‘IWIHIIIW“IRWIN“lfllll‘WlHlHHl)HIIHIHM

    

134
348
THS

  
 
 

m Itiiii’l‘iflliiiIflfliiiiliuml‘iinifil

3 1293 02058 6230

This is to certify that the

thesis entitled

Petrologic Constraints of Possible Reactions of Saline
Formation Water With Sediments at Equilibrium With
Fresh Pore Water, Saginaw Bay

presented by

Jennifer Ann Wilson

has been accepted towards fulfillment
of the requirements for

M.S. , Geological Science
degree in

 

 

 

11- -
Date 11 99

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

LIBRARY

Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11/00 animus-p.14

 

PETROLOGIC CONSTRAINTS OF POSSIBLE REACTIONS OF SALINE
FORMATION WATER WITH SEDIMENT S AT EQUILIBRIUM WITH FRESH PORE
WATER, SAGIN AW BAY

By

Jennifer Ann Wilson

A TI-HESIS

Submitted to Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Geological Sciences

1 999

ABSTRACT
PETROLOGIC CONSTRAINTS OF POSSIBLE REACTIONS OF SALINE
FORMATION WATER WITH SEDIMENTS AT EQUILIBRIUM WITH FRESH PORE
WATER, SAGINAW BAY
By
Jennifer A. Wilson

Brines seeping into Saginaw Bay are among the world’s most concentrated,
and mix with the fresh water of the bay.

This study examines four common cations: Na", K‘, Mg”, and Ca2+, as well as
HCO3'. Their concentrations have been measured in the pore waters of the bay’s
sediments. Two different cation-exchange capacity experiments were performed on
the sediment, as well as x-ray diffraction to identify the clay and carbonate
mineralogy of the sediment.

Carbonate dissolution is a major control on pore-water chemistry, based on
the ratio of Ca2+ + Mg2+ to HCOg‘, and therefore is probably a major control on
concentrations of Ca2+ and Mg“. However, there are some mixing effects for these
cations, since these cations do correlate well with Na” when Na” concentrations rise
above 70 mg/L. Na+ is controlled mostly by simple mixing, as it correlated well with
depth. K+ seems to be controlled primarily by sediment diagenesis, because the
concentration in the pore water did not change as a function of depth, as did the Na”,

but the source or sink of K" is not known.

ACKNOWLEDGMENTS

To paraphrase Mary Jo Pehl of MST3K: “I have mixed feelings about this thesis
project. On the one hand, its odd and vigorously depressing tone makes me want to step in
front of a bus. Then again, its sheer length, mass, heft, volume, and viscosity overwhelms
me, and it sits on my head, bouncing lightly up and down.” Of course, she was referring to
the Japanese Camera flying turtle movies, but I feel that it is appropriate nonetheless. This
is to everyone who assisted me during this harrowing time.

I’d like to thank my advisor, Duncan Sibley, for all the help he was able to give
me. I’d also like to thank Dr. Sharon Anderson, formerly of MSU’s Department of Crop
and Soil Sciences, for allowing me to use her lab, and for answering my endless questions
about CEC. I’d like to thank Dr. David Long and John Kolak for being patient with my
frequent requests for more water chemistry data. I’d especially like to thank Drs. Michael
Velbel and Danita Brandt, for providing me with the means to come back to E. Lansing to
finish the labwork on this project, and for listening to my constant griping.

I would also like to thank all my friends and family, but especially my friend Dave
Erazmus. Thanks for always paying for the telephone calls, encouraging me to call often,

and helping me to preserve my sanity.

iii

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

CHAPTER 1: INTRODUCTION
Purpose
Glacial History
Michigan Basin
Chemistry of Basin Water
Chemistry of Bay Water
Carbonate-dominated Systems

METHODS

RESULTS

DISCUSSION

CONCLUSION

REFERENCES CITED

iv

page v

page vi

page 1
page 1
page 3
page 4
page 6
page 8
page 17
page 20
page 50
page 59

page 61

LIST OF TABLES

Table 1: Major ion concentrations in the Great Lakes page 7
Table 2: r2 values for data versus depth page 23
Table 3: r2 values for total-dissolved solids versus ion-activity products page 25
Table 4: :2 values for calcite/dolomite data page 27
Table 5: Calcite/dolomite ratios page 32
Table 6: r2 values for cations versus N a+ concentrations in porewater page 34

Table 7: r2 values for cations versus Na+concentrations in porewater for SBZ4 page 39

Table 8: r2 for K’r concentrations in poewater versus Ca2+ and Mg2+ concentrations in

porewater page 43
Table 9: Concentrations of Ca“, Mg”, and alkalinity page 48
Table 10: r2 for Ca2+ and Mg“ water data page 49

LIST OF FIGURES

Figure 1: Saginaw Bay and sites of core removal page 2
Figure 2: X-ray diffractogram of SB21-1 Mg-saturated page 21
Figure 3: X-ray diffractogram of SB 1-1 Mg-saturated showing chlorite(004) peak and
kaolinite(002) peak page 22
Figure 4: Total CEC versus depth page 24
Figure 5: Na+ CBC versus Na+ concentration in porewater page 26
Figure 6: Total—dissolved solids versus ion-activity product for calcite page 28
Figure 7: Total—dissolved solids versus ion-activity products for dolomite-c and
dolomite-d page 29
Figure 8: Calcite/dolomite versus depth page 30
Figure 9: Calcite/dolomite versus depth without samples SB24-10, SB 18-4, SB1-3, and
SB 19-7 page 3 1
Figure 10: Calcite/dolomite versus depth for core SB24 page 35

Figure 11: Calcite/dolomite versus depth for core SB24 without sample SB24-10 page 36

Figure 12: Na+ concentration in the porewater versus depth page 37

Figure 13: N a+ concentration >70 mg/L versus depth page 38

Figure 14: Ca2+, Mg2+ and Cay/Mg2+ in porewater for for core SB24 versus

depth page 40

Figure 15: Concentration of K+ in core SB24 versus depth page 41

Figure 16: Concentration of cations in core SB24 versus depth without SBZ4-1 page 42

Figure 17: Concentration of Na+ versus depth for core 8824 page 44

vi

Figure 18: Concentration of Ca2+ versus concentration of Mg!2+ page 45
Figure 19: Concentration of Ca2+ versus concentration of Mg2+ in porewater with
concentrations in bay water added page 46

Figure 20: Concentration of Ca2+ versus Mg2+ for shallow poewater samples page 47

vii

INTRODUCTION

Purpose

Saginaw Bay presents an unparalleled opportunity to study sediment-water
interactions. Hypersaline brines from the Michigan basin are mixing with the fresh water
of the bay in the pores of recent lake sediments. Effects of simple mixing on the porewater
should be obvious, since the compositions of the two waters being mixed are so different;
therefore, effects of diagenesis on water chemistry should be easily distinguished from
these water mixing effects. The purpose of this research is to provide a mineralogic
constraint for the mineral-water interactions that influence pore-water chemistry in
sediments taken from Saginaw Bay. Pore-fluid chemistry was measured from gravity cores
of sediments from Saginaw Bay (Figure l). Combinations of mineral dissolution and
precipitation, and ion exchange reactions and simple mixing may result in a given pore-
water chemistry. Analysis of Ca2+ + Mg2+lHC03' in pore waters and XRD analysis of
sediments indicate carbonate dissolution within the sediments has a major impact on pore

fluid chemistry.

Glacial History

Saginaw Bay is a southwest extension of Lake Huron, located on the eastern side
of the lower peninsula of Michigan. Lake Huron is predominantly underlain by gently
dipping rocks of Paleozoic age. The Lake Huron basin was formed by glacial scouring of

relatively weak bedrock in pre-glacial valley systems (Larson, personal communication).

 

 

O
.8893 8311a
.8815

8816

S Li

 

 

 

 

Figure l. Saginaw Bay and sites of core removal

The Huronian basin is mostly underlain by easily erodible Devonian shales and limestones.
Saginaw Bay is primarily underlain by Late Mississippian age limestones (Michigan Fm.
and Bayport Limestone), with some Pennsylvanian sandstones and shales at the
southwestern end.

The glacial history of the Great Lakes basins is complex, with many different
intervals of ice advance and retreat affecting water levels in the basins, as well as direction
of drainage, sediment deposition, and scouring of the basins. The final water regression to

the present Lake Huron occurred around 3 ka (Eschman and Karrow 1985).

Michigan Basin

The Michigan basin is an intracratonic basin underlying the lower peninsula of
Michigan, parts of the upper peninsula of Michigan, Ohio, Illinois, Wisconsin, and Ontario.
During the Paleozoic, it was often occupied by marine waters. Of particular interest are the
deposits of Silurian, Devonian, and Mississippian ages. During these times the sea
occupying the Michigan basin was shallow, occasionally isolated from the open ocean.
Consequently, the rocks are dominated by chemically precipitated rocks, such as dolomite,
and also economically important deposits of halite and gypsum. It is from these formations
that brines are produced (Dorr and Eschman 1970, Wilson and Long 1993).

Rocks of Pennsylvanian age mark the end of the dominance of shallow seas in the
depositional history of Michigan. During this time, the land was emergent and stream
deposits dominate. The only rocks of more recent age occurring in the basin are some
Jurassic age red~beds. All these rocks are overlain by Pleistocene glacial deposits,

including Wisconsin age end moraines, outwash plains and glacial lake deposits.

Chemistry of basin water

The Saginaw Bay area has long been known to harbor saline, near-surface waters.
Tellam (1995), Weaver et al. (1995) and Kolak et al.(1999) found saline basin waters can
mix with meteoric water, and near-surface waters can have a complex history of multiple
water-mixing episodes. In their study of cross-formational fluid flow in the Michigan
basin, Weaver et al. (1995) proposed a two-stage mixing model based on trends in C1” and
stable isotopes (8180 and 82H). The first stage involves mixing a saline and isotopically
heavy (high 8180 and 82H) water with a dilute and isotopically light water. In their model,
45% brine mixed with 55% glacially recharged groundwater. This fluid then mixed with a
water of intermediate salinity and a more modern isotopic signature due to recent recharge
with modern meteoric water. Weaver et al. demonstrate that groundwater in shallow basin
environments is dynamic, with different episodes of fluid emplacement and mixing causing
modern chemistry and isotopic signatures.

Kolak et al. (1999) found anomalously low 5180 (between -10.00 %0 and -
16.00%0, versus between -8.00 %o and -10.00 %o for modern meteoric water) values in
water from cores taken from the Saginaw Lowland Area, suggesting the presence of waters
that were recharged during the last glaciation. These isotopically light waters tend to occur
in conjunction with high salinities. Values of 8180 taken from cores producing pronounced
salinity with increasing depth tend to have isotopic signatures that become lighter with
depth, indicating that the saline water mixed with glacial recharge water during the

Pleistocene.

The source of salinity in the shallow Michigan basin is probably formation water,
rather than from the dissolution of halite. The formation water of the Michigan basin is the
result of evapo-concentration of seawater (Wilson and Long 1993). Wilson and Long
(1985) determined formation waters are the probable source of saline waters in the
Michigan basin based on the Cl/Br ratio. Since bromine is preferentially partitioned into
the brine as halite crystallizes, halite has a Cl/Br ratio exceeding 3,000. Wilson and Long
(1985) found that the Cl/Br ratio of formation water is approximately 300, about that of
seawater. In the Saginaw Bay area, brines have been found to have Cl/Br ratios ranging
from 400 to 600 (Kolak et al. 1999). This disparity may be due to incomplete equilibration
of the partitioning of bromide between solid and liquid phases in a halite-dissolution
dominated system (Wilson and Long 1985). In the Saginaw Bay area, these near-surface
brines are very concentrated, even near the surface, with dissolved-solids commonly
exceeding 1,000 mg/L (Meissner 1993, Westjohn and Weaver 1996, Wilson and Long
1993). The major element composition of saline formation water in the Michigan basin has
been linked to evapo-concentration of sea water plus water-rock interactions involving
illitization and dolomitization (Wilson and Long 1986). Wahrer (1993) suggests the saline
waters are in equilibrium with respect to both illite and kaolinite. Many different chemical
reactions can produce the ions in the saline pore waters, and chemical modeling suggests
potential mineral equilibrium with halite, anhydrite, celestite, calcite, and dolomite (Wilson
and Long 1986). Meissner (1993) states that waters with high dissolved solids are Cl-SO4
dominated.

Studies in the Great Lakes area have documented salinization of near-surface

groundwater in the area, including in the Saginaw Bay region (Desaulniers et al. 1981,

Weaver et al. 1985, Long er al. 1988, Badalamenti 1992), evidently from upward transport
of brine-derived formation waters in the Michigan basin. Lacustrine sediments and clay-
rich lodgement tills seem to trap saline waters in the Saginaw Bay area, inhibiting
groundwater discharge in the Saginaw lowland area (W estjohn and Weaver 1996).
Recharge is also inhibited due to lacustrine sediments (Rheaume 1991).

Kolak et al. (1999) have determined that saline formation water is actively
introducing solutes into the sediments of Saginaw Bay. Many of the cores from which
sediment for this study were taken have been determined by Kolak et al. (1999) to have

pronounced chloride gradients with depth.

Chemistry of bay water

Lake Huron receives most of its inflow from Lakes Michigan and Superior.
Chloride, sulfate, and total dissolved solids (TDS) have increased with increasing
industrialization, with most of the water quality degradation focused on Saginaw Bay
(Beeton 1970). Weiler and Chawla (1969) found the major ion concentrations in Lake
Huron to be the following, in mg/L: Ca2+ = 28.1, Mg2+ = 6.7, Na+ = 3.2, K” = 0.84, or =
6.3, TDS = 118, alkalinity = 78.6, and pH = 8.0 (Table l). Saginaw Bay data are more
similar to the data for Lake Michigan (Kramer 1964), except that the concentrations of Na",
K+, and C1' are elevated. K+ concentrations are similar to those which Weiler and Chawla
(1969) found for Lakes Erie and Ontario. N 3+ and Cl‘ concentrations in Saginaw Bay are
lower than the values for Lakes Erie and Ontario, but higher than the values for Lakes
Michigan, Huron and Superior. The solute values that are higher than those for Lake

Huron as a whole are probably due to industrialization and the concurrent environmental

degradation in the area of the bay. Lake Superior has low concentrations of solutes

compared to the other Great Lakes, including Saginaw Bay. Total dissolved solids are only

52 mg/L, and CI' = 1.3 mg/L(Weiler and Chawla 1969).

 

 

 

 

 

 

 

 

 

Major Ion Concentrations in the Great Lakes

Ca2+ 1912+ Na+ K+ Cl- TDS alkalinity pH
Lake Huron 28.1 6.7 3.2 0.84 6.3 118 78.6 8.0
Lake Michigan 32 10 3.4 0.90 6.2 150 113 8.0
Lake Superior 13.2 2.7 1.3 0.54 1.3 52 51.8 7.8
Lake Ontario 40.3 8.1 12.6 1.35 27.5 194 92.8 7.9
Lake Erie 37.4 8.3 11.5 1.23 24.6 198 92.4 8.1
Saginaw Bay 32.1 9.2 9.3 1.44 13.4 112 8.2

 

 

 

 

 

 

 

 

 

 

Figure l. Saginaw Bay and sites of core removal

Kramer (1966) found that the Great Lakes fit a model involving the equilibrium of
calcite, dolomite, apatite, kaolinite, gibbsite, Na- and K-feldspars at 5° C., 1 atm. total
pressure and Pcm = 3.5 x 104 atm. The water of Saginaw Bay is saturated with respect to
calcite in that part of the bay closest to the main body of the lake, and supersaturated with
respect to calcite in the inner bay (Effler 1984). There are seasonal effects; calcite
saturation is higher in the summer, especially in the inner bay (Effler 1984). Weiler and
Chawla (1969) also point out that Saginaw Bay is a source of higher ion concentrations for

Lake Huron as a whole, with chloride concentrations, for instance, twice the normal value

for the lake.

Carbonate-dominated systems

The sediments at the bottom of Saginaw Bay, including both the overcompacted
glacial till and the more recent lake sediments, contain many minerals, including quartz,
feldspars, hornblende, micas and clays, and heavy minerals like magnetite, though pyrite is
mostly absent (Wood 1958). Sediments also contain detrital calcite and dolomite in
quantities ranging up to several weight percent (Wood 1964, Robbins 1986). These
minerals are predominantly in the fine-silt size range (Robbins 1986). Because calcite and
dolomite are more reactive than most silicates, it is likely that these minerals play a major
role in determining the chemistry of the water found in the sediment.

Models of stoichiometric dissolution provide a means of determining the origin of
a solution when that solution is the product of dilute waters in contact with soluble
minerals.

If the sum of the molar concentrations of Ca2+ and Mg2+ is plotted versus HCO3',
the slope of the least-squares best fit line can yield clues about diagenetic processes
operating in the sediment. If the primary control is COz—driven congruent or incongruent

dissolution of calcite and dolomite, the slope will be ~O.5, that is
(1) CaC03 + CaMg(C03)2 + 3C0; + 3H20 —> 2Ca2+ + Mg“ + 6HC03'
or

(2) CaMg(C03)2 + (:02 + H20 —> CaC03 + Mg2+ + 2HC03’

The ratio of Ca2+ + Mg2+/I-IC03' will be 1:2.

Some literature was reviewed in an attempt to constrain the reactions likely in
carbonate mineral dominated rocks and sediment. In their discussion of diagenesis driven
by interaction of crude oil with sediments in a sand and gravel aquifer in Bemidji,
Minnesota, Bennett et al. (1993) provide a thorough documentation of possible reactions
involving carbonate minerals and their effects on water chemistry. The study area is on the
Bagley Outwash plain, consisting of moderately calcareous, silty sand occurring as glacial
outwash and morainal deposits, overlying a clayey till. The site was contaminated by oil
from a burst pipeline. Despite extensive remediation at the site, there remains an oil body
floating on the water table, plus oily residue which coats soil particles in the unsaturated
zone.

Bennett et al. (1993) studied the groundwater at Bemidji, starting with native
groundwater, which is a dilute Ca—Mg-HCO3' water with a Ca/Mg ratio of 2.2:1 (zone I). It
then flows under the area contaminated by oil spray from the pipeline rupture (zone 1]).
They monitor the water as it flows under the oil body (zone I11), into the transition zone
(zone IV), where 02 is limited, and dilution, mixing and biogeochemical reactions reduce
contaminant concentrations. Finally, this water mixes with native groundwater (zone V).

According to Bennett et al.(1993), dissolution of Ca-bearing silicates by carbonic

acid, also produce HC03' as in this reaction of anorthite

(3) CaAIZSizog + 21120 + co2 + H+ —> Alzsi205(OH)4 + Ca2+ + Hco;

In this case, the ratio of Ca2+ to HCO3‘ is ~l. If there were both dissolution of carbonates
and significant dissolution of Ca-bearing silicates, the Ca2+ + Mg2+lHC03' would be >0.5.

However, equation 3 should be written as:

(4) CflAIzSIzOg + 3H20 + 2C02 -9 AIzSIzOs(OH)4 + C82+ + 2HCO3

This yields a Ca2+ + Mg2+IHC03' ratio of 1:2, making it impossible to tell the difference
between carbonate and silicate weathering using this ratio. The source of the extra H+ in
eqn. 3 is probably from organic acids from the degrading oil body. Bemer and Bemer
(1996; note p. 216-217) that dissociation of the carboxyl groups of organic acids, such as

humic or fulvic acid, can consume bicarbonate according to the following equation:

(5) (R-COOH) + HCO3' —) (R-COO)' + H20 + C02

This would cause the ratio of Ca2+ + Mg2+IHCO3' to become greater than 0.5. Bemer and
Bemer (1996) note that in rivers with high organic acid contents, HCO3' tends to be

- consumed, neutralizing the H+ produced by the dissociation of the carboxyl groups of the
dissociating acids. In rivers with a preponderance of organic matter over inorganic matter,
the sum of the charge on major organic cations will be greater than the charge on the major
inorganic cations. As HCOg' is consumed by reaction with organic acids, charge is
maintained by the dissociation products of the organic acids (Bemer and Bemer 1996). As
HCO3' is added as these rivers flow downstream, pH tends to increase, due to the

consumption of PF.

10

Bemer and Berner (1996), demonstrate that in reactions involving carbonate

dissolution with sulfuric acid, the ratio of Ca2+ + Mg2+/anion would be 2:3

(6) CaMg(C03)2 + st04 —> Co“ + Mg2+ + so.” + 2HC03'

The addition of sulfuric acid to a system could perturb the Ca2+ + Mg2+/HCO3' ratio, due to
both enhanced cation release due to acid weathering and consumption of HCOg' by titration
by H+ to carbonic acid (HzCog). In the situation in equation 6, the Ca2+ + Mg2+l HCOg’ is
1:], compared to Ca2+ + Mg2+/HC03' = 1:2 in other systems.

If there is significant dissolution of SO42' bearing evaporites (gypsum

(CaSO4-2H20) and anhydrite (CaSO4), primarily), Ca2+/SO42' + HCO3’ = 2:3

(7) CaCO3 + Caso. + (:02 —> 2Ca2+ + so} + 2HC03'

In situations where Ca2+ + Mg2+IHCO3’ <0.5, either Ca2+ and Mg2+ are removed,
or an excess of HCOg' is produced (Bennett et al. 1993). If an excess of HCO3' is produced
by oxidation of organic matter, acidity should be produced. This acidity, however, would
tend to consume HCO3', having no effect on the ratio of C212“ + Mg2+lHCO3I If excess
HCOg’ is produced by continuous replenishment of CD; as the gas is depleted by solution
of carbonates, then pH should increase (Langmuir 1971).

A potentially important way to remove Ca2+ and Mg2+ from the water is by ion

exchange (Hanor 1982, Bennett et al. 1993). If this occurs, then monovalent cations will

11

be displaced from clays. Thus, for this mechanism, any depletion in the divalent cations

must be matched by a concurrent increase in the concentrations of Na+ and K‘”, that is

(8) Nag-clay + Ca2+ ——> Ca-clay + 2Na+

Since H+ is also exchangeable (April at al. 1986, Sequiera 1991), concentration of this
cation may also increase due to exchange with Ca2+ and Mg“, with a concomitant lowering
of pH. The ratio of Ca2+ + Mg2+IHC03' would then become <0.5.

If an imbalance of HC03' to Ca2+ and Mg“ is due to an excess of the anion, then it

may be due to a redox reaction represented by the following half-reaction

(9) (CH) + 3H20 -—) HCO3' + 6H“ + Se'

where (CH) represents organic matter being oxidized. When 0 is the electron acceptor,
C02 is produced. Thus carbonic acid is produced, raising the acidity and leading to the
dissolution of carbonate minerals. When SO42" or N 03' is the electron acceptor, however,

HC03' is produced without a concurrent increase in acidity, as in these equations:

(10) 2(CH20) + 2H20 + 8042' + 2H+ -> 2HC03’ + H28 + 2H20 + 2H+

(11) 2(CH20) + 3H20 + NO3' + H+ —-) 2HC03' + NH4+ + 2H20 + H+

In this situation, the water chemistry is not controlled by carbonate equilibria.

12

When organic material is oxidized in freshwater environments, the main electron
acceptors are Fe“, Mn“, organic compounds, nitrate, sulfate, and C02 (Jones et al. 1984).
Lovley et al. (1989) studied iron reduction at Bemidji. Fe3+ is an important electron
receptor. When organic material is oxidized in the presence of an oxidized Fe mineral, the

following reaction takes place

(12) (CH) + 5Fe(OH)3 —> HCO3’ + 5Fe2+ + 9OH' + 3H20

This would result in an increase in pH.

Bennett et al. (1993) propose that the reaction that would produce an excess of
HCO3' , compared to the ratio of Caz++ Mg2+/I-ICO3’ = 0.5 expected for simple carbonate
dissolution, while buffering pH, would be a combination of the above reaction with calcite

precipitation, that is

(13) 0.2(CH) + Fe(OH)3 + 1.8Ca2“ + 1.6HC03' —> 1.8 CaC03 + 1=o2+ + 2.41120

This reaction does not change pH. It consumes both Ca2+ and HC03', but at a rate that is
different from that formed by dissolution of carbonate or Ca—silicates (in this case Ca2+ is
consumed faster than bicarbonate, which would yield a ratio <0.5).

The preceding reactions were also proposed to explain the chemistry of the Ca-
Mg-HC03' water type in the Cambrian-Ordovician aquifer system in the northern midwest
(Siegel 1989). The rock units from this study area are separated into five aquifers for the

purpose of this study. Aquifer 1 (the Mount Simon aquifer) is the deepest, and consists

l3

primarily of the Mount Simon Sandstone, which is a coarse-grained conglomeratic quartz
arenite. Aquifer 2 (the Ironton-Galesville aquifer), consists of the Ironton and Galesville
Sandstones, and in Wisconsin and Iowa, includes the Wonewoc Formation. These
stratigraphic units consist primarily of calcareous to dolomitic quartz arenite. The third
aquifer varies considerably in its composition over this study area, but consists mostly of
the Jordan Sandstone, the Prairie du Chien Group, and the St. Peter Sandstone. The Jordan
Sandstone is silica-cemented quartz arenite, grading into a dolomitic sandstone, and then
into sandy dolomite and dolomite. The Prairie du Chien Group is fossiliferous dolomite
and limestone interbedded with thin layers of sandstone, siltstone and shale. The St. Peter
Sandstone is a friable quartz arenite. Aquifer 4 is dominantly dolomites, limestones and
shales of Silurian to Devonian age hydraulically connected by fractures. Aquifer 5 consists
variously of Holocene surficial deposits, Pleistocene glacial deposits, and/or Cretaceous
shales and lignitic sandstones. The glacial drift is calcareous, due to its origin in Paleozoic
limestones and dolomites.

Siegel (1989) found that the slope of the plot of Ca2+ + Mg2+IHCO3' for aquifers 1,
3, and 5 was = 0.55, and that the water chemistry was therefore controlled by dissolution of
carbonate minerals. Aquifers 2 and 4 were not well studied, and there were few data for
these aquifers. Siegel (1989) also proposed the coupled dissolution of carbonates and
reduction of iron minerals as an explanation for a constant pH. He also pointed to
exchange of divalent for monovalent cations as a potential modifier of water chemistry,
though he thought it unlikely to be a dominant mechanism in his study area. In his study
area, clays having high CEC are not abundant. On a trilinear plot of the water chemistry in,

the data trend towards the Na apex of the plot. Siegel (1989) states that if CEC were a

14

dominant mechanism, then the data would trend parallel to the Ca—Na axis. Additionally,
Siegel (1989) states that the covariance of the molality of Na to the molality of (Ca + Mg)
in his data has a low r2 (r2 = 0.3) and a slope = +2.1. If CEC were a dominant mechanism,
then the r2 should be closer to unity, and the slope should approximate -2. However, while
the average of the aquifers studied fits the model of carbonate-dissolution dominated water
chemistry, when analyzed separately, individual aquifers do not fit the model as well.
Siegle’s Aquifer 5 fits the model best, with Ca2+ + Mg2+/I-ICO3' = 0.53, Ca2+lMg2+ = 3.38.
For this aquifer, Caz‘ilMg2+ is not as close to the ideal 2:1 ratio, as shown in Equation 1, as
this ratio in Aquifers 1 and 3. The excess of Ca2+ over Mg“ may reflect a preponderance
of calcite over dolomite in the sediment, a much higher rate of calcite dissolution, or non-
stoichiometric dissolution of dolomite. Aquifer 3 yielded Ca2+ + Mg2+II-ICO3' = 0.41 ,
Ca2+lMg2+ = 2.66. Aquifer 1 yielded the worst Ca2+ + Mg2+/I-ICO3' ratio (ratio = 0.32),
though Caz‘ilMg2+ = 2.37. For these aquifers, Siegel (1989) proposed cation exchange to
remove Ca2+ and Mg2+, releasing Na“, and causing the Ca2+ + Mg2+lHC03' to be less than
0.5.

Langmuir (1971) evaluated ground waters near State College, Pennsylvania. The
rocks underlying the area are Ordovician and Cambrian sedimentary rocks that are part of a
regional anticlinorium. Nittany Mountain is a synclinal ridge within this anticlinorium.
Most of the aquifers in the area are limestone and dolomite units. Langmuir studied both
wells and springs in the carbonate rocks. Wells are mostly dolomite, and springs mostly
issue from limestone. 0f 29 springs, 22 issue from limestone, and 7 from dolomite. 0f 29

wells, 20 are drilled in dolomite, and only 9 are in limestone.

15

Langmuir (1971), divides the springs into two categories. Group I springs issue
from near the foot of mountain slopes. These springs have a subsurface residence time
often exceeding 2-6 days. Discharge from these springs varies considerably depending on
rainfall. Group 2 springs are located downvalley, and act as regional groundwater
discharge points. These springs have subsurface residence times measured in months.
Discharge from these springs is more stable throughout the year than those of Group 1.
Group 1 springs tend to be lower in specific conductance, pH, and the saturation indices of
both calcite and dolomite than Group 2 springs.

Most of the wells in the study area tend to have a lag time from 5-10 days, with the
exception of some wells in upland areas. The aquifer units (the Gatesburg and Stonehenge
Formations) in these areas are under a hundred or more feet of residual soil. Consequently,
water level changes in these wells are gradual and show little response to rainfall. Many of
the wells and springs in the area are vulnerable to pollution from sources like road salt,
sewage plant discharges, open dumping, septic tank discharge, and contact with animal
wastes and inorganic fertilizers. Langmuir (1971) states that the important ions in
unpolluted carbonate groundwaters are Ca“, Mg“, and HC03'. N a+, Cl', 8042' or NO;'
concentrations in excess of 10 ppm have been introduced by pollution.

In general, Langmuir (1971), explained his data in terms of carbonate dissolution
controlled water chemistry. Langmuir did have 8042’ data for his wells, but the ratio of
Ca2+ to 8042' + HCOg' did not yield a ratio of 2:3. For the waters from the springs, Ca2+ +
MgZVHCog' = 0.59 and Caz+lMg2+ = 3.75. Averaged well waters yielded the same
cation/anion ratio, and Caz‘VMg2+ = 1.3. His data from wells in dolomite yielded Ca2+ +

Mgzvlicog' = 0.57, and Calm/1g“ = 0.95. The variance in these data is best explained by

16

the fact that most of the springs are in limestone and the wells mostly in dolomite.
Langmuir (1971) reported a range of Caz‘VMg2+ for the springs ranging from 0.8 to 11, and
a range of Ca2+lMg2+ in the wells from 0.6 to 3.9. The ratios that are less than 1 are
explained as probably due to incongruent dissolution of dolomite. The deviation of the
Ca2+ + Mg2+/HC03' from a ratio of 0.5 may be due to degassing, decreasing the amount of
bicarbonate in the water.

Overall, the carbonate-dissolution model seems to best fit the data of Bennett et al.
(1993). For the oxygenated spray zone (Zone 11) in the oil-contaminated aquifer in their
study, Ca2+ + MgZVI-ICOg' = 0.52 and Caz‘ilMg2+ = 2.26. They suggest that this is due to

carbonic acid evolved from mineralization of the crude oil.

METHODS

Sediment samples were collected by gravity core during the summer of 1994 from
aboard the research vessel Laurentian. The cores were collected using a Benthos® gravity-
corer loaded with polycarbonate core liners. The core was then sliced into varying numbers
of sections (ranging from 6 to 24), each of approximately one-foot length, the number of
sections dependent on the length of core. Each section was then squeezed to remove pore
waters from the sample, and the sediment was bagged and refrigerated at 4°C.

For clay-mineral x-ray analysis, 35 samples were oxidized with hydrogen peroxide
to remove organic material (Kunze and Dixon 1987). Carbonates were removed from these

samples with a Na-acetate buffer adjusted to pH 5 with acetic acid (Kunze and Dixon

17

1987). Samples were placed in settling tubes to remove the <2wn clay fraction (Whittig
and Allardice 1987). Clays were then prepared for x-ray diffraction by saturating them with
K and Mg (Whittig and Allardice 1987). Slides were then prepared by dropping a clay
suspension on glass slides and allowing them to dry. Slides were prepared for Mg-25°C,
Mg-G (glycol), K-25°C, and K-550°C. The Mg-G slides were prepared by placement in a
glycolation chamber.

When it was found that these samples have insufficient organic material to
interfere with x-ray analysis, the remainder of the samples were prepared, without organic
removal, by a method based on Drever (1973). For this preparation, four aliquots of the
<2um clay fraction were placed in a Millipore vacuum filter apparatus. Two were treated
with a MgClz solution, one with a KC] solution, and the last left untreated. All were rinsed
with distilled water. The filter was removed from the apparatus, placed clay-side down on
a glass slide, and pressed onto the slide using a small glass sample tube as a roller. One of
the MgClz treated samples was then placed into a vacuum chamber with an ethylene glycol
solution. X-ray diffractograms were also made to determine relative abundances of calcite
and dolomite in the sediments of the bay. Four samples were also taken from near the
mouth of the Saginaw River. Untreated samples were used to determine calcite and
dolomite content. The calcite/dolomite ratio was then determined by drawing a line across
the bottom of both the calcite peak at 29.5 9 20 and the dolomite peak at 31° 20 and
counting the number of mm. to the tops of the peaks, then dividing the number for calcite +
dolomite into the number for calcite. Each sample was run three times, with a precision of

0.08.

18

Samples for both clay analysis (52 samples) and calcite-dolomite analysis (27 core
samples and 4 river sediment samples) were x-rayed in a Rigaku Geigerflex with Cu Ka
radiation, 8 Ni filter, and a theta-compensating slit. The scan length for clay analysis was
from 2° to 32° 20, with a step size of .02° 20, and a 1 sec step’l. Analysis for calcite-
dolomite was from 265° to 34° 20, with a step size of .02° 20, and a 1 sec step'l. Analysis
was also run for pyrite from 54° to 575° 20, with a step size of .02° 20, and a 1 sec step".

Two CEC analyses were done. Total CEC was done according to the magnesium
saturation method (Bain and Smith 1994). The sample was saturated with MgClz, and the
adsorbed Mg2+ was displaced by BaClz. The amount of Mg2+ entrained in the resultant
solution was then determined by atomic-absorption spectrometry. Individual cation CEC
. (i.e. amount of Ca“, Mg”, Na”, and K+) was accomplished using an ammonium acetate
saturation method (Bain and Smith 1994). The NH40AC was adjusted to pH 7.0 with
acetic acid. The results for Ca2+ may be skewed in the presence of free calcite or gypsum,
though other methods are no better.

The data for pore water chemistry are borrowed from Kolak (Ph.D. Dissertation in
progress). Four major cations were analyzed - Na‘, K“, Ca2+, and Mg“. Relatively few
pore water samples (25 samples for Kolak, dissertation in progress, only 14 of which were
analyzed for this study) were analyzed for HCO3'.

Pore water results for this study were compared to groundwater data from the US.
Geological Survey-Water Resources Division, Michigan Regional Aquifers System
Analysis (RASA) study of Michigan basin aquifers (Meissner 1993). Saginaw Bay data in

this study were compared to stream data from the US. Geological Survey National Stream-

l9

Water Monitoring Networks. These stream data were collected from 1965 to 1995 (though

some streams have been monitored beginning in the 19705) (Alexander et al. 1997).

RESULTS

The sediments of Saginaw Bay are sticky mud, ranging in color from 10YR3/1 to
5Y5/2, with most of the samples at 10YR4/2, according to the Munsell Soil Color Chart.
There were no color trends with depth or with carbonate content of the sediment. All the
sediments contained carbonate minerals, and they all had at least a little organic matter.
These muds are sufficiently sticky that separating sand and silt from clay was very difficult.
Clays accumulated into roughly sand-sized particles when rinsed in a sieve, but these
would typically reveal themselves to be clay, as they were easily smeared between the
fingers. Several samples were noticeably sandier than most of the samples, and several had
pebbles of carbonate rock. None of the cores were able to penetrate the overcompacted
Wisconsinan till that underlies the Holocene lake sediments of the bay. Consequently, this
till was not analyzed for this study.

The clay mineralogy of the sediment samples was similar throughout the sample
area, both with depth and areally. The only major clay mineral type which was absent from
the samples was smectite, based on the absence of a peak at 18 A when treated with glycol.
There was a peak at 14 A, indicating chlorite. There was a 7 A peak which corresponded
to the chlorite (002) and the kaolinite(001). The chlorite (004) peak at 3.49 A was just

distinct enough to be distinguished from the kaolinite(002) peak at 3.54 A (Figure 2 and 3).

20

assess: 2 «mm «o gouge 3.x .N tame.

 

_ fl q d

<<.«~«fluqq«4444.4444444.«..d<<44444.

4o:
fiend

21

3.54A

3 .49A

 

 

 

   

 

IILIIIIILILILIII:
Figure 3. X-ray diffractogram of SB 1-1 Mg-saturated showing chlorite(004) peak and
kaolinite (002) peak

22

There was an illite peak at 10 A. It was not detrital biotite because there was a 5 A peak.
When treated with K+, the 14 A peak decreased, and the 10 A increased, indicating the
presence of vermiculite. The changes in those two peak heights were even more
pronounced when K—saturated and heated to 550°C, indicating the presence of hydroxy-
interlayered clays, probably hydroxy-interlayered vermiculite (HIV). There were no depth
trends in the 7-10 A (kaolinite-illite) or 14-10 A (chlorite/vermiculite - illite) peak height

ratios.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

r2 Values fofiJata Versus Depth ‘
Analysis r2

Total CEC vs. Depth 0.133
Na CEC vs. Depth 0.896
Na H20 vs. Depth 0.324
Na H20 <60 vs. Depth 0.735
K H20 vs. Depth 0.011
Ca H20 vs. Depth 0.001
Mg H20 vs. Depth 0.009
Ca/Mg H20 vs. Depth 0.013
K H20 SB24 vs. Depth 0.017
K H20 SB24 vs. Depth w/o 8824-1 0.647
Mg H20 8824 vs. Depth 0.12
__M_g H20 8824 vs. Depth w/o 3824-1 0.671
Ca H20 8324 vs. Depth 0.205
Ca H20 8824 vs. Depth w/o 8824-1 0.779
Na H20 8824 vs. Depth 0.997
Ca/Mg H20 8824 vs. Depth 0.244

 

Table 2. r2 values for data versus depth

 

The average total CEC likewise showed no trends with depth (Table 2 and Figure
4). Since the distribution of clay minerals and organic materials is unrelated to depth, so it

is likely that no depth trends would occur. The average CEC for all samples was 236.8

23

58c 38> omo 3o... .4 tame.

Soon
8a 8. 8? on o

 

 

a F
a . J a

.. om

 

 

 

 

00¢

24

930 101

mg/L. The CEC’s ranged from 45.6 mg/L to 376.5 mg/L. The CEC data yielded poor
reproducibility, with several samples yielding results differing by 50 mg/L, and some
differing by 100mg/L or more. This was probably exacerbated by only three analyses being
done per sample. There was no correlation between how poorly reproducible a sample was

and the salinity of the pore water at the depth from which the sample was taken.

 

r2 for TDS vs. IAP
Analysis r2
TDS vs. IAP-calcite 0.013
TDS vs. IAP-dolomite c 0.007
TDS vs. IAP d 0.008

 

 

 

 

 

 

 

 

Table 3. r2 for total dissolved solids versus ion-activity product

CEC analyses were also done for the individual cations Ca“, Mg“, N a", and K+ to
determine whether cations held on clays and organic materials might be perturbing water
chemistry by exchange; for instance, if Na‘ in the saline waters seeping into the sediments
from the basin were exchanging with Ca2+ adsorbed on the sediment. Little correlation was
found, except for Na CBC and Na H20 (Figure 5). Upon further investigation, it was
found that this CEC method was of little use when pore waters are as saline as these. The
concentrations of cations in the pore water were converted from mg/L to g/cc to a total
weight of cation per water sample. The same was done for the amount of cation extracted
from the sediment by CBC, and the results from the pore water were then subtracted from
the results from the CEC. The bulk of cations measured by CEC analysis were from

deposition onto the sediment surface as the samples were dried for this method, and it

25

58328 E eouubeooeoo .62 «88> ONO +az .m 0.5!..—

Owo a2
om on 00 on ov om ON 0 F o

 

 

. _ h _ P tr
. . 1. 1 l . . 1 o

400—.

. om?

. com

.omN

rcom”

. 0mm

Tomv

 

 

 

com

OZI-l alil
26

proved impossible to tease the amount of adsorbed cations from the data. Negative values

for weights of cations were not uncommon after “correction”.

 

 

 

 

 

 

 

 

 

 

 

 

 

r2 for Calcite/Dolomite Data
Analysis r2
Calcite/Dolomite vs. Na H20 0.198
Calcite/Dolomite vs. CalMg 0.010
Calcite/Dolomite vs. CalMg H20 wlo 24-10 0.166
Calcite/Dolomite vs. CalMg 8824 0.053
Calcite/Dolomite Vs. Depth 0.471
Calcite/Dolomite vs. Depth w/o 8824-10, 1-3A, 18-4A, 19-7A 0.592
Calcite/Dolomite vs. Depth 8824 0.588
Calcite/Dolomite vs. Depth 8824 wlo 8824-10 0.666

 

 

Table 4. r2 for calcite/dolomite data

When total dissolved solids were plotted versus the ion-activity products of calcite
and dolomite (Figs. 6 and 7, and Table 3), there was no correlation. When the
calcite/dolomite ratios for the lake sediments were plotted versus N a+ and Cay/Mg“, the
correlations were poor (Table 4). However, when plotted versus depth (Figure 8), there
were better correlations, especially when four anomalous samples were removed from the
sample set (Figure 9). These samples (SB24-10, SB1-3A, 8319-7, and SBl8-4A) were all
sandier, more calcareous samples, often with coarse pebbles of carbonate rock, and
certainly with sand-sized calcite particles. They all had relatively high calcite/dolomite
ratios (0.745 for SB24-10, 0.217 for SB1-3A, 0.66 for SB19-7, and 0.236 for SBl8-4A -
see Table 5). For calcite/dolomite versus depth with the four anomalous samples removed,

r2 = 0.592. Since clay and silt-sized particles of calcite would tend to dissolve faster than

27

 

 

 

 

0
~ 0
O

——1 b

i . . .

~ . '0 o
—i .

i o

l

i l

' 5
—; i

l l . l r T‘ *5

Figure 6. Total-dissolved solids versus ion-activity product for calcite .

28

700

500

'3
CD

 

 

 

 

 

 

 

 

 

 

 

H T A T I

10 -05 00 06
lAP-doic

i6 -l2 is
lAp~doid

Figure 7. Total-dissolved solids versus ion-activity products for dolomite-c and

dolomite-d

29

9o

 

No

0.0

.88 as? 2828383. a team

md

«b-

.0930
v.0

l4

«.0

«.0

_..o

1i-

 

 

8—

.. 8.

30

9.0

8-2mm
e5 .035 .3 _ mm 673mm 8388 «=er 58.8 as? 88262883 .s 83....

303.0
2.0 1.0 N '0 _..0 00.0 8.0 3.0 «40

 

 

 

 

00p

.:. 8P

31

 

Celene/Dolomite Ratios

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample dgpth calldol CalMgq
88 1-3A 59 0.217 2.68
88 14A 92 0.068 2.55
88 1-5A 120 0 2.25
88 4-1A 0 0 1.92
88 4-3A 60 0.037 1.66
88 44A 91 0.093 1.9
88 4-5A 116 0.129 1.62
88 4-50 140 0.093 1.55
88 4a-1A 0 0 2.51
88 4a-3A 37 0 2.01
88 4b-1 0 0 1.69
88 4b—2A 27 0 1.82
88 4b-4A 87.5 0.077 1.84
88 7-3A 47 0.028 2.05
88 18-1 0 0 2.18
88 18-3A 30 0 1.65
88 18-4A 53 0.236 2.65
88 19-1 0 0 2.18
88 19-3 61 0 1.86
88 19-4 91.5 0 1.91
88 19—6A 152.5 0.165 2.05
88 19-7 198 0.66 1.86
88 19-8A 213.5 0.179 2.02
88 21-1 0 0.068 2.73
88 21-6A 109 0.092 2.25
88 24-1 0 0 2.05
88 24-2A 6.5 0 1.83
88 24-3 28 0.048 1.68
88 24-4A 51 0 1.83
88 24-5A 70 0.042 1.61
88 24-6A 92 0.024 1.89
88 24-7A 119 0.04 2.01
88 24-8A 146 0.137 1.82
88 24-9A 171 0.25 1.94
88 24-10 203 0.745 1.85
8R 1 0 0
SR 2 0 1.33
8R 3 0 1.0
SR 4 0 0

 

 

 

 

Table 5.Calcite/dolornite ratios

32

larger particles, samples that had larger carbonate rock fragments would tend to have more
calcite. For core SB24 (Figure 10), calcite/dolomite versus depth yielded an r2 = 0.588,
which improved to 0.666 when SB24-10 was removed (Figure 11). This sample was
removed because it was a sandy sample with a very high calcite/dolomite ratio (0.745), and
made the sample set seem to have a better correlation of calcite/dolomite versus depth. than
the rest of the samples warrant.

The calcite/dolomite ratios for the sediments taken from Saginaw Bay were all less
than 1(T able 5). Samples which were all dolomite were said to have a calcite/dolomite
ratio of 0. The 4 samples taken at the sediment/water interface in the bay tend to be all
dolomite. Only SB21-1 had calcite (ratio = 0.068). The highest amount of calcite relative
to dolomite was found in sample SB24-10, which has a ratio of 0.745. Samples taken from
the Saginaw River (SR samples in Table 5), show marked heterogeneity. Two had no
calcite, and the other two had ratios of 1.0 or greater.

When all samples, including a plot of Ca2+lMg2+ in solution, were plotted versus
depth, no correlation was found, except Na+ (Figure 12 and Table 2). The correlation for
Na+ was weak, however (r2 = 0.324), and showed a two-pronged signature. Low-Na+
samples showed no correlation with depth. When low-Na+ were removed from the sample
set, the correlation improved to 12:0.735 (Figure 13). The high Na+ samples in figures 12
and 13 all come from two cores, SB-4 and SB-24. These resolve into two separate high-
Na+ trendlines.

When cations were plotted versus depth for the core SB-24 (Figure 14 and Table
2), a suspected site of a saline seep, the correlations with depth "tended to improve. This

was especially true when sample SB24-1 was removed from the sample set (Figure 15).

33

This sample shows an anomalously high value of all cations except N a+, despite being at
the sediment/water interface in the bay. The plot for K“ versus depth for this sample set
yields an r2=0.647 when 24-1 is removed. For Mg2+ vs. depth, r2=0.671 and for Ca2+ vs.
depth r2=0.779. The correlation for Na+ vs. depth for this sample set was r2=0.997, with
sample 24-1 still included in the calculation (Figure 16). For Ca2+/Mg2+ vs. depth,
r2=0.244.

Correlation of Na” with depth tended to be better than the correlations of other
cations with depth (Table 2), though in core SB24, correlations with depth tend to improve,
especially when sample SB24-1 is removed, as noted above. Correlation of the other
cations with Na+ was better than the correlations of these cations with depth, especially
when samples with Na+<70 mg/L were removed from the sample set (Table 6). These low-
Na+ samples tend to be like SB24-1: shallow samples with unusually high concentrations
of cations other than Na+. They tend to be the same samples in every data plot.
Correlations of cations like Ca2+ with Na+ in core SB24 tend to be better than correlations

in the sample set as a whole, especially when SB24-1 is removed (Table 7).

 

 

 

 

 

 

 

 

 

 

 

 

r2 Values for Cations vs. Na 112?
Analysis r2 Vain;
Ca H20 vs. Na H20 0.127
Ca H20 vs. Na H20 >70 0.730
Mg H20 vs. Na H20 0.310
Mg H20 vs. Na H20 >70 0.7731
K H20 vs. Na H20 0,141
K H20 vs. Na H20 >70 0,422
Ca + M20 vs. Na H20 0.059
‘Ca/Mg H20 vs. Na H20 0.206

 

 

Table 6. r2 values for cations versus Na+ concentrations in porewater

34

0.0

 

No

0.0

VNmm 28 cow 5000 «35> 838.3383 .3 ensur—
Guano—03.03200

0.0 to 0.0 «.0

.20

O

 

 

 

trade?

35

8.0

 

2.3mm 2.88 $er 3% 28 ea 58.. 38> 88283an ._ _ 23...
.0038

«.o etc 3 8...

r L
‘ ‘1

O

A

 

 

 

36

0mm

 

590 m§> 88322— 05 S noun—«888 .a2 .2 85E

oom

ammo

mem 9

om?

0

568

O

l—

T
O
to

1

L

O

. 00..

some

.oom

Temu

.000

a 000

e 8v

. 00v

 

 

000

021-1 2N

37

omN

CON

£30 name? 6&8 chA countenance 32 .m— 0.5mm..—

00—.

p
A

5&5

09.

on

 

 

vuflm o _

 

e on

L4 00—.

L

i one

.. omm

.T 09..

 

 

com

OZH 9N

38

Potassium in the pore waters showed a much better correlation with Mg2+ and Ca2+
than with N a+, even when the low-Na+ samples were removed (Table 8). The correlations
were even better for core SB24. There were no anomalous samples.

Samples analyzed for bicarbonate ranged in value from 140.2 mg/L HCOg' to 667.2 mg/L
HCO3' (Table 9). For most cores for which bicarbonate was measured, there is a trend of
increasing bicarbonate with depth, especially for SB-4 and SB-24. However, there was no
depth trend for SB20, and the trend is reversed for SB19, for which core bicarbonate
decreases with depth. Ca2+ versus Mg2+ in the pore waters correlated well (r2=0.864)
(Figure 17 and Table 10), as did these cations when the water of the bay was added to the
data set (r2=0.889) (Figure 18). The correlations were better still for core 8324 (r2=0.934)
(Figure 19). When these cations were added together and plotted versus HCOg' (Figure
20), they correlated well for both pore-water and for bay-water. The slope of the line for
this plot is 0.46. However, when the shallow (<70 cm.) pore-water samples for Ca2+ +
Mg“ versus HCO3' (Figure 21) were plotted separately (the ones corresponding to low-Na+
samples), the r2 dropped from 0.929 to 0.487. The slope for this plot is 0.22, probably
because most of the excess HCO3' is produced in the early stages of burial. Ca2+lMg2+

didn’t correspond at all when plotted versus HCOg’ (Table 10).

 

 

 

 

 

 

 

 

 

 

t2 Values for Cations vs. Na 8824
Analysis r2
K H20 vs. Na H20 8824 0.017
K H20 vs. Na H20 8824 wlo 8824-1 0.675
Ca H20 vs. Na H20 8824 0.024
Ca H20 vs. Na H20 8824 wlo 8824-1 0.798
MgHZO vs. Na H20 8824 0.121
Mg H20 vs. Na H20 wlo 8824-1 0.697

 

 

Table 7. 1'2 values for cations versus Na+ concentrations in porewater for S824

39

35

 

 

 

 

 

 

 

 

 

 

 

 

 

e
. e
3 0 . v o
o
o
25 °
+
N
in 2
+
N
0’ "
1
0.5
O __._.-__.-
O 50 fi_- ~
It!)
150 200 250
Deal:
160
140 0

 

 

 

 

 

 

 

 

 

 

+
N
0’
e
0
0° 0
40
20
o 4 - 4 - e J
o w 1m ’w 2m 2%
Dept
50
4
45
40 0 e
e
35 e
e
so
+
N 0
DD 25 °
2 .
20
e
15
10
5
o A a
0 so too 150 200 250

Fi . 2+ + 2+ + -
gure 14 Ca , Mg2 and Ca IMg2 1n porewater for core S824 versus depth

40

omN

588 38> 3% 28 a .ozo eoeaaeooeoo .2 8&8
586
com om r 09 on

 

 

1)—
_i_
.4

 

 

031-1 )1

41

4,5 ' -_.-_____,

 

 

 

 

 

 

 

 

 

 

 

 

 

290
0
20
o L” ~ -———-- A A A —A . .- —..4
O 50 100 150 200 250

-- M
Figure 16. Concentration of cations in core 8824 versus depth without SBZ4-1

42

 

r2 for K H20 vs. Ca ending
- l H

 

 

 

Analysis r2
K H20 vs. Ca H20 0.638
K H20 vs. Mg H20 0.552

 

K H20 vs. 05 H20 8824 0.801
K H20 vs. Mg H20 8824 0.797

 

 

 

 

 

Table 8. r2 for K+ concentrations in porewater versus Ca2+ and Mg“ concentrations in

porewater

Analysis of 27 bay water samples (converting data from mg/L to mmol/L) yields a
mean C8” + Mg2+/HC03’ for Saginaw Bay of 0.63. The mean Caz+/Mg2+ = 2.17. The
values for the same parameters for Lake Huron are 0.63 for mean Caz+ + Mg2+/l-ICO3' and
Caz‘ilMg2+ = 2.2. For the pore waters of the bay, mean Ca2+ + Mg2+IHCO3' = 0.46, with
mean Cap/Mg2+ = 1.997 (2.003 for samples for which HC03‘ was also analyzed). A t—test
performed on the bay water versus the pore water for Ca2+ + Mg2+IHC03' yielded T = -
10.25 and P = 0.0000. For Ca2+/Mg2+ for the pore waters versus the bay waters, a t-test
yielded t = -2.83 and P = 0.0073. The values for the t-test indicate that the bay water and
the pore waters probably represent different populations, as the differences are significant
at the 95% confidence interval.

The values of Cay/Mg2+ for 227 samples taken from the Michigan basin from

units from the Pennsylvanian on up for the RASA study yield a mean value of 1.92. For

43

omm

 

com

 

vmmm 88 com 50% ”88> .34 mo eouabeooeoo .2 9.53
588
on? 00? om

will
4

b
4

 

 

(NflJEN

on

83380080888 8%? 8.5 00 5:883:00 .3 0.5»...—

ON: a:
8 8 9. on on

or 0

 

O.

 

 

i 09.

i 00..

-. DON

 

 

omm

OZH 33

45

on

venue .08? >3 gougeooeoo
53> 88328 5 A3: 00558588 259, ES 00 covet—5280 .a— 25E
om... as.
00 on 9. on on o _.

_ _ . L _ p
. _

 

 

a . . .

 

l on

i 00—.

i oow

 

 

omN

OZH 93

46

835% 8836.80 30:85 .8 +~w2 382, :80 00 30088880 .3 25w...—

000....
v . m

p
a

4..

«r-

 

 

«P

n!—

 

no

5W+BO

-. me

 

0N

 

47

«€31 e8 em: .aao 88 883588 .a sea.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

383 8«««.4 9.38 «$8.« 39.8 «9.8 .8... 48.8 49.. 252
«to ««.« «.«o« «8.9 «9.8 32: «a... 3:4 «3.. «84mm.
495 «o«.« 38 .«E «3.8 .«ee. «35 «««.«e «8.9 24mm
83 «««.4 e.««« «8.. «8.3 «9.9 «a; «««.9. «8.9 786
83 «mo.« «.«9 :3 «8.8 «9.: 33 «8.3. «3 «9mm
«8.8 «««.« so: «3... 3;... 98.2 «38 «««.«« «no... «.93
«8... «o«.« «a: «9.9 «3.9. 3.2 ««e.o «8.8 «85 798
3...... 33 «.««« 33. «o«.«« .89 and e««.«« 33 «9mm
«33 «««.« «.«9 ««.. «8.9. «8a 32. «3.3 «8... 99mm
2.3 «8.... ««« «8.. 89.8 «2.9 «3 «««.«« «««.9 «-«mm
«98 «««.« ..«e« «8.9 5.8 «2.9 «8... «8.8 «4: «-«mm
58 «8.« «.««« 33 «5.8 «9.9 and «9.3. «o: Emw
«etc «8.2 «.«8 «a... «439 «2.3. 9.3 8.39 «8.« «4mm
«:5 Bee «.3... «3 «$9 «««.«« 98.9 «23: «8.« 3mm
«93 §«.« «.««« 33 46.8 «««.9 Red «5.9. 83 Ema
«d «88 e.««« 83 8«.«2 3.8 «E «3.9 8o.« «same
«23 83 38 «8.« «8.9 «5.9 Is «4.3. «a? 38w
«3... «.«.« :9 9.3 «8.9. «as 83 «B.«« Ed 396
93 «a? «.98 «e: «5.8 «3.9 «as «8.9. «.9 «-««mw
23 «5... «.3« «8.. «2.8 «3.9 23 «9.9. «2 79mm
«etc 33 «.9« «3.9 893.. «8.: 83 «8.9. «3 «.83
«£3 «mo.« «39 «8.9 «9.3 «one «««.o «2.8 «we «.83
«93 «3.4 «.««« «S.« «8.9 «.2 the «3.8 «3 398
Bed 83 «.««« «3.« 39: 49.9 «43 «8.8 33 «9mm
«33 ««.« «.F«« 3.. E..«« 48.9 .86 «9.8. «8.9 4.9mm
.«««.o «8.« «.«8 48.9 i «8.8 «9.9 :«d «8.3. «:9 9.9mm
ass «33:26 3525.....on .355552 SEE. as. + 8 3355.2 + 8 fine ms. .85: 2 do... no .55: «0 29:5

 

 

48

Ca2+ + Mg2+IHCO3' for these same samples, the mean value is 0.51. For Ca2+ +

Mg2+IHCO3' corrected for Ca2+ from gypsum (using 8042'), the mean value is 0.44.

 

 

 

 

 

r2 for Ca and Wig Data
Analysis r2
Ca H20 vs. Mg H20 0.864
Ca H20 vs. Mg H20 w/Bay H20 0.889
Ca H20 vs. Mg H20 8824 0.934

 

Ca + MgporeHZO vs. Alk. mmollL 0.929
Ca + Mg bayHZO vs. Alk. mmollL 0.812
Ca + Mg shallow poreH20 vs. HC03 0.487
CalMg poreH20 vs. H003 0.061
ea/Ma shallow poreH20 vs. H003 0.006

 

 

 

 

 

 

 

 

Table 10. r2 for Ca2+ and Mg2+ water data

There are many data available for rivers in the vicinity of Saginaw Bay from the
USGS National Stream-Water Monitoring Networks (Alexander et al. 1997). For the
Saginaw River, mean Cay/Mg2+ = 2.10. Ca2+ + Mg2+lHC03' for the Saginaw River = 0.74.
Ca2+ + Mg2+/HCO3' + 5042' = 0.55. For the Pigeon River, near Caseville, which is on
Saginaw Bay, Ca2+lMg2+ = 2.21. Ca2+ + Mg2+/HC03' = 0.56 and Ca2+ + Mg2*/Hco3' +
SO42' = 0.36, indicating a relatively large sulphate load for that river. Another river that
feeds into the Saginaw Bay is the Rifle River. Mean Cay/Mg“ = 2.35, and Ca2+ +

MgZVHCOg' = 0.59. Ca2+ + Mg2+IHC03' + 3042' = 0.49 for this river.

49

DISCUSSION

Despite the suggestions of Drimmie et al. (1992) that when the CEC of a sediment
exceeds the cation charge in solution, CEC can be a major control for the mobility of ions
in pore waters, it has proven difficult to get a handle on CBC effects in Saginaw Bay,
despite the relative abundance of vermiculite (a phyllosilicate with a relatively large CEC).
The waters seeping into the bay seem to be sufficiently concentrated that it was impossible
to tease relative abundances of cations adsorbed on the sediment out of the data. The CEC
experiments available are designed for soil scientists working in the vadose zone, where
water entering the material is likely to be dilute (rainwater, for instance), and seem to be
inadequate for saturated sediments in contact with less dilute solutions. The correlation of
K" with Ca2+ and Mg2+ is a positive one; as Ca2+ and Mg2+ increase in concentration in the
porewaters, so does K+. If CEC were primarily responsible for the cation concentrations,
then the relationship should be an inverse one (as Ca2+ or Mg2+ are adsorbed, two K+ ions
would be released). N a+ concentrations decrease as a function of depth, regardless of what
the concentrations of other cations are.

The bulk clay mineralogy of the sediments of Saginaw Bay is similar to that
reported for Lake Ontario, the clay fraction of the sediment containing illite (or muscovite),
vermiculite, kaolinite, and chlorite (Warren et al. 1996). Warren et al. (1996) calculated
the amount of dioctahedral mica in the sediments of Lake Ontario by dissolving Ca2+-
saturated clay-size separates in concentrated HF and HNO3. The digests were then
analyzed for K+ content by flame atomic absorption spectrophotometry (AAS). A K+

content of 83 g/kg was assumed for pure mica to represent the minimum content in the

50

clay-size separates. Vermiculite content was calculated by determining the total CEC after
saturation with KL After drying and resaturation with Ca2+, total CEC was recalculated.
The difference between the two CEC values was then used to calculate the vermiculite
content of the sediment, assuming an average CEC of 1,540 mmol(+)/kg. Amount of
kaolinite was approximated from the x-ray diffraction data, with the remaining clay
fraction attributed to chlorite. Though the clay mineralogy of Saginaw Bay sediments was
not quantified, peak heights on x-ray diffractograms indicated that chlorite was by far the
least abundant clay mineral present (except for entirely-absent smectite), whereas chlorite
was reported as the second-most abundant clay mineral in Lake Ontario sediments. The
other major difference was that no carbonates were reported in the clay fraction of Lake
Ontario sediments, whereas both calcite and dolomite are present in the <2}J.m. fraction of
Saginaw Bay sediments. Weiler (1973) also found montmorillonite in one of his sediment
cores from Lake Ontario, a clay mineral which is absent from all samples from Saginaw
Bay. Kramer (1962) noted the presence of small amounts of an iron sulfide mineral
(identified as “troilite”) in all clay bottom sediments from Lake Ontario. Iron sulfides were
largely absent from Saginaw Bay sediments, as also noted by Wood (1958). These
differences would seem to indicate a difference in source material for the sediments. The
sediments of Saginaw Bay are similar to the sediments of Lake Michigan, with the same
clay minerals represented, and including the increase in carbonate minerals with depth
(Callendar 1969). However, Callendar attributed the trend to changing source materials,
whereas this study suggests that changes in carbonate mineral content in Saginaw Bay
sediments are probably due to diagenesis, because of the fairly consistent decrease in

calcite going from deep to shallow sediments. Lake Superior sediments are very poor in

51

carbonate minerals, and are rich in chlorite compared to the other Great Lakes, including
Saginaw Bay (Callendar 1969). Differences in carbonate mineral and chlorite content in
Saginaw Bay from that of the sediments of Lake Superior is probably due to differences in
source rocks. Saginaw Bay’s sediment input is from areas with considerable amounts of
limestones, dolomites, and calcareous tills, whereas sediment source areas for Lake
Superior are primarily igneous and metamorphic rocks, with a correspondingly greater
amount of chlorite over Saginaw Bay, and a lesser amount of carbonate.

The high correlation of Na“, especially in sample cores SB24 and SB4, the sites of
probable saline seeps, with depth seems to indicate that N a+ concentrations are the result of
simple mixing of the dilute bay waters with the saline waters from the basin. Na+
consequently seems to be a good index cation for degree of dilution in fresh water. Kolak
et al. (1999) agree that cores SB24 and SB4 are probably sites for saline seeps, probably
due to fracturing of underlying units which allow upward transport of saline formation
waters. Indeed, core SB24 shows the best correlation of N a+ concentration with depth (r2 =
0.997). The two-pronged signature for Na“ versus depth is a result of saline gradients in
these two cores. Kolak et al. (1999) indicate that many of the cores have weak chloride
gradients, and 6180 similar to the modern water of the bay, unlike cores like SB24 and
SB4, which have strong chloride gradients and lighter 8180 values with increasing depth.

Ca2+, Mg”, and K+ were poorly correlated with depth, but well-correlated with Na+, as
- long as the Na+ concentrations were above '70 mg/L. Therefore, for water with Na+ > 70
mg/L, mixing dominates the concentrations of these cations, although an r2 = 0.779 for
Ca2+ versus depth for core SB24 (without 8324-1) suggests some diagenetic influence,

since if the situation were ure mixin , then the r2 should be closer to uni . Likewise,
P g

52

Mg2+ and K+ have r2 less than unity (0.671 for Mg2+ and 0.647 for K‘“). For this same core,
r2 for N 3+ = 0.997, even with sample SB24-1. Sample SB24-1 had high concentrations of
cations other than N a+. For instance, the concentration of Ca2+ in SB24-1 is 141.4 mg/L,
versus 57.1 mg/L in sample 8324-2, the next shallowest sample. Also, concentrations of
these cations remained high even in more dilute waters, as measured by Na+
concentrations. Thus, it seems that simple mixing is not the only mechanism controlling
these cation concentrations.

The depth-related trend of calcite/dolomite, with calcite becoming less abundant
towards the top of the sediments may be a result of diagenesis. It is difficult to eliminate a
change in source area as the cause of the trend. Some samples were viewed under a
scanning electron microscope, and etch pits were seen in calcite, but the pits may have
been the result of in situ weathering in the source area. The pore water chemistry is what
makes it seem more likely that this trend is the result of diagenesis. The sandy sediments
tend to have sand to pebble sized pieces of limestone in them, which disturbs the
calcite/dolomite depth trend, probably because large pieces of calcite take much longer to
dissolve than silt- to clay-sized particles. Sediments that only ever had particles of calcite
in the silt- to clay-sized range may have lost this calcite relatively rapidly, due to surface-
area enhanced dissolution rates.

As previously noted, if the sum of the molar concentrations of Ca2+ and Mg2+ is
plotted versus HCOg', the slope of the least-squares best fit line can yield clues about
diagenetic processes operating in the sediment. If the primary control is COz-driven
congruent or incongruent dissolution of calcite and dolomite, the slope will be ~0.5. The

ratio of Ca2+ and Mg2+ will be 2:1, if both calcite and dolomite dissolve stoichiometrically

53

and at equal molarities. This is kinetically unlikely, however, and the aggreement of these
data with this ideal ratio probably reflects the preponderance of dolomite in the sediment.
For none of the porewater samples in this study did the amount of calcite exceed that of
dolomite.

The chemistry of the pore water of Saginaw Bay supports the hypothesis that the
content of Mg2+ and Ca2+ is controlled by the dissolution of carbonates, especially when
combined with the depth trend in calcite/dolomite. The ratio of Ca2+ + Mg“ over HCO3'
averages 0.46 for the pore waters of Saginaw Bay, with a mean CaZJ'IMg2+ of 1.997. This
very nearly matches the ideal ratio for a water chemistry controlled solely by carbonate
dissolution. However, the Ca2+Mg2+ was not 1.0 for samples that contained only
dolomite, though these samples had ratios of these cations that were between that of the
bay water and the ideal ratio for congruent dissolution of dolomite (from 1.65 to 2.05, as
opposed to 2.17 for the bay). The only exception was SB 1-5A, which has a Cay/Mg2+ of
2.25, which is higher than that of the bay. These samples were all near or at the
sediment/water interface, so the concentration of calcium is curious. Perhaps this is due to
CEC release of Ca2+, the dissolution of Ca-silicates, or calcite dissolution (dolomite should
dissolve more slowly). River sediments analyzed for calcite/dolomite content were highly
variable in calcite content; a couple of the samples contained only dolomite. The mean
Ca2+lMg2+ for the Saginaw River is 2.0921, and mean Ca2+ + Mg2+IHCO3' for the Saginaw
River is 0.74, according to the USGS stream survey data, though when Ca2+ + Mg2+ is
divided by 8042’ + HCOg' for these data, the ratio = 0.55. The variable calcite content of
the streambed sediment may be a reflection of how recently the sediment has been

deposited and proximity to limestone boulders (no outcrops were visible in the areas that

54

samples were taken from, but limestone boulders were present in some areas to stabilize
banks). It was difficult to get samples far from shore, due mainly to the temperature of the
water (samples were taken in early spring) and instability of the river banks. The chemistry
of the river water suggests that its source areas must be dominated by limestones,
dolomites and calcareous tills, with sulfate input from evaporites and/or pollution (Bemer
and Bemer 1996).

By contrast, the studies of Langmuir (1971), Bennett et al.(1993), and Siegel
(1989) did not fare as well when compared to a pure carbonate-dissolution model. Bennett
et al. (1993) were modelling a situation where a large volume of organic contaminant was
perturbing the natural system. The acidity of the system at Bemidji increases in the vicinity
of the oil plume, suggesting that deviation from the ideal Cay/Mg2+ ratio of 2:1 is due to
an excess of Ca2+ generated by weathering of calcium-bearing silicates. Ca2+- bearing
silicate minerals like anorthite are devoid of magnesium and are easily weathered relative
to many other abundant silicates, such as micas, quartz, and sodic or potassic feldspars.
This would tend to skew the Caz‘VMg2+ ratio in the direction of Ca2+. Due to oxidation of
the crude oil, carbonic acid is produced, causing acid-enhanced weathering of all minerals.
Bennett et al. ( 1993) agree that this is probably occurring at this site. Langmuir’s (1971)
springs mostly issued from limestone, leading to the high amount of Ca2+ relative to Mg2+
for those waters (Caz‘VMg2+ = 3.75). The wells were mostly in dolomite. Consequently,
the ratio of CaZJ’IMg2+ is very close to the ideal 1:1 ratio for dissolution of dolomite only
(Ca2+lMg2+ = 1.3 for all wells, Ca2+IMg2+ = 0.95 for wells only in dolomite). The deviation
from the ideal Ca2+ + Mg2+IHCO3' of 0.5 may be due to degassing, decreasing the amount

of bicarbonate in the water. As water that is at equilibrium with carbonate minerals is

55

exposed to the atmosphere, such as when issuing from a formation as a spring, bicarbonate
tends to escape from the water as carbon dioxide gas. It is this process that changes
carbonate equilibrium in cave waters, leading to the deposition of speleothems.

Siegel (1989) had results for Ca2+ + Mg2+lHC03' that were close to ideal for his
Aquifer 5, which was in carbonate-rich Holocene deposits, Pleistocene glacial drift, and
Cretaceous rocks (the Dakota Formation, which consists primarily of marine shales,
siltstones and lignitic sandstones, with disseminated pyrite in its basal portion). There was
an excess of Ca2+ over Mg2+ in this aquifer (ratio = 3.38), possibly reflecting a greater
calcite content of the sediment, or non-stoichiometric dissolution. In this case, dissolution
of dolomite would have to preferentially release calcium. Incongruent dissolution of
dolomite, however, seems to usually release Mg“, transforming dolomite into calcite in the
process (Eqn. 2). It is difficult to constrain from his results, because he admits a
quantitative analysis of the mineral content of the rocks and sediments in his study area is
lacking. For his two deeper aquifers, Aquifer 3 (Ca2+ + Mg2+/HC03' = 0.41) and Aquifer 5
(Ca2+ + Mg2+/l-ICO3' = 0.32), there is either a process that is producing an excess of
bicarbonate over Ca2+ and Mg”, or that is consuming these cations without removing
HCOg'. A mechanism Siegel (1989) proposed to remove calcium and magnesium was
cation-exchange with Na+, as many of the water in the deeper aquifers are distinguished by
a dominance of this cation.

Furthermore, data from the US. Geological Survey-Water Resource Division,
Michigan Regional Aquifers System Analysis (RASA) study of groundwater in Michigan
yields numbers considerably closer to the data from Saginaw Bay pore waters than-the data

from the bay water. These waters are from the deep Michigan basin, and the mean Ca2+ +

56

Mg2+/Hco,' = 0.51 for 227 wells for the RASA study. The ratio fell to 0.44 after
correcting for probable gypsum dissolution, which would contribute Ca2+. The correction
was done stoichiometrically, using the concentration of SO42' in the water. All 8042’ was
assumed to be from gypsum dissolution, and an amount of Ca2+ equivalent to the amount
of 8042' was subtracted from the Ca2+ content of the water (gypsum is CaSO4-2H20).

The mean Ca2+ + Mg2+/I-IC03' for Saginaw Bay = 0.68. Though when plotted with
the mean ratio for the pore waters (Ca2+ + Mg2.+/HC03' = 0.46) the change in slope is minor
from the plot of just the pore water (Figure 12 and 13), the difference in Ca2+ +
Mg2+/HCO3’ between the water of Saginaw Bay and the pore waters is highly significant
with a t-test, T=-10.25 and P=0.0000. The mean Ca2+/Mg2+ for the RASA study yie1ded
1.92 (2.54 after gypsum correction). The mean Ca2+fMg2+ for the bay was 2.17. When this
mean is compared with the mean for the pore waters (Ca2+lMg2+ = 2.17), the t-test also
proves to be highly significant with T=-2.83 and P=0.0073.

The pCOz is higher in the pore waters (log pCOz = 1.84 - 2.57) than for the bay
waters (log pCOz = 3.10 — 3.39) (Kolak, dissertation in progress), which would enhance
carbonate dissolution, which may be a reason the pore waters fit a carbonate-dissolution
model better than the waters of the bay.

Overall, the waters of Saginaw Bay seem to be most similar to the chemistry of
river water in the area. The river waters have been found to have a Ca2+ + Mg2+IHCO3' =
0.63 (Alexander et al. 1997). The Ca2+lMg2+ varies from 1.1 to 2.14, dependent on flow,
with the lowest values in summer periods of low flow (Kolak, dissertation in progress).
The Cay/Mg2+ ratios in the river water seem to match the relative abundances of dolomite

and calcite in the river sediment. The bay is very similar in concentration to Lake

57

Michigan and (unsurprisingly) Lake Huron, though a study conducted in 1969 indicates
that the chemistry of the lakes has changed in the last 30 years, especially Lake Huron.

The ratios of Ca2+ + Mg2+/HCO3' from the bay resemble mean river water for
North America and the world as a whole, as calculated from numbers in Bemer and Bemer
(1996). The Great Lakes are in basins that have a lithology similar to that of the rocks that
most of North America’s rivers flow over (sedimentary, with a considerable amount of
limestone and dolomite), so the water chernistries are similar to that of North American
mean river water, though the Kolak (dissertation in progress) found that Ca2+lMg2+ could
be as low as 1.1 during low-flow periods in summer for rivers in the area. The Ca2+ +
Mg2+lI-IC03' (using their “unpollut ” numbers) for N. America is 0.60, with a Ca2+lMg2+
of 2.48. For the world mean river water, Ca2+ + Mg2+lHCO3' is 0.54, and Caz‘VMg2+ is
2.39. These numbers seem to indicate that carbonate-dissolution is a major influence on
water chemistry for fresh waters. When there is a relative abundance of Ca2+ over Mg“
(ratio >2), it may be due to combinations of reactions involving the dissolution of Ca-
bearing silicates, such as anorthite, which, as noted previously, produce an excess of Ca2+
over Mg2+ in a ratio greater than 2: 1, but is likely due to a preponderance of calcite over
dolomite. The relative closeness of the ratios from the Saginaw Bay to those for an ideal
carbonate-dissolution situation is probably due in part to the relative abundance of
carbonate rocks and carbonate-bearing sediments in the immediate vicinity of the bay, with
largely carbonate-bearing sediments underlying rivers draining into the bay.

K” was not completely studied for this project. It seemed to behave similarly to
Ca2+ and Mg2+. Like them, it correlated poorly with depth for the entire data set, but

correlation with depth improved in SB24, especially when sample 8324-1 was removed (in

58

this case, r2 = 0.647). Correlations of K‘“ with N a+ were worse than with calcium and
magnesium ( r2 = 0.422 for K’r vs. N a+ > 70, as opposed to r2 = 0.730 for Ca2+ and r2 =
0.773 for Mgz”), implying that potassium is less controlled by mixing effects than any of
the other cations in this study. When the activities of potassium over pH were plotted
versus the activity of dissolved silica (PLSiO4) using data from the 1985 EPA Survey of
Lakes Michigan, Erie and Huron, a potential disequilibrium with illite was found. Modern
waters plot squarely in the kaolinite stability field, whereas basin waters tend to plot in or
near the illite stability field (W ahrer 1993). The sediments in Saginaw Bay may be
undergoing kaolinization of illite, with a concurrent release of K+. K+ can also be released
by the decay of organic matter ( Bemer and Berner 1996), or by the weathering of K-

feldspars. Unfortunately, there were not enough data to constrain the possible reactions.

CONCLUSIONS

Overall, the pore waters in the sediments of Saginaw Bay are dominated by
stoichiometric dissolution of carbonates, as demonstrated by the ratio of Ca2+ +
Mg2+/I-IC03' = 0.46 and the calcite/dolomite data. Na+ seems to be controlled by simple
mixing of the dilute water of the bay and the brines seeping into the sediments from the
Michigan basin. The bay water is a distinct population from the pore waters, with an
excess of Ca2+ + Mg2+ over HCO3', probably from a combination of reactions from the

heterogeneous rocks that the rivers of the lower peninsula pass over. K+ is not from simple

59

mixing, but it was impossible to constrain the reactions that were producing the diagenetic

signature.

REFERENCES CITED

Alexander, R.B., Slack, J .R., Ludtke, A.S., Fitzgerald, K.K., and Schertz, T.L., Data from
selected US. Geological Survey National Stream Water-Quality Monitoring
Networks (W QN), USGS Digital Data Series DDS-37: Region 04 Great Lakes,
http://www.rvares.er.us gs. gov/qu96cd/html/qu/wq/region04. 1

April, R., Newton, R., and Coles, LT, 1986, Chemical weathering in two Adirondack
watersheds: Past and present-day rates, Geological Society of America
Bulletin, v. 97, p.1232-1238.

Badalamenti, LS, 1992, The geochemistry and isotopic chemistry of saline ground water
derived from near-surface deposits of the Saginaw Lowland, Michigan basin,
unpublished Master’s thesis, Michigan State University, 126p.

Bain, DC, and Smith, B.F.L., 1994, Chemical analysis. In Wilson, M.J., ed., §l_av
Mineralogy: chtroscopic and Chemical Deterrninative Methods, p. 311-315.

Beeton, A.M., 1970, Statement on pollution and eutrophication of the Great Lakes, Center
for Great Lakes Studies, University of Wisconsin-Milwaukee, Special Report No.
11, 35 p. .

Bennett, P.C., Siegel, D.E., Baedecker, M.J., and Hult, M.F., 1993, Crude oil in a shallow
sand and gravel aquifer-I. Hydrogeology and inorganic geochemistry, Applied
Geochemistry, v. 8, p. 529-549.

Bemer, E.K. and Bemer, R.A.,1996, Global Environment; Water, Air; and Geochemical
nglis, Prentice Hall, Upper Saddle River, NJ, 376 p.

Callendar, E., 1969, Geochemical characteristics of Lakes Michigan and Superior

61

sediments, Proceedings from the 12th Conference of Great Lakes Resources, p.
124-160.

Desaulniers, D.E., Cherry, J .A., and Fritz, P., 1981 , Origin, age, and movement of pore
water in argillaceous Quaternary deposits at four sites in southwestern Ontario,
Journal of Hydrology, v. 50, p. 231-257.

Dorr, J .A., and Eschman, DR, 1970, Geblogy of Michigan: Ann Arbor, The University of
Michigan Press, 476 p.

Drever, J .I., 1973, The preparation of oriented clay mineral specimens for x-ray diffraction

analysis by a filter-membrane peel technique, American Mineralogist, v. 58, no. 5-
6, p. 553-554.

Drimmie, R.J., Frape, S.K., Morden, MJ., Thomas, R.L., McNutt, R.H., and Bowins,
R.J., 1992, Chemical inputs to Lake Ontario from saline formation water below the
sediments, 35th Conference of the International Association for Great Lakes
Research, University of Waterloo, Waterloo, Canada, Abstracts, p. 90.

Effler, S.W., 1984, Carbonate equilibria and the distribution of inorganic carbon in
Saginaw Bay, Journal of Great Lakes Research, v. 10, no. 1, p. 3-14.

Eschman, BF, and Karrow, RR, 1985, Huron basin glacial lakes: a review. in Karrow,
PF, and Calkin, P.E., eds., Quaternary Evolution of the Greact Lakes, Geol. Assoc.
of Can. Spec. Paper 30: 79-93.

Hanor, J .S., 1982, Modification of the quality of water injected into Louisiana gulf coast
sands; effects of cation exchange, Environmental Geology and Water Sciences, v.
4, no. 2, p. 75-85.

Jones, J .F., Gardner, S., and Simon, B.M., 1984, Reduction of ferric iron by heterotrophic

62

bacteria in lake sediments, Journal of General Microbiology, v. 130, p. 45-51.

Kramer, J .R., 1964, Theoretical model for the chemical composition of fresh water with
application to the Great Lakes, Publication N o. 11, Great Lakes Research Division,
p. 147-160.

Kramer, J .R., 1966, Equilibrium models and composition of the great lakes,

Kolak, J .J ., Long, D.T., Matty, J .M., Larson, G.J., Sibley, D.F., and Councell, T.B., 1999,
Ground-water, large-lake interactions in Saginaw Bay, Lake Huron: a
geochemical and isotopic approach, Geological Society of America Bulletin, v.
111,no.2,p.177- 188.

Kolak, J .J ., untitled doctoral dissertation, Michigan State University, 1999.

Kunze, G.W., and Dixon, J.B., 1987, Pretreatment for mineralogical analysis. Section 5-4.
In A. Klute, ed., Methods of Soil Analysis Part I - Physical and mineralogical
methods, 2nd ed., Agronomy 9:97-99.

Langmuir, D., 1971, The geochemistry of some carbonate ground waters in central
Pennsylvania, Geochimica et Cosmochimica Acta, v. 35, p. 1023-1045.

Larson, G.J., personal communication, in person.

Long, D.T., Wilson, T.P., Takacs, MJ., and Rezabek, DH, 1988, Stable-isotope
geochemistry of saline near-surface ground-water: east-central Michigan basin,
Geological Society of America Bulletin, v. 100, p. 1568-1577.

Lovley, D.R., Baedecker, MJ., Longeman, D.J., Cozzarelli, I.M., Phillips, E.J.P., and
Siegel, D.I., 1989, Oxidation of aromatic contaminants coupled to microbial iron
reduction, Nature, v. 339, p. 297-299.

Meissner, B.D., 1993, The geochemistry and source for solutes in ground water from the

63

Pennsylvanian bedrock sequence in the Michigan basin, unpublished master’s
thesis, Michigan State University, 115 p.

Rheaume, SJ. 1991, Hydrologic provinces of Michigan, U.S.G.S. Water Resources
Investigations Report 91-4120.
Robbins, J .A., 1986, Sediments of Saginaw bay, Lake Huron: elemental composition and
accumulation rates, Special Report 102, Great Lakes Research Division.
Sequiera, R., 1991, A note on the consumption of acid through cation exchange with clay
minerals in atmospheric precipitation, Atmospheric Environment, v. 25A, no. 2, p.
487-490.

Siegel, D.I., 1989, Geocherrristry of the Cambrian-Ordovician aquifer system in the
northern midwest, United States, U.S.G.S Professional Paper 1405-D.

Tellam, J .H., 1995, Hydrochemistry of the saline groundwaters of the lower Mersey Basin
Permo-Triassic sandstone aquifer, UK, Journal of Hydrology, v. 165, p. 45-84.

Wahrer, M.A., 1993, The geochemistry and source of solutes in ground water from the
glacial drift regional aquifer, Michigan basin, unpublished Master’s Thesis,
Michigan State University, 111 p.

Warren, C.J., Drimmie, R.J., Kanters, W.A., and Frape, SK, 1996, Clay mineralogy of
Lake Ontario sediments, Journal of Great Lakes Research, v. 22, no. 4, p. 864-870.

Weaver, T.A., Frape, S.K., Cherry, J .A., 1995, Recent cross—formational fluid flow and
mixing in the shallow Michigan basin, Geological Society of America Bulletin, v.
107, no. 6, p. 697-707.

Weiler, RR, 1973, The interstitial water composition of the Great Lakes, 1. Western Lake

Ontario, Limnology and Oceanography, v. 18, no. 6, p. 918-931.

Weiler, RR, and Chawla, V.K., 1969, Dissolved mineral, quality of Great Lakes waters,
Proceedings of the 12th Annual Conference Great Lakes Resources, p. 801-818.

Westjohn, DB, and Weaver, T.L., 1996, Configuration of freshwater/saline—water
interface and geologic controls on distribution of freshwater in a regional aquifer
system, central lower peninsula of Michigan: US. Geological Survey Water-
Resources Investigation Report 94-4242, Lansing, Michigan, 44 p.

Whittig, L.D., and Allardice, W.R., 1987, X-ray diffraction techniques, Section12-3.3, In
A. Klute, ed., Methods of Soil Analysis, Part I-Physical and mineralogical
methods, 2nd ed., Agronomy 9:331-362.

Wilson, T.P., and Long, D.T., 1985, The behavior of bromide during the dissolution of
halite at 25 degrees C and 1 atrn., 97th Annual Meeting, Geological Society of
America, Boulder, Colorado, Abstracts, p. 697.

Wilson, T.P., and Long, D.T., 1986, The geochemistry of Michigan basin brines, Abstract
and paper presented American Chemical Society Annual Meeting, Anaheim,
California, Geochemistry Program.

Wilson, T.P., and Long, D.T., 1993, Geochemistry and isotope chemistry of Michigan
Basin brines: Devonian formations, Applied Geochemistry, v. 8, p. 81-100.

Wood, L.E., 1958, Bottom sediments of Saginaw bay, Michigan, Ph.D. dissertation,
Michigan State University, 325 p.

Wood, LE, 1964, Bottom sediments of Saginaw bay, Michigan, Journal of Sedimentary

Petrology, v. 34, no. 1, p. 173-184.

65

"Illlllllllllllllllll’