Sand

0.75
0.86
0.90
0.70
0.76
0.92
0.92

949
1224
1707
2844
3323
1888
1849

1612
2093
1707
4663
8046
1977
2932

1915
1957
2006
1978
1958
1929
1982

0.050
0.068
0.082
0.16
0.12
0.17
0.095

0.052
0.095
0.084
0.37
0.25
0.18
0.12

1971
1977
1990
1973
1960
1929
1988

Prior to the rise of ΣPAHs in the late 1800s and early 1900s mass 252 PAH dominate the core
profile in Avalon, Birch, Crystal and Round Lakes (Figure 4.2a-c & 4.2f). This is in contrast to
what has been found for the Fraser River Basin (Yunker et al., 2002) where the mass 178 and
202 PAHs were dominant. This may suggest a petrogenic origin for the mass 252 PAHs to these
80

lakes (Yunker et al., 1993). In Sand Lake (Figure 4.2g) mass 178 and 202 were greater than the
other PAHs which is consistent with the burning of grass, hardwood and softwoods (Fine et al.,
2001, Jenkins et al., 1996, Masclet et al., 1995). Masses 228 and 276 were for the most part the
least prevalent PAH prior to 1900 in all lakes which, a reflection of the lack of fossil fuel
combustion in the region at the time (Yunker and Macdonald, 2003).

81

¯
Round

(
!

!
(
Avalon

Muskegon

!
(

!
(

Lake

Crystal

!
(

(
!

“For interpretation of the references
to color in this and all other figures,
the reader is referred to the
electronic version of this
dissertation.”
15

Sand
Birch

!
(

Kilometers

Figure 4.1. Study lakes.

82

!
(

Otter

Avalon
Hg (mg/kg)
0

0.005

0.01 0.015

0.02 0.025

0.03 0.035

0.04 0.045

0.05

2000
1980
1960
1940
1920
1900
1880

178
202
228
252
276
Hg

1860
1840
1820
1800

0

50

100

150

200

250

300

350

400

450

PAH Concentration (ng/g dw)

Figure 4.2a. Concentration profiles of Hg and PAH isomer groups for Avalon Lake.

83

Birch
Hg (mg/kg)
0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

2000
1980
1960
1940
1920
1900
1880

178
202
228
252
276
Hg

1860
1840
1820
1800

0

100

200

300

400

500

600

PAH Concentration (ng/g dw)

Figure 4.2b. Concentration profiles of Hg and PAH isomer groups for Birch Lake.

84

Crystal
Hg (mg/kg)
0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

2000
1980
1960
1940
1920
1900
1880

178
202
228
252
276
Hg

1860
1840
1820
1800

0

50

100

150

200

250

300

350

400

450

500

PAH Concentration (ng/g dw)

Figure 4.2c. Concentration profiles of Hg and PAH isomer groups for Crystal Lake.

85

Muskegon
Hg (mg/kg)
2010

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

2000

1990

1980
178
202
228
252
276
Hg

1970

1960

1950

0

200

400

600

800

1000

1200

PAH Concentration (ng/g dw)

Figure 4.2d. Concentration profiles of Hg and PAH isomer groups for Muskegon Lake.

86

Otter
Hg (mg/kg)
2010

0

0.05

0.1

0.15

0.2
178
202
228
252
276
Hg

2000

1990

1980

1970

1960

1950

0

200

400

600

800

1000

1200

1400

1600

1800

2000

PAH Concentration (ng/g dw)

Figure 4.2e. Concentration profiles of Hg and PAH isomer groups for Otter Lake.

87

Round
Hg (mg/kg)
0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

2000
1980
1960
1940
1920
1900
1880

178
202
228
252
276
Hg

1860
1840
1820
1800

0

50

100

150

200

250

300

350

400

450

500

PAH Concentration (ng/g dw)

Figure 4.2f. Concentration profiles of Hg and PAH isomer groups for Round Lake.

88

Sand
Hg (mg/kg)
0

0.02

0.04

0.06

0.08

0.1

0.12

2000
1980
1960
1940
1920
1900
1880

178
202
228
252
276
Hg

1860
1840
1820
1800

0

200

400

600

800

1000

1200

1400

PAH Concentration (ng/g dw)

Figure 4.2g. Concentration profiles of Hg and PAH isomer groups for Sand Lake.

89

After the turn of the last century until recent times the masses 202, 252 and 276 PAHs
were dominant in most lakes. In Avalon, Round and Sand lakes all three of these mass groups
were elevated to near equal concentrations after their respective peaks (Avalon and Round) or
after 1940 (Sand). Mass 202 was greater after the ΣPAH peak in Avalon and after 1920 in Sand
Lake; whereas, mass 276 was greatest among PAH mass groups after ca. 1920 in Round Lake.
Birch was similar to Avalon and Sand lakes, but mass 202 was clearly the major constituent after
the peak in ΣPAH around ca. 1960. During the mid-century ΣPAH peak, 1950-1970, the masses
202 and 276 were dominant in Crystal Lake. The mass 202 PAHs were the major constituent in
Muskegon and Otter throughout the entire core (Figures 4.2d & 4.2e, respectively). Profiles of
PAHs from the Strait of Georgia also showed mass 202 to be the major constituent (Yunker and
Macdonald, 2003) and is associated with sources from urbanized or industrialized areas
(Lipiatou and Saliot, 1991, Yunker et al., 1993). These sources include the burning of gasoline,
diesel and coal which have also been shown to be important sources for masses 202, 252, and to
a lesser extent 228 and 276 (Oros and Simoneit, 2000, Rogge et al., 1993, Wang et al., 1999).

The significant correlation between the ΣPAH and Hg concentration profiles in these
lakes implies that throughout the lakes history a common influence on their concentrations.
However, there are several instances of coincident increases and decreases that can be either
unique or shared among the PAH mass groups. In Avalon Lake the slight concentration increase
across all mass groups during the mid-1990s may be related to the coincident increase in Hg
concentration during the same period. Curiously Hg does not peak until more recent times, ca.
1970, in Avalon while those of the PAH mass groups peaked much earlier, ca. 1920, and have
since declined. The Huron Portland Cement Company started near the town of Alpena (located

90

th

about 20 miles to the east of Avalon Lake) around the turn of the 19 century and by 1910 was
the largest cement plant in the world (NOAA, 2011). Cement kilns have long been recognized as
contributors of Hg to the environment (Munthe et al., 2007), that combined with the increased
demand for cement due to the construction of the first concrete roads in Michigan (Morrison,
1945) could be a likely reason for the earlier than expected peak of PAHs in Avalon Lake. In
Birch Lake the peak of Hg in ca. 1980 followed by decline until present is mimicked by most
mass groups. The trajectory of increase in Hg after 1920 in Crystal Lake closely resembles that
of mass groups 202 and 276 than the other groups. This would be consistent with an
urbanization of the watershed and the combustion of grasses for the clearing of agricultural land
(Fine et al., 2001, Jenkins et al., 1996, Masclet et al., 1995). Whereas the increase in Hg
between 1830 and 1890 more closely resembles the increases in masses 252 and 202. The
profiles of Hg from Muskegon and Otter lakes resembles all of the PAH masses suggesting a
common influence for all mass groups and Hg. In Round Lake the profiles of Hg and the PAH
masses were quite similar from 1860 until the 1920s when Hg becomes nearly constant.

Diagnostic PAH Ratios. The ratio of PAH isomers are used to identify sources. However, their
use as a diagnostic tool to differentiate sources can be confounded by differences in volatility,
water solubility, and adsorption (Zhang et al., 2005). These attributes can cause differential
environmental transport and thus alter ratios. To deal with this problem researchers generally
limit studies to PAHs with multiple isomers with interpretive support using multiple diagnostic
ratios (Wang et al., 2004, Yunker and Macdonald, 2003, Yunker et al., 2002). The isomer ratios
compare the least stable, or kinetic, isomer to the more stable, thermodynamic, isomer. When
combustion sources are predominant the least stable isomer is enhanced (Yunker and
91

Macdonald, 1995) thus, the PAH mass groups with the greatest range in stabilities between the
isomers have more potential to differentiate between petroleum and combustion sources (Yunker
et al., 2002). The PAH molecular masses with the greatest range of stability, in order of most
potential for discrimination to least, are 276, 202, 252 and 178 (Yunker et al., 2002). Examples
of specific diagnostic ratios include the following. A ratio of anthracene to phenanthrene plus
anthracene (An/178) less than 0.1 indicates a dominance of petroleum, e.g., naturally occurring
oil seeps or unburned gasoline, while ratios greater than 0.10 indicates combustion processes
(Yunker et al., 2002). Fluoranthene to fluoranthene plus pyrene (Fl/Fl + Py) less than 0.40
indicates petroleum, a value between 0.40 and 0.50 reflects liquid fossil fuels and ratios greater
than 0.50 are characteristic of grass, wood or coal combustion. Ratios of Benzo[a]anthracene to
benzo[a]anthracene plus chrysene (BaA/228) tend to be less than 0.20 for petroleum sources,
between 0.20 and 0.35 can indicate petroleum or combustion sources, and greater than 0.35 for
combustion sources. Values of the indeno[1,2,3-cd]pyrene to indeno[1,2,3-cd]pyrene plus
benzo[ghi]perylene (IP/IP + Bghi) ratio less than 0.20 are likely representative of petroleum
whereas those values greater than 0.5 imply grass, wood and coal combustion, while those values
in between 0.20 and 0.50 are characteristic of liquid fossil fuel combustion.

Prior to ca. 1930 Avalon, Birch and Sand lakes showed a predominance of petroleum
sources (Figure 4.3a-b & 4.3g); whereas, Crystal and Round (Figure 4.3c & 4.3f, respectively)
lakes showed influence from combustion sources as indicated by the An/178 ratio. Muskegon
and Otter (Figure 4.3d & 4.3e, respectively) lakes show evidence of only combustion sources,
likely due to the short time period covered by the sediment cores and industrial sources within
proximity of these lakes. The Fl/Fl+Py ratio reveals grass/wood/coal combustion for nearly all

92

sediment slices in all lakes except for Muskegon where fossil fuel combustion was found to be
important prior to the early 1980s while afterwards a predominant source could not be identified.
In Birch, Crystal, Round and Sand lakes the ratio of BaA/228 is lower at the base of the sediment
cores, indicative of petroleum sources, then increases towards petroleum combustion ratios
between 1880 and 1900 except for Sand Lake where the transition was later, ca. 1930.
Conversely in Avalon Lake, the BaA/228 ratio suggests combustion sources throughout much of
the core. In Otter Lake a combustion source was identified for the entire core; whereas a mixture
of petroleum and combustion sources were found for Muskegon Lake. The sources indicated by
the ratio of IP/IP+Bghi were mixed between grass/wood/coal combustion and fossil fuel
combustion. In Crystal, Round and Sand lakes the older portions of the core are characterized by
grass/wood/coal combustion sources. Round and Sand showed grass/wood/coal combustion
important recently in contrast to Crystal which like Muskegon Lake was dominated by fossil fuel
combustion. Avalon and Birch lakes IP/IP+Bghi ratios varied between fossil fuel and
grass/wood/coal combustion source. Conversely, Otter Lake seems dominated by
grass/wood/coal combustion.

93

Avalon
2000
1980
1960
1940
1920
1900
1880
1860
1840
1820
1800

0

0.5 0
An/178

0.5
Fl/Fl+Py

10

0.5
BaA/228

10

0.5
IP/IP+Bghi

Figure 4.3a. Selected ratios of PAH isomers in sediment of Avalon Lake. Vertical lines
represent specific PAH diagnostic ratios used to distinguish source, see text for specific
examples.

94

1

Birch
2000
1980
1960
1940
1920
1900
1880
1860
1840
1820
1800

0

0.5 0
An/178

0.5
Fl/Fl+Py

10

0.5
BaA/228

10

0.5
IP/IP+Bghi

1

Figure 4.3b. Selected ratios of PAH isomers in sediment of Birch Lake. Vertical lines represent
specific PAH diagnostic ratios used to distinguish source, see text for specific examples.

95

Crystal
2000
1980
1960
1940
1920
1900
1880
1860
1840
1820
1800

0

0.5 0
An/178

0.5
Fl/Fl+Py

10

0.5
BaA/228

10

0.5
IP/IP+Bghi

1

Figure 4.3c. Selected ratios of PAH isomers in sediment of Crystal Lake. Vertical lines represent
specific PAH diagnostic ratios used to distinguish source, see text for specific examples.

96

Muskegon
2010

2000

1990

1980

1970

1960

1950

0

0.5 0
An/178

0.5
Fl/Fl+Py

10

0.5
BaA/228

10

0.5
IP/IP+Bghi

1

Figure 4.3d. Selected ratios of PAH isomers in sediment of Muskegon Lake. Vertical lines
represent specific PAH diagnostic ratios used to distinguish source, see text for specific
examples.

97

Otter
2010

2000

1990

1980

1970

1960

1950

0

0.5 0
An/178

0.5
Fl/Fl+Py

10

0.5
BaA/228

10

0.5
IP/IP+Bghi

1

Figure 4.3e. Selected ratios of PAH isomers in sediment of Otter Lake. Vertical lines represent
specific PAH diagnostic ratios used to distinguish source, see text for specific examples.

98

Round
2000
1980
1960
1940
1920
1900
1880
1860
1840
1820
1800

0

0.5 0
An/178

0.5
Fl/Fl+Py

10

0.5
BaA/228

10

0.5
IP/IP+Bghi

1

Figure 4.3f. Selected ratios of PAH isomers in sediment of Round Lake. Vertical lines represent
specific PAH diagnostic ratios used to distinguish source, see text for specific examples.

99

Sand
2000
1980
1960
1940
1920
1900
1880
1860
1840
1820
1800

0

0.5 0
An/178

0.5
Fl/Fl+Py

10

0.5
BaA/228

10

0.5
IP/IP+Bghi

1

Figure 4.3g. Selected ratios of PAH isomers in sediment of Sand Lake. Vertical lines represent
specific PAH diagnostic ratios used to distinguish source, see text for specific examples.

100

In Avalon Lake the sediment ratios of Fl/Fl+Py and IP/IP+Bghi do not indicate the same
source for PAHs. The Fl/Fl+Py ratio can be interpreted as being from a combustion source
which is consistent with the PAH profile plots that revealed the cement plant in Alpena as the
most likely source for both Hg and PAHs. The conflict then is likely a result of sources mixing.
Flouranthene can be enriched in particulate transported long distances (Yunker and Macdonald,
2003) resulting in the higher Fl/Fl+Py ratios observed in Avalon Lake while the more local
source of IP would be from the combustion of fossil fuels, possibly the power generating station
located due west of the lake (Figure 4.4) which combusts natural gas or local vehicle traffic. The
latter is unlikely however since this area of the State is lightly populated. Although both natural
gas combustion and cement production are sources of Hg it is likely that cement production
would be more dominant in this case since the IP/IP+Bghi ratio hovers between fossil fuel and
biomass combustion and the Fl/Fl+Py clearly indicates biomass combustion. The cause for the
decoupling of the Hg profile from that of the PAH profile is unclear. Decreases in PAH could be
a result of decreased demand for cement coupled with new sources of limestone or coal fly ash
used in the production that had a higher Hg content resulting in further increases of Hg between
1920 and 1980.

The cause of concurrent peaks of Hg and PAHs in the sediments of Birch Lake in the
1990s could not be clearly identified using indicator ratios or profile plots. However, similar to
Avalon Lake, PAH ratios do not indicate the same source. The ratio of Fl/Fl+Py indicates a
combustion source while that of IP/IP+Bghi varies between fossil fuel sources and biomass
combustion. The mixture of sources is confirmed by the BaA/228 ratios which transition from
petroleum deep in the core towards combustion and petroleum sources coincident with the

101

stabilization of the Fl/Fl+Py and IP/IP+Bghi ratios. Since Birch Lake is located in a highly
agricultural landscape it is likely that the combustion source identified by the increased Fl is
from the increased combustion of coal from the greater Chicago, IL/Gary, IN area. The State
Line Power Plant along the Indiana/Illinois border, to the southwest of Birch Lake, was built in
the 1920s (Indiana, 2011) concurrent with the transition of the An/178 ratio from petroleum to
combustion and a change in the rate of Hg increase. During the peak of PAHs in the late 1980s
into the 1990s mass group 202 and 252 are dominant consistent with results found for areas
downwind of industrialized and urbanized areas (Yunker and MacDonald, 2003).

In Crystal Lake between 1820 and 1880 the Fl/Fl+Py and IP/IP+Bghi ratios indicate
biomass combustion, consistent with the burning and clearing of forest that occurred in this
region (Long et al., 2010). After about 1930, both of these indicator ratios become stabilized
near the ratio discriminating between fossil fuel and biomass combustion. This is likely the
result of sources mixing since Crystal Lake is within close proximity of several natural gas
consumers and immediately east of several coal fired power plants (Figure 4.4).

Muskegon Lake is located within a large metropolitan area and has a coal fired power
plant, along with other industrial activities, on its shoreline. One might assume that the PAH
diagnostic ratios would be dominated by a coal signal. However, the Fl/Fl+Py and IP/IP+Bghi
ratios both indicated fossil fuel combustion as the dominant source rather than coal. Ratios
indicating fossil fuel combustion could be due to a vehicular source which would be likely for
this urbanized region; vehicle emissions were found to contribute just under 10% of PAH load to
southern Lake Michigan (Simcik et al., 1999) but, it is unlikely that a vehicular source could

102

result in the concentrations of Hg in the lake sediments (Parsons et al., 2007). Therefore, a
potential source could be the natural gas and fuel oil electricity generating stations located in
southeast Wisconsin, about 90 miles west and west-southwest of Muskegon Lake, for example
the Milwaukee County Power Plant has burned natural gas or #2 fuel oil for electricity
generation since 1970 (Resources, 2011).

103

¯
F
)
F
)
(

(
(

)
!
(
F
(
)

(

(
)

)

)
)
(
(
F
F

)
F)
!
(
)
X
F
(

X
)
(

Lake
Coal
Natural Gas
Oil
Wood

(
X

(
X
(
((
X

F
F
F
)

F
(
!
(
X
)
!
X
(
)
X
F
)
X
FF
X
F
)
F
F F ( ( ) ))
X
)
(
F
F
)
FF
((
F
F

150

!F (
( F
(

Kilometers

)F
X
)F F
)F

F
F ( )
(F )
)) ! ) )
(F

F
F
)
)
)
)

F
X
)
F X
F
)

X
)
!
(

F

)
F
F
X

F
F
)
) F F
X
)
) F
FFX
F F
F
X
X
F
F
X
F FXX)
)) )
X
)
F
F
)
! F F
( F F
F
)
F
)

Figure 4.4. Industrial and utility energy emitters according to the 2002 Mercury Emissions
Inventory for the State of Michigan.
104

F
)
)
F
F
X
)
X

The diagnostic ratios from Otter Lake reveal a dominant source for PAHs is from
biomass combustion. Mercury and PAHs concentrations are highly correlated suggest a
common source for these contaminants; the proximity of Otter Lake to the coal fired power
plants (Figure 4.4) combined with diagnostic ratio data could indicate a common source, the
combustion of coal for power production.

In Round Lake, ratios of Fl/Fl+Py and IP/IP+Bghi indicate a biomass combustion source
throughout much of the core. Combustion sources in this area would likely be limited to the
burning of grasses and forests prior to 1900. Round Lake, located in the middle of the Upper
Peninsula, is not currently in close proximity to any industrial sources (Figure 4.4). A major
industrial activity in Upper Peninsula is mining; however, most historic mining related activities
have been located to the northwest on the Keewenaw Peninsula. The town of Fayette,
approximately 80 km south of the lake, would have been the closest industrial source. Located
on the shores of Little Bay De Noc between 1867 and 1891, Fayette was a major iron ore smelter
and used local timber to fuel the process (Michigan, 2011a). The peak of Hg and PAHs occurred
during the late 1920s into the 1930s, similar to Avalon Lake. And much like Avalon the higher
mass PAHs were predominant during the peak. This would be consistent with a biomass
combustion source (Yunker and Macdonald, 2003) during the industrial period of the town and
until recent times. Ratios tend to show biomass combustion after about 1900, which may be an
indication of the mining related activities occurring in Michigan’s Upper Peninsula at the turn of
the last century or the clearing of forests for agriculature or timber. The Hg profile was similar
to those of the PAHs until the peak and then became decoupled. The decoupling occurs at the
time of decline in the mining industry, due to the economic factors and maturity of the mines, in

105

the region. Diagnostic ratios after the peak ca. 1920 are consistent with the combustion of
biomass (e.g., coal) a likely source for Hg to Round Lake.

Sand Lake is located near the City of Detroit. Prior to 1900 Fl/Fl + Py and IP/IP + Bghi
ratios in the sediment indicate biomass combustion, consistent with biomass burning during
intense clear cutting of forests. Mercury also shows an increase prior to 1900 which is likely due
to the erodibility of forest cleared watersheds (Long et al., 2010). After 1900, Hg and PAH start
to increase which is most likely due to the the installation of the cement kiln in Cement City due
west of Sand Lake (Figure 4.5). The kiln was shut down in the early 1960s (Morrison, 1945).
After ca. 1940, the ratios of Fl/Fl + Py and IP/IP + Bghi were consistent with a biomass
combustion source. The predominance of the mass group 202 after 1920 suggests burning of
grass, hardwood and softwood; although it is usually accompanied by the mass group 178 (Fine
et al., 2001, Jenkins et al., 1996, Masclet et al., 1995). The presence of the mass group 202 may
be related to the use of marl at the Cement City cement manufacturer, which would contain
remnants of grasses and softwoods washed in from the marl source watershed. The slope of Hg
and PAH concentration profiles are similar between 1900 and 1950. After 1950, concentrations
of both contaminants continue to increase; however, Hg does so at a much higher rate. This may
be due to the decommissioning of the cement kiln and commissioning of 2 major coal-fired
power plants to the east of Sand Lake in the City of Detroit. The coincident peak and decline of
Hg and PAH in the 1980s and 1990s implies a common source or influence and would be
consistent with the implementation of abatement technologies required by the Clean Air Act.

106

¯
Round

!
(

!
(

!
(

Avalon

!
(
!
(

Lake
Cement Kiln

Muskegon

!
(

!
(

Crystal

Otter

!
(

!
(

Sand
150

Birch

!
(

Kilometers

!
(

!
(

Figure 4.5. Current cement kilns in the State of Michigan. The historic cement kiln located in
Cement City, MI is not shown but would have been located approximately 20 km due west of
Sand Lake.

107

Sediment chronologies of contaminants reflect activities within the watershed and
atmospheric sources that could be located within or outside of the watershed. This makes the
identification of sources challenging since the concentration recorded in the sediment profile
contains potential information from all sources. Earlier research attempts to assign sources based
on sediment chronologies have successfully used production and consumption records of
contaminants (Engstrom and Swain, 1997, Lockhart et al., 2000). This work compliments those
studies and helps to confirm sources identified using production/consumption methods.
Furthermore, this approach has demonstrated that sources for PAHs, both historic and modern, to
inland lakes throughout the State of Michigan vary by region. Assuming that PAHs and Hg
share common sources and pathways, then identifying these local to regional scale sources has
important implications for legislation seeking to limit the release of Hg to environment.
Furthermore, some of the historic sources of Hg may have not been identified using traditional
methods (e.g., cement kilns, Avalon and Sand lakes, and mining related activities, Round Lake).
A modern coal combustion source was found to be significant for lakes in the southern portion of
the State and in some cases can be correlated to plant start-ups. A coal combustion source could
not be identified for Muskegon Lake which has a coal-fired power plant on its shoreline, which
provides further insight into the local scale transport of Hg. This approach is limited; however,
to the availability and quality of PAH data and can be enhanced by including those PAHs
identified by others to have more discriminatory power (e.g., retene see (Yunker and Macdonald,
2003)).

108

REFERENCES

109

REFERENCES
Barrick, R.C., Prahl, F.G., 1987. Hydrocarbon Geochemistry of the Puget Sound Region .3.
Polycyclic Aromatic-Hydrocarbons in Sediments. Estuarine Coastal and Shelf Science
25, 175-191.
Bindler, R., 2003. Estimating the natural background atmospheric deposition rate of mercury
utilizing ombrotrophic bogs in southern Sweden. Environmental Science and Technology
37, 40-46.
Boutron, C.F., Vandal, G.M., Fitzgerald, W.F., Ferrari, C.P., 1998. A forty year record of
mercury in central Greenland snow. Geophysical Research Letters 25, 3315-3318.
Chester, R., Bradshaw, G.F., Ottley, C.J., Harrison, R.M., Merrett, J.L., Preston, M.R., Rendell,
A.R., Kane, M.M., Jickells, T.D., 1993. The atmospheric distributions of Trace-Metals,
Trace Organics and Nitrogen Species Over the North-Sea. Philosophical Transactions of
the Royal Society of London Series a-Mathematical Physical and Engineering Sciences
343, 543-556.
Di Leonardo, R., Vizzini, S., Bellanca, A., Mazzola, A., 2009. Sedimentary record of
anthropogenic contaminants (trace metals and PAHs) and organic matter in a
Mediterranean coastal area (Gulf of Palermo, Italy). Journal of Marine Systems 78, 136145.
Donahue, W.F., Allen, E.W., Schindler, D.W., 2006. Impacts of coal-fired power plants on trace
metals and polycyclic aromatic hydrocarbons (PAHs) in lake sediments in central
Alberta, Canada. Journal of Paleolimnology 35, 111-128.
Durnford, D., Dastoor, A., Figueras-Nieto, D., Ryjkov, A., 2010. Long range transport of
mercury to the Arctic and across Canada. Atmospheric Chemistry and Physics 10, 60636086.
Dvonch, J.T., Graney, J.R., Keeler, G.J., Stevens, R.K., 1999. Use of elemental tracers to source
apportion mercury in South Florida precipitation. Environmental Science & Technology
33, 4522-4527.
Dvonch, J.T., Graney, J.R., Marsik, F.J., Keeler, G.J., Stevens, R.K., 1998. An investigation of
source-receptor relationships for mercury in south Florida using event precipitation data.
Science of the Total Environment 213, 95-108.
Engstrom, D.R., Swain, E.B., 1997. Recent declines in atmospheric mercury deposition in the
upper Midwest. Environmental Science and Technology 31, 960-967.
Evers, D.C., Wiener, J.G., Driscoll, C.T., Gay, D.A., Basu, N., Monson, B.A., Lambert, K.F.,
Morrison, H.A., Morgan, J.T., Williams, K.A., Soehl, A.G., 2011. Great Lakes Mercury
Connections: The Extent and Effects of Mercury Pollution in the Great Lakes Region.
110

Biodiversity Research Institute. Gorham, MA. Report BRI 2011-18, 44 pages. Available
from: http://www.briloon.org/mercuryconnections/GreatLakes.
Fernandez, P., Vilanova, R.M., Martinez, C., Appleby, P., Grimalt, J.O., 2000. The historical
record of atmospheric pyrolytic pollution over Europe registered in the sedimentary PAH
from remote mountain lakes. Environmental Science & Technology 34, 1906-1913.
Fine, P.M., Cass, G.R., Simoneit, B.R.T., 2001. Chemical characterization of fine particle
emissions from fireplace combustion of woods grown in the northeastern United States.
Environmental Science & Technology 35, 2665-2675.
Fitzgerald, W.F., 1989. Atmospheric and oceanic cycling of mercury. Academic Press.
Fitzgerald, W.F., Engstrom, D.R., Mason, R.P., Nater, E.A., 1998. The case for atmospheric
mercury contamination in remote areas. Environmental Science & Technology 32, 1-7.
Gnandi, K., Bandowe, B.A.M., Deheyn, D.D., Porrachia, M., Kersten, M., Wilcke, W., 2011.
Polycyclic aromatic hydrocarbons and trace metal contamination of coastal sediment and
biota from Togo. Journal of Environmental Monitoring 13, 2033-2041.
Hites, R.A., Laflamme, R.E., Farrington, J.W., 1977. Sedimentary Polycyclic AromaticHydrocarbons - Historical Record. Science 198, 829-831.
Howe, T.S., Billings, S., Stolzberg, R.J., 2004. Sources of polycyclic aromatic hydrocarbons and
hexachlorobenzene in spruce needles of eastern Alaska. Environmental Science &
Technology 38, 3294-3298.
City of Hammond Indiana, 2011. History of State Line Generating Plant, Available online at
http://www.hammondindiana.com/history/stateline.htm. (Accessed: September 5, 2011).
Jaffe, D., Prestbo, E., Swartzendruber, P., Weiss-Penzias, P., Kato, S., Takami, A., Hatakeyama,
S., Kajii, Y., 2005. Export of atmospheric mercury from Asia. Atmospheric Environment
39, 3029-3038.
Jaffe, D., Strode, S., 2008. Sources, fate and transport of atmospheric mercury from Asia.
Environmental Chemistry 5, 121-126.
Jenkins, B.M., Jones, A.D., Turn, S.Q., Williams, R.B., 1996. Emission factors for polycyclic
aromatic hydrocarbons from biomass burning. Environmental Science & Technology 30,
2462-2469.
Kamman, N.C., Engstrom, D.R., 2002. Historical and present fluxes of mercury to Vermont and
New Hampshire lakes inferred from Pb-210 dated sediment cores. Atmospheric
Environment 36, 1599-1609.

111

Kannan, K., Johnson-Restrepo, B., Yohn, S.S., Giesy, J.P., Long, D.T., 2005. Spatial and
temporal distribution of polycyclic aromatic hydrocarbons in sediments from Michigan
inland lakes. Environmental Science & Technology 39, 4700-4706.
Keeler, G.J., Landis, M.S., Norris, G.A., Christianson, E.M., Dvonch, J.T., 2006. Sources of
mercury wet deposition in Eastern Ohio, USA. Environmental Science & Technology 40,
5874-5881.
Killin, R.K., Simonich, S.L., Jaffe, D.A., DeForest, C.L., Wilson, G.R., 2004. Transpacific and
regional atmospheric transport of anthropogenic semivolatile organic compounds to
Cheeka Peak Observatory during the spring of 2002. Journal of Geophysical ResearchAtmospheres 109.
Landis, M.S., Vette, A.F., Keeler, G.J., 2002. Atmospheric mercury in the Lake Michigan basin:
Influence of the Chicago/Gary urban area. Environmental Science and Technology 36,
4508-4517.
Lima, A.L.C., Eglinton, T.I., Reddy, C.M., 2003. High-resolution record of pyrogenic polycyclic
aromatic hydrocarbon deposition during the 20th century. Environmental Science &
Technology 37, 53-61.
Lindberg, S., Bullock, R., Ebinghaus, R., Engstrom, D., Feng, X.B., Fitzgerald, W., Pirrone, N.,
Prestbo, E., Seigneur, C., 2007. A synthesis of progress and uncertainties in attributing
the sources of mercury in deposition. Ambio 36, 19-32.
Lipiatou, E., Saliot, A., 1991. Fluxes and Transport of Anthropogenic and Natural Polycyclic
Aromatic-Hydrocarbons in the Western Mediterranean-Sea. Marine Chemistry 32, 51-71.
Lockhart, W.L., Macdonald, R.W., Outridge, P.M., Wilkinson, P., DeLaronde, J.B., Rudd,
J.W.M., 2000. Tests of the fidelity of lake sediment core records of mercury deposition to
known histories of mercury contamination. Science of the Total Environment 260, 171180.
Long, D.T., Eisenreich, S.J., Swackhammer, D.L. Volume III - Metal Concentrations in
Sediments of the Great Lakes; Government Printing Office, 1995.
Long, D.T., Parsons, M.J., Yansa, C.H., Yohn, S.S., McLean, C.E., Vannier, R.G., 2010.
Assessing the response of watersheds to catastrophic (logging) and possible secular
(global temperature change) perturbations using sediment-chemical chronologies.
Applied Geochemistry 25, 143-158.
Masclet, P., Cachier, H., Liousse, C., Wortham, H., 1995. Emissions of Polycyclic AromaticHydrocarbons by Savanna Fires. Journal of Atmospheric Chemistry 22, 41-54.
Menounou, N., Presley, B.J., 2003. Mercury and other trace elements in sediment cores from
central texas lakes. Archives of Environmental Contamination and Toxicology 45, 11-29.
112

State Of MIchigan, 2011. About Fayette Historic Townsite, Garden, Available online at
www.michigan.gov/dnr. (Accessed: September 5, 2011).
State of Michigan, 2011. Sediment Chemistry, Available online at
http://www.michigan.gov/deq/0,4561,7-135-3313_3686_3728-32365--,00.html.
(Accessed: September 5, 2011).
Morrison, P.C., 1945. Cement Plant Migration in Michigan. Economic Geography 21, 1-16.
Munthe, J., Bodaly, R.A., Branfireun, B.A., Driscoll, C.T., Gilmour, C.C., Harris, R., Horvat,
M., Lucotte, M., Malm, O., 2007. Recovery of mercury-contaminated fisheries. Ambio
36, 33-44.
Mzoughi, N., Chouba, L., 2011. Distribution of trace metals, aliphatic hydrocarbons and
polycyclic aromatic hydrocarbons in sediment cores from the Sicily Channel and the Gulf
of Tunis (south-western Mediterranean Sea). Environmental Technology 32, 43-54.
NOAA, 2011. Thunder Bay National Marine Sanctuary, Available online at
http://thunderbay.noaa.gov. (Accessed: September 5, 2011).
Obrist, D., Hallar, A.G., McCubbin, I., Stephens, B.B., Rahn, T., 2008. Atmospheric mercury
concentrations at Storm Peak Laboratory in the Rocky Mountains: Evidence for longrange transport from Asia, boundary layer contributions, and plant mercury uptake.
Atmospheric Environment 42, 7579-7589.
Oros, D.R., Simoneit, B.R.T., 2000. Identification and emission rates of molecular tracers in coal
smoke particulate matter. Fuel 79, 515-536.
Pacyna, E.G., Pacyna, J.M., Fudala, J., Strzelecka-Jastrzab, E., Hlawiczka, S., Panasiuk, D.,
2006. Mercury emissions to the atmosphere from anthropogenic sources in Europe in
2000 and their scenarios until 2020. Science of the Total Environment 370, 147-156.
Parsons, M.J., Long, D.T., Yohn, S.S., Giesy, J.P., 2007. Spatial and Temporal Trends of
Mercury Loadings to Michigan Inland Lakes. Environmental Science and Technology 41,
5634-5640.
Reidmiller, D.R., Jaffe, D.A., Chand, D., Strode, S., Swartzendruber, P., Wolfe, G.M., Thornton,
J.A., 2009. Interannual variability of long-range transport as seen at the Mt. Bachelor
observatory. Atmospheric Chemistry and Physics 9, 557-572.
Wisconsin Department of Natural Resources, 2011. Climate Change Guide e-Appendix,
Available online at http://dnr.wi.gov. (Accessed: September 5, 2011).

113

Rogge, W.F., Hildemann, L.M., Mazurek, M.A., Cass, G.R., Simoneit, B.R.T., 1993. Sources of
Fine Organic Aerosol 2. Noncatalyst and Catalyst-Equipped Automobiles and HeavyDuty Diesel Trucks. Environmental Science & Technology 27, 636-651.
Schuster, P.F., Krabbenhoft, D.P., Naftz, D.L., Cecil, L.D., Olson, M.L., Dewild, J.F., Susong,
D.D., Green, J.R., Abbott, M.L., 2002. Atmospheric mercury deposition during the last
270 years: A glacial ice core record of natural and anthropogenic sources. Environmental
Science & Technology 36, 2303-2310.
Simcik, M.F., Eisenreich, S.J., Lioy, P.J., 1999. Source apportionment and source/sink
relationships of PAHs in the coastal atmosphere of Chicago and Lake Michigan.
Atmospheric Environment 33, 5071-5079.
Sofowote, U.M., Allan, L.M., McCarry, B.E., 2010. Evaluation of PAH diagnostic ratios as
source apportionment tools for air particulates collected in an urban-industrial
environment. Journal of Environmental Monitoring 12, 417-424.
Sofowote, U.M., Hung, H., Rastogi, A.K., Westgate, J.N., Deluca, P.F., Su, Y.S., McCarry, B.E.,
2011. Assessing the long-range transport of PAH to a sub-Arctic site using positive
matrix factorization and potential source contribution function. Atmospheric
Environment 45, 967-976.
R version 2.6.2, 2008. R version 2.6.2
US EPA, 2011. Proposed Rule: National Emission Standards for Hazardous Air Pollutants From
Coal- and Oil-Fired Electric Utility Steam Generating Units and Standards of
Performance for Fossil-Fuel-Fired Electric Utility, Industrial-Commercial-Institutional,
and Small Industrial-Commercial-Institutional Steam Generating Units. US
EPA.Government Printing Office, Washington, D.C. 40 CFR Parts 60 and 63.
Usenko, S., Smonich, S.L.M., Hageman, K.J., Schrlau, J.E., Geiser, L., Campbell, D.H.,
Appleby, P.G., Landers, D.H., 2010. Sources and Deposition of Polycyclic Aromatic
Hydrocarbons to Western US National Parks. Environmental Science & Technology 44,
4512-4518.
Valente, R.J., Shea, C., Humes, K.L., Tanner, R.L., 2007. Atmospheric mercury in the Great
Smoky Mountains compared to regional and global levels. Atmospheric Environment 41,
1861-1873.
Wakeham, S.G., Schaffner, C., Giger, W., 1980. Polycyclic Aromatic-Hydrocarbons in Recent
Lake-Sediments 1. Compounds Having Anthropogenic Origins. Geochimica et
Cosmochimica Acta 44, 403-413.
Wang, G.D., Mielke, H.W., Quach, V., Gonzales, C., Mang, Q., 2004. Determination of
polycyclic aromatic hydrocarbons and trace metals in New Orleans soils and Sediments.
Soil & Sediment Contamination 13, 313-327.

114

Wang, Z.D., Fingas, M., Shu, Y.Y., Sigouin, L., Landriault, M., Lambert, P., Turpin, R.,
Campagna, P., Mullin, J., 1999. Quantitative characterization of PAHs in burn residue
and soot samples and differentiation of pyrogenic PAHs from petrogenic PAHs - The
1994 Mobile Burn Study. Environmental Science & Technology 33, 3100-3109.
Watras, C.J., Bloom, N.S., Hudson, J.M., Gherini, S., Munson, R., Claas, S.A., Marrison, K.A.,
Hurely, J., Wiener, J.G., Fitzgerald, W.F., Mason, R., Vandal, G., Powell, D., Rada, R.,
Rislov, L., Winfrey, M., Elder, J., Krabbenhoft, D., Andren, A.W., Babiarz, C., Porcella,
D.B., Huckabee, J.W., 1994. Sources and fates of mercury and methylmercury in
Wisconsin lakes. In: C.J. Watras, J.W. Huckabee (Eds.), Mercury Pollution: Intergration
and Synthesis. CRC Press, Ann Arbor, pp. 153-179.
Yohn, S., Long, D., Fett, J., Patino, L., 2004. Regional versus local influences on lead and
cadmium loading to the Great Lakes region. Applied Geochemistry 19, 1157-1175.
Yunker, M.B., Macdonald, R.W., 2003. Alkane and PAH depositional history, sources and
fluxes in sediments from the Fraser River Basin and Strait of Georgia, Canada. Organic
Geochemistry 34, 1429-1454.
Yunker, M.B., Macdonald, R.W., 1995. Composition and Origins of Polycyclic AromaticHydrocarbons in the Mackenzie River and on the Beaufort Sea Shelf. Arctic 48, 118-129.
Yunker, M.B., Macdonald, R.W., Cretney, W.J., Fowler, B.R., McLaughlin, F.A., 1993. Alkane,
Terpene, and Polycyclic Aromatic Hydrocarbon Geochemistry of the Mackenzie River
and Mackenzie Shelf - Riverine Contributions to Beaufort Sea Coastal Sediment.
Geochimica et Cosmochimica Acta 57, 3041-3061.
Yunker, M.B., Macdonald, R.W., Goyette, D., Paton, D.W., Fowler, B.R., Sullivan, D., Boyd, J.,
1999. Natural and anthropogenic inputs of hydrocarbons to the Strait of Georgia. Science
of the Total Environment 225, 181-209.
Yunker, M.B., Macdonald, R.W., Vingarzan, R., Mitchell, R.H., Goyette, D., Sylvestre, S., 2002.
PAHs in the Fraser River basin: a critical appraisal of PAH ratios as indicators of PAH
source and composition. Organic Geochemistry 33, 489-515.
Zechmeister, H.G., Dullinger, S., Hohenwallner, D., Riss, A., Hanus-Illnar, A., Scharf, S., 2006.
Pilot study on road traffic emissions (PAHs, heavy metals) measured by using mosses in
a tunnel experiment in Vienna, Austria. Environmental Science and Pollution Research
13, 398-405.
Zhang, X.L., Tao, S., Liu, W.X., Yang, Y., Zuo, Q., Liu, S.Z., 2005. Source diagnostics of
polycyclic aromatic hydrocarbons based on species ratios: A multimedia approach.
Environmental Science & Technology 39, 9109-9114.
Zheng, W., Lichwa, J., Yan, T., 2011. Impact of different land uses on polycyclic aromatic
hydrocarbon contamination in coastal stream sediments. Chemosphere 84, 376-382.
115

CHAPTER 5
ASSESSING THE NATURAL RECOVERY OF A LAKE CONTAMINATED WITH HG
USING ESTIMATED RECOVERY RATES DETERMINED BY SEDIMENT
CHRONOLOGIES.
Abstract
Deer Lake is an impoundment located near Ishpeming, MI, USA. Iron mining assay laboratories
located in Ishpeming disposed of Hg salts to the city sewer whose outfall was located along an
inlet to Deer Lake. An effort to remediate the system in the mid 1980s which consisted of
drawing down water in the impoundment in order to volatize Hg from the sediments did not
result in recovery of the system. Since the mid 1990s, the remediation strategy has been to allow
the continual burial of the contaminated sediments, i.e., natural recovery. The goal of this study
was to assess the effectiveness of this strategy. This was accomplished by investigating the state
of the system in terms of its recovery and estimating the time frame for recovery. Sediment
cores were collected in 2000 to determine historical trends in accumulation rates and
concentrations of Hg and other metals. Sedimentation rates and sediment ages were estimated
using

210

Pb. Event based dating (e.g., peak of

data due to non-monotonic features in the

210

137

Cs in 1963) was used to supplement

210

Pb

Pb profile and activities that were not at supported

levels at the base of the core. Selected results are that 1) drawdown significantly influenced
sedimentation patterns causing slopes for

210

Pb profiles that reflected the influx of older

sediment, 2) periods of iron production correlate to Hg loading indicating the point source for
contamination, a relationship not previously identified, 3) Hg:Al ratios indicate a recent change
to a watershed pathway for Hg loading and 4) Hg concentrations had decreased from their peak,
remain elevated, and were increasing after 1997. The cause of the recent Hg concentrations may
be related to influx of contaminated watershed soils or sediments. Estimating the time frame for
116

recovery is challenging in this system because the process of natural recovery seems to have
been arrested and deeper, uncontaminated sediments, were not recovered as a basis for reference.
However, a recovery to background conditions is likely not achievable since rates of Hg loading
to nearby lakes and the current rate of atmospheric deposition are greater than an estimate of
background conditions for Deer Lake. Assuming recovery continued after 2000, estimates of the
time required for recovery varied based on the system state used to define it (e.g., recent rates of
wet Hg deposition or Hg surface concentrations/fluxes from similar systems), but were less than
12 y. However, the recent increasing values of recovery indicators (e.g., Hg concentrations)
suggests that these estimates are conservative and will be longer if recovery remains arrested,
which may in part be due to the legacy of Hg contamination on the landscape. This study shows
that estimates of recovery of highly disturbed lake systems can be made in the absence of within
lake reference conditions by using comparisons to reference systems and challenges of
estimating ages from atypical

210

Pb activity profiles can be overcome in part using event-based

dating techniques.
Introduction.
Deer Lake is an impoundment located in the western Upper Peninsula of Michigan within
the Marquette Iron Range (Figure 5.1). An ecological assessment of the lake in 1979 determined
that fish and sediments were contaminated with Hg (MDNR, 1987). In 1981 a Michigan
Department of Natural Resources (MDNR) fish survey determined that fish contained up to 1.38
mg Hg/kg wet weight greatly exceeding the 1981 US average of 0.2 mg Hg/kg and the State of
Michigan Fish Consumption Advisory level (MDNR, 1987). Past gold and iron mining related
activities were considered to have led to the contamination of the system (MDNR, 1987). As a
result of the widespread contamination and threats to ecosystem and human health, the Deer
117

Lake area, including a portion of the Carp River (Figure 5.1), was designated an Area of Concern
(AOC) by the International Joint Commission in 1985 (USEPA, 2007b).

The first steps in the remediation of Deer Lake consisted of lake-level drawdown, fish
eradication, and then water level stabilization. In the fall of 1984 Deer Lake was drawn down to
its lowest level to facilitate Hg degassing from the sediments and to assist with fish eradication
(D'Itri et al., 1992, USEPA, 2007b). Fish eradication efforts included addition of a sharp grate
on the dam exiting Deer Lake (killing any fish that might pass through), net-and-kill, and
rotenone application; fish were not removed from the system. Prior to 1984 lake levels in Deer
Lake were allowed to fluctuate in order to control spring floods and flow through the Cliffs
Electric Service Company hydroelectric dam. Fluctuating lake levels might have led to
increased rates of Hg methylation in the sediments (D'Itri et al., 1992). Therefore, maintaining a
constant water level was included as part of the remediation strategy. In 1987 the lake was
refilled and restocked with yellow perch and walleye. It was thought that water level
stabilization would reduce the potential for the sediments to act as a source of methyl-Hg to the
water column, and reduce Hg burdens in fish. However, in 1999 Hg burdens in large fish were
still elevated with respect to other local lakes (USEPA, 2007b)

118

Figure 5.1. Deer Lake study area. Expanded view shows an areal photograph of the study area
clipped to the Deer Lake shoreline during the major drawdown. Yellow circles indicate
sampling locations. Also shown are the approximate locations of the Rope’s Gold Mine, Carp
River Dam and the City of Ishpeming WWTP Outfall.

119

Since initial remediation efforts failed, three alternative remediation strategies were
proposed: 1) dredging to remove the most contaminated sediments, 2) installing a gel-cap to
isolate the contaminated sediments from newly deposited sediments or 3) allowing new sediment
to accumulate and bury the contaminated sediment (natural recovery). To gain insight into the
feasibility of natural recovery, sediment cores were taken from the lake to assess current
conditions and if temporal trends of Hg loadings and concentrations indicated natural recovery.
Natural recovery processes would be indicated by Hg loadings showing a decreasing trend. If
recovery is occurring then the time to recovery could be estimated as the time required for the
surface sediments to have concentrations or fluxes similar to geochemical background. In
addition when the major pathway for contaminant is atmospheric transport as it is for Hg, then
another indicator for the state of recovery would be current atmospheric deposition rates.

History and Site Description
A complete history of the Deer Lake AOC is described in the 1987 Remedial Action
Plan (RAP) available from the U.S. Environmental Protection Agency (USEPA) AOC website
(USEPA, 2007a). Key dates for this study are highlighted in Table 1.

Deer Lake was directly impacted by two mining operations, the Rope’s Gold Mine
(RGM) and Cleveland Cliffs Incorporated (CCI). The RGM operated on the western shore of
Deer Lake from 1882-1899; Hg was used to amalgamate gold during this period. Tailings piles
from the mine still exist along the northwestern shoreline of Deer Lake (Figure 5.1); however,
most of these tailings are covered by sediments or the road way that separates Causeway Pond
from Deer Lake. CCI operated iron mines in the region and maintained analytical laboratories in

120

the City of Ishpeming. From 1929 to 1981 CCI used mercuric chloride to assay iron ore and
discharged the wastes into the City of Ishpeming sewer system. The sewer outfall for the City of
Ishpeming was located on Carp Creek (Figure 5.1), approximately 1 km upstream of Deer Lake.
CCI ceased this activity in 1981 after an ecological survey of the area determined that the
sediments and fish of Deer Lake were contaminated with Hg.

Deer Lake has undergone significant hydrologic changes over time. Prior to the
installation of a series of dams between 1912 and 1942 on the Carp River, Deer Lake was much
smaller (36 ha) than present day (367 ha). The first dam was installed by the RGM, near the
present dam site, to satisfy water needs for the mining process. The present dam was installed by
the Cliffs Electric Service Company. Between 1979 and 1984, the water level of Deer Lake was
regulated for flood control of spring run-off and power generation, resulting in annual water
level changes of approximately 2 m. It has been estimated that 52% of Deer Lake sediments
were exposed to the atmosphere during these annual water level fluctuations (MDNR, 1987).

The City of Ishpeming waste water treatment plant (WWTP) prior to 1986 consisted of
only primary treatment of the incoming waste stream and therefore did not remove dissolved
heavy metals or nutrients (D'Itri, 2008). Mercury wastes from CCI were in particulate form and
did, for the most part, settle out during treatment; however, heavy rains would cause flooding of
the treatment plant resulting in solid waste flowing into Carp Creek (D'Itri, 2008). Excess
nutrient load from the WWTP resulted in the eutrophication of Deer Lake with dissolved oxygen
supersaturation of the epilimnion and hypolimentic anoxia (MDNR, 1987). This has important
implications for the cycling of Hg in the water column and sediments of the lake. Methyl-Hg

121

production is favored and methyl-Hg bioaccumulation is significantly enhanced under anoxic
conditions (Eckley et al., 2005, Munthe et al., 2007, Watras et al., 1994). Therefore,
anthropogenic eutrophication may have lead to the increased rate of Hg methylation in Deer
Lake and consequently higher Hg burdens in fish. A new WWTP, serving the City of
Ishpeming, which has the capability of removing nitrogen and phosphorus, was put on-line in
1986. Reduced nutrient input should result in a reduced rate of eutrophication; this should also
result in decreased rates of Hg methylation by reducing the periods of anoxia in bottom waters.

Although it cannot be shown that the water level stabilization in Deer Lake resulted in
reduced rates of Hg methylation, recent work has shown that methyl-Hg production may be
enhanced during sediment resuspension (Sunderland et al., 2004, Tseng et al., 2001) and
drying/re-wetting cycles (Rumbold and Fink, 2006). The fluctuation in water level would also
expose areas of the lake bottom to periods of higher energy regimes that would lead to
movement of more sediment into depositional zones. Under stable water level conditions these
sediment would not otherwise be transported (Shotbolt et al., 2005). Thus, higher energy
regimes could result in the transport of Hg contaminated sediments from the littoral zone to the
profundal zone in Deer Lake, increasing the burden of sediments exposed to anoxic bottom
waters. The increased mass of Hg-contaminated sediments exposed to anoxic bottom waters
may have led to increased rates of methyl-Hg production and increased fish Hg burdens.

Sampling and Analytical Methods

122

In the summer of 2000 sediment cores were retrieved from the deepest portion of the
North and South basins of Deer Lake. Aerial photography indicated these sites had remained
under water during the major drawdown period (1984-1987). Sediment cores were collected
using an Ocean Instrument MC-400 Multicorer (www.oceaninstruments.com) deployed from the
US EPA R/V Mudpuppy. The MC-400 collects four simultaneous cores which were used for
porewater metals analyses, radiometric dating, and sediment metals analyses; the fourth core was
kept as an archive core. Polypropylene jars and polyethylene bottles used to collect and/or store
samples for trace metal analyses were thoroughly acid cleaned, using trace metal grades
hydrochloric (HCl) and nitric (HNO3) acids, using one of three procedures: 1) cold-acid
washing; cold 24hr 10% hydrochloric acid (HCl) cycle followed by a 24hr pure water (≥18 µScm) rinse; 2) hot acid washing, hot 24hr 15% HCl cycle followed by a 24hr pure water rinse; 3)
clean-acid washing, cold 24hr 1:1 HCl cycle followed by a hot 24hr 1:1 HNO3 cycle followed by
a 24hr pure water rinse. Cores intended for radiometric dating and sediment metals analyses
were sectioned on-shore immediately after retrieval at 0.5 cm increments for the first 8 cm and at
1 cm increments for the remainder of the core. The effects of core smearing during extrusion
were reduced by discarding those sediments in contact with the walls of the core tube (metals
core only). Sediment samples were placed in cold-acid-washed polypropylene jars and stored on
ice until return to the laboratory.

In the laboratory, sediment samples for elemental analysis were stored frozen until being
freeze-dried prior to microwave-assisted nitric acid digestion using EPA Method 3051 (USEPA,
2007c). Samples for radiochemical dating were shipped to the Freshwater Institute, Winnipeg,
Canada. Reference material NIST 8704 Buffalo River Sediment (RM8704) and procedural

123

blanks were analyzed with each digestion batch. Precision of the digestion procedure was
evaluated with NIST RM8704, Buffalo River Sediment, %RSD for 10 replicates was 9.0% and
6.0% for Al and Hg, respectively. Concentrations of metals in procedural blanks were
consistently lower than the quantification limit (MDL) of the analytical method (HgMDL = 1.27
ng/mL, AlMDL = 91.93 ng/mL). Digestates were diluted to 100 mL with pure water (≥18 µS-cm)
prior to filtration with acid-cleaned, cold 10% HNO3 followed by a 24hr pure water rinse, 0.40
µm Nucleopore filters. Total metal and Hg samples were filtered into separate hot acid washed
60 mL polyethylene bottles. Samples for Hg analysis were preserved by adding 200 µL of a 100
ppm gold chloride solution. All samples were stored in the dark at 4ºC prior to analysis.
Sediment iron and calcium concentrations were determined using atomic absorption
spectrophotometry (Yohn et al., 2002). All other sediment metal concentrations were
determined using a Micromass Platform (now Thermo) ICP-MS with hexapole collision cell
technology. A 2% HNO3 rinse solution was used for all sediment metals analyses, except for Hg
which used a 2.5% HNO3 + 2.5%HCl + 10 ppm gold chloride rinse solution. The HNO3/HCl +
gold solution was used to prevent memory effects of Hg in the mass spectrometer (Allibone et
al., 1999).

The radionuclide

210

Pb was used to determine sedimentation rates and assign dates to the
137

sediment cores and cesium-137 (

Cs) was measured in select slices of sediment from the North

Basin to determine the peak value in
137

137

Cs activity. Detailed methods of analysis for

210

Pb and

Cs can be found elsewhere (Wilkinson and Simpson, 2003). Cs-137, was determined by

direct counting of 5-10 g of sediment on a γ-spectrometer (Ge-Li) semiconductor detector.
124

Samples of 1-3 g were analyzed for
presence of a

209

210

Pb and radium-226 by leaching in 6 N HCl in the

Po tracer, autoplating Po onto a silver disk, and counting the disk on an α-

spectrometer to determine

210

Pb via its

210

Po daughter. Sediment slice ages were estimated

using the constant rate of supply (CRS) model (Appelby and Oldfield, 1978) or the regression
equation of excess

210

Pb activity versus total accumulated sediment dry mass using the constant

flux:constant sedimentation (CF:CS) (Golden et al., 1993) and segmented CF:CS models
(Heyvaert et al., 2000).

Results and Discussion

Effects of Lake Level Drawdown on
210

210

Pb Profiles. Vertical changes in the slope of excess

Pb versus depth are used to identify changes in sedimentation rates caused by natural

(Appleby et al., 1998) or anthropogenic processes (Heyvaert et al., 2000) and were used here to
investigate the effects of the major drawdown period in the mid to late 1980s on sedimentation
rates in Deer Lake. Higher sedimentation rates are indicated by more vertical slopes of the
excess

210

Pb profile. It was anticipated that during the drawdown period more material from the

littoral areas would be focused into the deeper portions of the lake because of the increased
energy regimes (Shotbolt et al., 2005). Such focusing would be reflected by an increase in slope
during the drawdown.

125

Vertical profiles of excess

210

Pb activities, in general, increased to the surface in both

sediment cores (Figure 5.2). However, in the South Basin

210

Pb activities increased until about

17 cm depth. Then between 17 and 10 cm there was a large deviation from the log-linear trend.
Deer Lake
South Basin

Deer Lake
North Basin

0

0

5

5
1988

10

10
1987
15

1981

20

Depth (cm)

Depth (cm)

15

25
30

20
25

1978

30

35

35

40

40

45

45

50 -2
10

-1

10
210
Excess
Pb (Bq/g)

10

50 -2
-1
0
10
10
10
210
137
Excess
Pb or
Cs (Bq/g)

0

210

137

Figure 5.2. Excess
Pb profiles (solid circles) from the South and North Basins and the
Cs
profile (open circles) from the North Basin of Deer Lake. Double arrows indicate the large
210
deviation in the excess
Pb profile; CRS model dates were assigned to the beginning and end
of the large deviation for their associated mass depths.

126

210

Maximum

Pb activity was at approximately 4 cm and then decreased towards the sediment-

water interface. In the North Basin

210

Pb activities generally increased until around 22 cm

depth. From 20 to 14 cm activities were lower and constant. Above 14 cm

210

Pb activities

increased and peaked at 3 cm and then decreased to the surface.

From our past work (Parsons et al., 2007, Yohn et al., 2004, Yohn et al., 2002) the
CF:CS and CRS models were considered the most appropriate methods of estimating dates for
sediment cores collected from inland lakes of Michigan. The CF:CS model assumes a constant
flux of

210

Pb and a constant sedimentation rate and more appropriately models core profiles that

exhibit a log-linear trend (Appleby et al., 1998, Robbins and Edgington, 1975). The CRS model
also assumes a constant flux of

210

Pb to the sediment surface but allows for changes in

sedimentation rate to account for the inventory of excess
the product of sediment dry mass and excess

210

210

Pb activity, where the inventory is

Pb summed for the entire core. This model

would be useful for interpretations of small scale deviations in the

210

Pb profile, such as those

above 15 cm in the South Basin. However, to calculate the CRS model unsupported
reach 99% equilibrium with the supporting

226

210

Pb must

Ra which occurs in approximately 150 y (Appelby

and Oldfield, 1978, Krishnaswamy et al., 1971, Robbins, 1978) which was not observed in either
sediment core.

127

Cesium-137 is a radionuclide introduced to the environment during the atmospheric
testing of nuclear weapons. Testing of these weapons peaked in 1963 thus supplying a
chronostratigraphic marker in lake sediments (for a review of chronostratigraphic markers see
(Appleby, 2001)). When the CF:CS model was used to estimate dates the peak in

137

Cs activity

from the North Basin core was assigned a date of 1956, almost a decade early. Furthermore, the
deviation in

210

Pb activity at 22 cm depth, which we assume to be a result of lake level

management or draw down occurred in 1973, five years before the active water level
management of the lake. It appears that the CF:CS model did not assign appropriate dates to
these events, and it would be useful to date the core with an alternative method. Since excess
210

Pb was detected in the bottom slices of the South and North basin sediment cores use of the

CRS model might be considered limited because the

210

Pb inventory could not be calculated

directly (Appelby and Oldfield, 1978). However, in an effort to more appropriately reflect the
changes in the trajectories observed in the excess
model was used. First, activities of excess
equilibrium with supported

210

210

210

Pb profile a modified approach to the CRS

Pb for the North Basin were extrapolated to

Pb. Resultant dates were then adjusted by adding

inventory (less than 1%) to the last core sample so that the peak of
Using this approach the deviation in the

210

137

210

Pb

Cs occurred in 1963.

Pb profile that occurred at 22 cm dated to 1978, the

start of the active water level management period. Activities of excess

210

Pb between 20 and 14

cm were constant indicating a zone where sediments have been mixed. Both the modified CRS

128

approach and the CF:CS model assigned a date of 1987, the year the lake was refilled, to the
upper portion of this zone.

Since the CRS dates for the North Basin more appropriately reflected the period of active
water level management period (1979-1984, Table 1) and the major drawdown and lake refilling
(1984-1987, Table 1) a similar modified CRS model approach was used to assign dates to the
South Basin core. The South Basin core was not submitted for
change in excess

210

137

Cs analysis so the trajectory

Pb at 17 cm was used to adjust dates with the assumption that the change

occurred in the late 1970s or early 1980s coincident with the start of active water-level
management of Deer Lake. Similarly to the North Basin, excess

210

added to the last core sample so that the 17 cm transition to lower

Pb inventory (~1%) was

210

Pb activities occurred in

1981. Doing so placed the upper limit of the mixing zone, at 10 cm, in 1988. This is similar to
the procedure used to correct for old date error (Binford, 1990). The CF:CS model assigned
dates of 1979 and 1993 to the 17 and 10 cm depths, respectively. Although the former date
appears to reflect the beginning of the water-level stabilization period the latter does not
correspond to the refilling of Deer Lake as might be expected and found for the North Basin.
Since the

137

Cs profile could not be used to verify the dates assigned by the CRS model, the

peak in stable Pb concentration was used to assess the accuracy of the assigned dates (Figure 5.3)
Stable Pb concentrations in sediment cores should peak in the mid 1970s consistent with the
removal of Pb from gasoline and this approach has been used to confirm sediment ages (AlfaroDe la Torre and Tessier, 2002, Yohn et al., 2004, Yohn et al., 2002). The Pb concentration
profile from the South Basin had two peaks: the first in the early 1960s and the second in the
129

early 1970s using the CRS model (Figure 5.3). The older peak had concentrations that were
greater, but following the more recent peak concentrations steadily decreased. We consider the
latter peak, occurring in 1973 (1971 using the CF:CS model), followed by the decline to be
indicative of the removal of leaded gasoline and therefore useful as a event-based dating tool. In
the North Basin the peak in Pb concentration occurred in the late 1970s (1974 using the CF:CS
model) (Figure 5.3) The approach of using a modified CRS model resulted in a more
appropriate reflection of age information obtained from the peak in stable Pb as well as events
influencing sedimentation rates in Deer Lake. Thus, the dates from the modified CRS approach
will be used here.

130

Deer Lake
South Basin

Deer Lake
North Basin

2000

2000

1995

1995

1990

1990

1985

1985

1980

1980

1975

1975

1970

1970

1965

1965

1960

1960

1955

1955

1950

1950

1945

1945

1940

1940

1935

1935

1930

1930

1925

0

0.5
Normalized
Concentration (mg/kg)

1

1925

Mercury
Lead
0

0.5
Normalized
Concentration (mg/kg)

1

Figure 5.3. Profiles of mercury (open markers) and lead (solid markers) normalized to their
respective maximums to highlight similarities between the profiles.

Trends of

210

Pb activities that are steep or approaching vertical would be anticipated as a

result of increased sediment fluxes and mixing during the draw down (e.g., (Appleby et al.,
1998, Robbins, 1982); however, the negative slopes were not. Sediment dry mass was higher
during this period consistent with an increased flux of sandy sediment from the littoral zones of
the lake (Figure 5.4), evidenced by sediment color and texture change. The negative slopes
indicate that sediments were diluted with older (i.e., lower

210

Pb activities) sediment. Dates

corresponding to the 17 and 22 cm depths were 1981 and 1978 in the South and North basins,

131

respectively, which is just prior to the active water level management period of Deer Lake (Table
1). It is highly probable that, during the water level management period and major drawdown,
older sediment from shallower non-depositional zones of the lake as well as shoreline sediments
were focused to the deeper areas where the samples were taken. The prolonged exposure of
lake-bottom sediments would 1) allow wind erosion of exposed sediments and 2) create zones of
higher energy at lake locations that would otherwise be at lower energy regimes (Shotbolt et al.,
2005). These processes could result in the transport of older sediment into the depositional zone
diluting the activity of

210

Pb. Lead can be influenced by post depositional mobility (Benoit and

Rozan, 2001); however, vertical patterns of

210

Pb activities would not be expected if this

processes were occurring and the correlation of

210

Pb deviations with human activities provides

strong evidence against the mobility of Pb in the sediment column.

132

Deer Lake
South Basin

Deer Lake
North Basin

0

0

5

5
1988

10

10
1987
15

1981

20

Depth (cm)

Depth (cm)

15

25
30

20
1978

25
30

35

35

40

40

45

45

50

0

5
10
Dry Mass (g)

50

15

0

5
10
Dry Mass (g)

15

Figure 5.4. Dry mass profiles from the South and North basins of Deer Lake. Double arrows
210
indicate the large deviation in the excess
Pb profile; dates are assigned to the beginning and
end of the large deviation for their associated mass depths.

The peak in excess

210

Pb activities occurred at about 4 and 3 cm depths in the South and

North basins, respectively. Near the surface

210

Pb activity values are often homogenized

because of biotubation and currents resulting in near vertical slopes (Robbins, 1982). Again, the
negative slopes imply that older sediment was deposited over and perhaps mixed with newer
material. However, lake levels have not been intentionally lowered as was interpreted to be the
cause for the negative slopes down core. Thus a possible explanation might be explained by
drought conditions that existed in the study area during the 1990s.
133

The Palmer modified drought index (PMDI) indicated that there was a period of
prolonged drought (PMDI < 0) between June 1997 and June 1999 (National Oceanic and
Atmospheric Administration, 2007). This was followed by a brief, four month, recovery (PMDI
between 0 and 2) between July 1999 and October 1999. Drought conditions then persisted until
January 2001. This prolonged drought may have resulted in lake level lowering and exposure of
lake-bottom sediment and the influx of older sediment, similar to what may have occurred during
the major draw down.

Impact of mining activities on mercury concentration in Deer Lake. In the South Basin,
temporal changes in sediment Hg concentrations were more variable and concentrations higher
than in the North Basin (Figure 5.5). This is most likely due to the proximity of this basin to
Carp Creek, the discharge location for the City of Ishpeming WWTP (Figure 5.1). Recently,
sediment Hg concentrations from both basins have remained constant albeit elevated above
background concentrations from other Upper Peninsula of Michigan lakes (Parsons et al., 2007).
Prior to this, in the South Basin, sediment Hg concentrations increased steadily from the late
1920s until the mid 1940s, declined until the late 1950s, and then slowly increased until the mid
1970s. During the mid to late 1970s sediment Hg concentration increased, remained elevated
and then rapidly declined until the early 1980s, consistent with the removal of the point source of
Hg contamination. Concentrations then increased until the late 1980s followed by decreased
concentrations until the late 1990s. In the North Basin, sediment Hg concentrations slowly
declined from the 1950s until the mid 1960s, then increased until the early 1980s. Since then,

134

sediment concentrations declined until the late 1990s.
Deer Lake
South Basin

Deer Lake
North Basin

2000

2000

1995

1995

1990

1990

1985

1985

1980

1980

1975

1975

1970

1970

1965

1965

1960

1960

1955

1955

1950

1950

1945

1945

1940

1940

1935

1935

1930

1930

1925

0

10
20
30
Concentration (mg/kg)
or Iron Produced (metric ton)

1925

0

10
20
30
Concentration (mg/kg)
or Iron Produced (metric ton)

Figure 5.5. Hg concentration (mg/kg) profiles from the South and North basins of Deer Lake
(open circles) and production of usable iron ore from Michigan (closed circles, metric
ton(x1000)).

Because of the long history of base-metal mining in Michigan’s Upper Peninsula the
impact of mining activities has been well documented (Kerfoot et al., 1999, Kerfoot et al., 2004,
Kerfoot and Lauster, 1994, Kolak et al., 1998, Kolak et al., 1999) and although CCI was
eventually found liable for the contamination of Deer Lake the RAP (MDNR, 1987) does not
provide geochemical evidence to link CCI’s mining activities to Hg accumulation in Deer Lake.
The role of the abandoned RGM, specifically the mine tailings on the northwestern shore of the
lake, to mercury contamination has been unclear. Since our sediment records can not provide
135

information about the potential contribution of the mine tailings to mercury contamination (i.e.,
do not contain sediment of adequate age), profiles of Hg concentrations were compared to
production of iron ore from the Upper Peninsula of Michigan (USGS, 2009). The assumption is
that an estimate of CCI’s activities can be deduced from these production records. In other
words, when production of CCI was producing more iron they would in turn being assaying
more iron resulting in an increased load of Hg to Deer Lake. Therefore, if the sediment Hg
profiles can be correlated to the production records of iron ore then this would tend to indicate
CCI’s activities as the primary source of contamination Deer Lake Hg concentration profiles and
usable iron ore production from the Marquette Range of Michigan’s Upper Peninsula (19402000) (USGS, 2009) are shown in Figure 5.5. Production of iron ore from the State of Michigan
was greatest in the 1940s, when iron was needed for wartime manufacturing, and 1970s, when
monetary inflation and other economic forces created renewed interest in mining iron ore
(Fenton, 1998).

In the South Basin, sediment Hg concentrations exhibited high concentrations in the mid
1940s coincident with higher iron production during the 1940s and Hg concentration generally
tracks the production of iron ore between 1950 and 1980 (Pearson’s r = 0.59, p = 0.001) (Figure
5.5) providing further evidence for a mining related source for the Hg contamination of Deer
Lake. Mercury concentrations and iron ore production appeared to be correlated after the early
1980s to 1990, but their trends become decoupled after 1990. In the North Basin, iron ore
production and sediment Hg concentration did not correlate as well as in the South Basin
between 1950 and 1980 (Pearson’s r = 0.52, p = 0.006) (Figure 5.5). A general decrease in iron
ore production corresponds to the decrease in Hg concentration between 1950 and 1965 and the

136

peak in Hg concentration in the late 1970s correlates with the peak in iron ore production during
this period providing evidence for an iron-mining related source. Mercury concentration and
iron ore production become decoupled after ca. 1985, slightly earlier than what was observed in
the South Basin but more consistent with the removal of the point source of Hg contamination.
The smoother North Basin Hg concentration profile and poorer correlation to iron ore production
are suggestive of a slow bleed of Hg contaminated sediments from the South Basin. This
evidence coupled with the decreasing concentrations of Hg after removal of the point source
indicate that the Hg in the North Basin were also related to the CCI mining activities rather than
contributions from the RGM mine tailings. The reason for the later decoupling in the South
Basin is unclear, but may be related to watershed disturbance and is discussed below.

The influence of watershed processes on mercury concentrations in Deer Lake. If the City
of Ishpeming sewer were the sole pathway of Hg to Deer Lake via Carp Creek then Hg
concentrations should have decreased after CCI stopped the disposal of mercuric chloride into
the sewer in 1981. In both basins Hg concentrations clearly declined after 1981 (Figure 5.5);
however, the decrease in the South Basin was rapid while in the North Basin more gradual. The
more rapid decline of Hg concentrations in the South Basin is likely due to the response of this
basin to the removal of the point source and the proximity of this Basin to the WWTP discharge
location. The gradual decline observed in the North Basin could then be interpreted as a muted
response to the point source removal combined with inter-basin mixing of contaminated
sediments.

137

Since watershed soils are known to be significant sinks of metals (Mason and Sheu,
2002), when disturbed by human activities or climatic changes they can be significant sources of
metals to aquatic ecosystems. Metal/aluminum concentration ratio and aluminum concentration
profiles are often used to infer and quantify the contribution of chemicals derived from terrestrial
material (i.e., watershed soils) into a lake (Bruland et al., 1974, Koelmans, 1998, Tuncer et al.,
2001, Yohn et al., 2002). Ratios of Hg:Al were calculated for both basins and are shown in
Figure 5.6 along with aluminum concentrations. In the South Basin Al concentrations steadily
decreased from the late 1920s until the early 1980s, increased rapidly until the late 1980s and
have since decreased (Figure 5.6). The Hg/Al ratio was relatively constant prior to 1935 then
increased until the mid 1940s, decreased until the mid 1950s and generally increased again until
the mid 1970s (Figure 5.6). After the mid 1970s the ratio rapidly increases, peaking in the late
1970s, then rapidly decreases until the mid 1980s. Ratio values then generally decline until the
mid 1990s and increase thereafter. The above trends are interesting in that they can be
interpreted to indicate two different pathways of Hg to the South Basin. Between 1930 and the
early 1980s the Hg/Al ratio and Al concentration profiles were not correlated, which clearly
indicates a point source for Hg, i.e., mercuric chloride from CCI, to Deer Lake during this
period. During the major draw down period Al concentrations increased while the Hg/Al profile
decreased suggesting a dilution of Hg concentration due to increased sediment focusing, which is
the likely reason for the sharp decline in Hg concentrations at the beginning of the draw down
(Figure 5.5). After the late 1980s, about the time Deer Lake was refilled, and until the mid 1990s
Al concentrations once again start to decline and appear coupled to the Hg/Al ratio which
suggests a terrestrial pathway for Hg. After 1995 the two profiles are again decoupled,
indicating an additional source for Hg to the lake. The profile of Pb was similar to that of Hg

138

between 1983 and 1995 suggesting a watershed pathway for Pb as well (Figure 5.3), the profiles
prior the early 1980s were not as similar indicating unique sources for each chemical.

Deer Lake
South Basin

2000

Deer Lake
North Basin

2000

1990

1990

1980

1980

1970

1970

1960

1960

1950

1950

1940

1940

1930

1930

Aluminum
7

Hg/Al (x10 )
1920

0

5000
Hg/Al or Al (mg/kg)

10000

1920

0

5000
Hg/Al or Al (mg/kg)

10000

Figure 5.6. Profiles of Al (open circles) and Hg:Al ratio (solid circles) for the South and North
basins of Deer Lake.

In the North Basin Al concentrations, in general, decreased between 1950 until 1970,
increased until ca. 1983, decreased rapidly then slowly increased between 1983 and 1987 and
had since been generally declining until the late 1990s (Figure 5.6). The ratio of Hg/Al was
relatively constant between the 1950s and mid 1960s, increased from 1965 until the early 1970s,
139

remained relatively constant or increased slowly until 1985, and had since declined. The trends
in Al and Hg:Al ratio of the North Basin differ from those of the South Basin. The most
apparent difference is the broad nature of the trends compared to the more abrupt changes
observed in the South Basin. The effects of the direct addition of mercuric chloride are most
evident between 1965 and 1972. From 1972 to 1985 both the Hg/Al ratios and Al concentrations
were high which might indicate a contribution to Hg levels from South Basin contaminated
sediment as well as direct inputs. The correlated decrease in Hg/Al ratio and Al concentrations
between 1985 and 1997 can be interpreted as burial of contaminated sediments by less
contaminated sediment. The profile of Pb in the North Basin was more similar to Hg than the
South Basin throughout the entire sediment core. This is indicative of the processes leading to
deposition of chemical to the North Basin i.e., transport of sediment from the South Basin.

These patterns are interpreted to indicate that the pathway for Hg has changed from being a point
source to that of a terrestrial source (i.e., Hg laden soils/sediment from the watershed or Carp
Creek sediments). The possibility of the Carp Creek sediments as a source/pathway has
important implications for the recovery of Deer Lake in that total recovery may not take place
until the sediments of Carp Creek have been remediated.

Estimating Recovery using Hg Accumulation Rates and Sediment Concentrations. The
recovery of a system is often defined as the condition when contaminant loading rates or
concentrations reach a specified threshold or reference value (Magar, 2001). Threshold values
are concentrations or fluxes of contaminants to the environment usually defined by state and
federal agencies as above which there are adverse effects on ecological and human health.

140

Because threshold values are defined independent of system processes, they cannot be used to
indicate recovery of disturbed watershed-lake systems. Reference values are estimates of the
geochemical background flux or concentration of a contaminant prior to human perturbation.
However, recovery to a reference value may not be achievable when current loading rates are
greater than pre-disturbance conditions. For example, 1999 wet deposition rates of Hg along the
border between Wisconsin and Michigan’s Upper Peninsula varied between 9.0 and 13.3 µg Hg
-2 -1

m y (Illinois State Water Survey, 2008) compared to the background flux of about 8.3 µg Hg
-2 -1

m y for 5 Upper Peninsula lakes (Parsons et al., 2007). Total deposition would be greater
due to dry deposition of Hg (Lindberg et al., 2007) which can equal that of wet deposition for
this region (Landis and Keeler, 2002). A more reasonable recovery condition would then be that
of similar systems, in this case surface concentrations/fluxes of nearby inland lakes or equal to
the atmospheric deposition rate of a contaminant. In this way the recovery of a system could
then be defined as when the system comes into balance (steady-state and equilibrium processes)
with current inputs (e.g., atmospheric and watershed inputs) (Long et al., 2010).

To estimate the effects of the natural recovery remediation strategy for Deer Lake focusing
corrected accumulation rates (FCAR) of Hg were calculated as:

Hg Acc Rate ( µg m

−2

Hg (mg kg −1 ) × Sed Rate ( g m −2 y −1 )
y )=
(1)
FF
−1

where FF is the focusing factor and is calculated as the ratio of the actual

210

Pb inventory to that

which is expected and Sed Rate is the sedimentation rate calculated from the profiles of excess
210

Pb. The expected inventory of excess

210

-1

Pb for this region is 0.574 Bq cm (Golden et al.,

141

1993). Since neither core was deep enough to observe
excess

210

210

Pb equilibrium with 226-Radium, the

Pb inventory was estimated by extrapolation (Parsons et al., 2007). The FF’s were

approximately 1.7 and 2.7 for the South and North basins, respectively and were similar to other
Michigan lakes which range from 1 - 3 (Parsons et al., 2007, Yohn et al., 2004).

On average, focusing corrected accumulation rates (FCAR) for Hg were higher in the South
Basin than the North Basin (Figure 5.7) and are higher than other Upper Peninsula lakes (Parsons
et al., 2007) consistent with the pollution history of Deer Lake. Mercury FCAR showed only
small changes in the North or South basin since 1997, and may have started to increase,
suggesting that natural recovery had ceased. Surficial FCAR were greater for the North Basin
-2 -1

-2 -1

(1167 µg Hg m y ) than the South Basin (719 µg Hg m y ) and both are much higher than
what would be expected for wet Hg deposition to this region (Illinois State Water Survey, 2008).
The reasons for the cessation of recovery are unclear but may be due to watershed disturbance or
leaching of Hg from the RGM tailings piles. Surficial Hg FCAR for other Upper Peninsula lakes
-2 -1

vary between 14 and 63 µg Hg m y (Parsons et al., 2007) and would be expected for Deer
Lake in the absence of point-source Hg pollution. To estimate the time required for surficial
-2 -1

sediment Hg FCAR in Deer Lake to achieve a flux of 20 µg Hg m y , the median surficial Hg
flux for 5 Upper Peninsula lakes (Parsons et al., 2007), a linear regression was made on Hg
FCAR starting in 1993 and 1991 in the South Basin and North Basin, respectively (Figure 5.8).
These dates were chosen because they represent the most recent peaks or slope changes in Hg
FCAR. It is assumed here that the regions of relatively constant FCAR represent a brief period
of arrested recovery and were not included in the regression and that the processes resulting in
142

clean sediment being deposited on top of the more contaminated sediment continue at the same
rate. Regression equations calculated from the FCAR profile were shifted to the most recent
sediment slice to calculate a rate of recovery, that is, if recovery were to begin again how long
would it take for fluxes/concentrations to be in the new system state. However, it is not certain if
or when recovery will resume since the initial recovery was arrested, indicated by relatively
constant or increasing FCAR since 1997 (Figure 5.8). Results of this regression suggest that
recovery in the South and North Basins would be approximately 3 and 7 years, respectively.
Using the average of the 1999 wet deposition for the four Wisconsin and the northeast Minnesota
2

Mercury Deposition Network sites (9.1 µg Hg/m /y, Illinois State Water Survey, 2008), chosen
for their proximity to Deer Lake, as the system state the same results were found. This indicates
that the rate of recovery for Deer Lake is rapid and that natural recovery appears to be a proper
remediation strategy for this system. However, there is no guarantee that the loading of Hg to
Deer Lake will remain constant and a longer term monitoring strategy should be adopted.

Calculating fluxes of contaminants permits comparisons among lakes and comparisons to
atmospheric deposition rates. However, fluxes do not provide direct information about potential
exposure of benthos to sediment contaminants. The concentrations of Hg in the surficial
sediments were about one order of magnitude greater than most Upper Peninsula lakes (Parsons
et al., 2006). Thus, it would be helpful to know when concentrations of Hg in the surficial
sediment of Deer Lake would be more reflective of other Upper Peninsula lakes. Similar to the
flux regression analysis above, a linear regression on concentration was made to determine the

143

Deer Lake
South Basin

Deer Lake
North Basin

2000

2000

1995

1995

1990

1990

1985

1985

1980

1980

1975

1975

1970

1970

1965

1965

1960

1960

1955

1955

1950

1950

1945

1945

1940

1940

1935

1935

1930

1930

1925

0

5000
10000
2
Hg FCAR (µg/m /a)

1925

0

1000 2000 3000 4000 5000
2
Hg FCAR (µg/m /a)

Figure 5.7. Focusing corrected Hg accumulation (µg/m2/y) rates for the South and North basins
of Deer Lake.

144

Figure 5.8. Hg FCAR profiles for the South and North basins of Deer Lake, 1980 to present. The
dashed line shows the linear model used to determine the time to recovery, equations used are
also shown. The dash-dot-dash line shows the model applied to predict the year of recovery.

145

Deer Lake
South Basin
2015

Deer Lake
North Basin
2015
y = -2.2619x + 2008.9

y = -1.0611x + 1998.5
2010

2010

2005

2005

2000

2000

1995

1995

1990

1990

1985

1985

1980

0

2
4
6
8 10 12
Hg Concentration (mg/kg)

1980

0

2
4
6
8 10 12
Hg Concentration (mg/kg)

Figure 5.9. Hg concentration profiles for the South and North basins of Deer Lake, 1980 to
present. The dashed line shows the linear model used to determine the time to recovery,
equations used are also shown. The dash-dot-dash line shows the model applied to predict the
year of recovery.

time it would take for Deer Lake sediments to have Hg concentrations similar to surficial
sediments of other Upper Peninsula lakes (Figure 5.9). Assuming that sedimentation rates
remained constant, to be similar to current Hg concentrations (0.16 mg Hg/kg; average of 5
Upper Peninsula lakes (Parsons et al., 2007)) the South Basin would require at least 3 y, whereas
the North Basin would require 12 y. Again this analysis points out the connection between the
two basins of the lake. That is, the estimated recovery to current concentration, wet deposition
and surficial fluxes were the same for the South Basin, while those recovery estimates for
146

concentration in the North Basin were longer than the other two system states. This suggests that
the recovery of the North Basin is coupled to the recovery of the South Basin due to the transport
of sediment from the South to the North basin.

Conclusions

The goal of this study was to assess the state of recovery of a Hg contaminated lake in which
remediation is through the natural burial of contaminated sediments. The results revealed two
previously undocumented assumptions 1) a correlation between iron ore production from the
Marquette Range of Michigan’s Upper Peninsula and Hg concentrations in the profiles clearly
demonstrates the influence of mining related activities on Deer Lake and 2) Al concentration and
Hg:Al ratio profiles show a change in pathway for Hg from a point source to a non-point source
(atmospheric deposition and watershed runoff) with cessation of mining activities.

This study also shows the challenges in interpreting the states of highly disturbed systems. In
this case, atypical

210

Pb profiles and lack of within lake historic uncontaminated sediments

posed challenges to age determination and a reference condition for recovery assessment
comparison. These problems were caused by human and climatically induced water-level
fluctuations which introduced older sediments into the profundal areas of the lake and the depth
of contaminated sediment. Sedimentation rates and sediment ages were determined via
modifications to the constant rate of supply (CRS) model, combined with event-based dating
techniques. Several definitions of system state (e.g., flux and concentrations) were used to
estimate the time required for recovery.
147

Although recovery to background conditions is often one of the types of goals for a remediation
program it could not be used here because current rates of atmospheric Hg deposition for the
region are greater than an estimate of background conditions for Deer Lake. Thus, Hg loadings
can only recover to atmospheric loading rates. Considering this and using the most recently
measured surface sediment as a starting condition once recovery begins, and remains
uninterrupted, it will take between 3 and 7 y for the South and North Basin, respectively, for
recovery based on a flux reference and 3 y and 12 y, respectively, based on a concentrations
reference. However, the recently increasing values of recovery indicators (e.g., Hg
concentration) suggest that these estimates are conservative and will be longer if recovery
remains arrested. This may be due in part to the legacy of Hg contamination on the landscape.

148

REFERENCES

149

REFERENCES
Alfaro-De la Torre, M.C., Tessier, A., 2002. Cadmium deposition and mobility in the sediments
of an acidic oligotrophic lake. Geochimica et Cosmochimica Acta 66, 3549-3562.
Allibone, J., Fatemian, E., Walker, P.J., 1999. Determination of mercury in potable water by
ICP-MS using gold as a stabilising agent. Journal of Analytical Atomic Spectrometry 14,
235-239.
Appelby, P.G., Oldfield, F., 1978. The calculation of 210Pb dates assuming a constant rate of
supply of unsupported 210Pb to the sediment. Catena 5, 1-8.
Appleby, P.G., 2001. Chronostratigraphic Techniques in Recent Sediments. In: W. M. Last, J. P.
Smol (Eds.), Tracking Environmental Change Using Lake Sediments. Kluwer Academic
Publishers, Boston, pp. 171-205.
Appleby, P.G., Flower, R.J., Mackay, A.W., Rose, N.L., 1998. Paleolimnological assessment of
recent environmental change in Lake Baikal: sediment chronology. Journal of
Paleolimnology 20, 119-133.
Benoit, G., Rozan, T.F., 2001. Pb-210 and Cs-137 dating methods in lakes: a retrospective study.
Journal of Paleolimnology 25, 455-465.
Binford, M.W., 1990. Calculation and uncertainty analysis of 210Pb dates for PIRLA project lake
sediment cores. Journal of Paleolimnology 3, 253-267.
Bruland, K.W., Bertine, K., Koide, M., Goldberg, E.D., 1974. History of Metal Pollution in
Southern-California Coastal Zone. Environmental Science & Technology 8, 425-432.
D'Itri, F., 2008. Personal Communication.
D'Itri, F.M., Evans, E.D., Kubitz, J.A., Ushikubo, A., Kurita-Matsuba, H., 1992. The remediation
of mercury contaminated fish in an artificial reservoir. International Journal of
Limnology 52, 301-317.
Eckley, C.S., Watras, C.J., Hintelmann, H., Morrison, K., Kent, A.D., Regnell, O., 2005.
Mercury methylation in the hypolimnetic waters of lakes with and without connection to
wetlands in northern Wisconsin. Canadian Journal of Fisheries and Aquatic Sciences 62,
400-411.
Fenton, M., 1998. Iron and Steel. In: U.S. Geological Survey Metal Prices in the United States
Through 1998. pp. 61-64.
Golden, K.A., Wong, C.S., Jeremiason, J.D., Eisenreich, S.J., Sanders, M.G., Hallgren, J.,
Swackhammer, D.L., Engstrom, D.R., Long, D.T., 1993. Accumulation and preliminary
inventory of organochlorines in Great Lakes sediments. Water Science and Technology
28, 19-31.
150

Heyvaert, A.C., Reuter, J.E., Slotton, D.G., Goldman, C.R., 2000. Paleolimnological
reconstruction of historical atmospheric lead and mercury deposition at Lake Tahoe,
California-Nevada. Environmental Science and Technology 34, 3588-3597.
National Atmospheric Deposition Program (NRSP-3), Illinois State Water Survey, NADP
Program Office, 2204 Griffith Drive, Champaign, IL 61820, Available online at
http://nadp.sws.uiuc.edu/mdn/. (Accessed: February 5, 1998).
Kerfoot, W.C., Harting, S., Rossmann, R., Robbins, J.A., 1999. Anthropogenic copper
inventories and mercury profiles from Lake Superior: Evidence for mining impacts.
Journal of Great Lakes Research 25, 663-682.
Kerfoot, W.C., Harting, S.L., Jeong, J., Robbins, J.A., Rossmann, R., 2004. Local, regional, and
global implications of elemental mercury in metal (copper, silver, gold, and zinc) ores:
Insights from Lake Superior sediments. Journal of Great Lakes Research 30, 162-184.
Kerfoot, W.C., Lauster, G., 1994. Paleolimnological study of copper mining around Lake
Superior: Artificial varves from Portage Lake provide a high resolution record.
Limnology and Oceanography 39, 649-669.
Koelmans, A.A., 1998. Geochemistry of suspended and settling solids in two freshwater lakes.
Hydrobiologia 364, 15-29.
Kolak, J.J., Long, D.T., Beals, T.M., Eisenreich, S.J., Swackhamer, D.L., 1998. Anthropogenic
inventories and historical and present accumulation rates of copper in Great Lakes
sediments. Applied Geochemistry 13, 59-75.
Kolak, J.J., Long, D.T., Kerfoot, W.C., Beals, T.M., Eisenreich, S.J., 1999. Nearshore versus
offshore copper loading in Lake Superior sediments: Implications for transport and
cycling. Journal of Great Lakes Research 25, 611-624.
Krishnaswamy, S., Lal, D., Martin, J.M., Meybeck, M., 1971. Geochronology of Lake
Sediments. Earth and Planetary Science Letters 11, 407.
Landis, M.S., Keeler, G.J., 2002. Atmospheric mercury deposition to Lake Michigan during the
Lake Michigan Mass Balance Study. Environmental Science and Technology 36, 45184524.
Lindberg, S., Bullock, R., Ebinghaus, R., Engstrom, D., Feng, X.B., Fitzgerald, W., Pirrone, N.,
Prestbo, E., Seigneur, C., 2007. A synthesis of progress and uncertainties in attributing
the sources of mercury in deposition. Ambio 36, 19-32.
Magar, V.S., 2001. Natural recovery of contaminated sediments. Journal of Environmental
Engineering-Asce 127, 473-474.

151

Mason, R.P., Sheu, G.R., 2002. Role of the ocean in the global mercury cycle. Global
Biogeochemical Cycles 16.
MDNR, M.D.o.N.R. Surface Water Quality Division. Remedial Action Plan for Deer Lake
Area of Concern; Government Printing Office, 1987.
Munthe, J., Bodaly, R.A., Branfireun, B.A., Driscoll, C.T., Gilmour, C.C., Harris, R., Horvat,
M., Lucotte, M., Malm, O., 2007. Recovery of mercury-contaminated fisheries. Ambio
36, 33-44.
National Oceanic and Atmospheric Administration, 2008. NNDC Climate Online, Available
online at http://www7.ncdc.noaa.gov/. (Accessed: October 27, 2008).
Parsons, M.J., Long, D.T., Yohn, S.S., Giesy, J.P., 2007. Spatial and Temporal Trends of
Mercury Loadings to Michigan Inland Lakes. Environmental Science and Technology 41,
5634-5640.
Parsons, M.J., Yohn, S.S., Long, D.T., Patino, L.C., Stone, K.A., 2006. Inland Lakes Sediment
Trends: Mercury Sediment Analysis Results for 27 Michigan Lakes. Lansing, MI. 47
pages. Available from:
Robbins, J.A., 1978. Geochemical and geophysical applications of radioactive lead. In: J. O.
Nriagu (Ed.), The Biogeochemistry of Lead in the Environment. Elsevier, New York, pp.
285-393.
Robbins, J.A., 1982. Stratigraphic and dynamic effects of sediment reworking by Great Lakes
zoobenthos. Hydrobiologia 92, 611-622.
Robbins, J.A., Edgington, D.N., 1975. Determination of recent sedimentation rates in Lake
Michigan using Pb-210 and Cs-137. Geochimica et Cosmochimica Acta 39, 285-304.
Rumbold, D.G., Fink, L.E., 2006. Extreme spatial variability and unprecedented methylmercury
concentrations within a constructed wetland. Environmental Monitoring and Assessment
112, 115-135.
Shotbolt, L.A., Thomas, A.D., Hutchinson, S.M., 2005. The use of reservoir sediments as
environmental archives of catchment inputs and atmospheric pollution. Progress in
Physical Geography 29, 337-361.
Sunderland, E.M., Gobas, F., Heyes, A., Branfireun, B.A., Bayer, A.K., Cranston, R.E., Parsons,
M.B., 2004. Speciation and bioavailability of mercury in well-mixed estuarine sediments.
Marine Chemistry 90, 91-105.
Tseng, C.M., Amouroux, D., Abril, G., Tessier, E., Etcheber, H., Donard, O.F.X., 2001.
Speciation of mercury in a fluid mud profile of a highly turbid macrotidal estuary
(Gironde, France). Environmental Science & Technology 35, 2627-2633.

152

Tuncer, G., Tuncel, G., Balkas, T.I., 2001. Evolution of metal pollution in the Golden Horn
(Turkey) sediments between 1912 and 1987. Marine Pollution Bulletin 42, 350-360.
Deer Lake River Area of Concern, Available online at
http://www.epa.gov/glnpo/aoc/drlake.html. (Accessed: January 30, 2008).
US EPA, 2007. Great Lakes Areas of Concern, Available online at
http://www.epa.gov/glnpo/aoc/. (Accessed: January 31, 2008).
USEPA. Office of Solid Waste and Emergency Response. Test Methods for Evaluating Solid
Waste, Physical/Chemical Methods; Government Printing Office, 2007c.
United States Geological Survey, 2009. Minerals Yearbook, Available online at
http://minerals.usgs.gov/minerals/pubs/myb.html. (Accessed: July 9).
Watras, C.J., Bloom, N.S., Hudson, J.M., Gherini, S., Munson, R., Claas, S.A., Marrison, K.A.,
Hurely, J., Wiener, J.G., Fitzgerald, W.F., Mason, R., Vandal, G., Powell, D., Rada, R.,
Rislov, L., Winfrey, M., Elder, J., Krabbenhoft, D., Andren, A.W., Babiarz, C., Porcella,
D.B., Huckabee, J.W., 1994. Sources and fates of mercury and methylmercury in
Wisconsin lakes. In: C.J. Watras, J.W. Huckabee (Eds.), Mercury Pollution: Intergration
and Synthesis. CRC Press, Ann Arbor, pp. 153-179.
Wilkinson, P., Simpson, S.L. Geological Survey of Canada. Radiochemical analysis of Lake
Winnipeg 99-900 cores 4 and 8; Ottawa, Ontario: Government Printing Office, 2003.
Yohn, S., Long, D., Fett, J., Patino, L., 2004. Regional versus local influences on lead and
cadmium loading to the Great Lakes region. Applied Geochemistry 19, 1157-1175.
Yohn, S.S., Long, D.T., Fett, J.D., Patino, L., Giesy, J.P., Kannan, K., 2002. Assessing
environmental change through chemical-sediment chronologies from inland lakes. Lakes
Reserv Res Manage 7, 217-230.

153

CHAPTER 6
SUMMARY AND RECOMMENDATIONS
Summary
The working hypothesis of this research is: local scale sources and watershed scale
processes have more impact on the rate of Hg accumulation in the aquatic environment than
regional to global scale sources. The hypothesis was investigated using the following
approaches: 1) local scale sources are reflected in the sediment record as discontinuities in spatial
scale analysis of anthropogenic inventories and Hg profiles should be unique to a lake or subregion, 2) watershed scale processes will impact the rate of modern recovery from Hg
contamination and variations in the recovery rate between watersheds can be correlated to
differences in watershed attributes and 3) diagnostic ratios of PAHs infer sources of Hg since
they share a common source and pathway.
Spatial analysis of anthropogenic inventories amongst Michigan inland lakes were
indicative of a regional source for Hg assuming the inventories found for Lakes Michigan and
Superior are representative of a regional source. However, there were notable exceptions to this
generalization, specifically lakes Houghton, Higgins, Gull. Additionally, lakes that are in close
proximity should have anthropogenic inventories that are similar if a regional source were
present. Elk and Torch and Higgins and Houghton share watershed boundaries, but have quite
different anthropogenic inventories more consistent with watershed scale sources or processes.
Inventories from Upper Peninsula lakes were consistent with those found for Lake Superior, but
were varied suggesting the influence of watershed scale processes.
In general anthropogenic accumulation of Hg in the study lakes first occurred in the late
1800s consistent with the habitation of the lower Great Lakes region by European settlers.
Increased anthropogenic accumulation continues until the middle or latter part of the previous
154

century. Lakes from sub-regions of the study area had similar profiles of anthropogenic Hg
accumulation suggesting sources that were common to these lakes. Episodic accumulation
events common among those lakes provide further evidence of more local scale processes.
Recent trend analysis showed that several lakes had increasing Hg accumulation none of which
were located near coal-fired power plants, a predominant source for Hg, providing further
verification for local-scale or watershed process influence for these lakes.
To examine the impacts of watershed attributes on the modern recovery of lakes from Hg
contamination slopes of recent declines in Hg concentration and accumulation were calculated
and compared to GIS delineated watershed attributes. Strong correlations were found for
agriculture, forest and wetland attributes. However, those attributes along highly conductive
flowpaths or within close proximity to the lake had the strongest relationships. Though
relationships between the accumulation of Hg to the environment and watershed attributes have
been studied previously, the relationship between watershed attributes and dynamic flowpaths
was unique to this study. What may be a more important finding of this study is the lack of
relationship between recovery rate and measures of atmospheric Hg such as sulfate deposition
and Hg emissions. This lack of relationship suggests that local scale sources and watershed
processes determine the recovery of lakes from Hg enrichment rather than regional decreases in
Hg emission. However, the lack of relationship may also be related to “legacy Hg” deposited to
the landscape and the lag time required for that Hg to reach the aquatic ecosystem. The
implications of this are it is then those activities that disturb the landscape that control
accumulation of Hg in inland lakes.
Diagnostic ratios of PAHs can often be used to differentiate sources of these
contaminants. Mercury and PAHs share common sources and pathways allowing these

155

diagnostic ratios to infer sources of Hg to the environment. Modern ratios of PAH showed that
combustion sources predominate across the State. The prevalence of combustion sources prior to
1900 in Avalon and Round lakes may be indicative of early industrial activity near these lakes,
i.e., cement manufacturing and iron smelting, respectively. Coincident with the ratios of An/178
and Fl/Fl+Py a coal combustion source was potentially identified for Birch Lake, i.e., coal fired
power plants in the Chicago/Gary industrial complex. A coal combustion source was also
presumed to cause the increase in Hg in Sand Lake, which was coincident with the
commissioning of power plants in southeast Michigan. In Muskegon Lake where a coal-fired
power plant resides on the lakes’ shoreline no coal combustion source was identified. Rather a
source consistent with fossil fuels was observed. This could have two explanations either a
vehicular or natural gas power plant source. The lake is in a highly urbanized area of the state
and lies due east of a natural gas fired power plant in Wisconsin. The latter is more likely due to
the elevated Hg levels observed in the sediment of Muskegon Lake and the geochemical cycle of
Hg. Furthermore, the lack of evidence of a biomass combustion source may be indicative of the
long range transport of Hg due to the height of power plant smokestacks. Unlike Muskegon
Lake, Otter Lake showed evidence of influence by biomass combustion throughout the entire
core, likely coal fired power plants due to its proximity to power plants fired with coal to the
north and east. Crystal Lake showed a varied response between fossil fuel and biomass
combustion, evidence of mixed sources to the lake.
Burial in deep sediments is often the fate of Hg emitted to the environment. This was the
case for Deer Lake where Hg was used to assay iron ore in the Cleveland Cliffs Laboratory and
dumped to the City of Ishpeming municipal sewer system. Deer Lake became an Area of
Concern when it was determined that fish in the lake had concentrations of mercury that

156

exceeded health limits along with other beneficial use impairments. A high concentration of Hg
in the sediments was suspected to be the primary cause of fish Hg. When the original
remediation attempt failed to lower fish tissue concentrations it was determined to allow the lake
to recover naturally by the accumulation of clean sediments. Although the point source had
been removed Hg concentrations were still not reflective of a recovering system. Using Hg:Al
ratios it was determined that a watershed or erosional source for Hg could be an important
contemporary source which could be contaminated sediments upstream or in the floodplain of
the creek used for wastewater discharge. Calculations based on the slope of modern
concentrations and accumulation rates indicate a quick, ~10 y, recovery to those concentrations
and accumulation rates of nearby Upper Peninsula lakes even though surficial concentrations are
an order of magnitude greater than nearby lakes. This has important implications for the use of
natural recovery as a remediation strategy.
This is the first research to show strong correlations between watershed attributes and the
recovery of inland lake systems and the first to use ratios of PAHs to infer sources of Hg. Much
of the previous work investigating watershed attributes has focused on the impact of specific
land cover types, investigated a narrow time period or used a limited number of lakes. The high
correlation between Hg recovery and watershed attributes gives promise to this approach when
choosing sediment archives for contaminant trend monitoring. It also has implications for
watershed managers attempting to manage daily loads of contaminants to environmental
systems. Inferring source of Hg using PAHs showed great promise in the elucidation of source
and source areas for these contaminants to the environment. Though several studies have
presented PAHs and heavy metal data, few have attempted to correlate the two measures of

157

environmental state. The powerful potential insight into sources of Hg using diagnostic ratios of
PAHs should not be overlooked in future Hg studies.
Recommendations
Below are suggestions for future work based on the outcome of this study.
1) The statistical outcomes shown in Chapter 3 demonstrate the effectiveness of this
approach; however, ground truth studies would be useful in determining the transport
mechanics in these systems. Highly dynamic flowpaths of a watershed can easily be
determined using GIS techniques. In depth mechanistic studies of Hg release and
transport from these flowpaths should be identified.
2) Apply the lake recovery models to more lakes. Based on the results of Chapter 2
there were a limited number of lakes that could be used to investigate the impact of
watershed attributes on recovery. The correlations should have widespread
applicability and should be tested using a larger sample size of lakes and potentially a
larger range of watershed attributes. A larger sample size may also allow us of more
powerful multivariate statistics.
3) Increase the number of organic tracers measured in the sediment core. The results of
using diagnostic tracers of PAHs was effectively demonstrated, but could be made
even more powerful by measuring those organics that are more discerning of source
including: alkyl PAHs, simonellite and retene. These organics have been used
successfully to delineate combustion and vegetation sources of PAHs to the
environment.
4) Ground truth the relationship between PAHs and Hg, i.e., correlation is not causation.
The high correlation observed between PAHs and Hg combined with knowledge of

158

their geochemical cycles provides confidence in stating that they share common
sources and pathways to the environment. However, a more detailed atmospheric
based study of the relationship between these contaminants and their sources and
pathways would provide further discriminating power to this approach.
5) Extend the use of diagnostic PAHs and recovery models to other metals. Mercury is
not the only metal of concern in the Great Lakes region and with the vast body of data
available the approach taken here for Hg could easily be extended to lead, and to a
lesser extent zinc, cadmium and copper.
6) Return to resample Deer Lake to evaluate the system state and verify predictions
made based on the linear models of recovery.
This research has many practical benefits that may extend beyond the research
community. The use of inland lake sediment chronologies to determine areas of
increased contaminant loading in a region cannot be overstated. They also provide
important information regarding the potential for exposure to recreational users of the
lake. Atmospheric deposition monitors are expensive to maintain and require a long-term
commitment for meaningful data. In addition, they do not incorporate information
concerning the overland transport of contaminants. Sediment chronologies provide a
long term and relatively inexpensive method for determining system state; however,
identifying the source or ratio of sources is difficult at best. The potential of an organic
tracer to help elucidate source is unique, but still may have difficulty determining
proportions of mixed sources. Watershed managers will find useful the contributions of
dynamic flowpaths to Hg recovery since reduction of loads to aquatic systems is the goal
of many of their programs. The costs of remediating contaminated sediment systems can
159

often soar into the hundreds of millions of dollars. A less costly alternative would be
natural recovery if the system can be managed in such a way to the make the strategy
feasible. Here we have demonstrated how sediment chronologies can be used to monitor
such a strategy.
The approach taken here has identified the importance of local sources and
determined the watershed processes that impact Hg accumulation in inland lakes. The
techniques used here can easily be translated to other regions, that is to say, to the extent
that a significant amount of lakes exist in the area to sample.

160

APPENDICIES

161

APPENDIX A

SEDIMENT CHEMICAL DATA

162

APPENDIX A. SEDIMENT CHEMICAL DATA

210

Sediment concentration,
Pb activities and dates, and core desciptions can be found in
Michigan Department of Environmental Quality reports including:
Cass, Gratiot, Gull, Elk and Higgins lakes:
Simpson, S.J., Long, D.T., Geisy, J.P., and Fett, J.D., 2000. Inland lakes sediment trends:
sediment analysis results for five Michigan Lakes (MI/DEQ/SWQ-01/030), Department of
Environmental Quality, Lansing, MI.
http://www.michigan.gov/deq/0,1607,7-135-3313_3686_3728-32365--,00.html
Crystal M and Littlefield lakes:
Yohn, S.S., Long, D.T., Giesy, J.P., Fett, J.D., and Kannan, K., 2001. Inland lakes sediment
trends: sediment analysis results for two Michigan Lakes (MI/DEQ/WD-02/115), Department of
Environment Quality, Lansing, MI.
Cadillac, Crystal B, Mullett, Paw Paw, and Whitmore lakes:
Yohn, S.S., Long, D.T., Giesy, J.P., Scholle, L.K., Patino, L.C., Fett, J.D., and Kannan, K., 2002.
Inland lakes sediment trends: sediment analysis results for five Michigan Lakes (MI/DEQ/WD03/052). Department of Environmental Quality, Lansing, MI.
Houghton, Hubbard, Imp, Round, Torch and Witch lakes:
Yohn, S.S., Parsons, M.J., Long, D.T., Giesy, J.P., Scholle, L.K., and Patino, L.C., 2003. Inland
lakes sediment trends: sediment analysis results for six Michigan lakes (MI/DEQ/WB-04/066).
Department of Environmental Quality, Lansing, MI.
Avalon, Birch, Muskegon, Sand and Shupac lakes:
Parsons, M. J.; Yohn, S. S.; Long, D. T.; Giesy, J. P.; Bradley, P. W., 2004. Inland Lakes
Sediment Trends: Sediment Analysis Results for Five Michigan Lakes (MI/DEQ/WB-06/003).
Michigan Department of Environmental Quality: Lansing, MI.
Crystal M04, George, Hackert, Otter and Round lakes:
Parsons, M. J.; Yohn, S. S.; Long, D. T.; Giesy, J. P.; Patino, L.C., 2005. Inland Lakes Sediment
Trends: Sediment Analysis Results for Five Michigan Lakes (MI/DEQ/WB-07/078). Michigan
Department of Environmental Quality: Lansing, MI.
Mercury analysis for all lakes above:
Parsons, M. J.; Yohn, S. S.; Long, D. T.; Patino, L. C.; Stone, K. A., 2006. Inland Lakes
Sediment Trends: Mercury Sediment Analysis Results for 27 Michigan Lakes (MI/DEQ/WB06/090). Michigan Department of Environmental Quality, Water Bureau: Lansing, MI.
Charlevoix, Cadillac05, Gull05, Nichols, Campau, Muskegon06 and White lakes:

163

Parsons, M. J.; Yohn, S. S.; Long, D. T.; Giesy, J. P.; Patino, L.C., 2005. Inland Lakes Sediment
Trends: Sediment Analysis Results for Seven Michigan Lakes (MI/DEQ/WB-09/069). Michigan
Department of Environmental Quality: Lansing, MI.

164

Table A1.

210

SAMPLE

DEER1-1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

Pb activities for Deer Lake South sediment core.
DRY ACCUMULATED POROSITY Thickness WATER
WT
DRY WT
(g)
(gm./sq.cm.)
(cm)
(%)
0.44
0.0062
1.00
0.44
98.6
0.55
0.0140
0.99
0.32
97.6
0.78
0.0250
0.99
0.40
97.3
1.20
0.0419
0.99
0.52
96.8
1.34
0.0608
0.99
0.50
96.3
1.45
0.0813
0.98
0.46
95.6
1.98
0.1092
0.98
0.45
94.1
1.68
0.1329
0.98
0.40
94.2
2.44
0.1673
0.97
0.46
92.9
2.22
0.1986
0.98
0.44
93.2
2.55
0.2346
0.97
0.45
92.4
2.51
0.2700
0.97
0.42
92.0
2.83
0.3100
0.97
0.45
91.7
3.36
0.3574
0.97
0.51
91.2
3.47
0.4063
0.96
0.45
89.9
4.14
0.4647
0.95
0.42
87.4
8.95
0.5910
0.95
0.93
87.5
9.35
0.7229
0.95
0.91
86.8
8.68
0.8454
0.96
0.95
88.1
7.03
0.9446
0.96
0.93
90.1
9.77
1.0824
0.95
0.95
86.8
3.45
1.1311
0.98
0.92
94.9
2.69
1.1690
0.99
0.96
96.1
3.60
1.2198
0.98
0.91
94.6
3.99
1.2761
0.98
0.98
94.4
165

Excess
Pb-210
(Bq/g)
4.88E-01
5.73E-01
5.77E-01
6.28E-01
6.18E-01
7.13E-01
7.91E-01
8.44E-01
7.12E-01
7.14E-01
6.19E-01
5.54E-01
5.12E-01
4.87E-01
3.25E-01
2.61E-01
2.12E-01
1.80E-01
1.94E-01
1.91E-01
1.49E-01
2.71E-01
2.33E-01
2.45E-01
3.06E-01

ACTIVITY ERROR
Pb-210
(+/- 1 SD)
(Bq/g)
(Bq/g)
5.28E-01
1.64E-02
6.13E-01
1.64E-02
6.17E-01
1.44E-02
6.68E-01
1.22E-02
6.58E-01
1.16E-02
7.53E-01
1.36E-02
8.31E-01
1.41E-02
8.84E-01
1.63E-02
7.52E-01
1.29E-02
7.54E-01
1.47E-02
6.59E-01
1.23E-02
5.94E-01
1.05E-02
5.52E-01
1.06E-02
5.27E-01
9.55E-03
3.65E-01
7.91E-03
3.01E-01
7.25E-03
2.52E-01
6.84E-03
2.20E-01
6.05E-03
2.34E-01
6.46E-03
2.31E-01
5.24E-03
1.89E-01
4.27E-03
3.11E-01
6.72E-03
2.73E-01
5.77E-03
2.85E-01
7.13E-03
3.46E-01
9.38E-03

Table A1 (cont’d)
SAMPLE

26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50

DRY ACCUMULATED POROSITY Thickness WATER
WT
DRY WT
(g)
(gm./sq.cm.)
(cm)
(%)
4.62
1.3413
0.97
0.88
93.0
4.78
1.4087
0.97
0.92
93.0
4.57
1.4732
0.98
0.90
93.1
4.10
1.5310
0.98
0.86
93.6
4.07
1.5885
0.98
0.90
93.9
4.16
1.6472
0.98
0.84
93.3
4.02
1.7039
0.98
0.81
93.3
3.89
1.7587
0.98
0.82
93.6
4.20
1.8180
0.98
0.90
93.7
4.03
1.8749
0.98
0.85
93.6
4.30
1.9355
0.97
0.81
92.9
5.44
2.0123
0.97
0.87
91.6
5.72
2.0930
0.97
0.88
91.4
6.89
2.1902
0.96
0.91
90.0
6.19
2.2775
0.97
0.87
90.6
5.99
2.3620
0.97
0.88
90.9
5.97
2.4462
0.97
0.90
91.2
7.47
2.5516
0.96
0.99
90.0
7.27
2.6542
0.96
0.89
89.3
6.95
2.7523
0.96
0.90
89.8
6.11
2.8385
0.96
0.81
90.1
6.29
2.9272
0.97
0.90
90.7
5.33
3.0024
0.97
0.86
91.7
7.51
3.1084
0.96
0.88
88.9
7.29
3.2112
0.96
0.92
89.6
166

Excess
Pb-210
(Bq/g)
3.16E-01
3.29E-01
3.04E-01
2.99E-01
2.68E-01
2.48E-01
2.10E-01
2.17E-01
2.36E-01
2.25E-01
2.10E-01
1.89E-01
1.67E-01
1.69E-01
1.72E-01
1.64E-01
1.57E-01
1.48E-01
1.40E-01
1.33E-01
1.27E-01
1.21E-01
1.16E-01
1.10E-01
1.04E-01

ACTIVITY ERROR
Pb-210
(+/- 1 SD)
(Bq/g)
(Bq/g)
3.56E-01
1.04E-02
3.69E-01
9.00E-03
3.44E-01
7.92E-03
3.39E-01
8.55E-03
3.08E-01
7.62E-03
2.88E-01
7.22E-03
2.50E-01
8.57E-03
2.57E-01
7.03E-03
2.76E-01
8.34E-03
2.65E-01
7.19E-03
2.50E-01
8.23E-03
2.29E-01
7.74E-03
2.07E-01
6.03E-03
2.09E-01
6.70E-03
2.12E-01
2.04E-01
1.97E-01
1.88E-01
1.80E-01
1.73E-01
1.67E-01
1.61E-01
1.56E-01
1.50E-01
1.44E-01

Table A1 (cont’d)
SAMPLE

51
52
53
54
55
56
57
58
59

DRY ACCUMULATED POROSITY Thickness WATER
WT
DRY WT
(g)
(gm./sq.cm.)
(cm)
(%)
7.71
3.3200
0.96
0.91
88.9
8.32
3.4374
0.96
0.91
88.1
8.55
3.5580
0.95
0.86
87.1
9.32
3.6895
0.95
0.91
86.7
9.65
3.8256
0.95
0.90
86.3
10.03
3.9671
0.95
0.91
85.9
10.22
4.1113
0.94
0.90
85.4
11.43
4.2726
0.94
0.93
84.5
12.37
4.4471
0.94
0.96
83.8

167

Excess
Pb-210
(Bq/g)
9.81E-02
9.21E-02
8.64E-02
8.05E-02
7.48E-02
6.93E-02
6.42E-02
5.42E-02
5.37E-02

ACTIVITY ERROR
Pb-210
(+/- 1 SD)
(Bq/g)
(Bq/g)
1.38E-01
1.32E-01
1.26E-01
1.20E-01
1.15E-01
1.09E-01
1.04E-01
9.42E-02
2.88E-03
9.37E-02
3.37E-03

Table A2.

210

SAMPLE

DEER2-1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

Pb activities for Deer Lake North sediment core.
DRY ACCUMULATED POROSITY Thickness WATER
WT
DRY WT
(g)
(gm./sq.cm.)
(cm)
(%)
1.38
0.0195
0.99
0.60
96.8
1.69
0.0433
0.97
0.31
92.7
2.02
0.0718
0.98
0.40
93.3
2.18
0.1026
0.98
0.44
93.3
2.06
0.1316
0.97
0.40
93.1
2.30
0.1641
0.97
0.42
92.7
2.94
0.2056
0.97
0.56
92.9
1.88
0.2321
0.97
0.33
92.3
2.36
0.2654
0.97
0.37
91.5
1.90
0.2922
0.97
0.31
91.9
2.38
0.3258
0.97
0.35
91.1
2.45
0.3603
0.97
0.43
92.3
2.74
0.3990
0.97
0.47
92.1
2.69
0.4369
0.96
0.35
90.0
2.46
0.4716
0.97
0.34
90.5
2.77
0.5107
0.97
0.42
91.3
5.69
0.5910
0.97
0.86
91.2
5.78
0.6725
0.97
0.84
90.9
6.06
0.7580
0.97
0.87
90.7
5.96
0.8421
0.97
1.02
92.2
7.24
0.9443
0.96
0.88
89.2
7.66
1.0523
0.96
0.94
89.3
7.81
1.1625
0.96
0.97
89.4
7.34
1.2661
0.96
0.94
89.7
6.87
1.3630
0.96
0.89
89.8
168

Excess
Pb-210
(Bq/g)
4.88E-01
5.73E-01
5.77E-01
6.28E-01
6.18E-01
7.13E-01
7.91E-01
8.44E-01
7.12E-01
7.14E-01
6.19E-01
5.54E-01
5.12E-01
4.87E-01
3.25E-01
2.61E-01
2.12E-01
1.80E-01
1.94E-01
1.91E-01
1.49E-01
2.71E-01
2.33E-01
2.45E-01
3.06E-01

ACTIVITY ERROR
Pb-210
(+/- 1 SD)
(Bq/g)
(Bq/g)
6.71E-01
1.38E-02
7.26E-01
1.66E-02
7.70E-01
1.65E-02
8.05E-01
1.65E-02
8.39E-01
1.35E-02
8.82E-01
1.47E-02
8.48E-01
1.48E-02
8.16E-01
1.44E-02
7.88E-01
1.46E-02
7.09E-01
1.51E-02
7.47E-01
1.43E-02
6.95E-01
1.24E-02
6.12E-01
1.17E-02
5.65E-01
1.01E-02
5.15E-01
9.49E-03
4.99E-01
9.20E-03
4.79E-01
1.19E-02
4.36E-01
1.38E-02
4.24E-01
1.15E-02
3.70E-01
9.30E-03
3.14E-01
6.58E-03
2.83E-01
6.45E-03
2.89E-01
6.64E-03
2.83E-01
6.59E-03
2.87E-01
5.87E-03

Table A2. (cont’d)
SAMPLE

26
27
28
29
30
31
32
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50

DRY ACCUMULATED POROSITY Thickness WATER
WT
DRY WT
(g)
(gm./sq.cm.)
(cm)
(%)
6.92
1.4606
0.96
0.88
89.7
7.17
1.5618
0.97
1.01
90.6
7.19
1.6632
0.96
0.94
89.9
7.55
1.7698
0.96
0.92
89.3
6.75
1.8650
0.97
0.97
90.8
5.68
1.9451
0.97
0.90
91.6
5.98
2.0295
0.97
0.97
91.8
5.53
2.1075
0.97
0.94
92.1
5.45
2.1844
0.97
0.94
92.3
5.60
2.2634
0.97
0.94
92.0
5.84
2.3458
0.97
0.88
91.2
6.73
2.4407
0.97
0.99
91.0
6.47
2.5320
0.97
0.94
90.8
6.37
2.6219
0.97
0.90
90.6
7.38
2.7260
0.96
1.00
90.2
6.42
2.8166
0.96
0.89
90.4
6.99
2.9152
0.96
0.96
90.4
6.82
3.0114
0.96
0.94
90.4
6.84
3.1079
0.97
0.96
90.5
5.79
3.1896
0.97
0.83
90.8
6.97
3.2880
0.97
0.98
90.6
6.35
3.3775
0.97
0.89
90.5
6.53
3.4697
0.96
0.90
90.4
6.82
3.5659
0.96
0.91
90.2

169

Excess
Pb-210
(Bq/g)
3.16E-01
3.29E-01
3.04E-01
2.99E-01
2.68E-01
2.48E-01
2.10E-01
2.17E-01
2.36E-01
2.25E-01
2.10E-01
1.89E-01
1.67E-01
1.69E-01
1.72E-01
1.64E-01
1.57E-01
1.48E-01
1.40E-01
1.33E-01
1.27E-01
1.21E-01
1.16E-01
1.10E-01

ACTIVITY ERROR
Pb-210
(+/- 1 SD)
(Bq/g)
(Bq/g)
2.78E-01
6.02E-03
2.88E-01
6.92E-03
2.79E-01
6.23E-03
3.00E-01
6.42E-03
3.58E-01
7.52E-03
3.58E-01
7.37E-03
3.50E-01
7.36E-03
3.39E-01
7.02E-03
3.21E-01
7.24E-03
3.04E-01
6.40E-03
2.89E-01
6.13E-03
2.93E-01
5.15E-03
2.81E-01
5.14E-03
2.64E-01
5.11E-03
2.50E-01
4.82E-03
2.43E-01
7.45E-03
2.15E-01
6.63E-03
1.81E-01
6.62E-03
1.96E-01
6.63E-03
1.78E-01
5.71E-03
1.62E-01
5.56E-03
1.77E-01
6.12E-03
1.71E-01
5.67E-03
1.77E-01
4.03E-03

Table A2. (cont’d)
SAMPLE

51
52
53
54
55
56
57
58
59

DRY ACCUMULATED POROSITY Thickness WATER
WT
DRY WT
(g)
(gm./sq.cm.)
(cm)
(%)
6.53
3.6580
0.96
0.88
90.2
7.10
3.7582
0.96
0.90
89.6
8.14
3.8730
0.96
0.95
88.8
8.63
3.9948
0.95
0.87
87.1
9.17
4.1242
0.95
0.93
87.3
9.23
4.2544
0.95
0.91
86.9
9.33
4.3860
0.95
0.93
87.0
9.69
4.5227
0.95
0.93
86.6
9.25
4.6532
0.95
0.92
87.0

170

Excess
Pb-210
(Bq/g)
9.81E-02
9.21E-02
8.64E-02
8.05E-02
7.48E-02
6.93E-02
6.42E-02
5.42E-02
5.37E-02

ACTIVITY ERROR
Pb-210
(+/- 1 SD)
(Bq/g)
(Bq/g)
1.94E-01
4.22E-03
1.73E-01
4.43E-03
1.52E-01
3.67E-03
1.42E-01
3.33E-03
1.41E-01
3.68E-03
1.38E-01
3.45E-03
1.27E-01
3.12E-03
1.09E-01
3.91E-03
1.10E-01
3.55E-03

Table A3. Metal concentrations for Deer Lake south core (mg/kg).
Sample
Deer 1-1
Deer 2-1
Deer 3-1
Deer 4-1
Deer 5-1
Deer 6-1
Deer 7-1
Deer 8-1
Deer 9-1
Deer 10-1
Deer 11-1
Deer 12-1
Deer 13-1
Deer 14-1
Deer 15-1
Deer 16-1
Deer 17-1
Deer 18-1
Deer 19-1
Deer 20-1
Deer 21-1
Deer 22-1
Deer 23-1
Deer 24-1
Deer 25-1

Depth
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
9.0
10.0
11.0
12.0
13.0
14.0
15.0
16.0
17.0

Sc
2.19
2.21
2.10
2.26
2.57
2.43
2.93
3.36
3.23
3.13
3.18
3.01
3.14
3.07
3.94
3.22
3.45
3.72
3.25
2.65
2.71
2.21
2.18
2.09
2.41

Ti
161.03
165.55
131.55
156.33
51.43
33.28
41.14
37.85
26.30
24.92
47.08
38.57
28.35
29.06
55.46
226.63
235.10
251.76
184.31
158.05
170.00
169.31
131.25
130.39
141.63

V
27.71
30.05
27.36
26.65
30.89
30.66
35.42
39.11
37.14
35.93
36.96
36.11
37.45
35.24
41.69
32.99
34.85
36.30
34.15
28.69
28.47
21.37
20.96
21.09
23.44

Cr
36.96
45.22
42.00
40.68
47.88
45.53
53.98
64.73
64.06
60.41
62.73
56.76
60.73
58.71
74.88
60.56
68.37
71.15
72.99
49.88
55.06
42.90
43.16
46.33
57.03

Co
7.04
6.90
7.92
8.04
8.50
9.59
9.86
11.48
11.92
10.43
9.67
9.75
10.20
11.10
12.29
9.20
9.64
9.83
8.92
6.94
6.55
5.36
4.82
5.26
5.72

171

Ni
29.11
27.21
29.52
32.02
32.23
33.48
36.06
43.39
44.40
41.48
36.88
31.43
32.30
32.99
41.82
30.73
32.90
35.07
31.28
24.67
24.67
21.85
21.68
20.88
21.27

Cu
151
155
110
76
68
54
74
79
59
64
66
220
48
76
78
87
117
104
86
79
161
77
103
114
156

As
31.85
32.45
34.52
35.03
33.06
39.92
41.61
33.54
35.94
37.60
38.82
36.64
36.11
43.29
40.51
28.43
27.10
26.92
24.25
16.68
21.14
21.76
23.70
24.74
27.96

Mo
1.66
1.54
2.08
2.04
1.90
2.16
1.81
2.14
2.97
2.99
2.94
2.89
2.83
3.64
3.17
2.02
1.98
1.76
1.96
1.74
2.57
3.70
3.95
3.76
4.60

Cd
1.03
1.03
0.95
1.00
1.01
0.93
1.06
1.37
1.24
1.26
1.51
1.43
1.77
2.11
2.61
2.35
2.43
2.53
2.14
1.65
1.86
1.32
1.58
1.65
2.30

Pb
46.67
47.64
45.77
45.77
50.89
46.02
49.30
57.68
62.95
63.52
80.29
84.26
96.58
97.51
123.88
103.11
106.32
131.30
123.37
96.34
103.96
79.96
83.08
91.66
117.50

Table A3. (cont’d)
Sample
Deer 26-1
Deer 27-1
Deer 28-1
Deer 29-1
Deer 30-1
Deer 31-1
Deer 32-1
Deer 33-1
Deer 34-1
Deer 35-1
Deer 36-1
Deer 37-1
Deer 38-1
Deer 39-1
Deer 40-1
Deer 41-1
Deer 42-1
Deer 43-1
Deer 44-1
Deer 45-1
Deer 46-1
Deer 47-1
Deer 48-1
Deer 49-1
Deer 50-1

Depth
18.0
19.0
20.0
21.0
22.0
23.0
24.0
25.0
26.0
27.0
28.0
29.0
30.0
31.0
32.0
33.0
34.0
35.0
36.0
37.0
38.0
39.0
40.0
41.0
42.0

Sc
2.37
2.29
2.56
2.46
2.38
2.54
2.28
2.56
2.22
2.63
2.48
2.53
2.52
2.76
2.54
2.70
2.39
3.04
2.91
2.77
2.78
3.21
3.20
3.10
3.37

Ti
138.38
141.41
165.85
55.95
213.98
209.94
190.86
64.13
161.77
182.13
163.73
156.32
155.26
194.58
70.51
39.10
55.34
81.90
40.39
100.22
199.46
68.08
85.72
95.19
173.52

V
23.83
22.29
22.79
23.36
23.16
25.48
22.46
24.78
21.18
25.76
25.42
26.26
24.85
27.00
25.42
28.62
25.84
30.73
29.44
28.02
28.70
32.44
32.77
32.70
35.78

Cr
54.50
54.55
56.20
55.44
53.03
57.62
61.65
76.24
70.55
121.67
119.79
101.81
79.73
98.35
70.81
54.49
50.27
54.15
50.27
47.76
48.83
57.67
57.97
52.21
53.96

Co
5.88
5.94
6.28
6.26
6.04
5.82
5.01
7.70
5.49
7.31
8.23
8.56
8.19
8.31
8.56
9.71
8.77
10.00
9.71
9.44
9.54
11.28
11.37
11.23
11.89

172

Ni
21.54
20.89
22.54
24.47
21.98
24.14
19.67
26.65
19.56
24.50
23.06
23.91
21.15
23.42
17.86
19.64
15.93
30.49
28.89
27.62
29.75
33.74
33.61
32.43
35.36

Cu
149
120
126
38
157
199
167
45
108
181
168
155
141
112
90
99
103
207
170
120
158
103
98
103
116

As
25.62
22.86
23.51
17.17
23.62
25.08
23.54
20.98
22.26
26.47
22.58
23.25
20.02
18.67
19.83
18.89
17.65
24.92
23.53
22.30
20.86
20.04
20.57
22.56
23.35

Mo
4.20
4.29
6.47
7.15
4.11
3.88
3.63
4.96
2.72
3.82
3.02
3.50
3.77
3.57
4.09
2.62
2.47
2.77
3.01
3.35
2.74
4.27
6.19
5.83
4.66

Cd
2.33
2.21
2.60
2.61
2.36
2.48
2.48
2.72
2.36
3.03
2.60
2.80
2.74
2.53
2.71
2.43
2.24
2.15
2.06
2.00
2.26
2.14
2.23
2.09
2.03

Pb
129.76
133.00
148.91
166.19
158.17
180.90
172.28
189.76
158.63
176.04
155.10
167.28
160.25
172.11
229.96
200.27
183.45
177.87
178.14
159.60
155.48
174.71
164.61
146.60
131.89

Table A3. (cont’d)
Sample
Deer 51-1
Deer 52-1
Deer 53-1
Deer 54-1
Deer 55-1
Deer 56-1

Depth
43.0
44.0
45.0
46.0
47.0
48.0

Sc
3.28
3.22
3.37
3.56
3.28
3.33

Ti
181.19
199.47
200.77
242.82
246.18
215.27

V
34.43
34.84
35.68
38.51
35.80
35.55

Cr
52.06
51.35
50.47
52.82
46.53
46.81

Co
11.82
11.41
11.87
12.49
11.60
11.47

173

Ni
35.32
33.55
34.20
36.48
33.56
49.88

Cu
90
104
116
134
103
100

As
25.67
24.57
25.03
26.58
25.97
25.30

Mo
4.15
3.53
2.82
2.97
2.81
2.48

Cd
2.15
2.12
2.13
2.27
2.29
1.99

Pb
126.03
120.12
114.52
118.71
111.10
107.23

Table A3. (cont’d)
Sample
Deer 1-1
Deer 2-1
Deer 3-1
Deer 4-1
Deer 5-1
Deer 6-1
Deer 7-1
Deer 8-1
Deer 9-1
Deer 10-1
Deer 11-1
Deer 12-1
Deer 13-1
Deer 14-1
Deer 15-1
Deer 16-1
Deer 17-1
Deer 18-1
Deer 19-1
Deer 20-1
Deer 21-1
Deer 22-1
Deer 23-1
Deer 24-1
Deer 25-1

Depth
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
9.0
10.0
11.0
12.0
13.0
14.0
15.0
16.0
17.0

Al
6686
6828
6357
6873
7838
7652
8690
10612
10124
9651
10553
10921
10957
11140
14544
12392
13376
18516
12592
10142
11169
7797
7388
7582
8007

Zn
172
168
160
175
197
187
218
251
251
240
282
267
319
406
471
342
358
370
323
237
249
184
232
233
290

Se
3.08
3.25
2.98
3.15
3.26
3.17
3.17
3.07
3.23
3.07
3.27
3.20
3.03
3.35
3.53
2.70
2.81
2.96
2.99
1.99
3.31
3.30
3.83
3.86
4.24

Sr
40.18
41.20
36.47
33.79
35.69
34.23
34.26
36.08
37.95
38.01
39.04
39.57
40.26
43.66
53.62
43.10
49.11
48.36
48.38
40.34
61.65
75.52
63.37
64.36
65.81

Mg
3344
3280
3165
3350
3839
3598
4396
5433
4903
4704
4555
4238
4701
4735
6421
5616
5997
6102
5787
5480
4968
3762
3212
3110
3339

K
1065
857
837
845
856
717
765
931
929
955
861
763
871
824
1104
796
845
968
711
567
606
581
490
469
499

174

Mn
3098
3603
3110
2570
2601
2728
3187
2994
2840
2633
2451
2320
2435
2233
2017
1386
1338
1640
1454
1228
1700
1091
1307
1135
1426

Ba
337
415
297
225
152
85
136
121
93
87
136
129
97
98
170
210
281
261
244
223
358
280
301
348
411

U
2.23
2.28
2.18
2.14
2.36
2.39
2.75
2.94
2.61
2.54
2.71
2.56
2.73
2.70
2.70
2.13
2.23
2.16
2.34
2.22
2.20
1.79
2.43
2.81
3.55

Fe
554
601
547
533
618
613
708
782
743
719
739
722
749
705
834
660
697
726
683
574
569
427
419
422
469

Ca
185
226
210
203
239
228
270
324
320
302
314
284
304
294
374
303
342
356
365
249
275
215
216
232
285

Hg
2.05
2.28
2.18
1.90
1.96
1.88
1.75
1.84
2.26
2.38
4.20
4.86
5.60
6.52
8.09
7.87
8.02
10.09
9.26
5.77
6.90
5.52
4.93
8.03
15.78

Table A3. (cont’d)
Sample
Deer 26-1
Deer 27-1
Deer 28-1
Deer 29-1
Deer 30-1
Deer 31-1
Deer 32-1
Deer 33-1
Deer 34-1
Deer 35-1
Deer 36-1
Deer 37-1
Deer 38-1
Deer 39-1
Deer 40-1
Deer 41-1
Deer 42-1
Deer 43-1
Deer 44-1
Deer 45-1
Deer 46-1
Deer 47-1
Deer 48-1
Deer 49-1
Deer 50-1

Depth
18.0
19.0
20.0
21.0
22.0
23.0
24.0
25.0
26.0
27.0
28.0
29.0
30.0
31.0
32.0
33.0
34.0
35.0
36.0
37.0
38.0
39.0
40.0
41.0
42.0

Al
8247
7536
8105
8308
7942
8728
8014
9011
7300
8946
9277
9175
8955
9678
9087
10406
9403
11060
10145
10177
10561
12086
12329
11658
12911

Zn
300
267
317
405
310
334
331
427
299
447
415
376
331
394
383
349
320
348
358
359
437
428
435
407
377

Se
4.42
4.27
5.39
4.98
5.79
3.56
3.12
2.85
3.00
3.76
3.65
4.49
4.19
4.13
5.55
4.40
4.35
4.39
4.10
4.77
3.77
4.03
4.51
3.62
3.79

Sr
66.51
65.73
72.31
64.32
70.21
77.84
77.43
81.61
67.72
75.14
70.23
73.77
67.90
65.15
71.08
68.60
63.35
73.89
78.83
74.91
76.23
75.71
75.45
72.58
71.36

Mg
3376
3276
3712
3622
3419
3930
3433
4178
3549
4623
4765
4872
5166
6106
4771
4899
4438
5064
5040
5203
5501
6583
6632
6597
7434

K
501
481
543
712
632
663
599
686
494
599
594
578
628
759
676
775
669
1394
1259
1256
1238
1312
1373
1280
1451

175

Mn
1185
781
791
798
857
1040
971
1002
1077
1563
2212
2246
1592
1352
1945
2050
1880
2423
2642
1868
1916
1892
1655
1866
1822

Ba
416
373
380
50
395
472
453
64
349
398
405
409
380
351
300
252
271
319
308
274
305
179
118
127
197

U
3.51
3.12
3.45
4.67
3.98
3.90
3.73
4.13
4.24
4.36
4.33
4.63
4.45
4.33
5.19
4.82
4.39
5.69
5.51
4.69
4.29
4.58
4.55
4.33
3.96

Fe
477
446
456
467
463
510
449
496
424
515
508
525
497
540
508
572
517
615
589
560
574
649
655
654
716

Ca
273
273
281
277
265
288
308
381
353
608
599
509
399
492
354
272
251
271
251
239
244
288
290
261
270

Hg
25.50
24.89
25.35
24.34
11.87
11.86
13.58
15.96
11.62
13.23
11.40
13.13
12.56
10.30
13.33
12.66
10.98
9.60
11.92
14.09
14.00
17.31
18.71
19.72
18.62

Table A3. (cont’d)
Sample
Deer 51-1
Deer 52-1
Deer 53-1
Deer 54-1
Deer 55-1
Deer 56-1

Depth
43.0
44.0
45.0
46.0
47.0
48.0

Al
12833
12461
13160
13791
12928
13024

Zn
382
349
347
364
393
384

Se
3.28
3.20
3.18
3.19
2.81
2.62

Sr
75.22
71.90
70.08
72.72
63.74
62.68

Mg
6763
6152
6307
6870
6339
5804

K
1289
1239
1294
1311
1168
1151

176

Mn
1687
1774
1985
1994
1883
1721

Ba
224
221
232
233
202
183

U
3.79
3.29
3.00
2.74
2.35
2.27

Fe
689
697
714
770
716
711

Ca
260
257
252
264
233
234

Hg
19.21
18.31
17.63
16.73
15.39
15.25

Table A4. Metal concentrations for Deer Lake north core (mg/kg).
Sample
Deer 1-2
Deer 2-2
Deer 3-2
Deer 4-2
Deer 5-2
Deer 6-2
Deer 7-2
Deer 8-2
Deer 9-2
Deer 10-2
Deer 11-2
Deer 12-2
Deer 13-2
Deer 14-2
Deer 15-2
Deer 16-2
Deer 17-2
Deer 18-2
Deer 19-2
Deer 20-2
Deer 21-2
Deer 22-2
Deer 23-2
Deer 24-2
Deer 25-2

Depth
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
9.0
10.0
11.0
12.0
13.0
14.0
15.0
16.0
17.0

Sc
2.68
2.55
2.34
2.29
2.43
2.51
2.58
2.54
2.50
2.54
2.58
2.49
2.52
2.41
2.48
2.56
2.52
2.46
2.56
2.63
2.73
2.73
2.70
2.51
3.16

Ti
232.1
230.9
223.6
205.4
203.9
226.7
229.6
210.7
217.0
207.7
218.0
198.4
222.2
195.9
200.1
217.0
196.7
206.3
228.5
230.4
257.5
259.2
252.0
214.3
254.3

V
35.57
34.81
32.03
31.74
33.37
34.38
34.16
32.58
32.37
32.72
31.77
30.92
31.31
30.22
31.01
30.85
30.45
30.20
31.17
30.84
32.12
31.40
31.95
29.64
36.70

Cr
62.07
59.88
54.13
54.68
55.67
59.59
58.64
56.06
58.38
61.47
55.24
58.41
57.69
53.74
55.13
56.33
55.54
55.58
54.34
57.17
60.00
60.36
62.74
57.92
73.96

Co
14.55
14.44
13.22
12.23
13.30
13.88
14.40
13.32
13.85
15.89
15.78
16.71
16.80
17.07
18.63
18.69
19.92
19.84
20.72
19.62
18.23
14.19
12.15
10.71
12.60

177

Ni
68.27
65.01
59.78
59.78
63.65
65.17
66.84
61.39
64.24
68.38
67.78
64.73
64.81
62.48
66.48
69.56
66.25
69.65
63.65
65.30
63.87
59.41
59.46
56.88
71.37

Cu
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
0.00
0.00
0.00

As
23.52
24.11
20.33
17.85
18.27
18.76
20.14
19.46
20.46
24.21
24.29
24.96
26.00
26.87
29.17
29.98
30.79
30.39
31.74
30.36
29.71
26.34
23.14
21.78
24.87

Mo
2.52
2.42
2.04
1.84
2.24
2.32
2.54
2.25
2.62
3.15
3.31
3.31
3.52
3.59
3.85
3.94
3.92
3.95
3.95
3.96
4.06
4.23
4.35
4.09
5.29

Cd
1.73
1.73
1.56
1.55
1.49
1.54
1.68
1.59
1.71
1.82
1.85
1.98
2.11
2.09
2.24
2.23
2.34
2.28
2.30
2.27
2.13
1.99
1.77
1.70
2.15

Pb
58.93
55.34
51.13
51.21
52.30
54.32
56.17
55.75
58.34
63.68
66.12
68.06
70.07
70.26
75.32
75.67
75.57
74.62
75.64
77.17
78.87
80.55
81.77
81.01
104.27

Table A4. (cont’d)
Sample
Deer 26-2
Deer 27-2
Deer 28-2
Deer 29-2
Deer 30-2
Deer 31-2
Deer 32-2
Deer 33-2
Deer 34-2
Deer 35-2
Deer 36-2
Deer 37-2
Deer 38-2
Deer 39-2
Deer 40-2
Deer 41-2
Deer 42-2
Deer 43-2
Deer 44-2
Deer 45-2
Deer 46-2
Deer 47-2
Deer 48-2
Deer 49-2
Deer 50-2

Depth
18.0
19.0
20.0
21.0
22.0
23.0
24.0
25.0
26.0
27.0
28.0
29.0
30.0
31.0
32.0
33.0
34.0
35.0
36.0
37.0
38.0
39.0
40.0
41.0
42.0

Sc
2.62
2.83
2.41
2.73
2.72
2.67
2.55
2.48
2.50
2.46
2.46
2.31
2.22
2.22
2.26
2.47
2.44
2.27
2.21
2.27
2.22
2.35
2.45
2.55
2.30

Ti
229.2
232.1
203.6
238.1
246.5
274.0
245.6
228.8
249.6
246.2
261.6
136.6
222.4
237.8
234.0
257.8
247.9
169.7
167.1
177.6
175.7
208.7
212.0
239.5
192.1

V
31.18
31.16
29.84
30.99
30.33
30.92
30.22
27.96
31.72
32.08
29.84
26.13
25.95
26.94
29.48
32.41
31.63
29.22
28.18
27.37
26.97
28.52
30.70
31.33
28.19

Cr
58.76
62.25
53.27
64.67
59.64
56.25
54.38
58.60
52.84
55.07
54.15
46.20
47.72
50.73
54.83
59.35
48.65
50.20
41.02
44.03
43.66
42.35
44.05
48.84
42.30

Co
10.51
11.09
10.73
9.34
9.12
9.12
8.64
8.19
8.75
10.69
7.90
6.39
5.58
5.47
6.07
6.73
6.29
6.40
5.90
5.60
5.49
5.26
5.95
6.16
5.49

178

Ni
58.54
65.35
59.45
61.72
61.05
61.93
59.96
57.50
60.45
62.50
57.86
53.37
54.44
49.86
50.02
52.22
50.97
50.60
47.40
47.50
47.56
47.59
48.92
50.81
46.36

Cu
0.00
74.98
45.18
56.71
53.71
55.08
40.88
51.58
52.59
57.37
50.36
43.04
51.39
56.27
52.34
46.92
53.61
40.76
53.20
41.51
53.50
48.23
44.28
58.73
45.43

As
19.96
21.98
22.45
20.60
19.94
19.64
19.63
17.63
19.36
20.99
18.32
18.63
18.99
19.96
20.72
22.25
19.92
18.25
16.70
15.61
14.55
16.99
16.51
16.87
15.76

Mo
4.38
4.21
3.88
4.42
4.28
4.08
4.09
4.13
3.91
4.37
3.77
4.02
3.91
3.69
3.41
3.59
3.49
3.38
3.59
3.40
2.97
3.22
2.79
2.93
2.84

Cd
1.83
1.86
1.67
1.72
1.63
1.61
1.47
1.50
1.57
1.52
1.44
1.17
1.15
1.02
1.04
1.02
0.93
1.05
1.29
0.98
0.91
0.94
0.95
0.95
0.82

Pb
86.98
86.74
74.76
88.26
87.42
83.09
78.09
79.42
77.86
78.46
76.29
79.96
79.98
70.18
64.11
66.33
59.98
62.94
61.91
59.66
57.40
64.10
64.23
65.17
52.70

Table A4. (cont’d)
Sample
Deer 51-2
Deer 52-2
Deer 53-2
Deer 54-2
Deer 55-2

Depth
43.0
44.0
45.0
46.0
47.0

Sc
2.50
2.49
2.43
2.36
3.11

Ti
247.4
208.3
211.3
233.7
275.9

V
30.29
28.97
31.38
32.01
32.81

Cr
44.00
43.82
46.80
43.87
42.29

Co
5.55
5.36
5.83
6.02
6.19

179

Ni
47.27
49.03
47.14
46.99
48.38

Cu
47.21
61.44
41.94
43.89
57.40

As
15.82
15.38
17.18
17.85
17.38

Mo
3.24
3.06
3.23
3.22
3.18

Cd
0.88
0.84
1.03
1.00
1.00

Pb
51.86
55.21
53.51
51.13
51.64

Table A4. (cont’d)
Sample
Deer 1-2
Deer 2-2
Deer 3-2
Deer 4-2
Deer 5-2
Deer 6-2
Deer 7-2
Deer 8-2
Deer 9-2
Deer 10-2
Deer 11-2
Deer 12-2
Deer 13-2
Deer 14-2
Deer 15-2
Deer 16-2
Deer 17-2
Deer 18-2
Deer 19-2
Deer 20-2
Deer 21-2
Deer 22-2
Deer 23-2
Deer 24-2
Deer 25-2

Depth
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
9.0
10.0
11.0
12.0
13.0
14.0
15.0
16.0
17.0

Al
11616
10966
10289
10144
10312
12743
11153
10528
11436
11158
11587
12838
11876
15683
11934
12082
12084
12053
12607
12387
12825
12393
12114
11260
14167

Zn
272
255
233
232
238
250
260
249
263
291
304
319
336
339
378
376
407
407
400
393
357
301
263
240
289

Se
4.29
4.32
4.11
4.31
4.93
4.65
4.69
4.71
4.78
5.03
4.52
5.02
4.39
4.91
5.22
4.79
5.03
4.70
4.62
4.54
4.65
4.33
4.69
4.76
6.22

Sr
42.09
40.02
37.86
36.97
37.29
36.39
35.90
33.76
35.11
36.14
35.74
35.59
36.48
36.07
38.96
39.19
40.85
48.31
43.34
46.06
46.52
45.03
45.61
45.30
57.77

Mg
4653
4171
3879
3910
3947
4091
4217
4010
4139
4272
4388
4374
4381
4231
4447
4431
4439
4406
4523
4593
4775
4870
4928
4906
6141

K
924
903
919
868
864
879
884
807
900
857
910
867
927
853
908
945
893
930
977
1020
1106
1090
1080
944
1122

180

Mn
1491
1604
1572
1334
918
789
853
948
981
822
781
817
803
801
868
947
1107
1346
1303
1250
1042
688
561
528
623

Ba
110.42
118.66
120.17
110.86
103.90
96.24
91.05
79.09
76.07
74.23
78.72
78.75
81.44
84.78
84.37
84.25
85.86
95.01
89.50
92.87
94.77
95.64
93.90
90.67
112.70

U
3.62
3.69
3.33
3.27
3.71
3.92
3.96
3.52
3.68
3.89
3.56
3.61
3.61
3.54
3.56
3.60
3.63
3.63
3.62
3.67
3.75
3.97
4.12
4.13
5.37

Fe
67515
69470
64758
63405
63028
63444
67870
70258
68015
58787
53258
54471
54852
54593
56149
57615
56271
56090
54879
52231
52154
49981
47968
45494
56539

Ca
16879
17367
16190
15851
15757
15861
16968
17565
17004
14697
13315
13618
13713
13648
14037
14404
14068
14022
13720
13058
13038
12495
11992
11374
14135

Hg
4.39
4.28
3.94
4.23
4.29
4.36
4.67
4.49
4.78
5.24
5.40
5.94
6.00
6.22
6.58
6.52
6.57
6.80
6.88
7.13
7.38
7.74
8.19
8.52
10.78

Table A4. (cont’d)
Sample
Deer 26-2
Deer 27-2
Deer 28-2
Deer 29-2
Deer 30-2
Deer 31-2
Deer 32-2
Deer 33-2
Deer 34-2
Deer 35-2
Deer 36-2
Deer 37-2
Deer 38-2
Deer 39-2
Deer 40-2
Deer 41-2
Deer 42-2
Deer 43-2
Deer 44-2
Deer 45-2
Deer 46-2
Deer 47-2
Deer 48-2
Deer 49-2
Deer 50-2

Depth
18.0
19.0
20.0
21.0
22.0
23.0
24.0
25.0
26.0
27.0
28.0
29.0
30.0
31.0
32.0
33.0
34.0
35.0
36.0
37.0
38.0
39.0
40.0
41.0
42.0

Al
12997
12032
11936
11589
11679
11718
11145
10660
10840
10806
10477
9571
9847
9759
10253
11398
11155
10101
10311
10186
10234
11540
11265
12026
11116

Zn
240
260
222
227
218
202
183
188
195
195
176
168
141
122
123
128
107
108
104
97
95
101
105
108
95

Se
5.21
5.35
5.20
4.98
4.84
4.88
5.05
4.81
4.84
5.03
4.85
4.39
4.31
4.12
4.55
4.98
4.80
5.19
5.25
5.22
4.74
4.76
4.42
4.54
4.16

Sr
48.11
50.23
55.10
49.46
48.41
47.89
49.12
46.69
58.02
55.10
53.21
68.31
47.86
50.16
52.63
58.29
58.25
59.88
60.70
60.64
60.69
62.74
65.83
68.88
67.90

Mg
5141
5077
4205
4969
4850
4639
4269
4303
4501
4261
4099
3560
3160
2992
3201
3554
3470
3481
3479
3535
3528
3799
3820
3965
3724

181

K
1023
801
650
835
856
938
839
799
834
809
859
777
741
773
799
899
875
653
657
697
704
825
825
1708
817

Mn
550
644
616
585
580
512
460
486
520
465
474
580
325
327
343
383
406
391
353
344
338
347
351
362
350

Ba
95.05
97.46
88.73
89.61
86.12
79.76
73.03
75.62
84.99
76.70
79.92
83.78
76.31
77.31
80.26
84.16
78.53
79.18
77.52
77.54
78.31
83.06
86.13
90.08
85.04

U
4.38
4.34
4.00
4.53
4.53
4.32
4.20
4.31
4.25
4.56
4.15
4.02
4.07
4.22
4.21
4.39
4.23
4.27
4.02
3.93
3.82
4.35
3.99
3.96
3.48

Fe
47179
48450
44824
47984
48069
49482
46837
42423
41890
49006
45230
40630
43492
47020
42164
44196
40806
40040
39057
37391
33898
37148
36475
37529
35720

Ca
11795
12112
11206
11996
12017
12371
11709
10606
10473
12251
11307
10157
6524
7053
6325
6629
6121
6006
5859
5609
5085
5572
5471
5629
5358

Hg
9.01
9.05
7.31
8.78
8.43
7.81
7.17
7.31
7.54
7.20
5.57
4.81
4.42
3.67
3.05
3.27
2.90
2.90
2.85
2.80
2.80
2.93
3.11
3.11
3.13

Table A4. (cont’d)
Sample
Deer 51-2
Deer 52-2
Deer 53-2
Deer 54-2
Deer 55-2

Depth
43.0
44.0
45.0
46.0
47.0

Al
12454
12483
11865
11678
12243

Zn
96
92
102
100
103

Se
4.24
4.20
4.13
3.81
3.93

Sr
72.11
74.98
74.07
73.24
73.20

Mg
4011
4096
3667
3483
3636

K
997
904
851
876
1024

182

Mn
357
359
345
337
342

Ba
91.11
97.53
93.03
90.86
93.36

U
3.48
3.27
3.48
3.38
3.24

Fe
38789
38099
39028
40287
38359

Ca
5818
5715
5854
6043
5754

Hg
3.20
3.25
3.70
3.75
3.81

APPENDIX B
SUPPORTING INFORMATION FOR CHAPTER 2

183

APPENDIX B. SUPPORTING INFORMATION FOR CHAPTER 2

Hackert

Crystal M

Littlefield

Cadillac

2005
2000
2000
1985
1970
1955
1940
1925
1910
1895
1880
1865
1850

0

20

40

0

20

40

0

20

40

0

50

100

2

FCAAR ( µg/m /yr)

210

Figure S1.
Pb dated focusing-corrected anthropogenic accumulation rates for four
Michigan lakes, dates older than 1850 were truncated to highlight more recent loading.
Note the change in scale on the x-axis.

184

Avalon

Shupac

George

2005
2000
2000
1985
1970
1955
1940
1925
1910
1895
1880
1865
1850

0

5

10

0

20

40

0

20

40

2

FCAAR ( µg/m /yr)

210

Figure S2.
Pb dated focusing-corrected anthropogenic accumulation rates for three
Michigan lakes, dates older than 1850 were truncated to highlight more recent loading.
Note the change in scale on the x-axis.
Dating Methodology
Methods for determining appropriate sediment ages for these cores were adapted from a
decade of work in the Great Lakes region (Golden et al., 1993, Kolak et al., 1998, Kolak et
al., 1999, Pearson et al., 1998, Pearson et al., 1997a, Pearson et al., 1997b, Simick et al.,
1996, Wong et al., 1995). Four different

210

Pb dating models were compared to determine

accurate dates for sediment cores collected in this study (Table S1). The constant flux:
constant sedimentation (CF:CS) model assumes a constant flux of

210

Pb to the sediment and

a constant rate of sedimentation (Golden et al., 1993). However, if clear sedimentation rate
changes occurred within the core indicated by changes in slope of the accumulated dry mass
versus excess

210

Pb plot then a segmented CF:CS (SCF:CS) model was used (Heyvaert et

185

al., 2000). If a mixing zone was present, appearing as a vertical straight line in the
accumulated dry mass versus excess

210

Pb plot, then the rapid steady state mixing model

(RSSM) (Robbins, 1982) may be used. When excess

210

Pb inventories could be calculated

the constant rate of supply model (CRS) was also evaluated (Oldfield et al., 1978). The
most appropriate dating model was chosen based on three separate qualifiers: date of peak
137

Cs activity (ca. 1963) (Walling and Qingping, 1992), age of oldest sediment containing

excess

210

Pb (ca. 1850), and stable Pb peak concentration (ca. 1970) (Alfaro-De la Torre

and Tessier, 2002) (Table S1). Maximum stable Pb concentration not placed in the mid1970s does not conclude an incorrectly dated sediment core if significant local watershed
sources exist. However, if the maximum stable Pb concentration does occur in the mid1970s then the age dating model was appropriate. Only core sections containing excess
210

Pb can be dated, therefore core sections without excess

210

Pb were dated by assuming

that sedimentation rates remained constant below this depth. For the CF:CS, SCF:CS and
RSSM model the sedimentation rate for the lowest excess

210

used. For the CRS model the average of the last five excess

Pb-containing sections was

210

Pb-containing core sections

was used.
Focusing factors were calculated for all lakes. When
inventories of
210

210

210

Pb was present in all core sections

Pb were estimated by extrapolating the regression equation of excess
2

Pb vs. accumulated dry mass (g/cm ). For lakes dated with the CF:CS model the entire

core section was used. Only the bottom core sections were used to determine the equation
186

when the SCF:CS was applied. Dry mass was estimated as the average dry mass for the
bottom two to twelve core sections. The final focusing factor was deemed appropriate when
additional core sections did not result in an appreciable change.

Dating in Littlefield Lake was difficult due to a disturbed

210

Pb profile; however, stable Pb

concentrations appeared to reach background concentrations when compared to other study
lakes. In order to date this sediment core, the core section containing the greatest stable Pb
concentration was assigned a median year of deposition of 1972. The total mass of sediment
was determined for core sections above the greatest stable Pb concentration and divided by
2

28 y (2000-1972) to determine the sedimentation rate, 693 g/m /yr. However, applying this
sedimentation rate to the entire core placed background concentrations of Pb much earlier
than other lakes. Therefore, a SCF:CS model was used to date this core with a
2

sedimentation rate of 693 g/m /y for core sections younger than 1972 and a sedimentation
2

rate of 252 g/m /y in core sections older than 1972. The focusing factor for Littlefield Lake
was determined as the average of Elk, Gull, Crystal M, and Higgins lakes.

187

Table S1:

210

Pb data for 26 Michigan lakes

210

137

Lake
Avalon
Birch

Model

Cadillac

CRS

CRS
CRS

Cass
Crystal B
Crystal M
Elk
George

CF:CS
CRS
CRS
SCF:CS
CF:CS

Gratiot

CF:CS

Gull
Hackert

SCF:CS
CRS

Higgins
Houghton
Imp

CF:CS
SCF:CS
CRS

Littlefield
Mullett

Pb
SCF:CS

Muskegon
Otter
Paw Paw
Round
Round D

CF:CS

Sand
Shupac
Torch
Whitmore
Witch

CRS

a

CF:CS
CRS
CRS

CRS
SCF:CS
SCF:CS
CRS

Indicates that the

b137
c

CF:CS

b

NM
1966
1977
1966
1968
b
NM
b

NM
1966
b
NM
1971
1978
b
NM
1965
c
Unclear
1963
1963
1971
1968
a
Disturb
1961
1969
1967
1961

Stable
Pb Date
1974
1982
1973

0.00357-0.764

76-398

1981
1966
1977
1972
1971

0.128-0.309
0.0213-0.636
0.00382-0.536
0.000489-0.637
0.101-0.931

3480
421-796
351-548
450, 980, 220
417

1968

0.000612-1.27

255

1977
1988

0.00147-0.372
0.0133-3.12

190,535
87-1116

1984
1969
1974

232
59,490,152
51-431

assigned
1973

0.00368-0.988
0.000489-0.795
0.00201-2.52
d
0.199-0.250
0.000893-0.426

252,693
980, 1550, 420

1968
1961
1974
1971
1937

0.0911-0.329
0.0768-1.02
0.0741-0.581
0.00486-1.01
0.00280-1.20

1711
933
828
154-442
137-1439

1997
1983
1983
1976
1961

0.00834-0.905
0.00296-3.38
0.000149-0.922
0.0198-0.822
0.0822-1.18

110-2173
99-1134
390, 1108, 329, 1280, 1504
552, 2188, 202
44-590

e

Sedimentation Rate Range
2

(g/m /y)
100-637
235-1871

Cs profile was disturbed and a clear peak in activity was not identified.

Cs was not measured in these lakes.

The

d

137

Cs
Date
1957
1969
a
Disturb

Pb
Excess
Range
(Bq/g)
0.0152-0.461
0.00162-1.10

137

Excess

Cs was not disturbed but a peak was not readily identifiable.

210

Pb values for Littlefield Lake were estimated

e

In the case of the CF:CS model the sedimentation rate applies to the entire core. The range
of sedimentation rates are provided for the CRS model. Sedimentation rates for cores dated
with the SCF:CS model are read from left to right representing sedimentation rates from the
top to the bottom of the core.
188

REFERENCES

189

REFERENCES
Alfaro-De la Torre, M.C., Tessier, A., 2002. Cadmium deposition and mobility in the
sediments of an acidic oligotrophic lake. Geochimica et Cosmochimica Acta 66,
3549-3562.
Golden, K.A., Wong, C.S., Jeremiason, J.D., Eisenreich, S.J., Sanders, M.G., Hallgren, J.,
Swackhammer, D.L., Engstrom, D.R., Long, D.T., 1993. Accumulation and
preliminary inventory of organochlorines in Great Lakes sediments. Water Science
and Technology 28, 19-31.
Heyvaert, A.C., Reuter, J.E., Slotton, D.G., Goldman, C.R., 2000. Paleolimnological
reconstruction of historical atmospheric lead and mercury deposition at Lake Tahoe,
California-Nevada. Environmental Science and Technology 34, 3588-3597.
Kolak, J.J., Long, D.T., Beals, T.M., Eisenreich, S.J., Swackhamer, D.L., 1998.
Anthropogenic inventories and historical and present accumulation rates of copper in
Great Lakes sediments. Applied Geochemistry 13, 59-75.
Kolak, J.J., Long, D.T., Kerfoot, W.C., Beals, T.M., Eisenreich, S.J., 1999. Nearshore versus
offshore copper loading in Lake Superior sediments: Implications for transport and
cycling. Journal of Great Lakes Research 25, 611-624.
Oldfield, F., Appleby, P.G., Battarbee, R.W., 1978. Alternative Pb-210 Dating - Results
from New-Guinea Highlands and Lough Erne. Nature 271, 339-342.
Pearson, R.F., Swackhamer, D.L., Eisenreich, S.J., Long, D.T., 1998. Atmospheric inputs of
polychlorinated dibenzo-p-dioxins and dibenzofurans in sediments of the Great
Lakes: compositional comparison of PCDD and PCDF in sediments. Journal of
Great Lakes Research 24, 65-82.
Pearson, R.F., Swackhamer, D.L., Eisenreich, S.J., Long, D.T., 1997a. Concentrations,
accumulations, and inventories of polychlorinated dibenzo-p-dioxins and
dibenzofurans in sediments of the Great Lakes. Environmental Science and
Technology 31, 2903-2909.
Pearson, R.F., Swackhamer, D.L., Eisenreich, S.J., Long, D.T., 1997b. Concentrations,
accumulations, and inventories of toxaphene in sediments of the Great Lakes.
Environmental Science and Technology 31, 3523-3520.
Robbins, J.A., 1982. Stratigraphic and dynamic effects of sediment reworking by Great
Lakes zoobenthos. Hydrobiologia 92, 611-622.
Simick, M.F., Eisenreich, S.J., Golden, K.A., Liu, S., Lipiatou, E., Swackhammer, D.H.,
Long, D.T., 1996. Atmospheric loading of polycyclic aromatic hydrocarbons to Lake
190

Michigan as recorded in sediments. Environmental Science and Technology 30,
3039-3046.
Walling, D.E., Qingping, H., 1992. Interpretation of caesium-137 profiles in lacustrine and
other sediments: the role of catchment-derived inputs. Hydrobiologia 235, 219-230.
Wong, C.S., Sanders, G., Engstrom, D.R., Long, D.T., Swackhammer, D.L., Eisenreich,
S.J., 1995. Accumulation, inventory, and diagenesis of chlorinated hydrocarbons in
Lake Ontario sediments. Environmental Science and Technology 29, 2661-2672.

191

APPENDIX C
METHODS FOR ACQUISITION OF WATERSHED CHARACTERISTIC DATA

192

APPENDIX C. METHODS FOR ACQUISITION OF WATERSHED
CHARACTERISTICS DATA

The following methods supplement those prepared by Sharon S. Yohn and were used in the
acquisition of the watershed characteristics used in Chapter 3. ArcGIS run on a Windows
platform was used for all GIS work.

Calculating Land Use in Watershed and Buffers
First, reclassifying the land use makes the process a bit easier. A reclass file has been saved
to the C:\Workspace file for both the IFMAP (C:\Workspace\ifmap_class) and 1978 MIRIS
(C:\Workspace\78_class) land use databases. To reclassify the data follow the steps below.
IFMAP
1. Add the buffer polygon file and the watershed land use raster to the active layer
2. In the ArcToolbox→Spatial Analyst→Reclass→Reclassify toolbox
3. Input Raster: The watershed land use raster
4. Reclass Field: Use the Value field
5. Press the Load.. button and find the proper reclassification table (IFMAP or MIRIS)
6. Run the Tool
7. Now you have a reclassified landuse raster. The table below defines how the raster
was reclassified.
Table C1. IFMAP Reclassification Scheme
Land Use Type
Old Classification
Urban
1-4
Agriculture
5,6,7,9
Upland Openland
10,12,13
Forest
14-22,24-26
Water
23
Wetland
27-30
Barren
31,32,35

New Classification
100
200
300
400
500
600
700

a. *Note the Forest classification includes lowland forest which by IFMAP
classifcations is defined as wetland. Following the methods of Yohn S. S.
2004 it will be defined as forest here.
8. Now that the land use has been reclassified we need to clip the watershed land use
raster with the buffer polygon. For this we shall set the analysis mask of the spatial
193

analyst. Setting the mask in spatial analyst will force all operations to be constrained
to the mask and thus must be reset or turned off in order to complete new mask
operations.
9. In the Spatial Analyst Toolbar choose Options
10. In the General Tab:
a. Working directory: Set this to the directory in which to save the files
b. Analysis Mask: Set this to the buffer polygon in the drop down menu
c. Analysis Coordinate System: Set this to the coordinate system of the active
data layer (Be sure to change the active data layer to the Michigan GEOREF
projection)
11. In the Extent Tab:
a. Analysis Extent: In the pulldown menu specify the buffer polygon
12. In the Cell Size Tab:
a. Analysis cell size: In the pulldown menu specify the reclassified watershed
landuse raster
b. Specify the cell size, generally you want to use the lowest cell size possible to
capture all of the detail, 1m is generally works well. 1m will not work well
with large watersheds due to restrictions in space and computing time.
13. Press Ok to exit the Options Window.
14. In the Spatial Analyst toolbar click on the Raster Calculator and double-click on the
reclassified watershed landuse raster and press Evaluate
15. Now a temporary file called “Calculation” has been created. To create a permanent
file right click on the Calculation file and press Make Permanent. This does not have
to be done in able to calculate areas but may be convienent if you want to revisit the
data. Instead of making the raster “Calculation” permanent you can join the results
of the Calculation to the watershed land use raster.
a. Create a file name and save it in the directory
16. Add the file you created when you made the Calculation data file permanent. Right
click on the file and open the attribute table.

194

a. The Value represents the land use classification scheme described above and
the count is the number of cells making up the land use. Multiplying the
count times the square of the cell size will result in the area in square meters
of land use. The attribute file can be exported to Excel for ease of
calculation.
17. With the attribute table open, click on the Options.. menu and click on Export.
a. Export: All records
b. Output Table: Specify the directory and name.
c. Click on the folder button to browse and change the file type to text.
d. Click Save
18. Open the new land cover file in Excel and multiply the count column times the
square of the cell size specified above (generally 30^2). These are the areas of land
use in the buffer around the lake.

195

MIRIS Land Use
The method for calculating 1978 MIRIS land use is different because the values assigned to the
land use classifications change from lake to lake.
1. Add the 1978 land use shapefile clipped to the watershed to the active layer
2. Convert the shapefile to a raster dataset in the ArcToolbox→Conversion Tools→To
Raster→Feature to Raster
a. Input Features: Select the 1978 land use shapefile
b. Field: Select the LABEL1 field
c. Output Raster: Specify the file name and location
d. Output Cell Size: Specify the cell size (1m generally works well for our
watersheds and simplifies the calculation of land use)
3. Follow step 9-14 from the IFMAP area calculations to create a raster of the land use in
the buffer.
4. Right click on the newly created raster and open the attribute table.
a. You will notice that the counts and label name are not preserved in the buffer
raster attribute file we need to join the fields in the buffer raster with the fields in
the 1978 land use watershed raster
5. A join can be made between the 1978 watershed land use file and the buffer file to
preserve the labels
a. Right click on the 1978 land use buffer raster and click on Joins and
Relates→Join…
i. You want to join attributes from a table
ii. Choose the Value Field in the 1. Drop down box
iii. Choose the 1978 watershed land use raster from the 2. drop down box
iv. Choose the Value Field in the 3. drop down box
v. Click OK

6. Now right click on the buffer raster file and the labels are now included in the file.
a. Export the attribute file to a text file
7. Open the newly created text file in Excel and multiply the counts times the square of the
cell size you specified when creating the raster.
8. These are the areas of 1978 land use categorized by the most general level of detail
a. Urban, Agriculture, Nonforested, Forested, Water, Wetlands, and Barren
*Note: Barren in the MIRIS database has a different meaning than Barren in the IFMAP. Barren
in MIRIS is defined as beaches and riverbanks or sand covered beaches.

196

Calculating Land Use/Land Cover Using the Inland Lakes Toolbox in ArcGIS 9.2
This method requires that the Proj_Fill_Fdir_Facc and Watershed_No_Lake_Polygon models
have been run.
Make sure that the ArcINFO license is installed and that the Spatial Analyst Extension is
activated
Details for installing the ArcINFO license can be found at:
http://support.esri.com/index.cfm?fa=knowledgebase.techarticles.articleShow&d=24741, it
requires changes to the registry, so be sure that you understand the implications of this and make
a backup of the current registry if necessary.
This methodology will
1. Calculate the LULC attributes of a watershed based on the 1978 MIRIS database and the
2001 IFMAP
2. Calculate the LULC attributes within a 100m buffer of the lake based on the 1978 MIRIS
database and the 2001 IFMAP
3. Calculate the Flow weighted LULC attributes of a watershed based on the 1978 MIRIS
database and the 2001 IFMAP
4. Calculate the Flow weighted LULC attributes within a 100m buffer of the lake based on
the 1978 MIRIS database and the 2001 IFMAP
Land Use/Land Cover delineated are:
Urban
Agriculture
Forest
Barren
Rangeland/Upland
Water
Wetlands

1978 MIRIS Land Use/ Land Cover
Open up the 1978_Landuse model in the Inland Lakes Toolbox
A dialog box similar to the one below will appear.
The following datasets will be needed:
1. County Land Use
a. These files are available from the MiDL website for each county. If the
watershed spans multiple counties the individual county landuse files will need to
be merged.
i. The merge tool can be found in ArcToolbox→Data
Management→General→Merge

197

2. No Lake Polygon
a. This is the watershed no lake polygon that was created using the
WS_NL_Polygon model, e.g., lakeshed_nl
3. Lake Shapefile
a. This is the lake shapefile created using the Proj_Fill_Fdir_Facc model, e.g.,
lakes_rivers_select.
4. Field
a. Select Level 1. This is the top level Anderson LULC codes for the MIRIS
database
5. Flow Length Raster
a. This is the flow length raster calculated using the WS_NL_Polygon model, e.g.,
lakeshed_fl
6. Extent
a. Browse to the lake_elev file in the NED geodatabase.
7. Environments
a. Click on the Environments button and open up the General Settings drop-down
box.
i. Change the current workspace to the lu1978 geodatabase.

2001 Land Use / Land Cover
Open the 2001_Landuse model in the Inland Lakes Toolbox
A dialog box will appear similar to the one above.
Four datasets will be required
1. Flow length Raster
a. The flow length raster calculated using the WS_NL_Polygon model, e.g.,
lakeshed_fl
2. WS_NL_Polygon
a. The watershed no lake polygon calculated using the WS_NL_Polygon model,
e.g., lakeshed_nl
3. Lake Shapefile
a. The lake shapefile calculated using the Proj_Fill_Fdir_Facc model, e.g.,
lakes_rivers_select. It will be located in the Scratch folder.
4. Lakename
198

a. The name of the lake, an abbreviation of the lake works best, e.g., cad for
Cadillac.
5. Environments
a. Change the current workspace to the lu2001 geodatabase.

199

Creating buffers around lake:
1.
2.
3.
4.
5.
6.
7.

Open lake only file – polygon or shape in the lake layer
In ArcToolbox open Analysis Tools→Proximity→Buffer
Input Features: Select lake only file from drop down menu or browse
Output Features: New file name
Select Distance: Be sure that a coordinate system is specified
Dissolve Type: Select All (This will force the new file to only contain one polygon)
Run the Tool

This creates a buffer extending to the specified distance around the lake but also includes the
lake itself. The lake now needs to be removed, accomplished by creating the union of the lake
and buffer file just created.
8. In ArcToolbox→Analysis Tools→Overlay→Union
9. Input Features: Select the lake only file and the buffer file created above
10. Output Features: The new file name
11. Run the Tool

200

Creating the Watershed No Lake Shapefile Using the Inland Lakes Toolbox in ArcGIS 9.2

***The watershed no lake file is required to calculate the land use/land
cover within the watershed, it erases the lake from the watershed.***
Make sure that the ArcINFO license is installed and that the Spatial Analyst Extension is
activated
Details for installing the ArcINFO license can be found at:
http://support.esri.com/index.cfm?fa=knowledgebase.techarticles.articleShow&d=24741, it
requires changes to the registry, so be sure that you understand the implications of this and make
a backup of the current registry if necessary.
This methodology will:
1. Create a watershed with the lake removed
2. Calculate the flow length clipped to the watershed
Open the Watershed_No_Lake_Polygon model in the Inland Lakes Toolbox
A dialog box will appear similar to the one shown above.
The following datasets will be needed
1. Watershed raster
a. The watershed calculated by the Proj_Fill_Facc_FDir model, e.g., lake_shed
2. FDIR Raster
a. The flow direction raster calculated by the Proj_Fill_Facc_FDir model, e.g.,
FlowDir_Fill1
3. Lake Shapefile
a. The lake shapefile, found in the scratch geodatabase, that was created using the
Proj_Fill_Facc_FDir model, e.g., lakes_rivers_select
4. Directory
a. Browse to the NED geodatabase, this is where all of the output will be stored.
i. Two files are created:
1. Lakeshed_nl – the watershed with the lake removed
2. Lakeshed_fl – the calculated flow length clipped to the watershed
5. Extent
a. Browse the the lake_elev file, this is the original file that was projected, e.g.,
lake_elev.

201

Creating the watershed_nl file in ArcMap 9.2
For many of the analyses it is desirable to have a coverage or shape file of the watershed without
the lake present (e.g., soils in the watershed, land cover, or slope). The watershed_nl file will
also be used to clip grids for flow weighted land cover and average slope calculations.
This method will do the following:
Create a lake_shapefile from the county TIGER lakes dataset
Convert the lake_shapefile to a raster dataset to use for clipping grids
Open up the county TIGER shapefile dataset, these can be downloaded from the MiGDL website
(http://www.mcgi.state.mi.us/mgdl/) making sure that the projection has been defined for the
lakes file, this may be necessary using Data Management Tools Toolbox→Define Projection
Open the watershed file
Open the attribute table of the TIGER lakes file and record the FID for the lake of interest
From the ArcToolbox navigate to Analysis Tools→Extract→Select
Input Features: TIGER lakes file
Output Features: create a name and location for the new lake file (e.g., white_lake)
Expression: click on the SQL box to the right
Double click on “FID”
Click the “=” button
Type in the expression for the FID for the lake, recorded above
From the ArcToolbox navigate to Conversion Tools→From Raster→Raster to Polygon
Input Raster: watershed raster (e.g., white_shed)
Field: ID
Output polygon features: Create a name and location for the new file (e.g., white_shedp)
From the ArcToolbox navigate to Analysis Tools→Overlay→Union
Input Features: watershed polygon file (e.g., white_shedp from above) and the lake file
(e.g., white_lake from above)
Output Feature Class: Create a name and location for the new union file (e.g.,
whiteshed_nl)
Keep the defaults for the remaining selections
Right click on the watershed_nl file (e.g., whiteshed_nl from above) and open the attribute table
In the Editor Toolbar select Start Editing
Select the rows in the watershed_nl file and press delete
In the Editor Toolbar select Stop Editing and save changes
From the ArcToolbox navigate to Conversion Tools→To Raster→Feature to Raster
Input Features: select the watershed_nl file (e.g., whiteshed_nl from above)
Field: ID
Output Raster: create a name and location for the output (e.g., whiteshed_nl)
Output cell size: navigate to the flowaccumulation file for the lake of interest (e.g.,
white_facc)

202

Delineating Watershed (Long Method) in ArcMap 9.2
This method uses NED elevation data available from the US Geological Survey. It compliments
the ArcINFO method but benefits in its GUI interface which is more easily used by traditionally
non-GIS users.
This method will
1. Project the NED elevation data set from the WGS_1983 standard to Michigan GeoRef
1983 (meters)
2. Fill any sinks in the data set
3. Calculate flow direction
4. Calculate flow accumulation
5. Create “pour-points” for watershed delineation
6. Delineate a watershed
Methodology
1.
2.
3.
4.

5.

6.

7.

8.

Obtain 1 arc second NED elevation data from USGS website http://seamless.usgs.gov
Open ArcMap and add the NED elevation data
Open the ArcToolbox if not already open
Navigate to the Data Management Toolbox → Projections and Transformation →
Raster → Project Raster
a. Input Raster: Select the NED elevation data set
b. Output Raster: Create a name (e.g., white_elev) and location for your new
projected elevation data set
c. Output coordinate system: click on the box to the right
i. In the new Spatial Reference Properties dialog box click Select
1. Choose the 1983 Michigan GeoRef (meters) projection from:
Projected Coordinate Systems→State Systems
d. Geographic Transformation: NAD 1983 to WGS 1984_5 (if necessary)
e. Resampling Technique: Cubic
f. Output cell size: 30.86 for 1 arc second DEM data
Navigate to Spatial Analyst Toolbox→Hydrology→Fill
a. Provide the input raster (e.g., white_elev from above)
b. Create a name and location for your new filled elevation data set (e.g.,
white_elevf)
c. Provide a Z-limit if appropriate (usually no Z-limit is required)
Navigate to Spatial Analyst Toolbox→Hydrology→Flow Direction
a. Select the filled elevation dataset (e.g., white_elevf from above)
b. Create a name and location for the flow direction raster data set (e.g., white_fdir)
Navigate to Spatial Analyst Toolbox→Hydrology→Flow Accumulation
a. Select the flow direction raster data set (e.g., white_facc from above)
b. Create a name and location for the flow accumulation raster data set (e.g.,
white_facc)
Double-click on the newly created flow accumulation raster data set
a. Click on the Symbology tab
203

b. In the Show: frame at left click on classified
c. Click on the classify button at the right
d. Change the classification to the standard deviation and change the Interval Size to
¼ Std Dev
e. Apply the changes (you may need to invert the colors or choose an appropriate
color ramp to see the output)
f. You should now see several lines of flow accumulation in the data frame, zoom
into the lake of interest and find the output (i.e., where flow accumulation is the
highest, can be found using the identity tool)
9. Open ArcCatalog and navigate to the folder where the raster files are saved.
a. Right click in the right most open pane
b. Click on New
i. Click on Shapefile
1. In the Create New Shapefile dialog box
a. Create a name for the shapefile (e.g., whiteout)
b. Make sure Feature Type: is point
c. Click EDIT and select the Michigan GeoRef 1983 (meters)
projection
10. Change the Spatial Analyst and Environment Settings to reflect the extent and cell size of
the flow accumulation raster data set
a. Click on the Spatial Analyst Toolbar→Options
i. click on the Extent Tab
1. Change the Analysis Extent: to the flow accumulation raster (e.g.,
white_facc)
ii. Click on the Cell Size tab
1. Change the Analysis cell size: to the flow accumulation raster
(e.g., white_facc)
b. In the Tools menu navigate to Options and click on the Geoprocessing tab, open
Environments…
i. In the General Settings drop down specify the Extent: to the flow
accumulation raster (e.g., white_facc)
ii. In the Raster Analysis Settings drop down box specify the Cell Size: to the
flow accumulation raster (e.g., white_facc)
c. Record the cell size, you will need this number in Step 11 below.
11. In ArcMap add the newly created shapefile (e.g., whiteout from above)
a. In the Tools Menu click on the Editor Menu if not already open
b. In the Editor Tools click start editing
i. In the dialog box select the file that contains the newly created shapefile
(e.g., whiteout from above)
c. Using the sketch tool place the newly created shapefile in the flow accumulation
line where it exits the lake
d. In the Editor Tools click stop editing and save the changes
e. Navigate to the Conversion Tools Toolbox→To Raster→Feature to Raster
f. Select the newly created and positioned shapefile
g. Create a new name for the raster (e.g., white_out)

204

h. Make sure the output cell size is the same size as the flow accumulation raster
data set, recorded in step 10 above
12. Navigate to the Spatial Analyst Tools Toolbox→Hydrology→Watershed
a. Select the flow direction raster (e.g., white_fdir)
b. Select the raster pour point (e.g., white_out)
c. Pour point field should be Value
d. Create a name for the output (e.g., white_shed)

205

Delineating Watersheds using the Inland Lake Toolbox in ArcGIS 9.2
Make sure that the ArcINFO license is installed and that the Spatial Analyst Extension is
activated
Details for installing the ArcINFO license can be found at:
http://support.esri.com/index.cfm?fa=knowledgebase.techarticles.articleShow&d=24741, it
requires changes to the registry, so be sure that you understand the implications of this and make
a backup of the current registry if necessary.
This methodology will:
7. Project the NED elevation data set from the WGS_1983 standard to Michigan GeoRef
1983 (meters)
8. Fill any sinks in the data set
9. Calculate flow direction
10. Calculate flow accumulation
11. Convert the lake shapefile to a raster dataset
12. Calculate the watershed using the lake as the pour point
In ArcCatalog create a new lake folder in the lakes_ned file folder found in the C:\Workspace
directory
In the new lake folder create four new file geodatabases
1. ned - for all of the elevation, filled elevation, flow direction, flow accumulation, lake
rasters, and watershed files
2. scratch – necessary for future landuse toolboxes
3. lu2001 – will contain all of the output from the 2001 landuse toolbox
4. lu1978 – will contain all of the output from the 1978 landuse toolbox

Methodology
Obtaining and projecting the NED data from USGS
13. Obtain 1 arc second NED elevation data from USGS website http://seamless.usgs.gov
14. Open ArcMap and add the NED elevation data
15. Open the ArcToolbox if not already open
16. Navigate to the Data Management Toolbox → Projections and Transformation →
Raster → Project Raster
a. Input Raster: Select the NED elevation data set
b. Output Raster: Create a name (e.g., white_elev) and location for your new
projected elevation data set (i.e., the ned geodatabase in the lake folder created
above)
c. Output coordinate system: click on the box to the right
i. In the new Spatial Reference Properties dialog box click Select
1. Choose the 1983 Michigan GeoRef (meters) projection from:
Projected Coordinate Systems→State Systems
d. Geographic Transformation: NAD 1983 to WGS 1984_5 (if necessary)
206

e. Resampling Technique: Cubic
f. Output cell size: 30 for 1 arc second DEM data

Using the Proj_Fill_Fdir_Facc toolbox

1. Double-click on the Proj_Fill_Fdir_Facc model in the Inland_Lakes toolbox and a dialog box
will appear.
a. Input the necessary features
i. Projected NED Raster is the raster downloaded and projected above
ii. Input Features is the county lakes_rivers TIGER95 shapefile, this may be a
problem since it appears that MiGDL has discontinued the use of the TIGER95
files
iii. Expression – this input relates to the lakes_rivers input that you selected in the
Input Features input line
iv. Click on the SQL button to the right
v. In the new dialog box
a. Scroll to “Name” in the menu to the left and double click on it
b. Press the equals sign
c. Scroll to the Lake Name in the menu to the left and double click on
it
d. Press OK
b. In the Field input, scroll down to AREA
c. In the Extent input, browse to the projected lake_elev file you created in the steps above
d. In the Directory input, browse to the NED geodatabase that was created in the steps
above
e. This steps insures that all data created are put into the NED folder
1. The lake_shapefile that is created is put into the scratch folder
2. All other datasets are put into the NED folder
f. Click OK
Depending upon the size of the elevation dataset, the tool may take up to ½ hour before
completing.

207

Determining Land Use in Watershed (ArcMap 9.2)
Land use for Michigan comes from two sources the IFMAP (Landsat 2001) and the MRBIS
1978 land use data. The ifmap data should be available on the local machine in
C:\Workspace\ifmap. There are two IFMAP rasters, lp and up, for the Lower Peninsula and
Upper Peninsula, respectively. The 1978 land use must be obtained from MiGDL by county at
the following website http://www.mcgi.state.mi.us/mgdl/ . This method assumes that you have
already created the watershed_nl file.
This method will:
1. Clip the land use data file to the watershed
2. Reclassify data to the highest level, needed to compare land use among datasets.
3. Create tables for export into Excel for data analysis
2001 land use IFMAP
1. In the ArcToolbox navigate to Spatial Analyst Tools→Extraction→Extract by Mask
a. Input Raster: IFMAP raster
b. Input Raster or feature mask data: watershed_nl raster
c. Output Raster: Create a name and location for the file (e.g., white_2001)
2. In the ArcToolbox navigate to Spatial Analyst Tools→Reclass→Reclassify
a. Input Raster the watershed landuse file (e.g., white_2001 from above)
b. Reclass field: Value
c. Reclassification: follow the reclassification scheme in the table below. The
reclassification can be saved and used for other lakes
Table C2. Landuse reclassification scheme
Land Use Type
Old Classification
Urban
1-4
Agriculture
5,6,7,9
Upland Openland
10,12,13
Forest
14-22,24-26
Water
23
Wetland
27-30
Barren
31,32,35

New Classification
100
200
300
400
500
600
700

d. Output raster: create a name and location for the new file (e.g., white_2001rc)
3. Right click in the newly created reclassified watershed land use file and open the attribute
table
a. The VALUE corresponds to the reclassified land use in the table above
b. The COUNT column is the number of cells in the watershed with a specific land
use
i. The product of COUNT and area of the cell is the total land use in the
watershed
c. Click Options: in the lower right hand corner of the attribute table and select
Export…
208

i. Export: All records
ii. Output Table: create a name and location for the file (e.g., white_2001)
the file can be saved as “.dbf” or “.txt”
4. Right click in the newly created reclassified watershed land use file and open
Properties…
a. Click on the Source tab
b. Note the cell size, this value squared will be the area of one raster cell
5. Open the exported table in Excel
a. Multiply the COUNT column by the area of the raster cell
b. Sum all of the areas
c. Compare the calculated land use to the area of watershed_nl file by calculating
the product of COUNT and the cell area for the watershed_nl file. COUNT can
be found in the attribute table of the watershed_nl file.
d. Area of the total land use in the watershed and the watershed should be
comparable
e. Save the file as an Excel worksheet
1978 MIRIS Land use
1. Download the 1978 MIRIS land use data from MiGDL
a. If multiple the watershed spans several counties it will require the individual
county files to be merged
2. From the ArcToolbox navigate to Data Management Tools→General→Merge
a. Input Datasets: select all of the county 1978 land use files
b. Output Dataset: create a name and location for the output (e.g., 1978 lulc)
3. From the ArcToolbox navigate to the Analysis Tools→Overlay→Intersect
a. Input Features: select the watershed_nl (e.g., whiteshed_nl) and merged land use
files (e.g., 1978 lulc from above)
b. Output Feature Class: create a name and location for the output (e.g., white_1978)
c. keep all of remaining selection at default values
4. From the ArcToolbox navigate to the Conversion Tools→To Raster→Feature to Raster
a. Input features: select the watershed landuse file (e.g., white_1978 from above)
b. Field: Select Level 1
c. Output Raster: create a name and location for the file (e.g., white_1978)
d. Output cell size: either enter the cell size desired (usually the same as the original
DEM) or navigate to the original projected DEM
5. Right click on the newly created watershed land use raster (e.g., white_1978 from step 4)
and open the attribute table
a. Export the data to a text file following step 3 from the IFMAP watershed land use
methodology above
6. Open the exported table in Excel
a. Multiply the COUNT column by the area of the raster cell
b. Sum all of the areas
c. Compare the calculated land use to the area of watershed_nl file by calculating
the product of COUNT and the cell area for the watershed_nl file. COUNT can
be found in the attribute table of the watershed_nl file.

209

d. Area of the total land use in the watershed and the watershed should be
comparable
e. Land use codes for 1978 land use
Table C3. Land use classifcations for the 1978 MIRIS dataset
Land Use
Level 1 Code
Urban
1
Agriculture
2
Rangeland
3
Forest
4
Water
5
Wetlands
6
Barren
7
f. Save the file as an Excel worksheet

210

APPENDIX D
ERROR ANALYSIS

211

APPENDIX D. Error Analysis

Absolute error in analytical methods can be derived from both random and systematic
sources. Sources of random errors are operationally unknown. They appear as spread in the data
during the analysis of a known analyte concentration and can often be described using Gaussian
statistics. Sources of systematic error are known and include: instrumental, personal and
method. Instrumental error occurs due to the wear, corrosion, or drift in electrical circuits over
time; it is usually overcome by the calibration of instruments prior to analysis. Personal error
can include such things as number bias and prejudice. Method errors occur due to non-ideal
chemical and physical behavior of reactants during analysis. Analysis of NIST standards, of
similar matrix to the samples of interest, can aid in determining method error. This report shall
outline the error in sediment mercury analysis from six sampling seasons (1999-2004) and
sediment major and trace element analysis from two sampling seasons (2003-2004).

Systematic errors will be assumed to instrumental in this analysis. Personal errors are
limited due to computer interfaces but still exist due to pipeting or weighing errors. Pipeting and
weighing errors generally considered gross errors and can be eliminated utilizing outlier analysis.
Outliers will be defined, by box plots, as . Method errors are generally minimized by calibration
curves which are run during each set of analyses.

Descriptive Statistics

The definitions and equations described below shall be utilized in the examination of error in
both mercury and major and trace element analysis. Equations presented below can be found in

212

any instrumental analysis chemistry textbook such as Skoog and Leary (1992). Absolute error,
E a , is defined as the deviation of the sample mean, x from the known analyte value, xt ,
E a = x − xt

(1)

where the right hand side of the equation includes both random and systematic components.
Random error, E r , for a small set of data can be estimated using

Er = x − µ

(2)

where µ is the population mean. In general the population mean is determined, and assumed to
be free of random error, by replicating the measurement 20 to 30 times and is calculated by
N

∑ xi
µ=

lim

N →∞

i =1
N

(3)

where xi is the value for the ith measurement and N is the number of measurements. The
standard deviation, σ, of the population is calculated
N

∑ ( xi − µ ) 2
σ=

lim

i =1
N

N →∞

(4)

The standard deviation of a set of samples differs from the population standard deviation by
replacing σ with s, µ with x , and N by N-1.

N

∑ ( xi − x ) 2
s = i =1

N −1

(5)

The arithmetic mean, x , will be used for sample means. The value of the population and sample
mean may differ. The sample mean usually represents a subset of the population or a smaller set

213

of measurements intended for comparison to the population. Standard errors shall also be
calculated to demonstrate the fluctuation in the standard deviation based upon the number of
samples
sm =

s
N

.

(6)

The coefficient of variation, CV,
CV =

s
× 100%
x

(7)

provides information on the magnitude of the standard deviation as a percentage of the mean and
is generally more useful than the standard deviation alone.

95% confidence limits are quantified using a t-statistic as
95% CL = x ±

ts
N

(8)

where t is obtained using a widely available statistical t-distribution table. Confidence limits can
also be calculated using the z-distribution table where z replaces t, σ replaces s and µ replaces
x.

Systematic Error (Bias)
Bias in analytical methods is developed due to systematic errors and can be demonstrated using
the t-statistic. The experimental difference between the sample mean and population mean is
calculated as

x−µ

214

(9)

and compared to the calculated difference at a predetermined confidence level as
±

ts
N

.

(10)

Bias is demonstrated if the experimental bias is greater than the calculated bias at a given
confidence interval. A similar method can determine bias in a population where

xt − µ
±

and

zs
N

(11)
(12)

replace Equations 9 and 10 respectively.
Limit of Detection
The detection limit (LOD) of an analytical method, determined here as the minimum
distinguishable signal, Sm, is a function of average signal of the blank, S bl , and a multiple, k, of
the blank signal, sbl

S m = S bl + ksbl .

(13)

Where blank data are available totaling a minimum of 20 to 30 measurements Equation 13 will
result in the detection limit, cm, as

S − S bl
cm = m
.
m

(14)

where m is the slope of the calibration line at the concentration of interest. In the absence of
reliable blank measurements the detection limit will be determined by linear regressions of
calibration standard concentrations versus instrument response. In these cases the slope,
intercept, and standard error of the estimate will be used to estimate the LOD.

215

Mercury
Mercury analysis was carried out on a Lumex Zeeman Corrected Atomic Absorption
Spectrophotometer with Thermal Decomposition attachment (TD-AA). Calibration of the
instrument was performed using either SRM 1633b (coal fly ash, Hg = 143.1 ng/g) or SRM 1515
(apple leaves, Hg = 44 ng/g) and the protocol outlined in Environmental Protection Agency
(EPA) method 7473 was followed. EPA Method 7473 dictates that a SRM be run every 10
samples; descriptive statistics will be derived based upon these measurements.
Descriptive Statistics and Bias
Table D1 and Table D2 show the population means and standard deviation of SRM 1515 and
SRM 1633b respectively. Based on the number of measurements for each SRM it is assumed
that the standard deviation and means of these measurements represent the population. Bias for
the population was determined using the z-distribution and was calculated using Equations 11
and 12; at 95% CL z is equal to 1.96.

216

Table D1 SRM 1515 Population Statistics
Population mean (ng/g)
Population standard deviation (ng/g)
CV (%)
sm
N
NIST accepted value (ng/g)
Bias

43.9
2.13
4.85
0.23
87
44
No bias demonstrated at 95% CL

Table D2 SRM 1633b Population Statistics
Population mean (ng/g)
Population standard deviation (ng/g)
CV (%)
sm
N
NIST accepted value (ng/g)
Bias

146.8
6.644
4.52
0.83
64
143.1
Presence of bias suggested at 95% CL

The presence of positive bias in SRM 1633b suggests that systematic errors influence the
measurement.

Table D3 shows the descriptive statistics for each lake. Sample means and standard deviations
are provided for each lake and are subsets of the population determined as subsets of the
population. The SRM used during analysis of the lake sediment mercury is noted in column 2.
Bias was calculated using Equations 9 and 10; the population mean was used to evaluate
Equation 9. It is evident from this analysis that only three lakes exhibited bias at the 95% CL,
Houghton, Littlefield, and Torch.

217

Table D3 Descriptive Sample Statistics and Bias
Lake
Houghton
Round
Avalon
Birch
Muskegon
Sand
Crystal-M04
George
Hackert
Otter
Round_D
Gull
Higgins
Cass
Crystal-M
Littlefield
Cadillac
Crystal-B
Hubbard
Mullett
Whitmore
Houghton
Imp
Paw Paw
Torch
Witch

SRM
1633b
1633b
1633b
1633b
1633b
1633b
1633b
1633b
1633b
1633b
1633b
1515
1515
1515
1515
1515
1515
1515
1515
1515
1515
1515
1515
1515
1515
1515

N
4
6
7
7
8
5
5
5
6
4
6
3
11
4
5
9
5
5
5
6
6
2
7
7
6
6

df
3
5
6
6
7
4
4
4
5
3
5
2
10
3
4
8
4
4
4
5
5
1
6
6
5
5

Average
(ng/g)
141.0
142.0
144.0
144.9
151.6
149.4
144.2
144.7
148.0
154.0
150.3
47.0
43.5
43.3
44.0
46.6
43.2
43.6
43.8
44.3
43.5
44.0
42.4
43.0
42.3
44.3

Std. dev
(ng/g)
3.16
5.90
3.92
3.85
9.74
2.30
4.60
4.32
5.33
10.8
5.57
1.66
1.86
1.50
1.87
2.55
2.05
1.67
1.30
1.75
2.17
1.41
1.62
2.08
0.82
2.07

CV
(%)
2.24
4.15
2.72
2.66
6.42
1.54
3.19
2.99
3.60
7.01
3.71
3.54
4.29
3.47
4.25
5.49
4.74
3.84
2.98
3.95
4.98
3.21
3.81
4.84
1.93
4.66

sm
1.58
2.41
1.48
1.45
3.44
1.03
2.06
1.93
2.18
5.40
2.28
0.96
0.56
0.75
0.84
0.85
0.92
0.75
0.58
0.71
0.89
1.00
0.61
0.79
0.33
0.84

Bias
Bias
Bias
Bias
-

Detection Limits
Detection limits, outlined in Table D4 and Table D5, were calculated using Equations 13 and 14
using the average blank signal obtained during sample runs. Slope was determined from plots of
mass Hg vs. Area. In Equation 13, k was assigned a value of three (REF). Differences in the
detection limit differed only slightly between the two SRMs. Detection limit concentrations are
calculated using the population average (e.g. if 250 mg of sediment are used the LOD is 0.687
ng/g).

218

Table D4 SRM 1515 LOD
Mean
Standard Deviation

Detection Limit (ng/g)

Mass Hg (pg)
141.8
20.3

Hg (fmoles)
706.9
101.2

250 mg
0.687

125 mg
1.37

Mass Hg (pg)
133.5
6.06

Hg (fmoles)
665.5
30.2

250 mg
0.577

125 mg
1.15

10 mg
17.2

Table D5 SRM 1633b LOD
Mean
Standard Deviation

Detection Limit (ng/g)

10 mg
14.4

Discussion
Results of the mercury analysis indicate that, run to run, very little systematic error has been
demonstrated and that detection limits are low. Population mean of SRM 1515 agreed well with
the accepted value and did not demonstrate bias. Conversely SRM 1633b did demonstrate bias
at a 95% CL. This may be an effect of improperly cleaned glassware. It is evident that during
the analysis of high organic carbon sediments that lenses encasing the reaction cell can become
contaminated. This reduces the amount of light penetrating the reaction cell which reduces
signal and increases the amount of Zeeman correction (Lumex, personal communication). Otter,
Sand, Muskegon, and Round_D lakes all visually appeared to contain high amounts of organic
carbon and these lakes have higher average values. It may be necessary to implement a protocol
for periodic lens cleaning during analysis of high organic carbon sediments. Individual runs
indicate that only three lakes demonstrate bias and that none of the lakes demonstrating bias
would be considered to have high organic carbon values. The test for bias by lake suffers from
using t-statistics versus the z-statistic for the population. Detection limits are lower than those
estimated by the manufacturer for this method (WWW.OHIOLUMEX.COM).
219

REFERENCES

220

REFERENCES

Skoog D. A. and Leary J. J. (1992) Principles of Instrumental Analysis, Fourth Edition. Saunders
College Publishing.
SPSS. (2000) SYSTAT 10.
www.ohiolumex.com.
Yohn S. S. (2004) Natural and Anthropogenic Factors Influencing Spatial and Temporal Patterns
of Metal Accumulation in Inland Lakes, Michigan State University.

221