Understanding Geochemical Recovery in Anthropogenically Disturbed Landscapes
By
Ryan Glenn Vannier

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Geological Sciences – Doctor of Philosophy
2014

ABSTRACT
UNDERSTANDING GEOCHEMICAL RECOVERY IN ANTHROPOGENICALLY
DISTURBED LANDSCAPES
By
Ryan Glenn Vannier
Environmental regulations have greatly reduced the output of anthropogenic chemicals
to the environment from the mid 1970’s to the present. However, lake sediment cores collected
in Michigan show temporal concentration profiles of chemicals considered toxic to humans and
wildlife to decrease at variable rates during the decades following the 1970s. This disparity may
be due factors affecting recovery such as watershed or regional land use, population, and
chemical production/consumption trends. For this study, sediment chemical chronologies
collected from inland lakes across Michigan were compared to determine temporal and spatial
trends of two organic chemicals: PCBs and DDT. These chemicals were selected for study as
they both represent banned chemicals with different utilization and dispersion patterns, as well
as, different chemical characteristics. Surface water chemistry was also studied in the Saginaw
Bay Watershed attempting to relate differences in the dissolved constituent make-up with land
use as a means of determining changes to the chemical pathways through the changing seasons.
Results for DDT show US production and usage to be the overriding determinants of the
general sediment profiles. However, results show significant variation amongst the individual
lakes in terms of temporal peak values and recovery since the ban of these chemicals. This is
thought to be the result of both independent watershed-scale and regional sources of DDT
recorded in the lake sediment. Some correlation to watershed %urban land-use is apparent but
land-use characteristics overall were poorly correlated to DDT concentrations, peak and onset
dates. DDT and its metabolites found in recent sediment (top 4 cm), as well as, sediment

accumulation rates of these chemicals in each sediment core are well correlated to latitudes of
Michigan with higher populations, even if the lake watershed itself wasn’t especially populace.
This suggests a significant regional component to DDT both historically and since its ban.
PCBs were elevated in the Lower Peninsula compared to the Upper Peninsula and
locations of high PCB inventories near urban areas were found. However, concentration and
accumulation peak dates increased with latitude of the lake possibly due to secondary mobility
and deposition away from sources. During high PCB production years, congener clusters show
clear localized patterns, but more recently, congener distributions suggest a more regional signal.
However, state-wide regional patterns appear to serve as a component to both recent and
historical data, despite the prominence of local and sub-regional trends during the peak PCB
production/consumption years.
The SBW study documented strong local fluctuations of solutes attributed primarily to
local proximity of urban centers and agricultural land use. However, the geochemical fingerprint
of a given land use varied in intensity throughout the year: agricultural land use on surface water
chemistry was most evident during a summer low flow event and a spring high flow event,
which may be a reflection of seasonal variations in the intensity of agricultural practices. Urban
land use exhibited its strongest fingerprint during winter high flow, likely due to road deicing salt
application.
Overall, we conclude that studies attempting to link chemical concentrations to a
particular land use should give consideration to coordinating surface water analytes to the
hydrologic regime and time of year the sample collection is to be undertaken. Further, the
regional component of banned chemicals continues to be represented in modern sediment,
despite several decades since cessation of production.

ACKNOWLEDGEMENTS

I would like to thank my advisor, David Long, for his guidance and insight throughout
this project. I would also like to give a personal thanks to Matthew Parsons for his many hours
spent helping to provide training on the lab protocols for this project and for assistance with the
ICP-MS. Thanks to my committee member Grahame Larson, for his thoughts and
encouragement during my time at Michigan State University. Thanks to my other committee
members: Tom Voice and Phani Kumar for their many insights into this project. Thanks to the
many graduate and undergraduate students helped with lab techniques, or provided guidance and
encouragement including Colleen McClean, Paul Grieve, and Amanda Robinson to name but a
few. A special thanks to the people at the Michigan Department of Environmental Quality, who
provided the funding for this project. Finally, I would like to thank my girls: Grace and Amy,
for all of their love, encouragement, and support.

iv

TABLE OF CONTENTS

LIST OF TABLES

vii

LIST OF FIGURES

ix

CHAPTER 1. INTRODUCTION
Overarching Research Hypothesis and Approach
General Methodology
REFERENCES

1
4
5
11

CHAPTER 2. HISTORICAL DISTRIBUTION AND RECOVERY OF DDT AS
INTERPRETED FROM MICHIGAN INLAND LAKE SEDIMENTS
Abstract
Introduction
Hypothesis and Test
Methodology
Study Setting
Sample Collection
Chemical Analysis
Age Dating
Results and Discussion
Spatial DDT/DDD/DDE Concentrations, Inventories, and Accumulation
Rates
Temporal DDT/DDD/DDE Trends
DDT Onset Trends
DDT Peak Values, Spatial and Temporal Relationship
DDT Trends in Recent Sediment
Sediment DDT Degradation and Half-Life Calculations
Sediment DDT Concentrations and Porewater Trace Metal Profiles
DDT Half-Life Calculations
DDT and Land Use
Conclusion
APPENDIX
REFERENCES

15
15
17
22
24
24
25
26
26
29
29
37
39
42
44
45
47
49
51
52
55
62

CHAPTER 3. LANDSCAPE RECOVERY FROM HISTORICAL PCB LOADINGS USING
SEDIMENT-CHEMICAL CHRONOLOGIES
70
Abstract
70
Introduction
72
Hypothesis and Test
80
Methodology
82
Sampling Methodology
82
Studying Setting
83
Age Dating
85
v

Results and Discussion
PCB Temporal Concentration Analysis
PCB Accumulation Rate Analysis
PCB Inventory Analysis
PCB Congener Analysis
Cluster Analysis
Conclusions
REFERENCES
CHAPTER 4. SEASONAL INFLUENCE ON LAND USE BIOGEOCHEMICAL
FINGERPRINTS IN SURFACE WATER BODIES
Abstract
Introduction
Methodology
Study Site
Sample Plan and Analysis
Watershed and Land-Use Distribution
Statistical Procedures
Results
Temporal Chemical Variability
Factor Analysis of Individual Sample Events
Factor Analysis of Combined Data Sets
Sodium, Bromide, and Chloride Sources in the SBW
Conclusions
REFERENCES

vi

88
88
101
104
107
111
119
122

133
133
135
140
140
141
145
146
147
147
148
155
157
167
170

LIST OF TABLES

Table 2.1: Physical characteristics of study lakes……………………………………………......56
Table 2.2: Decadal average DDT concentration of all study lakes………………………………57
Table 2.3: DDT Concentration Peak Date, Peak DDT Concentration, Total Sediment Core DDT
Concentration, Calculated DDT Half-Life, and R2 for Half-Life……………………….……….58
Table 2.4: Select data from 210Pb analysis, including the model used for dating, mixed depth,
surficial sedimentation rate, focusing factor (FF) and the age of the oldest section in the sediment
core……………………………………………………………………………………………….59
Table 2.5: GIS-determined 1970s watershed land-use percentages for all Michigan study lakes
organized by latitude……………………………………………………………………………..60
Table 2.6: GIS-determined 1990s watershed land-use percentages for all Michigan study lakes
organized by latitude……………………………………………………………………………..61
Table 3.1: PCB concentrations of all study lakes averaged by decade and concentrations
averaged by decade for each of the three regional zones. UP = Upper Peninsula, LP North =
Northern Lower Peninsula, LP South = Southern Lower Peninsula. Values are normalized to
peak concentration of the individual lakes for the table on the right .............................................93
Table 3.2: Focus-Corrected Accumulation Rates (ug/m2/yr) of all study lakes averaged by
decade and concentrations averaged by decade for each of the three regional zones. The highest
value in each category is colored grey. UP = Upper Peninsula, LP North = Northern Lower
Peninsula, LP South = Southern Lower Peninsula ........................................................................93
Table 3.3: Focus-Corrected PCB Inventories (ug/cm2) of all study lakes averaged by decade and
concentrations averaged by decade for each of the three regional zones. Highest value in each
category in grey. UP = Upper Peninsula, LP North = Northern Lower Peninsula, LP South =
Southern Lower Peninsula .............................................................................................................93
Table 4.1: Basic statistics organized by sample event .................................................................148
Table 4.2a: Factors for the RIV-I data collected 10-11 July 1995. Numbers in bold indicate
strong loading on a factor (loading > 0.75). Only loadings >0.5 or <-0.5 are shown ................152
Table 4.2b: Factors for the RIV-II data collected 20-21 July 1995. Numbers in bold indicate
strong loading on a factor (loading > 0.75). Only loadings >0.5 or <-0.5 are shown ................153

vii

Table 4.2c: Factors for the RIV-III data collected 1-5 September 1995. Numbers in bold indicate
strong loading on a factor (loading > 0.75). Only loadings >0.5 or <-0.5 are shown ................153
Table 4.2d: Factors for the RIV-IV data collected 21-25 February 1995. Numbers in bold
indicate strong loading on a factor (loading > 0.75). Only loadings >0.5 or <-0.5 are shown ...154
Table 4.2e: Factors for the RIV-V data collected 23-25 May 1995. Numbers in bold indicate
strong loading on a factor (loading > 0.75). Only loadings >0.5 or <-0.5 are shown ................154
Table 4.3: Factors for data from all sampling events combined. Numbers in bold indicate strong
loading on a factor (loading > 0.75). Only loadings >0.5 or <-0.5 are shown ...........................157
Table 4.4: Sample ID, location, and land use distribution for sampling of Saginaw Basin
Watershed, MI, USA. Values in bold indicate highest land use percentage. Dominant land use is
defined as land use percentage >50%. Those sample IDs without a land use >50% are defined as
‘mixed’. ........................................................................................................................................168

viii

LIST OF FIGURES

Figure 1.1: Project Lakes 1999 to 2010 ...........................................................................................9
Figure 2.1: Domestic Production and Consumption of DDT in the United States, 1950-1972.
The 1975 EPA report entitled DDT, A Review of Scientific and Economic Aspects of the
Decision To Ban Its Use as a Pesticide analyzed domestic production and consumption data
which showed peak production and consumption of DDT to be in 1963 and 1959 respectively.
Source: EPA-540/1-75-022, p.40 ...................................................................................................18
Figure 2.2: Locations of study lakes subject to DDT analysis from 1999 to 2010........................25
Figure 2.3: Temporal concentration profiles for six study lakes with highest peak total DDT in
order by peak concentration value .................................................................................................29
Figure 2.4: Summed concentrations of DDT and its breakdown products (DDD and DDE) in the
0-2 cm and 2-4 cm sediment depth intervals for Michigan inland lakes arranged by latitude ......31
Figure 2.5: Summed focus-corrected accumulation rates of DDT and its breakdown products
(DDD and DDE) in the 0-2 cm and 2-4 cm sediment depth intervals for Michigan inland lakes
arranged by latitude........................................................................................................................33
Figure 2.6: Average focus-corrected accumulation rates of DDT and its breakdown products
(DDD and DDE) since 1940 for Michigan inland lakes arranged by latitude ...............................34
Figure 2.7: Spatial distribution of average focus-corrected accumulation rates of DDT and its
breakdown products (DDD and DDE) since 1940 for Michigan inland lakes ..............................34
Figure 2.8: Spatial distribution of focus-corrected DDT inventories for Michigan inland lakes ..36
Figure 2.9: Focus-corrected DDT inventories for Michigan inland lakes arranged by latitude ....36
Figure 2.10: Decadal averages of Total DDT concentrations of 39 lakes in Michigan.................38
Figure 2.11: Normalized decadal averages of Total DDT concentrations of 39 lakes in Michigan
........................................................................................................................................................38
Figure 2.12: Excess 210Pb activity and 137Cs activity versus depth in Teal Lake, Michigan .........42
Figure 2.13: Year and magnitude of peak concentration for DDT in inland lakes of Michigan ...44

ix

Figure 2.14: Ratios of DDT/(DDD+DDE) as a function of time in selected Michigan inland lake
sediments........................................................................................................................................47
Figure 2.15: DDT/(DDD+DDE) ratio and porewater iron concentrations plotted as a function of
depth in Gull Lake .........................................................................................................................48
Figure 2.16: Half-life calculation plot for Higgins Lake ...............................................................50
Figure 2.17: 1970s Total DDT versus study lake watershed population .......................................51
Figure 2.18: Sources and water column cycling of DDT associated with suspended and settling
solids ..............................................................................................................................................53
Figure 3.1: A schematic of the biologically mediated transport of PCBs to the inland lake
ecosystem. Initially, PCBs are emitted into the environment from industrial, municipal, or
agricultural regions into the atmosphere or directly to surface water bodies. Much of the PCBs
in Michigan are transported by air to the Great Lakes where they can become widely distributed
around the entire Great Lakes Region. Biota in the higher trophic levels of the Great Lakes
ecosystem (including piscivorous birds and fish) bioaccumulate the PCBs primarily in fatty
tissues. The wildlife then travels to inland where the PCBs are then deposited inland by delivery
of fecal matter, molting, or mortality of the organism. From Blais et al., 2007 ...........................75
Figure 3.2: 44 lakes studied with full core subject to PCB analysis. Note that this is a subset of
the 47 individual lakes cored total, 3 of which only top 4 cm were analyzed for PCBs ...............84
Figure 3.3: Log 210Pb activity (Bq/g) versus accumulated dry mass (g/cm2) in Lake Gogebic.
Regression lines for the CF:CS (constant flux-constant sedimentation) model and the SCF:CS
(segmented constant flux-constant sedimentation) model are shown............................................86
Figure 3.4: Terrestrial ecoregions framework as defined by Albert (1995). The separate Upper
Peninsula ecoregions were considered together for this study due to the lack of suitable lakes
meeting the criteria needed to qualify them for this project ..........................................................88
Figure 3.5a: Temporal Concentration Profiles of lakes in the Upper Peninsula of Michigan since
the year 1900. .................................................................................................................................89
Figure 3.5b: Temporal Concentration Profiles of lakes in the northern portion of the Lower
Peninsula since the year 1900. The peak of Lake George was not included for clarity of the
lower concentration lakes ..............................................................................................................90
Figure 3.5c: Temporal Concentration Profiles of lakes in the southern portion of the Lower
Peninsula since the year 1900. The peak of lakes Thompson and White were not included for
clarity. Note the change in PCB Concentration scale of the X-axis relative to those of the
Figures 3.5a and 3.5b .....................................................................................................................92

x

Figure 3.6: Historical records of the sales/production volumes of PCBs and the similarity of
these time-varying trends to the accumulation rates of these chemicals in the sediments of Lake
Ontario (from Eisenreich et al., 1989) .......................................................................................... 94
Figure 3.7: Sum of PCB congener concentration (ng/g dry wt) in surficial sediments of inland
lakes across Michigan organized by latitude .................................................................................95
Figure 3.8: Comparison of averaged 210Pb dated decadal block of normalized Pb concentrations
from 1930s to present in three physiographic zones described by Albert et al., 1995. Note: not
all lakes were available for Pb inventory analysis as it requires recognizable background
(reference) trends of Pb which are not recognized in several cores ..............................................98
Figure 3.9: Comparison of averaged 210Pb dated decadal block of normalized PCB
concentrations from 1930s to present in three physiographic zones described by Albert et al.,
1995................................................................................................................................................99
Figure 3.10: 1970s PCB Accumulation Rates for all lakes organized by latitude. White Lake
peak removed for clarity ..............................................................................................................103
Figure 3.11: 2000s PCB Accumulation Rates for all lakes organized by latitude. White Lake
peak removed for clarity ..............................................................................................................103
Figure 3.12: PCB Inventories in all lakes organized by latitude. White Lake peak removed for
clarity. 1: From Schneider et al., 2001; 2: From Pearson et al., 1997. ........................................105
Figure 3.13: Focusing-Corrected PCB Inventories versus watershed % urban in the 1970s and
2000s. The top three study lake PCB inventories are labeled illustrating the PCB inventory
independence of % urban .............................................................................................................105
Figure 3.14: Average PCB congener fingerprints for worldwide production (as compiled by
Breivik et al., 2002), Michigan study lakes during maximum production, and Michigan study
lakes after PCB ban ......................................................................................................................109
Figure 3.15: Average PCB congener fingerprints for Michigan inland lakes divided into
subregions: UP = Upper Peninsula, ULP = Upper Lower Peninsula, LLP = Lower Lower
Peninsula. Figure 15a shows the average fingerprint for sediment dated to the time period of
maximum consumption/use of PCBs. Figure 15b shows the fingerprint for sediment dated to the
years since the PCB ban ...............................................................................................................110
Figure 3.16: Cluster analysis of compositional data: Landscape Peak Inventories (1950s to
1980s).......................................................................................................................................... 114
Figure 3.17: Cluster analysis of compositional data: Landscape Recovery Inventories (1990s to
present) ........................................................................................................................................ 115

xi

Figure 3.18: Cluster analysis of all compositional data. Original lake names refer to the lake data
from the historical peak loadings time slice (1950s to 1980s) and lake names followed by a ‘2’
refer to the lake data from the recent time slice (1990s to present) ............................................ 118
Figure 4.1: Location of Saginaw Bay Watershed, Michigan, and sampling sites .......................141
Figure 4.2: Stream hydrographs for the Rifle and Shiawassee Rivers ........................................143
Figure 4.3: Stream hydrographs for the Cass and Tittabawassee Rivers .....................................144
Figure 4.4: Example drainage area calculations for sample sites 9-4, 18-2, 27-8, 28-8, and 29-5 ....
......................................................................................................................................................145
Figure 4.5: Ratio of Na+ to Cl- (moles/L) in all samples collected in the SBW. Simple halite
dissolution would show a 1:1 molar ratio. The brine ratio is represented by 0.61 to 1 which is
typical of Michigan Basin brines (Wilson, 1989) ....................................................................... 159
Figure 4.6: Cl:Br ratios vs Cl- concentrations of all samples collected from the SBW. Lines
indicate potential sources based on Panno et al., 2006 ................................................................165
Figure 4.7: Triplot showing Glaciofluvial aquifer data collected by Wahrer et al., 2006 in the
SBW overprinted with RIV II, RIV III, RIV IV, and RIV V samples. RIV I was not included as
the analytes tested during this event did not include Ca, Mg, Na or K .......................................165

xii

CHAPTER 1. INTRODUCTION

Since the 1800’s, humans have altered ecosystems in North America more rapidly and
extensively than in any prior period of time in human history, largely to meet rapidly growing
demands for food, fresh water, timber, fiber, and energy (Fleischner, 1994; McNeill, 2000).
These demands have stressed ecosystems through activities such as caused by deforestation,
industrialization, urbanization, and other forms of both intentional and unintentional land use
change (Latimer, 2003). These activities have also led to environmental contamination affecting
surface water bodies through nutrient addition, sediment runoff, and the addition of air and
water-borne chemicals, all of which occur simultaneously and during a period of global climate
change (Cohen, 2004).
Efforts are being made to restore perturbed environments to a pre disturbance target or
time from of ‘minimal human impact’ as an established predetermined target (Vile et al., 2000).
Environmental managers are increasingly bound by legal requirements to define restoration
targets (Battarbee et al., 2005, Stoddard et al, 2006). Therefore there to define these targets there
is a need to understand the degree of disturbance and conditions prior to disturbance e.g.,
“reference” conditions (Stoddard et al., 2006).
One method for reconstruction of historical surface water quality and watershed histories
is in the analysis of lake sediment cores. It has been demonstrated that lacustrine sediments
function as one of the best preserved timelines of both watershed and regional environmental
disturbances available in a terrestrial setting (Last, 2001). These sedimentary records contain
chemical records or “signals” of changes in the environment during this time period both from

1

climate fluctuation, natural watershed-scale and regional processes and anthropogenic impacts.
Using these sediment cores provides a means of understanding both anthropogenic change to the
landscape and failure to recover to geochemical reference values despite over 30 years of
environmental legislation (Long, 2010). Such information is critical for land-use planning and to
better understand the effects of land use change on our aquatic environments (Schindler, 2008)
Historically, analysis of lake sediments has served to infer changes in vegetation or assess
the success of environmental laws or management. Lake sediments have also been successfully
employed for analysis of changes in trophic status (Heathwaite, 1994; Brezonik and Engstrom,
1998), lake redox condition by examination of the Fe and Mn sediment profiles (Mackereth,
1966; Davison, 1993) and anthropogenically induced pollution (Norton et al., 1992; and
Engstrom et al., 2007). However, the distribution of anthropogenic chemicals remains poorly
understood and some aspects relating land use and surface water impacts remains elusive (Boyle,
1998). Due to the poorly understood biogeochemistry and toxic nature often associated with
them, trace metals and synthetic chemicals released to the environment remain a source of
concern for environmental managers.
The State of Michigan occupies a large geographic footprint in the Great Lakes region.
The landscape is characterized by a diversity of land use and land cover types. It is a glaciated
terrain and as such contains many inland lakes that hold sediment records from their birth at the
end of the Ice Age (approximately 10,000 years ago) to the present. The combination of the
diversity of land use and lakes affords an opportunity to study and assess environmental
disturbance and recovery from disturbance. Preliminary observations of data collected from
selected inland lakes sediment in Michigan have shown some lakes to recover fully from large
anthropogenic perturbation, some to establish a new geochemical stable state different from its

2

initial state, and some to remain in a state of temporal change (Long, 2010). Questions that can
be asked are: can several decades since the cessation in chemical production of toxic chemicals
provide time enough for landscape recovery of these compounds, if not, can we forecast a time
when these chemicals will not be present?

3

Overarching Research Hypothesis and Approach
Given environmental legislation, erosion control measures, and the variety of measures
undertaken to keep surface water bodies in Michigan pristine, pollution recovery should be
observed in those surface water bodies. This study attempts to provide some elucidation both on
the degree of recovery of these chemicals since their historical peaks in consumption and usage,
and the association of land use to recovery (or lack thereof). The overarching hypothesis for this
research then is: the recovery of surface water systems from environmental contaminants is
driven not only by the production and consumption of the contaminant chemicals but the
physical relationship to land use. To test this hypothesis, three methodologies will be employed:
1.)

Develop temporal patterns of anthropogenic chemicals in relationship to both
their production and consumption peaks and the years associated with their ban.

2.)

Perform degradation half-life calculations when capable.

3.)

Determine the statistical correlation of these chemicals to historical and modern
land use.

These steps are described in more detail in the following sections.

4

General Methodology
Fifty eight sediment cores have been collected from 47 individual inland lakes distributed
throughout Michigan since 1999 as part of the Michigan Department of Environmental Quality
(MDEQ) Michigan Inland Lakes Trends project (http://www.michigan.gov/deq/0,4561,7-1353313_3686_3728-32365--,00.html ; Figure 1, below). Lakes were selected in collaboration with
MDEQ and chosen to cover a large spatial distribution, different land uses and population
densities within the watershed. Collection of sediments was conducted from the deepest portion
of each lake using an MC-400 Lake/Shelf Multi-corer deployed from the Monitoring Vessel
Nibi. The M/V Nibi was designed and has successfully provided access to both major and
remote inland lakes throughout Michigan. The MC-400 is designed to collect four simultaneous
cores from a single location. Collected sediment cores were described and examined for color,
texture, and signs of zoobenthos. Cores were then extruded and sectioned at 0.5 cm intervals for
the top 8 cm, and at 1 cm intervals for the remainder of the core.
Upon collection, sediments were frozen, freeze-dried and digested by nitric acid in a
CEM-MDS-81D microwave (Hewitt and Renyolds, 1990). Standard reference material (NIST
RM 8704 Buffalo River Sediment) and procedural blanks were processed to test for accuracy and
contamination. The concentrated-acid digests were filtered through an acid-washed, e-pure
(Barnstead) rinsed 0.40 μm polycarbonate filter. Samples were then analyzed using a either a
Micromass Platform inductively coupled, plasma, mass spectrometer with hexapole technology
(ICP-MS-HEX)(all analyses prior to September 2013) or, the more recently implemented
Thermo-Fisher ICP-Q (all analyses after September 2013). Sediments were analyzed for a suite
of metals and metalloids that typically included Mg, Al, K, Ca, Ti, V, Cr, Mn, Fe, Co, Ni, Cu,
Zn, As, Sr, Mo, Cd, Ba, Pb, and U. Data was gathered from as many elements as possible during

5

the ICP-MS analysis but are confined to the specific capabilities of the instrument used and the
available elemental standards at the time of analysis. Other elements such as Li, Be, B, Na, Se,
Ag, Sb, Cs, Tl, and Bi were occasionally able to be quantified with the ICP-Q instrument and
data from these elements is made available when possible.
Any freeze-dried sediment remaining from the microwave digestion procedure was
potentially subject to Total Kjeldahl nitrogen (TKN) analysis or Loss-on-Ignition (LOI) analysis.
TKN is designed to test the sediment for the sum of organic nitrogen, ammonia (NH3), and
ammonium (NH4) in the sediment. LOI analysis approximates the organic carbon content of the
sediment. Both analyses could only be performed if enough material was left from the above
described acid digestion procedures as the sediment used for LOI and TKN was taken from the
same core. As such, lakes with this data available vary.
210

Pb was measured on one sub-core from each lake to determine porosity, accumulated

dry mass, sedimentation rates, sediment ages and focusing factors (Freshwater Institute in
Winnipeg, Manitoba, Canada). Several models exist to determine sediment ages from 210Pb
activities, and sediments were dated using the constant flux: constant sedimentation rate model
(CF:CS) (Golden et al., 1993), segmented CF:CS (SCF:CS) (Heyvaert et al., 2000), rapid steady
state mixing model (RSSM) (Robbins, 1982), and the constant rate of supply model (CRS)
(Sanchez-Cabeza et al., 2000). The CF:CS model assumes a constant sedimentation rate
throughout the core. The RSSM model also assumes a constant sedimentation rate, but also
allows for a mixed zone. The SCF:CS model allows for more than one sedimentation rate, and
accounts for a mixed zone. The CRS model determines a different sedimentation rate for each
sample based on the inventory of excess-210Pb. Dating models were verified using 137Cs, stable
Pb peak and presence of excess 210Pb.

6

Sedimentation rates in each lake were determined using all models possible for that lake,
and then the models were evaluated to ascertain which was the most appropriate for determining
sediment ages. There is no consensus as to which model is more appropriate in all cases
(Oldfield and Appleby, 1984), and several factors were considered when choosing a model.
Visual examination of the 210Pb profile gave some insight into the most appropriate model to be
used. The RSSM or CRS models are more appropriate for lakes with large mixing zones, and
the SCF:CS or CRS models are more appropriate for lakes with clear changes in sedimentation.
Additionally, this study uses two other indicators to determine the most appropriate model to
use: profiles of 137Cs activity and stable lead concentration profiles.

137

Cs is an artificial

radionuclide that was produced by atmospheric testing of nuclear weapons in the late-1950s and
early-1960s. The peak level of fallout occurred in 1963, and therefore the peak activity in the
sediment should occur in the early 1960s (Walling and Qingping, 1992). The second indicator is
the stable lead peak. Stable lead has an historical pattern of deposition that is very consistent
among lakes, with lead concentrations increasing from the mid-1800s to the early to mid-1970s,
and decreasing to the present. The peak in lead concentrations in the mid-1970s is consistent
enough to use for dating verification (Alfaro-De la Torre and Tessier, 2002, Callender and
vanMetre, 1997). Excess-210Pb should not be present in sediment slices older than ca. 1850,
therefore, dating models that place sediment slices, containing excess 210Pb, older than ca. 1850
are suspect. The dating model with the most appropriate date for both the 137Cs peak (1963-64),
stable lead peak (early to mid-1970s), and assigned appropriate dates to the presence of excess
210Pb is chosen.
Focusing factors were also determined from 210Pb analysis. Sediment focusing occurs
when fine-grained sediments in a lake are eroded from higher energy erosional zones near the

7

shore of the lake, transported through transitional zones (where deposition and erosion occur
episodically) and deposited in depositional zones (Downing and Rath, 1988, Hakanson, 1977).
This process of focusing occurs to different extents among lakes, and must be accounted for by
using the focusing factor before comparing inventories and accumulation rates among lakes.
Another sub-core was sectioned and submitted to Michigan State Universities Aquatic
Toxicology Laboratory for analysis of organic contaminants. Unlike the metals and 210Pb subcores, the organics core was sectioned at 1 cm increments for the entire core length. There was
insufficient material for analysis in the topmost sediments, so the first two sections were
combined, and the third and fourth sections were combined. The combined samples represent 2
cm intervals through the entire core. The samples were analyzed for PCBs, PAHs, and pesticides
(Khim et al., 1999a, Khim et al., 1999b). A portion of the sediment was dried at 100ºC to
determine moisture content.

8

!
Emily A

A
Gogebic

A

Gogebic

!

Gratiot

Round L

Witch

XF Teal
XAImpEmily X
X Imp
X

± ±

Gratiot

X
X

Witch

X

Round D

D

Round L

X

Round D

D

Mullett

Geneserath F
Charlevoix

Platte

Mullett

E#Charlevoix
k Black

E

Torch
k

^
A
!Elk
^Torch

A !
Crystal B # F Elk
Platte

! Summer 1999
" Summer 2000
Sampling
Year
#

!
^
X"
#
D
E^
lX
AD
E
R

kl
FA

0

Summer 2001

Summer 1999

Summer 2002

Summer 2000

Summer 2003

Summer
2001
Summer
2004
Summer
2002
Summer
2005
Summer
2003
Summer
2006
Summer
2004
Summer
2007
Summer
2005
Summer
2008
Summer
2009
Summer
2006
Summer 2010

55

X

Avalon
X Avalon

Clear

X

#

#
Hubbard
Shupac
Hubbard
Crystal B #k
Shupac
!
D George
Cadillac
Higgins
D George
Cadillac
Hackert
D
Hackert

^

D

X

^Higgins

!k

^Houghton
^Houghton
"
"
Nichols
E E
Littlefield
Littlefield
Fremont
Nichols
Otter Otter
k
Crystal M
l l
White
White
"
"
!
D
Crystal
M
!
D
Muskegon
Muskegon
k Long
Davison
R
E Campau
Morrison
E Campau
R ! CassCass
Whitmore
Thompson
Cora

Paw Paw # R

!

F

A

Gull Duck
!

Paw Paw
#
Birch

GullBird

X RBirch

Summer 2007

110

#

Klinger

R

X

#

A

Thompson
#

XSand

!

Whitmore

XSand

220 Kilometers

Figure 1.1: Project Lakes 1999 to 2010

To investigate the relationship of water quality to present day land use, water chemistry
from selected rivers in the Saginaw Bay Watershed of Michigan were compared to land use. The
comparisons were done synoptically for five flow events that ranging from high flow to low flow
conditions. Comparisons were made using factor analysis.
9

The results of this research are summarized in the following three chapters of this
dissertation. These chapters are:


Chapter 2: Historical Distribution and Recovery of DDT as Interpreted from Michigan
Inland Sediments



Chapter 3: Landscape Recovery from Historical PCB Loadings Using Sediment
Chemical Chronologies



Chapter 4: Seasonal Influence on Land Use Biogeochemical Fingerprints in Surface
Water Bodies

10

REFERENCES

11

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.
Battarbee, R. W., John Anderson, N., Jeppesen, E., & Leavitt, P. R., 2005. Combining
palaeolimnological and limnological approaches in assessing lake ecosystem response to
nutrient reduction. Freshwater Biology, 50(10), 1772-1780.
Boyle, S. J., Tsanis, I. K., & Kanaroglou, P. S. 1998. Developing geographic information
systems for land use impact assessment in flooding conditions. Journal of water resources
planning and management, 124(2), 89-98.
Brezonik, P. L., & Engstrom, D. R., 1998. Modern and historic accumulation rates of phosphorus
in Lake Okeechobee, Florida. Journal of Paleolimnology,20(1), 31-46.
Callender, E., vanMetre, P.C., 1997. Reservoir sediment cores show US lead declines.
Environmental Science and Technology 31, A424-A428.
Cohen, A.S., 2004. Paleolimnology: The History and Evolution of Lake Systems. PALAIOS 19,
184-186(2004).
Davison, W. Iron and manganese in lakes., 1993. Earth-Science Reviews 34, 119-163.
Downing, J.A., Rath, L.C., 1988. Spatial patchiness in the lacustrine sedimentary environment.
Limnology and Oceanography 33, 447-458.
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(6), 2467-2483.
Fleischner, T.L., 1994. Ecological costs of livestock grazing in western North America.
Conservation Biology 8, 629-644.
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.
Hakanson, L., 1977. The influence of wind, fetch, and water depth on the distribution of
sediments in Lake Vanern, Sweden. Canadian Journal of Earth Science 14, 397-412.
12

Heathwaite, A.L., 1994. Chemical fractionation of lake sediments to determine the effects of
land-use change on nutrient loading. Journal of Hydrology 159, 395-421.
Hewitt, A.D. & Reynolds, C.M., 1990. Dissolution of metals from soils and sediments with a
microwave-nitric acid digestion technique. Atomic Spectroscopy 11, 1425-1436
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.
Khim, J., Kannan, K., Villeneuve, D., Koh, C., Giesy, J., 1999a. Characterization and
distribution of trace organic contaminants in sediment from Masan Bay, Korea: 1.
Instrumental analysis. Environmental Science and Technology 33, 4199-4205.
Khim, J., Villeneuve, D., Kannan, K., Lee, K., Snyder, S., Koh, C., Giesy, J., 1999b.
Alkylphenols, polycyclic aromatic hydrocarbons (PAHs), and organochlorines in sediment
from Lake Shihwa, Korea: Instrumental and bioanalytical characterization. Environ
Toxicol Chem 8, 2424-2432.
Last, W.M. & Smol, J.P., 2001. An Introduction to Physical and Geochemical Methods Used in
Paleolimnology. Tracking Environmental Change Using Lake Sediments Volume 2
Physical and Geochemical Methods 1-5.
Latimer, J. S., Boothman, W. S., Pesch, C. E., Chmura, G. L., Pospelova, V., & Jayaraman, S.,
2003. Environmental stress and recovery: the geochemical record of human disturbance in
New Bedford Harbor and Apponagansett Bay, Massachusetts (USA). Science of the total
environment, 313(1), 153-176.
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(1), 143-158.
Mackereth, F.J.H., 1966. Some chemical observations on post-glacial lake sediments.
Philosophical Transactions of the Royal Society of London Series B 250, 165-213.
McNeill, J.R., 2000. Something New Under the Sun: An Environmental History of the
Twentieth-Century World (Global Century Series). Nature 407, 674-675(W.W. Norton &
Company).
Norton, S. B., Rodier, D. J., van der Schalie, W. H., Wood, W. P., Slimak, M. W., & Gentile, J.
H. 1992. A framework for ecological risk assessment at the EPA. Environmental
Toxicology and Chemistry, 11(12), 1663-1672.

13

Oldfield, F., Appleby, P.G., 1984. Empirical testing of 210Pb-dating models for lake sediments.
In: E. Y. Haworth, J. W. G. Lund (Eds.), Lake Sediments and Environmental History.
University of Minnesota Press, Minneapolis, pp. 93-124.
Sanchez-Cabeza, J.A., Ani-Ragolta, I., Masque, P., 2000. Some considerations of the Pb-210
constant rate of supply (CRS) dating model. Limnology and Oceanography 45, 990-995.
Stoddard, J. L., Larsen, D. P., Hawkins, C. P., Johnson, R. K., & Norris, R. H., 2006. Setting
expectations for the ecological condition of streams: the concept of reference
condition. Ecological Applications, 16(4), 1267-1276.
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.

14

CHAPTER 2
TEMPORAL AND SPATIAL PATTERNS OF DDT LOADINGS IN A LARGE
GEOGRAPHIC REGION OF THE GREAT LAKES WATERSHEDT

Abstract
Concentrations of DDT, DDD, and DDE were measured in sediment cores collected from
39 inland lakes in Michigan for examination of spatial and temporal trends in accumulation and
degradation. It was hypothesized that the trends of DDT observed in the sediment cores would
reflect the US production and consumption trends with an increase to peak consumption and use
in the late-1950s and early 1960s followed by a decrease of DDT to the present. While it is clear
that US production and usage are the overriding determinants of the general DDT profiles,
results show significant variation amongst the individual lakes in terms of temporal peak values
and recovery since the ban of these chemicals. This is thought to be the result of both
independent watershed-scale and regional sources of DDT recorded in the lake sediment. Some
correlation to watershed %urban land-use is apparent but land-use characteristics overall were
poorly correlated to DDT concentrations, peak and onset dates. However, DDT and its
metabolites found in recent sediment (top 4 cm), as well as, sediment accumulation rates of these
chemicals in each sediment core are well correlated to latitudes of Michigan with higher
populations, even if the lake watershed itself wasn’t especially populace. This suggests a
significant regional component to DDT both historically and since its ban. However, ratios of
DDT to its breakdown products suggest some lakes maintained watershed-scale sources in the

15

years after the DDT ban. Degradation half-life calculations of DDT in Michigan lakes were
determined to average approximately 20 years, similar to reported DDT half-lives of others.

16

Introduction
Dichlorodiphenyltrichloroethane (DDT) is one of the most well-known and widely
distributed synthetic pesticides. The persistent nature of DDT has prevented complete
elimination from the food web even though production of this chemical has ceased in the US for
decades. In fact, DDT was detected in almost all human blood samples tested by the Center for
Disease Control in 2005 (Eskenazi et al., 2009) and the Food and Drug Administration continue
to detect DDT in routine food tests (USDA, 2006). As the fate of many of these chemicals is to
wash to surface water bodies in the regions in which they are applied, continued work needs to
be completed to ascertain the chemicals availability for absorption by aquatic organisms. Here
we use inland lake sediment cores to provide some elucidation to DDTs continued presence in
sediment dated throughout the chemicals use history in Michigan.
Although first synthesized in 1874 by Othmar Zeidler at the University of Strassburg,
Germany (Zimmerman and Lavine, 1946), it was not until the 1930s that scientists discovered its
insecticidal properties (WHO, 1979). During the 1930’s potent pesticides were sought after for
malaria and crop-pest control. In the autumn of 1939, Paul Muller, working in the laboratories
of Johann Rudolph Geigy in Basle, Switzerland synthesized the same compound was originally
developed by Othmar Zeidler years before (Zimmerman and Lavine, 1946; West and Campbell,
1952; Brooks, 1974). In 1942, the Geigy Company in New York received its first shipment of
DDT from Geigy, Switzerland and then submitted DDT to the United States Department of
Agriculture for testing. DDT was used extensively during the Second World War among allied
troops and proved effective in controlling diseases such as malaria and typhus (ATSDR, 2002).
By 1943, DDT was in commercial production at the Cincinnati Chemical Works, and in early
1944, DuPont, Merck, and Hercules Powder Company also went into commercial production of

17

DDT (Zimmerman and Lavine, 1946). Aside from its effectiveness in controlling biting insect
populations that spread disease, DDT was also widely used as a form of agricultural pest control.
From 1945 until it was banned, DDT was one of the most widely used pesticides for the control
of insects on agricultural crops (ATSDR, 2002).

Figure 2.1: Domestic Production and Consumption of DDT in the United States, 1950-1972. The 1975 EPA
report entitled DDT, A Review of Scientific and Economic Aspects of the Decision To Ban Its Use as a Pesticide
analyzed domestic production and consumption data which showed peak production and consumption of
DDT to be in 1963 and 1959 respectively. Source: EPA-540/1-75-022, p. 140

By 1959, the US use of DDT peaked with approximately 35,771 tons being used annually
after which use began to decline until its ban in 1972 (USEPA, 1975; WHO, 1989)(Figure 1).
The decline following the peak use in 1959 was the consequence of growing knowledge amongst
users of its human toxicity (first reported in 1946) and by the growing number of ecological
studies of its effects on the populations of fish and birds of prey (USEPA, 1975; WHO, 1979).
In the United States, the actual ban of DDT was initiated in 1962 with the publication of Rachel
Carson’s book Silent Spring. The book cataloged the environmental impacts of environmental
DDT distribution and warning of unforeseen affects to ecological and human health as a result.
This led to increased environmental awareness of persistent chemicals and, in combination with
the growing scientific evidence of DDT persistence and harmful traits, resulted in political
18

responses banning its use in the Great Lakes basin. As of January 1, 1973, all uses of DDT in
the United States were largely suspended with the exception of emergency public health uses
(Meister and Sine, 1999; ATSDR, 2002). Although DDT was produced by companies in Mexico
and China well into the 2000’s, no companies in the United States manufacture DDT for use in
other countries (Meister and Sine, 1999). Amongst other countries using the chemical, the actual
date of the ban of DDT varies substantially, with some countries continuing production into the
21st century. As of this writing, it believed that India is the only country still manufacturing
DDT and is also its largest consumer (Van Den Berg, 2009; Stockholm Convention, 2010).
Since the ban of DDT, scientists have measured the declining environmental
concentrations in studies regarding the Great Lakes in the tissues of birds and fish, as well as
lake sediment. These studies have shown that the concentration of DDT observed in the
environment follow closely the production and use history (Bierman and Swain, 1982; Oliver et
al., 1989). Once in the environment, breakdown of DDT occurs slowly as it was formulated to
be highly stable. Literature review of DDT half-life of in soil is exceptionally broad, with early
research reporting half-lives on the order of 3-5 years with loses caused by metabolism to DDE
and DDD, vaporization or erosion (Bierman and Swain, 1982; Matsumara et al., 1971; Hiltbold,
1974; Voerman and Besemer, 1975). Later research showed that DDT remains in soil and
sediment much longer than initially reported and are lost on the order of several decades through
metabolism and vaporization (Oliver et al., 1989; Dimond and Owen, 1996; Harris et al., 2000).
In surface water bodies, degradation and loss of DDT includes sedimentation,
volatilization, photolysis, and both aerobic and anaerobic biodegradation (ATSDR, 1994). The
environmental stability of DDT stems from its strong carbon-chlorine covalent bonds which is
resistant to enzymatic breakdown by microorganisms. Persistence is further perpetuated by the

19

fact that DDT has low aqueous solubilities (Fiedler and Lau, 1998). As DDT has a strong
tendency to adsorb to surfaces, it often enters aquatic systems already attached to soil or organic
matter particles (Corolla, 1945). It has been observed that particulate settling to be responsible
for 61-95% of the DDT loss rate in the Great Lakes (Bierman and Swain, 1982) and it is
assumed that particulate settling to play a large role in loss of DDT from inland lake water
column as well. This allows study of lake sediment to reflect changes in use of many pesticides
and act as a chemical tape recorder of its presence or use in a given watershed.
Common breakdown products of DDT include include dichlorodiphenyldichloroethylene
(DDE) and dichlorodiphenyldichloroethane (DDD), which are also highly persistent and have
similar chemical, physical, and toxicological properties (WHO, 1989). DDE has no commercial
use but DDD has been used as an insecticide for contact control of leaf rollers and other insects
on vegetables and tobacco and is no longer produced commercially. DDE is the result of
dehydrochlorination of DDT. This oxidative process is considered the primary aerobic pathway
for dehalogenation of DDT (Mohn and Tiedje, 1992). Aerobic dehydrochlorination of DDT
simultaneously removes the hydrogen and chlorine from the aliphatic portion of the molecule. A
number of studies conducted in the 1960’s and 1970’s have found that DDE was the major
degradation product of DDT in such aerobic mediums as ocean water (Patil et al., 1972) and
aerobic soils (Guenzi and Beard, 1968). Under reducing conditions, reductive dehalogenation is
the major mechanism for the conversion of DDT to DDD (Fries et al., 1969). This is considered
the primary degredation pathway of DDT under anaerobic conditions such as in lake or ocean
sediments (Jensen et al, 1972; Zoro et al. 1974). Reductive dehalogenation effectively
dechlorinates the DDT compound by the removal of a chlorine atom directly from the biphenyl
ring and adds a hydrogen atom in its place. In anaerobic sediments, such as the application here,

20

this process is usually limited by the availability of electron acceptors. The process may slow
upon depletion of the more common natural electron acceptors (e.g. NO3-, SO4 2-).
As with many pesticides, these processes of transformation of DDT to a less chlorinated
product also results in a reduction in toxicity of the chemical (Mohn and Tiedje, 1992) and this is
reflected in the sediment quality guidelines and threshold effect levels published by many
agencies for these chemicals (Burton, 2002). However, given the toxicological nature and the
environmental persistence, DDD and DDE are still widely considered toxic compounds, and, as
such, the conversion of DDT to either DDE or DDD in sediments cannot be considered as a
detoxification step (ATSDR, 2002). For these reasons and their known continued presence, both
DDD and DDE were measured along with DDT for this project.

21

Hypothesis and Test
In general, the importance of this study lies in the persistence of organochlorine
contaminants in the environment and in understanding their fate since regulations controlling
them have developed. This is particularly important with DDT with which its hydrophobicity
and solubility in lipids allow it to accumulate to harmful concentrations in biota. This study
seeks to answer questions related to an on-going legacy with DDT in the environment: is the
chemical still present in measurable amounts? If so, can we project the trends to a date when it
will not be present? It is hypothesized that the sediment profiles of DDT observed in the inland
lake sediment would reflect the U.S. consumption and production of these chemicals. Peak dates
of production/consumption of DDT were 1962 and 1959, respectively (Figure 1). These peaks
should be mimicked by the temporal concentration and accumulation rate profiles of the
sediment cores, with increasing values in the years prior to these peaks and progressively lower
values after these peaks.
To test this hypothesis, sediment cores from 39 inland lakes were collected from across
the State of Michigan representing lakes within various types of land use, watershed
characteristics, and trophic status. Once collected the sediment cores were taken ashore where
they were divided into 2 cm segments, stored on ice for transportation to the laboratory, and
concentrations of DDT, DDD and DDE were then measured. Additional sediment cores were
processed similarly and used for radiometric dating via 210Pb and metals analysis. DDT, DDE,
and DDD concentrations, total inventories, and, with the addition of 210Pb results, accumulation
rates of PCBs were then calculated for all lakes sampled.
If the hypothesis proves true: 1) temporal DDT, DDD and DDE concentration and
accumulation rate profiles would peak close to the production/consumption timeframe described

22

in the 1975 USEPA report, 2.) increasing trends should be observed in sediment dated from the
1940’s time period to the production/consumption peak, and 3.) a decreasing trend in DDT
should be observed in the years following the production/consumption time frame, both as the
chemical itself is phased out of use and the watersheds in which it is applied hydraulically
remove the chemical.

23

Methodology
Study Setting:
The State of Michigan covers 58,528 sq miles of which land takes up 56,954 sq miles and
inland water approximately 1,573 square miles. Sediment cores were collected from 39 lakes
throughout the State of Michigan from 2000 to 2010 (Figure 2) as part of the Michigan
Department of Environmental Quality’s (MDEQ) Inland Lakes Sediment Trend Monitoring
program. The number of inland lakes in Michigan exceeds 10,000 (MDNR, 2013) but lakes
were chosen for sediment sampling based on a set of criteria including: available boat access to
the lake, sufficient depth to preclude significant bioturbation of sediment (generally lakes >25
feet depth), avoidance of lakes potentially mined for marl, and maximizing spatial range of the
project.
To maximize the geographic extent of the project, sediment core collection ranged from
inland lakes located along its southern border (e.g. Klinger Lake, Latitude: 41.80) to lakes
located at its northern extent (e.g. Gratiot Lake, Latitude: 47.36) (Figure 2). This wide
geographic distribution of lakes sampled represents a variety of watershed land use types ranging
from largely urban near major cities (e.g. Cass Lake, Oakland County) to those that are remote
with watersheds almost completely forested (e.g. Emily Lake, Houghton County). Further, lakes
were sampled to represent a large range of physical characteristics. These include but are not
limited to: physical size (depth, area), watershed size, proximity to Great Lakes, and trophic
status. Table1 is included at the end of this report detailing a list of the lakes sampled and some
physical characteristics for this research.

24

Figure 2.2: Locations of study lakes sampled from 1999 to 2010.

Sample Collection:
Four replicate sediment cores from 40 to 59 cm in length were collected from the deepest
portion of each lake using a MC-400 Lake/Shelf Multi-corer deployed from either the EPA
Research Vessel Mudpuppy or the MDEQ Monitoring Vessel Nibi. One entire core was
dedicated to analysis of a suite of organic compounds including polychlorinated biphenyls
(PCBs), polycyclic aromatic hydrocarbons (PAHs) and a suite of pesticides which included
DDT, DDD, and DDE. In the field, each core was extruded and sectioned into two centimeter
intervals through the entire core. Once sectioned, the samples were placed on ice until submittal
to Aquatic Toxicology Laboratory at the Department of Animal Science at Michigan State
University. Samples were stored at -20°C until analysis. Analytical methods are described in

25

Khim et al., 1999. A separate core from each lake was sectioned and submitted to the
Freshwater Institute in Winnipeg, Manitoba, Canada, for measurement of 210Pb to determine
sedimentation rates, sediment ages, and focusing factors. The use of 210Pb dating for these cores
in discussed in detail elsewhere (Yohn et al., 2004).

Chemical Analysis:
Analytical methods for DDT and other organic contaminants have been described
elsewhere (Khim et al., 1999). Briefly, 40 g sediment (wet) was homogenized with 160 g of
anhydrous sodium sulfate and extracted with methylene chloride and hexane (3:1, 400 mL) in a
Soxhlet apparatus for 16 h. Sulfur was removed by treating the extracts with activated copper
granules. Extracts were then concentrated and passed through a glass column packed with
activated Florisil (10 g). First fraction, eluted with 100mL hexane contained p,p’-DDE . Second
fraction eluted with 100 mL of 20% methylene chloride in hexane contained p,p’-DDD and p,p’DDT. TCMX and PCB 30 were spiked as internal standards to check for recoveries of analytical
procedure. Pesticides were quantified on a Hewlett-Packard 5890 series II gas chromatograph
equipped with a 5972 series mass spectrometric detector. The MS was operated in selected ion
monitoring mode by use of the molecular ions selective for the individual compounds. Other
QA/QC procedures include matrix spikes and analytical blanks. DDT could not be quantified in
Black, Clear, Whitmore07 Lakes due to excessive interferences (likely sulfate) and DDT samples
collected from White and Muskegon Lakes were not able to meet adequate QA/QC recovery.
Age Dating:
The 210Pb profiles for these sediment cores varied greatly lake to lake. This variation can
be attributed to the land use changes altering the rates in which erosive material is washed to the

26

lakes and nutrient-driven trophic changes. In general, lakes exhibited a relatively stable
background prior to the mid-1800’s when deforestation began to impact erosion within the
watershed, significantly increasing the sedimentation rates of the lake. The specific timing and
scale of the increase in sedimentation rate is addressed on an individual lake basis with a best fit
model. Each model applied to a 210Pb profile was evaluated as to its appropriate fit with 137Cs
and mass spectrometer determined stable lead peak as independent age markers. The 210Pb determined average sedimentation rate is offered with other lake characteristics in Table 1. The
oldest (deepest) dated sediment within the cores varied from A.D. 890 (Lake Geneserath) to
1972 (Morrison Lake). Despite the wide range of sedimentation rates, most sediment cores were
able to reach sediments dated to the early 1800’s. It should be noted that sediment deeper than
the presence of excess- 210Pb cannot be directly dated and dates older than this were determined
via extrapolation using the assumption that sedimentation rates remain constant below this depth.
As mentioned, this study uses two other indicators to assist in determining the most
appropriate model to apply to the 210Pb data. These indicators include profiles of 137Cs activity
and stable lead concentration profiles (208Pb) of which the peak dates are known and ubiquitous
amongst lakes in this region.

137

Cs is an artificial radionuclide that was produced by atmospheric

testing of nuclear weapons in the late 1950s and early 1960s. The peak level of fallout occurred
in 1963, and therefore the peak activity in the sediment should occur in the early 1960s (Walling
and Qingping, 1992). The second indicator is from mass spectrometry analytical results for
stable lead (208Pb). This stable lead isotope has an historical pattern of deposition that is very
consistent among lakes, with lead concentrations increasing from the mid-1800s to the early to
mid-1970s, and decreasing to the present. The peak in lead concentrations in the mid-1970s is
consistent enough to use for dating verification (Alfaro-De la Torre and Tessier, 2002, Callender

27

and VanMetre, 1997). Excess-210Pb should not be present in sediment slices older than ca. 1850,
therefore, dating models that place sediment slices, containing excess 210Pb, older than ca. 1850
are suspect. The dating model with the most appropriate date for both the 137Cs peak (1963-64),
stable lead peak (early to mid-1970s), and assigned appropriate dates to the presence of excess
210

Pb is chosen.

28

Results and Discussion
Spatial DDT/DDD/DDE Concentrations, Inventories, and Accumulation Rates:
DDT or it breakdown products (DDD and DDE) is detected in all 39 study lakes. While
sample concentrations vary considerably, nearly all elevated values are noted in lower portion of
Michigan (Figure 2). Whitmore, Paw Paw, Otter, Davidson, Cass, and Cora Lakes contain the
highest peak concentrations close to 1960 for each lake except Davidson, which has a peak
concentration in 1974 (Figure 3). Overall, the highest concentrations of any individual sample
analyzed for DDT, DDD, and DDE is observed in Lake Cora with values of 366, 1456 and 1064
ng/g dry weight (respectively) occurring in sediment dated to 1966. Lowest concentrations
amongst the individual study lakes are those of Lake Gogebic (Upper Peninsula) with only DDE
detected at its highest value of 1.19 ng/g dry weight in sediment dated to 1983.

Figure 2.3: Temporal concentration profiles for six study lakes with highest peak total DDT in order by peak
concentration value.

29

Concentrations of DDD, DDE, and DDT were summed in the top two samples collected
from each core (Figure 3). These measurements are of particular concern as concentrations from
the top sample intervals most directly affect benthic organisms. The concentrations are shown in
a south to north distribution by latitude of the coring sites in Figure 4. Most populous latitudes
(as determined by U.S. Census Bureau 2000 Census Data) are shown in the grey portion of the
graph. Visual inspection of the graph shows relatively low values of DDT and breakdown
products in the four southernmost lakes, followed by elevated values in the more populous
lower-middle and middle latitudes of Michigan. The upper latitudes of the Lower Peninsula of
Michigan and the Upper Peninsula generally show reduced values relative to the populace
centers of the State located in the lower and middle latitudes.
The relatively reduced concentrations of DDT or its products in the northern Lower
Peninsula lakes could be a function of the ultra-oligotrophic nature of these lakes. The reduced
water column nutrient concentrations in these lakes results in reduced productivity (e.g.
phytoplankton growth) which drives a portion of the sedimentation rate. With a reduced flux of
this organic material that these chemicals are often sorbed to, it is expected that DDT is more
likely to avoid sedimentation in these lakes (e.g. Elk, Crystal).

30

47.22º N
North

41.80º N
South

Figure 2.4: Summed concentrations of DDT and its breakdown products (DDD and DDE) in the 0-2 cm and
2-4 cm sediment depth intervals for Michigan inland lakes arranged by latitude.

Many of the lakes with elevated surficial DDT results (Cora, Davidson, Otter, Hackert,
Higgins and Emily) also show an increase from the 2-4 cm interval to the 0-2 cm interval.
Overall, this increasing trend in sediment concentration between the two samples is observed in
21 of the 39 lakes studied. However, interpretations regarding concentrations can be misleading
without accounting for changes in sedimentation rate and the differences in the physical
characteristics of the sediment both temporally through the core and between individual study
lakes (e.g. quantity of organic material versus clastic material). Because of these differences,
DDT concentration chronologies are difficult to compare between cores. For example, if over
time there is an overall increase in total sedimentation rate, but the flux of DDT to the sediment
remains the same, the concentration patterns in the sediment core will appear to be reduced.
Thus, supplemental to temporal DDT chronologies based solely on concentration data, the
sources and pathways of these chemicals to lake sediment are supported with inspection of the
spatial trends of DDT accumulation rates, which accounts for the changes in sedimentation rate,

31

and inventories, which is the total mass of DDT in the core per unit area. Accumulation rates of
DDT and its breakdown products are calculation made based on the measured concentration of
the pesticide in a given sample with the 210Pb-determined mass sedimentation rate.
Accumulation rates of DDT and its breakdown products were calculated as follows:

Accum (ng / cm 2 / yr )  Csed (ng / g )  W ( g / cm 2 / yr )
Where Accum = Pesticide accumulation rate (ng/cm2/yr);
Csed= concentration of pesticides in surficial sediment (ng/g dry wt);
and W= mass sedimentation rate (ng/cm2/yr) based on 210Pb dating.

The accumulation rates are then focus-corrected (calculated as accumulation rate divided
by a 210Pb-determined focusing factor) (Golden et al., 1993). Sediment focusing of fine sediment
and associated chemicals is a process of redistribution laterally from non-depositional zones to
depositional zones via wave or wind energy (Golden et al., 1993). Sediment focusing occurs in
different intensities largely based on lake morphology. This prevents easy comparison of the
chemical trends between lakes. To accommodate this, the focusing factor is calculated as the
ratio of the integrated 210Pb inventory measured in the sediment to the expected inventory from
210

Pb deposition from the atmosphere (15.5 pCi/cm2 ; from Golden et al., 1993).
The accumulation rates of the DDT summed with DDE and DDD for the top two samples

from each lake is shown as Figure 5, the average accumulation rate for Total DDT in the whole
core is shown in Figure 6, and the spatial distribution of average accumulation rates is shown
state-wide as Figure 7. Overall, the south to north pattern of average accumulation rates both in
the top samples and the whole core is similar in distribution to the average surficial concentration
data shown in Figure 5. Again, it appears that the populace middle and lower-middle portions of
the State harbor the most elevated values.
32

Lake Geneserath accumulation data shows elevated accumulation rates of total DDT in
the top two samples of this lake relative to lakes of similar latitude (Figure 5). This lake is
located on Beaver Island approximately 30 miles northwest of mainland Michigan and has a
small watershed with low population. Because of its location and watershed size, this lake is
expected to very low atmospheric inputs of DDT. However, this lake contains anomalously high
accumulation rates in surficial sediment with an apparent increasing trend to the surface.
Examination of the remaining portions of the core reveals that Lake Geneserath has an
anomalously low sedimentation rate in the deeper portions of the core. Thus the top samples
represent most of the environmental history of the lake and its watershed. As a result, analysis of
the top samples from this lake is misleading when comparing to other lakes. The difference is
reflected in the whole core average accumulation rate (Figure 6), where the difference between
this lake and others of similar latitude are not as profound.

Figure 2.5: Summed focus-corrected accumulation rates of DDT and its breakdown products (DDD and
DDE) in the 0-2 cm and 2-4 cm sediment depth intervals for Michigan inland lakes arranged by latitude.

33

Figure 2.6: Average focus-corrected accumulation rates of DDT and its breakdown products (DDD and DDE)
since 1940 for Michigan inland lakes arranged by latitude.

Figure 2.7: Spatial distribution of average focus-corrected accumulation rates of DDT and its breakdown
products (DDD and DDE) since 1940 for Michigan inland lakes arranged by latitude.

To further examine the spatial distribution of DDT loadings, focusing-corrected
inventories are calculated for each study lake and compared. DDT inventories represent the

34

mass sum of the chemical per unit area through its entire environmental history as represented in
the sediment core and are considered independent of individual lake trophic status and
sedimentation rates. The utility of determining DDT inventories is to ascertain the degree at
which atmospheric deposition of these compounds play a role in their distribution. If local
sources alone drive the DDT flux to the surface water bodies, DDT inventories would be
expected to have significant inventory differences in watersheds located close to one another. If
atmospheric deposition plays a large role in DDT distribution, regional scale inventory gradients
should dominate the spatial distribution.
Inventories of DDT and its breakdown products in sediment cores are calculated as:

Inv(ng / cm 2 )  i [Csed (ng / g )  (1   )   ( g / cm 3 )  d (cm)]
Where Inv= total chemical dry mass in a core (ng/cm2);
Csed = sediment concentration;
φ= porosity;
ρ = bulk density;
d = thickness of sediment increment.
The spatial distribution shown in Figure 8 appears to reflect a signal generated by local
sources overprinted by a component of regional atmospheric inputs. Like the concentration and
accumulation rate data, the study lake inventory appears to generally correlate with population
within Michigan. Lakes along the Indiana/Ohio border show relatively low inventories
compared to those in the more populace latitudes but not as low as some of the lakes in the UP or
in northern lower Michigan. This could be interpreted as indication of extra-watershed inputs
from upwind sources such as the areas near Chicago, Kalamazoo, and Lansing. Supporting this
interpretation is other remote lakes with measurable DDT inventories such as Lake Geneserath
35

and Emily Lake which, without atmospheric inputs, would not likely contain any measurable
DDT within their sediments. Finally, Teal Lake (Figure 2) is noted as representing an
anomalously high inventory for the Upper Peninsula of Michigan. However, the Teal Lake
watershed contains the city of Ishpeming and is generally much more populace than the other
lakes studied in the Upper Peninsula.

Figure 2.8: Spatial distribution of focus-corrected DDT inventories for Michigan inland lakes arranged by
latitude.

36

Figure 2.9: Focus-corrected DDT inventories for Michigan inland lakes arranged by latitude.

Temporal DDT/DDD/DDE Trends
On average, temporal changes in concentration of DDT, DDD, and DDE follows the
hypothesized trend of an increase to peak production/consumption followed by a decrease in the
decades following. Figure 2.10 and 2.11 illustrate the decadal averages of DDT concentrations
state-wide with a noted increase in the 2000’s decade sediment relative to that of the 1990’s.
Given that some of the most impacted lakes with regard to DDT exhibit this increase (e.g. Cora
Lake), the data is normalized to peak value for comparison in Figure 2.11. Here a continued
decrease is observed, however slight, in the 1990s and 2000s decadal concentrations (normalized
values of 0.41 and 0.37, respectively). The 1990s and 2000s decadal average concentrations
represent a large reduction in the rate of recovery. Possible reasons for this include bioturbation,
atmospheric transport of DDT, diffusion, and benthic recycling. These processes are discussed
in Temporal DDT/DDD/DDE Trends below.

37

Figure 2.10: Decadal averages of Total DDT concentrations of 39 lakes in Michigan.

Figure 2.11: Normalized decadal averages of Total DDT concentrations of 39 lakes in Michigan.

Wong et al., 1995 and Eisenreich et al., 1989 conducted studies in which sediment cores
were collected from within Lake Ontario. A DDT trend similar to findings in this study was
found in these cores with peak value reflecting peak production/consumption and recent
sediment maintaining measurable concentrations to the present. However, on an individual lake
basis in this study, some variation in the DDT profiles is noted. Many lakes did not exhibit some
aspect of the hypothesized temporal trend or those observed by Wong or Eisenreich. The
differences between those lakes exhibiting an excursion from the hypothesized trend and studies
conducted in the Great Lakes lies in the difference in scales. The Great Lakes watershed
38

occupies a massive area of the US Midwest and Ontario, Canada, with mixing that prevents any
single DDT release event from manifesting itself within the sediment record. A normalized
average of DDT concentrations in all the study lakes across Michigan provides a signal similar in
many respects to that reported by both Wong and Eisenreich because it’s integrating DDT
concentrations across a large area, similar in size to a Great Lake watershed. As the individual
study lakes here occupy watersheds much smaller than those of the Great Lakes, these systems
are much more prone to recording local inputs of DDT. As a result, section will describe the
discrepancies from the hypothesized trends in three parts: DDT onset trends, DDT trend peak,
and recent DDT trends.

DDT Onset Trends:
It is expected that lake sediment dated to time before significant use of DDT in the
United States (mid-1940’s) to contain no detectable concentrations of DDT. However, 16 of the
39 lakes studied show low but detectable concentrations of DDT in the 1930s dated sediment or
earlier. This could possibly be attributed to the smearing effect occurring upon corer insertion
into the benthic sediments. However, measures are taken to remove sediment in direct contact
with the coring tube and the coring tube assembly is cleaned between lakes. The field
methodology designed to prevent this makes these early detections seem improbable.
Post-depositional diffusion of the pesticides offers another potential pathway for this
phenomenon. Based on calculations made by Wu and Gschwend (1988), the effective porous
media diffusivity for compounds such as DDT controls their flux in lake sediments. Eisenreich
et al. (1989) used calculations developed by Thomann and Di Toro (1983) and from Geankoplis
(1972) to calculate diffusion fluxes in chlorinated hydrocarbons in Lake Ontario sediment.

39

Diffusion rates of 2.3 x 10-9 cm2/s were calculated for DDT (approx. 0.06 cm2/yr). These values
match those calculated here using the equation put forth by Thomann and Di Toro (1983).
Based on calculations made by Wu and Gschwend (1988), the effective porous media diffusivity
for compounds such as DDT controls their flux in lake sediments. Eisenreich et al. (1989) used
calculations developed by Thomann and Di Toro (1983) and from Geankoplis (1972) to calculate
diffusion fluxes in chlorinated hydrocarbons in Lake Ontario sediment. Diffusion rates of 2.3 x
10-9 cm2/s were calculated for DDT (approx. 0.06 cm2/yr). These calculation match calculations
here using the equation put for by Thomann and Di Toro (1983).
Deff = (Dmol φ2) / (1 + Kp ρb / φ)
where Dmol = is the aqueous solution molecular diffusivity (cm2/sec);
φ is sediment porosity;
φ2 is a measure of the tortuous path of solute migration in a porous medium;
Kp is the solid-solution partition coefficient (cm3/g); and
ρb is the sediment bulk density (gsolid/cm3water).
Assuming Dmol is approximately 32 cm2/yr; as calculated from the Wilke-Chang and OthmerThaker equations (Geankoplis, 1972)
ρb is averaged from all lakes 0.3 g/cm3;
φ is averaged from all lakes to 0.93; and
Kp is estimated at 2.3 x 103 cm3/g after Eisenriech et al., 1989.
then Deff is approximately 0.04 cm2/yr
This calculation results in approximately 2.1 cm of potential movement from the onset of DDT
downward to older sediment.

40

Another calculation proposed by Edgington et al., 1991 shows similar results. Penetration depth
(z) is calculated here from the equation:
z = (4 x Dc x t)1/2
where Dc = effective diffusion coefficient in pore water (cm2/year); and t, time (46 years, average
time between sample collection date and the maximum production year of 1959).
Using the above calculated diffusion coefficients for DDT-related compounds of 1.3 x
10-9 cm2/s (0.04 cm2/year) these calculations account for approximately 2.7 cm of downward
movement of DDT in the sediment cores. This calculation can account for several of the early
detections of DDT and related compounds but is too small a distance to account for all.
Other possible influences on the early detection of DDT is in the mixing of sediments
caused by feeding of oligochaete worms on sediment particles on which the DDT compounds are
sorbed. These worms have been shown to play an important role in the mobilization of
contaminants (Karckhoff and Morriss, 1985). Biological disruption such as oligochaete worms
have the potential to rapidly distribute contaminants (mm/day as opposed to mm/month for
diffusion) if a significant worm population is present. Oligochaete worms and other benthic
organisms serve to smear the sediment record of DDT with a similar effect as diffusion: a
reduced peak value but earlier onset dates and later recovery dates. Some correlation is observed
between lakes displaying a significant mixing zone and early detections of DDT (Figure 12,
Table 2). It is also noted that early detections of DDT occur exclusively in mesotrophic and
eutrophic lakes (e.g. Cadillac, Campau, Emily, Klinger Hackert, Paw Paw, Teal, Thompson, and
Long Lakes) and not observed in those lakes classified as oligotrophic (e.g. Black, Clear,
CrystalB, Elk, Shupac) (See Fuller and Minnerick, 2008 for trophic classification) lending some

41

credibility to the argument that early detections are a function of an biological-sourced smearing
effect.

DDT Peak Values, Spatial and Temporal Relationship
The average total DDT concentration in lake sediments nearly tripled in value from the
1940s average total DDT concentration of 57 ng/g to an average value of 146 ng/g in sediment
dated to the 1950s. This is expected as DDT saw its first significant widespread use during this
time period and these decades overall saw the largest increases in total DDT concentrations.
Another increase is then observed to the peak decade (1960s) reflecting the period of peak
production and consumption of the chemical in the United States (Figure 13). Specifically, the
dates of the peak concentration of total DDT from the 39 lakes is 1966 with an average peak total
DDT value of 229 ng/g (Table 3). This is followed by a decrease in the years following the peak
production/ consumption to a reduced set of values to the present. On average, the
concentrations detected in the study lakes confirm the hypothesis that the sediment core records
should reflect the historic usage patterns of these chemicals.

Figure 2.12: Excess 210Pb activity and 137Cs activity versus depth in Teal Lake, Michigan.

42

On an individual basis, nearly all the lakes studied elicited a definitive peak value in the
trends of DDT, DDD, and DDE amongst the sectioned samples of the sediment core. No clear
latitudinal trends are recognized from the peak date data. However, there appears to be a delay
in the peak values for those several lakes located at higher latitudes (Figure 13). It is possible
that the delayed peak values for nearly all lakes studied represents the lag time in the erosion of
soil and organic particles the DDT is bound to from its location on the watershed to the lake
sediment. This seems unlikely as half-life results in soil studies are typically shorter than those
of lake sediment (Varca et al., 1994; Gustafson, 1989; Andrea et al., 1994).
An alternative explanation may lie in post-depositional processes of DDT and its
association with organic material susceptible to decay. If an association exists between organic
material and DDT (or its congeners), it’s possible that these compounds have a post-depositional
surface enrichment similar to the findings in regard to trace metals in El Bilali, 2002. Here and
in other works (Gobeil et al, 1999) it was described that many trace metals are subject to a
mechanism where they are released from the Fe and Mn oxides or organic material they are
sorbed to upon burial as a consequence of two processes: organic matter decay and oxide
dissolution. These metals can be subject to rerelease into the water column where they are
resorbed to other settling materials or released to porewater where, upon continuing
sedimentation, compaction causes upward migration. Either recycling mechanism resulted in
apparent surficial enrichment of the metals in the sediment, not increased loadings. It seems
possible that a similar process could be affecting DDT. Since any additional loading of the
chemical after its ban in the early-1970s is highly unlikely, it is plausible that this “recycling”
process in surficial sediments may cause these apparent late peaks in concentrations.

43

Figure 2.13: Year of peak concentration for DDT in inland lakes of Michigan.

DDT Trends in Recent Sediment
An increase in DDT concentrations in recent (2000s dated sediment) relative to the 1990s
dated sediment is noted in 14 of the 39 lakes studied. Several other lakes are noted as having
equivalent or only marginal decreases between the 1990s sediment and the 2000s dated sediment
(Table 2). It is known that, despite a worldwide ban on DDT for any use but disease vector
control, continued use is described in India, North Korea, and possibly elsewhere (Imphal Free
Press, 2008; van den Berg, 2009). Since DDT has been banned in North America and
considering the magnitude of present fluxes to Michigan surface water bodies, these residues
could be attributed to the result of atmospheric transboundary transport. However, it seems
unlikely that small usage of these chemicals from areas of great distance from Michigan could
generate the atmospheric deposition to produce such a significant increase in concentration in
recent sediments. Other research has shown benthic recycling (Baker et al., 1991) and past
44

impacted sites such as Michigan fruit orchards (Hermanson et al., 2007) to continue to be
significant sources of DDT into the recent.

Sediment DDT Degradation and Half-Life Calculations:
The large reduction in the recovery rates observed in many of the lakes provide cause for
an investigation of the relationship between existing quantities DDT in the sediment with the
quantities of it breakdown products, DDD and DDE. Rapaport et al. (1985) made measurements
of DDT in rain and snow in eastern North America and reported DDT inputs of atmospheric
transboundary transport from neighboring countries where DDT was not banned until a much
later date. An attempt to quantify the amount of this “new” DDT that has entered Michigan
surface water bodies is made by examination of the DDT/DDE+DDD ratio trends through time.
Similar methods has been used by others (Harner et al., 2004; Gingrich et al., 2001) to determine
proximity to source and relative age of organochlorine pesticide deposition. In Harner et al.,
2004, low values of DDT/DDE were interpreted as an aged DDT source such as previously
treated agricultural soil. Others have shown the relatively low ratios of DDT to its breakdown
products (generally < 1.0) to be prevalent in areas likely to receive inputs of these chemicals by
long range atmospheric transport such as at Integrated Atmospheric Deposition Network sites
(Cortes et al, 1998, Audette et al., 2008) and the Canadian Arctic (Halsall et al., 1998). In
comparison, elevated ratios (generally >1.0) have been reported in regions in which DDT has
recently been applied such as Belize, Central America (Alegria et al., 2000).
For this study, results of these calculations are generally consistent with the manner
expected since applications of this chemical have been illegal for many decades. The
proportions of DDT to DDE + DDD decrease with time in nearly all lakes studied, and the

45

smallest ratios (reduced DDT concentrations in relation to elevated DDE+DDD concentrations)
occurred most often close to the sediment-water interface (younger sediment). The temporal
changes in the ratio of DDT to DDE+DDD for selected lakes are shown in Figure 14. Duck
Lake in Figure 14a displays the trend indicative of most lakes studied with an overall decreasing
ratio of DDT to its breakdown products from older to recent sediments. However, several lakes
showed a reversal of the trend at periods of time after the ban of these chemicals. Examples of
this are shown in the profiles of DDT to its breakdown products in 1990s dated sediment of
Morrison Lake and Cass Lake (Figure 14b and 14c, respectively), where a brief event occurred
in the watershed history causing a “peak” in the DDT/product ratios. This may be indicative of a
new source of DDT to the watershed or possibly the disturbance of watershed features (e.g. soils)
containing DDT that is then hydraulically transported to the lake. Since lakes exhibiting this
trend do not lie adjacent to one another these trend reversals in the DDT/product ratios is
interpreted to be sourced at a watershed (local) scale. Interpretation of peaks in the DDT ratio
trends is difficult due to the coarse resolution of the sampling intervals but it appears at least
three other lakes exhibit this trend: Davidson, Bird, and Thompson Lakes (Figures 14d, 14e, 14f,
respectively).
A third trend observed in the DDT/product ratio analysis of this study is gradual increase
to recent from the older portions of the core to the recent. This is exhibited most obviously in
Paw Paw08 Lake, Fremont Lake and Lake Cadillac05 (Figure 10g, 10h, 10i, respectively).
Several other lakes show evidence of this trend as well, including CrystalM Lake, Lake
Geneserath, Gull Lake, and Higgins Lake. With this observation in at least 6 of the lakes
studied, most notably Lake Geneserath, a remote lake on Beaver Island in northern Lake

46

Michigan, suggests the potential for an atmospherically derived component of new DDT to
Michigan.

Figure 2.14: Ratios of DDT / (DDD+DDE) as a function of time in selected Michigan inland lake sediments.

Sediment DDT Concentrations and Porewater Trace Metal Profiles
All study lakes were subject to pore water extraction from a replicate sediment core and
subject to pore water trace metal analysis via ICP-MS. The sediment pore water is defined as the
interstitial water between sediment grains often used for analysis of the redox stratification in
lake sediment (Berner, 1980; Davison, 1993; Urban et al., 1997, Koretsky et al., 2006). Iron can
be used as a surrogate for sediment oxygen concentrations as Fe (III) form oxides in the presence
of oxygen but reductively dissolve to produce the more soluble Fe (II). Thus, it is expected that
47

dissolved Fe to be low near the oxygen-rich surface sediments relative to the suboxic deeper
sediments, expected to be enriched in dissolved Fe.
While many of the lake pore water samples are in the process of analysis or interpretation
at the time of this writing, preliminary results from Gull Lake pore water show concentrations of
iron to be elevated in deeper sediments (> 10 cm depth) and the sediment at these depths do not
contain detectable concentrations of DDT or its break down products (Figure 15). In sediment
above approximately 10 cm, ratios of DDT to its metabolites undergo an increase to the surface.
This suggests that, once buried, reductive dechlorination is the primary pathway for DDT loss
from the sediment in this lake. It can be hypothesized that the observed surficial increase of
DDT noted in many study lakes is attributed to the absence of reductive dechlorination in
surficial sediment as oxygen penetration into the sediment prevents a sufficient microbial
community from performing the necessary dechlorinating breakdown.

Sediment DDT/DDD+DDE (Dimensionless)
0.00

0.60

1.20

1.80

2.40

3.00

0

Sediment Depth (cm)

10

Legend

20

DDT/(DDD+DDE)
PW Fe

30

40

50
0

150

300

450

600

Porewater Iron Concentrations (ng/mL)

Figure 2.15: DDT/(DDD+DDE) ratio and porewater iron concentrations plotted as a function of depth in Gull
Lake.

48

DDT Half-Life Calculations:
As described above, DDT is highly persistent in the environment. Degradation of DDT
in soil occurs by photolysis or biologically-mediated degradation (either aerobic or anaerobic).
Other environmental losses of DDT occur by runoff and volatilization (ATSDR, 2002). A
review of the literature reveals a substantial variation in the calculated half-lives for DDT in
soils. This is likely dependent upon soil-specific conditions such as organic matter content,
percent fine clastic material, oxidation-reduction regimes, etc. Some reported half-life
calculations for DDT are as short as 3-5 years (Bierman and Swain, 1982; Matsumara et al.,
1971). Others report longer half-lives of up to 15 years (USEPA, 1989; Augustijin-Beckers et
al., 1994). Still others estimate longer half-lives of approximately 20 years (Martijn et al., 1993).
Upon entering a surface water body, the DDT half-life is expected to be considerably
shorter with reported values of 56 days in lake water and 28 dates in river water (USEPA, 1989).
Aside from degradation, primary loss pathways include volatilization, photolysis, and
sedimentation (ATSDR, 2002). If DDT avoids breakdown in the water column of a lake and
becomes sedimented, the half-life expands as the compound is readily stored by adsorption to the
organic and clay particles or in the tissues of benthic organisms. Field and laboratory studies in
the UK demonstrate very little breakdown of DDT in sediment after a 46 day study was
concluded (WHO, 1989). Typical half-life calculations of DDT in sediment range from 8 to 20
years and has been shown to follow a first-order conversion (Wolfe et al., 1977; Oliver et al.,
1989).
An estimate of DDT half-life can be performed using the measured concentrations of
DDT and its breakdown products (Oliver et al., 1989). Plotting ln [DDT / (DDT + DDD +
DDE)] versus time is used because of the changing amount of total DDT in the various core

49

segments. As the presence of both DDT and its breakdown products is required for the
calculation, the half-life of DDT in lake sediment is able to be performed for 15 of the 39 lakes
studied. Other lakes only contained detectable breakdown products in its sediment (Elk, Clear,
Black, Klinger, Gogebic, Platte, Whitmore, Gratiot, Muskegon, White, Nichols, Avalon, George,
Sand) or were increasing values of ln [DDT / (DDT + DDD + DDE)] with time (Campau, Birch,
CrystalM, Hackert, RoundD) preventing the calculation from being performed. A representative
example of the half-life calculation plot is shown as Figure 2.16. The half-life calculation for the
15 lakes averaged 20.31 years (Average R2 = 0.69). It is not clear why the DDT half-life
calculations reflect a longer environmental persistence on average than other studies. It appears
that the lakes where half-life calculations tend to be longer (>40 yrs) are correlated to lakes of
elevated sedimentation rates. This suggests the possibility that sedimentation of DDT is
enhanced in more eutrophic lakes by the increased flux to the sediment of organic material
potentially harboring these chemicals.

Figure 2.16: Half-life calculation plot for Higgins Lake.

50

DDT and Land Use:
Based on prior research (Harner et al., 2004; Hermanson, et al., 2007; Wong et al., 2009),
it was expected that DDT concentrations in lake sediment to be correlated to anthropogenic land
use types (%urban and %agriculture) or watershed population. However, the data showed only
weak correlations to any type of land use and the strongest correlation to watershed population
density in the 1970s (R2 = 0.32) (Figure 2.17). This generally agrees with Wong et al, 2009
which showed highest DDT concentrations to be located at urban sites in the 1970s. However,
the correlation is weak possibly due to the mixed use DDT both in an agricultural setting to
control a wide range of insect crop pests but also in range lands and urban settings to control
mosquito populations.
It is also possible that the summed core trends reflect the remains of the regional signal of
DDT application. The latitudes exhibiting elevated summed core concentrations are also those
lying along the more populace belt from the Chicago, IL region east to Detroit, MI. This region
of the Michigan hosts many of the larger cities in the state including Jackson, Lansing, Grand
Rapids and Kalamazoo. Given the relatively weak correlations between DDT and specific
watershed land-use, it’s possible that significant extra-watershed sources along the lower-central
latitudes could provide supplemental inputs of DDT.

Figure 2.17: 1970s Total DDT versus study lake watershed population.

51

Conclusion
Concentrations of DDT, DDD, and DDE are measured in sediment cores collected from
39 inland lakes in Michigan for examination of spatial and temporal trends in accumulation and
degradation. It is hypothesized that the temporal trends of DDT in the sediment cores would
coincide with the reported U.S. production and use trends of the chemical. Peak use and
production of DDT occurred in the U.S. in the late 1950’s and early 1960’s, respectively
(Woodwell, 1971) and DDT values should peak in this timeframe as well with reduced values in
older and more recent sediment.
Overall, temporal profiles of DDT in the lake sediment typically reflect the peak
production/consumption timeframe of the compound, validating the hypothesis. However, a
significant portion of the lake sediment data show DDT results to peak outside of the
production/consumption peak time frame illustrating the independence of the historical use of
DDT at the watershed scale. Many lakes exhibit concentration data of both DDT and its
breakdown products in 210Pb dated sediment prior to its known use in the United States.
Potential explanation include bioturbation of the sediment or molecular diffusion of the
chemicals but neither explanation seems sufficient given the depth of the cores and the distance
over which the chemicals have traveled in some lakes. Also, many lakes show recent increasing
trends in both DDT and its break down products.
While it is clear that US production and usage are the overriding determinants of the
general DDT profiles, results show significant variation of peak concentrations and dates of the
peaks amongst the individual lakes. Concentrations appear to be related to historic population
trends on a regional scale, with the most elevated concentrations, accumulation rates and
inventories of DDT in the southern, more populous, portion of the State. However, specific
52

watershed land use or population did not exhibit exceptionally strong relationships. The
prevalence of “new” DDT in sediment dated to time after the chemicals ban is suggestive of
some remaining local or watershed scale sources supplementing the larger regional signal
amongst the lakes.
Degradation half-life calculations of DDT in Michigan lakes were determined to average
approximately 20 yrs, which is similar to those reported by others. Preliminary pore water
results point to anaerobic dehalogenation as the primary means for breakdown of these
chemicals. This occurs at varying depths in the sediment profile dependent upon the site specific
enrichment of oxygen at the benthic boundary.

Figure 2.18: Sources and water column cycling of DDT associated with suspended and settling solids

It is proposed that cycling of the DDT compounds associated with decaying organic
material both in the water column and recently deposited in the sediment serves as a potential
cause for this phenomenon. It is possible that extra-watershed sources continue to play a role in
the cycle of these chemicals (Harmonson, et al., 2007; Giesz, et al., 2008), however, they are not
53

likely to prevent the full recovery in such a significant portion of the lakes studied. Rather, a
combination of small quantities of extra-watershed sources, continued volatilization from
historically impacted sites within a given watershed (Monosmith and Hermanson 1996), and the
continued cycling of DDT in the water column and recent sediment all interact to produce the
observed signal (Figure 2.18). Similar studies in the Great Lakes have shown organic
contaminants to have a residence time of the 2-3 times the particulate residence time due to the
effects of resuspension and mixing of surficial sediments by aquatic organisms (Eisenreich,
1989).

54

APPENDIX

55

Table 2.1: Physical characteristics of study lakes.

Sampling

Counties

Lake
area

Lake

year

of watershed

(km2)

(m)

section

Surficial
Sedimentation
rate
(g/m2/yr)

Avalon

2003

Montmorency

1.5

21.3

1790

296

3.1

1.2

29.6

1824

540

2.7

Birch

2003

Cass

Bird

2008

Black

2009

Hillsdale
Cheboygan, Montmorency,
Otsego, Presque Isle

Cadillac

2005

Campau
Cass M
Charlevoix

2005

Clear

2009

Cora

2008

CrystalB
CrystalM
Davidson

Sampling
depth

Oldest

Watershed
area
(km2)

4.2

19.5

1916

759

4.2

41.0

14.0

1627

550

1382.5

Wexford, Missaukee

4.7

8.2

1825

396

47.6

2005

Kent

0.5

16.2

1848

2100

4.7

2008

Oakland
Antrim, Charlevoix, Ottawa,
Emmitt

5.2

36.6

1856

965

46.8

69.8

33.5

1653

707

765.1

Montmorency

0.5

27.1

1809

294

4.4

Van Buren

0.9

18.9

1733

369

6.3

2009

Benzie

39.3

49.7

1657

555

106.2

2004

Montcalm

2.9

16.8

1804

559

12.1

Oakland, Lapeer

0.2

21.6

1675

232

1.9

Calhoun
Grand Traverse, Antrim,
Kalkaska

2.4

15.6

1792

453

7.9

31.3

58.8

1398

442

130.9

Houghton

0.2

27.9

1800

104

0.8

26.5

1941

2054

40.5
23.5

2008

Duck

2010

Elk

2010

Emily

2007

Fremont

2009

Newaygo

3.2

Geneserath

2010

Charlevoix

1.9

15.5

795

273

George

2004

Ogemaw

0.8

26.2

1932

417

5.1

Gogebic

2007

Gogebic

51.8

7.62

1611

337

454.1

Gratiot

2007

Keweenaw

5.8

23.8

1837

444

27.0

423

61.7

Gull

2005

Kalamazoo, Barry

8.2

33.5

1674

Hackert

2004

Mason

451

2008

St. Joseph

15.5
21.9

1840

Klinger

0.5
3.3

1800

464

1.5
48.8

Long

2009

Ionia, Montcalm

1.4

18.0

1892

367

22.3

Ionia

1.3

10.3

1973

2590

27.8

35.7

1708

801

718.0

Morrison

2008

Mullett

2001

Cheboygan, Otsego

70.3

Muskegon

2006

Muskegon, Newaygo

16.8

14.5

1965

1607

885.0

Nichols

2005

Newaygo

0.64

19.1

1972

193

1.0

Otter

2004

Genesee, Lapeer, Tuscola

0.3

36.9

1945

933

3.4

27.7

1923

828

30.0

Paw Paw

2008

Berrien, Van Buren

3.7

Platte Lake

2007

Benzie County

10.2

28.8

1910

1885

374.6

16.0

1852

270

2.0

RoundD

2004

Delta

1.8

Sand

2003

Lenawee

1.8

17.3

1864

441

24.5

Shupac

2003

Crawford

0.4

30.4

1829

261

2.2

Teal

2010

Marquette

2.0

10.8

1816

323

5.5

17.0

1698

815

55.5
5.6

Thompson

2007

Livingston

0.61

Whitmore

2007

Washtenaw, Livingston

2.7

20.8

1876

534

White*

2006

Muskegon

10.4

21.6

1873

977

* A watershed was not delineated for White Lake.

56

Table 2.2: Decadal average DDT concentration of all study lakes.
2000
1990
1980
1970
1960
1950
1940
1930
1920
1910
1900

Avalon
4.8
3.4
4.3
5.9
6.5
6.9
5.1
4.7
4.4
2.8
0.9

Birch
4.6
8.5
36.3
62.7
92.5
11.5
0.2
0.0
0.0
0.0
0.6

2000
1990
1980
1970
1960
1950
1940
1930
1920
1910
1900

Crystal
B09
17.6
26.4
18.8
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

Crystal
M04
64.0
38.4
37.7
58.6
45.5
21.6
10.3
5.8
3.4
2.5
2.6

2000
1990
1980
1970
1960
1950
1940
1930
1920
1910
1900

Gratiot07
0.0
0.0
11.8
10.7
10.4
9.1
8.0
0.0
0.0
0.0
0.0
PawPaw08
38.2
52.9
86.4
123.7
207.8
375.6
189.6
38.8
8.1
7.1

2000
1990
1980
1970
1960
1950
1940
1930
1920
1910
1900
Notes:

Bird
4.8
4.5
11.2
18.4
44.2
24.3
8.1
0.6
0.0
0.0
0.0

Black
7.8
6.0
6.7
10.7
9.2
4.2
1.2
0.0
0.0
0.0
0.0

Cadillac05
11.7
9.4
11.1
12.8
14.5
20.9
39.4
66.2
63.9
32.6
16.6

Campau
32.5
39.0
73.4
73.4
166.6
180.0
67.8
55.7
55.8
87.5
68.7

Cass08
40.8
68.0
113.7
230.2
772.1
145.3
19.5
0.0
0.0
6.5
0.0

Charlevoix
3.7
7.1
12.0
23.0
31.7
4.8
16.8
10.3
0.0
0.0
0.0

Clear
2.7
11.9
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

Cora
1526.4
967.2
1311.9
2773.5
2888.1
1965.5
668.0
268.7
106.5
47.2
24.1

Davidson

Duck

Elk10

Emily

Fremont

Geneserath

George

Gogebic

118.7
85.6
204.4
787.0
662.6
323.3
29.0
9.3
0.0
0.0
0.0

50.7
92.0
116.9
49.4
15.5
9.0
7.9
6.9
7.4
3.7

2.0
12.7
17.5
1.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0

84.3
35.5
25.9
39.1
68.1
55.7
34.3
15.1
2.8
29.4
2.8

38.4
33.1
59.1
78.1
52.0
60.5
2.7
0.8

210.2
124.6
84.2
10.8
0.0
0.0
0.0
4.5
0.0
0.0
0.0

8.7
9.9
19.8
59.7
78.0
17.9
1.3
0.0

2.1
1.9
2.8
3.8
4.4
2.3
0.0
0.0
0.0
0.0
0.0

Gull05
9.4
24.3
32.3
57.5
143.0
162.7
138.3
59.9
26.1
0.0
0.0

Hackert
62.8
41.4
54.2
93.2
87.2
15.4
1.5
1.1
1.0
1.2
1.6

Higgins09
109.0
77.8
102.9
85.4
102.0
42.6
14.5
7.1
0.0
0.0
0.0

Klinger
10.3
10.5
14.0
40.9
44.9
24.7
32.5
12.9
4.7
0.0
0.0

Long
13.8
35.7
50.8
64.9
50.8
10.3
4.2
4.5
0.8
0.5
0.3

Morrison
29.6
45.0
62.6
93.0

Muskegon
13.2
12.4
18.7
30.8
76.8
115.3

Nichols
23.8
44.6
62.2
99.5
74.1
59.2
21.4
0.0
0.0
0.0
0.0

Otter
95.8
122.9
159.2
234.6
417.8
28.4

Platte
7.5
15.2
33.0
47.3
5.3

RoundD
11.1
12.4
12.2
8.6
11.2
9.9
20.6
19.2
12.6
5.7
2.8

Sand
14.7
16.6
18.2
18.5
18.8
15.9
5.6
3.7
2.1
2.0
0.5

Shupac
19.3
15.1
21.6
31.3
27.9
16.1
1.3
1.2
0.1
0.0
0.1

Teal
35.1
58.8
23.3
36.0
20.3
10.5
8.7
12.4
12.4
9.4
4.8

Thompson
12.7
16.9
28.8
53.8
77.6
240.4
46.9
3.5
0.0
3.4
2.5

White
37.9
42.9
55.6
150.3
302.2
248.8
184.7
43.4
0.0
0.0
0.0

Whitmore07
32.0
94.1
77.4
82.9
151.6
303.5
81.8
23.3
9.4
12.9
15.9

Concentrations below the detection limit are treated as a concentration of 0.0
Due to high sedimentation rates, some lake cores did not acquire sediment dated to early 20 th century.
Concentrations in bold represent detected values prior to significant world-wide use of DDT.

57

Table 2.3. DDT Concentration Peak Date, Peak DDT Concentration, Total Sediment Core DDT
Concentration, Calculated DDT Half-Life, and R2 for Half-Life.

Lake

Latitude

Gratiot
Emily
Gogebic
Teal
RoundD
Geneserath
Black
Charlevoix
Clear
Avalon
Elk
Shupac
Platte
CrystalB
Higgins
George
Cadillac
Hackert
Nichols
Fremont
White
CrystalM
Muskegon
Otter
Long
Davidson
Morrison
Campau
Thompson
Cass
Whitmore
Gull05
Duck
Paw Paw
Cora
Sand
Birch
Bird
Klinger
Average

47.22
46.86
46.54
46.51
46.15
45.60
45.45
45.26
45.13
45.10
44.87
44.73
44.70
44.67
44.49
44.40
44.24
43.98
43.73
43.45
43.38
43.26
43.23
43.22
43.11
42.89
42.87
42.83
42.61
42.60
42.44
42.40
42.39
42.21
42.20
42.03
41.87
41.83
41.80

Peak
Date

Peak
Conc

Surficial
sedimentation rate
(g m-2 yr-1)

1985
1964
1958
2000
1945
2007
1965
1964
1969
1962
1988
1975
1967
1960
1965
1970
1931
1972
1977
1971
1968
1971
1952
1963
1973
1974
1974
1920
1956
1956
1957
1952
1989
1948
1966
1981
1967
1961
1952
1966

11.78
68.01
4.36
127.33
20.57
210.19
10.73
31.69
16.23
7.34
17.52
31.34
33.38
43.72
101.96
135.38
66.23
93.17
102.51
84.07
302.16
58.53
133.87
711.45
69.71
787.01
102.09
179.99
274.65
879.43
334.68
162.73
137.08
446.24
2888.14
19.02
132.74
63.57
44.88
229.37

444
104
337
323
270
273
550
707
294
296
442
261
1885
555
214
417
396
451
193
2054
977
554
1607
933
367
232
2590
2100
815
965
534
423
453
828
369
441
540
759
464
1364.15

a

Average focus-corrected
DDT accum. rate since
1940’s
(μg m-2 yr-1)

DDT
Inventory
(ng cm-2)

0.8
2.2
0.1
5.3
1.1
11.2
2.0
3.5
0.5
1.2
1.2
1.9
10.2
3.1
6.6
6.3
2.1
5.4
4.2
51.9
16.2
12.0
26.7b
43.2
46.2
46.1
63.2b
35.2
13.4
81.1
22.0
18.5
20.8
35.8
123.5
2.7
9.2
12.2
14.4
18.2

2.61
27.56 a
0.31
56.99 a
10.17
42.99 a
10.52
19.52
4.82
6.26
7.41
9.48
110.72
42.84
41.35
40.68
35.59 a
29.80
36.01
594.09 a
354.65 a
52.40
88.85 a
475.54 a
348.50 a
232.21
258.22 a
232.76
82.83
800.52
205.32
26.70
165.64 a
398.42 a
730.01
12.19
41.26
66.89
41.11
107.3

t1/2

R2

13.9
20.4

0.51
0.85

44.1
10.4

0.87
0.71

49.9

0.63

6.5

0.96

45.9
10.2
14.3
29.2

0.11
0.64
0.57
0.51

9.6

0.71

10.3
18.4
16.3

0.95
0.6
0.81

5.3

0.76

20.3

0.68

= Indicates Total DDT concentrations were detected above 1 ng/g to the bottom of the sediment core. DDT concentrations may exist below
the depth of the core taken and the DDT Inventory calculations may be under represented.
b
= Sediment core dating showed deepest sediment to age younger than 1940s, thus average DDT accumulation rate may be subject to some
error.

58

Table 2.4. Select data from 210Pb analysis, including the model used for dating, mixed depth, surficial
sedimentation rate, focusing factor (FF) and the age of the oldest section in the sediment core.

Lake
Avalon
Birch
Bird
Black
Cadillac
Campau
Cass
Charlevoix
Clear
Cora
CrystalB
CrystalM
Davidson
Duck
Elk
Emily
Fremont
Geneserath
George
Gogebic
Gratiot07
Gull05
Hackert
Higgins
Houghton
Imp
Klinger
Long
Morrison
Muskegon
Nichols
Otter
Paw Paw
Platte
RoundL
RoundD
Sand
Shupac
Teal
Thompson
Torch
Whitmore
White
Witch

Model
CRS
CRS
CRS
SCF:CS
CRS
SCF:CS
CRS
CRS
CF:CS
SCF:CS
SCF:CS
CRS
CRS
CRS
CF:CS
CRS
CRS
CRS
CF:CS
CF:CS
CF:CS
CRS
CRS
CRS
SCF:CS
CRS
SCF:CF
SCF:CS
CF:CS
CF:CS
CRSd
CF:CS
CF:CS
CF:CS
CRS
CRS
CRS
CRS
CF:CS
CRS
CRS
SCF:CS
SCF:CS
CRS

Surficial
Sedimentation
rate
(g/m2/yr)
296
540
759
550
396
2100
965
707
294
369
555
559
232
453
442
104
2054
273
417
337
444
423
451
214
165
119
464
367
2590
1607
193
933
828
1885
317
270
441
261
323
815
365
534
977
269

Approximate
mixed depth (cm)
4
3
1.5
0
17
4.5
0
1.5
2.5
3.5
4
0
4
2.5
0.5
0
3.5
3
9
4
4
0
0
1.5
8
3
1.5
1.0
3.5
3
0
8
3
6
7
9
0
1
5.5
8
0
3
3
6

FF
1.5
1.7
1.7
1.7
1.8
1.6
3.2
2.0
2.4
2.1
3.5
1.6
3.5
1.8
2.1
2.7
4.1
1.7
2.1
1.3
1.5
1.7
1.9
2.5
1.2
1.5
2.4
2.5
2.0
2.9c
2.3c
3.5
2.7c
3.4
2.3
2.4
1.8
2.0
2.4
1.5
2.4
2.3
1.5c
1.7

Oldest section
1790a
1824a
1916 a
1627 a
1825a
1848a
1856
1653a
1809a
1733 a
1657 a
1804
1675 a
1792 a
1398 a
1800 a
1941
795
1932
1611a
1837a
1674a
1840
1663 a
1715a
1745a
1900
1892
1973
1965
1972d
1945
1923
1910
1851
1852
1864
1829a
1816a
1698a
1493a
1876
1873
1767a

a = Estimated dates based on extrapolation
b = focusing factor could not be calculated for Littlefield Lake, so the average focusing factor of all lakes sampled previously (except Cass
Lake) was used
c = Estimated focusing factor based on extrapolation
d = An extrapolation of the 210Pb integral was used

59

Table 2.5: GIS-determined 1970s watershed land-use percentages for all Michigan study lakes organized by
latitude.
Urban %

Agriculture %

Forest %

Range %

Wetland %

Avalon

Lake

2.9

70.0

25.7

1.2

0.1

Birch

2.2

33.0

42.2

19.9

2.8

Bird

10.6

66.6

19.8

2.6

0.1

Black

1.1

5.5

80.1

8.6

3.8

Cadillac

19.9

41.1

24.6

10.3

4.2

Campau

17.8

28.1

22.2

27.2

1.9

Cass

62.0

2.5

8.8

14.3

12.5

Charlevoix

4.0

20.5

57.0

15.7

2.0

Clear

20.6

0.0

76.6

1.5

0.0

Cora

19.4

48.1

23.3

4.9

4.0

Crystal B

12.4

13.0

52.1

21.3

1.2

Crystal M

11.1

58.6

17.0

5.0

8.3

Davison

14.1

17.2

26.8

38.6

2.7

Duck

6.2

43.1

30.2

11.4

7.7

Elk

6.4

28.2

41.4

20.0

4.0

Emily

0.0

0.0

93.2

0.0

5.5

Fremont

20.5

58.2

13.2

3.1

3.7

Geneserath

0.3

0.0

84.3

0.4

8.0

George

9.7

23.1

40.0

24.3

2.8

Gogebic

1.2

0.0

92.1

0.5

5.0

Gratiot

0.9

0.0

98.5

0.0

0.6

Gull

10.0

61.1

14.6

8.2

6.0

Hackert

21.5

31.9

28.5

16.0

2.1

Higgins

15.0

0.5

77.6

3.5

3.4

Houghton

7.2

1.6

75.8

5.4

10.0

Imp

7.1

0.0

88.1

0.0

4.8

Klinger

6.8

57.0

21.5

5.4

5.9

Littlefield

1.9

9.4

70.5

14.0

4.3

Long

5.7

39.5

42.0

4.6

7.5

Morrison

6.0

74.4

13.9

1.9

3.7

Mullett

3.0

7.4

71.4

15.3

2.9

Muskegon

11.4

71.0

4.1

13.2

0.3

Nichols

7.3

0.0

84.2

0.0

1.6

Otter

9.3

8.3

50.3

17.3

14.8

Paw Paw

17.8

30.8

35.8

13.5

2.1

Platte

2.3

9.9

63.5

22.0

1.2

Round

4.8

13.9

68.9

11.5

1.0

Round_D

9.9

0.0

89.1

0.0

1.0

Sand

20.5

11.8

50.2

15.2

2.3

Shupac

0.9

0.5

83.6

13.9

1.1

Teal

27.0

0.0

59.7

6.2

2.6

Thompson

22.9

25.7

17.4

29.8

3.8

Torch

7.4

23.6

45.1

22.0

1.9

White

3.3

24.2

58.9

9.5

2.8

Whitmore

22.4

23.9

24.1

20.8

8.8

Witch

1.8

3.0

81.8

6.8

6.6

60

Table 2.6: GIS-determined 1970s watershed land-use percentages for all Michigan study lakes organized by
latitude.
Lake

Urban %

Agriculture %

Upland %

Forest %

Wetland %

Barren %

Avalon

8.8

0.0

12.2

71.2

1.2

0.2

Birch

6.4

48.3

7.8

27.8

4.4

0.1

Bird

8.3

52.8

8.4

23.2

6.5

0.0

Black

1.4

3.0

16.2

69.9

8.3

0.1

Cadillac

10.7

26.6

18.8

35.1

5.7

1.7

Campau

10.4

31.9

13.7

32.7

9.2

0.0

Cass

24.8

0.5

18.9

37.5

11.0

0.1

Charlevoix

3.4

17.3

14.8

61.0

2.3

0.3

Clear

3.3

0.0

11.4

82.2

1.1

0.1

Cora

7.2

49.7

9.0

31.0

2.3

0.1

Crystal B

5.5

9.4

22.3

60.0

0.9

0.3

Crystal M

20.0

56.7

3.5

13.7

3.3

0.1

Davison

3.5

21.0

13.1

50.7

11.2

0.0

Duck

17.0

49.8

6.9

21.4

4.6

0.0

Elk

4.2

20.4

18.6

53.6

1.8

0.2

Emily

1.8

0.0

0.6

89.5

2.1

0.0

Fremont

9.9

47.8

12.7

22.8

4.2

0.8

Geneserath

0.1

0.0

6.7

74.2

18.2

0.0

George

5.0

6.3

17.4

59.9

9.4

0.2

Gogebic

1.2

0.0

1.8

86.6

8.4

0.2

Gratiot

0.9

0.0

6.2

89.3

3.0

0.1

Gull

5.6

59.8

6.1

21.0

4.9

0.1

Hackert

5.9

33.8

10.2

42.3

4.3

0.2

Higgins

6.9

0.2

15.8

72.2

3.7

0.2

Houghton

3.7

0.8

14.6

62.2

16.4

0.2

Imp

3.5

0.0

0.6

89.2

0.5

0.3

Klinger

5.8

53.4

5.3

27.1

5.5

0.2

Littlefield

2.9

2.3

12.6

75.6

4.6

0.1

Long

4.8

30.6

9.0

31.1

23.4

0.1

Morrison

6.4

72.2

4.5

11.9

4.7

0.2

Mullett

2.8

7.2

17.7

68.1

3.2

0.2

Muskegon

7.0

17.2

18.0

47.4

7.0

0.4

Nichols

0.9

0.0

2.0

86.4

5.4

0.2

Otter

8.3

4.8

16.7

49.6

17.9

0.0

Paw Paw

9.2

38.6

13.2

31.6

6.4

0.0

Platte

2.6

8.7

19.6

66.7

1.3

0.2

Round

0.4

7.2

9.2

72.9

9.3

0.3

Sand

8.5

6.9

13.4

62.6

7.0

0.0

Shupac

2.8

0.0

18.5

74.8

2.8

0.0

Teal

20.0

0.0

5.5

69.8

1.3

0.1

Thompson

17.4

26.3

15.7

32.0

0.4

0.3

Torch

4.2

18.6

19.4

54.3

1.5

0.2

White

3.0

17.4

16.2

58.1

0.3

4.1

Whitmore

7.6

35.9

12.5

32.7

10.4

0.0

Witch

3.0

0.0

2.8

87.2

0.0

2.1

61

REFERENCES

62

REFERENCES

Agency for Toxic Substances and Diseases Registry (ATSDR)/US Public Health Service,
Toxicological Profile for 4,4'-DDT, 4,4'-DDE, 4, 4'-DDD (Update). 1994. ATSDR.
Atlanta, GA. http://www.atsdr.cdc.gov/.
Agency for Toxic Substances and Disease Registry (ATSDR), 2002. Toxicological Profile for
DDT, DDE, and DDD. U.S Department of Health and Human Services.
http://www.atsdr.cdc.gov/toxprofiles/tp.asp?id=81&tid=20 Retrieved 06/18/13.
Agency for Toxic Substances and Disease Registry (ATSDR), 2008. Addendum to the
Toxicological Profile for DDT, DDE, and DDD. U.S Department of Health and Human
Services. http://www.atsdr.cdc.gov/toxprofiles/tp.asp?id=81&tid=20 Retrieved 06/18/13.
Agrochemicals Desk Reference: Environmental Data. 1993. Lewis Publishers. Chelsea,
Michigan
Aislabie, J. M., Richards, N. K., & Boul, H. L., 1997. Microbial degradation of DDT and its
residues—a review. New Zealand Journal of Agricultural Research, 40(2), 269-282.
Alegria, H. A., Bidleman, T. F., & Shaw, T. J., 2000. Organochlorine pesticides in ambient air of
Belize, Central America. Environmental science & technology,34(10), 1953-1958.
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.
Andrea, M. M., Tomita, R. Y., Luchini, L. C., & Musumeci, M. R. 1994. Laboratory studies on
volatilization and mineralization of 14c‐p, p'‐DDT in soil, release of bound residues and
dissipation from solid surfaces. Journal of Environmental Science & Health Part
B, 29(1), 133-139.
Audette, C. V., Hulting, M. L., Basu, I., Brice, K. A., Chan, C. H., Dryfhout-Clark, H., ... &
Neilson, M. (2008). Atmospheric deposition of toxic substances to the Great Lakes:
IADN results through 2005 (pp. 978-0). Environment Canada.
Augustijn-Beckers, P.W.M., Hornsby, A.G. and Wauchope, R.D. 1994. SCS/ARS/CES Pesticide
Properties Database for Environmental Decisionmaking II. Additional Properties
Reviews of Environmental Contamination and Toxicology, Vol. 137
Augustijn-Beckers, P.W.M., Hornsby, A.G. and Wauchope, R.D. 1994. SCS/ARS/CES Pesticide
Properties Database for Environmental Decisionmaking II. Additional Properties
Reviews of Environmental Contamination and Toxicology, Vol. 137.

63

Baker, J. E., S. J. Eisenreich & B. J. Eadie, 1991. Sediment trap fluxes and benthic recycling of
organic carbon, polycyclic aromatic hydrocarbons, and polychlorinated biphenyl
congeners in Lake Superior. Environ. Sci. Technol. 25: 500–509.
Berner RA: Early Diagenesis. Princeton, NJ: Princeton University Press; 1980.
Brooks, G.T. 1974. Chlorinated Insecticides: Volume I Technology and Application. CRC
Press, Inc. Cleveland, Ohio.
Bierman, V. J., & Swain, W. R. 1982. Mass balance modeling of DDT dynamics in Lakes
Michigan and Superior. Environmental Science & Technology, 16(9), 572-579.
Burton Jr, G. A. 2002. Sediment quality criteria in use around the world. Limnology, 3(2), 65-76.
Callender, E., vanMetre, P.C., 1997. Reservoir sediment cores show US lead declines.
Environmental Science and Technology 31, A424-A428.
Carollo, J. A. 1945. The removal of DDT from water supplies. J. Am. Water Works. Assoc.,
37: 1310-1317.
Cortes, D. R., Basu, I., Sweet, C. W., Brice, K. A., Hoff, R. M., & Hites, R. A. (1998). Temporal
trends in gas-phase concentrations of chlorinated pesticides measured at the shores of the
Great Lakes. Environmental science & technology, 32(13), 1920-1927.
Cramer, J. Model of the circulation of DDT on earth. 1973. Atmospheric environment Oxford
England 1994 7, 241-256(1973).
Davison, W. (1993). Iron and manganese in lakes. Earth-Science Reviews,34(2), 119-163.
Dimond, J. B., and R. B. Owen. 1996, Long-term residue of DDT compounds in forest soils in
Maine, Environ. Pollut., 92, 227–230.
Eadie, B. J.; Robbins, J. A. In Sources and Fates of Aquatic Pollutants; Hites, R. A., Eisenreich,
S. J., Eds.; Advances in Chemistry 216; American Chemical Society: Washington, DC,
1987; pp 320-364
Edgington, D. N., Klump, J. V., Robbins, J. A., Kusner, Y. S., Pampura, V. D., & Sandimirov, I.
V. 1991. Sedimentation rates, residence times and radionuclide inventories in Lake
Baikal from 137Cs and 210Pb in sediment cores.
El Bilali, L., Rasmussen, P. E., Hall, G. E. M., & Fortin, D. 2002. Role of sediment composition
in trace metal distribution in lake sediments. Applied Geochemistry, 17(9), 1171-1181.
Eisenreich, S. J., Capel, P. D., Robbins, J. A., & Bourbonniere, R. 1989. Accumulation and
diagenesis of chlorinated hydrocarbons in lacustrine sediments. Environmental Science &
Technology, 23(9), 1116-1126.

64

Eskenazi, B., Chevrier, J., Rosas, L. G., Anderson, H. A., Bornman, M. S., Bouwman, H., ... &
Stapleton, D. 2009. The Pine River statement: human health consequences of DDT use.
Environmental Health Criteria 9: DDT and its derivatives, World Health Organization, 1979.
Fiedler, H., Lau, C. 1998. Environmental Fate of Chlorinated Organics. In: Schuurmann G,
Markert, B.(eds) Ecotoxicology. Wiley, p 317.
Fries, G.R., Marrow, G.S.; Gordon, C.H. 1969. Metabolism of o.p’-DDT by rumen
microorganisms. Journal of agricultural and food chemistry 17: 860-862
Fuller, L.M., and Minnerick, R.J., 2008, State and regional water-quality characteristics and
trophic conditions of Michigan’s inland lakes, 2001–2005: U.S. Geological Survey
Scientific Investigations Report 2008–5188, 58 p. Date Posted: December 17, 2008:
[http://pubs.water.usgs.gov/sir20085188/]
Geankoplis, C. J. 1972. Mass transfer phenomena. Holt Rinehart and Winston, NY, USA.
Gingrich, S. E., Diamond, M. L., Stern, G. A., & McCarry, B. E. 2001. Atmospherically derived
organic surface films along an urban-rural gradient.Environmental science &
technology, 35(20), 4031-4037.
Gobeil, C., Macdonald, R. W., & Smith, J. N. 1999. Mercury profiles in sediments of the Arctic
Ocean basins. Environmental science & technology,33(23), 4194-4198.
Golden, K. A., C. S. Wong, J. D. Jeremiason, S. J. Eisenreich, G. Sanders, J. Hallgren, D. L.
Swackhamer, D. R. Engstrom, and D. T. Long. 1993. Accumulation and preliminary
inventory of organochlorines in Great Lakes sediments. Water Science Technology. 28:
19-31.
Guenzi, W. D., and Beard, W. E. 1967. Anaerobic biodegradation of DDT to DDD in soil.
Science. 156; 1116-1117.
Guenzi, W. D., and Beard, W. E. 1968. Anaerobic conversion of DDT to DDD and
Aerobic Stability of DDT in Soil. Soil. Soil Sci. Soc. Amer. Proc. 32: 522-524.
Gustafson, D. I. 1989. Groundwater ubiquity score: a simple method for assessing pesticide
leachability. Environmental toxicology and chemistry, 8(4), 339-357.
Halsall, C. J., Bailey, R., Stern, G. A., Barrie, L. A., Fellin, P., Muir, D. C. G., ... & Pastukhov,
B. (1998). Multi-year observations of organohalogen pesticides in the Arctic
atmosphere. Environmental Pollution, 102(1), 51-62.
Harner, T., Bidleman, T. F., Jantunen, L. M., & Mackay, D. (2001). Soil—air exchange model of
persistent pesticides in the United States cotton belt. Environmental Toxicology and
Chemistry, 20(7), 1612-1621.

65

Harner, T., Shoeib, M., Diamond, M., Stern, G., & Rosenberg, B. (2004). Using passive air
samplers to assess urban-rural trends for persistent organic pollutants. 1. Polychlorinated
biphenyls and organochlorine pesticides. Environmental Science & Technology, 38(17),
4474-4483.
Harris,M. L., L. K.Wilson, J. E. Elliott, C. A. Bishop, A. D. Tomlin, and K.V.Henning. 2000,
Transfer of DDT and metabolites from fruit orchard soils to American robins (Turdus
migratorius) twenty years after agricultural useof DDT in Canada, Arch. Environ.
Contam. Toxicol., 39, 205– 220.
Hermanson, M. H., Moss, D. J., Monosmith, C. L., & Keeler, G. J. 2007. Spatial and temporal
trends of gas and particle phase atmospheric DDT and metabolites in Michigan: Evidence
of long‐term persistence and atmospheric emission in a high‐DDT‐use fruit
orchard. Journal of Geophysical Research: Atmospheres (1984–2012), 112(D4).
Hill D. W., and McCarty, P. L. 1967. Anaerobic degradation of selected chlorinated
hydrocarbon pesticides. J. Water Pollut. Control Fed. 39: 1259-1277.
Hiltbold, A. E. 1974, Persistence of pesticides in soil, in Pesticides in Soil and Water, edited by
W. D. Guenzi et al., pp. 203–222, Soil Sci. Soc. Of Am., Madison, Wis
"Concern over excessive DDT use in Jiribam fields". The Imphal Free Press. 2008-05-05.
Archived from the original on 2008-12-06. Retrieved 2014-03-18.
Jeremiason, J. D., Hornbuckle, K. C., & Eisenreich, S. J. (1994). PCBs in Lake Superior, 19781992: decreases in water concentrations reflect loss by volatilization. Environmental
science & technology, 28(5), 903-914.
Jensen, S.; Gothe, R.; Kindertedt, M.O. 1972. Bis-(p-Chlorophenyl)-Acetonitrile
(DDN), a new DDT derivative formed in anaerobic digested sewage sludge and Lake
Sediment. Nature. 240:421-422.
Karickhoff, S. W., & Morris, K. R. 1985. Impact of tubificid oligochaetes on pollutant transport
in bottom sediments. Environmental science & technology,19(1), 51-56.
Koretsky, C. M., Haas, J. R., Miller, D., & Ndenga, N. T. (2006). Seasonal variations in pore
water and sediment geochemistry of littoral lake sediments (Asylum Lake, MI,
USA). Geochemical transactions, 7(11).
Khim, J.S. et al. Characterization and Distribution of Trace Organic Contaminants in
Sediment from Masan Bay, Korea. 1. Instrumental Analysis. Environmental Science
& Technology 33, 4199-4205(1999).
Ko W.H., Lockwood J. L., 1968. Conversion of DDT to DDD in soil and the effects of these
compounds on soil microorganisms. Canadian Journal of Microbiology. 14:1069-1073.

66

Martijn, A., Bakker, H., & Schreuder, R. H. 1993. Soil persistence of DDT, dieldrin, and lindane
over a long period. Bulletin of environmental contamination and toxicology, 51(2), 178184.
Matsumura, F., Patil, K. C., and Boush, G.M. 1971. DDT metabolized by microorganisms from
Lake Michigan. Nature (London). 230:325-326.
Meister, R. T., & Sine, C. 2009. Crop protection handbook. Meister Media Worldwide.
Mohn W.W., Tiedje J. M. 1992. Microbial reductive dehalogenation. Microbiol Rev.
56:482-507.
Monosmith, C. L., & Hermanson, M. H. 1996. Spatial and temporal trends of atmospheric
organochlorine vapors in the central and upper Great Lakes.Environmental science &
technology, 30(12), 3464-3472.
Muir, D. et al. Spatial trends and historical profiles of organochlorine pesticides in Arctic lake
sediments. Science of The Total Environment 160-161, 447-457(1995).
Oliver, B. G., M. N. Charlton & R. W. Durham, 1989. Distribution, redistribution, and
geochronology of polychlorinated biphenyl congeners and other chlorinated
hydrocarbons in Lake Ontario sediments. Environ. Sci. Technol. 23: 200–208
Patil, K.C., F. Matsumura, and G.M. Boush. 1972. Metabolic transformation of DDT, dieldrin,
aldrin, and endrin by marine microorganisms. Environ. Sci, Thechnol. 21: 397-399.
Quensen J.F., Tiedje J.M., Jain M.K., and Mueller, S. A. 2001. Factors controlling the rate of
DDE dechlorination to DDMU in palos verdes margin Sediments under anaerobic
conditions. Environ. Sci. Technol. 35; 286-291.
Quensen, J. F., Mueller, S. A., Jain, M. K., & Tiedje, J. M. 1998. Reductive dechlorination of
DDE to DDMU in marine sediment microcosms. Science, 280(5364), 722-724.
Patil, K.C., F. Matsumura, and G.M. Boush. 1972. Metabolic transformation of DDT, dieldrin,
aldrin, and endrin by marine microorganisms. Environ. Sci, Thechnol. 21: 397-399.
Rapaport, R. A., Urban, N. R., Capel, P. D., Baker, J. E., Looney, B. B., Eisenreich, S. J., &
Gorham, E. 1985. “New” DDT inputs to North America: atmospheric
deposition. Chemosphere, 14(9), 1167-1173.
Rawn, D.F. et al. Historical contamination of Yukon Lake sediments by PCBs and
organochlorine pesticides: influence of local sources and watershed characteristics.
Science of the Total Environment 280, 17-37(2001).
Report of the Third Expert Group Meeting on DDT, UNEP/POPS/DDT-EG.3/3, Stockholm
Convention on Persistent Organic Pollutants, November 12, 2010.

67

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(3), 285-304.
Stockholm Convention on Persistent Organic Pollutants. 2010. Report of the Third Expert Group
Meeting on DDT, UNEP/POPS/DDT-EG.3/3,
http://chm.pops.int/Programmes/DDT/Meetings/DDTEG32010/tabid/1108/mctl/ViewDet
ails/EventModID/1421/EventID/116/xmid/4037/language/en-US/Default.aspx
Retrieved: 05/20/13.
Thomann, R. V., & Di Toro, D. M. 1983. Physico-chemical model of toxic substances in the
Great Lakes. Journal of Great Lakes Research, 9(4), 474-496.
Urban, N. R., Dinkel, C., & Wehrli, B. 1997. Solute transfer across the sediment surface of a
eutrophic lake: I. Porewater profiles from dialysis samplers. Aquatic Sciences, 59(1), 125.
U.S. Environmental Protection Agency. 1975. DDT, A Review Of Scientific And Economic
Aspects Of The Decision to Ban Its Use As A Pesticide. Washington, DC. Found at:
http://www2.epa.gov/sites/production/files/documents/DDT.pdf Retrieved 04/13/13.
U.S. Environmental Protection Agency. 1989. Environmental Fate and Effects Division,
Pesticide Environmental Fate One Line Summary: DDT (p, p'). Washington, DC.
U.S. Department of Agriculture, 2006, Pesticide Data Program Annual Summary Calendar Year
2005. http://www.ams.usda.gov/AMSv1.0/getfile?dDocName=STELPRDC5049946
Retrieved 03/26/2014
Van Den Berg, H. Global Status of DDT and Its Alternatives for Use in Vector Control to
Prevent Disease. Environmental Health Perspectives 117, 1656-1663(2009).
Van Metre, P. C., Callender, E., & Fuller, C. C. 1997. Historical trends in organochlorine
compounds in river basins identified using sediment cores from
reservoirs. Environmental Science & Technology, 31(8), 2339-2344.
Van Metre, P.C. and B.J. Mahler. 2005. Trends in Hydrophobic Organic Contaminants in Urban
and Reference Lake Sediments across the United States, 1970-2001. Environmental
Science & Technology 39(15), 5567-5574
Varca, L. M., & Magallona, E. D. 1994. Dissipation and degradation of DDT and DDE in
Philippine soil under field conditions. Journal of Environmental Science & Health Part
B, 29(1), 25-35.
Voerman, S., and A. F. H. Besemer 1975, Persistence of dieldrin, lindane and DDT in a light
sandy soil and their uptake by grass, Bull. Environ. Contam. Toxicol., 13, 501– 505.
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.

68

Wolfe, N. L., Zepp, R. G., Paris, D. F., Baughman, G. L., & Hollis, R. C. 1977. Methoxychlor
and DDT degradation in water: rates and products. Environmental Science &
Technology, 11(12), 1077-1081.
Wong, C. S., Sanders, G., Engstrom, D. R., Long, D. T., Swackhamer, D. L., & Eisenreich, S. J.
1995. Accumulation, inventory, and diagenesis of chlorinated hydrocarbons in Lake
Ontario sediments. Environmental science & technology,29(10), 2661-2672.
Wong, F., Robson, M., Diamond, M. L., Harrad, S., & Truong, J. 2009. Concentrations and
chiral signatures of POPs in soils and sediments: a comparative urban versus rural study
in Canada and UK. Chemosphere, 74(3), 404-411.
World Health Organization (WHO). 1989. Environmental Health Criteria 83, DDT and its
Derivatives--Environmental Effects. World Health Organization, Geneva.
World Health Organization (WHO), 1979. Environmental Health Criteria 9: DDT and its
Derivatives, Geneva.
Woodwell, G.M., Craig, P.P. & Johnson, H.A. DDT in the biosphere: where does it go? Science
174, 1101-1107(1971).
West, T. F. and Campbell, G. A. 1950. DDT and Newer Persistent Insecticides, revised 2nd ed.,
Chapman and Hall, London.
Wolfe, N. L., Zepp, R. G., Paris, D. F., Baughman, G. L., & Hollis, R. C. (1977). Methoxychlor
and DDT degradation in water: rates and products.Environmental Science &
Technology, 11(12), 1077-1081.
World Health Organization, 1979. Environmental Health Criteria 9: DDT and its Derivatives,
Geneva.
Wu, S. C., & Gschwend, P. M. 1988. Numerical modeling of sorption kinetics of organic
compounds to soil and sediment particles. Water Resources Research, 24(8), 1373-1383.
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.
Available from: <http://www.deq.state.mi.us/documents/deq-wb-swassedimenttrend0203finalreport.pdf>.
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(7), 1157-1175.
Zimmerman, O.T., Lavine, I. 1946. DDT- Killer of Killers, Industrial Research Service, Dover,
New Hampshire.
Zoro, J. A., Hunter, J. M., and Eglington, G. 1974. Degradation of p, p’-DDT in Reducing
Environments. Nature. 247:235-237.

69

CHAPTER 3
LANDSCAPE RECOVERY FROM HISTORICAL PCB LOADINGS USING SEDIMENTCHEMICAL CHRONOLOGIES

Abstract
Previous studies have shown atmospheric transport to be a dominant pathway for PCB
loadings to the environment and that urban areas typically serve as major sources of these
chemicals. A significant segment of the existing research to date in the Great Lakes region
mainly involves the Great Lakes themselves. These water bodies occupy an enormous footprint
on the North American continent and results on PCB studies here are more reflective of a coarse
regional signal of PCB distribution than local inputs. More detailed understanding of loading
patterns is needed for a more complete understanding of sources of PCBs and the current state of
environmental recovery. In this study, sediment records were studied from 44 inland lakes
across the State of Michigan representing a variety of watershed land use and trophic
characteristics to better understand the spatial distribution and changes over time of PCB
loadings.

210

Pb dated PCB concentrations, accumulation rates and inventories from each lake

were compared to GIS-determined watershed land-use characteristics. Overall, sediment PCBs
were elevated in the Lower Peninsula compared to the Upper Peninsula and locations of high
PCB inventories near urban areas were found. However, concentration and accumulation peak
dates increased with latitude of the lake possibly due to secondary mobility and deposition away
from sources. Individual PCB congener concentrations were subject to cluster analysis via Jump
5.0 software. During high PCB production years, congener clusters show clear localized

70

patterns, but more recently, congener distributions suggest a more regional signal. However,
state-wide regional patterns appear to serve as a component to both recent and historical data,
despite the prominence of local and sub-regional trends during the peak PCB
production/consumption years.

71

Introduction
Polychlorinated biphenyls (PCBs) have been studied for decades as the hazards
associated with exposure of these chemicals to human and ecosystem health has been described
(Carpenter, 2006; Mozaffarian et al., 2006; Longnecker et al., 1997; Cordle et al., 1978; Nisbet
and Sarofim, 1972). Results from these studies show that, although manufacture and use of
PCBs has been banned for over 30 years, they still remain present in significant quantities in the
environment. A common fate of many environmentally-persistent chemicals in freshwater
systems is ultimately to become deposited to areas of low energy such as a lake bottom.
Contaminated sediments can directly impact bottom-dwelling organisms and represent a
continuing source for toxic substances in the aquatic food wed. This affects wildlife and humans
through food or water consumption. Therefore, an understanding of the trends of these toxic
chemicals accumulation in the environment is necessary to assess the current state and potential
future problems of surface water quality. Here we use chemical analysis of inland lake sediment
cores to provide some elucidation to PCBs continued environmental presence dated throughout
the chemicals use history in Michigan.
PCBs are a group of synthetic chemical compounds characterized by a pair of benzene
rings with between 2 and 10 attached with chlorine atoms of various configurations (Pohanish,
2008; USEPA, 2003; ATSDR, 2000). Historically, PCBs were first synthesized in 1881 but not
manufactured commercially until 1929. The chemistry of these compounds provides for
excellent heat conduction, low flammability, chemical stability and inertness. As such, PCBs
saw widespread use as coolants and lubricants in electrical equipment such as transformers and
capacitors (USEPA, 1013; Callahan, et al., 1983). In addition, PCBs were also used as coatings
on electrical wiring and components, pesticide extenders, cutting oils, flame retardants, and

72

hydraulic fluid providing the potential for many anthropogenic sources. Additional
complications arise in PCB dispersal when considering different congeners of PCBs were used at
different times and for different applications. For example, Aroclor 1260 and 1254 were
predominately used in the U.S. prior to 1950, Aroclor 1242 was the primary mixture used in the
1950s and 1960s, and Aroclor 1016 has been the most widely used since the 1970s (UNEP,
1997). Overall, annual U.S. production peaked in 1970 with a volume on the order of 1.5 million
tons (Voogt and Brinkman, 1989). Thereafter, PCB manufacture was ended in 1979 in the
United States as evidence of its causing cancer in animals and humans (Tanabe, 1988).
All 130 of the commercially-used chlorine-biphenyl arrangements of PCBs share a
generalized set of similar physical properties. They are odorless, tasteless and typically clear to
slightly yellow and viscous. All PCBs share the characteristic of having very low water
solubilities and low vapor pressures at room temperature, and have high solubilities in solvents,
oils, and fats. This lipophilicity property of PCBs is attributed to the significant bioaccumulation
in the food chain. Significant differences amongst the PCB congeners are usually attributed to
the degree of chlorination of the biphenyls. Physically, the more chlorinated the congener of
PCB, the more viscous and yellow-colored it becomes. Chemical differences as chlorination
increases amongst congeners of PCBs include increasing melting point and lipophilicity and
decreasing vapor pressure and water solubility decrease. As a result of the differential
chlorination, lighter PCB congeners are much more apt to volatilize to the atmosphere than the
heavier congeners. These properties provide for some confounding environmental behavior for
scientists attempting to generalize PCBs as a whole. Despite the differences amongst individual
PCB chemistries, their general presence can still be measured globally in various environmental
media even though production and use of PCBs has been banned in many countries decades ago

73

(Muir et al., 2000; Xing et al., 2005; Macgregor et al., 2010). The transport, distribution,
environmental behavior, and global fate of PCBs, is therefore of continuing concern.
As described above, PCBs can be linked to many aspects of anthropogenic presence.
Since the start if its commercial manufacture, PCBs have entered into the environment from
accidental spills during transport and handling, leaks from electrical transformers or other
equipment utilizing PCBs as a coolant, or by incineration of wastes containing PCBs (Voogt et
al., 1989; Rachdawong and Christensen, 1997). Also, redistribution of PCBs already in the
environment has been shown to be a contributing source (ATSDR, 2000). Even though the
production and use of PCBs has been discontinued in the United States, PCBs are still released
during some industrial processes. It is known that PCBs may be formed whenever a carbon
source and chlorine are combusted together, such as during municipal and hazardous waste
incineration (Bergman et al., 1984; Brown and Ganey, 1995; Alcock et al., 1999). Accidental
production is estimated to account for over 50 tons of PCBs released into the environment
annually. The USEPA has identified 80 chemicals which have a production means prone to
producing PCBs as a byproduct (Callahan et al., 1983).

74

Figure 3.1: A schematic of the biologically mediated transport of PCBs to the inland lake ecosystem.
Initially, PCBs are emitted into the environment from industrial, municipal, or agricultural regions into the
atmosphere or directly to surface water bodies. Much of the PCBs in Michigan are transported by air to the
Great Lakes where they can become widely distributed around the entire Great Lakes Region. Biota in the
higher trophic levels of the Great Lakes ecosystem (including piscivorous birds and fish) bioaccumulate the
PCBs primarily in fatty tissues. The wildlife then travels to inland where the PCBs are then deposited inland
by delivery of fecal matter, molting, or mortality of the organism. From Blais et al., 2007.

Once released into the environment, the compositions of commercial PCB mixtures are
altered through processes such as volatilization and other kinds of partitioning, chemical or
biological transformation, and preferential bioaccumulation (Figure 1). The degree to which
these processes occur is highly dependent upon the degree of chlorination of the biphenyl
molecule. Generally, the lower chlorinated congeners of PCBs are more apt to volatilize to the
atmosphere and are generally present in elevated concentrations relative to the higher chlorinated
congeners, especially in northern latitudes (Halsall et al., 1999; Harner et al., 1998; Ockenden et
al,. 1998). The more highly chlorinated congeners tend to adsorb more strongly to sediment and
soil (Smith et al., 1990; Oliver and Niimi 1988). All PCBs rely upon the high bond strength of
the carbon-chlorine covalent bond for its environmental stability and long half-life (8 to 10
years) (Hillery et al., 1997), particularly if buried in soil or sediment (Sinkkonen and Paasivirta,
2000). The most critical step in the decomposition of PCBs is the cleavage of this carbon75

chlorine covalent bond. In lake sediments, this dechlorination usually occurs through hydrolysis
or through reductive dechlorination of the compound.
In addition to the aforementioned carcinogenity, compounds in the PCB group share a
structural similarity to, and toxic mode of action as, dioxin, with health effects such as endocrine
disruption and neurotoxicity. Health concerns have served as the impetus for development in the
understanding the sources, fate and transport of PCBs and has been the goal of many bodies of
research since the 1970s accordingly (Schuster et al., 2011; Schuster et al., 2010; Wong et al.,
1995; Hillery et al., 1997; Muir et al., 1996; Eisenreich et al., 1989; Murphy and Rzeszutko,
1977; Nisbet and Sarofim, 1972). PCBs history as an environmental contaminant begins in its
first documented detection of fish in Swedish lakes (Jensen, 1966). Since this time, findings in
the UK show that local urban sources cause PCB concentrations to be amplified 5-10 times
compared to rural regions (Schuster et al., 2010; Sun et al., 2007; Totten et al, 2004). Other
studies link environmental PCB concentrations specifically to more populace areas of the U.S.
and decrease away from these locations (Shen et al., 2006; Golden et al., 1993). The same trend
was also found in the European atmosphere with a general spatial trend showing a decline in
concentration from high population density in the United Kingdom to low population density in
Norway (Schuster et al., 2010).
In the Great Lakes region of the United States, research has shown that urban areas such
as Chicago, Illinois and Milwaukee, Wisconsin serve as major sources of PCBs to the Great
Lakes (Sun et al., 2007; Hafner and Hites, 2003; Buehler and Hites, 2002; Strandberg et al.,
2001; Eisenreich et al., 1981). Also, most PCB inputs are shown to be locally sourced and
atmospheric appears to be the dominant pathway for the transfer of PCBs from land to natural
waters (Offenberg et al. 2005, Hoff et al, 1996; Sweet et al, 1993; Swackhamer et al, 1988;

76

Eisenreich et al. 1979, 1981). It was estimated that 55-90% of the PCBs entering Lake Superior,
Huron, and Michigan and 7-13% of the PCBs entering Lakes Erie and Ontario come from
atmospheric inputs (Hoff et al, 1996; Strachan et al., 1979, Strachan and Eisenreich, 1990;
Eisenreich et al, 1979, 1981). PCB emissions in the Chicago atmosphere lead to dramatically
increased atmospheric concentrations and deposition to southern Lake Michigan compared to the
regional signal (Offenberg et al., 2005). This indicates that near urban areas, localized, urban
sources are relatively more dominant than long-range atmospheric transport.
Due to the high degree of PCB emissions, the Great Lakes region of the United States has
been closely monitored. From 1991 to 1997, studies have shown the change in atmospheric PCB
concentration at the city of Chicago and Lakes Superior, Michigan, and Erie (Simcik et al. 1999).
Gas-phase concentrations were found to have decreased in Chicago and near Lake Michigan and
Erie, but remained fairly constant near Lake Superior. Inputs of PCBs to the Great Lakes region
are influenced heavily by atmospheric transport and deposition (Franz et al. 1998, Jeremiason et
al. 1998). A strong correlation between sediment accumulation of PCBs and dry deposition to
Lake Michigan has also been observed and suggests that dry deposition may account for most of
the particulate PCBs accumulating in the sediments of Lake Michigan (Franz et al. 1998).
In most research, PCBs appear in the sediment record around 1930 in the Great Lakes
region, at the time they were first produced, and remain at very high levels between 1960 and
1980 (Helm et al., 2011; Hu et al., 2011; Golden et al, 1993; Eisenreich et al., 1989; Oliver et al.,
1989; Rapaport and Eisenreich, 1988). This peak is followed by a decrease after 1980 when they
were banned by the US government. Concentrations in soils and sediments have followed a
downward trend over time and appear to have reached a steady state concentration in several
locations (Bopp et al. 1998; Van Metre et al. 1998; Lead et al. 1997). Review of past water

77

monitoring studies indicate that detectable concentrations of PCBs are widely found in surface
waters but their low solubility generally prevents them from reaching high concentrations,
especially in groundwater (USEPA, 1980). PCB concentrations have also been shown to be
decreasing in surface water bodies since the late 1970s due to the cessation of production and
manufacturing (Anderson et al., 1999; Jeremiason et al., 1998).
Lake sediment acts as strong sinks for chemicals like PCBs because of their persistence
and hydrophobicity (Blais and Muir, 2001). As described above, lake sediment records the
historical PCB concentration as the particles that they are absorbed to settle out and are deposited
at the bottom of a lake and, therefore, provide an excellent means of determining a chronological
sequence of deposition. However, these sediments also record both atmospheric inputs and
inputs from the surrounding watershed and differentiating between these two inputs remain an
issue. Much of the existing research includes collected and dated sediment cores with measured
total PCB concentrations at different depths of the cores to construct temporal profiles of PCBs
(Muir et al, 1996; Golden et al, 1993; Eisenreich et al, 1989). Studies of total PCB temporal
profiles do adequately reflect the production and usage history of PCBs (Helm et al, 2011; Hu et
al 2010; Golden et al, 1993; Eisenreich et al, 1989; Oliver et al, 1989; Rapaport et al, 1988) but
in some cases fail to account for the differences in chemistry amongst the congeners.
The concentration differences through time among different lakes can utilize temporal
profiles to include a spatial component to these studies. Around 1970, total PCBs concentration
in Lake Superior sediment peaks at 15 ng/g while in Lake Ontario it peaks at 800 ng/g. The
concentrations had then decreased to about 7 and 200 ng/g respectively in lake sediment dated
approximately to 1990 (Golden et al., 1993). It has also been shown that the percentage of lower
chlorinated congeners (tri to penta) decreased with depth, while hexachlorobiphenyls remained

78

fairly constant throughout and the concentration of highly chlorinated congeners increased with
depth (Oliver et al., 1989). It was determined that the deposition rate of PCBs from the
atmosphere to the Great Lakes is approximately equal to the amount evaporating from the lakes
to the atmosphere. This suggests PCB cycling in the Great Lakes ecosphere may have reached a
steady state (Simcik et al., 2000). These studies of PCB sediment records in the Great Lakes
have provided much in the way of understanding large-scale regional trends of these chemicals.
Studying inland lake sediment cores has the benefit of providing insight into the PCB
inputs on a smaller scale, when observed on the individual lake basis. Also, data from a number
of smaller lakes across a region can provide regional signals in the same vain as the Great Lakes
studies when results from several lakes in proximity to one another are compared. This strategy
has provided understanding in the behavior of many anthropogenically sourced trace metals
(Parsons et al., 2007; Yohn et al., 2002). However, few studies have examined the spatial and
temporal trends of PCBs using many inland lakes across a large region (Golden et al., 1993).
The ability to examine the results of distinct PCB congeners on a large scale can provide insight
into the composition of the emissions, which may be quite variable on a local scale (Wania and
Su, 2004). Applications of such studies provide insight into global fractionation, which results in
changes in the composition of persistent organic pollutant mixtures with latitude (von Waldow et
al., 2010; Wania and Mackay, 1993) and PCB dechlorination, where more highly chlorinated
congeners dechlorinate to lower chlorinated congener through time (Magar et al, 2005).

79

Hypothesis and Test
The hypothesis driving this research is that during the time of high PCB production
(1970s) the influence of population centers on loadings will cause a localized spatial pattern in
PCB distribution with elevated concentrations in watersheds occupied primarily by urban land
use. With the prohibition of manufacture and use of these chemicals and with the direct evidence
of a steady air concentrations of PCBs persisting in the Great Lake region to the present
(Strandberg et al., 2001), it follows that recent trends should exhibit PCB loadings more
reflective of a regional signal, with larger distribution gradients between available sources of the
chemicals independent of watershed boundaries. Ancillary to this hypothesis is that congener
data should provide a gradient across the State of Michigan with the heavier PCB congeners
located more closely to primary sources in the southern, more populace end of the state, and the
light PCB congeners primarily located in the northern portion of the state (Upper Peninsula).
To test these hypotheses, sediment cores from 44 inland lakes were collected from across
the State of Michigan representing lakes within various land use, watershed characteristics, and
trophic status. The sediment-chemical chronologies were studied in these cores. If the
hypotheses is true, then: 1) total PCB inventories in Michigan inland lakes would show a south
to north decreasing pattern due to larger population and more industries in the south, and would
show a relatively higher PCB concentrations in the urban/industry area, 2) temporal PCB
concentration profiles would show significant differences during high PCB production years
among different lakes, with higher concentration near urban areas and lower concentration in the
remote areas but the difference among lakes would become smaller in recent years, and 3) the
relative abundances of the PCB congeners in sediment samples during the period of high
production should show a characteristic change from south to north. This change would be

80

evident during the high production period then in more recently deposited sediment, when
loading and atmospheric mixing would work to reduce the intensity of the change.

81

Methodology
Sampling Methodology
The number of inland lakes in Michigan exceeds 11,000 (MIDNR, 2013) and lakes
chosen amongst these for sediment sampling were based on a set of criteria including but not
limited to: availability of boat access to the lake, sufficient depth to preclude significant
bioturbation of sediment (generally lakes >25 feet depth), avoidance of lakes potentially mined
for marl or had other significant sediment disturbances, and maximizing spatial range of the
project. Sediment core collection was conducted using an MC-400 Multi-corer modified for
inland lake use. This instrument was deployed from the M/V Nibi built to house the multi-corer
and designed for access to shallow water entry and exit.
Once collected, cores intended for 210Pb and metals analyses were extruded and
sectioned at 0.5 cm intervals for the top 8 cm, and at 1 cm intervals for the remainder of the core.
The 210Pb analyses were conducted by Freshwater Institute in Winnipeg, Manitoba, Canada and
determined porosity, accumulated dry mass, sedimentation rates, sediment ages and focusing
factors. Dating models were verified using 137Cs, the temporal peak of the stable 208Pb
concentrations and presence/absence of excess 210Pb.
Another sub-core was sectioned and used for analysis of a suite of trace persistent organic
pollutants (POPs) that included PCBs. Unlike the 210Pb sub-core, it was sectioned at 2 cm
increments for the entire core length to provide sufficient sediment mass for the required
analyses. Several lakes sampled prior to 2003 were sampled again in subsequent years both in
order to recover the POP analysis not described earlier and to provide a measure of quality
assurance for other analytes (210Pb, trace metals, and others). Lakes sampled twice include Lake
Cadillac, Cass Lake, Crystal Lake (Montcalm County), Crystal Lake (Benzie County), Elk Lake,

82

Gratiot Lake, Gull Lake, Higgins Lake, Muskegon Lake, Paw Paw Lake, and Whitmore Lake.
For clarity, lakes with similar names will include an abbreviation of the county in which it is
located (e.g., Round D, D for Delta County). Lakes that have been revisited will include the year
of resampling (e.g. Gratiot07 is Gratiot Lake in Keweenaw County sampled in 2007).
Analytical methods for PCBs have been described elsewhere (Khim et al, 1999). Since a
2 cm sediment increment provided inadequate mass, equal amount of sediment from each sample
were pooled in the laboratory to obtain adequate material representing each layer. A total of 40 g
of wet sediment was then weighed and extracted for analysis. Moisture content was determined
by drying 5 g of wet sediment overnight at 100C. The sediment was homogenized with 120 g of
anhydrous sodium sulfate and extracted with dichloromethane and hexane (3:1, 400 mL) in a
Soxhlet apparatus for 16 hours. Extracts were then concentrated to 1 mL. The sample cleanup
was done using an SPE cartridge packed with Florisil (Waters, Sep-Pak Florisil 6 mL). The
cartridge was eluted with 10% acetone in hexane (9 mL). The eluent was concentrated to 1 mL
and placed in gas chromatograph vial for analysis of PCBs. Di through Deca polychlorinated
biphenyls were added as internal standards. PCBs were quantified using bracketed calibration
generated by averaging two calibration curves that were run before and after each set of 10
samples to better adjust to changes in instruments during analytical runs.

Study Setting
Sediment cores were collected from 44 individual inland lakes from 1999 to 2010 (Figure
1) as part of the Michigan Department of Environmental Quality’s (MDEQ) Inland Lakes
Sediment Trend Monitoring program. As described above, 11 lakes were sampled twice which
totals 58 separate sample events throughout the State of Michigan. Several lakes early in the

83

project were not analyzed for POPs such as PCBs. The total usable sediment cores for this
research then totaled 44 (Figure 2).

Figure 3.2: 44 lakes studied with full core subject to PCB analysis. Note that this is a subset of the 47
individual lakes cored total, 3 of which only top 4 cm were analyzed for PCBs.

Michigan covers 58,528 sq miles of which land takes up 56,954 sq. miles and inland
water approximately 1,573 square miles. Sediment core collection ranged from lakes located
along its southern border (e.g. Klinger Lake, Latitude: 41.80) to lakes located at its northern
extent (e.g. Gratiot Lake, Latitude: 47.37). This wide geographic distribution of lakes sampled
represents a variety of watershed land use types (Figure 2) ranging from largely urban near major
cities (e.g. Cass Lake, Oakland County) to those that are remote with watersheds almost
completely forested (e.g. Emily Lake, Houghton County). Further, lakes were sampled to
represent a large range of physical characteristics. These include but are not limited to: physical

84

size (depth, area), watershed size, proximity to Great Lakes, and trophic status. Table 1 (below)
is included detailing a list of the lakes sampled and some physical characteristics for this
research.
Land use data were obtained from the 2001 IFMAP/GAP Lower and Upper Peninsula
data sets (MDNR, 2003) for the watershed of each study lake. Land use was classified using the
upper-most level Anderson classification and resulted in 7 fundamental land use types: Urban,
Agriculture, Forest, Barren, Upland, Water, and Wetland. Watersheds were delineated from the
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, 2011).
Age Dating
The radioactive isotope 210Pb was used to date sediments from each lake. Several models
exist to determine sediment ages from 210Pb activities, and sediments were dated using the
constant flux: constant sedimentation rate model (CF:CS) (Golden et al., 1993), segmented
CF:CS (SCF:CS) (Heyvaert et al., 2000), rapid steady state mixing model (RSSM) (Robbins,
1982), and the constant rate of supply model (CRS) (Sanchez-Cabeza et al., 2000). For all
models, only sediment layers containing excess-210Pb could be dated. The ages of sediment slices
not containing excess-210Pb were estimated by extrapolation using the assumption that
sedimentation rates remain constant below the existence of excess 210Pb.
Sedimentation rates in each lake were determined using all models possible for that lake,
and then the models were evaluated to ascertain which was the most appropriate for determining
sediment ages. Profiles of 137Cs activity and stable lead concentration profiles were used as
indicators to determine the most appropriate model (Robbins, 1978). Focusing factors were also
85

determined from 210Pb analysis. Sediment focusing occurs when fine-grained sediments in a lake
are eroded from higher energy erosional zones near the shore of the lake, transported through
transitional zones (where deposition and erosion occur episodically) and deposited in
depositional zones (Downing and Rath, 1988; Hakanson, 1977). This process of focusing occurs
to different extents among lakes, and must be accounted for by using the focusing factor before
comparing inventories and accumulation rates among lakes. Focusing factors (FF) were
estimated at each site from the ratio of unsupported 210Pb inventory measured in the core to the
210

Pb inventory expected from atmospheric deposition (Muir et al., 1996; Golden et al., 1993).

Figure 3.3. Log 210Pb activity (Bq/g) versus accumulated dry mass (g/cm2) in Lake Gogebic. Regression
lines for the CF:CS (constant flux-constant sedimentation) model and the SCF:CS (segmented constant
flux-constant sedimentation) model are shown.

In addition to information regarding sediment age, rates of sedimentation and general
quality of the sediment core can be determined from 210Pb analysis. Results from lakes used for
this study show no indication of significant disturbance in the sediment cores. However, results
do indicate some sediment mixing near the surface of the core. The surficial sediment mixing is
86

likely influenced by benthic organisms (bioturbation). This mixing zone is determined by
plotting accumulated dry mass versus log activity of 210Pb, where the slope is related to
sedimentation rates. The non-linear portion near the surface is considered the mixing zone
(Figure 3). The estimated mixed depth is listed in Table 1.

87

Results and Discussion
PCB Temporal Concentration Analysis
Due to the scale and scope of the Michigan Inland Lake Sediment Trends Project, lakes
were considered in three groups based on Albert, et al. (1986, 1995) regional classification of
Michigan (Figure 4). These lakes were grouped based on the relative homogeneity of climate
and physiography identified in the landscape ecosystems. In their work, Albert, et al. (1995) had
the Upper Peninsula separated into two distinct zones. However, since the amount of lakes
meeting the depth and accessibility criteria needed to be utilized in this study were much less
than those in the lower portions of Michigan, the Upper Peninsula here will be considered as a
whole.

Figure 3.4: Terrestrial ecoregions framework as defined by Albert (1995). The separate Upper Peninsula
ecoregions were considered together for this study due to the lack of suitable lakes meeting the criteria
needed to qualify them for this project.

88

Figure 3.5a: Temporal Concentration Profiles of lakes in the Upper Peninsula of Michigan since the year
1900.

Lakes in the Upper Peninsula elicited PCB concentrations generally with temporal
concentrations below 50 ng/g dry wt. The only exception is Teal, which has its highest value in
the most recent sediment slice at 229.42 ng/g (Figure 5a). There appears to be no clear historical
trend found in these lakes such as an increase to the time frame of peak production of PCBs
followed by a decrease to recent sediments. Due to the remote locations of U.P. lakes, the
sediment concentrations are assumed to be representative of uncontaminated ecosystems
dominated by atmospheric deposition with the noted exception of Teal Lake, which has much
more urban land use within its watershed than any other U.P. lakes studied. Land use for Teal
was 27.01% and 20.01% urban in the 1970s and 1990s MiGDL data layers (respectively). The
next largest watershed percent urban was Imp Lake with 7.14% 1970s and 3.46% 1990s.
Overall, the average concentrations of PCBs in U.P lakes are relatively low compared to
lakes in the southern areas of the State and the most important decade for PCB deposition here is
the 2000s (Table 2). Both the spatial remoteness of these lakes from significant sources and the

89

number of years since the peak production and ban of PCBs points to atmospheric deposition is
as being a significant source of PCBs in this region.

Figure 3.5b: Temporal Concentration Profiles of lakes in the northern portion of the Lower Peninsula since
the year 1900. The peak of Lake George was not included for clarity of the lower concentration lakes.

In the northern portion of the Lower Peninsula, the temporal profiles vary significantly
from lake to lake (Figure 5b). Like the Upper Peninsula, lakes here generally maintained values
below 50 ng/g with the exception of George, Cadillac and Houghton. These lakes did have the
presence of relatively elevated % urban watershed land use compared some many of the other
lakes in this region (9.66%, 19.93%, 7.23% urban 1970s, respectively). However, many other
lakes in this region with significant % urban such as Hackert (21.53%), Higgins (15.02%), and
Fremont (20.46%) did not exhibit concentration peaks as large. Also, lakes in this region of
Michigan appear to exhibit earlier peak trends relative to study lakes in the U.P. These trends
are interpreted as being the result from closer proximity to source than the more remote lakes in
the U.P.

90

It appears, in this region of Michigan, land use is either not one of the fundamental
drivers of PCB loadings or one of several other contributing watershed or regional features
serving as sources of PCBs to these lakes. Table 3 shows the lake sediment of lakes in this region
to be, on average, lower than those lakes in the Upper Peninsula despite the proximity of larger
urban centers (e.g. Traverse City, Alpena, etc.). It is believed that trophic status of the lakes in
this region to play an important role in the sediment contaminant record. Lakes studied in this
region such as Torch, Crystal (Benzie County), Clear, Higgins, Black, Elk and Charlevoix are all
known to be oligotrophic (low water nutrient concentrations) (Fuller and Minnerick, 2008).
Thus, the sediment is lacking much of the organic material PCBs would normally be adsorbed to.
It is possible that this reduces the flux of PCBs to the sediment. The decadal block most
important to deliver of PCBs to the sediment in this region is the 1980s. Since this is the decade
after PCBs were prohibited, it is possible that the PCB signal in this portion of Michigan is
derived from a combination of atmospheric deposition from sources further south and west along
with point sources of PCBs in the individual watersheds.

91

Figure 3.5c: Temporal Concentration Profiles of lakes in the southern portion of the Lower Peninsula since
the year 1900. The peak of lakes Thompson and White were not included for clarity. Note the change in
PCB Concentration scale of the X-axis relative to those of the Figures 3.5a and 3.5b.

The Southern Lower Peninsula region is characterized by significantly higher
concentrations of PCBs than the lakes described above (Table 2). Much like the Northern Lower
Peninsula region, these temporal trends do not appear to closely follow watershed land use as
anticipated. For example, the two lakes with the highest concentration peaks, Thompson and
White, at much different land use characteristics. Thompson had 1970s % urban calculated at
22.90% and White had the same land use calculated at 3.33%. If should be noted that White
Lake has historically been an EPA Area of Concern and remediation steps have been taken to rid
the lake of many contaminants contained in the sediment, including PCBs (MDNR, 2013;
USEPA, 2013). Thus, it appears that PCB concentrations detected in lake sediment are more
closely tied to particular industries that utilize the chemical than they are a particular land use.

92

Table 3.1: PCB concentrations of all study lakes averaged by decade and concentrations averaged by decade
for each of the three regional zones. UP = Upper Peninsula, LP North = Northern Lower Peninsula, LP
South = Southern Lower Peninsula. Values are normalized to peak concentration of the individual lakes for
the table on the right.
Average Decadal Lake Concentrations
(ng/g)

2000s
1990s
1980s
1970s
1960s
1950s
1940s
1930s

Decadal Lake Concentrations Normalized to
Peak

All Lakes

UP

LP North

LP
South

26.87
39.47
79.39
83.11
24.45
40.04
2.85
3.12

33.49
24.80
9.93
13.50
10.79
9.98
6.14
7.96

18.34
21.12
41.71
6.39
4.35
1.87
1.97
2.09

29.87
59.84
138.39
172.60
48.62
83.71
1.40
0.56

All Lakes

UP

LP North

LP
South

0.31
0.41
0.37
0.26
0.23
0.11
0.10
0.11

0.60
0.47
0.30
0.38
0.31
0.32
0.22
0.21

0.36
0.46
0.40
0.17
0.27
0.07
0.10
0.17

0.14
0.35
0.37
0.27
0.16
0.04
0.01
0.01

2000s
1990s
1980s
1970s
1960s
1950s
1940s
1930s

Table 3.2: Focus-Corrected Accumulation Rates (ug/m2/yr) of all study lakes averaged by decade and
concentrations averaged by decade for each of the three regional zones. The highest value in each category is
colored grey. UP = Upper Peninsula, LP North = Northern Lower Peninsula, LP South = Southern Lower
Peninsula.

2000s
1990s
1980s
1970s
1960s
1950s
1940s
1930s

AVERAGE
0.64
0.63
0.93
1.39
1.41
1.09
0.45
0.09

UP
0.84
0.70
0.89
0.43
0.25
0.12
0.14
0.22

LP North
0.32
0.33
0.55
1.06
0.51
0.59
0.08
0.04

LP South
0.82
0.87
1.29
2.07
2.82
1.90
1.10
0.04

Table 3.3: Focus-Corrected PCB Inventories (ug/cm2) of all study lakes averaged by decade and
concentrations averaged by decade for each of the three regional zones. Highest value in each category in
grey. UP = Upper Peninsula, LP North = Northern Lower Peninsula, LP South = Southern Lower Peninsula.

2000s

All Lakes
4.47

UP
8.24

LP North
2.10

LP South
4.88

1990s

6.33

6.51

3.30

8.80

1980s

8.72

7.47

5.46

11.99

1970s

14.05

2.78

11.25

21.15

1960s

15.37

2.83

4.46

30.66

1950s

12.18

1.12

5.26

23.90

1940s

10.67

1.47

0.63

25.30

1930s

1.36

2.43

0.44

1.72

93

Figure 3.6: Historical records of the sales/production volumes of PCBs and the similarity of these timevarying trends to the accumulation rates of these chemicals in the sediments of Lake Ontario (from
Eisenreich et al., 1989)

The decadal averages of these lakes exhibit average concentrations to be greatest in
1970s, closest to the peak U.S. sales of these chemicals as described by Eisenreich et al, 1989
(Figure 6). This is probably due to the fact that the lake watersheds of the Southern Lower
Peninsula region are generally more populace and closer to the sources of PCBs for both point
sources and regional atmospheric deposition. However, on an individual basis, Michigan inland
lakes only occasionally appear to have recorded a clear signal of U.S. sales of PCBs as seen by
Eisenreich et al., 1989 in the Great Lakes.
Care should be taken when comparing concentrations of organic contaminants among
lakes. Due to the low amount of sediment in the top-most sections of the core, the top two slices
were combined, and slices 3 and 4 were combined, to create two samples. As a result of
combination and variable sedimentation rates among lakes, each sample represents a range of
94

deposition years. For example, Lake Geneserath (sampled in 2010), had an extremely low
sedimentation rate for much of the described core. Thus, the surficial sediments represented a
larger range of sediment ages and might be expected to have higher concentrations of PCB
contamination than Cass Lake, which had a much higher sedimentation rate (sampled in 1999).
As Michigan hosts inland lakes with a wide range of trophic status, watershed sizes, and
morphologies, sedimentation rates are quite variable. Many of the nutrient-poor, oligotrophic
lakes reflect low sedimentation rates such as Clear Lake with a sedimentation rate of 294
g/m2/yr. Conversely, other lakes, typically eutrophic in nature, exhibited elevated sedimentation
rates such as Morrison Lake with a sedimentation rate of 2,590 g/m2/yr. (See Fuller and
Minnerick, 2008 for Michigan lake trophic classifications).

Figure 3.7: Sum of PCB congener concentration (ng/g dry wt) in surficial sediments of inland lakes across
Michigan organized by latitude.

95

Due to the wide range of sediment accumulation within these lakes, it is difficult to draw
conclusions simply from examination of surficial sediments.

As expected, PCB totals measured

in the top 4 cm allowing for graphical comparison of all 44 sample events show a wide range of
PCB total concentrations (Figure 7). This is reflective of the rate at which the lakes are
accumulating both sediment and the actual contaminant under study. A more adequate
assessment of PCB impacts to a surface water body may be in its accumulation rate or total
inventory of the PCB quantity. PCB accumulation rate accounts for the sedimentation rates and
sediment focusing while PCB inventories measure the total quantity of PCBs per unit area.
These measurements and how they are calculated are described further in the PCB Accumulation
Rate Analysis and PCB Inventory Analysis below.
PCB concentrations in the top portion of sediment are of special concern as they provide
an estimation of the amount of PCBs potentially available for redistribution (i.e., are not
permanently buried) into the water column by sediment-water exchange (as described by Wong
et al, 1995). Surficial sediments are of also much more directly accessible to benthic biota, and
thus, the remainder of the aquatic food web. The PCB surficial concentrations appear quite
variable across Michigan with elevated quantities found as far south as Cora Lake (latitude
42.20) and as far north as Teal Lake (latitude 46.51). There does not appear to be a significant
latitudinal trend associated with the distribution of elevated PCB concentrations and lakes
exhibiting high surficial concentrations maybe receiving continued inputs from watershed
sources. Comparing the 0-2 cm core slice and the 2-4 cm core slice in most of the impacted
lakes (such as Cora, Teal, and White Lakes) show increasing trends of PCB concentrations to the
surface sediments.

96

White Lake in particular is anticipated having elevated values of PCBs throughout its
core as it has been listed as an EPA Area of Concern since 1985 (EPA, 2013[4]). The Occidental
(Hooker) Chemical Company and the Whitehall Leather Company were the primary sources of
contamination and measures were taken to remove sediments associated with the two
contaminant “hot spots” in 2002. A variety of organic chemicals has historically been
discharged to White Lake including PCBs.
Examination of PCB concentrations to the maximum depth of the cores (approximately
40 to 60 cm) usually provides the entire PCB temporal history of the given lake. The only
exceptions of are those lakes with high sedimentation rates. High sedimentation rates prevent
the recovered sediment from the coring device from reaching the lakes “background” values (i.e.
prior to anthropogenic disturbance) and a longer (deeper) core is needed. The “Oldest Section”
column of Table 1 illustrates this by noting the 210Pb determined date of the deepest recovered
sample in the given lake.
Averaging decadal concentrations for the entire database reveals that the most significant
increases occur in lakes from 1940s to 1950s and from 1960s to 1970s. The historic peak is then
observed, on average, to occur in the 1970s as expected with a significant average concentration
following into the next decade. This could be representative of the lag time of soil and
waterborne PCB inputs into the environment during the 1970s (again, PCBs were banned in
1979), continued input of PCB containing equipment and unlined landfills, and continued
atmospheric rainout of extra-watershed PCBs. This is followed by recovery to the present
possibly due to the removal of most sources and the continued breakdown released PCBs.
Finally, it is noted that trace concentrations of PCBs appear to exist in the years prior to its
starting use. This could possibly be attributed to the smearing effect occurring upon corer

97

insertion into the benthic sediments. However, measures are taken during field sampling to
remove sediment in direct contact with the coring tube and the coring tube assembly is cleaned
between lakes. The field methodology designed to prevent this makes these early detections
seem improbable. Post-depositional mobility of the PCB compounds offers another possibility
for their early detections. The PCB concentrations found in sediment dated to earlier than 1930
were comprised primarily of the less chlorinated or “lighter” and thus more mobile, fraction of
the PCB congeners.

Figure 3.8: Comparison of averaged 210Pb dated decadal block of normalized Pb concentrations from 1930s to
present in three physiographic zones described by Albert et al., 1995. Note: not all lakes were available for
Pb inventory analysis as it requires recognizable background (reference) trends of Pb which are not
recognized in several cores.

98

Figure 3.9: Comparison of averaged 210Pb dated decadal block of normalized PCB concentrations from 1930s
to present in three physiographic zones described by Albert et al., 1995.

PCB spatial trends are compared to lead (Pb) inventories to examine correlations in their
spatial distribution. Pb inventories are calculated by combining the total sediment abundance of
Pb per unit area (i.e. by considering vertically integrated sediment inventories). Corrections are
applied to the Pb inventories to account for the watershed erosional inputs and sediment focusing
(Discussed in PCB Inventory Analysis below). This allows for conclusions to be drawn about the
anthropogenic-specific quantity of Pb atmospherically deposited to each lake. Work on trace
metals distribution and histories in Michigan lakes show that lead (Pb) inventories across
Michigan can be roughly divided into three separate zones (Vannier et al., 2013). These zones
coincide with the three climate and physiographic zones as described by Albert et al., 1995
(Figure 4). Given the known sources of Pb to the environment, anthropogenic inputs to
Michigan surface water bodies are more likely related population densities. The southern-Lower
Peninsula zone is the most populous area of the State and contains the most elevated Pb
99

inventories, the Upper Peninsula the lowest, and the northern half of the Lower Peninsula has Pb
inventories generally between these two (Vannier et al., unpublished data).
As shown in Figure 8, despite relative latitude or Pb inventory, Pb concentrations peak in
the 1970s. This is attributed to the effectiveness of environmental legislation (e.g. Clean Air Act,
Clean Water Act) as this period coincides with the time these major legislative bodies were
enacted. However, comparing inventories to the peak PCB concentration dates in these lakes
reveals a gradient from south to north, with the normalized average PCB concentration in each
zone tending to peak earlier in the south than in the northern extents of the State (Figure 3.8).
If the PCB distribution is driven by direct atmospheric deposition to the northern latitude
lakes, the profiles of these lakes should follow the PCB sales history with an observable peak in
the late 1960s. However, lakes in the northern extents of the State have a tendency to peak in the
most recent decade and could be interpreted as evidence for secondary deposition of PCBs from
its primary source locations (near populace areas) to more remote locations (e.g. Upper
Peninsula, Michigan). This was hypothesized by others (Wania and Dugani, 2003; Wania and
Mackay, 1995, 1996) that a gradual movement of PCBs northward, caused by temperaturedependent partitioning results in elevated PCB concentrations in relatively recent surficial
sediments (dated to 1980s and early 1990s). Hassanin et al., 2005 also observed decreasing
percentages of lighter congeners with increasing latitude while the opposite is true for heavier
congeners. Such observations have been explained by the global fractionation theory (Wania and
Mackay, 1995).
In general, spatial and temporal total PCB concentration results correspond well with
previous reports (Hu et al., 2010; Golden et al., 1993; Eisenreich et al., 1989; Rapaport et al.,
1988). The low PCB concentrations in Lake Superior which is close to the Upper Peninsula

100

reflect atmospheric inputs and the higher concentrations in Lake Michigan and Ontario reflect
regional and localized atmospheric deposition (Golden et al., 1993).

PCB Accumulation Rate Analysis
Accumulation rate of PCBs in the sediment cores were calculated as follows:

Accum (ng / cm 2 / yr )  Csed (ng / g )  W ( g / cm 2 / yr )
Where Accum = PCB accumulation rate (ng/cm2/yr),
Csed= concentration of PCB in surficial sediment (ng/g dry wt), and
W= mass sedimentation rate (ng/cm2/yr) based on 210Pb dating results.
Accumulation rates of PCBs were then corrected for sediment focusing which is
calculated as raw accumulation rate / 210Pb determined focusing factor. Focusing factor is the
calculated ratio of the unsupported, or atmospherically derived, 210Pb inventory to the expected
210Pb inventory (Eisenreich et al., 1989; Simcik et al., 1996). It is a positive, unitless ratio
accounting for the tendency of lake sediment to become redistributed from shallow portions of
the lake (via wind or water movement energy) to the deeper portions of the lake (after Golden et
al., 1993). Simcik et al. (1996) selected 15.5 picocuries cm-2 as the expected value for the entire
region. This concept was used in studies of eastern Lake Ontario sediments (Eisenreich et al.,
1989) and Lake Superior sediments (Jeremiason et al., 1994). In this work, the value selected by
Simcik et al. (1996) is used to retain uniformity in all lakes studied. As sediment focusing at the
point of maximum depth where the samples are collected always results in a net gain of material
and, when the correction is applied, the concentration is less than raw value.
Figure 10 and 11 illustrates the general reduction in the rate at which inland lakes in
Michigan are accumulating PCBs since the 1970s: the averaged focusing corrected accumulation
rates for sediment in the 1970s vary from approximately 0 to 9 ng/cm2/yr (average 1.33
ng/cm2/yr). In the 2000s, these values have been reduced to 0 to 5 ng/cm2/yr (average 0.63
101

ng/cm2/yr). This reduction is expected as the 1970s accumulation rates encompassed the
production peak year (around 1970). It appears from these figures that atmospheric fallout is the
primary pathway of PCBs to most lakes in the northern extents of the State. This is concluded
due to the relatively low watershed populations, the general remoteness of the lakes (e.g. low %
urban), and the similarities between lake accumulation rates as one would expect from a mixed
atmospheric signal. This is considerably less than those lakes located along the more populace
latitudes to the south giving indication of point-source or sub-regional PCB pollution providing a
significant quantity PCB to these lakes, both historically and in the present. It also appears that
atmospheric fallout of PCBs hasn’t decreased substantially in many lakes as lakes in the extreme
southern and northern extents of Michigan are receiving nearly as much (in some cases more)
PCB inputs currently than they have been in the 1970s.
As with the spatial analysis of PCB concentrations, PCB accumulation rates also follow a
general trend of peaking temporally later as latitude increases (Table 3.3). This is consistent
with the interpretation that the observed trend is a result of distillation of PCBs evaporating and
condensing at higher latitudes due to temperature differences. It is also possible that this effect
represents the lag time between the deposition of atmospherically derived PCBs to the watershed
and the hydraulic washing of the chemicals to the lake sediment. However, watersheds of the
lakes studied in the Upper Peninsula average to be much smaller than those of the Lower
Peninsula (approximately 170 sq mi versus 54 sq mi), thus, it would be anticipated that the lag
time effect would be reduced compared to those lakes in lower Michigan.

102

Figure 3.10: 1970s PCB Accumulation Rates for all lakes organized by latitude. White Lake peak removed
for clarity.
1
: From Eisenreich et al., 1989.

Figure 3.11: 2000s PCB Accumulation Rates for all lakes organized by latitude. White Lake peak removed
for clarity.

103

PCB Inventory Analysis
PCB inventories are calculated by combining the total sediment abundance of PCBs per
unit area (i.e. by considering vertically integrated sediment inventories). Similar to both
concentrations and accumulation rates of PCBs, the inventories vary significantly amongst the
lakes. Inventories of PCBs in sediment cores were calculated as follows:

Inv(ng / cm 2 )  i [Csed (ng / g )  (1   )   ( g / cm 3 )  d (cm)]
Where Inv = total PCB dry mass in a core (ng/cm2),
Csed = sediment concentration,
 = porosity,
ρ = bulk density,
d = thickness of sediment increment.

Focusing-corrected inventories were calculated as sediment PCB inventories divided by
the 210Pb calculated focusing factor. Focusing-corrected PCB inventories represent the historical
sum of all PCB loadings to a given lake per unit area. The advantage of utilizing inventories in
this study is that, unlike raw concentrations or accumulation rates, inventories integrate the
whole core thereby removing the influence of sediment slice thickness and errors in sample
dating. However, even with the focusing corrections, the inventories from the southern to the
northern extent of Michigan in the lakes studies show results similar to those of the 2000s PCB
accumulation rates (Figure 3.11). Peaks are noted in several of the lakes with significant urban
land use in the watershed (e.g. Thompson, Teal, and Muskegon). A few lakes are noted as
having elevated inventories but very low recent accumulation rates (e.g. Teal) indicating that
historical inputs of PCBs to the lake are much greater than current inputs. The reverse is true
comparing Cora where inventories of PCBs are relatively low compared to other lakes but
accumulation rates are relatively high. This is a result of recent trends in Cora for PCB
accumulation rates being elevated but, historically, PCBs do not appear to have impacted the
104

lake significantly.

Figure 3.12: PCB Inventories in all lakes organized by latitude. White Lake peak removed for clarity.
1
: From Schneider et al., 2001; 2: From Pearson et al., 1997.

Figure 3.13: Focusing-Corrected PCB Inventories versus watershed % urban in the 1970s and 2000s. The
top three study lake PCB inventories are labeled illustrating the PCB inventory independence of % urban.

105

Residential neighborhoods have been shown to account for more than 70% and 45% of
PCB spills from transformers and capacitors (respectively) and industrial areas approximately
14% and 29% (Voogt and Brinkman, 1989). Other studies have linked environmental PCB
detections with various anthropogenic activities (Breivik et al., 2002) and atmospheric PCB
studies have clearly shown ties to urban land use (e.g. Hafner and Hites., 2003; Harner et al.,
2004; Jaward et al., 2004). This study appears to reflect these finding as sedimentary PCB
inventories and accumulation rates match of the lake watersheds occupied with high % urban
land use exhibited elevated PCB inventories or accumulation rates. Teal Lake, for example, was
determined as having approximately 27% urban land use in the 1970s and approximately 20% in
the 1990s. This watershed was much more urban and had higher PCBs inventories than any
other in the Upper Peninsula of Michigan. The correlation between urban land use and elevated
PCB inventories and accumulation rates also appears to carry through with Paw Paw, Cass and
Thompson lakes (17.75%, 61.97%, and 22.90% 1970s urban, respectively). However, the
correlation weakens between watershed % urban and elevated PCB values weakens in some
lakes such as White Lake, the most impacted lake studied but with little 1970s watershed %
urban (3.33% 1970s urban). Others lakes with elevated PCB accumulation rates and inventories
but relatively low % urban include Otter, Muskegon, and George (9.26%, 11.40%, and 9.66%
1970s urban respectively). Figure 3.13 is shown to illustrate the low correlation between %
urban land use and PCBs in the lakes studied both historically (1970s) and recently (2000s).
Clearly, point sources of PCBs continue to play a significant role in the accumulation of
PCBs both historically and in the present. For example, White Lake is known to be and
Environmental Protection Agency Area of Concern originally due to organic solvent
contaminated groundwater migrating to the lake from the Occidental Chemical Site (formerly

106

Hooker Chemical Company) along with eight other sites affecting the lake with contamination.
(MDEQ, 2013; USEPA[3], 2013). On the other hand, atmospheric fallout from extra-watershed
sources must play a significant role as well with many lakes containing measurable PCBs in their
sediment with relatively low % urban land use (e.g. Morrison Lake, 6.00% 1970s urban land
use). Further, focus-corrected PCB inventories of each decadal time block were calculated and
presented in Table 3.4. Much like the sediment accumulation rates of PCBs, a temporal stepwise
inventory peak proceeds from the southern portions of Michigan (closer to potential source) to
the much less populous Upper Peninsula.

PCB Congener Analysis:
The term “polychlorinated biphenyl” is actually a generalization for a class of compounds
manufactured in a similar fashion through progressive chlorination of biphenyl. With few
exceptions, PCBs were manufactured and sold as mixtures that may contain as many as 70-100
PCB congeners. A PCB congeners is any single, uniquely defined chemical in the PCB category
(EPA, 2013[2]). The name of a given congener specifies the total number of chlorine atoms that
are attached to the biphenyl rings. For example, dichlorobiphenyl is a congener with two
chlorine atoms. There are 209 individual PCB congeners, most of which contain from one to ten
chlorine atoms. Total concentration of PCBs is simply the sum of the congeners for each
sample. The compositional data of an individual sample is the term used to describe the
concentrations of the individual congeners
PCBs were manufactured under the common name “Aroclor” by the Monsanto
Corporation, the single largest producer of the chemical in the United States. In general, each
PCB is assigned a four digit number detailing the number of carbon atoms in the phenyl rings

107

(first two numbers) followed by the percentage of chlorine by mass in the mixture (last two
numbers). For example, Aroclor 1254 is a PCB mixture containing 12 carbon atoms and is 54%
chlorine by weight. To simplify, the nomenclature used herein will include the PCB congener
chemical name without the position indication of the chlorine atom. For example, 4,4’Dichlorobiphenyl will simply be noted as ‘dichloro’ or simply ‘di’ indicating a biphenyl
structure with two chlorine substituents. The importance of identifying individual PCB
congeners lies in the differences of their physical/chemical properties. Their varying degree of
chlorination effects properties such as solubility, vapor pressure, Henry’s law constant, and the
tendency to sorb to particles can cause significant differences in environmental behavior between
individual congeners (Mackay, 2010; Hites, 1987). PCBs also vary in their susceptibility to
dechlorination which has the potential to complicate PCB trend interpretation and source
identification. Some long-term studies of contaminated sediments have shown specific bacterial
groups to transform particular PCB congeners in a distinctive dechlorination pattern (Brown et
al., 1987; Brown et al., 1984). Often, instead of completely deconstructing the PCB molecule,
bacterial-mediated dechlorination will work to create less chlorinated congeners from PCBs as
chlorines are selectively removed. However, this process depends heavily on site specific
conditions and overall chemistry of the individual PCBs present (Bedard and Quensen, 1995),
preventing easy generalization of degradation trends.
In general, meta-substituted chlorines are dechlorinated more readily followed by parasubstituted chlorines. The presence of chlorines adjacent to one another (i.e., “flanked” metaand para-substituted chlorine) increases the susceptibility to dechlorination. A double-flanked
meta-substituted congener (i.e., with also a chlorine in the adjacent para- and ortho-positions) is
particularly susceptible to dechlorination of the chlorine in the meta position (Klasson and Just,

108

1999; Karcher et al., 2007). The most heavily chlorinated PCB congeners (e.g., hepta-, octa-,
nona- and deca-chlorobiphenyls) tend to be less susceptible to dechlorination than the less
chlorinated congeners (e.g., tri-, tetra-, penta-, and hexa-chlorobiphenyls) (NAVFAC, 2012).
The relative distribution of the individual PCB congener concentrations in a given
sediment sample or suite of samples can be used to develop a PCB “fingerprint“. Applications
of fingerprinting PCB composition to reconstruct the original source fingerprint have been
successful in the Magar, et al. 2005 work examining PCB contaminated sediements at the Lake
Hartwell Superfund Site. However, only the recent sediment resembled that of the known source
and deeper sediment was found to be significantly dechlorinated and no longer resembled the
known source mixtures. Given the differential rates of dechlorination of individual congeners,
the data used for this project could not be used in that capacity, especially given the likelihood of
multiple sources of the compounds into the environment.

Figure 3.14: Average PCB congener fingerprints for worldwide production (as compiled by Breivik et al.,
2002), Michigan study lakes during maximum production, and Michigan study lakes after PCB ban.

Inspection of the spatial and temporal trends of the PCB congeners was undertaken with
to provide clarification in the differences of the decadal peaks in concentration and
accumulation. The average congener fingerprint of all lakes studied in this research is calculated
109

for the period of peak production and usage of PCBs (1960s to 1980s) and during the timeframe
in which the surface water bodies should be recovering (1990s to present). In Figure 3.14, these
calculations are compared to the global average PCB congener fingerprint as compiled by
Breivik et al., 2002. Neither time period matches perfectly with the global production
fingerprint. However, it would be expected that if atmospheric deposition is now driving inputs
in all lakes compared to the consumption/use decades of PCBs, atmospheric mixing and
secondary deposition would cause the recovery time frame to much more closely match that of
worldwide production. The differences must be attributed to continued local inputs of PCBs.
Since PCBs with different congener profiles were produced at different times, at different
locations, and often for different purposes, the continued presence of PCB sources allow
Michigan to maintain a unique fingerprint with regard to PCB pollution fingerprints.

A

B

Figure 3.15: Average PCB congener fingerprints for Michigan inland lakes divided into subregions: UP =
Upper Peninsula, ULP = Upper Lower Peninsula, LLP = Lower Lower Peninsula. Figure 15a shows the
average fingerprint for sediment dated to the time period of maximum consumption/use of PCBs. Figure 15b
shows the fingerprint for sediment dated to the years since the PCB ban.

When the data specific to Michigan inland lakes is compiled and averaged by sub-region,
it becomes apparent that PCBs dispersed to surface water bodies during the maximum
consumption use period varied substantially by location (Figure 3.15a). Upper Peninsula lakes
exhibited elevated concentrations of hexachloro PCBs while lakes in the Upper Lower Peninsula
110

showed an abundance of decachloro PCBs. Increasing similarities are seen during the recovery
time period. This time period was calculated by averaging the fingerprints of PCBs from
sediment dated from the 1990s to the present. In order to minimize signal mixture, the year of
the PCB ban (1979) was not used to account for potential lag times of PCB inputs into the
watershed to be deposited to the lake sediment. This time period shows increased similarities in
the average congener signal amongst the PCBs present. This gives some indication that
individual sources have become less important than the historical maximum production time
period as a more ubiquitous signal is being generated. Comparing Figure 3.15b to the worldwide
production calculations in Figure 3.14 still shows significant differences from the worldwide
average production signature. Thus, the signal generated by the recovery time period can be
considered a regional signal for Michigan and the immediate area only, not a global pattern.

Cluster analysis
To examine differences and similarities in the relative abundances of PCB congeners
measured at various locations; cluster analysis was performed on the PCB congener data using
JMP 5.0 (JMP, 2005). Cluster analysis refers to a general category of algorithms designed to
classify samples into groups of similar characteristics, i.e., the relative abundances of the PCBs.
The most common type of cluster analysis is the hierarchical cluster analysis (HCA) which is an
exploratory method that, used here, includes calculations describing the similarity amongst the
individual congener results of the sediment samples. The most common similarity metrics are
based on distances calculated in multivariate space. When the distances are small, it is implied
that there is greater similarity between all the present congeners representing two individual
samples. Dissimilar samples will be separated by larger distances.

111

For this study, twenty PCB congeners were measured and analyzed. Examination of the
literature has shown there to be a large number of multivariate statistical methods applied to
environmental data (Howell, 2007; Skrbic and Durisic-Mladenovic, 2007; Onuska and Davies,
1991). It is suggested that use of log-ratio transformation is useful for describing compositional
data (such as those data used here) as it avoids many problems associated with the analysis and
interpretation of raw compositional data (Howell, 2007). All congener concentration values
were expressed as the proportion of total PCB mass (by weight), where PCB mass total is
defined as the sum of congeners. This is termed percentage standardization. These data are often
described as “compositional data” as they have the feature that the proportions of pollutants in a
sample sum to 1. Compositional data have been shown to have problems with correlation
coefficients and standard multivariate techniques which lead to problems with the interpretation
of simple summary statistics (Pearson, 1897). Therefore, log-ratio transformation of the
compositional data was also used as proposed by Aitchison (1986). The log-ratios are calculated
by dividing the compositional data of each PCB congener in each lake by the geometric mean of
the twenty PCB congeners of that lake.

Centered Log-ratio = log (pij / g(pj))
Where pij = the proportion of PCB congener i in Lake j,
And g(pj) = (p1jp2j . . . ,pdj)1/20
The congener compositional data and log-ratio transformed data of two particular time
horizons were calculated and used for the cluster analysis. High PCB emission years from 1950s
to the 1980s were both averaged and summed to represent the years with significant quantities of
PCBs cycling in the environment. The 1990s and the 2000s were then averaged and summed to
represent the years of environmental recovery. Six, eight, ten and twelve clusters were studied
for both data sets and similar patterns were found. The ten cluster analysis shows the clearest
112

patterns and only these results are presented. Likewise, as the summed and averaged data provide
near-identical clustering results, only the summed (or inventory) data from the respective time
horizons are presented.
The analysis itself indicates which lakes are similar in the relative abundance or
fingerprint of the congeners. Examining the relationships amongst the clusters assist identify
common sources. This approach has been successfully employed by Weiss et al., 1994 to cluster
PCB signatures amongst more polluted sites, adjacent sites and background sites. A hierarchical
cluster approach was also used in PCB congener analysis of sediment cores collected in Lake
Hartwell Superfund site (Pickens County, SC) in the characterization of PCB sources and
dechlorination patterns in lieu of direct comparisons of chromatograms (Magar, 2005).
The results of the cluster analysis are shown in Figure 3.16 and 3.17 where similar colors
reflect similar fingerprints. In the 1950s to 1980s 210Pb dates sediment, the congener cluster
pattern of compositional data shows several major clusters and a number of smaller clusters of
two or three lakes (Figure 3.16). The cluster analysis revealed an increased quantity of clusters
of smaller groups of lakes appears to be consistent with a combination of atmosphericallyderived and local point source pollution. Most of the lakes occupying the smaller clusters are
those within significant urban land use watersheds (e.g. Teal Lake in the UP, Cass Lake in the
Detroit area). However, interspersed throughout lakes shown in Figure 3.16 are lakes with
watersheds that are much less populous all sharing similar congener patterns (Figure 13A inset).
These lakes share two prominent clusters in the dendrogram and are interpreted as receiving
atmospherically-derived PCB inputs. As the atmosphere mixes the signals from extra-watershed
sources, inputs of PCBs into these more remote lakes become much more uniform than those
lying close to the source. Both the UP and the LP have a set of lakes sharing clusters in Figure

113

3.16.

Figure 3.16: Cluster analysis of compositional data: Landscape Peak Inventories (1950s to 1980s)

114

Figure 3.17: Cluster analysis of compositional data: Landscape Recovery Inventories (1990s to present).

Since the lakes sharing these clusters are spread across a fairly vast geographic area, it
could be interpreted that sources for these PCB inputs lie outside the State of Michigan. The
distance required for atmospheric mixing and deposition to occur across such an area points to
regional sources lying across Lake Michigan (e.g., Green Bay, Milwaukee, and Chicago). As
most of Michigan’s UP lakes cluster differently, their regional sources are likely separate (e.g.
Duluth, Minneapolis) from those in the LP but likely not farther than the American Midwest.
Otherwise, the atmosphere would have sufficient time/distance to provide a ubiquitous signal for
lakes in both the UP and the LP.

115

The cluster analysis performed on the Landscape Recovery Inventories shows a much
more uniform pattern of PCB congeners across the entire State. One clear cluster pattern appears
to dominate (yellow ‘plus’ signs) followed by a secondary pattern situated primarily in the
central portion of the State (red squares). As this change from the Landscape Peak Inventory
analysis is to a smaller number of cluster groups over a wider geographical area is consistent
with the hypothesis. However, the reasons for the particular clustering patterns remain unknown.
For example, it is not immediately clear why the suite of lakes represented by the ‘red squares’
clustered together throughout the Lower Peninsula, other than the fact that they all represent
primarily agricultural watershed land use. As to why these lakes do not correlate with the
agriculture dominated southern portion of the State is not clear. It appears that many of the small
clusters or clusters represented by an individual lake lie within watersheds of elevated population
or urban land use (e.g. Thompson, Muskegon), likely representing inputs at the sub-regional or
watershed scale. Others are lakes known to have significant loadings of PCBs to the lake itself
by known watershed industrial practices (White). These lakes are interpreted to be still within a
recovery phase as the watershed and lake appear to continue cycling through previous PCBs
loadings.
Clustering all both the historical and recent PCB inventory data was performed to provide
insight into the significance of previously deposited PCBs in modern sediment. Hypothetically,
if lakes from the peak depositional years correlate with the same lakes as those in the present, it
could be interpreted that the source of the PCBs has not changed since this period. Since PCBs
dechlorination occurs at different rates, clustered lakes cannot be the result of previous PCB
loadings continuing to cycle in the lakes sediment and water column. Results show lakes
correlating in the same clade for both time slices (Peak Inventory and Recovery Inventory)

116

include Teal, Charlevoix, Muskegon, and Avalon lakes. White lake from both time slices
occupied the same clade but at a high geometric x distance, probably the result of cleanup efforts
of the lakes sediment. Many of the lakes from the more recent time slice correlate to one
another, likely an artifact of similarities because of atmospheric mixing as observed in Figure
3.18.

117

Figure 3.18: Cluster analysis of all compositional data. Original lake names refer to the lake data from the
historical peak loadings time slice (1950s to 1980s) and lake names followed by a ‘2’ refer to the lake data
from the recent time slice (1990s to present).

118

Conclusion
Sediment records from 44 inland lakes widely distributed in the State of Michigan were
studied to better understand PCB spatial and temporal patterns. The hypothesis driving this
study is that the high PCB production years (e.g., 1960s) and the influence of population centers
on loadings will cause a localized pattern in PCB loadings compared to recent times (e.g.,
2000s), where PCB loadings should be more reflective of a regional signal. The regional signal
is expected to dominate more remote lakes and these lakes are anticipated to be largely
composed of the lighter, more mobile, fraction of the PCB congeners. To test the hypothesis,
total PCB inventories, temporal concentrations and accumulation rates of all the lakes were
studied along with the congener makeup of these PCB values. These data were subject to both
temporal and spatial evaluation across the State of Michigan in lake sediment dated from the
1930’s to present.
It is clear from the data that the State is in a general recovery trajectory in terms of both
PCB concentration and accumulation rates in recent lake sediment relative to sediment deposited
during the high PCB production/consumption years. PCB values are generally a fraction of these
values in the recent compared to what they had been in previous decades with few exceptions,
notably lakes in the Upper Peninsula. Temporally, concentration and accumulation rate values
appear to reflect consumption data only close to the populous portion of the State, and peaks
typically occur later with increased latitude. The exceptions to this observation are likely caused
by inputs overwhelmed by local point sources (e.g. White Lake).
The concentration profiles of most lakes exhibit PCB production and usage history, onset
after 1940, peak during 1960-1980, and decrease after 1980. During high PCB production years,
the total concentrations and accumulation rates among the lakes were significantly different.

119

Lakes near urban areas have higher PCB concentrations and accumulation rates than lakes in
remote areas. In recent years, those differences appear to be smaller. PCB inventories show a
regional pattern and a localized pattern, with higher inventories in the Lower Peninsula than in
the Upper Peninsula and with higher inventories in urban areas than in rural areas.
Congener cluster analysis was performed on select time horizons of the data set. These
time horizons included 210Pb-dated sediment from years that include elevated PCB production
and consumption in the United States (sediment dated from 1950s to 1980s) and sediment from
years of landscape recovery from the previous period of PCB loading (sediment dated 1990s to
2000s). Results from the congener analysis revealed the period of elevated PCB loadings to
cluster into primarily local or sub-regional groups overprinted by a regional signal. As the
regional signal was different in the Upper Peninsula relative to the remainder of Michigan, it is
possible that this cluster is sourced from industrialized west coast of Lake Michigan. Much of
the recent time horizon of sediment clustered into one large, state-wide group with several
smaller groups and individual lakes clustering separately. It is not clear what the most important
characteristics are of this secondary group. Lakes not clustered with these two groups have a
significant % urban component to their watershed (e.g. Thompson) or are formerly highly
impacted lakes (e.g. White Lake). This is likely indicative of continued recovery of the
watersheds from historical PCB loadings.
Clustering all the data together provides some insight into lakes that continue to carry
similar congener signatures in both recent and historical sediment. As inland lake sediment
subject to both local and regional inputs of PCBs will likely result in the presence of several
Aroclor mixtures that are dechlorinated to some degree, it can be inferred that the same lakes
could not cluster together unless they continue to receive PCBs from the same source currently

120

as in the past. This appears to be true for Avalon, Charlevoix, Muskegon, Teal and, to a lesser
degree, White Lake.
Overall, the results indicate that during the time of high PCB production, the influence of
population centers on loadings caused a spatial gradient in PCB distribution from the more
populous areas of the State to the more remote areas. Localized, urban sources have been
relatively more dominant than long-range atmospheric transport during this period but an
atmospheric component to the PCB loadings is almost certainly present. In recent years; PCB
loadings appear to be dominated by a regional signal with long-range atmospheric transport
serving as the primary source of PCB loading to many of the lakes in the recent. The data also
show both PCB concentrations and accumulation rates in to peak later with increasing latitude.
Comparing this to atmospherically derived stable Pb profiles that have a strong tendency to peak
during the early to mid-1970s points to the possibility of secondary mobilization and deposition
of much of the observed PCBs in the northern extents of the state. Difficulties arise when
attempting to correlate specific watershed attributes (e.g. land use, population) to the PCB data.
This is probably the result of PCB use being more closely tied to particular industries that utilize
the chemical than they are a particular land use. However, utilization of GIS data layers
characterizing land use and population appear to have a weak correlation to PCBs because of the
ubiquitous use of the chemicals in both residential and industrial practices. The correlation
between watershed factors and the PCB data of all the lakes studied remains weak probably due
to continued independent local inputs to some lakes overprinted by a regional signal and further
complexed by secondary mobilization of the chemicals.

121

REFERENCES

122

REFERENCES

Agency for Toxic Substances and Disease Registry (ATSDR), 2000. Toxicological profile for
Polychlorinated Biphenyls (PCBs). Atlanta, GA: U.S. Department of Health and Human
Services, Public Health Service.
Aitchison, J., 1986. The statistical analysis of compositional data. Chapman & Hall, Ltd.
Albert, D.A., Denton S.A., and Barnes, B.V., 1986. Regional landscape ecosystems of Michigan.
School of Natural Resources, University of Michigan. Ann Arbor, Michigan, USA.
Albert, D.A., 1995. Regional landscape ecosystems of Michigan, Minnesota, and Wisconsin: a
working classification (Fourth Revision: July 1994). North Central Forest Exp. Station.
Forest Service—U.S. Dept. of Ag. General Technical Report NC-178. Northern Prairie
Wildlife Research Center
(http://www.npwrc.usgs.gov/resource/1998/rlandscp/rlandscp.htm) Retrieved 04/12/2013
Alcock, R.E., Sweetman, A.J., Juan, C-Y, 1999. The intake and clearance of PCBs in humans – a
generic model of lifetime exposure. Organohalogen Compounds 44:61-65.
Anderson, D. J., Bloem, T. B., Blankenbaker, R. K., Stanko, T. A. 1999. Concentrations of
polychlorinated biphenyls in the water column of the Laurentian Great Lakes: Spring
1993. Journal of Great Lakes Research, 25(1), 160-170.
Bedard, D.L. and J.F. Quensen. 1995. Microbial Reductive Dechlorination of Polychlorinated
Biphenyls, In: Microbial Transformation and Degradation of Toxic Organic Chemicals;
Young, L.Y. and Cerniglia, C.E., Eds; Wiley-Liss, New York. 127-216.
Bergman, A., Hagman, A., Jacobsson, S., Jansson, B., & Ahlman, M. 1984. Thermal degradation
of polychlorinated alkanes. Chemosphere, 13(2), 237-250.
Blais, J.M., Muir, D. C., 2002. Paleolimnological methods and applications for persistent organic
pollutants. In Tracking environmental change using lake sediments (pp. 271-298).
Springer Netherlands.
Blais, J.M., Macdonald, R.W., Mackay, D., Webster, E., Harvey, C., Smol, J.P., 2007.
Biologically Mediated Transport of Contaminants to Aquatic Systems. Environmental
Science & Technology 41 (4), 1075-1084
Bopp, R. F., Simpson, H. J., Olsen, C. R., Trier, R. M., & Kostyk, N. 1982. Chlorinated
hydrocarbons and radionuclide chronologies in sediments of the Hudson River and
Estuary, New York. Environmental Science & Technology,16(10), 666-676.
123

Breivik, K., Sweetman, A., Pacyna, J. M., & Jones, K. C. 2002. Towards a global historical
emission inventory for selected PCB congeners—a mass balance approach: 1. Global
production and consumption. Science of the Total Environment, 290(1), 181-198.
Brown, J.F., Wagner, R.E., Bedard, D.L., Brennan, M.J., Carnahan, J.C., May, R.J., and
Tofflemire, T.J., 1984. PCB transformations in upper Hudson sediments. Northeastern
Environ. Sci. 3: 167–179.
Brown, J.F., Bedard, D.L., Brennan, M.J., Carnahan, J. C., Feng, H., and Wagner, R.E.,
1987a. Polychlorinated biphenyl dechlorination in aquatic sediments. Science. 236: 709–
712.
Buehler, S. S., & Hites, R. A. 2002. Peer Reviewed: The Great Lakes' Integrated Atmospheric
Deposition Network. Environmental science & technology, 36(17), 354A-359A.
Callahan, M. A., Hammerstrom, K. A., & Schweer, G., 1983. Present PCB uses and their
potential for release to the environment. In Proceedings of the PCB-Seminar, The Hague,
The Netherlands (pp. 28-30).
Carpenter, D. O. (2006). Polychlorinated biphenyls (PCBs): routes of exposure and effects on
human health. Reviews on environmental health, 21(1), 1-24.
Cohen, A. S., 2003. Paleolimnology: the history and evolution of lake systems(p. 500). Oxford:
Oxford University Press.
Cordle, F., Corneliussen, P., Jelinek, C., Hackley, B., Lehman, R., McLaughlin, J., ... & Shapiro,
R., 1978. Final Report of the Subcommittee on Health Effects of Pcbs and Pbbs: Human
exposure to polychorinated biphenyls and polybrominated biphenyls. Environmental
health perspectives, 24, 157.
Downing, J. A., & Rath, L. C., 1988. Spatial patchiness in the lacustrine sedimentary
environment. Limnol. Oceanogr, 33(3), 447458.
Eisenreich, S. J., Hollod, G. J., & Johnson, T. C., 1979. Accumulation of polychlorinated
biphenyls (PCBs) in surficial Lake Superior sediments. Atmospheric
deposition. Environmental Science & Technology, 13(5), 569-573.
Eisenreich, S. J., Looney, B. B., & Thornton, J. D., 1981. Airborne organic contaminants in the
Great Lakes ecosystem. Environmental Science & Technology, 15(1), 30-38.
Eisenreich, S. J., Capel, P. D., Robbins, J. A., & Bourbonniere, R., 1989. Accumulation and
diagenesis of chlorinated hydrocarbons in lacustrine sediments. Environmental Science &
Technology, 23(9), 1116-1126.
ESRI, 2011. ArcGIS Desktop: Release 10. Redlands, CA: Environmental Systems Research
Institute
124

Franz, T. P., Eisenreich, S. J. & Holsen, T. M. (1998). Dry Deposition of Particulate
Polychlorinated Biphenyls and Polycyclic Aromatic Hydrocarbons to Lake Michigan.
Environmental Science & Technology 32, 3681–3688
Fuller, L.M., and Minnerick, R.J., 2008, State and regional water-quality characteristics and
trophic conditions of Michigan’s inland lakes, 2001–2005: U.S. Geological Survey
Scientific Investigations Report 2008–5188, 58 p. Date Posted: December 17, 2008:
[http://pubs.water.usgs.gov/sir20085188/]
Golden, K. A., Wong, C. S., Jeremiason, J. D., Eisenreich, S. J., Sanders, G., Hallgren, J., ... &
Long, D. T., 1993. Accumulation and preliminary inventory of organochlorines in Great
Lakes sediments. Water Science & Technology, 28(8-9), 19-31.
Hafner, W. D., & Hites, R. A., 2003. Potential sources of pesticides, PCBs, and PAHs to the
atmosphere of the Great Lakes. Environmental science & technology, 37(17), 3764-3773.
Håkanson, L., 1977. The influence of wind, fetch, and water depth on the distribution of
sediments in Lake Vänern, Sweden. Canadian Journal of Earth Sciences, 14(3), 397-412.
Halsall, C. J., Gevao, B., Howsam, M., Lee, R. G. M., Ockenden, W. A., & Jones, K. C., 1999.
Temperature dependence of PCBs in the UK atmosphere. Atmospheric
Environment, 33(4), 541-552.
Harner, T., Shoeib, M., Diamond, M., Stern, G., & Rosenberg, B., 2004. Using passive air
samplers to assess urban-rural trends for persistent organic pollutants. 1. Polychlorinated
biphenyls and organochlorine pesticides. Environmental science & technology, 38(17),
4474-4483.
Helm, P. A., Milne, J., Hiriart-Baer, V., Crozier, P., Kolic, T., Lega, R., ... & Reiner, E. J., 2011.
Lake-wide distribution and depositional history of current-and past-use persistent organic
pollutants in Lake Simcoe, Ontario, Canada. Journal of Great Lakes Research, 37, 132141.
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 & technology, 34(17), 3588-3597.
Hillery, B. R., Basu, I., Sweet, C. W., & Hites, R. A., 1997. Temporal and spatial trends in a
long-term study of gas-phase PCB concentrations near the Great Lakes. Environmental
science & technology, 31(6), 1811-1816.
Hites, R. A., & Eisenreich, S. J., 1987. In Sources and fates of aquatic pollutants (No. CONF850942-). American Chemical Society, Washington, DC.; pp 394-469

125

Hoff, R. M., Strachan, W. M. J., Sweet, C. W., Chan, C. H., Shackleton, M., Bidleman, T. F., ...
& Schroeder, W. H., 1996. Atmospheric deposition of toxic chemicals to the Great Lakes:
A review of data through 1994. Atmospheric Environment, 30(20), 3505-3527.
Howel, D., 2007. Multivariate data analysis of pollutant profiles: PCB levels across
Europe. Chemosphere, 67(7), 1300-1307.
Hu, D., Martinez, A., & Hornbuckle, K. C., 2011. Sedimentary records of non-Aroclor and
Aroclor PCB mixtures in the Great Lakes. Journal of Great Lakes Research, 37(2), 359.
Jaward, F. M., Meijer, S. N., Steinnes, E., Thomas, G. O., & Jones, K. C., 2004. Further studies
on the latitudinal and temporal trends of persistent organic pollutants in Norwegian and
UK background air. Environmental science & technology, 38(9), 2523-2530.
Jensen, S., 1966. "Report of a new chemical hazard". New Sci. 32: 612.
Jeremiason, J. D., Hornbuckle, K. C., & Eisenreich, S. J., 1994. PCBs in Lake Superior, 19781992: decreases in water concentrations reflect loss by volatilization. Environmental
science & technology, 28(5), 903-914.
Jeremiason, J. D., Eisenreich, S. J., Baker, J. E., & Eadie, B. J., 1998. PCB decline in settling
particles and benthic recycling of PCBs and PAHs in Lake Superior. Environmental
science & technology, 32(21), 3249-3256.
JMP, Version 5. SAS Institute Inc., Cary, NC, 1989-2005.
Khim, J. S., Kannan, K., Villeneuve, D. L., Koh, C. H., & Giesy, J. P., 1999. Characterization
and distribution of trace organic contaminants in sediment from Masan Bay, Korea. 1.
Instrumental analysis. Environmental science & technology, 33(23), 4199-4205.
Klasson, K.T. and E.M. Just. 1999. “Computer model for prediction of PCB dechlorination and
biodegradation endpoints,” 5th International Symposium on In Situ and On-Site
Bioremediation. 19–22 April, 1999 San Diego, CA.
Karcher, S.C., J.M. VanBriesen, and M.J. Small. 2007. “Numerical method to elucidate likely
target positions of chlorine removal in anaerobic sediments undergoing polychlorinated
biphenyl dechlorination,” J. Environ. Engineer. 133, vol. 3, 278–286.
Lead, W. A., Steinnes, E., Bacon, J. R., & Jones, K. C., 1997. Polychlorinated biphenyls in UK
and Norwegian soils: spatial and temporal trends. Science of the Total
Environment, 193(3), 229-236.
Longnecker, M. P., Rogan, W. J., & Lucier, G., 1997. THE HUMAN HEALTH EFFECTS OF
DDT (DICHLORODIPHENYLTRICHLOROETHANE) AND PCBS
(POLYCHLORINATED BIPHENYLS) AND AN OVERVIEW OF

126

ORGANOCHLORINES IN PUBLIC HEALTH 1. Annual review of public health,18(1),
211-244
Macgregor, K., Oliver, I. W., Harris, L., & Ridgway, I. M., 2010. Persistent organic pollutants
(PCB, DDT, HCH, HCB & BDE) in eels (Anguilla Anguilla) in Scotland: Current levels
and temporal trends. Environmental Pollution, 158(7), 2402-2411.
Magar, V. S., Johnson, G. W., Brenner, R. C., Quensen, J. F., Foote, E. A., Durell, G., ... &
Peven-McCarthy, C., 2005. Long-term recovery of PCB-contaminated sediments at the
Lake Hartwell superfund site: PCB dechlorination. 1. End-member
characterization. Environmental science & technology, 39(10), 3538-3547.
Mackay, D., Shiu, W.Y., Ma, K.C., & Lee, S.C., 2010. Handbook of physical-chemical
properties and environmental fate for organic chemicals. CRC press.
Michigan Department of Environmental Quality (MDEQ), 2013. White Lake Area of Concern.
Retrieved from http://www.michigan.gov/deq/0,1607,7-135-3313_3677_15430_57456--,00.html
Michigan Department of Natural Resources (MDNR), 2013. Lake Maps by County. Retrieved
from http://www.michigan.gov/dnr/0,1607,7-153-30301_31431_32340---,00.html
Michigan Department of Natural Resources (MDNR), 2003. IFMAP/GAP land cover. In:
MDNR (Ed.), Lansing, Michigan
Mozaffarian, D., & Rimm, E. B., 2006. Fish intake, contaminants, and human health. JAMA: the
journal of the American Medical Association, 296(15), 1885-1899.
Muir, D. C., Omelchenko, A., Grift, N. P., Savoie, D. A., Lockhart, W. L., Wilkinson, P., &
Brunskill, G. J., 1996. Spatial trends and historical deposition of polychlorinated
biphenyls in Canadian midlatitude and Arctic lake sediments. Environmental science &
technology, 30(12), 3609-3617.
Muir, D., Riget, F., Cleemann, M., Skaare, J., Kleivane, L., Nakata, H., ... & Tanabe, S., 2000.
Circumpolar trends of PCBs and organochlorine pesticides in the Arctic marine
environment inferred from levels in ringed seals. Environmental science &
technology, 34(12), 2431-2438.
Murphy, T. J., & Rzeszutko, C. P., 1977. Precipitation inputs of PCBs to Lake
Michigan. Journal of Great Lakes Research, 3(3), 305-312.
Naval Facilities Engineering Command (NACFAC), 2012. A Handbook for Determining the
Sources of PCB Contamination in Sediments. Technical Report TR-NAVFAC EXWCEV-1302. Retrieved from:
http://cluin.org/download/contaminantfocus/pcb/pcb_sediment_handbook.pdf

127

Nisbet, I. C., & Sarofim, A. F., 1972. Rates and routes of transport of PCBs in the
environment. Environmental Health Perspectives, 1, 21.
Ockenden, W. A., Prest, H. F., Thomas, G. O., Sweetman, A. & Jones, K. C., 1998. Passive air
sampling for PCBs : field calculation of atmospheric sampling rates by Triolein containing
semi-permeable membrane devices. Environmental Science & Technology 32, 1538–1543.
Offenberg, J. H., & Baker, J. E., 1997. Polychlorinated biphenyls in Chicago precipitation:
enhanced wet deposition to near-shore Lake Michigan. Environmental science &
technology, 31(5), 1534-1538.
Offenberg, J., Simcik, M., Baker, J., & Eisenreich, S. J., 2005. The impact of urban areas on the
deposition of air toxics to adjacent surface waters: A mass budget of PCBs in Lake
Michigan in 1994. Aquatic sciences, 67(1), 79-85.
Oliver B.G., Niimi A.J., 1988. Trophodynamic analysis of polychlorinated biphenyl congeners
and other chlorinated hydrocarbons in the Lake Ontario ecosystem. Environmental
Science & Technology 22, 388-397.
Onuska, F. I., & Davies, S. 1991. Multivariate observations of the distribution of polychlorinated
biphenyls in environmental compartments of two harbours. International journal of
environmental analytical chemistry, 43(2-3), 137-150.
Totten, L. A., Gigliotti, C. L., VanRy, D. A., Offenberg, J. H., Nelson, E. D., Dachs, J., ... &
Eisenreich, S. J., 2004. Atmospheric concentrations and deposition of polychorinated
biphenyls to the Hudson River Estuary.Environmental science & technology, 38(9), 25682573.
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 & technology, 41(16),
5634-5640.
Pearson, K., 1896. Mathematical Contributions to the Theory of Evolution.--On a Form of
Spurious Correlation Which May Arise When Indices Are Used in the Measurement of
Organs. Proceedings of the Royal Society of London, 60(359-367), 489-498.
Pearson, R. F., Swackhamer, D. L., Eisenreich, S. J., & Long, D. T. 1997. Concentrations,
accumulations, and inventories of toxaphene in sediments of the Great Lakes.
Environmental science & technology, 31(12), 3523-3529.
Pohanish, R.P., 2008. Sittig's handbook of toxic and hazardous chemicals and carcinogens.
Access Online via Elsevier.
Rachdawong, P., & Christensen, E. R., 1997. Determination of PCB sources by a principal
component method with nonnegative constraints. Environmental science &
technology, 31(9), 2686-2691.
128

Rapaport, R. A., & Eisenreich, S. J., 1988. Historical atmospheric inputs of high-molecularweight chlorinated hydrocarbons to eastern North America. Environmental science &
technology, 22(8), 931-941.
Robbins, J.A., Edgington, D.N., & Kemp, A.L.W., 1978. Comparative 210Pb, 137Cs, and pollen
geochronologies of sediments from Lakes Ontario and Erie. Quaternary Research, 10(2),
256-278.
Sanchez-Cabeza, J. A., Ani-Ragolta, I., & Masque, P., 2000. Some considerations of the 210Pb
constant rate of supply (CRS) dating model.Limnology and oceanography, 45(4), 990995.
Schneider, A. R., Stapleton, H. M., Cornwell, J., & Baker, J. E. 2001. Recent declines in PAH,
PCB, and toxaphene levels in the northern Great Lakes as determined from high resolution
sediment cores. Environmental science & technology, 35(19), 3809-3815.
Schuster, J. K., Gioia, R., Breivik, K., Steinnes, E., Scheringer, M., & Jones, K.C., 2010. Trends
in European background air reflect reductions in primary emissions of PCBs and
PBDEs. Environmental science & technology, 44(17), 6760-6766.
Schuster, J. K., Gioia, R., Moeckel, C., Agarwal, T., Bucheli, T. D., Breivik, K., ... & Jones, K.
C. (2011). Has the burden and distribution of PCBs and PBDEs changed in European
background soils between 1998 and 2008? Implications for sources and
processes. Environmental science & technology, 45(17), 7291-7297.
Škrbić, B., & Đurišić-Mladenović, N. (2007). Principal component analysis for soil
contamination with organochlorine compounds. Chemosphere, 68(11), 2144-2152.
Simcik, M. F., Zhang, H., Eisenreich, S. J., & Franz, T. P., 1997. Urban contamination of the
Chicago/coastal Lake Michigan atmosphere by PCBs and PAHs during
AEOLOS. Environmental science & technology, 31(7), 2141-2147.
Sinkkonen, S., & Paasivirta, J., 2000. Degradation half-life times of PCDDs, PCDFs and PCBs
for environmental fate modeling. Chemosphere, 40(9), 943-949.
Simcik, M. F., Basu, I., Sweet, C. W., & Hites, R. A., 1999. Temperature dependence and
temporal trends of polychlorinated biphenyl congeners in the Great Lakes atmosphere.
Environmental science & technology, 33(12), 1991-1995.
Simcik, M. F., Hoff, R. M., Strachan, W. M., Sweet, C. W., Basu, I., & Hites, R. A., 2000.
Temporal trends of semivolatile organic contaminants in Great Lakes
precipitation. Environmental science & technology, 34(3), 361-367.
Shen, L., Wania, F., Lei, Y.D., Teixeira, C., Muir, D.C., & Xiao, H., 2006. Polychlorinated
biphenyls and polybrominated diphenyl ethers in the North American
atmosphere. Environmental pollution, 144(2), 434-444.
129

Smith, A. G., Francis, J. E. & Carthew, P., 1990. Iron as a synergist for hepatocellular carcinoma
induced by polychlorinated biphenyls in Ah-responsive C57BL/10ScSn mice.
Carcinogenesis 11, 437–444.
Stern, G. A., Halsall, C. J., Barrie, L. A., Muir, D. C. G., Fellin, P., Rosenberg, B., ... &
Pastuhov, B., 1997. Polychlorinated biphenyls in Arctic air. 1. Temporal and spatial
trends: 1992-1994. Environmental science & technology,31(12), 3619-3628.
Strachan, W.M.J., & Huneault, H., 1979. Polychlorinated biphenyls and organochlorine
pesticides in Great Lakes precipitation. Journal of Great Lakes Research, 5(1), 61-68.
Strachan, W.M., & Eisenreich, S.J., 1990. Mass balance accounting of chemicals in the Great
Lakes. Long Range Transport of Pesticides. Chelsea, Michigan, USA: Lewis Publishers,
291-301.
Strandberg, B., Dodder, N. G., Basu, I., & Hites, R. A., 2001. Concentrations and spatial
variations of polybrominated diphenyl ethers and other organohalogen compounds in
Great Lakes air. Environmental Science & Technology, 35(6), 1078-1083.
Sun, P., Basu, I., Blanchard, P., Brice, K. A., & Hites, R. A., 2007. Temporal and spatial trends
of atmospheric polychlorinated biphenyl concentrations near the Great
Lakes. Environmental science & technology, 41(4), 1131-1136.
Swackhamer, D. L., McVeety, B. D., & Hites, R. A. (1988). Deposition and evaporation of
polychlorobiphenyl congeners to and from Siskiwit Lake, Isle Royale, Lake
Superior. Environmental science & technology, 22(6), 664-672.
Sweet, C. W., Murphy, T. J., Bannasch, J. H., Kelsey, C. A., & Hong, J., 1993. Atmospheric
deposition of PCBs into green bay. Journal of Great Lakes Research, 19(1), 109-128.
Tanabe, S., 1988. PCB problems in the future: foresight from current knowledge. Environmental
Pollution, 50(1), 5-28.
United Nations Environment Programme . "Proceedings of the Subregional Awareness Raising
Workshop on Persistent Organic Pollutants (POPs), Bangkok, Thailand". November 2528th, 1997.
(http://www.chem.unep.ch/pops/POPs_Inc/proceedings/bangkok/FIEDLER1.html).
United States Environmental Protection Agency (USEPA). 1980. Ambient water quality criteria
for polychlorinated biphenyls. Washington, DC: U.S. Environmental Protection Agency,
Criteria and Standards Division, Office of Water Regulations and Standards. EPA 440/580-068. Retrieved 06/07/2013.
United States Environmental Protection Agency (USEPA). 1996. "PCBs: Cancer Dose-Response
Assessment and Application to Environmental Mixtures."
http://www.epa.gov/pcb/pubs/pcb.pdf. Retrieved 06/07/2013.
130

United States Environmental Protection Agency (USEPA). 2003.
http://www.epa.gov/osw/hazard/tsd/pcbs/pubs/congenertable.pdf. Retrieved 09/24/2013.
United States Environmental Protection Agency (USEPA). (2013[1]).
http://www.epa.gov/wastes/hazard/tsd/pcbs/about.htm. Retrieved 09/24/2013.
United States Environmental Protection Agency(USEPA). (2013[2]):
http://www.epa.gov/epawaste/hazard/tsd/pcbs/pubs/effects.htm Retrieved 06/07/2013.
United States Environmental Protection Agency (USEPA). (2013[3]):
http://www.epa.gov/ogwdw/pdfs/factsheets/soc/pcbs.pdf. Retrieved 06/07/2013.
United States Environmental Protection Agency (USEPA). (2013[4]):
http://www.epa.gov/greatlakes/aoc/whitelake/. Retrieved 07/23/2013.
United States Geological Survey (USGS), The National Map Seamless Server, Available online
at: http://seamless.usgs.gov/index.php. (Accessed: September 25, 2011).
Vannier, R.G., Robinson, A., Long, D.T., and Giesy, J.P., 2013. Temporal Changes to Trace
Metal Loadings to Michigan Inland Lakes. Abstract presented at the 2013 Geological
Society of America North-Central Section, Kalamazoo, Michigan.
Van Metre, P. C., Wilson, J. T., Callender, E., & Fuller, C. C., 1998. Similar rates of decrease of
persistent, hydrophobic and particle-reactive contaminants in riverine
systems. Environmental science & technology, 32(21), 3312-3317.
von Waldow, H., MacLeod, M., Scheringer, M., & Hungerb hler, K., 2010. Quantifying
remoteness from emission sources of persistent organic pollutants on a global
scale. Environmental science & technology, 44(8), 2791-2796.
Voogt, P. D. E., & Brinkman, U. T., 1989. Production, properties and usage of polychlorinated
biphenyls. Halogenated biphenyls, terphenyls, naphthalenes, dibenzodioxins and related
products. Amsterdam: Elsevier, 3-45.
Wania, F., & Mackay, D., 1993. Global fractionation and cold condensation of low volatility
organochlorine compounds in polar regions. Ambio, 10-18.
Wania, F., & Mackay, D., 1995. A global distribution model for persistent organic
chemicals. Science of the Total Environment, 160, 211-232.
Wania, F., & Mackay, D., 1996. Peer reviewed: tracking the distribution of persistent organic
pollutants. Environmental Science & Technology, 30(9), 390A-396A.

131

Wania, F., & Dugani, C. B., 2003. Assessing the long‐range transport potential of
polybrominated diphenyl ethers: A comparison of four multimedia models.Environmental
Toxicology and Chemistry, 22(6), 1252-1261.
Wania, F., & Su, Y., 2004. Quantifying the global fractionation of polychlorinated
biphenyls. AMBIO: A Journal of the Human Environment, 33(3), 161-168.
Weiss, P., Riss, A., Gschmeidler, E., & Schentz, H. 1994. Investigation of heavy metal, PAH,
PCB patterns and PCDD/F profiles of soil samples from an industrialized urban area
(Linz, Upper Austria) with multivariate statistical methods. Chemosphere, 29(9), 22232236.
Wong, C. S., Sanders, G., Engstrom, D. R., Long, D. T., Swackhamer, D. L., & Eisenreich, S. J.
1995. Accumulation, inventory, and diagenesis of chlorinated hydrocarbons in Lake
Ontario sediments. Environmental science & technology,29(10), 2661-2672.
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
& Reservoirs: Research & Management, 7(3), 217-230.
Xing, Y., Lu, Y., Dawson, R. W., Shi, Y., Zhang, H., Wang, T., ... & Ren, H., 2005. A spatial
temporal assessment of pollution from PCBs in China. Chemosphere, 60(6), 731-739.

132

CHAPTER 4
SEASONAL INFLUENCE ON LAND USE BIOGEOCHEMICAL FINGERPRINTS IN
SURFACE WATER BODIES

Abstract
Surface water quality in many areas of the United States has declined over the past
several decades due to increases in loads of dissolved chemicals, some of which are approaching
toxic thresholds for aquatic biota. Land use characteristics and temporal loading characteristics
associated with the sources of the chemicals have yet to be ascertained in many cases. Five
synoptic sample events targeting seasonal and hydrologic variation in the Saginaw Bay
Watershed (SBW), Michigan were collected during the years 1995-96 to evaluate the
associations between land use and dissolved concentrations of major ions, such as calcium,
magnesium, sodium, etc. Significant local variations in stream chemistry were documented with
change in season and hydrologic regime, but when considering the entire SBW, the
concentration ranges for dissolved constituents generally were similar across all sampling events.
Stream chemistry data was combined with Geographic Information System (GIS)-derived land
use characteristics in a single database for use in an R-mode factor analysis to examine potential
relationships between land use and surface water chemistry across the entire SBW. Strong local
fluctuations of solutes were documented, attributed primarily to local proximity of urban centers
and agricultural land use. The lowest concentrations of solutes were observed in association
with forested land use. The fingerprint of agricultural land use on surface water chemistry was
most evident during a summer low flow event and a spring high flow event, which may be a

133

reflection of seasonal variations in the intensity of agricultural practices. Urban land use
exhibited its strongest fingerprint during winter high flow, likely due to road deicing salt
application. The SBW as a whole did not exhibit strong temporal fluctuation of sodium and
bromide concentrations despite considerable application of salt applied as a road-deicer in the
winter months. Cl/Br ratios are analyzed as tracers of surface water processes. Overall, we
conclude that studies attempting to link chemical concentrations to a particular land use should
give consideration to coordinating surface water analytes to the hydrologic regime and time of
year the sample collection is to be undertaken.

134

Introduction
Land use change has been shown to adversely affect the chemistry of aquatic ecosystems
(Fitzpatrick, 2007; Lindeman, 2007; Deocampa, 2004, Grimm et al., 2003; Wayland, 2003).
Dissolved solids concentrations is one aspect of aquatic chemistry profoundly affected by
anthropogenic alterations to land use (e.g., deforestation) (USEPA, 2000). Dissolved solids have
been linked to various impacts on surface water integrity such as decreased fish communities
(Meador and Goldstein, 2003) and trophic quality of wetland communities (Crosbie and ChowFrasier, 1999). With dissolved solids concentrations increasing over the last several decades in
surface water bodies of the Great Lakes region (Susanna and Chen, 2002; Richards et al., 1993),
identifying the sources of the dissolved solids has become increasingly important, especially
given the number of other stressors that may collectively contribute to overall degradation of
environmental quality and ecosystem integrity within the Great Lakes region (Allan et al., 2013).
Many of the increases in dissolved solids are directly attributable to anthropogenic practices such
as road salting (Kaushal et al., 2005; D’Itri, 1992) and nutrient addition to agricultural lands
(Gordon et al., 2010, Hofmann et al., 2010). Understanding sources and fate of these chemicals
is a need important to address, given continued expansion of urban areas, the likely expansion of
practices such as road deicing and the fact that these chemicals not flushed out of SBW in near
term may recharge shallow groundwater and raise baseflow concentrations.
The SBW in central Michigan is an area susceptible to the adverse effects of the
increasing dissolved load in the aquatic ecosystem, given the significant changes in land use
from pre-settlement to the current mix of land-use types (Einheuser et al., 2013; Giri et al.,
2012). Water quality in the SBW has declined over the last several decades, prompting studies
to identify potential pollutant sources and to develop strategies to mitigate or reduce these

135

sources (Science Subgroup of the Great Lakes Regional Working Group, 2013; Selzer, 2008;
Public Sector Consultants, Inc., 2002). Although changes in stream chemistry can be attributed
to land use (Riseng et al., 2010; Johnson et al., 1997), separating the relative contributions
natural sources (e.g., natural landscape erosion), anthropogenically-enhanced natural processes
(e.g., accelerated weathering of soils due to agricultural tilling), and direct anthropogenic
additions (e.g., dispersal of road deicers) is challenging.
Previous studies of the SBW have evaluated general river water quality but lacked
extensive spatial coverage to determine solute sources (e.g., Wood, 1970; Tiffany et al., 1969) or
focused primarily on nutrient concentrations (Giri et al., 2013, Dodds and Oakes, 2006; Johnson
et al., 1997). As urbanization and agricultural land use in the SBW continues to change, there
exists the potential for the increase in chemical fluxes associated with both agricultural and urban
land use. Use of agricultural chemicals, septic systems and landfills, and road-deicing agents
associated with these land use types have been shown to result in elevated concentrations of
dissolved solids in surface water bodies (Barber et al., 2006; Baker, 2005; Allan and Johnson,
1997). Fingerprinting the sources of these dissolved solids for any particular watershed has
challenges arising from spatial and temporal variations in the fluxes of chemical from the
landscape to the surface water bodies (Hildebrandt et al., 2008). Research in temporal variations
has shown conflicting views in the magnitude of the flux of chemicals from the landscape: some
studies exhibit highly variable chemical concentrations through time (Clow et al., 1996) and
other show water chemistry to be relatively constant (Pionke et al., 1999). Cyclical variations in
natural environmental conditions (e.g., air temperature, precipitation form/quantity) might be
expected to influence both the individual chemicals that constitute the dissolved load of a surface
water body and the total concentration of the dissolved solids. Also, variations in watershed

136

biomass and geology have been shown to complicate the use of cation and anion species for
environmental process interpretation (Drever, 1997). Finally, the presence of natural sources of
dissolved solids, such as soil or rock-water interactions and basin brines, elevated levels of
dissolved solids could be erroneously attributed to anthropogenic inputs. The upward flux of
solutes from subsurface brines of this region has been shown to potentially alter the geochemical
cycling of chloride or metals (Kolak et al., 1999). Thus, associations of dissolved solids
concentrations to land use types in watersheds like the SBW is complicated by these natural
variations and conclusions as to a biogeochemical fingerprint of a particular land use type may
change throughout the year.
Temporal heterogeneity in the flux of dissolved solids to surface water bodies may be
further modified by temporal variations in anthropogenic activities. Significant landscape
disturbance variations cycling over a given year in predominately agricultural areas are likely to
include soil tilling, fertilizer application and crop harvesting. Urban sources are also expected to
contribute a cyclical load of dissolved solids due to road deicer application in the colder months
and the tendency for major construction projects to occur in the warmer months.
Spatial challenges in identifying sources of dissolved solids could include the effects of
distance traveled by water at the specific sample collection location. Head water/low order
streams could be expected to generate a stronger connection to land use than higher order
streams, which integrate chemical signatures from a larger area and heterogeneity in land use
characteristics. The amount of chemical processing and alteration of water as it is transported
through soil or bedrock, prior to discharging to surface water may also vary as the relative
contributions of baseflow change. Dissolution, precipitation, biological activity within the root
zone, anion sorption, and exchange all potentially affect dissolved solids introduced to the

137

groundwater (Drever, 1997), thereby complicating interpretation of source and biogeochemical
fingerprint.
The study reported herein focuses on comparing GIS-calculated land use from 72 sample
locations to water-quality data from five synoptic sampling events obtained from July 1995 to
May 1996. Interpretation of complex data sets including geospatial components frequently
includes the use of multivariate statistical methods. The large number of variables resulting from
such studies is difficult and time consuming to thoroughly analyze. Our approach here is to
apply factor analysis (also referred to as principal component analysis) on the data set generated
by both laboratory results from surface water sampling and the GIS-calculated land use
associated with each sample locations. Factor analysis is a statistical method used to describe
variability among observed, correlated variables in terms of a potentially lower number of
unobserved, uncorrelated variables called factors. Factor analysis returns a quantitative
assessment of the strength of a series of factors in explaining the variance of variables in the
dataset (Gorsuch, 1974). Thus, a detailed biogeochemical fingerprinting can be developed to
identify how the chemical variables behave amongst the different land use characteristics
(Wayland et al., 2003). This methodology has been applied been successfully applied in a
number of related works (Fitzpatrick, 2007; Wayland et al., 2003; McGuire et al., 2000; Drever,
1997; Long et al., 1992; Long et al., 1988).
This research was driven by the hypothesis that periods of elevated runoff should produce
a much clearer biogeochemical signature of respective land use types than periods of low flow.
The primary goal of this paper is to provide an optimal time of year for collection of surface
water samples used to represent a particular land use with an ancillary goal of discerning the
relative contributions of these chemicals by their respective land use types. In this study, we

138

provide an evaluation of spatial and temporal variations of water quality in the SBW along with
statistical association (factor analysis) of the chemical results with GIS-calculated land use.

139

Methodology
Study Site
This study was conducted in the Saginaw Bay Watershed (SBW) in east-central Michigan
(Figure 1). The SBW is Michigan's largest watershed (22,556 km2), includes portions of 22
counties, and contains the United States’ largest contiguous freshwater coastal wetland system.
The presence of saline fluids in the SBW hydrogeologic system has been documented as far back
as 1838 (Houghton, 1838). This phenomenon has created an excellent location for a history of
studies regarding ground-water, large-lake interactions (Kolak, 1999; Mandle and Westjohn,
1989; Wahrer et al., 1996). These studies, along with those focused on toxic chemicals (e.g.,
Kannan et al., 2008) and pollution loadings (e.g., He et al., 2006) have provided a significant
literature body describing the hydrologic system of the SBW. Bedrock lithology (e.g., extent of
Michigan confining unit) does limit ground-water discharge in the case of the Saginaw Bay itself
(Hoaglund et al., 2004), but in the case of SBW, ground-water surface water interaction is likely
dominated by the Glaciofluvial aquifer (i.e., a shallow circulating system), and bedrock does not
constrain this exchange. It is likely that glacial material (lodgment till, etc.) limits bedrock
aquifer contributions to SBW. Wahrer et al. (1996) supports this interpretation: only a small
portion of Glaciofluvial aquifer has total dissolved solids (TDS) measurements in excess of
10,000 mg/L. For most of glaciofluvial aquifer underlying SBW, TDS measurements are below
1,000 mg/L. Given this lack of communication between bedrock aquifers (and associated
brines/saline ground-water) and surface water bodies in the SBW, brine inputs will not likely
affect land use fingerprints, allowing river chemistry to more adequately represent the
biogeochemistry of land use.

140

Figure 4.1: Location of Saginaw Bay Watershed, Michigan, and sampling sites.

Sample Plan and Analysis
Biogeochemical data were collected during five sets of synoptic river water sampling
events in the SBW from July 1995 to May 1996. Each set of river water samples was collected
with the intention of providing a large “snapshot” of stream chemistry by collection of samples
from a large number of sample sites in a short period of time (Eyre and Pepperell, 1999; Fetter,
1994). Use of this sampling methodology (provided flow remains relatively constant) eliminates
discharge-related chemical variability and allows for the spatial relationships of land use to
stream chemistry to be explored (Grayson et al., 1997). Each synoptic sampling event was
141

designed to target different hydrologic flow regimes (i.e. high flow, low flow) to allow
evaluation of temporal variations in solute concentrations (Figure 4.2 and 4.3). Sampling events
RIV-I and RIV-III were collected during summer, low-flow events. Sample event RIV-II is a
summer, high-flow event. RIV-IV is a winter, high-flow event and RIV-V is a spring, high-flow
event. Hydrologic flow regimes were monitored prior to each sampling event via U.S.
Geological Survey gaging stations located within the watershed. All samples were collected
within 3-4 days for each event to minimize potential artifacts from a change in flow regime.
The samples collected in this study were filtered (0.45 µm) and analyzed for major
cations (calcium, magnesium, sodium, and potassium), anions (chloride and bromide) and
dissolved silica (as SiO2). Filtered samples for major cations were acidified in the field to pH<2
with ultra-pure nitric acid. All samples were stored on ice during transport to the laboratory, and
refrigerated thereafter. During one sampling event (RIV-III), a subset of samples was subjected
to further analyses (pH, alkalinity, and T, all measured on site; and sulfate, measured in the
laboratory) to support geochemical modeling efforts to delineate mineral-water interactions that
might influence river chemistry. Processing of analytes intended for geochemical modeling (pH,
alkalinity, and temperature) was conducted immediately in the field. Samples intended for
dissolved major cations and anion analysis were field filtered with disposable 0.45 um filters and
laboratory analysis was typically initiated within 72 hours following completion of the sampling
event. Sample processing and analytical procedures are described in greater detail in Kolak
(2000). Stream chemistry data were then combined with GIS-derived land use characteristics in
a single database for use in an R-mode factor analysis to examine the relationships between land
use and surface water chemistry.

142

Figure 4.2: Stream hydrographs for the Rifle and Shiawassee Rivers

143

Figure 4.3: Stream hydrographs for the Cass and Tittabawassee Rivers

144

Watershed and Land Use Distribution
Land use and sampling points were related by GIS-based calculation of the up-gradient
area contributing to a given sample location (Figure 4.4). The polygon layer created in these
calculations was then intersected with land use data and proportions of each land-use type were
exported to Excel for use in statistical analysis. Land use data were obtained from the 2001
IFMAP/GAP Lower Peninsula dataset (Michigan Department of Natural Resources, 2003) for
each study catchment within the SBW. Land use was classified using the upper-most level
Anderson classification and resulted in 7 land use types: (urban, agriculture, upland/openland,
forest, water, wetland, and barren). Average land use in the study catchments closely reflected
that of the total SBW. Land use/land cover in the SBW is agricultural (55%) followed by
forested (26%) land use. Urban land use comprises only approximately 5% of the land use
studied.

Figure 4.4: Example drainage area calculations for sample sites 9-4, 18-2, 27-8, 28-8, and 29-5.

145

Statistical Procedures
Interpretations of factors in the context of biogeochemical sources and processes can be
difficult (Drever, 1997). Varimax rotation was used as this solution simplifies the interpretation
of the factors allowing variables to identify with a single factor. A loading close to +/- 1
indicates strong correlation between a variable and a factor, while loadings close to 0 indicate
weak correlation (Davis, 1986; Evans et al., 1996). Variables that exhibit a rotated loading
greater than 0.5 was considered moderately loaded on a factor, while variables greater than 0.75
were considered strongly loaded on a factor. In this research, factor analysis for each sample
event individually and for all sample events together was performed using SYSTAT software
(Wilkinson, 2008). A chemical fingerprint (chemical grouping) was sought for each land use
type using this methodology.
Examination of the literature has shown there to be a large number of multivariate
statistical methods applied to environmental data (Howell, 2007; Skrbic and Durisic-Mladenovic,
2007; Onuska and Davies, 1991). It is suggested that use ranked data is useful for describing
compositional data (such as those data used here) as it avoids many problems associated with the
analysis and interpretation of raw compositional data (Howell, 2007). Rank order has the
greatest influence on the correlation matrix upon which factors are based (Gorsuch, 1974) and as
factor analysis should be most sensitive to rank order, all factor analysis reported in this paper
were reported on ranked data.

146

Results
Temporal Chemical Variability
The RIV-I sample event was only tested for bromide, chloride, and sodium. The
remaining sample events were subject to an expanded suite of analytes to include major cations.
The summary statistics for all sample events is presented as Tables 2. It appears from the data
that elevated dissolved concentrations were not confined to either low or high flow regimes,
rather, RIV-I (high flow) and RIV-I (low flow) showed average dissolved concentrations of
many of the chemicals studied in excess of the other sample events. This is true for bromide,
chloride, sodium, magnesium, and SiO2 but not for potassium (RIV-I and RIV-IV had highest
average). This variability was not expected as all high flow events were expected to have the
lowest concentrations due to dilution by meteoric water. Concentrations of several dissolved
ions, including chloride, potassium, and sodium, varied within each sample event by an order of
magnitude or more between the minimum and maximum levels. Calcium was not as temporally
variable possibly related to the dissolution of carbonate minerals from widespread glacial
deposits present across the portion of Michigan occupied by the SBW (Phillips, 1988; Long et
al., 1988). Geochemical modeling provided by Kolak (2000) found that RIV-III samples
throughout the SBW were at equilibrium or supersaturated (SI >0) with respect to calcite,
supporting this interpretation. In addition to the glacial geology of the SBW, agricultural
applications (e.g. field liming) may also have a significant influence on the prevalence of Ca2+ in
the measured samples given the prevalence of agricultural land use in the SBW (see Table 1).

147

Table 4.2. Basic statistics organized by sample event.

RIV-I Sampling (Summer Low Flow)
Variable

RIV-II Sampling (Summer High Flow)

No. of Samples

Min.

Mean

Max

CV

No. of Samples

Min.

Mean

Max

CV

-

50

1.80

44.41

156.00

0.80

64

1.90

55.31

689.00

1.62

Br-

50

0.11

0.44

2.75

0.99

64

0.13

0.58

7.36

1.89

49

0.50

18.90

94.64

1.04

Cl

+

Na

64

0.80

19.00

118.00

0.87

Ca2+

64

22.80

79.31

125.00

0.30

2+

Mg

64

7.00

26.63

40.40

0.29

K+

64

0.90

4.56

16.40

0.58

SiO2

55

0.10

3.33

8.10

0.46

Temp
pH
Alkalinity

RIV-III Sampling (Summer Low Flow)

RIV-IV Sampling (Winter High Flow)

Variable

No. of Samples

Min.

Mean

Max

CV

No. of Samples

Min.

Mean

Max

CV

Cl-

68

4.80

78.51

800.00

1.49

72

4.70

30.82

164.00

0.89

Br-

68

0.10

0.72

9.65

1.86

72

0.11

0.35

1.35

0.65

Na+

68

3.30

31.41

204.00

1.20

72

0.50

12.10

95.50

1.28

Ca2+

68

41.50

88.65

169.00

0.28

72

0.40

46.40

135.00

0.46

Mg2+

68

8.50

30.48

52.00

0.31

72

4.00

15.08

38.40

0.47

K+

68

0.70

4.57

13.40

0.63

72

1.10

6.50

35.20

0.72

71

1.10

2.60

5.98

0.44

SiO2

68

0.30

3.23

18.60

0.79

Temp

59

8.14

18.63

25.90

0.16

pH

14

7.35

7.90

8.14

0.03

Alkalinity

14

231.00

291.08

412.00

0.17

Variable

No. of Samples

Min.

Mean

Max

CV

RIV-V Sampling (Spring High Flow)
Cl-

63

1.70

37.48

255.00

0.88

Br-

63

0.06

0.37

4.73

1.56

Na+

63

0.20

13.40

80.50

0.98

Ca2+

63

22.00

79.92

135.00

0.28

Mg2+

63

6.00

22.21

35.00

0.27

K+

63

0.30

3.48

8.80

0.43

SiO2

63

0.64

2.84

4.36

0.27

Temp
pH
Alkalinity

Factor Analysis of Individual Sample Events
Factor analysis results for individual sampling events are shown Table 3a through Table
3e. The seasonal and hydrologic characteristics of the five analyses are not identical although

148

some similarities can be observed amongst them. Two sample events were conducted during
summer low flow (RIV-I and RIV-II) to observe the chemical characteristics of the surface water
body during the maximum contribution of groundwater (baseflow). The km2 variable is included
to ascertain whether watershed size would correlate to contribution of dissolved solids.
Chloride/bromide and chloride/sodium concentration ratios (Cl:Br, Cl:Na, respectively) were
included as a means of attributing salt concentrations to source. These were calculated as ratios
of the concentrations determined by the analytical results of the respective ions. Waters affected
by halite dissolution tend to have an elevated Cl:Br ratio, often exceeding 3,000 (Wilson and
Long, 1985). Waters influenced by a brine source, such as those in the Michigan basin, have a
Cl:Br ratio of 400 to 600 (Wahrer et al., 1996; and Meissner et al., 1996 ).
The RIV-I data set was limited in analytical scope and the only elemental correlation
made was between chloride, bromide, and sodium, which were all positively loaded together.
These analytes do not exhibit a strong statistical correlation with any particular land use
indicating they are potentially ‘disconnected’ from land use during summer baseflow hydrologic
regimes. However, Cl/Br concentration ratios behave differently and exhibited factor loading by
itself in this data set.
The RIV-II data set was collected during the recession limb of a summer high flow event
for most of the rivers included in this study. However, the hydrologic response to the
precipitation event driving the summer high flow in these rivers was not uniform across the
entire SBW. For example, the Cass River (Figure 25; Kolak, 2000) illustrates an instance of a
minimal hydrographic response to the precipitation event. Given the large spatial area
incorporated in this project, this is an important consideration as differential precipitation and

149

snow melt has the potential to disrupt consistent biogeochemical associations between river
chemistry and land use.
Despite this, agricultural land use showed positive correlation with the Cl:Br ratio, as
well as calcium and magnesium. These chemical loadings may reflect agricultural control via
landscape manipulation (e.g., tilling) and flushing from the larger hydrologic event, possibly
including the influence of fertilizer application as well. As magnesium is often considered a
proxy for erosion or landscape denudation (Long et al., 2010, Tipper et al., 2008), this can be
interpreted as the season exhibiting elevated erosion rates from this land use type. The reason
for a positive correlation between agricultural land use and Cl:Br ratio is unclear as much of the
relic road salt applied during the previous colder months is expected to have been flushed from
the system. Septic effluent inputs from these rural localities offer a possible explanation as city
sewer systems commonly associated with the more population-dense urban environments.
Sodium concentrations measured during RIV-I showed the weakest correlation to Br and Cl,
suggesting the possible influence of other significant sources such as natural erosional sources
or, more likely, the lack of correlation is the result of differences in environmental behaviors
(e.g. Na adsorptive tendencies versus chloride conservative tendencies).
The RIV-III data set, like RIV-I, was also collected during a summer low flow period, but
with an expanded suite of analytes (Table 2). Here agricultural land use is moderately correlated
to calcium and magnesium, and strongly with potassium and alkalinity. Cl, Br and Na again
correlated strongly with one another under a separate factor. Cl:Br ratios were moderately well
correlated with both agriculture and urban land use types. The statistical disconnect between the
ratios of these elements from their given concentrations is difficult to explain as one would
expect a correlation between the two. Interestingly, the ratios appear to be driven by agricultural

150

land use, suggesting a common source amongst these sample location even if the concentrations
are variable.
Finally, this sampling event had a statistical correlation between watershed size and
dissolved SiO2. This result seems logical as the larger size of a catchment area would tend to
have deeper groundwater flow paths, enabling more mineral dissolution to occur than areas of
low groundwater retention. This same factor was also negatively correlated to pH which also
seems plausible as the smaller the catchment size, the less opportunity for waters in that
catchment to interact with the terrain carbonates that can raise pH.
The RIV-IV sampling was performed during a winter high flow event in the SBW. Here
results show a negative correlation between agriculture and SiO2. The implications of the
negative correlation here is unclear. The factor analysis exhibits strong loadings of urban in
relation to Cl, Br, and Na suggesting application of road salt as a deicer during the winter months
to play a dominant role in the quantities of salts in surface water bodies during this time of year.
Calcium and Mg correlated well together under Factor III with SiO2 factoring moderately.
However, these analytes did not correlate strongly to any particular land use type. This means a
common source for the cations and SiO2 but, whatever the source, is not restricted to a single
land use. Finally, %wetlands and %water show moderate control over the Cl:Br ratios. This
suggests the possibility of groundwater-surface water interaction playing a role in the presence of
the two elements, likely bromide, as the Cl:Br ratio is much lower in area brines (400 to 600)
than in the significantly more pure halite used as a road deicer (1,200 to 10,000) (Kolak et al.,
1999).
Factor loading for RIV-V (spring high flow) exhibited moderately strong correlation
between agricultural land use and Ca, Mg, and K. This result was expected initially for all

151

sampling events but only observed in loadings greater than 0.5 in RIV-III (summer low flow)
and RIV-V. Since the RIV-III sampling event was conducted in early September and RIV-V
was conducted in late May, these sampling events may have followed particular agricultural field
activities (fertilizer distribution, planting, harvesting, etc.) possibly contributing to soil
disturbance and run off from agricultural fields. The RIV-V data set also exhibited the typical
factor connection between Na, Br, and Cl. The elemental salt components were expected to
factor strongly with an urban environment as this landscape continues to hydraulically rid itself
of the road deicers built up during the winter. However, it appears the smaller hydrologic peaks
in the watersheds in the time period of RIV-IV and later, in the month of March (refer to Figures
2 and 3), removed road deicers enough that it did not agree as strongly with an urban
environment. There was a moderate correlation between Cl:Br ratios and %urban, suggesting
some influence of this landscape type over the presence of these ions in the watershed, but it is
not clear if this signature is sourced from relic road salt or other sources such as landfill and
septic effluent.
Table 4.2a: Factors for the RIV-I data collected 10-11 July 1995. Numbers in bold indicate strong loading
on a factor (loading > 0.75). Only loadings >0.5 or <-0.5 are shown.
RIV-I Sampling (Summer Low)

VARIABLE

I

FACTORS
III
IV

II

V

VI

0.769

Urban
Agriculture

-0.922

Upland Openland

0.915

Forest

0.908

Water
-0.925

Wetland
0.931

Barren
0.900

Total Km2
Cl

0.911

Br

0.929

Na

0.933
0.958

Cl:Br

152

Table 4.2b: Factors for the RIV-II data collected 20-21 July 1995. Numbers in bold indicate strong loading
on a factor (loading > 0.75). Only loadings >0.5 are shown.
RIV-II Sampling (Summer High Flow)
FACTORS
VARIABLE

I

II

III

IV

Agriculture

VI

0.703

Upland Openland

0.684
-0.750

Forest
0.773

Water
Wetland

-0.714
0.933

Barren

0.936

Total Km2
Cl

0.946

Br

0.962

Na

0.683

Cl:Br

0.723

Ca

0.653

Mg
K

V

0.875

Urban

0.729
0.638
0.775

SiO2

Table 4.2c: Factors for the RIV-III data collected 1-5 September 1995. Numbers in bold indicate strong
loading on a factor (loading > 0.75). Only loadings >0.5 are shown.
RIV-III Sampling (Summer Low Flow)
FACTORS
VARIABLE

I

II

III

IV

V

VI

0.943

Urban
Agriculture

0.939

Upland Openland

-0.850

Forest

-0.920
-0.806

Water
Wetland

0.694

Barren

-0.809

Total Km2

-0.890

Cl

-0.953

Br

-0.960
-0.890

Na
Cl:Br

0.537

Ca

0.710

Mg

0.688

K

0.786

0.577

-0.869

SiO2
-0.792

T (°C)
pH
Alk (HCO3-)

-0.555
0.780

153

0.583

Table 4.2d: Factors for the RIV-IV data collected 21-25 February 1995. Numbers in bold indicate strong
loading on a factor (loading > 0.75). Only loadings >0.5 are shown.
RIV-IV Sampling (Winter High Flow)
FACTORS
VARIABLE
Urban

I

Agriculture

-0.943

Upland Openland

0.911

Forest

0.951

II
0.853

III

IV

V

Water
Wetland

VI

-0.625
0.575

-0.566
0.943

Barren
Total Km2

0.983

Cl

0.978

Br

0.923

Na

0.958

Cl:Br

0.539

0.671

Ca

0.926

Mg

0.882

K
SiO2

0.548

0.577

Table 4.2e: Factors for the RIV-V data collected 23-25 May 1995. Numbers in bold indicate strong loading
on a factor (loading > 0.75). Only loadings >0.5 are shown.
RIV-V Sampling (Spring High Flow)
FACTORS
VARIABLE

I

II

III

Urban
Agriculture

0.935

Upland Openland

-0.888

Forest

-0.882

Water
Wetland

IV

V

VI

0.506

0.594
-0.543

-0.582
0.902

Barren

0.979

Total Km2
Cl

0.954

Br

0.911

Na

0.949
0.850

Cl:Br
Ca

0.797

Mg

0.745

K
SiO2

0.606
0.944

154

Factor Analysis of Combined Data Sets
Combining datasets from all five sampling events is included to investigate the statistical
loadings amongst the variables across an entire year. Though this would likely obscure seasonaldependent observations, more persistent influences have the potential to be revealed. Factors for
rank-ordered data from the combined sampling events are shown in Table 6. RIV-II included the
additional analytes temperature, pH, and alkalinity. Samples collected for RIV-I were analyzed
for an abbreviated suite of constituents: Cl, Na, and Br. The combination of data sets resulted in
a total of 317 individual samples. Proportions of agricultural land use, along with alkalinity,
exhibited strong loadings on Factor I; K was moderately loaded on this factor. Factor I can
possibly be explained by increased dissolution of soil minerals as a result of soil cultivation. This
process could increase concentrations dissolved calcite (CaCO3), dolomite (CaMg(CO3)2) and
potassium feldspar (KAl2SiO8). Because of this, it was questioned by Ca or Mg was not strongly
loaded as well on Factor I. More likely, K correlations with agricultural land use are the result of
application of potassium-containing fertilizers. This explanation agrees with Panno’s findings
(Panno et al., 2006) where a mass balance calculation was performed showing consistencies
between KCl application and water chemistries analyzed from tile drains of croplands.
Ca and Mg were shown to be loaded strongly on Factor VI which also has the agricultural
land use type loaded at 0.425, a little below the threshold of its reported value in the tables of
0.5. All other land use types were reported as a weaker negative value (closer to 0 than -0.5) on
Factor VI. Other researchers have noted similar loadings of alkalinity, Ca, and Mg together
under a given factor and interpreted the results as mineral weathering (Fitzpatrick et al., 2007;
Wayland et al., 2003; Puckett and Bricker, 1992; Schott and Van der Wal, 1992).

155

Factor II had strong loadings of Na, Cl, and Br together. As suggested previously,
sources of these ions could include road deicers, landfill leachate, septic effluent, and possibly
even subsurface brine interaction with surface water. A closer examination of potential source of
these ions is included in the next section. Factor III showed %water on the landscape to be
loaded with temperature and pH while inversely loaded with SiO2. Dissolved SiO2 and pH are
more difficult to explain. One possible explanation is that groundwater in the region is more
basic has a higher pH due to mineral-water interactions (e.g. calcite dissolution) that have the
potential to raise pH. Surface water has the potential to have a lower pH due to CO2 dissolution
in the surface water and the precipitation. Factor IV shows urban land use correlating to the
Cl:Br ratio. However, Cl, Br, and Na are all loaded together on a different factor (II) suggesting
a disconnection between urban land use and Na.
The individual analyses show that Ca, Mg generally seem to load on same factor with
agriculture (aside from winter high flow which can possibly be explained by frozen ground and
strength of urban signature. However, in the combined analysis of all data sets, Ca and Mg are
disconnected from agriculture. It is not clear if this disconnection is real or an artifact of the
strength of urban influence. If this is the case, the combined analysis may signify that the
combined analysis works to obscure interpretations of results.

156

Table 4.3: Factors for data from all sampling events combined. Numbers in bold indicate strong loading on
a factor (loading > 0.75). Only loadings >0.5 are shown.

RIV-I – RIV-V: Combined Data

VARIABLE

I

II

III

Urban

IV

V

VI

Agriculture

0.765

Upland Openland

-0.720

Forest

-0.760

Water

0.577

Wetland

-0.852

Barren

-0.887

Cl

0.971

Br

0.980

Na

0.898

Cl:Br

0.533

0.579

Ca

0.889

Mg
K

0.860
0.699

SiO2

-0.795

T (°C)

0.570

pH

0.874

Alk (HCO3-)

VII

0.920

0.593

0.927

Sodium, Bromide, and Chloride Sources in the SBW
Results from the statistical analyses performed indicate that Na, Br, and Cl were
significantly correlated to one another within each season the samples were collected. It has
been shown by others that ground and surface water contamination by road deicers is localized to
these waterbodies near roadways and the extent of the salt contamination is directly proportional
to the mass of salts applied (Foos, 2003). Given the chemically conservative nature of Cl and Br,
the anions have the potential to provide excellent fingerprints of urban land use (Smart et al.,
2001). However, these ions do not appear to have a strong relationship with any given land use
with the noted exception of urban during RIV-IV. This begs the question, what is the source of
these ions in the SBW?
157

If the source of these ions were simply attributed to road deicer application, Br would
likely not have been as strongly loaded with Na and Cl (Panno, 2006). Likewise, if wastewater
were a strong influence on these ions, a significant quantity of samples should plot above the
halite line as septic systems and animal waste are typically enriched in Na relative to Cl
(Vengosh and Keren, 1996). However, it has been observed in studies of deicing salts (Shanley,
1994; Driscoll et al., 1991) and waste water (Vengosh and Keren, 1996) that the ratio of Na to Cl
is reduced due to uptake of Na via cation exchange whereas Cl is relatively conservative. Thus,
the molar ratio of Na:Cl in halite is initially 1:1 and should decrease due to cation exchange as
the dissolved salt moves through the subsurface. The data show a generally reduced presence of
Na relative to Cl (Figure 4.5) and, while the samples collected from the SBW plot are scattered,
they appear to plot primarily along the brine ratio of 0.61 to 1. Given that this study
encompassed both small and large catchment areas above a given sample location (Smallest =
2.64 km2, Largest = 912.46 km2), exchange alone probably did not account for the nearly
exclusive loss of Na relative to Cl. Rather, it is possible that a localized influence of
groundwater from the glaciofluvial aquifer with enriched salinity from subsurface brines played
a role in dissolved salts present.
Also supporting the argument of cation exchange as the likely large loss of Na ions from
the surface water is in the trend lines plotted for each of the sampling events. These plots are
shown to overlay the halite and brine ratio lines in Figure 5. One would expect cation exchange
to provide the largest loss of Na during summer low flows when the potential mineral-water
interaction is at its peak. Conversely, cation exchange should be the least during winter high
flow events due to the increased amount of surface runoff. The latter is shown to be true with the
winter high flow (RIV IV) to overall exhibit the closest values of Na and Cl to that of halite.

158

However, the former argument is not supported by the data with a summer high flow (RIV II)
showing the most suppression of Na versus Cl. It’s possible that brine influenced areas of the
glacialfluvial aquifer are providing sufficient salts to confound the chemical results, raising the
Na concentrations relative to the Cl in the summer low flows (RIV I and RIV III). The RIV II
trend line is shown to plot much lower (largest loss of Na or gain of Cl) than the other event
trends possibly the result of maximized cation exchange in shallow subsurface hydrologic flow
and interactions with active landscape biota, while the contribution of salinized deeper aquifer
waters are diluted.

Figure 4.5: Ratio of Na+ to Cl- (moles/L) in all samples collected in the SBW. Simple halite dissolution would
show a 1:1 molar ratio. The brine ratio is represented by 0.61 to 1 which is typical of Michigan Basin brines
(Wilson, 1989).

159

Another technique used to reconstruct history or provenance of water systems is in the
examination of Cl:Br ratios. Cl and Br behave conservatively in both ground and surface water
and ratios of these two elements typically remain unaltered until a water of a dissimilar ratio is
mixed into the system (Davis et al., 1998). Thus, most small catchment basins and shallow
ground water tables reflect the Cl and Br concentrations of regional precipitation. In these cases,
Cl:Br ratios are usually between 80 and 160. Oil field waters typically have ratios between 200
and 400; sewage: 300 to 600; waters directly affected by halite dissolution: 1,000 to 10,000
(Davis et al., 1998). Cl:Br ratios versus Cl concentrations of all samples collected from the SBW
are shown as Figure 6. Lines indicate approximate regions where groups of samples have a
tendency to plot as depicted by Panno et al. (2006). The vast majority of the samples collected
plot in the ‘landfill leachate’ range of the figure. Given that urban land use of the watershed
studied amounted to merely 5% and the land use majority being agricultural and forested (55%
and 26%, respectively), it seems unlikely that landfill leachate could provide the Cl:Br chemistry
necessary for such ubiquitous results of the surface water samples collected. The Cl:Br ratios
would likely present a chemistry representative of a mix of sources. However, the tendency of
the Na/Cl ratio to plot along the brine signal (Figure 4.5) gives indication that a process is
present diminishing the Na concentrations or enriching the Cl concentration. With this in mind,
it seems more likely that this signal is representative of a mixture of halite inputs along with
other sources/processes (e.g. cation exchange) on a basin-wide scale. These processes of
enrichment of Cl or loss of Na have suppressed the Na:Cl ratios and manipulated the Cl:Br ratios
to provide surface water chemistry that may be interpreted as influenced by basin brines via
ground water-surface water interactions. However, the Na:Cl ratios of the independent sample
events show the event with ratio closest to halite dissolution (Na:Cl ratio = 1.0) occurred during

160

the RIV-I sampling (Summer Low Flow I). The Na:Cl ratio mean for RIV-I was the highest at
0.63 and the Na:Cl ratio for RIV-IV (Winter High Flow) was the lowest at 0.33. While RIV-I
averaged to have higher concentrations of both Na and Cl, concentrations of Na were elevated
with respect to Cl for this event. This result was not anticipated because it would seem that road
deicers in the winter and cation exchange during periods of low flow (larger contributions of
baseflow) would provide the opposite mean ratios. The results give some indication that the
elevated levels of Cl relative to Na observed during the high-flow events are related to surface
runoff and shallow groundwater; not from interaction with basin brines.
Independent observations of the Cl:Br ratios from each sampling event also show the
highest flow events showed the lowest Cl:Br ratios. RIV-III (the second Summer Low Flow)
resulted in the highest mean Cl:Br ratio of approximately 119.02 (dimensionless) while the RIVIV (Winter High Flow) resulted in the lowest mean Cl:Br ratio of 81.73. The remaining sample
events followed a similar pattern (RIV-I = 108.77, RIV-II = 102.34, RIV-V = 109.40). This
finding is counterintuitive as waters influenced by halite dissolution generally have a Cl:Br ratio
in excess of 3,000 (Wilson and Long, 1985). If halite road deicers alone dictated the ratio of
Cl:Br in the winter high flow observations, mean Cl:Br ratios during this sampling event should
exceed samples collected during other periods. Given the agricultural land use majority, the
potential exists for the legacy contributions of methyl bromide (a soil fumigant) to complicate Br
chemistries (methyl bromide was phased out by the EPA in 2005 [EPA, 2014]). However,
application of this compound would be used during the crop growing season exclusively,
promoting Br release to the environment during this period or briefly afterward. Ratios of Cl:Br
measured in river samples should be the most reduced then during the spring or summer months,
but again, the opposite is observed here. More likely, additions of KCl used primarily for

161

making liquid starter fertilizers in agricultural areas serve as the confounding source for Cl:Br
ratio observations.
It appears that simple assignment of Na-Cl-Br provenance in the SBW is confounded by
many factors. Others have proved successful in delineating similar solute sources in watersheds
(Herczeg et al., 1993), but, complications arise in the SBW given the variation of land uses,
seasonal change in fertilizer and road deicer application. These, along with the potential
presence of basin brine interaction with the surface water, prevent an easy definition to linking a
specific Na-Cl-Br chemistry with a particular land use. Supplementing the already extensive
data set with other geochemical analyses may be necessary to provide a definitive source of these
specific solutes. Panno suggested more distinctive separations of sodium and chloride sources to
be possible with analysis of boron (B) and iodine (I) as the septic effluent and effluent-affected
ground water samples plot near the origin, whereas animal waste is enriched in B and I2 (Panno
et al. 2005). Reviewing the data as a whole, there are several potential geochemical mixing
pathways. However, if based on spatial analysis of land use and amount of agricultural land use
in SBW, perhaps agricultural drainage seems most plausible, but more data are needed to support
this statement.
Despite this, synoptic sampling events in the SBW did document seasonal trends in
dissolved solids amongst the expanded sets of analytes. The factor analysis shows many of the
expanded set of water chemistry analytes to be directly attributed to land use but only on a
seasonal basis. Land-use analysis of the data primarily associates the strongest biogeochemical
fingerprint with urban centers and agricultural land use. The lowest concentrations of solutes
were observed in association with a high proportion of forested land use within a catchment.

162

Overall, the hypothesis was proven true and the relative strength of a biogeochemical
fingerprint for a particular land use type to be dependent upon season. Agricultural land use
exhibited its most significant fingerprint during a summer low-flow event and a spring high-flow
event, both of which possibly coincided with agricultural practices involving disturbance of the
soils during this time. This implies that the recent disturbances of the soil by agricultural activity
allows for manifestation of a discernable biogeochemical fingerprint in water bodies local to
these activities. The agricultural fingerprint included calcium, magnesium and potassium for
several synoptic events during of the study, as well as, alkalinity in the RIV-III analysis. The
RIV-IV analysis (winter high flow) was the only other high flow measurement that did not
correlate agricultural land use to any dissolved solids (negatively correlated to silica), however
the surface landscape at this latitude is typically frozen during much of the winter months
possibly preventing significant erosion from this land use.
Urban land use was well correlated to Cl, Br, Na, and Cl:Br ratio in the winter sample
event (RIV-IV) only. As this suite of chemicals is strongly associated with urban land use only
during winter months it can be interpreted as likely sourced from road deicer application. Urban
land use type was not factored to water chemistry at any other period with the noted exceptions
of a more moderate correlation to Cl:Br ratios during one summer low flow (RIV-III) and the
spring high flow (RIV-V). Several studies have shown that anthropogenically sourced Na and Cl
in groundwater and surface water bodies is often attributed to urban land use (Panno et al., 2006;
Mason et al., 1999; Buttle and Labadia, 1999). Our observations allow for a similar conclusion
in this study but the degree to which urban environments influence dissolved Na and Cl levels
appears to be heavily dependent on the season of which the sample is collected. It is unclear
why Br is linked to %urban as halite distributed as a road deicer would have little Br included in

163

it. Organic matter decomposition could contribute some Br to the surface water confounding
results, but it is difficult to conclude this as a reasonable source during the winter months. The
inclusion of Br with the suite of cations is possibly an indication of the influence of regional
baseflow chemistry (which, in turn, may be influenced by subsurface brines) on the surface
waters. This interference may be the reason %urban land use does not correlate well with Na
and Cl except winter months harboring discreet events (e.g. road deicing) producing a strong,
localized signal overprinting that of the background signal. It is also interesting to note that
%urban land use did not correlate with Na, Cl, and Br during the spring high flow but did
correlate to the Cl:Br ratio, indicating that the watersheds seem to rid themselves of the largest
quantities of salt deicers applied during the winter months, yet retain some control over the salt
concentrations into the spring.
Figure 4.6 shows the plotting ranges of different Cl and Br source (from Panno, et al.,
2006). Cl:Br ratio is plotted on the y-axis and Cl concentration along the x-axis. The data
collected from this study is plotted over the source divisions showing that many of the of the data
gathered plot in the range of landfill leachate boundary with brines and animal waste. This
figure alone would indicate that urban and agricultural centers to dominate the Cl Br chemistries
of the SBW surface water. However, it is clear that this conclusion is true only seasonally.
Figure 4.7 shows the triplot of the glaciofluvial aquifer studied by Wahrer et al., 1996.
As TDS increases (up to >1,000 mg/L), water chemistry shifts from mostly Ca-dominant, to no
dominant, to mostly Na+K dominant facies. Overplotting Wahrers data is samples collected
from four of the five RIV events. RIV I was not included as it was simply Na, Br, and Cl data.
In SBW, most samples (even those with TDS approaching 1,000 mg/L) plot in Ca-dominant to
no-dominant facies (especially at higher TDS). This may indicate a disconnect between SBW

164

and the Glaciofluvial aquifer in which case surficial inputs (landscape, surface runoff, interflow,
etc.) are a more important influence on hydrochemical facies

Figure 4.6: Cl:Br ratios vs Cl- concentrations of all samples collected from the SBW. Lines indicate potential
sources based on Panno et al., 2006.

Figure 4.7: Triplot showing Glaciofluvial aquifer data collected by Wahrer et al., 2006 in the SBW
overprinted with RIV II, RIV III, RIV IV, and RIV V samples. RIV I was not included as the analytes tested
during this event did not include Ca, Mg, Na or K.

165

A combination of all available data into a statistical model provides an normalization for
seasonal variation in activities associated with a given land use average set of correlations.
These calculations were performed for the observation of any persistent (e.g. multiple seasons or
the full year) biogeochemical signatures. The combined data set shows agriculture to be
correlated most often with K and alkalinity. Again, this correlation is interpreted as the result of
landscape disturbance and fertilizer distribution. Urban land use is closely correlated to the
Cl:Br ratio showing that urban environments elicit some control over this ratio. However,
observations made amongst the individual sample events show the correlation between %urban
and the Cl:Br ratio to exist only the winter and spring high flows.
Changes in the biogeochemical fingerprint associated with land use throughout the year
could also be the result of significant changes in the amount of surface water chemistry
interaction with the environment. As discussed, there has been documented evidence of ion
exchange due to interaction of a dissolved solid with its environment. Thus reduction in the
Na:Cl ratios have been shown to decrease with increasing surface water residence time and
distance traveled as groundwater (Vengosh and Keren, 1996; Shanley, 1994; Driscoll et al.,
1991). This suggests the variability of Na, Cl, and Br preventing a statistical assignment to any
particular land use is driven by regional baseflow and, aside from road salting, there is not a
significant regional land use signal competing with the baseflow chemistry. This is shown by the
strong statistical correlation of %urban to Na, Cl, and Br only during the winter high flow.

166

Conclusion
Overall, we conclude that studies attempting to link chemical concentrations to a
particular land use should give consideration to coordinating surface water analytes to the
hydrologic regime and time of year the sample collection is to be undertaken. The seasonal
changes in surface water chemistry are clearly modified by anthropogenic activities, particularly
during high flow regimes, when baseflow chemistry is overwhelmed with land use signals. This
study provides further evidence that surface water studies quality must give careful consideration
to both spatial and temporal factors before field activities are executed.
Continued research in this vane is needed as modeling scenarios in similar watersheds in
Michigan have shown deicing practices to have a more widespread effect on watershed
chemistry than anticipated and there could be a time lag on the order of decades or more before
the water chemistry impacts of past chemical distributions are realized (Boutt et al., 2001). This
study also documents several samples exceeding chloride concentrations of 250 mg/L,
illustrating the need to better understand Cl sources as the potential for adverse effects on aquatic
ecosystems exists

167

Table 4.4: Sample ID, location, and land use distribution for sampling of Saginaw Basin Watershed, MI,
USA. Values in bold indicate highest land use percentage. Dominant land use is defined as land use
percentage >50%. Those sample IDs without a land use >50% are defined as ‘mixed’.
Sample ID

Latitude

Longitude

%
Urban

%
Agriculture

%
Upland

%
Forest

%
Water

%
Wetland

%
Barren

Dominant
Land Use

DTL
1-1
1-1a
1-5
1-5a
2-1
2-4
3-1
3-1a
3-2
4-1
4-2
4-2a
4-3
5-1
5-1a
5-2
5-3
5-4
5-6
5-7
5-8
8-2
8-2b
8-4
9-1
9-1b
9-2
9-2a
9-3
9-3b
9-4
9-7
12-1
17-1
17-2
18-1
18-2
20-1
20-7
21-1
21-2
21-3
22-1
22-1b
22-2
22-4
22-4a
23-1
23-2
26-2
26-2a
26-3
27-1
27-1a
27-2
27-8

43.6753
44.3721
44.3775
44.3342
44.3675
44.0331
43.9450
43.8844
43.9050
43.8944
43.8069
43.7969
43.8010
43.8442
43.8925
43.8636
43.7925
43.7544
43.6674
43.6675
43.6386
43.6299
43.7408
43.7619
43.6458
43.6647
43.6690
43.7569
43.7570
43.6783
43.6700
43.6922
43.6983
43.5533
43.3429
43.3722
43.2430
43.3072
42.8760
43.0864
42.8950
42.8855
42.9425
43.0242
43.0082
42.9783
43.0931
43.0619
42.8302
42.8625
42.9325
42.9353
42.9340
43.1442
43.1872
43.1482
43.1158

-84.7081
-84.1061
-84.1056
-84.0636
-84.0517
-84.0678
-84.0210
-84.0256
-84.0210
-83.9379
-84.0864
-84.0461
-83.9640
-84.0264
-84.1856
-84.1864
-84.2272
-84.1447
-84.1670
-83.9701
-83.9778
-84.0503
-85.1450
-85.1478
-84.9775
-84.7878
-84.6910
-84.7064
-84.6690
-84.6875
-84.6828
-84.6275
-84.7550
-85.1055
-84.5080
-84.4301
-84.5829
-84.3694
-83.8673
-84.1397
-83.6578
-83.7542
-83.8580
-83.9100
-83.9303
-83.9878
-84.0267
-84.0897
-83.5615
-83.5152
-83.4206
-83.3797
-83.3315
-83.1722
-83.1350
-83.2333
-83.8598

5.4
2.7
1.4
2.5
2.6
3.3
2.4
4.5
3.1
6.6
3.8
3.6
4.9
3.2
2.0
2.1
0.3
1.8
4.0
3.7
7.2
6.9
2.5
1.8
3.6
5.1
5.8
5.9
5.8
5.9
5.8
4.9
6.2
2.8
5.0
4.3
3.6
5.5
4.5
3.8
4.1
8.9
12.1
9.9
11.1
2.7
3.2
3.2
6.1
3.8
1.3
4.5
3.6
1.3
2.5
1.4
21.4

86.2
25.6
11.6
7.1
6.6
8.7
41.5
24.0
55.9
57.9
23.2
33.9
56.6
53.6
2.5
7.6
0.7
8.9
34.0
31.0
65.6
58.9
39.4
20.2
86.5
74.6
78.9
79.2
67.0
81.9
78.5
71.7
79.7
39.3
85.9
89.7
81.6
86.3
70.7
77.4
48.8
47.6
51.0
71.6
68.9
93.4
88.5
92.1
15.3
17.7
22.8
40.9
49.2
43.9
53.3
15.5
33.5

2.6
10.0
11.1
17.3
17.9
16.0
20.5
20.9
7.3
9.0
13.8
16.5
9.4
10.9
23.6
19.5
9.6
13.8
15.8
13.9
7.0
9.0
32.9
16.6
3.9
3.6
4.9
4.0
10.4
3.3
5.1
6.1
3.0
19.9
1.9
1.2
2.6
1.5
3.5
2.2
11.7
10.4
9.5
4.9
5.1
1.0
0.6
0.5
13.0
11.0
10.2
11.7
8.9
9.3
7.7
8.2
14.0

3.8
60.7
71.2
63.0
63.4
49.0
32.8
41.6
31.3
22.3
57.1
44.0
26.7
30.6
64.1
65.0
67.2
67.4
43.3
46.3
17.5
22.3
23.4
45.1
5.1
10.7
7.4
6.7
11.9
5.5
7.7
13.1
6.6
34.4
5.6
3.6
9.7
4.7
17.5
14.7
29.3
24.6
22.1
12.3
13.0
2.6
6.3
3.5
49.3
56.4
52.0
37.2
30.8
37.5
28.3
58.3
25.2

0.1
0.0
0.9
2.1
1.9
0.0
0.0
0.1
0.0
0.2
0.0
0.0
0.2
0.0
0.0
0.0
5.9
1.0
2.5
0.5
0.2
0.3
0.0
8.6
0.0
0.1
0.1
0.1
0.7
0.1
0.1
0.2
0.2
0.6
0.0
0.0
0.0
0.2
0.2
0.0
0.0
0.4
1.5
0.0
0.0
0.1
0.0
0.0
2.9
2.4
0.3
0.5
0.3
0.0
0.5
1.7
1.0

1.8
1.0
3.7
8.0
7.5
23.0
2.5
8.8
1.6
3.7
1.9
1.8
2.0
1.4
7.8
5.7
16.3
7.0
0.3
4.4
2.2
2.3
1.8
7.6
0.8
5.8
2.6
4.0
4.0
3.1
2.6
3.9
4.0
3.0
1.6
1.0
2.5
1.6
3.5
1.8
6.0
7.6
3.8
1.0
1.7
0.1
1.3
0.8
7.1
8.6
13.2
5.2
7.1
7.9
7.7
14.9
4.5

0.1
0.1
0.0
0.1
0.1
0.0
0.2
0.1
0.8
0.2
0.1
0.1
0.2
0.3
0.0
0.0
0.0
0.1
0.0
0.3
0.2
0.2
0.0
0.0
0.1
0.1
0.2
0.1
0.2
0.2
0.2
0.1
0.3
0.0
0.0
0.0
0.0
0.1
0.1
0.1
0.1
0.4
0.0
0.3
0.1
0.1
0.0
0.0
6.2
0.0
0.0
0.0
0.0
0.1
0.0
0.0
0.4

Agriculture
Forest
Forest
Forest
Forest
Mixed
Mixed
Mixed
Agriculture
Agriculture
Forest
Mixed
Agriculture
Agriculture
Forest
Forest
Forest
Forest
Mixed
Mixed
Agriculture
Agriculture
Mixed
Mixed
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Mixed
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Mixed
Mixed
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Mixed
Forest
Forest
Mixed
Mixed
Mixed
Agriculture
Forest
Mixed

168

Table 4.4 (cont’d):
Sample ID

Latitude

Longitude

%
Urban

%
Agriculture

%
Upland

%
Forest

%
Water

%
Wetland

%
Barren

Dominant
Land Use

27-8a
27-8b
27-9
28-1
28-2
28-3
28-4
28-5
28-6
28-7
28-8
29-lake
29-1
29-2
29-3
29-3a
29-3b
29-4
29-4a
29-4b
29-5
29-6
31-5b
32-6

43.1333
43.1172
43.2062
43.2708
43.4300
43.5445
43.5323
43.7237
43.6643
43.6428
43.5675
43.9704
43.7022
43.7022
43.5997
43.6564
43.6564
43.7008
43.8019
43.7680
43.9380
43.8283
43.5360
43.4083

-83.7522
-83.8356
-83.7987
-83.0238
-83.0665
-82.9757
-83.0442
-82.9330
-82.9997
-82.9997
-83.1128
-83.2149
-83.1000
-83.1740
-83.3612
-83.3056
-83.3389
-83.3946
-83.2889
-83.3690
-83.2427
-83.1000
-83.7113
-83.6486

46.2
19.3
19.5
2.9
3.1
2.2
3.1
2.5
3.1
2.0
2.7
0.7
2.9
3.2
3.1
3.3
3.2
4.0
3.7
4.4
3.9
7.9
6.3
6.2

14.6
31.1
34.0
80.0
87.6
68.9
84.5
77.6
66.8
54.4
75.7
0.5
79.8
84.3
86.8
78.9
88.1
93.7
95.8
91.4
81.8
67.3
90.1
86.3

12.1
16.6
15.3
4.5
1.7
5.2
3.1
3.6
5.8
6.1
4.4
2.5
2.8
3.6
2.0
3.9
2.0
0.3
0.1
1.1
2.7
5.2
0.9
2.1

24.4
29.5
28.2
9.9
5.4
17.6
6.6
10.8
16.7
23.7
13.1
82.9
10.2
6.5
6.6
11.3
5.1
1.3
0.1
2.3
8.6
13.4
1.9
4.7

0.5
0.4
0.4
0.0
0.0
0.0
0.0
0.1
0.1
0.0
0.0
0.0
0.0
0.0
1.5
0.0
0.0
0.0
0.0
0.0
0.0
0.2
0.1
0.0

2.1
2.9
2.4
2.6
2.2
6.1
2.5
5.4
7.5
13.8
4.1
13.4
4.2
2.4
0.0
2.6
1.5
0.4
0.1
0.6
2.9
5.8
0.6
0.5

0.1
0.1
0.2
0.1
0.0
0.1
0.1
0.1
0.1
0.0
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.3
0.2
0.1
0.1
0.2
0.0
0.1

Mixed
Mixed
Mixed
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Forest
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture

169

REFERENCES

170

REFERENCES

Allan, J.D., P.B. McIntyre, S.D.P. Smith, B.S. Halpern, G.L. Boyer, A. Buchsbaum, G.A.
Burton, Jr., L.M. Campbell, W.L. Chadderton, J.J.H. Ciborowski, P.J. Doran, T. Eder,
D.M. Infante, L.B. Johnson, C.A. Joseph, A.L. Marino, A. Prusevich, J.G. Read, J.B.
Rose, E.S. Rutherford, S.P. Sowa, and A.D. Steinman. 2013. Joint analysis of stressors
and ecosystem services to enhance restoration effectiveness. Proceedings of the National
Academy of Sciences, v.110, no.1, p.372-377.
Allan, J., & Johnson, L. 1997. Catchment‐scale analysis of aquatic ecosystems. Freshwater
Biology
Baker, A. 2005. Land use and water quality. Encyclopedia of Hydrological Sciences., 37(1), 107111.
Barber, L. B., Murphy, S. F., Verplanck, P. L., Sandstrom, M. W., Taylor, H. E., & Furlong, E.
T. 2006. Chemical loading into surface water along a hydrological, biogeochemical, and
land use gradient: A holistic watershed approach. Environmental science &
technology, 40(2), 475-486.
Boutt, D.F., D.W. Hyndman, B.C. Pijanowski, and D.T. Long. 2001. Identifying potential land
use-derived solute sources to stream baseflow using ground water models and GIS.
Ground Water, v.39, no.1, p.24-34.
Buttle, J.M., and C.F. Labadia. 1999. Deicing salt accumulation and loss in highway snow banks.
Journal of Environmental Quality 28, no. 10: 96-99.
Clow, D.W., M.A. Mast, and D.H. Campbell. 1996. Controls on Surface Water Chemistry in the
Upper Merced River Basin, Yosemite National Park, California. Hydrological Processes
10, 727-746.
Crosbie, B. and P. Chow-Fraser. 1999. Canadian Journal of Fisheries and Aquatic Sciences, 56:
1781-1791, 10.1139/f99-109
D’Itri F.M. 1992. Deicing chemicals and the environment. Lewis Publishers.
Davis, S.N., D.O. Whittemore, and J. Fabryka-Martin, J. 1988. Uses of Chloride/Bromide Ratios
in Studies of Potable Water. Ground Water 36:338-350.
Deocampa, D.M. 2004. Hydrogeochemistry in the Ngorongoro Crater, Tanzania, and
implications for land use in a World Heritage Site. Appl. Geochemistry. 19, 755-767

171

Dodds, W. K., & R.M. Oakes 2006. Controls on nutrients across a prairie stream watershed: land
use and riparian cover effects. Environmental Management, 37(5), 634-646.
Drever, J.I. 1997. The Geochemistry of Natural Waters. Prentice Hall, New Jersey
Driscoll, C.T., R.M. Newton, C.P. Gupala, J.P. Baker and S.W. Christensen. 1991. Adirondack
Mountains. In: D.F. Charles (Editor), Acidic deposition and aquatic ecosystems: Regional
Case Studies. Verlag-Springer, New York.
Einheuser, M.D., A.P. Nejadhashemi, L. Wang, S.P. Sowa, S.A. Woznicki. 2013. Linking
biological integrity and watershed models to assess the impacts of historical land use and
climate changes on stream health. Environmental Management, v.51, no.6, p.1147-1163.
United States Environmental Protection Agency. (2014, February 18). The Phaseout of Methyl
Bromide. Retrieved: 06/23/2014. From the Environmental Protection Agency website:
http://www.epa.gov/ozone/mbr/
Evans, C.D., T.D. Davies, P.J. Wigington Jr., M. Tranter, and W.A. Kretser. 1996. Use of factor
analysis to investigate processes controlling the chemical composition of four streams in
the Adirondack Mountains, New York. Journal of Hydrology 185:297-316.
Eyre, B.D. and Pepperell, P. 1999. A spatially intensive approach to water quality monitoring in
the Rous River catchment, NSW, Australia. Journal of Environmental Management 56,
97-118.
Fetter, C.W. 1994. Applied Hydrogeology. Prentice Hall, New Jersey
Fitzpatrick, M., D.T. Long, and B. Pijanowski. 2007. Exploring the effects of urban and
agricultural land use on surface water chemistry, across a regional watershed, using
multivariate statistics. Applied Geochemistry 22:1825-1840.
Foos, A. 2003. Spatial distribution of road salt contamination of natural springs and seeps,
Cuyago Falls, Ohio, USA. Environmental Geology 44, 1:14-19.
Giri, S., A.P. Nejadhashemi, and S.A. Woznicki. 2012. Evaluation of targeting methods for
implementation of best management practices in the Saginaw River Watershed. Journal of
Environmental Management, v.103, no.1, p.24-40.
Gordona, L.J., C.M. Finlayson, and M. Falkenmarka. 2010. Managing water in agriculture for
food production and other ecosystem services. Agriculture Water Management 97,4:512519.
Grayson, R.B., C.J. Gippel, B.L. Finlayson, and B.T. Hart. 1997. Catchment-wide impacts on
water quality: the use of “snapshot” sampling during stable flow. Journal of Hydrology
199, 121-134.

172

Grimm NB, S.E. Gergel, W.H. McDowell, E.W. Boyer, C.L. Dent, P. Groffman, S.C. Hart, J.
Harvey, C. Johnston, E. Mayorga, M.E. McClain, and G. Pinay. 2003. Merging aquatic
and terrestrial perspectives of nutrient biogeochemistry. Oecologia 137: 485-501.
Gorsuch, R.L., 1974. Factor Analysis. W.B. Saunders Company, Philadelphia.
He, C., and T.E. Croley. 2006. Spatially modeling nonpoint source pollution loadings in the
Saginaw Bay Watersheds with the DLBRM.
Herczeg, A.L., H.J. Simpson, and E. Mazor. 1993. Transport of soluble salts in a large semiarid
basin: River Murray, Australia. Journal of Hydrology 144: 59-84.
Hildebrandt A, M. Guillamón, S. Lacorte, R. Tauler, and D. Barceló. 2008. Impact of pesticides
used in agriculture and vineyards to surface and groundwater quality (North Spain).
Water Research 42: 3315-3326
Hinton, M., S. Schiff, and M. English. 1993. Physical properties governing groundwater flow in
a glacial till catchment. Journal of Hydrology 142: 229-249
Hoaglund, J.R., J.J. Kolak, D.T. Long, and G.J. Larson. 2004. Analysis of modern and
Pleistocene hydrologic exchange between Saginaw Bay (Lake Huron) and the Saginaw
Lowlands area. Geol. Soc. Am. Bull., 116: 3–15.
Hoard, C.J., L.M. Fuller, and L.R. Fogarty. 2009. Analysis of water-quality trends for selected
stream in the Water Chemistry Monitoring Program, Michigan, 1998-2005. U.S.
Geological Survey Scientific Investigations Report 2009-5216, 60p.
Hofmann J, M.Venohr, and D.Behrendt. Opitz. 2010. Integrated water resources management in
central Asia: nutrient and heavy metal emissions and their relevance for the Kharaa River
Basin, Mongolia. Water Sci Technol. 62(2):353-363.
Houghton, D. 1838. Report of the State Geologist. Michigan Geological Survey. Lansing, MI, 3
Hwang, C. K., Cha, J. M., Kim, K. W., & Lee, H. K. 2001. Application of multivariate statistical
analysis and a geographic information system to trace element contamination in the
Chungnam Coal Mine area, Korea. Applied geochemistry, 16(11): 1455-1464.7
Johnson, L., C. Richards, G.E. Host, and J.W. Arthur. 1997. Landscape influences on water
chemistry in Midwestern stream ecosystems. Freshwater Biology 37:193-208.
Kannan, K., S.H. Yun, A. Ostaszewski, J.M. McCabe, D. Mackenzie-Taylor, and A.B. Taylor.
2008. Dioxin-Like Toxicity in the Saginaw River Watershed: Polychlorinated Dibenzo-pDioxins, Dibenzofurans, and Biphenyls in Sediments and Floodplain Soils from the
Saginaw and Shiawassee Rivers and Saginaw Bay, Michigan, USA. Archives of
Environmental Contamination and Toxicology, 8:9-19

173

Kaushal, S.S., P.M. Groffman, G.E. Likens, K.T. Belt, W.P. Stack, and V.R. Kelly. 2005.
Increased salinization of fresh water in the northeastern United States. Proceedings of the
National Academy of Sciences, 102:13517–13520.
Kolak, J.J., D.T. Long, J.M. Matty, G.J. Larson, D.F. Sibley, T.B. Councell. 1999. Groundwater, large-lake interactions in Saginaw Bay, Lake Huron: A geochemical and isotopic
approach. Geological Society of America Bulletin, 116:3-15.
Lenahan, M.J., K.L. Bristow, and P. de Caritat. 2011. Detecting induced correlations in
hydrochemistry. Chemical Geology v.284, no.1-2, p.182-192.
Long, D. N.E. Fegana, W.B. Lyons, M.E. Hines, P.G. Macumber, and A.M. Giblin. 1992.
Geochemistry of acid brines: Lake Tyrrell, Victoria, Australia. Chemical Geology 96:3352.
Long, D.T., T.P Wilson, M.J. Takacs, D.H. Rezabek. 1988. Stable-isotope geochemistry of
saline near-surface ground water: East-central Michigan Basin. Geological Society of
America Bulletin 100:1568-1577
Machavaram, M.V., D.O. Whittemore, M.E., Conrad, and N.L. Miller. 2006. Precipitation
induced stream flow: An event based chemical and isotopic study of a small stream in the
Greta Plains region of the USA. Journal of Hydrology, v.330, no.3-4, p.470-480.
Mandle, R. J., & Westjohn, D. B., 1989. Geohydrologic framework and ground-water flow in the
Michigan basin. In Regional Aquifer Systems of the United States: Aquifers of the
Midwestern Area. Papers Presented at 24 th Annual AWRA Conference and Symposium
November 6-11, 1988, Milwaukee, WI. AWRA Monograph Series (No. 13).
Mason, C.F., S.A. Norton, I.J. Fernandez, L.E. Katz. 1999. Deconstruction of the chemical
effects of road salt on stream water chemistry. Journal of Environmental Quality 28, no.
8: 59-60.
Meador, M.R. and R.M. Goldstein, 2003. Assessing water quality at large geographic scales:
relations among land use, water physicochemistry, riparian condition, and fish community
structure. Environmental Management 31:504-517.
Meissner, B.D., D.T. Long, and R.W. Lee. 1996. Selected geochemical characteristics of ground
water from the Saginaw Aquifer in the central Lower Peninsula of Michigan. U.S.
Geological Survey Water-Resources Investigations Report 93-4220, 19 p.
McGuire, J. T., E.W. Smith, D.T. Long, D.W. Hyndman, S.K. Haack, M.J. Klug, and M.A.
Velbel. 2000. Temporal variations in parameters reflecting terminal-electron-accepting
processes in an aquifer contaminated with waste fuel and chlorinated solvents. Chem.
Geol. 169: 471-485.

174

Michigan Department of Natural Resources, 2003. IFMAP/GAP land copver. In: MDNR (Ed.),
Lansing, MI
Panno, S.V., K.C. Hackley, H.H. Hwang, S.E. Greenberg, I.G. Krapac, S. Landsberger, and D.J.
O'Kelly. 2006. Characterization and identification of Na-Cl sources in ground water.
Ground Water 44:176-187
Phillips, R.A. 1988. Relationship between glacial geology and streamwater chemistry in an area
receiving acid deposition. Journal of Hydrology 10:263-273.
Pionke, H.B., W.J. Gburek, R.R. Schnabel, A.N. Sharpley, and G.F. Elwinger. 1999. Seasonal
flow, nutrient concentrations and loading patterns in stream flow draining an agricultural
hill-land watershed. Journal of Hydrology 220:62-73.
Public Sector Consultants, Inc. 2002. Targeting Environmental Restoration in the Saginaw
River/Bay Area of Concern (AOC): 2001 Remedial Action Plan Update. Report to the
Great Lakes Commission: Lansing, MI, 88 p. Accessed 19 September 2013 from:
http://www.pscinc.com/Portals/0/Publications/Saginaw_Bay/2002_RAP/RAPupdatereport
_2002.pdf
Puckett, L.J. and O.P. Bricker. 1992. Factors controlling the major ion chemistry of streams in
the Blue Ridge and Valley and Ridge physiographic provinces of Virginia and Maryland.
Hydrologic Processes, 6:79-98.
Richards, C., L. Johnson, and G. Host. 1993. Landscape influences on habitat, water chemistry,
and macroinvertebrate assemblages in Midwestern stream ecosystems. NRRI Technical
Report No. TR-93-109. 74 pp.
Riseng, C.M., M.J. Wiley, P.W. Seelbach, and R. J. Stevenson. 2010. An ecological assessment
of Great Lakes tributaries in the Michigan Peninsulas. Journal of Great Lakes Research
36,3:505-519
Rose, S. 1996. Temporal environmental isotopic variation within the Falling Creek (Georgia)
watershed: implications for contributions to streamflow. Journal of Hydrology 174:243261.
Schott, P.P. and J. Van der Wal. 1992. Human impact on regional groundwater composition
through intervention in natural flow patterns and changes in land use. Journal of
Hydrology 134:297-313.
Science Subgroup of the Great Lakes Regional Working Group. 2013. Great Lakes Restoration
Initiative: Adaptive Science-Based Framework for Great Lakes Restoration – Public
Comment Draft, 32 p. Accessed 19 September 2013 from:
http://greatlakesrestoration.us/pdfs/20130521-glri-adaptive-science-based-framework.pdf

175

Selzer, M.D. 2008. The Michigan Department of Environmental Quality Biennial Remedial
Action Plan Update for the Saginaw River/Bay Area of Concern. The Michigan
Department of Environmental Quality, Lansing, MI, 36 p. Accessed 19 September 2013
from:
http://www.glc.org/spac/pdf/rapupdates/Final%20Saginaw%20River%20Bay%20AOC%2
0RAP%20Update.pdf
Shanley, J.B., 1994. Effects of ion exchange on stream solute fluxes in a basin receiving highway
deicing salts. Journal of Environmental Quality 23:977-986.
Smart R., White C. C., Townend J. and Cresser M. S. 2001. A model for predicting chloride
concentrations in river water in a relatively unpolluted catchment in north-east
Scotland. Science of the Total Environment, 265, 131–141.
Susanna, T.Y. and W.C. Chen. 2002. Modeling the relationship between land use and surface
water quality. Journal of Environmental Management 66:377-393.
Tiffany, M.A., J.W. Winchester, and R.H. Loucks. 1969. Natural and pollution sources of iodine,
bromine, and chlorine in the Great Lakes. Journal Water Pollution Control Federation
41:1319-1329.
Tipper, E.T., A. Galy, and M. Bickle. 2008. Calcium and magnesium isotope systematics in
rivers draining the Himalaya-Tibetan-Plateau region: lithological or fractionation control?
Geochimica et Cosmochimica Acta 4,15:1057-1075.
United States Environmental Protection Agency. 2000. The Quality of Our Nations Waters: A
Summary of the National Water Quality Inventory. EPA841-S-00-001, United States
Environmental Protection Agency, Washington, D.C.
Vengosh, A. and R. Keren. 1996. Chemical modifications of groundwater contamination by
recharge of treated sewage effluent. Journal of Contaminant Hydrology 23:347-360.
Wahrer, M.A., D.T.Long, and R.W. Lee. 1996. Selected geochemical characteristics of ground
water from the Glaciofluvial aquifer in the central Lower Peninsula of Michigan. U.S.
Geological Survey Water-Resources Investigations Report 94-4017, 21p.
Wayland, K.G., D.T. Long, D.W. Hyndman, B.C. Pijanowski, S.M. Woodhams, and S.K. Haack.
2003. Identifying relationships between baseflow geochemistry and land use with
synoptic sampling and R-mode factor analysis. Journal of Environmental Quality 32:180190.
Wilkinson, L. (2008). Systat 12 [computer software]. Systat Software Inc., Chicago, IL
Wilson, T.P., and D.T. Long. 1985. Geochemistry and isotope chemistry of Michigan Basin
brines: Devonian formations. Applied Geochemistry 8:81-100.

176

Wilson, T.P. 1989. Origin and Geochemical Evolution of the Michigan Basin Brine. Ph.D.
Dissertation, Michigan State University, East Lansing, MI.
Wood, W.W. 1970. Chemical quality of Michigan streams. U.S. Geological Survey Circular
634, 21 pp.
Zimmer, M.A., S.W. Bailey, K.J. McGuire, T.D. Bullen. 2012. Fine scale variations of surface
water chemistry in an ephemeral to perennial drainage network. Hydrological Processes,
14 p. doi: 10.1002/hyp.9449

177