IN F O R M A T IO N T O U S E R S
T he m o s t a d va n ce d te c h n o lo g y has been use d to p h o to ­
g ra p h a n d re p ro d u ce th is m a n u s c rip t fro m th e m ic ro film
m aste r. U M I f ilm s th e t e x t d ir e c t ly fr o m th e o r ig in a l o r
copy s u b m itte d . T h u s , some th e s is a n d d is s e rta tio n copies
are in ty p e w r ite r face, w h ile o th e rs m a y be fro m a n y ty p e
o f c o m p u te r p rin te r.
T he q u a lit y o f t h is r e p ro d u c tio n is d e p e n d e n t u p o n th e
q u a lity o f th e copy s u b m itte d . B ro k e n o r in d is tin c t p r in t ,
c o lo re d o r p o o r q u a lit y illu s t r a t io n s a n d p h o to g ra p h s ,
p r in t b le e d th ro u g h , s u b s ta n d a rd m a rg in s , a n d im p ro p e r
a lig n m e n t can a d v e rs e ly a ffe c t re p ro d u c tio n .
I n th e u n lik e ly e ve n t th a t th e a u th o r d id n o t send U M I a
com plete m a n u s c rip t a n d th e re are m is s in g pages, these
w i l l be n o te d . A ls o , i f u n a u th o riz e d c o p y r ig h t m a t e r ia l
had to be rem oved, a n o te w i l l in d ic a te th e d e le tio n .
O versize m a te ria ls (e.g., m aps, d ra w in g s , c h a rts ) a re re ­
p ro d u ce d b y s e c tio n in g th e o r ig in a l, b e g in n in g a t th e
u p p e r le ft-h a n d c o rn e r a n d c o n tin u in g fro m le f t to r ig h t in
eq ua l sections w it h s m a ll overlaps. E a ch o r ig in a l is also
p h o to g ra p h e d in one exposure a n d is in c lu d e d in reduced
fo rm a t th e ba ck o f th e book. These a re also a v a ila b le as
one exposure on a s ta n d a rd 3 5 m m s lid e o r as a 17" x 23"
b la c k a n d w h it e p h o to g r a p h ic p r i n t f o r a n a d d it io n a l
charge.
P h o to g ra p h s in c lu d e d in th e o r ig in a l m a n u s c r ip t have
been re p ro d u c e d x e r o g r a p h ic a lly in t h is copy. H ig h e r
q u a lit y 6" x 9" b la c k a n d w h it e p h o to g ra p h ic p r in t s a re
a va ila b le fo r a n y p h o to g ra p h s o r illu s tr a tio n s a p p e a rin g
in th is copy fo r a n a d d itio n a l charge. C o n ta c t U M I d ir e c tly
to order.

University Microfilms International
A Bell & Howell Information C om pany
300 North Zeeb Road, Ann Arbcr, Ml 48106-1346 USA
313/761-4700 800/521-0600

O rder N um ber 9012086

Origin and geochem ical evolution o f th e M ichigan basin brine
Wilson, Timothy Peter, Ph.D.
Michigan State University, 1989

C o p y rig h t © 1 9 8 9 b y W ils o n , T im o th y P e te r. A ll rig h ts reserved.

UMI

300 N. Zeeb Rd.
Ann Arbor, MI 48106

ORIGIN AND GEOCHEMICAL EVOLUTION
OF THE
MICHIGAN BASIN BRINE

By
Timothy Peter Wilson

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Geological Sciences
1989

ABSTRACT
ORIGIN AND GEOCHEMICAL EVOLUTION
OF THE
MICHIGAN BASIN BRINE
By
Timothy Peter Wilson

Chemical and

isotopic data were collected on 126 oil­

field brine samples and were used to investigate the origin
and geochemical evolution of water in 8 geologic formations
in

the

Michigan

basin.

The

data

were

analyzed

using

graphical, thermodynamic modeling, and statistical methods.
Cl/Br

ratios

suggest

that

the

Michigan

basin

brines

originated from evapo-concentrated seawater.

Two groups of

brine

brine

are

found

in

the

basin,

the

Na-Ca-Cl

in

the

upper Devonian formations, and Ca-Na-Cl brine from the lower
Devonian and Silurian aged formations.
Devonian Berea,
from

seawater

brine

evolved

Traverse,
concentrated
by

aluminosilicate
bacterial

and Dundee
into

halite

or

by

formations
halite

precipitation,

reactions,

action

the

Water in the upper

and

CaS04

the

originated

facies.

This

dolomitization,

removal

of

precipitation.

S04

The

by

stable

isotopic composition (D, 0) is thought to represent dilution
of evapo-concentrated seawater by meteoric water.
brine

origins

sabkhas

and

are:

lagoons,

(1)

seawater

reflux

down

evaporation
into

the

Possible
in

coastal

sediments,

and

Timothy P. Wilson
later

dilution

by

meteoric

water,

or

(2)

residual

fluids

squeezed from the underlying Devonian and/or Silurian salts
during compaction mixed with dilute water

in the overlying

formations but maintaining saturation with halite.
Water

in the

lower

Devonian Richfield,

Detroit

River

Group, and Niagara-Salina formations is very saline Ca-Na-Cl
brine.

Cl/Br

concentrated

suggest

through

the

it

originated

halite

and

into

from
the

seawater
MgS04

salt

facies, with an origin linked to the Silurian and Devonian
salt

deposits.

increased

the

Dolomitization
Ca/Na,

and

aluminosilicate

halite

precipitation

reactions

removed

K,

and bacterial action or CaS04 precipitation removed S04 from
this brine.

Additional CaCl2 enrichment may have occurred

from the diagenesis of the Salina A-l potash salt.
isotopes

reflect

the

seawater

origin,

but

may

Stable

have

been

affected by carbonate equilibria at higher temperatures.
Water chemistry

in the Ordovician Trenton-Black River

formations indicates dilution of evapo-concentrated seawater
by fresh or seawater.
Ordovician

seawater,

Possible
present-day

saline end-members
upper

Devonian

Ca-Cl brine from the deeper areas in the basin.

include

brine,

or

Copyright by
TIMOTHY PETER WILSON
1989

ACKNOWLEDGMENTS
I would like to acknowledge the help given by the
following companies and individuals:
The Michigan oil and gas industry, especially the Dart
Oil and Gas Company, Shell Oil, Amoco Oil Company,
Mossbacher Production Company, Wiser Oil, Miller Oil,
and Wenner Production Company.
The faculty and staff at the Department of Geological
Sciences, Michigan State University.
Argonne National Laboratory.
Dr. Alden Carpenter, Chevron Research Co.
I would also like to acknowledge the following agencies
that funded this project:
The American Chemical Society Petroleum Research Fund.
Geological Society of America.
American Association of Petroleum Geologists.
Amoco Production Company.
Shell Western Exploration and Production Company.
Department of Energy Educational Programs,
Argonne
National Labortories.
Department
of
Geological
Sciences,
Michigan
State
University.
I would like to express my sincere thanks to the
teachers and friends that have helped me in this work:
Dr. David Long, Dr. Grahame Larson, Dr. Duncan Sibley,
Dr. William Cambray, Dr. John Wilband, Dr. William
Cooper, Dr. Richard Cooper, Dr. Richard Heimlich, Mike
Takacs, Dale Rezabek, Jim Tolbert, Joe McKee, and Dr.
Douglas Lee.
Finally, I would like to express my heartfelt gratitude
to my wife Janet, and to Dave.
Without your support,
friendship, and understanding, this work would not have
been possible.

TABLE OF CONTENTS

Chapter 1: Geochemistry of water in
Devonian aged formations, Michigan Basin
Introduction ............................................ 1
This S t u d y .............................................. 3
Study A r e a ......................................... 4
Analytic M e t h o d s ........................................ 15
Geochemical R e s u l t s ..................................... 19
Distribution of brine chemistry .................. 21
Major ion c o m p o s i t i o n ............................ 25
Geochemical o r i g i n ..........................
. . .36
Geochemical evolution ............................
37
Water-rock i n t e r a t i o n s ....................... 4 0
D o l o m i t i z a t i o n ................................40
Alumino-silicate reactions
, . ..49
Other s a l t s ...................... . ........... 51
Model for brine evolution
based on geochemical d a t a ........................ 53
Stable isotope results ................................. 54
Interpretation
...............................57
D i s c u s s i o n ......................................... 60
Saline endmember ............................
61
Apparent modern-day meteoric water ......... 62
Linking the isotopic and chemical data . . . 64
Model for brine evolution
based on isotopic d a t a ............................ 71
Summary
....................................... 75
Chapter 2: Origin and evolution of water in the NiagaraSalina and Ordovician aged formations,
Michigan Basin
Introduction ............................................ 78
Study Area, Niagara/Salina reefs ...................... 82
Reef h i s t o r y ...................................... 83
Salina s a l t s ....................................... 87
Ordovician formations ............................
89
M e t h o d s .................................................. 92
General r e s u l t s ..........................................94
Niagara/Salina formation r e s u l t s ................. 97
Ordovican formation results ...................... 101
Isotopic r e s u l t s ................................... 105

TABLE OP CONTENTS (cont'd.).

Geochemical evolution of
Niagara/Salina b r i n e ............................... 105
C a - M g - S r ....................................... 107
P o t a s s i u m ...................................... Ill
Mass balance model
............................... 113
Enrichment in CaCl2 . . . . = > ..................... 119
Isotopic evolution . . ............................. 127
Model for the origin and evolution
of Niagara/Salina formation w a t e r s ................ 134
Origin of Trenton-Black River
formation brine
........................................ 137
St. Peter Sandstone w a t e r ...........
143
Isotopic evolution ..................................... 14 5
148
C o n c l u s i o n s ........................
Estimation of Br in s a l t ..........................
151
Appendix A. Statistical evaluation of Michigan Basin brines 152
This s t u d y ..........................
153
Univariate s t a t i s t i c s ....................................154
Data distribution . . . . . .
.................... 154
Average brine composition ......... .
158
Comparision with seawater ........................ 163
Formation comparision ............................. 166
Correlation c o e f f i c i e n t s ........................... 169
Multivariate s t a t i s t i c s ..................................172
Interpreting f a c t o r s ............................... 174
Past work
.......................... . . . . . 1 7 4
Factor patterns
...................................181
Q-mode methods and r e s u l t s ......................... 185
R-mode factor methods ............................. 186
R-mode results
........................... 187
Effect of pH v a r i a b l e .............................. 211
Effect of partialling out s a l i n i t y ................. 211
Effect of rotation method .......................
212
Interpretation of R-mode r e s u l t s .................. 213.
D i s c u s s i o n .............
216
C o n c l u s i o n s ...............................................217
Appendix B. M e t h o d s ........................................... 219
Brine s a m p l i n g ....................................... . . 2 1 9
Field a n a l y s i s .............
220
D i s c u s s i o n .............
221
Analytic m e t h o d s ...................
. . 223
Sample preparation
........................... 225
Ca, Mg, S r ...........
227
Na, and K
......................................... 227
Rb, Cs, L i .......................................... 227
vi i

TABLE OF CONTENTS (cont'd.).

S i l i c a .............................................. 228
B o r o n .............................................. 229
NH4N ................................................ 229
C h l o r i d e ............................................ 230
B r o m i d e ............................................ 230
I o d i n e .............................................. 231
S u l f a t e ............................................ 232
TDS and d e n s i t y ................................... 232
Oxygen and hydrogen i s o t o p e s ...................... 233
Strontium i s o t o p e s ................................. 235
Subsurface temperature and pressure ............. 2 36
Analytic Error-charge b a l a n c e .......................... 239
Error e s t i m a t e ..................................... 240
Appendix C. Analytic d a t a ............................... 248
Bibliography ............................................

viii

2 58

LIST OF TABLES

Table 1-1.

Chapter 1
Estimated formation volumes
in the Michigan basin.................

8

Table 1-2.

Summary of Devonian formation geology,
Michigan basin................................. 11

Table 1-3.

Components measured and analytic methods. . .18

Table 1-4.

Average composition of formation waters.

Table 1-5.

Apparent meteoric water compositions . . . .

Table 1-6.

Results of mixing example..................... 66

. .20
56

Chapter 2
Table 2-1. Average composition of Niagara/Salina and
Ordovician Formation waters, Michigan basin. 95
Table 2-2. Evolution of Niagara/Salina brine .........

115

Table 2-3. Bromide concentrations characteristic
of potash m i n e r a l s .......................... 12 3
Table 2-4.

Modeled brine derived from potash salt
compared with N/S brine and seawater . . . .124

Table 2-5.

Isotopic fractionation factors for
dolomite-water .............................. 131

Table 2-6.

Modeled mixtures of Trenton-Black River
brine and s e a w a t e r ........................... 141

Table A-l.

Appendix A
Results of chi-squared t e s t ................ 155

Table A-2.

Summary s t a t i s t i c s .......................... 159

Table A-3.

Student's t-test r e s u l t s ............

168

Table A - 4 . Correlation coefficient matrix of
log-transformed d a t a .........................171
Table A-5.

R-mode factor analysis results ............
ix

189

LIST OF TABLES (cont'd.)*

Table B-l.

Appendix B
Components measured and analytic methods.

Table B-2.

Variability in 180/160 and D/H . . . . . .

Table B-3. Analytic precision and sample comparision.

. 226
234
241

Table B-4. Glassware t o l e r a n c e ........................ 243

x

LIST OF FIGURES

Chapter 1
Figure 1-1.

Generalized tectonic map
of the Michigan Basin a r e a ..................... 5

Figure 1-2.

Paleozoic section of the Michigan basin

Figure 1-3.

Paleo-geography of the Dundee Formation,
from Gardner ( 1 9 7 4 ) ..........................

14

Sample locations .............................

16

Figure 1-4.

...

6

Figure 1-5. Specific gravity (g/cm3) versus production
elevation (m) in the basin. The elevation of
the Great Lakes is approximately 180m. . . . .22
Figure 1-6. Br (mg/1) and SO* (mg/1) in the
Traverse and Dundee formation w a t e r s ......... 24
Figure 1-7. Ternary diagram showing percentages of
Ca-Mg-Na
(mole percent) in Michigan
basin brines.....................................26
Figure 1-8. Traverse and Berea Formation water chemistry
(log mg/1) compared with evapoconcentrated seawater (dashed line, data
from McCaffrey et al, 1988; and Carpenter,
1978). Average Niagara-Salina formation water
shown as ( O ) A: log Cl-log Br, B: log Na-Br,
C: log K-Br, D: log Ca-Br, E: log Mg-Br,
F: log Sr-Br .................................... 27
Figure 1-9.

Dundee Formation water chemistry
(log mg/1)
compared with evaporating
seawater. Average Niagara-Salina formation
water shown as (O ) A: i°9 Cl-log Br,
B: log Na-Br, C; log K-Br, D: log Ca-Br,
E: log Mg-Br, F: log Sr-Br...................... 29

LIST OF FIGURES (cont'd.).

Figure 1-10. Richfield and Detroit River formation
water chemistry (log mg/1) compared
with evaporating seawater. Average NiagaraSalina formation water shown as ( O) .
A: log Cl-log Br, B: log Na-Br, C: log K-Br,
D: log Ca-Br, E: log Mg-Br, F: log Sr-Br . . 31
Figure 1-11.(A) Log MC12 (meq/1) vs. log Br (mg/1).
(B) Log MCI, corrected for charge balance
vs. log Br f m g / l )
Figure 1-12.
Figure 1-13.

Log MCI,

34

(meq/1) vs. log Cl (mg/1).......... 35

87Sr/86Sr of Michigan basin brines vs.
producing formation age.
Also shown is
the variation of the 8 Sr/86Sr ratio of
seawater during the Phanerozoic Eon (Burke
et al. , 1982).................................

38

Figure 1-14.

Log MCI, (meq/1) vs. log Br (mg/1) in
formation waters of W. Canada, Mississippi,
and Illinois basins.......................... 41

Figure 1-15.

Results of dolomitization model. Log Ca
(mg/1) vs. log Br (mg/1). Filled
symbols are measured Ca, open symbols
are predicted C a ............................ 4 3

Figure 1-16:

Histograms of saturation indices
(log lAP/Ksp) for Michigan basin
formation waters ............................

Figure 1-17.

46

(A) del D°/oo (SMOW) concentration scale,
and (B) del D°/oo (SMOW) activity scale,
vs. del 1 0°/oo (SMOW) of Michigan basin
waters.
Also shown is the best-fit line
to the data, and two examples of
seawater composition during evapo-concentration,
from Holser (1979) and Pierre (1982) . . . . 55

Figure 1-18. del aD°/oo (SMOW) vs. log Br (rag/1)
in all Michigan basin waters.
Also shown
is the best fit line (r2=.8) to the
Devonian formation data, and sample M56
from %Clayton et al. (1966)
................ 59

LIST OF FIGURES

(cont'd.)-

Figure 1-19. Possible mixing scenarios between
Dundee formation waters and AMMW.
(A) Log Cl (mg/1) vs. log Br (mg/1), and
(B) del D°/oo vs. del 180°/oo (SMOW) . .

65

Figure 1-20. del 180 °/oo (SMOW) vs. estimated
formation temperature.
Shown as a dashed
line are the del 180 values for water
predicted to be in equilibrium with calcite
of -1,
-5.75, and -8 °/oo P D B ............... 70
Figure 1-21. Plot of 1/Sr (mg/1) vs. 87Sr/86Sr of
the Michigan basin brines..................... 73
Chapter 2
Figure 2-1. Stratigraphic column of the Michigan basin . .79
Figure 2-2. Ternary diagram showing percentages of
Ca-Mg-Na (mole percent) in
Michigan Basin brines ........................

81

Figure 2-3. Diagenetic history of Niagaran reefs in
Michigan, from Sears and Lucia (1982) . . . .

84

Figure 2-4. Proposed model for freshwater flushing
of Niagaran reefs in Michigan, from
Gill (1977).
Recharge may have entered
the Niagaran rocks along marginal arches,
flowed down into the basin, and dissolved
salt from reefs along the basin margins.
. . 86
Figure 2-5. Salina A-l salt stratigraphy, after
Matthews and Egleson (1977) . . .

88

Figure 2-6. Sample location map.
St. Peter Sandstone
sample from (A), and the Case (1945) sample
from (B)

90

Figure 2-7. Ternary diagram showing percentages of
Ca-Mg-Na (mole percent) in Niagara/Salina
and Ordovician aged formations in
the Michigan b a s i n ............................. 96

x i ii

LIST OF FIGURES (cont'd.).

Figure 2-8. Niagara/Salina formation water (log mg/1)
compared with evapo-concentrating seawater
(dashed lines, data from McCaffrey et al.,
1988; and Carpenter, 1978).A: Log Cl-log Br,
B: Log Na-Br, C;
Log K-Br, D: Log Ca-Br,
E: Log Mg-Br, F:
Log S r - B r .................. 98
Figure 2-9. Log MC12 (meq/1)
vs. log br (mg/1) and
log MCI2 (meq/1)
vs. log Cl (mg/1) in
the Niagara/Salina and Ordovician formation
waters, compared with evapo-concentrating
seawater (dashed line, data from McCaffrey
et al., 1988; and Carpenter, r9TS7— ............102
Figure 2-10. Trenton-Black River and St. Peter
Sandstone formation water (log mg/1) compared
with evapo-concentrating seawater (dashed
lines, data from McCaffrey et a l ., 1988;
and Carpenter, 1978).
A: Log Cl-log Br,
B: Log Na-Br, C: Log K-Br, D: Log Ca-Br,
E: Log Mg-Br,
F: Log Sr-Br .
103
Figure 2-11. del D °/oo (activity scale) versus
del 180 °/oo SMOW in the Niagara/Salina
and the Ordovician formation waters,
Michigan basin. Also shown is the global
meteoric water line (GMWL) from Craig (1969),
and the best-fit line to all Michigan
basin waters collected as a part
of this study................................. 106
Figure 2-12. Result of dolomitization model (log Ca
vs. log Br, mg/1) for Niagara/Salina
formation waters. Square symbols are measured
Ca values, circles are predicted Ca. . . . .108
Figure 2-13.

Histograms of saturation indices
(log lAP/Ksp) for the Niagara/Salina
and Trenton-Black River s a m p l e s ........... 110

Figure 2-14.

Log K vs. log Br (mg/1) in seawater brines
from the Laguna Madre, Texas (Long and
Gudramovics, 1983) ........................ 112

xi v

LIST OF FIGURES (cont'd.).

Figure 2-15. Calculated del 180 °/oo (PDB) values for
dolomite and calcite in isotopic equilibrium
with Niagara/Salina formation waters at
subsurface temperatures of present-day,
present-day+23°C, and 80°C. Bars show range
of values from four different dolomitewater fractionation equations discussed in
text.
Also shown are ranges of del
O
reported in late diagenetic dolomites
and calcites, whole rock (W.R.), A-l
Carbonate dolomites, and Niagaran dolomites
and calcites in Michigan, from Cercone and
Lohmann (1987) and Sears and Lucia (1982) . 133
Figure 2-16. Br (mg/1) in Trenton-Black River formation
brines from the Albion-Scipio tre n d ......... 138
Figure 2-17. Cl-Br (log mg/kg) and relative
Ca-Mg-Na composition calculated to result
__£rom mixing of N/S sample #2099 (point 1)
with seawater concentrated to gypsum
saturation (point 0). Numbers
indicate seawater/brine ratio.............
Figure 2-18,

142

Calculated del -*-80 °/oo (PDB) values for
dolomite in isotopic equilibrium with
Ordovician formation waters at
subsurface temperatures of present-day,
present-day +23°C, and 80°C.
Also shown are
ranges of isotopic values for fracture, cap,
and regional dolomites in the Trenton-Black
River formations, from Taylor (1982) . . . .147
Appendix A

Figure A-l,

Histograms of Na and Cl (mg/1)
concentrations..........
. .

156

Figure A-2. Log-probability plots of Na and Cl (mg/1)
concentrations ...............................

157

Figure A - 3 . Average Michigan formation brine
composition (log mg/1) compared with
seawater (dashed line). Seawater data
from McCaffrey et al. (1988)......... ..

164

xv

LIST OF FIGURES

Figure

Figure

Figure

Figure

Figure

Figure

Figure

l—4.

(cont'd.)*

R-mode factor analysis results
(Varimax rotation) from Egleson and Quario
(1969) for Sylvania Sandstone formation
waters, Michigan Basin ......................

175

.-5. R-mode factor analysis results from Hitchon
et al. (1971), for West Canada basin waters.
TOP: varimax rotation, BOTTOM: bi-quartum
oblique rotation ............................

176

.-6. R-mode factor analysis results from Long
et al. (1986) for near-surface saline
groundwaters in Michigan ....................

178

-7. R-mode factor analysis results, case 1.
Only samples with pH variable are used,
pH variable included,
TDS not partialled out . ....................

199

-8. R-mode factor analysis results, case 2.
Ail samples are used, pH variable included,
TDS partialled o u t ........... ...............

202

-9. R-mode factor analysis results, case 3.
All samples are used, pH variable
not included, TDS partialled out ...........

205

-10. R-mode factor analysis results, case 4.
All samples are used, pH variable not
included, TDS not partialled out .........

208

Appendix B
Figure

Figure

-1. Alkalinity titration curves, pH vs
volume of acid added (ml). Curve A,
Richfield brine sample, curve B,
Niagara/Salina formation sample,
curve C, a near-surface ground
water from Michigan..........................

224

-2. Formation temperature vs. production
elevation (meters). Also shown is a
geothermal gradient of 23°C/km, starting
at 10oc at 33m.
Data from Vugranovich (1986)...............

238

xvi

CHAPTER

Geochemistry of water in Devonian aged formations,
Michigan Basin.
INTRODUCTION
The origin and evolution of sedimentary basin brine has
been

the

1933;

focus

Case

of

et

study

a l .,

White, 1965; Torrey,
1966b;

Collins,

Carpenter,
chemistry
interest

of

1942;

De

number
Sitter,

Collins,

Hanor,

these

because

a

of

years

1947;

White,

1982).

unique
of

1975;
Besides

fluids,

their

basin

waters

involvement

in

1985;

base-metal mineralization

1982),

1970;

(Land,

diagenesis (Friedman and Sanders,

origin

and

1967;

Billings

Carpenter et a l . , 1974;

Cercone and Lohmann,

1987).

evolution

of

1964),

Domenico and Robbins,

(Roedder,

evaporite-cycling

are

the
of

hydrocarbon

basin hydrology

(Bethke,

1978;

understanding

(Degens and Chilingar,

Bush,

1957;

Carpenter,

production and migration

1969;

(Russell,

1966; Dickey, 1966; Graf et al., 1966a,

1975;

1978;

for

deep

1985),
et al.,

Long and Angino,

1987),

and

sediment

1967; Bein and Land, 1983;

In spite of this interest,
water

in

many

basins

the

remains

unresolved.
Any
chemical

model
and

for

brine

isotopic

origin

must

geochemistry

of

explain
the

both

the

water.

The

chemical evolution of sedimentary basin brine is thought to
reflect

two,

(Lerman,

1970).

albeit

overlapping,

groups

of

processes

First are the concentrating processes that

are responsible for the high salinity of these waters,
include

evaporation,

evaporite

1

dissolution,

and

and

shale

membrane

filtration.

Second,

are

the modifying

processes

that occurred during the evolution of the water, and include
dissolution-reprecipitation,
activity,

mixing

of

waters,

processes may mask the

ion
and

exchange,
shale

chemical

biological

filtration.

record of

These

formation water

origin (Chave, 1960).
The isotopic

evolution of sedimentary basin water may

also occur by two general processes.
proposed

that

reflects that
rocks,

isotopic

meteoric

composition

water

has

of

flushed

(1966)

basin

water

through

basin

and was subsequently modified by exchange with

minerals
al.

the

Clayton et al.

(Clayton et al.,

(1964),

et al.,

1966).

Knauth and Beeunas

1969? Hanor,

composition of

some

Alternatively,
(1986),

rock

Degens

and others

et

(Hitchon

1982), have proposed that the isotopic
formation water

reflects

mixing,

more

specifically, the dilution of evapo-concentrated seawater by
meteoric water.
isotopic
these

Hicchon

et a l .

(1969)

values of basin waters

processes

including

suggested that the

reflects

mixing

of

a combination

marine

with

of

meteoric

waters, exchange with rocks, and shale membrane filtration.
Several studies have demonstrated these general models
and

have

linked

(Carpenter,
Beeunas,
(Billings

1978 ?

brine

Stoessell

1986, Dutton,
et

a l .,

chemistry

1969,

and

1987),
Dressel

to

Moore,
mixed
and

residual
1983;

seawater

Knauth

and

seawater-freshwater
Rose,

1982;

Spencer,

1987), and freshwater origins (Clayton et a l ., 1966; Hitchon
et al., 1971; Kharaka and Berry,

1974? Kharaka et al., 1973;

3
Bassett

and

isotopic

Bentley,

evolution

1983).
of

However,

formation

the

water

chemical

in

many

and

areas,

including the Michigan basin, remains unresolved and may not
fit these models.
inability

to

One principle reason for this may be the

explain

both

the

chemical

and

isotopic

geochemistry of brine by a single m o d e l .

THIS STUDY
The

Michigan

understanding
Michigan
Similar
brine,

of

other

although

Michigan

of

shallow depths
have

been

Beecker,
Sorensen
al.,

al.

for

1966),

Segall,

evolution,

basin

because

time

(Lane,

Ca-Cl
The

as

are

they
solids

have
they

Egleson

1899;

in

high

Ca

exist

at

Querio,

al.,

1966;

Clayton

et

a l .,

1966b),

their

evolution

Cook,

and

1975;

and

is

Although the Michigan brines

1945;

et

because

brine.

dissolved

1945) , they

the

basins.

water

however,

total

and

in

sedimentary

unique

(Case,

some

Case,

origin

basin

role

contain Na-Cl

having

in the basin.

Graf

important

Michigan

somewhat

mg/1

et

1940;
and

and

formations

some

studied

1966a;

isotopic

basins,

640,000

(Graft

an

inter-cratonic

are

saline,

has

brine

some

brines

extremely

content

both

typifies
to

excess

basin

rem ins

1914 ;
1969;

Graf

geochemical
to

be

et
and

clearly

identified.
The goal of this study is to determine the geochemical
and

isotopic

chemical and
54

oil

evolution of the Michigan basin brines.
isotopic composition of brines

wells

from

6

Devonian

producing

collected
formations

The
from
in

Michigan are used,

in combination with brine chemistry from

Niagara/Salina and Ordovician aged formations in the basin,
which are discussed further in Chapter 2.
processes

for

brine

evolution

evapo-concentration,
rock

that the brine

considered

shale membrane

interactions.

A

tenant

in each

The geochemical
include

filtration,

followed

and

in this

formation may have

seawater
water-

research

evolved

is

through

different processes,

therefore,

the brine chemistry must be

evaluated

possible)

on

(whenever

an

individual

formation

basis, rather than collectively in a whole basin manner.
It

became

agreement

evident

exists

between

evapo-concentrated
unexpected,
deposits.

as

early

in

the

this

study

that

Michigan

brine

chemistry

seawater.

the

basin

This
is

noted

Based on this observation,

evapo-concentrated

seawater

brine,

reactions

in

study

is

used

this

is
for

a

not
its

close
and

totally
evaporihe

a framework model of

modified
(Wilson

by

and

water-rock
Long,

1986,

Wilson and Long, 1987) .

STUDY AREA
The

Michigan

sedimentary basin,
basin

is

shown

a

mature,

intercratonic

between

the

Canadian

Shield

and

The
the

and is outlined by the arches and platforms
Paleozoic

sediments

exist at the deepest point in the Michigan basin,

and range

from

in

is

occupying an area over 2 00,000 km 2 .

located

Illinois Basin,

Basin

Figure

Jurassic

Paleozoic

1-1.

to

section

Over

Cambrian
starts

4000m

in

with

of

age

basinal

(Figure

1-2).

Cambrian

The

sandstones

CANADIAN SHIELD
p-G

CHIGAN BASIN
V '

M

i

ILLINOIS
BASIN

Figure

1-1 : G e n e r a l i z e d
area .

M

P

GENERALIZED
TECTONIC MAP
OF THE
MICHIGAN BASIN

CINCINNATI
ARCH |

tectonic

map

AREA

of

the M i c h i g a n B a s i n

6

MICHIGAN BASIN
E LLS W O R TH 81.

PERIOD

SUBSURFACE
NOMENCLATURE

■
_

"~ r"*
ANTRIM 3M.

JURASSIC

TRAVERSE 0ROUP

M A R S H A L L S3

DUNDEE LM S.

6000'

®

f ~ ~ D E f R O I T R.

BELL SH

AMHERSTBURO

3 Y L V A N IA SS.
10,0 00 '
-A -.

SILURIAN

BOIS B LA N C

CARB

BASS IS L A N D

C A B O T H E A D SM

U TIC A SH

1 5 ,0 0 0 '

C A M B R IA N

of
Figure

!_2

;

P a le o z o ic

s e ctio n

the

M ich ig an

b asin

7
overlain by Ordovician

limestone-dolostones,

which

are

the

oldest rocks to exhibit the basin shape (Nunn et a l ., 1934).
Over 550m of Silurian evaporites and carbonates follow and
include

dolomite,

anhydrite,

halite,

and

Devonian aged carbonates and evaporites,
paper,

follow

and

are

overlain

by

potash
the

a

minerals.

focus of this

thick

layer

of

Carboniferous to Jurassic aged shales and clastic sediments.
The basin
sandstones,

is estimated to contain 47% carbonates,

18% shales,

and Landes, 1955).

and 12% evaporites by volume

23%

(Cohee

Estimated present-day formation volumes

(Table 1-1), measured from maps in Gardner (1974) and Curran
et al.

(1981),

indicate that

sediment exists in the basin.

7.4xl04 km3 of

Devonian

aged

Assuming an average porosity

of 10% and neglecting evaporite layers,

the Devonian rocks

in Michigan contain approximately 3.5xl016 liters of water.
The

present

approx imately

geothermal

22°C/Km

(Nunn

gradient
et

al.,

in
1984),

the

basin

but

measured

temperatures may vary considerably from this gradient
Appendix

B).

Hogarth

(1985)

used

demonstrate that the paleo-geothermal
to the

present day gradient,

but that

conodont

(see

color

gradient was
Paleozoic

is

to

similar

sediments

may have been buried almost 1 km deeper in the past.

Other

studies have suggested both deeper burial and higher paleogeothermal gradients for Michigan (Cercone,1984).

8

TABLE 1-1
Estimated formation volumes in the Kichigan Basin.

AGE&
REF.

FORMATION OR
GROUP________

Km

SAMPLED

xlO^

ABSOULTE AGE
END DURATIO
X l O ” XlO^

(1) CARBONIFEROUS
GRAND RAPIDS GROUP
BAY PORT LMS
MICHIGAN FM
UNCONFORMITY
MARSHALL SANDSTONE
COLDWATER SHALE
SUNBURY SHALE
BEREA SANDSTONE
BEDFORD SHALE

UNCONFORMITY

46

376

23

399

16

0.46
1.72
0.08
0.08
0.14

(1&2) DEVONIAN
ELLSWORTH SHALE
ANTRIM SHALE
TRAVERSE GROUP
TRAVERSE LMS
BELL SHALE
NATIONAL CITY GYPSUM
DUNDEE FORMATION
ROGERS CITY
REED CITY ANHYDRITE
LUCAS or
DETROIT RIVER
HORNER MEMBER TOTAL
HORNER SALTS
IUTZI MEMBER
MELDRUM TOTAL
RICHFIELD MBR.
FILER SANDSTONE
BOIS BLANC-SYLVANIA

300

1.42
1.20

1.31
0. 04
0.66

0.14
09
53
32
48
18
0.89

9

Table 1-1 (cont'd.).

AGES
REF.

FORMATION OR
GROUP

Km3
XlO4

SA
SAMPLED

ABSOULTE AGE
END DURATIO
X lO °

(1)SILURIAN
BASS ISLAND
SALINA GROUP
G UNIT
F EVAPORITE
E EVAPORITE
D EVAPORITE
C CARBONATE
B EVAPORITE
A-2 CARBONATE
A-2 EVAPORITE
A-l CARBONATE
A-l EVAPORITE
BROWN NIAGARAN

0.08
1.46
0.27
0.10
0.17
0.91
0.35
0.84
0.34
0.55
8.79

(1) ORDOVIVIAN
UTICA SHALE
PRARIE DU CHEIN

0.96
8.88

xlOc

415

26

411

21

.88

*

*

★

UNCONFORMITY

Key:
1: Curran et al. (1981).
2: Gardner (1974).
3: End and duration of geologic periods in Michigan
reported in 106 years, From Nunn et al., (1984).
SAMPLED: * indicates formations sampled in this study.

10

Formation waters
Sandstone,

the

reported

Traverse

on here

Group,

the

are

from the

Dundee

Formation,

Detroit River Sour Zone and Richfield Member
Formation,

and the

summary

the

of

thickness,

Sylvania

general

and

Sandstone

lithology,

depositional

environments

be

found

Eschman,

in

Gardner

(1974),

rocks

transgressive-regressive
restricted
1974).

Matthews

in Michigan

marine

2) .

textures,

for

Devonian

of

Door

of

the

Devonian

in

and

(1984).
a

that
the

shallow

water

followed

the

Silurian

(Gardner,

The Kaskaskia unconformity is considered the

boundary

A

and

(1977),

represent

sequence

conditions

1-

Lucas

Further descriptions can

(1981), and Montgomery et al.

Devonian

the

of the

(Figure

structure

formations is given in Table 1-2.

Berea

Michigan

lower

(Gardner,

1974).

Overlying this unconformity is the Sylvania Sandstone which
represents the start of a transgressive stage of the Early
Devonian

(Gardner,

fluvial

1974) .

transported

currents

sand

The
which

and

was

deposited

carbonates

of

the

Bois-Blanc

(Gardner,

1974).

The

cyclic

Sylvania
was

is

reworked

concurrent
and

a

by

with

Amherstburg

evaporite-carbonate

wind

and

marine
offshore

Formations

sediments

of

the

Amherstburg and Lucas Formations were then deposited during
the Middle Devonian when the basin became restricted.

The

cyclic deposition is best developed in the Richfield Member
of

the

Lucas

Formation,

alternating anhydrite

which

consists

and dolomite

layers.

of

over

60m

of

In the central

11

TABLE 1-2
SPMMARY OF DEVONIAN FORMATION GEOLOGY
STRATIGRAPHIC
UNIT & MAX.
THICKNESS
Traverse
Group
800'

LITHOLOGY

STRUCTURE
TEXTURE

Bio-calcarenite
bioherms, biostroms
limestone, few
wackestones and
anhydrite and
grainstones
and porous
dolomites on
west margin,
gray shale facies on
east margin.*

Dundee
dark limestone with irregular beds and
Formation laminar,impure
banks, bioturbated
400'
anhydrite and
wackestones and
porous secondary
grainstones
dolomites on west.*

ENVIRONMENT
OF DEPOSITION
shallow shelf
with muddy influx
from east. Lagoons
on west margin

subtidal to shallow
shelf, with lagoons,
and shabkas on basin
margins

Horner
Member
800'

massive salt with
interbeds of
anhydrite and dark
sour carbonates

impure dolomitic and
layered anhydrites,
cloudy-clear salt
layering

peripheral shabkas
and lagoons, and
basin centered
evaporites

Iutzi
Member
150'

massive anhydrite
at base, dolomite
and some limestones
wide spread at top

massive and impure,
dolomitic anhydrites
and dark micritic
carbonates

periphal shabkas and
penesaline lagoons

Richfield syngenetic dolomicrite with
Member
100 '
interbedded,
displacement
anhydrites.

oolitic, laminated
mudcracks, decussate
interclasts, nodular
mosaic anhydrites

shabka and
penesaline lagoons

Amherstburg
Member
300'

Dark, carbonaceous
wackestones

Bioherms, biostromes
shelf to basin,
bioturbate, bioclastics, increasingly
and pelletal
restricted in
upper portion.

Sylvania
300'

White quartzose
sandstone, with
secondary quartz
overgrowths,
interbedded with
carbonates.

cross beds, planar
to irreg. bedding,
dessication cracks,
frosted and polished
grains

dunes, beach, and
bar, open marine
along hinge line

12

TABLE 1-2 (cont'd.).
STRATIGRAPHIC
UNIT & MAX.
THICKNESS
Bois Blanc
700'

LITHOLOGY

STRUCTURE
TEXTURE

Cherty bioirregular to thin bedded
calcaranite,
mottled wackestone
tripolitic chert

ENVIRONMENT
OF DEPOSITION
open marine
carbonate shelf
silica from
adjacent desert
land mass

* Carbonates are generally limestone on east, and dolomite on south and
west basin margins. They are darker and finer grained in basin center.
SOURCE: Gardner, 1974

13
basin

area

the

found and

Detroit

includes

River

basin

deposit

carbonates

margins

as

a

underlying Salina

(Lucas

the Horner Member,

halite-anhydrite-dolomite
Detroit River

Group

are

result

salts

Formation)

a up

to

(Matthews,

later

thick

1977) .

highly brecciated
of

2 00m

The

around

dissolution

(Landes et a l ., 1945).

is

the

of

the

After

the

deposition of these evaporites and carbonates, a slow return
of

normal

marine

conditions

occurred

Traverse sediments were deposited.
of a biostromal
found

in

off-shore

extensive
margins
shales

shelf

sabkhas

during
cap

deposition

the
that

and Eschman,
The

areas.
and

this

anhydrite
margins,
also

1974;
1981;

existed

and

through

around

the

the

carbonates

suggested

1-3).

signal

formation

the

Upper
onset

that
basin

Devonian

of

clastic

Carboniferous

Matthews,
Montgomery

especially
in

mineralogy

but also includes

chert,

generally

occur

Dundee

(Door

1981).

sandstone,

Eschman,

(1974)

(Figure

and

continued

Devonian

(Gardner,

with argillaceous

lagoons

section

the

Both formations consist

Gardner

time

calcite and dolomite,
quartz

facies,

and

and

increase
towards

lenticular

(Montgomery, et al., 1984).

dominated

anhydrite,

Nowak,

al.,
in

the

1984).

abundance
west

zones

1978;

shales

Door

and

Dolomite
near

(Gardner,

throughout

the

and

basin

1974),
the

by

halite,

illitic-chloritic

1977;
et

is

but

basin

14

DUNDEE

LAGOON
OPEN MARINE

Figure

1-3.

P a l e o - g e o g r a p h y of the
after Gardner (1974).

Dundee

Formation,

15

Oil

production

from Michigan

Devonian

rocks

has

been

important since the early 1900's and occurs in a broad band
extending
(Cohee

from

and

the

central

Landes,

basin

1955;

westward

Montgomery

and

et

southward

al.,

1984).

Hydrocarbon production occurs from bioherms and anticlines,
dolomitized

porosity

zones,

and

from

fracture

related

porosity (Montgomery et al., 1984).
ANALYTIC METHODS
Fifty-four Devonian formation brines were collected as
a

part

of

a

included

waters

formations.
obtained
including

larger

from

on

Michigan

Silurian

basin

and

brines

which

Ordovician

aged

The samples were supplemented with 32 analyses

from

various

Michigan

(M.D.N.R.) open
Figure 4.

study

oil

company

Department

file data.

and

of

Sample

government

Natural

locations

files

Resources

are shown

Charge balance errors are less than 5%.

in

Sampled

intervals were confirmed by operating personal and drilling
records,

and

by

geophysical

plugging by salts
such,

recently
those

flushing

of

possible.

Well

in Michigan and

oil-wells

is

a

as

common

Care was taken to avoid sampling wells that were
flushed,

located

long use

when

is a serious problem

freshwater

practice.

logs

wells

near

for brine

brine

from

fields

disposal

under

wells.

disposal, Sylvania

water-flood,
Because

or

of

its

Sandstone brine

was

not collected in this study although some data were obtained
from M.D.N.R.

files.

•D U N D E E

• TRAVERSE

□ RICHFIELD

oBEREA

• D E TR O IT R.
Figure

1-4:

Sample

locations

17
Samples

were

collected

directly

from

the

well

head,

filtered through glass wool and Watman #1 paper filters, and
stored

in pre-rinsed plastic bottles.

analysis were collected by diluting

Samples

raw brine

HN03 , using Class A volumetric pipettes.
were

collected

for

anion

formeldehyde for S04 .
D/H,

180/160,

analysis

for cation
50% with

5%

Undiluted samples

and

preserved

with

Untreated samples were collected for

and

87Sr/86Sr

analyses.

The

pH,

Eh,

temperature, and when possible, alkalinity, were measured in
the

field

Collins

following

Lico

et

(1975) .

NBS buffers

standardization.

Formation

a l ., (1982), Wood (1981),

and

(pH 4 and 7)were used for pH
temperature "> were

calculated

following Vugrinovich (1986) using 10°C at 33m plus 23°C per
each km of depth thore-after.
laboratory

procedures

used

The components measured and
are

listed

in

Table

Inorganic analysis was done at the Geochemical
Michigan State University.

1-3.

Laboratory,

Stable isotope ratios

(180/ 160,

D/H) were measured at the Environmental Isotope Laboratory,
University

of

Waterloo,

Laboratories, New Jersey.
equilibration
distillation
Deuterium

with

Sofer and Gat
Argonne

were

(1975).

National

and

at

Teledyne

180/160 analysis utilized 72 hour

C02 , and

technique

values

Canada,

D/H

analysis

described in
corrected

Fritz

for

et

a
al.

activities

87Sr/86Sr analysis

Laboratory.

used

was

Standard

(1986).
following

preformed

at

chromatography

techniques using 2N double distilled HCL and Dowex
resin (Stueber et al., 1984) were used to

complete

(8x-2 00)

18

TABLE 1-3
Components measured and analytic methods.
COMPONENT
PH
Ca, Mg, Sr
Na, K
Rb, Cs
Li
Cl
Br
I
B
Si
n h 4n

so4
Alkalinity
Density
TDS

METHOD
electrometric
flame emmision
w / 1 :10 of 87g/l
LaClo
flame emmision
w/1:10 of 25.4g/l
NaCl or KC1
flame emmision
w/1:10 of 25.4g/l
Na-K-Cl
flame emmision
w/1:10 of 25.4g/l
Na-K-Cl
Mohr titration
colorimetric
bromide oxidation
colorimetric
w/carminic acid
colorimetric
following extraction
potentiometric titration
gravimetric
potentiometric titration
pyconometer @25°C
and by calculation
SPG=log TDS * 7, 102x10 -7
calculation

DILUTION
none
1:2000

REFERENCE
1
1

1:2000
1 :26

1:200

none
1:40
1:2
1:10

1
2
1
1

1:200

1:2
none
none
none

4
5
1
5

+ 0.996

References:
1: Brown et al.,
(1979) Methods for Determination of
Inorganic Substances in Water and Fluvial Sediments.
U.S.G.S. Water Resources Investigation Book 5, Chapter
A-l, Washington, D.C.
2: Presely, B.J. (1971), Part I: Determination of selected
minor and major inorganic constituents, in: Ewing, J.I.
and others, Initial Report of the Deep Sea Drilling
Project, Volume VII, Part 2: Washington, D.C., U.S.
Govt. Printing Office, p. 1749-1755.
3: Schrink, D.R., (1965) Determination of silica in sea
water using solvent extraction, Anal. Chem., 37,
764-765.
4: Collins, A.G., Cassaggno, J.L., and Macy, V.W. (1969)
Potentiometric determination of ammonium nitrogen in
oil-field brines, Environmental Science and Technology,
3, 274-275.
5: A.P.I. (1968) Recommended Practice for Analysis of
Oil-field Waters, American Petroleum Institute,
Washington, D.C., 2nd edition, 58p..

19

separate Sr. Rb/Sr ratios are <0.001, suggesting that Rb has
not appreciably enriched 87Sr.

GEOCHEMICAL RESULTS
Sample locations,
all

samples

Appendix

collected

of

this

ranges, and

well data,
in

and analytic results

this

report.

A

concentration

study

are

summary

ratios

are

of

listed

in

geometric

listed

for
the

means,

in Table

1-4.

Total dissolved solids in all samples are in excess of 300
g/1.
Sulfate,

alkalinity,

and pH values reported

waters deserve some discussion.

for these

Sulfate concentrations are

very low, much less than seawater values, with some samples
having non-detectable S04 .

Alkalinity

is highly variable,

ranging from undetectable values up to several hundred mg/1.
Several

factors may cause errors

measurements.
Na

For example,

poisoning

may

liquid junction potentials

introduce

electrode measurements made

considerable

1984).

and

uncertainty

in high salinity water

and Pytkowicz, 1973; Millero,
Dickson,

in the pH and alkalinity

in

(Hawley

1979; Harvie and Weare,

1980;

Alkalinity may be also affected by borate

and organic acids (Willey et al., 1975).

Measurement of the

organic

has

acid

successful
personal

because

comm.).

found

during

value

of

acids

content

3.5,

of

of

salinity

However,

alkalinity
perhaps

(Appendix

these

B).

a

waters

interferences

slight

titrations

indicating
Eh

not
(B.

inflection
near

the

the

of

been

Fisher,

point

apparent

presence

measurements

yet

of

these

was
pH

organic
waters

TABLE 1-4
Average composition of formation waters
COMPONENT

BEREA

TRAVERSE

DUNDEE

RICHFIELD DETROI'

Cl
Br
Ca
Mg
Na
K
Sr
Rb
Li
B
Si

195000
1400
451C0
8250
63000
640
2000
4.
8.
3.
74
12
45
23
315000
2980
3

172000
1100
29000
6600
68200
1660
1060
4.
29.
29
3.
125
50
65
11
282000
1840
30

171000
1000
24500
5100
71000
1640
780
3.
24.
21
3.
110
30
180
10
278000
1660
40

194000
2100
65200
8700
34200
7200
1970
12.
46.
119
2.
230
160
1200
23
296000
4020
13

n h 4n

hco3
so4
I
TDS
MC12
n

212000
3400
88000
11800
23100
13400
2680
38
93
226
2
550
180
10
38
355000
5420
2

Key: All values are geometric means in mg/1.
MC12 = Ca + Mg + Sr - 0.5HC03 -S04 as meq/1
n = number of samples analyzed for major components

21

resulted in highly unstable values that drifted considerably
with time.

Eh most likely

and Crooper,
Kharaka

et

1959;

al.,

controlled by dissolved Fe

Back and

1980;

Barnes,

Stumm

1965;

and Morgan,

Langmuir,
1981)

(Hem
1971;

because

in

non-treated samples, minerals thought to be Fe-oxide or FeS
(depending on formation) would precipitate after a few days
time.
DISTRIBUTION OF BRINE CHEMISTRY
Formation

waters

are

often

density

or

chemically

stratified in basins, a characteristic that may reflect mass
movement

or

addition,
reflect

diffusion

(Land,

spatial

variations

different

geochemical

areas of a basin.

1987;
in

Han o r ,

brine

processes

1984).

In

chemistry
operating

might
within

A summary of the distribution of brine

chemistry in the Michigan basin is given here.
Figure 1-5 shows the brine density
versus

production

elevation

for

(specific gravity)

all

formation

collected in this study, including those
Ordovician
salinity,

formations.

saline

in Michigan

sea-level, and
brine

surface.

brine

from Silurian and

density,

and

therefore

show a only a slight linear increase with depth.

Surface elevations
above

The

waters

exists

as
at

this
shallow

range between
diagram
depths

200

and

shows, very
(<500m)

300m

dense,

below

land

Although the Michigan basin waters appear density

stratified,

they

increase

salinity

in

do

not
with

demonstrate
depth

the

strong

characteristic

of

linear
other

sedimentary basins such as Illinois (Graf et al., 1966), and

(m asl)
ELEVATION
PRODUCTION

0-

.

-

?

Oo<

A

BEREA
• TR AVERSE
o DUNDEE
m RICHFIELD
A DETR O IT R.
□ N IA G A R A /
SALINA
O O R DO VICIAN

00

-1000

■

° s > m wa

■

a

J

□

-2000 tj_
1.0

A

I.!

1.2

1.3

1.4

SPECIFIC G RAVITY
3

Figure

1- 5 : S p e c i f i c g r a v i t y ( g / c m ) v e r s u s p r o d u c t i o n
e l e v a t i o n (m ) in the b a s i n . T h e e l e v a t i o n of
t h e G r e a t L a k e s is a p p r o x i m a t e l y 1 8 0 m .

23

the West

Canada

sedimentary basin

(Hitchon

et

al.,

1971).

This probably reflects the structural shape of the basin and
the

fact that

samples

from different

formations

collected in a single location in the basin.
formation water,

cannot be

The Niagaran

for example, was sampled only from near the

basin margins (Chapter 2).
Bromide can be used to represent spatial variability in
salinity and chemistry of the formation waters.

Bromide is

selected because the graphical treatment presented below is
based

on

plotting

brirs

chemistry

versus

Br,

and

because

bromide is strongly related to salinity (log TDS = 0 .357 * log
Br + 4.33, r2 = .85)
major elements
(Rb,

B,

and

(Cl,

and covaries directly with many of the
Ca, Mg,

K, Sr,

I) , deuterium,

distribution, in

a general

and Na) , minor elements
1 ft
0.

and
manner,

Therefore,

summarizes

the

the

Br

spatial

variation of the brine chemistry within the basin.
Bromide concentrations
data

increase

towards

similar distribution
waters,

but

the

the

in the combined Traverse-Dundee
basin

center

is suggested
limited

for

data

interpretations from being made.

(Figure

the

other

prevent

1-6).

A

formation
conclusive

It appears that the brine

chemistry is related to location and

formation depth, with

higher concentrations occurring in the deeper, central-basin
areas.

Sulfate does

(Figure

1-6),

found

follow this

but generally

in samples

basin margins.

not

collected

higher
from

general

distribution

S04 concentrations

shallower

depths

near

are
the

TRAVERSE ft

BROMIDE

.DUNDEE
\ • SAMPLE
.> l o c a t io n

Figure

1-6:

Br (mg/1) and
T r a v e r s e and D u n d e e

] g / l ) in t h e
so
formation waters.

25

MAJOR ION COMPOSITION
Chloride is the dominant anion in all the samples,
as Figure 7 shows,
The Berea,

either Na or Ca are the dominant cation.

Traverse,

and Dundee brines are Na-Ca-Cl brine,

while the Richfield,
are Ca-Na-Cl

and

water,

Sylvania,
similar

and Detroit River Sour

to many

of the

Zone

Niagara/Salina

formation brines (Chapter 2).
Collins

(1975)

and

Carpenter

(1978)

demonstrate

how

formation water chemistry can be compared with evaporating
seawater
origins
brine

in

order

to

and evolutions

studies

(Carpenter,

have

1979;

and

Rose,

1982;

The

major

ion

Figure 8 to

make
of

interpretations

brine

since

to

chemistry.

utilized

Rittenhouse,

Br/Cl

for

1967;Collins,

Spencer,

1987;

chemistries

are

10.

as

Dutton
compared

The trend lines

1987;

possible

A number

of

this

purpose

1975;

Dressel

and

with

others).

seawater

for evaporating

in

seawater

chemistry are from McCaffrey et a l . (1988), which agree well
with

Black Sea data

of Zherebtsova

and Volkova

(1966)

and

Carpenter (1978). Also shown are the average compositions of
the Niagara/Salina formation water samples from (Chapter 2).
The

B r ,C l , and

concentrated

Na

in

seawater, while

and Sr highly enriched over

the

brines

compare well

with

Mg and K are depleted and Ca
the seawater trend.

A strong

linear relationship exists between log Ca and log Br in the
Traverse and Dundee samples (slopes of 0.80 to 0.97, r2=0.77
to 0.78,

respectively)

r2=0.90 to 0.95).

and log Sr-Br (slopes of 1.0 to 1.2,

Magnesium also plots linear with log Br,

B: BEREA
T : TRAVERSE
D: DUNDEE
RF= RICHFIELD
DT:DETROITR.
SY:SYLVANIA
N /S : NIAGARA/
SALINA
0 : TRENTON/
BLACK RIVER

REDT.SY

Na

Co

Figure

1-7 : T e r n a r y d i a g r a m s h o w i n g p e r c e n t a g e s of C a - M g - N a
( m o l e p e r c e n t ) in M i c h i g a n b a s i n b r i n e s .

I

to

(Ti

27

Figure

1-8.

T r a v e r s e and B e r e a F o r m a t i o n w a t e r c h e m i s t r y (log
mg/1) compared with e v a p o - c o n c e n t r a t e d seawater
( d a s h e d l i n e , d a t a f r o m M c C a f f r e y et al , 1988 ;
and
Carpenter,
1978).
Average
Niagara-Salina
f o r m a t i o n w a t e r s h o w n as ( O ) A: l o g C l - l o g B r ,
B: l o g N a - B r , C: l o g K - B r , D: l o g C a - B r , E: l o g
M g - B r , F: l o g S r - B r .

28

TRAVERSE
& BEREA
5.S

B
u
O 4-5
o

•TR A V E R S E

J 4

A BEREA

_____L

J

8.5

B

o

5

Z

*>\

0 .4 .5

o

\

4

4.5

C

©

* 3.5

o>
© 3
-1

AA

2.5
5.5
o

5

«_>

O 4.5

•if*

-I

4.5

5 4
o>
J 35

© 3.5

•ii

V)

o>

*

• ••

j 25

1.5

_L
2

2.5

3

Log B r (m g /!)

Figure

8.

3.5

29

Figure

1 -9 .

Dundee
Formation
water
chemistry
(log
mg/1)
compaed
with
evaporating
seawater.
Average
N i a g a r a - S a l i n a f o r m a t i o n w a t e r s h o w n as ( O
) A:
l o g C l - l o g B r , B: l o g N a - B r , C: l o g K - B r , D: l o g
C a - B r , E: l o g M g - B r , F: l o g S r - B r .

30

DUNDEE
8.6

-----

•
e

s'

o
o
J

4
8.6

o B
Z
o

-I
4.5

-'" V

4
''

y: 3.5

u»
O

• •»«•••
% •

3

2.5

4.5

o
o
D» 4
O
_l

3.5

Z X jl

4.5

I

4

o>
3.5
^

3 5h
F

e.
E

3

•w
o
J_______ L

1.5

2

25

3

Log Br (mg /l)

Figure

9.

35

31

Figure

1-10.

Richfield
and
Detroit
River formation water
c h e m i s t r y (logmg/1) c o m p a r e d with e v a p o r a t i n g
s e a w a t e r . A v e r a g e N i a g a r a - S a 1 lna f o r m a t i o n w a t e r
s h o w n as (
)
.A: l o g C l - l o g B r , B: l o g N a - B r ,
C: l o g K - B r , D:
l o g C a - B r , E: l o g M g - B r , F: l o g
Sr-Br.

o

32

RICHFIELD
& DETROIT RIVER

0.5

03

4.5

4.5
AA

5.5

? 4.5

4.5

3.5

• RICHFIELD
3.5

A DET. R.

2.5

1.5

2.5

3.5

L oq Br lm g/1)

Figure

10.

33

but best fit lines have slopes between 0.4 and 0.5

(r2=0.76

to 0.79) .
Carpenter

(1978)

suggested

that

a

further

test

of

formation brine origin would be to compare MC12 with Br and
MCI 2

Cl.

represents

the

sum

of

divalent

cation

charge

balanced by Cl, MC12= Ca + Mg + Sr - 0.5*HC03 - S04 (meq/1).
This value
carbonates,

is not affected by dissolution-precipitation
dolomite,

approximately

sulfates minerals,

equals

Br

saturation is reached,

in

seawater,

or halite.

and

of

MCL2

until

carnallite

follows the relationship:

Log MC12 =

Log Br + 0.011 (Carpenter, 1978).

with

Log MCI2 is compared with log Br in Figure 1-lla,

and

Cl

but

in

Figure

1-12.

The

above, the MCl2-Br line

data

plot

for seawater,

parallel

and

suggest

to,
that

an

excess of divalent cations or a depletion in Br exists from
expected seawater values.

Because the MC12 value

includes

the combined error in measuring five components versus the
analytic
charge
analyses

error

in

balance

by

measuring
adjusting

B r , MC12
for

was

excess

corrected
charge

the

(MC12 '= MC12-(EPM cations)+ (EPM anions)), which is

shown in Figure 1-llb.

An excess of MC12 is still evident

in spite of this correction.

The log C1-MC12 relationship

(Figure 1-12) gives some insight into this excess.
similar to the Cl-Br plot
the

in

for

MCL2-C1

plot

in matching the

apparently

evapo-concentration than

reflects

seawater trend,

higher

are predicted by

Although

degrees

Cl-Br.

This

most evident for some Richfield-Detroit River brines,

of
is

which

34

MICHIGAN BASIN
i—

—

—

log

MCL2 (meq/D

a ®

SEAWATER TREND

L OG Br (mg/I)

M C I2

(m e q /l)

4 -------1-------1------- 1-------r
y

y

y

yO
y

25

y

% DEVONIAN
y'

Log

ONIAG.-SAL.
j_________i

1.5

2

25

3

3.5

1

J

L o g Br ( m g / l )
Figure

1-11.

(a) L o g M C I
( m e q / l ) vs. log Br ( m g / l ) .
MCI
corrected
for c h a r g e
b a l an ce vs.
(mg71).

( b) L o g
log
Br

35

5 .5
*
✓✓

5h

✓
*✓

*

•T R A V E R S E -

Log

C! ( m g / I )

A B EREA
l l _________

S_______ L______ ! ____

_J______ l

5.5

*
*

/

*

•D U N D E E

f

4.5
5.5

±1

I_____ I_____ I_____ 1_____

5

/
/

4

✓

.

5

/

i_

A•

✓

r

• RICHFIELD O D E T . RIVER

4
1.5

2

2.5

3

3.5

4

4.5

Log M C I2 ( m e q / l )
Figure

1-12.

Log

MC12

(meq/l)

vs.

log

Cl

(mg/l).

36
have MCL2 values identical to seawater concentrated past the
start of MgS04 salt precipitation.

GEOCHEMICAL ORIGIN
As shown in Figures 1-8 to 1-10,

a very good agreement

exists between the seawater compositions and the Cl-Br and
Cl-MCI2 values in these Michigan brines, suggesting that the
geochemical
seawater.

origin

of

the

brines

is

evapo-concentrated

Based on Br concentrations, the Michigan brines

have apparently evolved from seawater concentrated from the
start of halite precipitation into the MgS04 salt facies.
Close inspection of Figures 1-8 to 1-10 shows that some
of the samples with lower Br concentrations plot below the
seawater
This

trend

line,

dilution

suggesting

appears

production depth, as
were

collected

(Figure 1-6).
margins

related

samples

Devonian

seawater or

lower

from shallow depths

1-6)

might

brines

have

freshwater,

with the seawater line.

not

found in Michigan.

brines

margins

greatly

of the

dissolution

diluted

samples

with

agree well

Certainly, the large scale dilution

Mississippi Gulf Coast brines
Pennsylvania

basin

shows however, that

found in the West Canada sedimentary brines

the

concentrations

CaS04

data

been

as most

Br

and

found near the basin

reflect
The

diluted.

location

near the

The higher S04 values

(Figure

have been

to well

with

resulting from infiltration.
the

they

(Spencer, 1987),

(Carpenter et a l ., 1974),

(Dressel

and Rose,

1982)

is

and
not

37
GEOCHEMICAL EVOLUTION
While the Cl/Br suggest the brines originated from the
evapo-concentration
expected

of

seawater

seawater,

concentrations

the
of

differences

Ca,

Mg,

K

from

and

Sr

demonstrate that the brines have evolved by other processes;
most important may be water-rock reactions and mixing.

The

extent of evolution is illustrated by the 87Sr/86Sr of the
brines

and

results

by

of

MC12

enrichment

a preliminary

study

described
of the

earlier.

strontium

The

87Sr/86Sr

isotopic composition of the brines are shown in Figure 1-13,
where

the

samples

isotopic

curve

geologic

age

below,
of

are

from

plotted

Burke

al.

of the producing

evapo-concentrated

Sr,

et

therefore,

reactions with

the

on

carbonates

in
or

(1982)

formation.

seawater

Sr

the

seawater

according
As

should be

these

strontium

brines

to

the

is discussed
almost devoid
must

other minerals.

A

reflect

generally

good agreement exists between the seawater curve and many of
the samples,

especially for samples from the Richfield and

Dundee suggesting that the brines have reacted with Devonian
aged carbonates to gain Sr.
several

facts,

seawater
Paleozoic.

was

most

important

apparently
However,

This conclusion is tempered by

the

several

the

same

many

times

of the

samples

radiogenic

than

87Sr/86Sr

of

during

the

(Traverse

and

formations)

curve,

perhaps the result of reactions with shale minerals
1984).

more

that

Berea

(Stueber et al.,

are

is

the

seawater

Reaction with aluminosilicates

is

-.7100

.7100

•^SEAWATER
^ RA NG E

0.7090

709 0

•/
v
\

0)

«

(0
00

e
i
<9

-.7 0 8 0

/

0 .7 0 8 0 -

N

\J

>_

0}

fCO

a
0.7070

0.7060

■"O—

<=
m ■**
H </>

aa

•.7070

.7060
P e n n Miss.

300

Dev.

Sil.

Camb.

Ord.

400

600

500

A G E my.
B u r k e et al.
(1 9 8 2 )

Figure

1-13:

8 ? S r / 8 6 Sr of M i c h i g a n b a s i n b r i n e s vs. p r o d u c i n g
formaxion
A l s o s h o w n is t h e v a r i a t i o n o f
the
Sr/
Sr
ratio
of
seawater
during
the
P h a n e r o z o i c Eon ( B u r k e et a l . , 1982).

39

also

consistent

with

the

depletion

of

K

from

expected

seawater concentrations (Figures 1-7 to 1-9).
The

enrichment

in MC12 over

expected

seawater

values

(Figure 1-11) also demonstrates that the brines have evolved
from seawater.

One explanation might be that MC12 is not

enriched, but rather, Br is deficient in these w a t e r s .
mechanisms

are

for example,

known

that

bromide

is

deplete

Br

considered

from natural

conservative

Few

waters,

in

evapo-

concentrating seawater as it is removed only after potashmagnesia

salts

precipitate

and Krejci-Graf

(1963)

(Holser,

1979).

(in Van Everdingen,

that Br and Cl may exchange onto clays.

Kozin
1968)

(1960),

suggested

However, the effect

of anion exchange on Br-Cl in formation waters

is unknown,

and it seems unlikely that exchange could noticeably deplete
Br from these waters.
by the

large dilution

concentrated

brine

difference.

Finally,
required

may

Although

analytical error magnified
for the analysis

explain
correcting

decreases the MCl2-Br difference
the MCL2-Br difference

to

from

an

excess

of

of

for

the

MCL2-Br

charge

balance

it does not eliminate

Therefore,
result

some

of highly

(Figure

divalent

1-11)

it.

is assumed

cations

in

these

waters, balanced by Cl or anions other than S04~ and C03= .
Divalent

cation

enrichment

reactions

involving

may

aluminosilicate

or perhaps by shale filtration.
from

mixing

with

result

MCL2

rich

from

water-rock

or evaporite minerals,

Enrichment may also result

water

from

other

Although the MC12 enrichment appears large

formations.

in the Michigan

40

brines

it

is

not

uncommon

for

sedimentary

exhibit divalent cation enrichment.
West

Canada

basin

waters

the Illinois basin brines

water

to

Figure 1-14 shows that

(Hitchon

Mississippi Gulf Coast brines

basin

et

al.,

1971),

(Carpenter et al.,

(Graf et al.,

1966)

the

1974),

and

are enriched

in MCl2 over seawater.
WATER-ROCK REACTIONS
Differences between the Ca, Mg, Sr, K, and MC12 content
of these brines and evapo-concentrated seawater may reflect
water-rock

interactions.

explain

these

Reactions

differences

aluminosilicate

reactions,

considered

include

to

dolomitization,

ion-exchange,

mineral diagenesis (Collins, 1975; Land,

here

and

evaporite

1987; Hanor,

1982).

DOLOMITIZATION
The
(Figure

strong
1-8

(Collins,

to

relationships
1-10)

1975;

between

is suggestive

Carpenter,

Ca-Br

and

of dolomite

1978).

Mg-Br

equilibria

Carpenter

(1978)

demonstrated that an approximate 1:1 relationship between Ca
and

Br

results

seawater,

rather

from
than

dolomitization
from

shale

by

evapo-concentrated

filtration

(Anderson

et

al., 1966).
The influence of dolomitization on the brine chemistry
was evaluated as follows.

The difference between Mg in each

sample and equivalently concentrated seawater was calculated
based on the measured Br in each sample.
was assumed to result

only

This Mg deficiency

from dolomitization

and

so was

converted to a predicted Ca concentration based on a 1 for 1

41

W CANADA

3.5

2.5

MISS

3.5
•f

IL L IN O IS

8*

2.5

1.5

2

2.5

3

3.5

4

Log Br ( m g / l )
Figure

1-14.

L og M C l ^ ( m e q / l ) vs. log Br ( m g / l )
w a t e r s of W. C a n a d a , M i s s i s s i p p i ,
basins.

in f o r m a t i o n
and Illinois

42

mole

replacement.

The predicted

Ca

are

adjusted

for

CaS04 dissolution, and then compared with the measured Ca in
Figure 1-15.
Generally,

a

very

good

agreement

exists

between

the

predicted and measured Ca, suggesting that dolomitization by
concentrated seawater
This

explains

suggests that the

Ca

and Mg

in

these

1:1 relationship between

waters.

Ca

and

Br

simply reflects Ca taking the place of the Mg in seawater,
which originally

co-varied directly with

concentration.

An

interesting

Br due

feature

of

to

evapo-

the

Mg-Br

relationship is that Mg/Br do not approach 1, as do

Ca/ B r .

If

before

Mg

depletion

and

Ca

enrichment

occurred

evaporation, then both Ca and Mg would plot
slope

of

l.

The

dolomitization

Ca-Mg

occurred

chemistry

concurrent,

(vs. Br)

might

with,

near a

reflect

or

that

after evapo-

concentration.
If

dolomitization

concentration,

then

occurred

the

concurrent

seawater

trend

with

line

evapo-

for Mg

in

Figures 1-8 to 1-10 may not truly represent the chemistry of
evaporating seawater in carbonate basins.

Rather, the paths

that Ca and Mg in these brine plot along may represent the
path

seawater

calcite

is

follows

maintained

when

equilibrium

during

with

evaporation.

dolomite

and

Alternatively,

dolomitization may have occurred after evapo-concentration;
the more

highly

dolomitization

concentrated brines were
and

thus

show

a

This may reflect a greater ability

greater

involved
depletion

in more
in

Mg.

43

T3

5.5

4.5

• TR AVE RSE

3.5

A BEREA
jlZ

4.5

9 DUNDEE
-J

4.5

• RICHFIELD

3.5

A DET. RIVER

1.5

2

2.5

3

3.5

4

Log Br ( m g /l )
Figure

1-15.

Results
of d o l o m i t i z a t i o n m o d e l .
Log
Ca
( m g / l ) v s . l o g Br ( m g / l ) . F i l l e d s y m b o l s a r e
m e a s u r e d Ca, o p e n s y m b o l s are p r e d i c t e d Ca.

44

of highly

saline waters

salinity waters

have

to dolomitize,

simply

or that the higher

interacted

to

greater

(Figure

1-15)

extents

with formation minerals.
The dolomitization model
Ca in some samples,
samples.

Combined

measurements may
all.

especially in the less saline Traverse
analytic

explain

More likely,

of

error

some

in Ca,

of this

Fe-rich

from

seawater

dolomite

or Mg-clays

by

underprediction

it may be large.
(Ordovician

Age)

difference,

(2CaC03 +

but

Ca

by

would

the

Br
not

thus

formation

the Mg

Fe

=
of

depleted

resulting
The

Fex

in

an

content

of

is presently not known,

For example,
are

the

reduce

mod e l .

in Michigan

Mg ^ 1-Xj +

Both

dolomitization,

of

Devonian dolomites

S04 , and

Explanations to considered are the

CaMgFe(C03)2) or Mg-clay minerals.
Fe-rich dolomite

Mg,

other process have added Ca or depleted

Mg from these brines.
formation

under-predicts

but

Trenton Formation dolomites

reported

to

contain

over

8%

FeC03

(Taylor, 1982).
To

further demonstrate dolomite equilibria,

the brine

chemistries were modeled using the PHRQPITZ program (Plummer
et

al.,

which

1989).

uses

PHRQPITZ

Pitzer's

coefficients,

is

a specific

equations

for

interaction

determining

following Harvie and Weare

(1980).

model

activity
Activity

coefficients of Na and Cl are adjusted for temperature,
are

temperature

temperature
constants.

is

invariant
considered

for
in

the

other

adjusting

ions.

but
Only

equilibrium

The Devonian brines of Michigan are a good test

45

for calculating dolomite equilibria
solutions

because:

coexisting

1)

and

calcite

common

brines appear to have been affected by dolomitization,

and

have

been

of their
in

suspected

contact

with

these

are

the

because

throughout

ionic strength

2)

3)

minerals

dolomite

in high

formations,

seawater
the

origin,

formation

the

brines

minerals

for

considerable time.
The results of the chemical modeling when the measured
pH and alkalinity values are used initially show the brines
are

slightly

saturation
dolomite
the

undersaturated

index

=

-0.71,

saturation

possible

with

calcite

s2= 0 .80)

index=-l.149,

errors

in

pH

and

and

(average

calcite

dolomite

s2= l .357).
alkalinity

(average

Considering
measurements,

chemical modeling was repeated assuming calcite equilibria.
Under

this

assumption

index of 0.017

an

average

dolomite

disequilbrium

(s2=0.230) is calculated, suggesting that the

brines are in equilibrium with dolomite (Figure l-16a).
Strontium
over

is

seawater

and

similar to that

highly

enriched

follows

of Ca

a

(Figures

great contrast to seawater,

in

1:1

the

Michigan

relationship

1-8 to

1-10) .

brines

with
This

is

Br,
in

which is devoid of Sr at these

levels

of

evapo-concentration

(Zherebtsova

and

Volkova,

1966).

The high Sr content of these brines is probably the

result of carbonate mineral reactions which can be evaluated
using Sr/Ca ratios.

Sass and Starinsky

that distinct ranges of Sr/Ca result
dolomitization

(1979)

demonstrated

in brines affected by

of calcite or aragonite,

the transformation

46

40
! DOLOMITE

ANHYDRITE

30
>*
o
§20
cr
a>
10

EL

XL
j »i
"1

"6

-.2 O .2

.6

-3-2-1

LOG lAP/Ksp

0

I

2

3

LOG JAP/Ksp

30
H A LITE

25
*20
. 15

10

a_

-3-2-10

1

LOG lAP/Ksp

Figure 1-16:

2

3

Eh.

- 3 - 2 - 1 0 1 2 3
LOG lAP/Ksp

Histograms of saturation Indices (log
for Michigan basin formation waters.

lAP/Ksp)

47

POLYHALITE

16
S Y L V IT E
I
I

12
>*
o
c
a>
3 6
cr
°
a>

-18 -14 -10 - 6 - 2

2

6

LOG lAP/Ksp

-4

-3

-2

-|

o

LOG lAP/Ksp

24
CARNA LLITE

20

>*16
o
c
a>

g
-12
a)

XL
•12

J]

a

-9 -6 - 3
0
LOG lAP/Ksp

Figure 1-16:

Histograms of
saturation
Indices
(lor
lAP/Ksp)
for
Michigan
basin
formation
w a te r6 •

48

of

aragonite

calcite.

to

calcite,

and

solution-reprecipitation

of

The Devonian formation waters have molar Sr/Ca of

0.01 to 0.023, with values independent of formation.

These

values agree with the Sr/Ca predicted for dolomitization of
aragonite,

and

calcite.

in a few cases,

Dolomitization

carbonate

diagenesis

history,

before

reactions.
important

the

If
then

solution-reprecipitation

of

affected

aragonite
the

aragonite

suggests

brines
was

early

solution-reprecipitation
a

Sr-rich

calcite

of

precursor

that

in

affected

their

by

other

calcite
is

of

was

suggested.

The use of Sr/Ca ratios to interpret seawater derived brines
such as these is somewhat tenuous, however, as these brines
would have

been

associated

all phases of diagenesis.

with

carbonate

minerals

during

The Sr/Ca may therefore,

be the

sum of the many different reactions that affect carbonates
and only appear to reflect the diagenetic reaction involving
the largest exchange of Sr.
The

O *7

Sr/° Sr

#

ratios

presented

earlier

(Figure

1-13)

suggest that these brines may have reacted with carbonates
of similar geologic age to the producing formations of the
brines.

This

conclusion

is

tempered

however,

because

seawater has had similar Sr ratios many times throughout the
Paleozoic.

Until more data are collected on the Sr isotopic

ratios of the basin minerals, such conclusions are tenuous.

49

ALUMINOSILICATE REACTIONS
Aluminosilicate mineral

reactions may be

important

in

explaining the potassium depletion and the MC12 enrichment.
Reactions
formation

might
of

include

albitization

K-feldspar,

al., 1973; Merino,

1975;

illite,

and

of

plagioclase,

chlorite

(Kharaka

et

Land and Prezbindowski, 1981).

Authigenic feldspars have been found in carbonate rocks
(Kastner, 1979), but have rarely been reported in petrologic
studies

of

personal

Devonian

c o m m .).

carbonate

Most likely,

rocks

in

Michigan

(Sibley,

authigenic K-minerals

would

be associated with elastics in the basin such as the Berea
and

Sylvania

feldspars
clear

Berea

Nowak

(1978)

reported

in Paleozoic shales of the basin,

if

minerals.

Sandstones.

these

were

considered

Authigenic(?) muscovite

formation

by

Sawtelle

but

authigenic
is also

(1958).

Na
it

or

and

K

is not
primary

reported in the
Because

feldspar

abundances and petrogenesis in Michigan basin rocks has not
been

documented,

it

is

not

possible

at

this

point

to

quantify their effect on K in these brines.
Most likely, the formation of illite and the subsequent
reaction with carbonates has caused the potassium depletion,
for

example:

2K+

+

CaC03

+

3Al2Si20 5 (O H )4

2KAl2 (AlSi3 )O1 0 (OH)2 + Ca2+ + 4H20 + C02 .

Illitic

=

shales

dominate the upper Devonian and Carboniferous rocks, and are
present

throughout

dispersed minerals

the

section

(Nowak,

1978).

as

discrete
Gardner

layers

(1974)

and

reports

upper Traverse Group contains up to 80% shale in the central

50

basin.

The

interaction

of

a

K-rich

seawater

brine

with

clays may help explain the dominance of illite in the Upper
Devonian shales, which are presently at shallow depths
m)

and

low temperatures

kinetically

(<22°C), conditions

unfavorable

for

the

(<500

thought

to

be

smectite-to-illite

transformation (Burst, 1969; Perry and How e r , 1970; Hower et
a l ., 1976).
then

the

If the shales originated as expandable

interaction with

K-rich

hypersaline

explain the dominance of illite in the basin.

waters

Devonian

formations

which

contain

brine

than

in

the Dundee

and

from the

elastics,

example, K depletion is greater in the Berea,
Sylvania

might

Illitization

is also supported by the depletion of K in brines
upper

clays,

for

Trav e r s e , and

Richfield-Detroit

River samples (Figures 1-7 to 1-9).
The reaction given above demonstrates illite diagenesis
may

be

tied

into

m o d e l , similar

the

excess

to that

MC12 .

used for

An

elemental

Ca-Mg,

balance

was attempted

to

determine if illitization could explain the K depletion and
the

"excess"

Ca

this reaction,

not

explained

the depletion

produces 1 mole of Ca.
seawater
large

(based on Br)

number of

still

are

aluminosilicate

reactions

amount of K depleted from

found to

this
is

diagenesis may

much of the excess Ca.

During

moles of K from seawater

follow

Generally however,

Whether

illite

dolomitization.

is plotted versus the "excess" Ca,

exists.

suggest

of 2

When the

samples

2;1 molar relationship.

by

not

excess

the predicted

an excess of Ca

represents

clear,

explain the

a

but

the

other
results

K depletion

and

51

OTHER SALTS
The

diagenesis

of

evaporite

minerals

may

important in the evolution of Michigan brines.
abundant
both

throughout

near

the

central-basin

the

western
area

Devonian
margins

(Gardner,

the

1974).

in

basin

As

be

Anhydrite is

formations

of

also

an

Michigan,

and

in

the

example,

the

"massive anhydrite" layer at the base of the Ituzi Member of
the Lucas Formation is over 30m thick and contains some 103
Km3 of anhydrite (Table 1-1).

Reactions involving anhydrite

or

could

other

evaporite

minerals

have

supplied

Ca

and

possibly enriched MC12 in the basin waters.
Reactions to consider

include the gypsum to

anhydrite

transformation and the replacement of gypsum by glauberite
or polyhalite.

During the gypsum-anhydrite conversion, each

lm3 gypsum liberates 0.486m3 of CaS04 saturated water which
can

dissolve

evaporite

minerals

significance
considering

anhydrite

of
that

(Borchert

this
the

"massive anhydrite"

along

with
and

potential

Ca

hypothetical

layer

other
Muir,

supply

more

soluble

1964).
is

The

realized

transformation

from gypsum would have

of

by
the

liberated

some 7.8X10-*--*- liters of CaS04 enriched water.
The replacement of anhydrite or gypsum by glauberite or
polyhalite liberates Ca rich fluid by:
=

Na2Ca(S04)2

(glauberite)

+

2CaS04 (H20)2 + 2Na2+

Ca2+

+

2(H20),

and

4CaS04 (H20)2 + Mg+ + 2K+ = K2MgCa2 (S04 )4 (polyhalite) + 2Ca+
+ 8(H20).

Inspection of Figure 1-8 to 1-10 shows that the

Na concentrations generally match seawater values,

although

52

random samples show a depletion in Na.

The Na depletion may

reflect glauberite formation, and perhaps ion exchange of Na
for Ca.
forms

Glauberite is not a primary evaporite mineral,
during

the

"breakdown

of

association with NaCl solutions"
Glauberite has

yet

to be

gypsum-anhydrite

(Borchert and Muir,

reported

in Michigan,

but
in

1964).

but

it

is

found in many evaporite deposits such as the Ochoa Series in
Texas
to

(Borchert and Muir,

replace

(Dellwig,

gypsum in

1955) ,

and

1964).

the

may,

Polyhalite has been found

Silurian
in

part,

salts
explain

observed K depletion in the Devonian brines.
reactions
water,

may

not

produce

large

of

amounts

Michigan

some

of

the

Although these
of

Ca

they represent an additional mechanism

enriched

for divalent

cation enrichment (MC12) in brine.
In
phases,

order

to

determine

done

routine

Plummer et

of

equilibrium.

as

saturation

Histograms

of

common

Figure 1- 15.
in

brines

indices

minerals in

attempted

other
in

earlier using

(1989),
had
of

indices
the

by

and

this
the

Michigan

calcite

effect

non-carbonate
IAP/Ksp)
basin

study.

PHRQPITZ

assumed

little

(log

mineral

of

the

minerals.
calculated

are

shown

in

The model results suggest that the brines may
with

in addition

are

al.

saturation

equilibrium

celestite

described

This assumption

calculated

be

control

equilibrium modeling was

Modeling was

for

the

to

apparently

halite,

anhydrite,

dolomite described
undersaturated

polyhalite, carnallite, and sylvite.

gypsum,
earlier.

with

and
The

glauberite,

53

MODEL FOR THE BRINE EVOLUTION BASED ON GEOCHEMICAL DATA
The

chemical

data

and

geologic

information

presented

above suggest that brines in the Devonian Formations of the
Michigan basin evolved from evapo-concentrated seawater.
Devonian

seawater

origin

history and geology,
evaporative

is

consistent

as Devonian rocks

conditions.

This

is

with

the

in Michigan

A

basin
reflect

illustrated

by

the

dispersed and bedded evaporite minerals in the basin and by
the coastal environments interpreted to have existed in the
basin

(Gardner,

1974;

see Table

1-2).

Evapo-concentration

of seawater would have occurred in the coastal
lagoons located around the basin margins
perhaps

in

stands.

the

central-basin

area

sabkhas

and

(Figure 1-3),

and

during

lower

sea

level

Dense, Na-Mg-K-Cl-S04(?) seawater brine would have

refluxed down into the carbonate sediments towards the basin
center, mixing with and displacing less dense water residing
in the sediments.

The more highly concentrated brine would

have migrated farther down into the

central-basin area as

suggested

(Figure

by

underlying
trapped

the

Br

Devonian

the

brine

distribution

and

Silurian

and

prohibited

occurring.

Water-rock reactions,

after

brine

the

seawater

generation

composition.

dolomitization,
aluminosilicate

which

salt

which

deeper

modified

Important

diagenesis,

beds

The

would

have

reflux

from

occurring both during and

then

affected

1-6).

the

the

original

reactions

include

Ca,

affected

Mg,

and

potassium

Sr,
and

MC12 , and sulfate reduction or precipitation of CaS04 during

54

dolomitization,,

which

removed S04 .

Sulfate reduction is a

common processes known to operate in areas such as sabkhas,
and may have played a more important role in the basin than
previous studies have suggested (Matthews and Querio,

1974).

STABLE ISOTOPE RESULTS
Figure 1-17 shows the
(vs.

SMOW)

in

this study.
al.

the

Devonian

The results

(1966).

D/H and

Also shown
Craig

180/ 160 isotopic ratios

formation

waters

collected

in

are similar to those of Clayton et
is the global

(GMWL)

from

(1961),

during

evapo-concentration

and
up

two
to

meteoric water

pathways

halite

for

line

seawater

saturation,

from

Holser (1979) and Pierre (1982), respectively.
The

Michigan

brines

plot

very

near

the

end

evaporation trend lines and extend back to the GMWL
l-17a).

The

intermediate

Berea,

of

Traverse,

the more

saline

and

Dundee

Richfield,

of

the

(Figure

samples
Detroit

plot
River,

and Niagara-Salina samples and the GMWL (Figure l-17b).

The

isotopic composition of "apparent modern-day meteoric water"
(AMMW)

is

found

where

intersect the GMWL.
be

calculated

selected

and

best

fit

line

to

the

data

Four AMMW values and "basin" lines can

(Table
if

the

D

1-6),
is

Correlation coefficients

depending

corrected

for

on

the

activity

show a wide range

formation
( 6 a D) .

of goodness-of-

fit for calculated lines, with the best correlation obtained
using £ aD vs. S 180 in Devonian formation water samples only.
AMMW values are near estimated present-day average meteoric
water falling on mid-Michigan

(S180 = - 7 .3°/oo,

6D=-50.6°/oo,

(c o n c.)

55

• DEVONIAN
SMOW

D SMOW

• /
J

del

cP® .

O N IA G A R A /
SALINA
□ O RDO VIC IA N

/

D S M O W (act.)

80
40

i— i— ~i— i— i— ~~i

i

i

t

i |

i

r

O BEREA
M TRAVERSE
® DUNDEE
A RICHFIELD
A D E T R O I T R.
O N IA G A R A /
S A L IN A
□ O RDO VICIAN

B

O

,« 4

-40

* a

del

-80
-120
-16

Figure

1-17-

-12

-8
del

-4
0
4
180 S M O W

8

( b )
del D°/oo (SMOW) concentration scale, lb) and
del D /oo (SMOW) activity scale, vs. del
0°/oo
(SMOW) of Michigan basin waters.
Also shown Is
the best-fit line to the data, and two examples
of
seawater
compositions
during
evapoconcentration,
from Molser
(1979) and Pierre
(1982).

TABLE 1-5
Apparent meteoric water compositions.

BEST FIT LINE

r—

,„AMMW
18
/^a
del— 2^0^
del D
.

n

A: del CD = 1.912

del 180 + -34.537

0.46

-7.32

-48.53

40

B: del aD = 3.197

del 180 + -22.909

0. 67

-6,85

-44.82

40

C: del CD = 2.508

del 180 + -33.314

0. 61

-7.89

-53.10

29

D: del aD = 3.489

del 18() + -23.791

0.74

-7.49

-49.93

29

E: del CD = 3.646

del 18() + -38.59

0.79

-11.16

-79.28

26

Equation Key:
A: all data, del D on concentration scale
B: all data, del D on
activity scale
C: Devonian formation water data only,
del D on concentration scale
D: Devonian formation water data only,
del D on activity scale
E: data from %Clayton
et al. (1966), del D
on concentration scale
AMMW: Apparent modern-day meteoric water, calculated from
intersection of best-fit line with the GMWL
n: number of samples

57

Long et al, 1989).

Also listed in Table 1-6 is the equation

for the best fit line

from

Clayton et al.

Clayton et al.

used

samples mainly from the Devonian

formations,

their

calculated
However,

(19966)

for

line

is

Devonian

similar to

samples

the Clayton et al.

the

from

(1966)

(1966) . Because

best-fit

this

present

lines
study.

line intersects the GMWL

at slightly more negative values than present-day meteoric
water.

INTERPRETATION
Based partially on the §D-S180 values in Michigan basin
brines,
had

Clayton

et a l . (1966)

flushed the basin

argued

that meteoric

and that Michigan brines

part, by shale membrane filtration.

waters

formed,

in

A meteoric water origin

was suggested because the best fit line to the isotopic data
intersects

the

precipitation
calling

GMWL

at

falling

a
on

SD value
Michigan.

for the evolution

of

similar
Their

freshwater and

to

modern-day

model however,
seawater

into

brine, is not consistent with the evaporated seawater origin
interpreted from the chemical data in this present study.
Knauth

and

composition

of

dilution

of

evolution

Beeunas
many

(1986)

sedimentary

evapo-concentrated

of

meteoric

proposed

water

that

basin

seawater
into

brine.

the

brines
rather
The

isotopic
reflects
than

the

isotopic

composition of seawater first apparently increases in D and
0 during evaporation until approximately gypsum saturation
is reached

(4x concentration).

the isotopic values decrease;

With continued

evaporation

due in part to back exchange

58

with

the

atmosphere

and

changing

H20

activities.

The

isotopic composition of seawater after halite saturation is
reached

(lOx concentration,

lines in Figure 1-17)
and

may

be

highly

near the end of the evaporation

is however,

variable.

generally not predictable

The

dilution

of

a

residual

seawater brine by AMMW then causes formation waters to plot
along the best-fit "basin" line (Knauth and Beeunas,

1986).

The isotopic composition of the Michigan brines may be
explained by Knauth and Beeunas

(1986)

model.

Perhaps the

best evidence for this is the lower salinity of the samples
plotting close to the GMWL.

This is demonstrated by Figure

1-18, which shows £ aD plotted versus log Br

(mg/1)

for the

brines

strong

linear

collected

relationship
conservative

in

this

(r2=0.8)

study.

that

components

The

exists

shows

that

between

these

samples

with

negative &D values have lower Br and salinity.
this

and

the

spatial

distribution

shown

in

two
more

Considering
Figure

l-6a,

water with lower Br and more negative D values are produced
from near the basin margins.
the

isotopic

data

for

the

Thus, at first consideration,
Devonian

brines

are

consistent

with evapo-concentrated seawater that has been diluted near
the basin margin by
member waters

infiltrating meteoric water.

and the extent

entirely evident.

The

end

of dilution are however,

not

(a c t.)
del D %oSMOW

• DEVONIAN
✓

- 160

0

1

2

o

M5 6

O NIA G .-SAL.

3

4

Log Bfi (mg/i)
Figure

1-18 .

del
D /oo
(SMOW)
vs.
log
Br
(mg/1)
in a l l
Michigan ba^in waters.
A l s o s h o w n is t h e b e s t f i t l i n e ( r = . 8 ) to t h e D e v o n i a n f o r m a t i o n d a t a ,
and s a m p l e M56 from % C l a y t o n et a l . (1966).

60

DISCUSSION
Although the Knauth and Beeunas model is accepted here,
an inconsistency occurs in that while the isotopic data may
represent dilution,

the chemical data apparently does

not.

If dilution has been masked in the Cl and Na data shown in
Figures

1-8

between

Br

to

1-10,

and

Ca,

then
Sr,

the

Mg,

co-varying

K,

and

MC12

relationships
may

represent

dilution and not enrichment by evapo-concentration.
is indeed the case,
the

upper

formed

by

composition

then the isotopic data might show that

Devonian

brine

dilution

of

of

If this

(Berea,

saline

around

Traverse,

water

6 D=-20°/oo,

having

Dundee)

an

$ 180=2°/oo,

some of the more concentrated Richfield,
Niagara/Salina brines.

and

isotopic

similar

to

Detroit River,

or

In other w o r d s , the brine in upper

formations of the basin may have evolved from water derived
from deeper

formations.

agreement between

Two

explanations

for the

lack of

isotopic and chemical data are possible,

either the isotopic data do not represent dilution by AMMW,
or

halite

equilibria

has

removed

dilution from the Na-Cl data.

chemical

evidence

of

Failure for the isotopic data

to reflect dilution might mean the

isotopic composition of

the brine is primarily that of evapo-concentrated seawater,
or

alternatively,

the

components

expected

to

behave

conservatively during dilution (D,0,Br), in reality, do not.
In order
end

to explore

member

waters,

these
both

questions
saline

further,
and

the

AMMW

possible
must

be

characterized, and the question of what would be required to

61

link

the

chemical

and

isotopic

data

together

must

be

addressed.
SALINE END MEMBER
In contrast to
composition
difficult

of

its chemical

the

to

saline

end

quantify.

composition,

member

The

water

isotopic

the

isotopic

(seawater)

composition

is
of

evaporating water is a function of the initial composition,
humidity, rate of evaporation,
phase,

and

degree

of

mixing

isotopic values in the vapor
(Lloyd,

1972 and 1975; Nadler and Margatiz,
complex

interplay

of

these

1966;

Sofer

1980).

factors, the

and

Gat,

Because of the
isotopic

path

of

seawater during evaporation may be unique for any basin, and
may vary within a basin.

This variability

is demonstrated

by the two different evaporation paths shown in Figure 1-17,
and by the results of Nadler and Magritz
the

isotopic

evaporation.
saturation,
scale)

composition
Their

brines

data

Mediterranean

show

that

after

seawater

during

reaching

halite

residual seawater brine had a 8 D

composition

seawater

of

(1979) who measured

source.
are

precipitation

more

negative

Considering

concentrated
(lOx),

far

very

than

that
past

little

the

most
the

can

(concentration
non-evaporated

of

the

start

be

said

of

Michigan
halite

about

the

isotopic composition of the parent seawater bitterns for the
Michigan

brines.

seawater

before

present-day SMOW.

Additionally,
evaporation

may

the
have

isotopic
been

It has been suggested,

values

of

different

from

for example,

that

seawater may have been 1 to 2°/oo lighter in lsO during the

62

Devonian (Fritz, 1971; Popp et a l . , 1986).

A change of this

magnitude in both £ 180 and S D would make Devonian seawater
plot very near the main group of Michigan brines in Figure
1-17.

Thus,

the

isotopic

composition

of

Michigan

basin

brines may be nearly that of the parent seawater from which
they

were

derived,

seawater brines

or

were

alternatively,

diluted with

if

Devonian

concentrated
seawater,

then

dilution may not be visible in the' isotopic data.
APPARENT MODERN-DAY METEORIC WATER
The isotopic and chemical nature of AMMW must also be
considered.
must

have

A wide variety of meteoric water
fallen

on

the

basin

throughout

compositions

geologic

time.

Thus, the chemical and isotopic composition of AMMW may not
be that

of

"modern-day11 meteoric water.

between D and Br
AMMW

(Figure 1-18)

composition.

considered
deuterium

may

enriched

by

exchange

organic

in
with

matter

Hitchon and Friedman,

1969).

v s

with a

. S 180

SD

of

(Table
-50°/oo,

1-6) .

clays,
decay

water

is

generally

waters,

(Fritz

although

and

et

may

a l .,

The best fit line to the

br

1986;
D-Br

=.8), somewhat better than

This

line predicts
(Table

that AMMW

1-6) , should

(525 mg/1 using the concentration

The source for a Br enriched meteoric water of this

composition is unknown.
a

Br,

gases,

the AMMW value

contain about 450mg/l Br
scale).

like

formation

data shows a good correlation (r
for SD

relationship

can be used to help evaluate

Deuterium,

conservative

The

by

salt

To evolve meteoric water into such

dissolution,

saturation

with

halite

63

containing uncommonly high concentrations of Br is reguired,
in

this

case,

at

least

1260ppm

Br.

describe halite recrystallization

Recent

models

that

(Stoessell and Carpenter,

1986) and the effects of diffusion during the interaction of
saturated
(Wilson

brine
and

with

Long,

halite

1985),

may

but

explain

the

Br

enrichment

occurrence

of

these

processes in natural systems has not been demonstrated.
The

D-Br

line

negative

6 D

meteoric

water,

This

value

rather

values

is

than

isotopically
(Figure
not

in

1-18)

at

Br

low

example,

light
taken

values

of

is

back

when

water.

also

the

derived

by

fit

line

of

water

Infiltration

supported

best

very

Br=lmg/1.

sample

from Clayton et a l . (1966),

determining

to

characteristic

glacially

meteoric

water

extends

8 D=-195°/oo

characteristic

present-day

1-18)

used

for

(Figure

of
M56

which was

sho w n .

This

Traverse water sample was collected at the extreme western
edge of the basin
meltwater

(18N,

17W)

(gD=-109 °/oo,

S 180 = -13.11

which had dissolved halite
minimum)

(Graf et

best-fit

line,

al.,

and is thought to be glacial

(115 g/1)

1966).

suggesting

°/oo,

Br=

134 ppm)

and anhydrite

(3.6 g/1

Sample M56

that

dilution

light water causes the D-Br relationship.
member water has a more negative

plots
by

near

the

isotopically

If the dilute end

isotopic composition then

AMMW then a much smaller amount of meteoric water is needed
to explain the

isotopic trend,

which may help explain the

lack of dilution in the Cl-Na d a t a .

64

LINKING ISOTOPIC AND CHEMICAL DATA
Considering

the

problems

involved

in

defining

the

chemical and isotopic compositions of the end member waters,
it

may

not

be

possible

to

link

together in a single model.
would be

required

if the

the

two

types

of

A question therefore,

two types

of

data

data

is what

represent

the

same evolutionary path?
This

question

is

explored

in

Figure

1-19,

where,

because the isotopic composition of the saline end member is
not known,

the following model was considered.

Two Dundee

formation water samples (water A, #3029, and water B, #3081)
are

used,

along

Table 1-5
collected

with

apparent

meteoric

water

(AMMW)

The composition of these two samples, which were
from

the

basin

center

and

margin,

respectively,

and various mixtures are listed in Table 1-6.
of

a

from

single

saline

end

member

by

AMMW

If dilution

explains

these

samples, then continued dilution of sample A by AMMW should
account

for

the

chemical

intermediate sample B.

and

isotopic

composition

of

the

Bromide and deuterium are the only

elements that can be considered conservative

in mixing,

as

Cl could be supplied by halite dissolution and 180 could be
altered by reactions with carbonates and other minera’s.
First, the deuterium in sample B is predicted based on
dilution
bromide.

of

sample

Based

(Br=0)

predicts

Figure

l-19a

on

A,

B r , the

that

after

with

a

dilution

sample
50%

dilution

mix

B

would

of

calculated
sample

plot

at

h

by

point

using
AMMW
C

on

(Br=640mg/1, Cl= 9 2 ,500mg/l).

65

2

2.5

3

3.5

Log Br (mg/l)

xA
AMMW
®— B

-100
-15

0

5

del 1 8 0 (SMOW)
Figure

1-19.

PoBBible mixing scenarios between Dundee brines
and AMMW. (A) Log Cl ( m g / l ) v s . log Br ( m g / l ) ,
and (B) del D /oo vs. del
0 ° / o o (SM OW )

66

TABLE 1-6
Results of nixing example.

POINT

SD

Cl

S'J^ O c

Br

1.06

1240

185.000

6.21

620

92.500

C
-39.6
-35.8 -3.53
-3.42
(not corrected for brine salinity)

620

158.000

0.5

C
-39.6
-3.53
(corrected for brine salinity)

620

92.500

0.5

D'
-52.7
-52.6 -7.76
-7.73
(not corrected for brine salinity)

22

3330

0.18

D
-52.6
-7.76
(corrected for brine salinity)

19

2775

0. 15

A
-26.1
(sample 3029)

-18.6

0.43

B
-52.6
(sample 3081)

-48.6

-6.53

-

AMMW
-4 9.9 -7.4 9
(activity scale from line D, table 5)

variable

AMMW -53.1
-7.89
0
(concentration scale, from line C, Table 5)

variable

Key: Br and Cl as mg/l, del D and del 180 as °/oo, SMOW.
Point C = composition resulting from dilution of Br
jo Dn= composition resulting from dilution of D
E°ink
'D, “ O = deuterium and oxygen on concentration scale.
aD, 180a = deuterium and oxygen on activity scale.
f = fraction of brine sample A in mixture, 1-f
fraction of AMMW
in mix.

67

Point C clearly does not match sample B in Cl,

but if AMMW

had

dissolution

dissolved

63.5

g/1

halite,

or

if

halite

occurred after mixing to increase Cl to 131,000 mg/l,
both Br and Cl

in sample B can be explained.

shown in Figure l-19b,
S180=-3.22°/oo

in

(concentration

•
enrichment

S cl80

over

1Q
of S o

However,

as

the mixture has a 6 cD=- 3 9 .6°/oo and
scale,

activity following Sofer and Gat,
enriched

than

by

water

S 180

1972),
B.

which

Whether
,
mineral

carbonate

corrected

from

is 3 permil

this

reflects
,

reactions

is

discussed below.
Next, the Br in sample B is predicted based on dilution
of sample A, with dilution calculated using deuterium.
calculating

the

isotopic

and dilute water,

composition

of mixtures

of

When
brine

it is important to recognize that a given

volume of a highly concentrated brine contains significantly
fewer water molecules
example,

than

the

same

amount

a liter of brine A contains about

molecules in a liter of pure water.

of AMMW.

For

7 0% of the H 20

Consideration of this

is made in the isotopic mixing calculation by using:
F* A * (

Dbrine)+ (1-F)*i*(

Dfreshwater)=

where F is the fraction of brine mixed,
of

water

compared

molecules
with

in

freshwater

the

and A is the ratio

concentrated

(water

Dmix

molecules

brine
per

solution
kg

brine

solution / water molecules per kg freshwater). The value of
A in AMMW is assigned 1.

The results of calculated mixing

considering this factor are listed in Table 1-6.

68

The mixing of 15% sample A with 85% AMW can predict the
S CD in sample B.

Because sample B does not fall exactly on a

straight mixing line between sample A and AMMW, sample B can
be matched only

in

8CD,

and only approximately matched

S180.

This mix plots at point D in Figure l-19b.

point

D plots

jpD=-52.6,

in Figure

l-19a at

S c 180= - 7 .76°/oo).

in

However,

Br=19mg/1, Cl=2775mg/1

Because

the

final

would depend on the amount of halite dissolved,

Cl

(

content

the mixture

9

may plot
mixture

anywhere
chemistry

along

the

would

line

reflect

D-D .

In this

case

halite

dissolution,

the
and

clearly does not match the composition of sample B.
In sum,
Br

and

&CD

it does not appear possible to predict both the
in

sample

B

by

using

a

single

Similar results are obtained when mixing

mixing

samples that have

SCD more intermediate between sample A and AMMW,
is noted that the closer the
water is to sample A,

ratio.

however, it

SCD value of the intermediate

the closer the agreement becomes with

dilution based on Br.
These calculations suggest that if dilution of a saline
end

member

brine,

water

by

AMMW

explains

the

Devonian

as suggested by the isotopic data,

formation

then it must have

been accompanied by halite dissolution and

180 enrichment.

Halite

considering

dissolution

is

certainly

presence of halite in the basin,

possible,

the

and the geologic evidence

that Salina salts were extensively dissolved from around the
basin

margins

1945).

The

during

next

the

Devonian

consideration,

time

therefore,

(Landes

et

is whether

al.,
180

69

enrichment has occurred
.
minerals.
(1966),

from reactions with oxygen bearing

•
1Q
Enrichment of xo0 was proposed by Clayton et al.
who

suggested

that

isotopic

•

I

carbonate minerals causes the

Q

,

g o

m

equilibrium
,

with

,

brines to shift

the meteoric water line (Figure 1-17).

from

Hitchon and Friedman

(1969) calculated that West Canada formation waters may have
been enriched by as much as +9°/oo in
equilibration
demonstrated
•

increase
Knauth

with
that

pore-waters

1
xo0

during

ft

m
and

Beeunas

formation waters

Isotopic

versus

in

(1986)

estimated

in carbonates

Land

(1980)

will

rapidly

show how

180

enrichment

and
of

from gypsum dewatering as well

1986).

equilibrium
Figure

rocks.

mineral re-crystallization,

can result

(Knauth and Beeunas,

evaluated

carbonate

S180 as a result of

1-20,

formation

with

carbonate

where

S180

values

temperatures.

minerals
are
Data

is

plotted
on

the

isotopic composition of Devonian carbonates in Michigan are
severely lacking, however, Gardner (1974) reports an average
180/160

of

-5.75°/00

Detroit River Group.
Clayton

et

calculate

al.

water

(PDB)

for micritic

(alpha)

from

the

This values is similar to that used by

(1966) (-5.39°/oo PDB) and can be used to
1o
t
§ 0
as a function of temperature using:

alpha* < S18°rock+100<» = < 6 18‘W e r » ' 1°°0 > •
factor

calcite

is calculated using:

- 2.89 , from O'Neil and Epstein

The

fractionation

103ln alpha=2.78xl06T-2

(1966).

Figure 1-20 shows

the brines exhibit a wide spread of 180/ 160 ratios over the
narrow range of formation temperatures,

and plot across the

70

- I P D B ,,'

-5.75 _ .
PDB^
/

/

/

□

/

£

o

r-

□

A

s
/°A
t/> 0
o

-2

v

(X>
Q)

°

-4

/
/
V

<P

/

/

/
/

PDB

ABEREA
©TRAVERSE
• DUNDEE
□ RICHFIELD
■ DETROIT R.

/

/

/

/

/* O
-/•
/
•

/

/

/

• * /

. / •
,

"O

/

/

/

-

/
•/
-6 h / •
£*
i

20

30

i

40

JL
50

I

60

70

TEMPERATURE°C
Figure 1-20.

del
0 °/oo (SMOW) vs. estimated formation
Jgmperature. Shown as a dashed line are the del
0 values
for water
predicted
to be
in
equilibrium with calclte of -1, -5.75,
and -8 /oo PDB.

71
calculated equilibrium line.

Calculated values fall between

lines for calcite of -l°/oo PDB and -8°/oo PDB.

Although

the data seem to follow the general trend of the equilibrium
•

•

,

1 Q

lines, there is not a single

,

xo0 value for calcite that can

explain the range of brine compositions.

A better fit would

be obtained if (1) higher temperatures (paleo-temperatures?)
are used for samples collected from the basin interior,
(2)

the

fractionation

temperature.

factor

(alpha)

increases

Non-uniform higher temperatures

more

or

with

in the basin

interior may reflect that brines in these areas were buried
deeper

in

the

past.

The

explanation

of

a

fractionation

factor that increases more as a function of temperature than
the one used here might indicate that dolomite equilibrium
plays

a

role

equations

in

controlling

for dolomite-water
than

these

fractionation

apparently

have

Further information on basin mineral and cements and

isotopes

in the brines

needed before

this

system

as

1982).

completely addressed.

calcite-water

waters,

slopes

are

the

in

larger

paleo-temperatures

for

180/ 160

(Land,

question

can

be

However, the enrichment of the oxygen
by

carbonate

equilibrium remains

a

possibility.
MODEL FOR THE BRINE EVOLUTION BASED ON ISOTOPIC DATA
The isotopic evolution of Michigan formation waters is
an enigmatic problem.
brines

largely

signature
dilution

retain

One explanation is that the Michigan
their

original

seawater

isotopic

(perhaps modified by water-rock equilibria) , and
has

been

limited,

primarily

to

near

the

basin

72

margins.

This

explanation

geochemical

model

presented

applicable

for

brines

in

is

consistent

earlier,

the

and

lower

with

the

be

most

would

Devonian

formations

(Richfield, Detroit River).
The isotopic data can also be interpreted to show that
large

scale

dilution

of

a

occurred.

Considering this,

the

Devonian

upper

concentrated brine,

brines

seawater

by

AMMW

originated

by

mixing

of

with AMMW.

reflect dilution,

concentration.
remove

In this case the co-

has

of

been

characteristic
carbonate

(Figures

and not enrichment by evapo-

Additionally, halite eguilibria is needed to

evidence

dilution

higher

from lower Devonian or

varying relationships between Ca, Sr, K, Mg and Br
1-7 to 1-9)

has

a second scenario may be that

perhaps derived

Niagara/Salina formations,

brine

limited

of

rocks

dilution

from

to

Cl/Br

where

Br

and

remains

evapo-concentration; and

have

have affected Mg.

increased

180

in

Na/Br

ratios;

at

values

reactions

the

brines

with

and

may

Each of these conditions appear possible.

If mixing explains the elemental

and

isotopic trends,

then evidence of mixing would be expected in the Sr isotopic
ratios

when

plotted

versus

Sr

content

(Figure

1-21).

Mixing of two waters having different isotope ratios and Sr
concentrations should result in a straight line in a plot of
87Sr/86Sr

vs

1/Sr

(Faure,

1986).

The

amount

of

data

available limits the interpretations that can be made, but a
direct
Rather,

linear
the

mixing

data

trend

appear

to

is not
group

observed
by

in

formation

this
in

plot.
a

band

O BEREA
V TRAVERSE
© DUNDEE
A RICHFIELD
A D E T R O IT R.
O N IA G A R A /
S ALINA
□ O R D O V IC IA N

□

.710
CO
CO
00
\
i>_
CO

\
I

:o o ' ,

.709

J

V

\

oo

o

K\m
v>#

I
' *.<

.708

(A
.707

i

i

i

t

_|

5x10*

I ■ ■ I j

i i i

r3
1x10
1 /S r(m g /I)

Figure

1-21:

P l o t of 1 / S r ( m g / l ) v s .
M i c h i g a n basin brines.

i

74

having

similar

isotopic

values

independent of Sr concentration.

(0.7082

to

0.7092)

One interpretation is that

this diagram does show dilution of a Richfield-Detroit River
or Niagara/Salina formation brine,by a dilute water having a
isotopic
ground

composition

water

in

similar

Michigan

to

may

the

have

brine.
these

Near-surface
characteristics

because their Sr composition would be governed by reactions
with Paleozoic
example,

carbonates

in the bedrock and t i l l .

Lake Huron water

ratio of 0.7086

is reported to have

an

a 87Sr/86Sr

(Faure et a l ., 1963), similar to many of the

deep brine samples reported here.
by water-rock reactions,
(McNutt et al.,

As

Enrichment in 87Sr caused

especially those

involving

shales

1987; Stueber et al., 1984), may explain the

dispersion from a linear mixing trend.
Several

other

observations

lend

support

to

a

evolution for the upper Devonian formation brines.
it

is

questionable

that

evaporative

mixing
First,

conditions

in

the

Traverse and Dundee formations ever reached the high degrees
needed to

explain

the

Br

in the brines.

Conditions

that

favor seawater evapo-concentration into the late halite and
MgS04

salt

facies

is

certainly

more

consistent

with

the

Salina geology in Michigan.
Additionally,

the Niagara-Salina and Richfield-Detroit

River formation waters are highly enriched
1-11, Chapter 2).

in MC12

(Figure

Although the mineral reactions discussed

previously might explain some MCl2 enrichment, especially in
light of the enrichments found in other basins

(Figure 1-4),

75

dilution of a MC12 enriched brine derived from the Silurian
salts

is

also

possible.

Based

volumes listed in Table 1-1,

on the

measured

formation

fluids representing less than

10% of the present volume of Silurian salts can account for
the volume

of

Devonian

porosity

of

Silurian

salts

compaction

10%).

Any

that

may

formation waters
remnant

has

have

brine

migrated

caused

the

(assuming
squeezed

upwards
MC12

average

from

during

sediment

enrichment.

These

fluids would have mixed with dilute waters present
Devonian

sediments,

margins.

or

were

later

It is not known however,

have moved

upwards

compacting

basin

suggests

a

to

diluted

near

as

to

basin

if such dense brine could

Michigan (Bethke,

modification

in the

the

the Devonian sediments in

such

the

a

slowly

1985),

which

this scenario.

The

MC 1 2

enrichment may have simply been carried along through time
with

each

not

based

incursion of
on

suggested
remained

chemical

that
in

seawater

the

highly

into the basin.

evidence,

Briggs

concentrated

central-basin

are

resided

Devonian.

in

the

lower

areas

of

al.

seawater
of

incursions of seawater in the Silurian.
have

et

Although

the

(1980)

bitterns

Michigan

between

Similar fluids may
basin

during

Burial and compaction may later move the

the

fluids

towards the basin margins, mixing with meteoric water.
SUMMARY
The data presented here suggest that Devonian formation
brines

in

seawater.

Michigan
This

is

originated
supported

by

from
the

evapo-concentrated

halide

chemistry,

as

76

well as by relationships between the other elements and Br.
Although

somewhat

dilution

of

meteoric

more

an

equivocal,

isotopic

evapo-concentrated

water.

The

original

data

seawater

support

bittern

evapo-concentrated

by

seawater

chemistry was modified by extensive water-rock interactions.
Most

important

appears

to

be

dolomitization

and

aluminosilicate reactions such as illitization.
Two scenarios are presented to explain the chemical and
isotopic
with

origin.

Devonian

from

the

Either

the

brines

formed

formation rocks or they

underlying

Detroit River,

formations,

are

such

or Niagara-Salina.

syngenetically
in part derived

as

the

Richfield,

The data do not seem to

unequivocally distinguish between these origins.
Several
favor

a

chemical

syngenetic

formation water,
important
degree

are

of

the

Sylvania-Detroit
suggest

an

brines.

evolution

difference

concentration

in

conditions

observations

for

the

however,

upper

Devonian

in

reached

cation
by

Most

composition

the

upper

and

Devonian

(Dundee-Traverse-Berea) brines, versus the
underlying
River

waters

Silurian).

evolution

similarity

(Sawtelle,

formation

and

independent

The

geologic

and make this the preferred m o d e l .

the

formation brines
water

and

in

1958)

for

These
these

chemistry
suggests

and
that

(Richfielddifferences

two
lack
Berea

groups

of

evaporitic
Formation

waters perhaps migrated from the upper Devonian formations.
The

fact

(Richfield,

that

the

lower

Detroit River)

Devonian

formation

brines

are more highly concentrated Ca-

77
Cl solutions that have similar isotopic compositions that do
not suggest dilution,

and are associated with Devonian and

Silurian evaporite deposits supports that these waters share
a common origin,

possibly being remnant brine derived

from

the evaporite deposits.
In

the

syngenetic

origin

model,

the

brines originated during deposition of

upper

Devonian

formation sediments

in coastal sabkhas and lagoons as well as during restricted
periods of the Devonian.

Dense seawater brines would have

refluxed down into the lower areas of the basin and reacted
with the
layers

formation

would

downward.
isotopic

have

carbonates
kept

The brines
composition,

the

and clays.

brines

from

Underlying
refluxing

further

largely retain their parent
perhaps

being

modified

by

salt

seawater
carbonate

equilibria, and some were later apparently diluted to minor
degrees with meteoric water.
In

the

second

scenario,

Devonian formations originated,
from

the

Divalent

lower

Devonian

cations

in

the

or

the

brines

in part,
perhaps

fluids

would

in

the

upper

from fluids derived

the

Silurian

have

been

salts.

enriched

either as the result of the reactions suggested earlier,
by reactions involving the Silurian potash deposits.
fluids would have moved
compaction

and

mixed

into the

or

These

Devonian sediments during

with meteoric

water.

In

this

case,

the diluted bitterns have maintained equilibrium with halite
and carbonate rocks.

CHAPTER 2

Origiii and evolution of water in Niagara-Salina and
Ordovician aged formations, Michigan Basin

INTRODUCTION
Sedimentary basins contain a variety of sediment types
reflecting many geologic histories,

and as such, no single

rock type or diagenetic history will
basin.

Similarly,

characterize a single

it is possible that no single origin or

evolution can explain the chemistry of all waters within
basin.
with

Although
sediments

formation

water
and

chemistry

may

evolve

mixing,

the

water

by

of a basin may

retain a chemical

by

a

reactions

within
signature

reflects an origin or evolution unique to the basin.

each
that
This

suggests that to best determine their geochemical origin and
evolution,

basin waters should be studied on an

individual

formation basis whenever possible.
The Paleozoic rocks in the Michigan basin
reflect
basin,

the

many

geologic

conditions

that

(Figure 2-1)

existed

in

the

such as periods of shallow and deep water carbonate

deposition,
deposition,

hyper-saline

evaporite

and periods of emergence.

deposition,

clastic

The origin of saline

waters contained within the basin has been linked to periods
of intense seawater evaporation

(Chapter 1).

Under a more

detailed examination however, differences in water chemistry
are observed between the individual formations of the basin.
Figure 2-2

illustrates the the major cation composition of
78

79

MICHIGAN BASIN

PERIOD

SUBSURFACE
NOMENCLATURE

JURASSIC

A-2 CARB.

P E N N SYLVAN IA N

M ISSISStPIAN

A-2 SALT

M A R S H A L L SS
E LLS W O R TH
ANTRIM f w

B E R E A 8S

A -l

6000'
TR A V ER SE
CROUP

D E V O N IA N

is ®

B E L L SH

DUNDEE Fw
OET. R IV E R
CRO UP

NIA6ARAN

S YLV AN IA SS

C A B O T H E A D SH,
M A N IT O U L IN P O L .
10,0 0 0 *
S A L IN A

S IL U R IA N

A -2

GROUP

RICHMOND

•A* I

U T IC A

O RDOVICIAN

SH

CABOT HE AO SH

C Q L H N CWQQD j

UTICA SH

TRENTON
BLACK RIVgR

H.

TRENTO N
B LA C K RIVER
P R A IR IE
DU C H IE N

St. PETER SS

1 5 .0 0 0 *

PRARIE du CHEIN
CA M BRIAN

Figure

2-1.

S tra ti g r a p h ic

column

of

the

Michigan

basin.

80

brine

from

chemistry

the

different

formations

in

Michigan.

in the basin ranges between Na

water as a function of formation.

Brine

and Ca dominated

These differences reflect

the different evolutions of the water within each formation,
and

might

be

explained

by

formation

history,

mineralogy,

hydrology, and geochemical reactions.
This

paper

Ordovician

and

basin.

These

history,

and

chemistry.

reports
Silurian

on

formation

aged

formations

formations

as

Figure

Silurian

differ

2-2

in

waters
in

the

their

demonstrates,

reefs

reduce

and

their

are

and

CaCl2 brine

known

Michigan

in

their

and
water

formations produced water mainly

encased

hydrologic

formations,

the

geology

Niagaran aged reefs that encircle the basin.
buried

in

produced

in

of

1945).

the
In

These deeply

that

connection

some

(Case,

salts

significantly

with
most

from

overlying

highly

contrast,

saline

Ordovician

aged Trenton and Black River formations produce water from a
major fracture system in the basin,
at

shallower

depths

than,

the

located outside of,
Niagaran

reefs.

and
The

Ordovician rocks in this area are not capped by the Salina
salts and the fault system allows ready connection to other
formations.
that
being

is

These Ordovician formations contain NaCl brine

less

more

saline

similar

than
to

formations in the basin

the

water

Niagara/Salina
from

(Figure 2-2).

the

formations,

upper

Devonian

This paper proposes

that water in the Silurian and Ordovician formations of the
Michigan basin originated from evapo-concentrated seawater,

B : B EREA
T : TRAVERSE
D ! DUNDEE
R F : RSCHFIELD
D T :D E T R O IT R .
S Y :SYLVANIA
N / S : N IA G A R A /
S A L IN A
0 : TRENTON/
BLACK RIVER

Ca
Figure

2 2.

Na

T e r n a r y d i a g r a m s h o w i n g p e r c e n t a g e s of C a - M g - N a
( m o l e p e r c e n t ) in M i c h i g a n B a s i n b r i n e s .

00

82

similar to the origin proposed for water in other formations
in the

basin

(Chapter

1) .

Both

formations

retain

unique

chemical signatures, however, the Ordovician formation brine
suggests

an

formation
their

evolution

water

by mixing,

retains

seawater

a

origin

while

chemical

and

an

the

Niagara/Salina

signature

evolution

reflecting

by

water-rock

reactions.

STUDY AREA
NIAGARA/SALINA REEFS
Formation waters studied here are produced from middle
Silurian

aged

Niagaran

carbonate

reefs

(Figure

2-1).

Because the reefs are built upon the Niagaran dolomites and
intersect the

Salina

the true producing
Therefore,
brines.

these

5000

1979).

formation

samples

Three

the reefs

with

acres,

diagenetic and

mostly

Carbonate unit within
of the water

are

termed

They

salts,

in question.

Niagara/Salina

individual

averaging

reef

are

have

highly

80

acres

been

criteria

located nearest

water,

about

types

geologic

reefs occupying

(N/S)

is

located

contain

more

(Gill,

porosity plugged with
innermost

pinnacle

dolomitized,

and

salt.

reefs

size

1979).

and

and

have
The

which

is
only

First

have

on
are

contain

little

or

The second reef
less

dolomitized.

minor

third group

contain

(Gill,

based

the basin margins which

basinward

hydrocarbons

in

from <1 to

identified

none of their porosity plugged with salt.
group

is

the

Niagaran reefs encircle the basin in a belt some 16

to 40 km wide,
over

A-l

mostly

amounts

includes
gas

or

of
the
are

83

barren,

are slightly dolomitized,

and in some cases,

their

porosity is completely plugged with salt.
REEF HISTORY
The Niagaran

reefs

have

interacted with

water types during their history
Friedman,

1977; Gill,

2-3 outlines the

(Fisher,

1977;

1979; Sears and Lucia,

Nurmi

1980).

Reef

Niagaran

Lockport

of
and

Figure

reef history and is from Sears and

(1980).

Lucia

growth began after deposition of the Middle
Formation carbonates,

stable ramp platform in the basin (3-1).
the

a variety

shallower

basin

margins,

in

deeper water as pinnacle reefs.

which

formed

a

Reefs formed along

mid-shelf

areas,

and

During reef growth,

in

frame

building organisms that formed the bulk of the reefs built
upwards in response to rising sea level.
Sometime
about

45m,

after

seawater

Dolomitization

isolation

pinnacle

levels

of

the

during this period
basin

the

reefs

lowered

reefs is

reached

heights

exposed

the

and

occurred

by a mixed water system (3-II).

Extreme

during

which

to

reefs.

have

followed

thought

time

concentrated

seawater interacted with the reefs as the Salina A-l
were deposited around the reefs
in sea level

and

deposited

first

base

(3-IV).

two stages,
levels,

and

(3-III).

conglomerates
over

the

of

salts

the
and

stromatolites,

A-l Carbonate
later

salts

A subsequent rise

allowed thinly laminated algal

mudstones,

of

around

to

the

be
reef

The A-l Carbonate was apparently deposited in
over the reefs during the periods
inter-reef

during

lower

stands.

of high
Tidal

sea

flats

84

o .

I

n

REEF GROWTH

EXPO SURE AND D O LO M lTIZATlO N
IN M IXING ZONE

o
7X AI

I

R fS T R /C T fD * A * * £

"*0*[RC*P°'

* { F l UX

«7*
m

BPIHf

ftA T

!

DEPO SITIO N OF A -l EVAPORITE

AHnrt)*!T£

12

D E PO SITIO N OF A -l CARBONATE OVER
REEFS AND IN IN TE R -R E E F AREAS

amtrDRift

rJ O d i F IA T

SA LT

S A T UR A T CD

ZJ£0K~
~ T 7
2

DOLOMITIZATION BY REFLUX FROM
A-l CARBONATE TID A L FLAT

3ZI

S a lt

m

Figure

2-3.

DEPOSITION OF A-2 EVAPORITE

S a tu ra tc d

BRmcs

COMPACTION AND SALT PLUGGING

D i a g e n e t i c h i s t o r y of N i a g a r a n
M i c h i g a n , from Sears and Lucia

r e e f s in
(1982).

85

formed

during

low

dolomitization
inter-reef

seawater

of the

areas

stands

and

resulted

upper portions of the

by

refluxing

seawater

in

reefs

and the

brine

(3-V).

Following this,

the A-2 and ensuing Salina salts buried the

reefs

Salt

(3-VI).

saturated

brines

are

thought

to

have

infiltrated the reefs during the deposition of the A-l
A-2

salts

(1977)

(3-VII)

and

suggested

deposition

of

during

that

the

A-l

salt compaction.

concentrated
salts

was

McCollough

seawater

responsible

and

from

the

the

salt

for

plugging of porosity in the Niagaran ree f s .
Although a variety of waters,
during

their

early

history,

interacted with the reefs

including

fresh-water, normal

marine, and hypersaline seawater, fluid movement through the
reefs

occurred

later

as w e l l .

Gill

freshwater entered the Niagaran

(1979)

suggested

rocks at the basin margins

sometime during the Late Silurian to Middle Devonian
and flushed the outer
2-4) .

After

migrating
reefs

fluids

that

and

cements, equant
temperature

filling

salts

hydrocarbons

were

became

time,

(Figure

removed, up-dip
entrapped

in

the

In addition, Cercone and Lohmann (1987)

late

(geopetal

reefs opening their porosity

pore

(Gill, 1979).

reported
reefs

the

that

burial

diagenetic

sediment,

pyrite,

calcite

(>80°C)

spar)

deep

formed

basinal

assemblages
bituman,
as

a

brines

result
that

in

the

dolomitic
of

high

migrated

through the Niagara and the A-l Carbonate formations.

SAL T
PLUGGED

Woter too
concentrated
to d i i s o l v a

PARTLY
REOPENED

REOPENED R E E F S

RECHARGE

1 LOCKPORT
2 S A L I NA A - l
3 S A L I N A A - 2 SALT
4 S A L I N A GROUP

hal i t e
WATER F LOW

Figure

2-4.
Proposed
model
for
freshwater
flushing
of
Niagaran
reefs
in M i c h i g a n ,
from
Gill
(1977).
Recharge
may
have
entered
the
Niagaran
rocks
along
marginal
arches,
flowed
down
into
the
b a s i n , a n d d i s s o l v e d s a l t f r o m r e e f s a l o n g the
bas in m a r g i n s .

87

SALINA SALTS
The Salina salts, especially the A-l salt, may play an
important
brines.

role

in

the

generation

of

the

Over 500m of Salina salts exist

Michigan

Basin

in the Michigan

basin in 6 individual units (the Salina A to F salts) .
salt

origin has

been

linked to the

restricted

The

conditions

caused, in part, by the Niagaran reef system described above
(Dellwig,
include
1955;

halite,

Kunasz,

nodules

Minerals

reported

in

anhydrite,

sylvite,

carnallite

1955).

1970),

(Nurmi

chalcedony

and

polyhalite
Friedman,

the

(Dellwig,

1977),

Salina

(Dellwig,

1955),

authigenic

salts

Ca-borate
quartz

and

(Gill, 1979), hematite, pyrite, and both illitic-

chloritic shale (Lounsbury, 1963).
The

Salina A-l

salt

formation in the basin

is the only known potash

(Nurmi and Freidman,

1980;

Sonnenfeld,

1985).

salts

is

equivocal. Matthews

somewhat

The

geologic

1977;

history

(1970)

bearing
Elowski,

of

and

these

Matthews

and Egleson (1974) thought the A-l salts represent the start
of

a

megacycle

of

deep-water

evaporite

deposition

in

the

basin, while Nurmi and Freidman (1977) suggested that the A1 potash salts precipitated in a highly desiccated,
marine

basin.

Figure

2-5

illustrates

the

shallow

geology

and

stratigraphy of the Salina A-l unit in relationship to the
Niagaran reefs and the Salina A-l carbonate.
(a mixture
cover

an

of NaCl
area

of

and KC1)
33,700

bearing units

km2 , and

occur

The sylvinite

of the A-l
in

multiple

salt
beds

between lm' and 4m thick that coalesce into a single bed 1520m thick near the northwest area of the basin

(Matthews and

88

t h in n e d

SYLVITE

SYLVITE

A* 1 C A R B O N A T E
A -1 P O T A S H
ZO NE
A -1 S A L T ,

5C m

2 0 km

Figure

2-5.

Salina
„ . „ ,A_1 salt s t r a t i g r a p h y
a n d E g l e s o n ( 1 9 7 7 ).
’ after

Matthews

89

Egleson,
sylvite

1974) .
is

In

thinned

the

and

northern

truncated

area
by

that intersects the A-l carbonate.

of

an

the

erosional

are

apparently

encased

the

surface

A notable feature of the

A-l salt is the lack of Mg-K-S04 evaporites,
layers

basin

completely

observation led Matthews and Egleson

as the sylvite

in halite.

(1974)

This

to suggest that

the Michigan evaporites belong to the class of highly MgS04
deficient deposits categorized by Braitsch (1971).

ORDOVICIAN FORMATIONS
Hydrocarbon
formations

and brine production

has,

until

recently,

Albion-Pulawski-Scipio
(Figure 2-6) .
fault

zone

upwarp

now

attributed
area

located
has
to

(Ells,

Trenton
composed

and
of

(wackestone)
minerals
and

This

trend

along

an

Black

synclinal

wells

in

the

form

Hydrocarbon

production

Formations,

dolomite,
1982).
are

finely
locally

both

open

rich

to

of
is
of

in

long
This

crest,

the

fault

from

the

which

are

limestone

argillaceous

interconnection

both

samples are termed the Trenton-Black River
this study.

its

crystalline

Because of their
often

35 mile

along

the

Michigan

basin.

fracturing

River

aged

along

Southern

and

fossiliferous,
and

centered

in

upwarp

dolomitization
1967).

been

is a highly dolomitized,

a general

(Taylor,

because

fault

from Ordovician

zones,

(TBR)

these

brines

in

NORTH
••

Jackson
I

s

•
i

o

^

• •* ,

^

0

;

/
^

Calhoun

^

J

'
V , J
t --------- ) - % U -

i

Branch

I

J

X1

— r 1-

- ± _

Lenawee

Hillsdale

_____I

Figure

2-6.

Sample l o c a t i o n map.
St. P e t e r S a n d s t o n e
f r o m (A), a n d the C a s e ( 1 9 4 5 ) s a m p l e f r o m

sample
(B).

91

The faulting, or reactivation of pre-existing faults in
this

area

apparently

created

dolomitized the fault zone
this

fracture

dolomites
Taylor,
thick,

occur

dolomite,

two

ferroan dolomite

in the

upper

section

basin

margins

1982).

in the

and

grades

for

1982).

fluids

generations

(Cohee et al.,

zone,

that

In addition to

other

of

1958;

approximately

of the

referred to as cap dolomite,

(Taylor,
found

A

exists

(Taylor,

in the Trenton rocks

1982).

dolomite,
the

related

channels

Trenton.

15m
This

is present only near

downward

into

limestone

A second generation of regional dolomite is

extreme western and

southwestern

area

of the

basin, where Trenton rocks are almost completely dolomitized
(Taylor,

1982).

The fracture related dolomite differs from

the cap and regional dolomites, as it is non-ferroan,

coarse

grained,

(1975)

and

suggested
between

xenotropic

that

fracture

Silurian

and

temperature of 80°C.
suggested

by

(Taylor,
related

Devonian

1982).

Shaw

dolomitization

time,

at

a

occurred

minimum

fluid

This is also the minimum temperature

Cercone

and

Lohmann

(1987)

for

the

late

diagenetic fluids that affected the Silurian reefs.
Gas

has

recently

Ordovician

St.

formations,

from

Peter
the

been

produced

Sandstone
northern

area

and
of

from
Prairie
the

the

lower

du

Chien

basin.

These

formations contain generally quartz sandstone and dolomitic
sandstones.

92

METHODS
Seven

brine

formations

were

samples

collected

southwest Michigan,
obtained

from

from

from

and were

Michigan

the St.

the

supplemented
of

trends,

reefs
and

in

a

also

collected.

from

M.D.N.R.

(1945,

listed

locations.
balance

in Table

All

errors

the

from

These
files,

Natural

and

A-2

and

Resources

the

made
than

analyses
Figure
in

2-6

this

5%,

from
reef

(#2020)

were

with

9

reported
shows
study

although

in

southern

Carbonate

supplemented

2-4).

less

of

analyses

Sixteen samples

northern

the

were

analyses
of

5

was collected from a gas well

both

sample

trend

one sample of water from

Missaukee County in central Michigan.
Niagaran

with

River

and from files of various oil

In addition,

Peter Sandstone

Trenton-Black

Albion-Scipio

Department

Geological Survey open files,
producing companies.

the

analyses
by

Case

the

sample

have

charge

some

of

the

collected analyses had charge balance errors between +6% and
+8%.

Comparison

with

analyses

made

in

the

present

study

shows that samples with poor charge balances differed mainly
in Cl.

These collected analyses were used only in graphical

treatments of the data.

l:The sample well is the Dart Edwards 7-36 (T22N, R7w, sec.
36).
The drillers log records this well as being finished
in the Prairie du Chien Sandstone.
Sibley (personal
communication) suggests that the producing formation is
really the St. Peter Sandstone, or perhaps the Bruegers
Sandstone.

93

Brine sampling was preformed in the manner explained in
detail in Chapter 1.
the well
were

Samples were collected directly

head whenever possible,

filtered

through

Field measurements

glass

included

possible, alkalinity.

allowed

wool

and

to

separate,

Watman

temperature,

pH,

from

#1
Eh,

and

filters.
and

when

The N/S samples were diluted in field

to avoid possible problems with salt precipitation,
occurrence in Niagara/Salina reef wells.

a common

At each location,

one sample was diluted 50% with 5%HN03 for cation analysis
and

another

analysis.

diluted

50%

with

distilled

H 20

for

anion

Hydrocarbon-brine mixtures would occasionally not

separate

in

the

field,

and

in

these

cases,

combined

hydrocarbon-water samples were collected in plastic bottles,
sealed,

and

allowed to

separate

in the

laboratory,

before

being treated as described above.
Because

of

the

hydrocarbon

and

salinity

characteristics,

fresh water and other additives are

introduced

N/S

build-up.

into

to

prevent

salt

and

paraffin

Every attempt was made to avoid sampling any well

injected

with

occurred.

additives,

Additional

sampling TBR wells.
hydrocarbons
several

wells

often

since

years

formation.

brine

or

to

sample

problems

were

The Albion-Scipio
the
has

early
been

1930's,

before

injection

encountered
trend has
and

re-injected

for
into

with

produced
the

past

the

TBR

The limited number of samples obtained from the

Trenton-Black

River

formation

reflects

the

difficulty

in

94

finding production wells sufficiently distant from injection
wells to insure representative samples could be collected.
The

samples

and gravimetric

were

analyzed

methods

using

reported

in

A.A.S,

Chapter

titrimetric,
1.

Five

N/S

brines and 3 TBR brines were further analyzed for the stable
isotope

ratios

Isotope

of

D/H

Laboratory,

Ontario.

and

180/160

University

of

at

the

Environmental

Waterloo,

Waterloo,

Results are reported in 6°/oo values normalized to

SMOW.

Measured

activity

and

isotopic

concentration

values
scale

were

converted

following

between

Sofer

and

Gat

(1972) .
GENERAL RESULTS
Well
listed

information,

in Appendix

compositions
waters,

of

and also

analytical,

C.

the

Table

2-1

and isotopic results are
is a summary

Niagara/Salina

lists the one St.

collected in this study.

and

of

average

Ordovician

Peter Sandstone

TBR

sample

Geometric means are listed because

the data were found to be log-normally distributed.
The Niagara/Salina brines are highly concentrated with
TDS

values

typically

in

excess

of

400

g/1.

The

high

salinity is best exemplified by the Niagaran brine reported
by Case

(1945)

that has

a T.D.S.

of

640

g/1

(Table

2-4) .

The TBR waters are less concentrated with salinities between
300 and 360 g/1.
of Na,

Ca,

and Mg

Figure 2-7 shows the relative precentages
(mole percentage)

and Ordovician waters used
that most of the N/S

in this

samples,

in the Niagara/Salina
study,

and the St.

and demonstrates
Peter Sandstone

TABLE 2-1
Average composition of Niagara/Salina and Ordovician formation
waters, Michigan Basin

Component

NIAGARASALINA

Cl
Br
Ca
Mg
Na
K
Sr
Rb
Li
B
Si

210,000
2400
74,000
11,500
31,400
9250
2190
24
60
127
2
550
90
50
25
345,000
4720
26

n h 4n
hco3

so4
I
TDS
mci2

n

ORDOVICIAN
(TBR onlv).
126,000
1030
22,800
5130
30,000
3680
750
8
38
13
3
80
40
205
11
207,000
1580
11

ORDOVICIAN
St. Peter
250,000
3100
90,100
9710
28,200
19,500
3160
58
44
107
20
247
10
-

20
404,000
5367
1

Key: All values as mg/1, except MCl2 (meq/1).
HC03 : Alkalinity as HC03
MCI-, = Ca + Mg + Sr - 0.5HC03 - S04
n = number of analyses for major components
St. Peter Sandstone sample, #8040

50%

□ NORTHERN N /S
O SOUTHERN N/S
A TRENTON-B.R.
O St. PETER
9 CASE SAMPLE

%
□

Figure

2-7 .

o

□

□

3

8? (?

*4*.

T e r n a r y d i a g r a m s h o w i n g p e r c e n t a g e s of C a - M g - N a
( m o l e p e r c e n t ) in N i a g a r a / S a l i n a and O r d o v i c i a n
a g e d f o r m a t i o n s in the M i c h i g a n b a s i n .

97

sample are Ca-Cl brine, while the TBR brines are Na-Cl.
N/S samples are separated by reef trend
and

as

can

contain

a

be

seen,

higher

the

relative

southern

reef

samples.

southern

reef

samples

elements,

discussed below.

in

these

waters,

components
Case

and

(1945)

however,

in

samples

percentage

Differences
are

some

also

of

appear

Ca

to

then

between

observed

south),
be

do

the

northern

and

in

the

other

Chloride is the dominant anion

while

samples

these

northern

(north vs.

The

S04

and

samples

alkalinity

are

below

are

minor

detection.

The

show anomously high alkalinity values,

may

be

related

to

dissolved

organic

acids

(Case, 1945; Willey et al.,1975).
NIAGARA/SALINA FORMATION RESULTS
Previous work

(Chapter 1)

has shown that the Michigan

basin brines may have evolved from evapo-concentrated water
and

in

following

samples
Figure

are
2-8

that

compared
(d-f).

observation,

with

The

the

evapo-concentrated

into

seawater

seawater concentration trend

shown are from data in McCaffrey et al.
(1978).

Niagara/Salina
in

lines

(1988) and Carpenter

The data shown in the graphical plots are separated

northern

reef

trend

and

southern

reef

trend

samples

with the mid-basin samples (sample #2020, and the Case, 1945
analysis)

combined

generally

good

with

agreement

brines and seawater

the
is

precipitation,

evident

(Figure 2-8a).

trend samples plot on the
MgSQ4

southern

while

trend

between

samples.
Cl-Br

in

the

Several of the northern

seawater line past the
several

A

samples

start

having

of

lower

98

Figure

2-8.

Niagara/Salina
formation waters
(log mg/1)
compared
with
evapo-concentrating
seawater
( d a s h e d l i n e s , d a t a f r o m M c C a f f r e y e t al. 1 9 8 8 ;
a n d C a r p e n t e r , 1 9 7 8 ) . A: L o g C l - l o g B r , B: L o g
N a - B r , C: L o g K - B r , D: L o g C a - B r , E: L o g M g - B r ,
F: L o g S r - B r .

99

NIAG ARA/S ALINA

6
5.5)

O
O.

5

o

J

-1.5

3.5
B

a

Z

-1.5

5 •
4 .5

o>
o

3 .5

5 .5

5

o
O
CT 4 .5
O

3
c 45
2
0
O» 4
S5

*

O
£

3 .5

P A
CO

o
>
o
_l
1.5

2

2 .5

3

3 .3

Log Br ( m g / l )

Figure

2 -8

100

amounts of Br plot below the seawater line.
sample

plots

above

the

seawater

trend

The Case (194 5)

line

and

has

a

Br

concentration similar to seawater concentrated to the K-salt
facies.
The

log

Na-Br

plot

also

demonstrates

the

close

agreement with seawater and a division between northern and
southern

reef

trend

southern

reef

samples,

plot

a

as

However,

cluster

samples
and

very

(Figure
some

near

of
the

2-8b).
the

Most

northern

seawater

of

the

samples,

trend

line.

several northern samples are depleted in Na or Br,

and plot as outliers

from this cluster.

This

figure

also

demonstrates how the N/S brines plot as clusters on the loglog

plots

rather

characteristic

of the

than

as

other

distinct

linear

formation waters

trends

in the

basin

(Chapter 1).
Figure

2-8 (d-f)

shows

that

Ca,

Mg,

and

Sr

co-vary

directly with Br, with Ca and Sr being highly enriched and
Mg depleted from expected seawater values,
similar

to

those

(Chapter 1).
and

Sr-Br

found

in

the

characteristics

Devonian

formation

brines

A near 1:1 relationship exists between Ca-Br
in

formation waters

most

samples,

(Chapter 1).

similar

to

other

Michigan

However, Mg values generally

plot more as a cluster below the seawater line rather than
in

a

linear

trend.

Potassium

plots

in

two

group, generally the southern reef samples,
K and plot below the seawater trend line.

groups,

one

are depleted in

The second group,

101

generally the northern samples and the Case

(1945)

sample,

plot very near the seawater trend line.
An

important

concentrated

relationship

seawater

MCl2=Ca+Mg+Sr-HC03-S04

used

brine

in

is

(in meq/1,

examining

MC12 ,

defined

Carpenter,

remains

of

approximately

affected

by

seawater,
l.

the

precipitation-dissolution,

MC 1 2

and during

MC12 (meq/1)/Br(mg/1)

MC12 is conservative

dolomitization,

as:

1978).

represents the divalent cations balanced by Cl,
evapo-concentration

evapo-

CaC03 ,

CaS04 ,

and

is not

and

and sulfate reduction.

halite
When the

log MCI2 values are plotted versus log Br (Figure 2-9),

the

N/S

are

brine

appear

to

be

either

enriched

depleted in Br from concentrated seawater.

in

MC12

or

This is also the

case for the TBR samples and the Devonian formation water in
the basin
Cl

(Chapter 1).

(Figure

seawater

2-9b), the N/S brine apparently agree with

concentration

(Figure 2-8a).
— MCI 2 may
waters,

When log MC12 is plotted versus log

appear

line,

similar

to

the

Br-Cl

the
plot

Chapter 1 discusses some of the reasons why
to

be

enriched

or

Br

depleted

in

these

but the agreement with seawater in both the MC12-C1

and Br-Cl

plots

suggests

that MC12 values are enriched

in

these waters.
ORDOVICIAN FORMATION RESULTS
The
seawater
the

Ordovician
in Figure

seawater

Cl-Br

(Figure 2-10a).

formation
2- 1 0 (a-f).
line

and

A similar

samples

are

compared

with

The TBR samples plot below
extend

down

relationship

is

toward

seawater

found when

the

102

Log

MCI2

(meq/1)

4.5
4

3.5
3
2.5
1.5

2

2.5
3
3.5
Lo g Br ( m g / l )

4

4.5

Log Cl(mg/I)

6

5.5
□ NORTHERN N / S
O SOUTHERN N / S

5

A T R E N T O N -B . R .
4.5

• St. PETER
O CASE SAMPLE

4
2

Figure

2-9.

2.5
3
3.5
4
Log MCI2 (meq/1)

4.5

L o g M C I 2 (meq/1) vs. l og b r (mg/l) a n d log M C I 2
(meq/1) vs. log Cl (mg/l) in t h e N i a g a r a / S a l i n a
and O r d o v i c i a n form ati on waters, compared w i t h
evapo-concentrated
seawater
(dased
line,
data
from
McCaffrey
et
a l . , 1988;
and
Carpenter,
1978)

103

Figure

2-10.

Trenton-Black
River
and
St.
P et e r
Sandstone
formation
w a t e r (log mg / l) c o m p a r e d w i t h e v a p o c o n c e n t r a t i n g s e a w a t e r ( d a s h e d l i n es , d a t a f ro m
M c C a f f r e y et al . 1988 ; and C a r p e n t e r , 1978 ). A:
Log C l - l o g Br, B: Log N a - B r, C: Log K -B r , D: Log
C a - B r , E: Log M g - B r , F: Log Sr-Br.

104

ORDOVICIAN

4.5
5.5

D

«>3 5

5.5

4.5

• TRENTON
O S t. PETER

3

2

1.5

3.5

L o g B r (mg/l)

Fig u r e
V

2-10.

4

105

other cations are plotted versus Br, and in the MCL2-C1 plot
(Figure 2-9b) .
this

study

Peter

to plot

brine

samples,

The TBR

does

but

are the only samples

consistently
not

rather

plot
it

in this

with

the

resembles

collected

manner.

The

Trenton-Black

many

of

the

in
St.

River

N/S

brine

samples by closely resembling seawater at high Br content.
ISOTOPIC RESULTS
As

a part

of

this

study,

the

stable

isotopic

ratios

180 / 160 and D/H in 7 Niagara/Salina samples and 5 Ordovician
formation samples were measured.

The results are listed in

Appendix C, and are plotted on Figure 2-11.
the global-meteoric water line (GMWL)
the best

fit

line

for all

and

to

the

right

of

from Craig (1961), and

Michigan basin

from this study (Chapter 1).

Also shown is

formation waters

The N/S brine data plot below

the meteoric

water

line,

with

most

samples plotting towards the enriched (positive 8 18o values)
end of the Michigan basin best-fit

line,

from the

(Richfield

lower

Devonian

formations

River formations, Chapter 1).
and

the

St.

Peter

intermediate

and

Detroit

The Trenton-Black River water

sample (#8040)

between

similar to water

the best

plot

fit

in

line

a

tight

and

cluster

the

global

meteoric water line.
GEOCHEMICAL EVOLUTION OF NIAGARA/SALINA BRINES
The
lines,

Cl/Br

agree

suggesting

well with

that

originated

from

previously,

the N/S

the

the

Niagara/Salina

evapo-concentrated
samples

seawater

concentration

formation

seawater.

appear to plot

as

As

brine
noted

clusters

on

106

50

(MOWS)aeiap

MICH.
BASIN
LINE

-50
AMMW

• NIAGARA/SALINA
O T R E N T O N - B L A C K R.'
O S t . PETER

100
-10

0

del 1 8 0 (S M O W )
Figure

2-11.

del D /oo S M O W ( a c t i v i t y s c a l e ) vs. del
0 °/oo SMOW
in
the
Niagara/Salina
and
Ordovician
formation
waters,
M i c h i g a n basin.
Also
shown
is
the g l o b a l
meteoric
water
l i n e ( G M W L ) f r o m C r a i g ( 1 9 6 9 ) , a n d the b e s t - f i t l i n e to all
M i c h i g a n b as i n f o r m a t i o n w a t e r s c o l l e c t e d in this s t u d y .

107

log

concentration

plots,

perhaps

suggesting

that

formation contains water with a more homogeneous
than

observed

in

the

However,

differences

northern

reef

suggesting

and

that

distribution

Devonian
in

of

chemistry

southern
the

formations

reef

are

sampling

chemistry

(Chapter

observed

Niagara/Salina

homogeneity

reflects

locations

for

this

between
samples,

the

the

1).

limited

N/S

water

compared with the wide distribution for the Devonian.
Although

the

Cl/Br

suggest

an

evaporative

seawater

origin, differences exist between Ca, Mg, Sr, K and MC12 in
the samples when compared with seawater.
reflect

the

geochemical

evolution

of

These differences

the

water

resulting

from water-rock reactions or mixing.
Ca-Mq-Sr
Dolomitization appears to be one of the more important
reactions in the evolution of Michigan brine, as it explains
the

Ca

and Mg

in these

enriched in Ca
concentrated

The N/S

brine

is

highly

(8d), while Mg is depleted from equivalently
seawater

dolomitization
demonstrate

waters.

(Figure

presented

that

the

in

2-8e).

Chapter

deficiency

in

The
1

Mg

can
from

model
be

used

for
to

concentrated

seawater can explain most of the measured Ca, based on a 1
to

1

mole

replacement

during

dolomitization.

The

ca

concentrations predicted by this model are compared with the
measured values
for the

most

in Figure

part,

agree

2-12.
with

The predicted
the

measured

Ca

Ca values,
supporting

that dolomitization by seawater explains the Ca and Mg

in

5.5

t

1 "i— i

1 " i■ i

r -■-?— |— -t ~~T " I

r

B

W

□
o

g 4.5

o

<0

O

4

W

□ MEASURED
O PREDICTED

■

2.5

3

3.5

Log Br (mg/l)
Figure

2 - 1 2 . R e s u l t of d o l o m i t i z a t i o r m o d e l ( l o g C a v s . l o g
B r , m g / l ) for N i a g a r a / S a l i n a f o r m a t i o n w a t e r s .
S qu ar e s ymbols are m e a s u r e d Ca values, c ir c l e s
are predicted C a .

108

o
"* 3 5

109

this water.

For the Case

(1945)

few lesser concentrated brines,
predicted values.

A

similar

Michigan formation water

analysis as well as for a

an excess of Ca exists over
excess

is

observed

in

other

(Chapter 1), and is interpreted to

show that non-stoichiometric replacement of Mg for Ca during
dolomitization has occurred,
Mg must be considered.

or that other sources of Ca or

Other Ca sources might include a Ca

enriched parent seawater or the addition of Mg by exchange
or clay diagenesis.
Strontium is highly enriched in N/S brines over what is
expected for evaporating seawater
plot

in

a

cluster

very

near

(Figure 2-8f), and values

the

1:1

line.

Sass

and

Starinsky (1979) suggested that the dissolved Sr/Ca might be
used to
that

indicate the type of carbonate mineral

supply

0.009

to

Sr

0.022,

Starinsky

(1979)

to

brines.

which

The

according

Sr/Ca
to

or reaction

(molar)

range

the model

of

Sass

from
and

suggested that dolomitization of aragonite

and solution-reprecipitation of calcite have supplied Sr to
these waters.
Thermodynamic calculations also support that the waters
are

in

halite,

equilibrium
and

saturation

in

some

indices

calculated only

with

dolomite,
cases,

(log

along

sylvite.

IAP/Ksp)

for

indices were computed using the PHRQPITZ
uses

anhydrite,

Histograms
these

of

minerals,

for the N/S samples collected and analyzed

in this study, are shown in Figure 2-13.

which

with

Pitzer's

equations

for

Mineral saturation
computer routine,

activity

coefficients

DOLOMITE

HALITE
30
25

30
20

3 20

10
5

tn
1 -0. B -0.2

0. 2

0
-3

0.6

2

0

LOG 1AP/KSP

2

LOG 1AP/KSP

ANHYDRITE

SYLVITE

2D •

u 15 •
c
01

D
t
r
01

£10

■

5

fl rilTh

0

-3-2-1

0

rn_p
1

LOG 1AP/KSP

Figure

2-13.

2

3

-A

3

-2

0

LOG 1AP/KSP

H i s t o g r a m s of s a t u r a t i o n i n d i c e s ( l o g I A P / K s p )
f o r the N i a g a r a / S a l i n a w a t e r s a n d T r e n t o n - B l a c k
River samples .

Ill

(Plummer et

al.,

were

on

based

thereafter
makes

no

1989).
10°C

(Chapter

at
1;

adjustment

alkalinity data were

Estimated
30m,

and

formation
a

gradient

Vugranovich,
for

1986).

pressure.

first used

temperatures

The

of

23°C/km

The

program

measured

in the modeling,

pH-

and

then

the modeling was repeated assuming that calcite equilibrium
controls pH and alkalinity.

A similar range of saturation

indices are calculated under both conditions,

although

average dolomite saturation indice is reduced slightly,

the
from

0.49 to 0.06, when calcite saturation is assumed.
Potassium
Potassium values plot below the seawater concentration
trend line for many of the samples

(Figure 2-8c) , which

similar to how many of the other Michigan basin
waters plot

(Chapter 1).

the N/S brines however,

is

formation

The K depletion is less severe in
and does not occur in all samples.

The depletion of K from formation water is often interpreted
to represent illitization of clay minerals upon deep burial.
However,

depletion

seawater brine.

may

For

start

example,

early

in

Figure

the

2-14

history
shows

the

of

a

K-Br

concentrations in a seawater sabkha brine from Laguna Madre,
Texas

(Long

and

Gudramovics,

3.983) .

This

water

was

collected from the upper few meters of the sabkha sediments,
and exhibits a depletion of K from seawater values,
the

result

of

reactions

with

clay

minerals.

perhaps

This

data

illustrates that the chemistry of a seawater brine may start
to evolve soon after entering sediments,

and that

seawater

t

5

i

n

r

4.5
*«■*
\

05

4

E

Z

3 -5

,
5

112

05
o

-j

2.5

SEAWATER

2
1.5

!_____
2

!_________ I_________ I
2.5
3
3.5

4

Log Br (mg/l)

Figure

2-14.

L o g K v s. l o g Br ( m g / l )
in s e a w a t e r b r i n e s
from
Laguna
Madre,
Texas
(Long
and
G u d r a m o v i c s , 1983)

113

in

natural

settings

may

quickly

deviate

from

the

evapo-

concentrated seawater chemistry measured in salt pans.
The depletion of K from the Niagara/Salina brine might
be due to

illitization and authigenic

feldspar diagenesis.

The Niagaran rocks and the Salina salts contain illite and
chlorite in shale partings and as thin,

black shale

in

1963;

the

Nurmi

lower
and

Salina

Friedman,

observed

in

the

rocks

(Lounsbury,

1977).

Authigenic

Trenton

Nowak,

feldspar

formations,

layers
1978;

has

(Sibley,

been

personal

communication), and may be present in the Niagaran rocks as
well.
MASS BALANCE MODEL
The extent

to which these mineralogical

explain the brine
balance model

reactions

chemistry can be evaluated using

(Collins,

1975;

Carpenter,

1978).

can

a mass

The model

starts with seawater and attempts to derive the composition
of

average

sample

N/S

(Table

degrees

(based

brine
2-6a)
on

and
from
Br)

a

southern

seawater
by

removal

by

reduction

(or

resulting

from

CaC03

Although

will

it

calculated

used,

CaS04

dolomitization),

precipitation.
be

by

is

for an actual

the

(#2098)
to

equal

dolomitization,

aluminosilicate diagenesis,

precipitation

equally,

trend

concentrated

considering

recrystallization of aragonite,
Ca

reef

resulting
mineral

and

from

precipitation

halite

"average"

important

that

sample.

Sample

dissolution-

brine
an

sulfate

composition

evolution

#2098 was

be

chosen

114

because

its

Br

content

is

similar

to

the

average

N/S

composition.
Many
type,

assumptions

the

most

for

might

halite

Carpenter,

being

that

seawater

1987)

include

clay

a

account.
by

model

this

the

only

exchange,

recrystallization
and

the

the

interaction

analytical

error

Other
organic

(Stoessell
of

saturated brine (Wilson and Long, 1984).
such

is

of

and that Br acts conservatively.

Br

breakdown,

into mass balance models

important

source of bromide,
sources

go

and

halite

with

To correctly apply

must

be

taken

into

The two analyses were corrected for charge balance

distributing

the

error

proportionally

among

components

relative to their concentrations (corrected compositions are
also listed in Table 2-2).
initial

seawater

composition

solution of the brines.
this
al.

example

Finally,

it is assumed that the

represents

the

actual

parent

The seawater compositions used

are taken directly

from data

in McCaffrey

in
et

(1988).
The

mass

balance

results

(Table

2-2)

show

that

each

liter of average N/S brine might have evolved from seawater
by the formation of 192g of dolomite,

the recrystallization

or dolomitization of 213g of aragonite (containing 10,OOOppm
S r ) , and the formation of 56g of illite.
followed either of two evolutions.
the

precipitation

35,840mg of Cl.

of

59g

of

halite

Alternatively,

Na and Cl may have

In order to explain Na,
is

required,

the dissolution

removing
of

51g of

halite can be used to explain Cl, but in this case, 43,300mg

115
TABLE 2-2
EVOLUTION OF NIAQARA/BALINA BRINE

Ll
Component
Br
Cl
Ca
Mg
Na
K
Sr

Average
N/S brine
2400
210,000

74,000
11,500
31,400
92,500
2190
50
90
4720

so 4
HCO 3
mci2

Average
corrected
2420
215,900
72,000
11,200
30,560
9000
2130
50
90
4562

#2098
2560
230,000
79,600
16,700
40,200
8780
2300
35
350
5393

#2098
corrected
2650
247,800
77,200
16,200
39,000
8500
2230
35
350
5233

Corrected: corrected for charge imbalance, all values in mg/l,
except MCI2 (meq/1).

lb.
Evolution of average Niagara/Salina brine (values in mg/l)
______________ Ca______Mg______Sr
Seawater
(Br=2440)
Dolomitization
(192g)

48,200

14,500

53,800

185,000

2820

1.1

5500

Halite
(59g)

-64,350
-23,240

36,000
62,900

11,200 2130

Average...
brine

72,000

11,200 2130

Excess a)

36,200
9120

b)

64,400

2130

(**)C03
-26,870
precipitation resulting from S04 reduction

Predicted
Brine a)..
b)..

MC12/

61,010 -37,000

Recrystal-970
ization of Aragonite
(*)Illite
diagenesis
(56g)

K_______Na________ Cl_____ S04

9000

30,560

-35,840

149,200

50

1.1
1.7

9000

30,560

215,960
66,800
66,800

50

1.9

116

TABLE 2-2 (Cont'd.)

Ci.

Evolution of Sample #2098 (values in mg/l)
________ :
______ Ca________ Mg____£r_____K________ Ha______ Cl
Seawater
(Br=2630)

52,600

15,500

46,900

185,500

SQ4

MC12/Br

70,300

1.0

Dolomitization 60,025 -36,400
(189g)
Recrystal-1020
ization of aragonite
(*)Illite
diagenesis
(71g)

2230

3590

7000

(**)C03
-29,340
precipitation resulting from S04 reduction

-70,300

Halite
-7900
-12,200
(20g)______________________________________________________________________ _
Predicted
brine a)
b).....

33,300
62,600

16,200 2230

8500

39,000

173,300

35

1.2
1.8

Measured
brine

77,200

16,200 2230

8500

39,000

247,800

35

2.0

Excess a)....
b ) ....

43,900
14,600

74,500
74,500

Key: Seawater: Composition of seawater based on measured Br.
Predicted brine is sum of reactions.
MC12/Br: MC1 2 (meq/l)/Br (mg/l), where MCl2=Ca+Mg+Sr-S04-0.5HC03
(*) : The aluminosilicate reaction used is:
2K^ + 3Al2 Si 20 5 (0H)4 + CaC03 =
Ca
+ 2KA121AlSi3 )0j0 (OH) 2 + C02 + 4H2<3
(**) : Amount of Ca removed by CaC03 precipitation
resulting from S04 reduction.
Predicted Ca values:
a): concentration when
resulting from S04
b ) : concentration when
resulting from S0 4

Ca is removed by CaC03 precipitation
reduction.
Ca is not removed by CaC0 3 precipitation
reduction.

Excess: measured amount minus predicted amount for case a and b.

117

(1.88m)
the

of Na must be removed from solution to account for

measured

Na.

This

later

evolutionary

scheme

is

not

listed in Table 2-2, for reasons discussed below.
A

similar

evolution

dolcmitization of
223g

of

aragonite,

The

Na

precipitation

of

calculated

189g of carbonate
and

which is about 1.5
sample.

is

for

rocks,

the formation

of

sample

#2098,

recrystallizing
71.3g of

illite,

times the amount required by the average
chemistry
20g

might

of

be

halite,

explained

or

by

the

alternatively,

if

dissolution of 103g of halite is used on to balance Cl, then
48,300mg of Na must be removed to account for the measured
Na (this not listed in Table 2-2).
These reactions

appear to account

for most

components

in the brine,

however,

a large excess of Ca and Cl need to

be explained,

and

role

the

of halite

dissolution must be established.
a

Ca

excess

predictions

of

during sulfate reduction,
or

by

(0.9m/l)

Ca

over the

model

is removed with S04

(either

CH20 + S04= + Ca2+ =

CaS04 mineralization

during

precipitation

66,800mg/l

(1.88m/l),

moles/1 excess
is

not

in Ca.

accounted

concentrated
illitization,

(to balance

for

in the

removal,

+

is

in

balancing

excess
the

by
0.9

a significant amount of CaCl2

seawater has been
sulfate

M g 2+

In the evolution by

Na) , Cl

approximately
Thus,

CaC03+ H 2S ;

dolomitization,

2CaC03 + S042- = CaMg(C03 )2 + CaS04 ) .
halite

versus

The average N/S sample has

36,000mg/l

(line a) , when

precipitation

evolution

involved

scheme

where

in dolomitization,

and halite precipitation

118
A

decrease

explain

in

some

the
of

amount

the

of

extra

Ca

Ca,

removed
but

not

by

sulfate

the

extra

might

Cl.

A

possible source discussed below may be the input of a CaCl2
rich brine, which would have caused the halite precipitation
to occur (Holser, 1979; Braitsch,

1971).

The possibility remains that Cl should be balanced

in

the model by halite dissolution, with removal of the added
Na by exchange
composition,

onto

exchange

dissolution could
the

predicted

predicted

Ca

72,OOOmg/1.
halite

clays

add

Ca

for Ca.

of

Na

For the
generated

37,800mg of Ca,

(line

a;

from

which when

36,OOOmg/1),

very

near

the

average

Thus,

it

would

appear

dissolution

average

accompanied

that

added to
in

composition
an

cation

evolution

a
of
by

exchange

could

This evolution however,

would

require the dissolution of a large amount of halite,

5lg/l

explain the brine evolution.

by

halite

results

brine

brine

of halite for the average brine and 100g/l for sample #2098.
Dissolution of halite

is unlikely because: (1)

a

seawater

brine saturated with halite was used as the parent solution,
and would not be able to dissolve halite,
Na/Br

ratios

do

not

affected the brines,
support

that

formation

and

support
and

meteoric

halite

dissolution

has

(3) the stable isotope data do not

water

dissolved

that

(2) the Cl/Br and

or

salt.

seawater
The

has

evolution

entered
calling

the
on

halite dissolution

is therefor dismissed and the

by dolomitization,

S04 removal by either reduction or CaS04

precipitation,

aluminosilicate

reactions,

and

evolution

halite

119

precipitation

associated

with

an

input

of

CaCl2

is

suggested.
ENRICHMENT IN CaClg
Although

evapo-concentration

of

seawater

may

explain

the origin of these brines, the enrichment in CaCl2 suggests
that

a

more

complex

evolution

must

be

considered.

This

enrichment in CaCl2 also represents the enrichment in MC12
observed in Figure 2-9.

The N/S samples, on average, have a

MC12 (meq/l)/Br(mg/l)

of about

2 with values

2.7,

expected

ratio

far

seawater
MC12

from

the

(Carpenter,

197 8).

(as CaCl2 ) is by the

carbonate

minerals;

CaCl2 (aq)

+

1978).

of

1

ranging up
in

One mechanism

evaporating

for

increasing

reaction of aluminosilicate

2KC1(aq)

+

CaC03 +

2KA12 (AlSi3)01 Q (0H2) +

3Al2Si2C>5 (OH) 4

4H20 +

C02

to

and
=

(Carpenter,

Similar reactions can be written using Na-feldspars.

The maximum CaCl2 enrichment obtainable from these reactions
occurs

when

all

K+

in

eguivalent amount of Ca
maximum value
removed

from

of

1.2.

the

24*

seawater

is

replaced

with

an

, and would increase MCl2/Br to a

Potassium

Michigan

brine,

has not

been

so the

upper

completely
limit

of

MCl2/Br=l.2 helps illustrate the extreme enrichment of MC12
in these waters.

Exchange of Na for Ca on clay minerals,

non-stoichiometric
dolomitization,

replacement

of

considered

are

for

Mg

during

and lack of complete removal of Ca as CaS04

minerals are possible explanations
balanced by Cl.

Ca

for some of the extra

Ca

Three other sources of CaCl2 that can be
the

diagenesis

or

dissolution

of

MgCl2

120

evaporites,

and a pre-existing enrichment in CaCl2 in the

seawater parent to the brines.
The

diagenesis

or

dissolution

of

evaporite

minerals

such as carnallite might generate MgCl2 solutions that would
be

subsequently

altered

to

CaCl2

by

dolomitization.

Although previous work has suggested the potash deposits in
Michigan are primary salts (Matthews and Egleson,
origin

may

considered

be
a

equivocal.

primary

Sylvite

seawater

evaporite

"dynamo-polythermal precipitation"
Draitsch,

1971;

Garrett,

1970;

is

1974), the

not

generally

mineral

during

(Borchert and Muir,

Sonnenfeld,

1985),

1964;

although

some potash deposits are suggested to be primary in origin
(Sonnenfeld,

1985).

Recent thermodynamic modeling by Harvie

and Weare (1980) supports that sylvite is not primary during
normal seawater evaporation conditions,
and

fractional

crystallization

both in equilibrium

evaporation

schemes

schemes do not consider carbonates, however).

In short, the

"physical chemistry of evaporating seawater does not
sylvinite

(KC1 plus NaCl)

normal circumstances"
Sylvite

can

6H20,

or

undersaturated
may

go

before

through

(Garrett, 1970).

also

form

by

the

thermal

the

seawater

(Garrett,

1970).

of

reactions

cycles
as

breakdown

of

KMgCl3 *6H20 = KCl^s j + MgCl2

through

stabilizing

favor

being directly precipitated under

carnallite during burial:
+

(both

leaching

these
sylvite

of

carnallite
Potash

deposits

during

(Sonnenfeld,

by

burial

personal

communication) , which liberate large amounts of MgCl2 brine

121

and heat

(Garrett,

1970).

These MgCl2 solutions would have

been quickly altered to CaCl2 upon reacting with carbonates
minerals.
Whether the sylvite is a primary or secondary mineral
deposit

in Michigan

geochemical
(1974)

study

suggested

is

of

not

clear.

these

the

Br

In

salts,

content

the

only

Matthews

of

the

reported

and

salts

Egleson
show

the

sylvite is primary, but other interpretations of their data
are possible.

The Br

in the sylvite

of the

one

reported

core from the mid-basin area in Michigan decreases from 4 340
to

2320

ppm

over

the

top

0.5m

of

underlying 13m of the potash core,
potash averages 2100 ppm.
Kuhn

(1968)

reports

for

the

core.

however,

In

bromide

the

in the

These values are similar to those
descendent

which are listed in Table 2-3.

or

secondary

sylvite,

A more detailed examination

of these salts is warranted, but the possibility exists that
the

sylvite

in

Michigan

may

have

originated

from

a

carnallite precursor.
A

second

source

of

MgCl2

fluids

would

incongruent dissolution of the carnallite.
the stratigraphic relationships

of the A-l

be

from

Figure 2-5 shows
sylvinite

in the basin, which is also shown in cross section.
northern

area

of

the

basin

the

the

sylvinite

is

salts
In the

thinned

by

erosion and is truncation by an erosional surface of the A-l
Carbonate.

Interaction

concentrated

seawater

and

would

have

of

would

altered

the
have
any

carnallite
dissolved
remaining

with
the

lesser­

carnallite

carnallite

to

Table 2-3
Bromide concentrations characteristic of potash minerals
Genetic
Type_______

Description

Carnallite
Br CPPJIlL

Sylvite
Br .(ppm).

Primary

Crystallization
from seawater

3000-5000

3000-4000

Descendent

Crystallization
from second cycle
brines

1700-3000

1500-3000

Secondary

Alteration at depth 1000-2000
by brines and
temperature

1000-2000

Posthumuous

Late alteration
by groundwater

1000-1700

Not applicable

From Kuhn

(1968)

123
descendent sylvite

(Garrett,

1974;

Krauskopf,

1979).

Both

the dissolution and alteration of carnallite would produce
MgCl2-rich brine.
In

order

to

estimate

the

composition

of

the

fluids

derived from the alteration or dissolution of carnallite and
involved in dolomitization, the composition of a solution in
equilibrium
dolomite,

with

and calcite was

(Plummer et
using

a

carnallite,

a l . 1989)

temperature

formation temperature.
carnallite-halite

halite,

calculated.

was

of

sylvite,

used

50°C,

The

to make

the

anhydrite,

PHRQPITZ

this

approximate

model

calculation,
present-day

Bromide was estimated based on the

mixture

being

the

only

source

of

Br

(please see discussion at the end of this paper) , using an
initial Br content of 3000 mg/kg in the carnallite, which is
within the

range for primary carnallite

(Kuhn,

1968).

PC02

was not controlled in these calculations.
The results of this calculation are listed in Table 25, and suggest that this hypothetical bittern produced from
potash salts is enriched in Ca, Na and Cl over concentrated
seawater, characteristics common to many N/S samples.
importantly,

the

approximately 2.5,
many

of

the

N/S

Although

the

MC12

calculated
which

has

and
is

the

Case

variable

concentration chosen for the carnallite,
support that
mechanism.

a

MCL2/Br

is also similar to the values

samples
value

bittern

More

enrichment of CaCl2

(194 5)

based

on

of
in

sample.
the

Br

these calculations

(MC12 ) can occur by this

TABLE 2-4
Modeled brine derived from potash salt
compared with N/S brine and seawater
Component

Modeled

Case-1

Case-2

pH
aH20
TEMP °C

6.18
0.49
50

3.55

3.55

65a

65a

Ca
Mg
Na
K
Cl
so4
ALK
Br
mc12
TDS
MCl2/Br

158,800
20,470
9400
44260
395,000
5
36.8
3830°
9607
631,522
2.7

253,000
8830
26
26,200
494,600

212,600
10,100
4878
21,800
432,900
nil *
2000°
4068
9550
579328
2.3

nil

1482
4290
10867
642798
2.5

Vi

Seawater

nil
84,200
17,100
21,900
208,000
103,000
nil
4180
4782
438,000
1.14

Key: All values as mg/kg, except MC12 (meq/kg).
Modeled brine
is a solution saturated with
sylvite,
carnallite, halite, anhydrite, dolomite, and calcite.
Case
samples are from the Niagara/Salina formation, from the
Gulf Salina-1 well, sec. 34, T. 15N, R. 4E.,Michigan,
permit # 10551.
Case-1 sample was collected from a drill
stem test, and
the Case-2 sample was collected at the
surface from the flowing well. From Case (1945).
a: Estimated temperature based on depth,
b: Reported to include organic acids,
c: Estimated, see appendix.

This

example

is

intended

only

to

point

out

that

the

enrichment in MC12 (as CaCl2 ) may be the result of reactions
involving potash salts,
that

the

Niagara/Salina

from these salts.
based

although

on

formation

water

originated

solely

This is most likely not the case because

present-day

insufficient

it is tempting to suggest

volumes

sylvinite

of

brine

volumes

would

be

(Chapter

generated.

1) ,
The

Salina A-l salt contains between 510 and 720 km3 of sylvite
(depending on %KC1 chosen), representing an average of 2790
km3 of primary carnallite.

The isothermal breakdown of lm3

of carnallite produces 0.22m3 of sylvite,
622

and approximately

liters

of water (Borchert and Muir, 1964), therefore,
1C
,
some 1.7x10
liters of water may have been liberated from
the present-day A-l

sylvite.

This volume

represents

less

than 25% of the total pore volume in the underlying Niagaran
formation
assuming

rocks
a

formation

similar

porosity

brine,

originated

(estimated

from

of

such

chemistries

10%) .

as

Salina
and

volume

the

of

If
St.

seawater

the

Peter
(as

salinities),

volumes of solution are required.

7.9xl04
deep

km3

Ordovician

formation

suggested
then

and

by

even

brine,
their
larger

While it does not appear

realistic that the Niagara/Salina brine originated only from
the breakdown of carnallite, they may have been supplemented
by this process.

A third source of CaCl2

(MC12 ) could be enrichment of

the parent seawater of the brines,
concentration.
explain

Enrichment before evapo-concentration might

for example,

between

MCl2-Br

criteria,

occurring before evapo-

in

the

strong

Figure

parallel relationship

2-9.

It

is

not

clear

what

if any, in brine chemistry can be used to test the

hypothesis that the brines originated from a CaCl2 enriched
parent
might

seawater.
be

potash

However,

consistent

salts

with

proposed

an

"anomalous” parent

the

by

model

of

Matthews

the

and

seawater

Silurian

Egleson

A-l

(1974).

Anomalous is used here to signify a difference in chemistry
from the
is used
taken

seawater measured
in this

from

Matthews

study

McCaffrey

deficient

in MgS04

for the
et

and Egleson

in evaporative salt pans,

al.

reference

(1988)

and

(1974) suggested

because of

seawater

the

lack

associated with the A-l potash unit.

and

Carpenter

that

which

(1978) .

seawater
of MgS04

In "normal"

was

was

salts

seawater

evaporation schemes, the evapo-concentration from the halite
to KC1 facies results in the deposition of large volumes of
MgS04

salts

originated

(Garrett,
from

a

1970).

"normal"

In

fact,

seawater,

then

if

the

potash

roughly

equal

volumes of MgS04 and KCl salts should be found in the basin.
The

lack

layers,

salts

on

either

side

of the

sylvite

and the thin or non-existent anhydrite layers which

separate
Carbonate
1977),

of these

the

A-l

salts

from

the

(Matthews and Egleson,

suggest

that

seawater

in

Niagaran

1974;
the

Nurmi

and
and

Michigan

the

A-l

Friedman,
basin

was

127
deficient
Egleson,

in

MgS04

1974).

during

the

Silurian

(Matthews

and

One explanation for the lack of S04 salts

is that dolomitization and sulfate reduction occurred near
the basin margins,

and perhaps outside of the basin margins

to deplete MgS04 from the seawater that entered the central
Michigan basin during the Silurian.

While these

reactions

would not enrich the seawater in absolute amounts of CaCl2 ,
other

water-rock

aluminosilicates,

reactions,
may

have

increased

evapo-concentration occurred.
"anomalous"
salts

those
MC12

involving

values

before

If the seawater chemistry was

and produced primary sylvite without the MgS04

(carnallite)

earlier,

perhaps

then

it

and
may

the

have

MgCl2

also

solutions

been

enriched

described
in

CaCl2 .

However, an "anomalous" seawater is not required to explain
the

evaporite

example,

salts

proposed

in

the

a model

basin.

for the

Garrett
origin

of

(1970)

for

sylvite

that

calls on the direct precipitation of carnallite from normal
seawater, without the associated MgS04 salt.

The evolution

of the potash salts in Michigan may have been quite complex,
but

it

is

precursor,

possible
or

they

that
may

they

have

evolved

evolved

from

from

an

a

carnallite
"anomalous"

seawater source.
ISOTOPIC EVOLUTION
Two general models for the evolution of stable isotopes
in sedimentary basin brines have been advanced.
al.

(1966)

basin waters

interpreted
to

the

represent

stable

the

isotope

Clayton et

composition

of

flushing of meteoric waters

128
through
host

the

basin,

formation

along

with

carbonate

isotopic

minerals.

presented by Knauth and Beeunas
1964),

is

that

represents

the

the

dilution

meteoric water.
however,

isotopic

This

of

equilibration

An

(1986)

with

alternative

and others

composition

of

latter explanation

(Degens,

basin

evapo-concentrated

view,

brines

seawater

by

is accepted here,

it is not possible to correlate isotopic dilution

with chemical dilution in these waters.

The Niagara/Salina

samples do not exhibit either isotopic or chemical evidence
of dilution.

A conclusion that can be reached is that the

N/S waters retain an isotopic composition reflecting evapoconcentrated seawater (Chapter 1).
A

number

interpreting

of

the

factors
isotopic

need

to

composition

thought to be derived from seawater
Beeunas,

1986) .

seawater

is

The

isotopic

controlled

by

be

considered

of

formation waters

(Chapter 1; Knauth and

composition

many

when

of

variables,

evaporating

and

as

such,

seawater may exhibit a wide range of isotopic values during
evapo-concentration.

This

Beeunas

reported

(1986),

inclusions,

who

Pierre

et

al.

was

illustrated
values

(1984),

and

by

from

Knauth

halite

Holser

and

fluid

(1979),

who

reported values from sabkhas and lagoon waters concentrated
up to halite saturation, and Nadler and Magaritz

(1979), who

reported

evaporation

values

for

Mediterranean

seawater

pans concentrated

past halite saturation.

found

isotopic

that

the

composition

in

This later study
of

seawater

after

reaching halite saturation is more negative in SD than the

129

seawater source.

The Niagara/Salina brine

far past the start of halite saturation,
have

evolved

unlike

the

from

a

examples

seawater
cited

brine

above.

is concentrated

and therefore,

having

a

may

composition

Additionally,

ancient

seawater may have had an isotopic composition different from
present

seawater.

For

example,

Brand

and

suggest that Ordovician seawater was about
1ft
in ■LO0 than present ocean water.
an

extent

then

the

seawater

shifted towards more negative
In sum,
the

Veizer

(1980)

-5.5°/oo lighter

.
If seawater varied to such

evaporation

path

would

have

S18o, and perhaps £ D values.

it is not possible to constrain with any certainty

isotopic

composition

of

the

parent

seawater

of

the

Niagara/Salina brine.
It is also difficult to imagine that the basin waters
have not reacted during their history with the carbonate and
evaporite
function

minerals
of

ubiquitous

water-rock

to

ratios,

the

basin.

carbonate

Although

reactions

a

would

enrich the brines in 180 (Land, 1982; Clayton et a l ., 1966).
If

isotopic

isotopic

equilibrium

composition

of

can
the

be
water

demonstrated
can

not

then

be

the

entirely

primary.
The extent of

isotopic equilibrium between the brines

and carbonate minerals can be evaluated in a general manner.
Such an evaluation was

attempted by Clayton

et a l . (1966)

who attempted to demonstrate that Michigan basin
waters are in isotopic equilibrium with calcite.
al.

(1966),

however,

made

assumptions

formation
Clayton et

concerning

mineral

130

180 values

and formation temperatures and did not consider

that dolomite
isotopic

equilibrium may play a role

composition

of

the

water.

in altering the

Recent

data

help

constrain the temperature and mineral isotopic compositions
in the basin and allow for a more refined re-examination of
this question.
Determining
dolomite

is

if

of

isotopic

course

equilibrium

hindered

because

exists

of

the

with

lack

of

knowledge on oxygen fractionation during the low temperature
diagenesis of dolomite
(1980)

discusses

describe

(Land,

four

1980? Carpenter,

equations

dolomite-water

(Table

oxygen

1980).

2-3)

isotope

Land

thought

to

fractionation

developed by Northrop and Clayton (1966), O'Neil and Epstein
(1966),

Shepard

(1970).

The following average fractionation factor equation

can

calculated

be

and

Schwarcz

from

(1970),

these

four

and

Fritz

and

equations:

Smith

103

In

ALPHA=3.14xl06T—2 - 2.00, where ALPHA=(103 + S18Odol)/ ( 103 +
1ft
% 0water) .

•

Another

limitation

temperatures

needed

for

temperatures

derived

from

color

alterations

suggest

such

a

fluid
two

is

the

knowledge

calculation.
inclusions

Paleo-

and

paleo-temperatures

of

conodont
may

have

been important in the evolution of these rocks and waters:
present-day

+

present-day

gradients

Cercone,
fluid

2 3°C,

1984),

inclusions

representing
(Nunn

and 80°C,
in

the

et

burial

al.,

1

1984;

km

deeper

Hogarth,

1985;

a minimum temperature found
late

diagenetic

dolomites

at

in

from
the

TABLE 2-5
Isotopic fractionation factors for dolomite-water

10

In ALPHA(dolomite-water)
3.20xl0^T“2
3.34xlO°T“2
3.23xlO^T
2.78 x 106T “2

+

1.50
3.34
3.29
0.11

(Northrop and Clayton, 1966)
(O'Neil and Epstein, 1966)
(Shepard and Schwarcz, 1970)
(Fritz and Smith, 1970)
t—1
U>
H

All equations corrected to be consistent with Friedman and
O'Neil (1977), from Land, (1988).
Average fractionation factor:
10 In ALPHA(doiomite-water)

= 3 *14 x 1°6t 2 “ 2.00

For calcite:
10 In ALPHA (Calcite-water) = 2.78xlO^T 2 - 2.89
(Friedman and O'Neil,

1977)

132

Niagaran

reefs

(Cercone

and

Lohmann,

1987)

and

in

the

Trenton (Shaw, 1975).
The

isotopic

composition

of

dolomite

and

calcite

in

equilibrium with several of the northern reef water samples
was

computed

using

the

estimated

present-day

temperature,

and the two paleo-temperatures.

rock values

are

(°/00

P.D.B.)

compared
of

in

Figure

Niagaran

2-15

and

carbonates reported by Cercone and Lohmann
and

Lucia

(1982)

for northern

The resulting

with

Salina

formation

S 180 values

A-l

formation

(1987)

and Sears

reef rock samples.

Similar

calculations were made for calcite using the fraction factor
from O'Neil

and Epstein

(1966),

Cercone and Lohmann (1987).
the

average

values

the

four

data

from

The rock value calculated using

fractionation

from

and the mineral

factor

and

fractionation

the

factor

range

of

equations

rock
are

shown for reference by the bracket lines.

As indicated in

Figure

be

2-15,

the

brines

do

not

appear

to

in

isotopic

equilibrium with any of the reported dolomites at presentday temperatures.

Whether equilibrium with calcite

is not clear however,
in

isotopic

occurs

but some of the samples appear to be

equilibrium

with

bulk

rock

limestone

values

the

dolomite

values

reported by Sears and Lucia (1980).
With

increased

temperature,

predicted to be in equilibrium with the water increase and
approach
bulk

the

measured

Niagaran

considering

dolomite

dolomite.

that

the

This

values,
increase

fractionation

especially
is

factor

not

for

the

unexpected

decreases

with

I— a— a--- a*--- o— I 80 °

|—

o— ■--- o*

— ■— oo— J PD. + 2 3 ‘

|— #o-

o

80°

□ ♦ a o-o P D . + 2 3 *

— jPresent
Day temp.

D>— ®~CD Present
Day temp.

late dolomite H
whole rock |---- 1
I

Niagaran |------------ j

LJ
-15

Figure

I 1 1 1__L J I 1 L J ! I
-10
-5
0
del ls0 % o (PDB) dolomite

I I L

I

1 late calcite
133

A l - C 0 3 I—

SAMPLE
O 2100
B 2097
□ 2092
O 2078
• 2020

Niagaran

L _ 1 I 1 1__ 1_I__L.

-15

1__1_1__1

■

1

1 > 1

I

-10
-5
0
del 180 % . (PDB) calcite

2-15.
C a l c u l a t e d del
0 ° / o o ( P D B ) v a l u e s f o r d o l o m i t e a n d c a l c i t e in
isotopic
equilibrium
with
N i a g a r a / S a 1 in_
formation
waters
at
Eubsurface
temperatures
of
present-day,
p r e s e n t - d a y + 2 3 C,
and
80 C.
Bars
show
range
of
values
from
four
different
do l o r n i t e - w a t e r
f r a c tjigO n a t i o n
equations
discussed
in
text.
A l s o s h o w n a r e r a n g e s of
del
0 values
reported
in
late
diagenetic
dolomites
and
calcites,
w h o l e r o c k ( W . R . ) , A - l C a r b o n a t e d o l o m i t e s , a n d N i a g a r a n d o l o m i t e s and
calcites
in M i c h i g a n ,
from
Cercone
and
Lohmann
(1987)
and
Sears
and
Lucia (1982).

134
increasing

temperature.

values

in

using

independently

suggest

N/S

that

However,

carbonates

are

similar

determined

equilibrium

temperatures has

the

with

affected the

fact
to

that

those

measured

calculated

paleo-temperatures
dolomite

isotopic

at

might

higher

composition

paleoof

the

brines.
Because of the uncertainty in fractionation factors for
dolomites

and

temperatures,

the

implications

calculations are somewhat tenuous.

of

Clearly more research is

needed to test whether isotopic equilibrium exists
the brine and host rock minerals and cements.
can be

established,

evapo-concentrated
refined.
suggest

The

then

seawater

results

that

the

of

carbonate

simple
by

model

dilution

water

calculations

equilibrium

at

between

If equilibrium
of

meteoric

these

these

must

do

of
be

however,

warmer

paleo-

temperatures may have affected the isotopic composition of
the brine,

and that

re-equilibration

may not have altered the water
equilibration
(1986)

is not

model,

composition of

and

$180 values.

considered
further

during

basin

Water-mineral

in the Knauth

suggests

formation waters may

cooling

that

and

the

be decoupled

Beeunas
isotopic
from

the

chemical evolution (Chapter 1).

MODEL

FOR

THE

ORIGIN

AND

EVOLUTION

OF

NIAGARA/SALINA

FORMATION BRINES
The
linked,

origin

of

Michigan

basin

Ca-Cl

brine

has

been

in the past, to geochemical processes such as shale

membrane filtration (Graf et al., 1966).

The data presented

135

here

suggest

originated

that

from

the

Niagara/Salina

evapo-concentrated

formation

seawater.

water

The

brine

appear to have been derived from seawater concentrated
past halite saturation,
the Na content

thus halite saturation

of the brine.

far

has lowered

Dolomitization has

likewise

removed Mg from the parent seawater and replaced it with Ca.
Additional reactions

involving aluminosilicates and perhaps

reactions with evaporitic salts has added CaCl2 , and either
S04

reduction

Dolomitization
processes
seawater

that
brine

or

CaS04

and

mineralization

halite

precipitation

transformed
into

the

the

Ca-Na-Cl

have

removed

appear

original
brine

now

to

be

S04
the

Na-Mg-Cl-S04
found

in

the

basin.
The high Br levels in the N/S samples suggest that the
parent

seawater

was

concentrated

The origin of the brines may,
Salina

A-l

salts,

the

minerals in the basin.

only

to

K-salt

therefore,
formation

be

precipitation.
linked to the

containing

potash

During deposition of the A-l salts,

dense seawater bitterns may have entered the Niagaran rocks
along the reef margins and may have moved downward through
the accumulating halite

into the underlying Niagara

rocks.

During later Salina salt deposition, seawater derived brines
may have

infiltrated

the

Niagaran

rocks,

however

it would

seem unlikely that brines generated during younger periods
of

evaporation

could

accumulation of halite.

reflux

through

the

thickening

Later compaction of the salts may

136
have been very important in forcing brine into the adjacent
formations.
As
samples
For

previously

discussed,

agree

well

example,

very
many

from expected

some

with

northern

of

the

northern

evapo-concentrated

samples

seawater values,

a

are

not

seawater.

depleted

characteristic

many of the southern reef samples.

reef

in

common

K
to

This may indicate that

the northern waters were derived in part directly from the
salt during
the

its

spatial

locations,
Figure

This

relationship

the

2-5

compaction.

A-l

potash

shows

where

observation may

between
salts,

the

N/S

the

and

northern

the

waters

reflect

A-l

were

reef

Carbonate.
sampled

in

relation to the different salt facies in the Salina A-l salt
and

the

vertical

salts.

The

Carbonate

in

relationship

sylvite
close

is

between

thinned

proximity

and

to the

the

different

intersects

northern

the

A-l

ree f s , while

the underlying halite thins below the potash salts
area.

A-l

in this

Fluids derived from the A-l salts during compaction,

would have had an easy pathway for migration either through
the

A-l

Niagaran

Carbonate,

or

carbonates.

alternatively,
The

short

downward

migration

would

into

the

minimize

water-rock interactions, explaining for example, the lack of
K

depletion.

located

near

depleted in K,

In
the

contrast,
A-l

salts

the

southern

and

contain

suggestive of water-rock

reefs
water

are

not

generally

interactions.

Any

fluids squeezed from the potash salts would have had ample

137

opportunity to react with formation minerals before entering
the southern reefs.

ORIGIN OF TRENTON-BLACK RIVER BRINE
The Trenton-Black River formation samples show chemical
evidence of mixing between evapo-concentrated seawater and a
more dilute water, possibly seawater or freshwater.

Whether

this mixing represents dilution of a pre-existing brine by
dilute

water,

or

enrichment

brine is not clear.

of

dilute

(salinity)

by

migrating

Figure 2-16 shows the Br values in TBR

waters from the Albion-Scipio area.
Br

water

argues

against

from production operations.

The smooth variation in

localized

dilution

resulting

This distribution does not rule

out that regional infiltration of dilute water has occurred
in this area of the basin.

Alternatively, the fact that the

deeper

Ordovician

formation

sample

(St.

Peter

Sandstone,

#8040)

is more saline suggests that an up-dip migration of

brine from the deep basin may have also occurred.
If

the

explained,
Three

history

the

TBR

brines

is

to

be

fully

then the end member waters must be constrained.

saline

geology

of

of

end

the

members

area;

are

seawater

possible

considering

the

(perhaps

Ordovician),

the

present-day Niagara/Salina formation water, and the presentday Devonian

formation water.

Extrapolation

of the

data back to the seawater trend line (Figure 2-10a)
that
start

the

saline

of halite

precipitation.

end

member

precipitation

was
and

concentrated
the

start

Cl-Br

suggests

between

the

of MgS04-salt

If the TBR brines originated from Ordovician

(400

Figure

2-16.

Br ( m g / 1 ) in T r e n t o n - B l a c k R i v e r
f r o m the A l b i o n - S c i p i o t r e n d .

formation

138

Br ( m g / I )

brines

139

seawater, then they were concentrated to this degree before
dilution

occurred.

However,

this

seawater

composition

is

also similar to the present upper Devonian formation brine,
both

in

its

salinity

and

suggesting that Devonian
down

into

the

possible

TBR

because

in

its

Na-Ca-Cl

chemistry,

formation brine may have migrated

formations.
the Salina

Downward
salts

do

migration

not

cover

is
the

Ordovician formations in the Albion-Scipio area, the AlbionScipio fracture system is known to extend up to the Devonian
formations (Carpenter, personal communication), and areas of
possible basin recharge have been identified
central

area

Vugranovich,
The
collected
formation

of

Michigan

(Wilson

in the

south-

Long,

1986;

and

1986).

chemistry

of

from

central

the

sample,

the

#8040)

Ordovician
basin

might

formation

area

suggest

(the

St.

brine
Peter

however, that

the

previously proposed waters may not be the saline end member
water.
that

This
is

samples
might

sample

similar
(Table

have

a highly

many

2-1).

had,

alternatively,

in

is

that

respects

This

before
Ca-Cl

concentrated
to

suggests

dilution,
rich

a

waters

the

that

have

brine

Niagara/Salina
the

Ca-Cl

Ca-Cl

TBR

waters

chemistry,
migrated

or

into,

and mixed with more dilute waters in the Trenton-Black River
formations.
Therefore, a third hypothesis is that the TBR formation
brines

are

mixtures

of

Ca-Na-Cl

altered to a Na-Ca-Cl composition.

brine

which

have

been

The feasibility of this

140

hypothesis

can

be

tested

using

Niagaran/Salina sample #2099,
from the
member.

southern

reef

thermodynamic

a typical N/S brine collected

area,

was

used

for

the

saline

end

Seawater from Nordstrom et a l . (1979), concentrated

to gypsum saturation

(and in equilibrium with calcite)

used for the dilute end member composition.
simply

a

convenient

standard

that

precipitation needed to remove Ca.
several brine/freshwater ratios

log plot

of

Cl-Br

Figure 2-17.

and

dolomite,

and

with these

the

and any
The

symbols.

In the

end

(1.,

0.9,

the

results

of

second

positions

and

then

from

this

case,
(end
the

CaS04

0.75,

0.5,

0.25,

and are shown on a log-

member

other mineral

equilibration

different

maximizes

as mole percentages

anhydrite,

mixing.

without

This choice is

of

Ca-Mg-Na

The mixing was done using two scenarios,

equilibrating

was

The results of mixing at

.1, 0) are presented in Table 2-6,

by

modeling.

waters

with

maintaining
that

case
the

first

calcite,

equilibrium

shown

end members

members

first

supersaturates
are

plot

in

in

case), and

upon

as

solid

were

mixed

slightly
mixing

was

calculated without maintaining equilibrium with any mineral.
These results are shown as open symbols on this diagram.

In

either case anhydrite was predicted to supersaturate in all
mixtures.
The

results

explained by

(Figure

2-17)

show that

brine/seawater mixing

at

Cl/Br

ratios

ratios

between

are
0.2 5

and 0.5, while the relative percentages of Ca-Mg-Na in these
brines

suggest

a brine/seawater ratio near

0.25.

This

is

141
TABLE 2-6
Modalad mixtures of Niagara/Balina brine and seawater

All values mg/kg
Brine/
Seawater
ratio

0-:l

.1: .9

PH
aH2o
Ionic
Strength
TEMP

7.33
0.932
2.246

5.15
0.914
2.940

Ca
Mg
Na
K
Cl
so4
ALK
Br

4581
1080
39015
1446
70122
3400
16
244

30

30

.25:75
4.75
0.882
4.025
30

.5: .5
4.35
0.819
5.874
30

.75:.25
4.05
0.748
7.736
30

.9: .1
3.88
0.703
8.851
30

1— :0
3.77
0.672
9.598
30

11136
2580
39605
222
87275
1555
701
550

21590
4742
40490
3376
113000
809
606
1010

39888
7998
41969
5306
155900
338
378
1777

59026
10829
43445
7236
198800
147
219
2544

70845
12343
44332
8394
224500
90
156
3004

78853
13282
44922
9165
241600
65
124
3311

13. 1
5.0
78 .9

21.6
7.8
70.6

31.6
10 .4
58 .0

38 .7
11.7
49.6

42.0
42 .0
12 .0

44.0
12 .2
43.8

Mole percentages
%Ca
%Mg
%Na

6.2
2.4
88.4

Mineral Balance
Mineral Phase fmoles/kai
Anhydrite -0.066
Dolomite -0.14 6
Calcite
0.292

Key:

-0.0759
-.112
0.229

-0.0685
-0.0648
0.134

-0.0480
-0.0005
0.004

-0.0247
0.046
-0.092

-0.010 -0.0001
0. 067
0. 078
-.133 -0.155

Composition of various mixtures of seawater evaporated
to gypsum saturation and a Niagara/Salina brine from the
southern reef trend. Results listed are for the case
when equilibrium is maintained with anhydrite, dolomite, and calcite.
* Seawater concentrated to gypsum saturation
**: Niagara/Salina brine sample #2099
Mineral balance: += mineral dissolution, - = mineral precipitation

5.5
.25
4.5

Log

Cl ( m g / k g )

142

3.5
1.5

2

2.5
3
3.5
L o g Br ( m g / k g )

4
O MEASURED
A A P R E D IC T E D

50%Mg

.25
Ca

Na

Figure 2-17.
Cl-Br
(log
mg/kg)
and
relative
Ca-Mg-Na
composition calculated to result from mixing of
N/S
sample
(12099
(point
1)
with
seawater
concentrated
to gypsum
saturation
(point
0).
Numbers indicate seawater/brine mix ratio.

143
is

the

case

regardless

reestablished

or

not.

whether
Some

of

mineral
this

equilibrium

difference

might

is
be

explained by a lower Ca/Na ratio in the saline end member or
a difference
water

in temperature,

(seawater)

56g halite/kg)
result

of

may

have

or alternatively,

dissolved

before mixing,

Nevertheless,

(approximately

or had a lower Ca/Na

evapo-concentration

precipitation.

halite

that dilute

past

the

start

of

as

the

gypsum

the results show that mixing

of Ca-Na-Cl

brine with

seawater

for the TBR

formation waters.

is a possible
Mixing with

explanation

seawater would

result in CaS04 , dolomite, and calcite saturation (depending
on mix ratio), and if these minerals precipitate, Ca removal
would

follow.

reactions
reported

Geologic

have

occurred

anhydrite

is

evidence
is

given

the

last

that
by

some

Miller

of

these

(1988),

diagenetic

mineral

who
to

precipitate in the Trenton-Black River rocks in the AlbionScipio trend.
St. PETER SANDSTONE WATER
The one brine sample collected from the Ordovician St.
Peter sandstone

formation allows only a limited

of the water chemistry in the deep basin.
collected

from

a

gas

producing

from

a

depth

well
of

in

This sample was

northern

over

evaluation

lower

3000m.

Michigan

Although

the

reliability that can be placed on interpretations made from
a single

brine

analysis

is

limited,

this water deserves some attention.
saline

(404

g/1

TDS)

Ca-Na-Cl

the

unique

nature

of

This water is a highly
(Figure

2-7)

brine.

Inspection of Figure 2-10 demonstrates that this sample has
a Cl,

Na,

seawater
trend

K and Br chemistry that match evapo-concentrated
very

line

well,

although

for Cl.

Mg

it

plots

above

is highly depleted

the

and

seawater

Sr greatly

enriched over expected seawater values.

Application of the

dolomitization

shows

depletion

model

discussed

resulting

from

stoichiometric

explain most the observed Ca.
is

interpreted

seawater

that

to

have

evolved

earlier

dolomitization

from

highly

dolomitization.

reported

to

(Sibley,

personal

clastic

1) , show

samples

in

the

the

the

formation

initial

St.

authigenic

Peter

in

K

northern
however,

is

the Berea

for

other
example

from

seawater

reef

trend

values.
of

depleted

Indeed,

K

minerals

from

strikingly

of

that has been

and brine

are not

seawater content.
sample

lack

aluminosilicate

basin,

a depletion

from

Niagara/Salina
from their

in

can

This water is

sandstone formation,

communication),

formations

(Chapter
Many

contain

Mg

concentrated

The

depletion is curious and somewhat unexpected.
produced from a clastic

the

Thus, except for K, the water

evolved

by

that

the

in

K

chemistry

similar

chemistry observed in the Niagara/Salina samples

the

to

the

(Figure 2-

8) , including the lack of K depletion and 87Sr/86Sr ratios
(Chapter 1).

This water sample may therefore, be related to

the Niagara/Salina water and possibly originated during the
deposition
into

the

of

the

Niagara/Salina

underlying

Ordovician

salts and
formation.

migrated
Until

down
more

145

samples are recovered and analyzed however,

this remains

a

preliminary interpretation.
ISOTOPIC EVOLUTION
The isotopic
the

TBR

brine

data

do

not

originated

support
from

the

CaCl2

suggestion
brine

that

source.The

stable isotope values in the TBR samples plot as a cluster
intermediate
11).

of the best-fit

SMOW

(Figure

2-

This is not where a water that evolved by mixing of a

CaCl2 Niagara/Salina
end

line and the

of the

seawater.

best-fit
The

type

brine,

line

isotopic

plotting

would
values

plot
of

the

at

when
TBR

the

enriched

mixed
brine

with
plot

intermediate between a dilute formation brine (plotting near
the

GMWL)

and

SMOW

or seawater

at gypsum

saturation,

supporting this hypothesized origin.

The TBR samples

plot

sandstone

intermediate

of

the

St.

Peter

sample

also
and

SMOW, which might be consistent with an evolution of brine
migrating up from deeper areas of the basin and mixing with
SMOW.

But making this

interpretation based on the

isotopic

composition of a single brine sample is tenuous.
A relationship between £D values and Br (salinity ) was
found in the Devonian

formation water of Michigan

1) , but as the following list shows,
is observed in the TBR samples:

(Chapter

no clear relationship

146

SAMPLE

— 0°/oo_____Br(mq/1^
delD^ /oo_____del— ^

6074

-26.40

-1.85

1080

6094

-24.30

-1.99

1160

6095

-26.70

-1.76

1250

The lack of correlation might reflect the small number
of

analyses

available,

or

that

both

end member

have had a similar isotopic composition.

water

Why mixing is not

represented in D values remains a question that,
limited

amount

of

data,

does

not

seem

may

given the

answerable

at

this

time.
It may also be possible that the isotopic values have
been masked by reactions with carbonate minerals,
for 180.
and

the

in

equilibrium

same

method

Niagara/Salina waters.
values

least

In order to investigate this, the S 18o of dolomite

calcite

using

at

with

Ordovician

the

and

the

brine

was

temperatures

calculated

described

for

Figure 2-18 compares the calculated

ranges

dolomites

with

of

S ^ 80

(Taylor,

in

the

1982).

three

types

Average

of

isotopic

values are -6.8°/oo for the regional dolomites, -7.8°/oo for
the

cap

dolomites,

dolomites,

and

-9.0°/oo

respectively

(Taylor,

values

for minerals

agree

with

temperatures.

any

for

the

1982).

fracture
The

calculated

in equilibrium with the brines
of

However,

these

minerals

at

related

do not

present-day

the brine values are near those

in

so°c
h&-j PD. +23°C

1—

PresentDay

■J R E G I O N A L

J

•
? ° I I ----- |-------------- F R A C T U R E
□ .6094
1
1
O 6095
|---1 C A P

I t « I ! I » i i i I L.J L J
-15

-10

-5

I I I
0

del ,80 %o(PDB)

Figure

2-18.
Cal cu la te d del
0 ° / o o ( P D B ) v a l u e s f o r d o l o m i t e in
isotopic
equilibrium
with
Trenton-Black
River
formation
w a t e r s a t^ s u b s u r f a c e
t e m p e r a t u r e s of p r e s e n t - d a y , p r e s e n t d a y + 2 3 C, a n d 8 0 C.
A l s o s h o w n a r e r a n g e s of i s o t o p i c
values
for
fracture,
cap,
and
regional
dolomites
in
the
Tre nt on -B la ck River formation, from Taylor (1982).

148
equilibrium

with

regional

present-day+23°C,

equivalent

present-day geothermal
be

near

isotopic

dolomite

at

represent

to

equilibrium

at

burial

gradient.

temperatures

temperature.

dolomite

temperatures

1km

deeper

of

at

the

The brine also appear to
with

near

the

80°C,

fracture

but

not

related

at

higher

Thus the isotopic compoition of the brine may

equilibrium

with

formation

minerals

and

not

the

mixing indicated by the chemical data.
CONCLUSIONS
The

following

conclusions

are

made

concerning

the

Niagara/Salina and Trenton-Black River formation brines.
1)

The Niagara/Salina

formation water originated

from

evapo-concentrated seawater, concentrated through the halite
facies to the MgS04 facies and
salt facies.

in some cases,

into the K-

This suggests that their origin is linked to

the Salina evaporitic seas

in Michigan,

deposition of the Salina A-l
refluxed

into

underlying

period.

Additional

salts.

Niagaran

brine

may

and perhaps to the

Dense brine may
formations

have

been

during

squeezed

have
this

out

of

these salts during salt compaction.
2)

Dolomitization

appears

to

be

the

most

important

reaction to have affected the brine, as it explains the CaMg chemistry.

Dolomitization in conjunction with extensive

halite precipitation were the mechanisms generating the CaNa-Cl

chemistry

from

illitization

or

water

carbonate

while

seawater.

K-feldspar

Reactions

diagenesis

reactions

removed

supplied

Sr.

K

such

as

from

the

Reactions

149

involving aluminosilicates and/or late stage evaporites have
apparently

increased

the

CaCl2

content

of

the

brine

from

seawater values.
3) Although their origin is similar, the Niagara/Salina
waters

show

differences

formation brines
which

they

within

the

their

in the basin,

plot

Niagara/Salina

in

on

log

brine

chemistry

reflected

other

in the manner

concentration

diagrams.

may

show

differences

in

samples

collected

formation,

from

in

in
The

chemistry

from

different

areas of the basin.
4)
Black

Brine

produced

River

concentrated
saline

end

from

Formations
seawater
member

the

evolved

by

more

dilute

with

waters

Ordovician

are

aged

mixing

Trenton-

of

water.

evapo-concentrated

evapoPossible
seawater

(Ordovician?), Devonian formation brine, or Ca-Cl brine such
as found in the Niagara/Salina formations.

It is not clear

at this time what the true end member waters are.

However,

thermodynamic modeling suggests that mixing of Ca-Cl brine,
such as the Niagara/Salina brine,

and seawater can explain

the chemistry of the Trenton-Black River water.
5)

The

isotopic

evolution

of the

formation waters

is

generally consistent with the model proposed by Knauth and
Beeunas
questions
dilution

(1986)

for

remain.
in

both

evaporating
There

sets

of

is

seawater,

little

formation

isotopic
waters,

although

many

evidence
even

for

though

chemical evidence supports it for the Ordovician formation.

The

isotopic

brines

composition

might

reflect

the

isotopic

however,
from the

chemical

isotopic

composition

equilibrium

with

temperatures.
formation
evolution.

of

the

the

Niagara/Salina

original

evolution may

evolution.
of

the

formation

One
brines

seawater
have

has

carbonates

composition,

been

suggestion
been
at

formation

independent
is

that

affected

higher

the
by

paleo-

In sum, the isotopic evolution of these deep

waters

may

be

independent

of

the

chemical

151

Estimation of Bromide in salt
The
composition
of
a water
in
equilibrium
with
carnallite,
sylvite,
halite,
anhydrite,
dolomite,
and
calcite was calculated using the PHRQPITZ computer routine
(Plummer et al., 1989). This calculation was done for a
temperature
of
50°C,
representative
of
present-day
temperatures in the basin.
The pC02 was not fixed in this
example.
This computation is rather straight-forward using
this computer program, however, Br must be estimated.
The
only source for Br in this hypothetical brine is the
carnallite, and any halite that dissolves during carnallite
recrystallization.
The initial carnallite may have had a Br concentration
ranging from 3000ppm to 5000ppm (Kuhn, 1968).
A lm3 volume
of carnallite (1.67x10 kg) that converts to sylvite (0.22m3)
releases 36082 moles of H 20 (649.5kg).
If the carnallite
has a Br content of 3000ppm, then 5.01xl06mg of Br are
released
into
649.5kg
of
water,
resulting
in
a
Br
concentration of 7713mg/kgs (kgs=kg of solution).
This concentration must then be corrected for Br uptake
by re-precipitating sylvite and halite. An input of Mg-Ca-KCl
into
solution
decreases
the
solubility
of
halite
dramatically
(Holser,
1979),
there-by removing Br from
solution in addition to that removed by the sylvite.
Br
uptake by sylvite and halite was considered in this example,
although the amount of Br removed by halite is calculated to
be negligible. Using a distribution coefficient of d = 0 .73,
(Braitsch, 1971), the reprecipitated sylvite is calculated
to contain 5630ppm Br.
The total amount of sylvite re­
precipitated is 43 6kg. This removes 2.45xl06mg of Br from
solution,
leaving
a
Br
concentration
in
solution
of
3930mg/kg.
The calculated sylvite has a Br content higher than the
maximum
reported
value
in
Michigan
potash
(4340ppm).
However,
considering
the
variability
in
distribution
coefficients and that fact that carnallite may dissolve
incongruently, the results are encouraging.
If a second
generation
of
sylvite
precipitates
from
the
remaining
bittern (3800ppm Br) , then it is predicted to have a Br
content of 2800ppm, near the Br values measured in the
Michigan potash by Matthews and Egleson (1974).

APPENDIX A

APPENDIX A

STATISTICAL EVALUATION OF MICHIGAN BASIN BRINES
Natural
chemical

water

can

makeup.

exhibit

a

types

and

The

wide

variation

amounts

of

in

its

dissolved

components will reflect a water's origin and the geochemical
processes involved in its evolution.
water chemistry
Drever

(1988),

are discussed by

reactions,

biological
operate,
chemistry

water

activity.
combining

and

ion

Often
to

a

create

reflected by

exchange

mixing,

and

number
a

of

doing

so,

help

eguilibria,
adsorption,

these

structure
between

and

processes

in

the

water

concentrations

Statistical tests can be a useful

in elucidating these often complicated
in

(1981),

evapo-concentration,

correlations

of dissolved components.
tool

Stumm and Morgan

and others, and include mineral

dissolution-precipitation,
REDOX

Processes that control

to

demonstrate

the

relationships,
processes

that

control water chemistry.
With a goal
geochemical
evaluation

of

further characterizing the

evolution
of

the

of

the

brine

Michigan

chemistry

was

basin

origin

and

brine,

an

undertaken

univariate and multivariate statistical methods.

using

Univariate

methods are first used to characterize Michigan basin brine
chemistry,

compare

it

with

seawater,

and

then

average brine composition between formations.
means

and

standard

Student's t-test

deviations

are

to

compare

To do this,

calculated,

and

the

is used to test mean compositions between
152

153

formations.

The multivariate methods of R and Q-mode factor

analysis

are

then

used

attempt

to

find

test

data

statistical

controlling

the

results

compared

are

to

brine

homogeneity,

evidence

chemistry.

with

for

The

results

from

and

to

processes

factor

analysis

previous

research

which used factor analysis to study saline ground w a t e r s .
THIS STUDY
The

population

studied

in

this

Michigan basin formation waters.
were collected and analyzed
as

described

population

in

Chapter

was

investigation

is

all

Samples of this population

for major and minor components

1

subdivided

and
by

Chapter

2.

producing

The

sample

formation

for

statistical testing whenever possible.
Statistical evaluations, especially of natural systems,
are

often

based

on

several

underlying

assumptions.

The

principle assumption made in this study is that the samples
are random and representative of the parent population,
assumption typical of ground water studies.
demonstrate

the

possible

short-coming

in

two

Two examples

this

assumption.

The oil wells from which the samples were collected are not
random

in

the

basin,

but

are

located

conditions favor hydrocarbon entrapment.
coincide wi t h , for example,
high

porosity

characteristics
representative
possible

that

or

unique

that

might

of

the

waters

structural
mineralogy,
make

the

formation

in

are

where

These areas might
features,
or

associated
general.

density

geologic

zones

have

other

water
It

stratified

of

is

non­
also

within

154

formations.
traps,

Because hydrocarbons reside

water

produced by

with

oil

non-representative

wells.

Considering

in up-dip areas

of

density

be

this,

might

oil-well

samples

may not be random or representative of the formation water
as a whole.

UNIVARIATE STATISTICS
DATA DISTRIBUTION
Univariate parametric statistical methods are based on
a

normal

samples

distribution
(Koch and

demonstrated
normally

that

of

Link,

1971).

many

distributed

the

chemical
A

number

geochemical

(Ahrens,

variables
of

data

1954;

studies

sets

Koch

and

in

are

Link,

the
have
log-

1971,

Hitchon et a l ., 1971, Long et a l ., 1986). Before statistical
test can be correctly applied therefore, the distribution of
the sample population must be determined.
(X*)

test

both the

xs used to test
raw data and

log

The chi-squared

for population
(base 10)

normality,

using

transformed d a t a . The

results of this test at the 5% significance level or better
(TABLE

A-l),

distributed,

indicate

that

the

except for Na and C l .

data

are

lognormally

Histograms

(Figure A-l)

of Na and Cl show these elements follow a negatively skewed
log-normal
break

distribution, and

in

slope,

populations
is not

lognormal

characteristic

(Figure A - 2 ) .

clear,

but

log-probability

because

distribution,

The cause
the

albeit

of

two

show

a

overlapping

for this distribution

Na-Cl
skewed,

transformed for use statistical tests.

plots

histograms
all

data

suggest
were

a

log-

155

RESULTS
C O MP ON EN T

Br
Log
Cl
Log
Na
Log
Ca
Log
Mg
Log

X

1550
3.12 1
18000
Cl
5.241
55900
Na
4. 706
43900
Ca
4.555
7610
Mg
3.835
K
4860
Log K
3.472
Sr
1480
Log Sr
3.067
Rb
13.0
Log Rb
0.85
Cs
2.5
Log C b
0.17
Li
39.6
Log Li
1.51
B
92.
Log B
1.67
NH4N
266.
Log NH4 2.26
HCO
80.5
Log H C O .. 1.60
50
157 .
Log SO^ 1.72
17.2
Log I
1.12
51
3.2
Log Si
0.42
TDS
294000
Log TDS 5.452
Br

TA B LE A-l
OF C H I - S Q U A R E D TEST

SD

DF
SIGNIFICANCE
_________ LEVEL

865
0.262
39025
0.135
21400

50.5
20.7
12.9
33.9

0 .2 10

26900
0.298
3500
0.210

5250
0.443
1010
0.321
16.4
0.50
2.4
0.49
23.1
0.31
109.
0.58
255 .
0.38
96.6
0.58
190.
1.18
13.5
0.32
2.7
0.25
65900
0.138

10
7
8
5

2. 18x10 17
-3
4 .2 1x 10
0.113
-6
2.48x10

21.2

10

0.020

44.4
67.8
14.4
30.7
6.80
137 .

9
10
7
9
7
9

1.17x10
-10
1.17x10
0.044
-4
3.39x10
0.450

12.8

8

0.120

38 .6
8.6
89.2
4.753
32.1
17.811
25.2
24.8
42.0
2.84
71.2

7
8

2.29x10
0.380

6

0

7
5

0
-6

12.2

7

0.690
5.62x10
-3
6.72x10
-3
2. 95x10
-3
3.22x10
-8
5.85x10
0.724
-13
2.367x10
0.093

133.
17.9
95.5
85.8
62.5
7.9
3 6.6
9 .9
16.6
34.4

6

0

9
7
9

0.035

6

9
8

5
5
6

10

9
4
6

9
5

0

-14

1 . 1 10x10
-9
1 . 189x10
0.539

2.2 2x10~
0.128
0.055
-6
2.012x10

KEY :
H q ■= no evidence to suggest that sample is not from a normal
population
S i gnificance level of acceptance;
a ■= 2.5%
b = 5%
c •= 10%
d = <1%
X = mean value,
SD ■= standard d ev iation, X' = c hi - s q ua r ed
value, DF “ degrees of freedom

30

50

25

40

20
30

I*
E

20

= 10
10

5
0

m m r r i ________
.................

3.5

4

i

»

4.5

t

i

i

i

i

t

5

i

i

0

i

4.5

5.5

A-l.

Histograms

5.5

Log Cl ( m g / l )

L o g Na ( m g / l )

Figure

5

of

Na

and

Cl

(mg/l)

concentrations.

Lo g Na ( m g / l )

Figure

A-2.

L og -p ro bability plots
c oncen trfftions.

_J____ »

»

i

iJ

4.4

0

OJ

5.6

4.7

L o g Cl ( m g / l )

of Na and Cl (mg/l)

158

AVERAGE BRINE COMPOSITION
Table A-2
means

lists

the

formation water

means

(geometric

and an unbiased estimator of the me a n ) , maximum and

minimum concentrations,

variances

sample

listed

sizes.

Also

(unbiased estimator), and

are

seawater

concentration

ratios, defined as: SWR= average brine value/seawater value,
and evaporation factors,

defined as E F = (concentration ratio

of a given constituent/concentration ratio of B r ) .
values
are

are given

from

Brewer

Seawater

in parentheses next to each element,
(1975).

Averages

and

standard

and

deviations

calculated from log transformed data represent the geometric
means and geometric variances,

respectively.

The geometric

mean may be the preferred estimator of the population means
and

variance

when

deviation/mean)
and Link
means
and

and
have

unbiased

the

is less than 1.2,

(1971).

exist

smallest

estimators.

that

variation

estimators

are

sampling

These

of

(standard

a rule suggested by Koch

However, other

variances
the

coefficient

of population

statistically
error

statistics

variance

are

termed

unbiased
of

all

minimum

variance unbiased estimators. Generally, unbiased estimators
of

means

are

preferred

if

the

lognormal

skewed, as is observed for Na and Cl values

distribution

is

(Gilbert, 1987).

The unbiased estimates of the population means and variances
are calculated using:
(2y)

m =

[exp (y) ] X n (1/2 Su 2 ) and V=exp

[Xn (su 2/ 2 ) ]-Xn [su 2 (n-2) / (n-1)] where y=geometric mean,

su 2= geometric variance,

n=number of observations,

and the

159

TABLE A-2
SUMMARY STATISTICS
TRV
BEREA
(29)
n
(3)
Cl (19,350 mq/kq)
194000
172000
A
175000
B
195000
10000
15600
C
176000
D
13600
209000
196000
E
SWR 10. 0
8.1
EF
0.48
0.54

DD
(40)

RP
(13)

DT. R
(2)

N/8
(26)

TBR
(11)

171000
173000
3430
120000
212000
8.8
0. 59

194000
194000
2070
75800
242000
10.0
0.32

212000
212000
3720
179000
251000
10. 9
0.22

210000
214000
8180
117000
265000
10.8
0.33

12600
12800
6219
91000
16300'
6.5
0.42

Br (67 mq/kq)
1400
A
B
1460
C
295
940
D
E
1840
SWR 20 .8
EF
1

1100
1240
164
590
2340
14 .9
1

1000
1090
74 .0
315
2390
14 .9
1

2100
2300
308
1050
4210
31.3
1

3400
3400
345
3060
3750
50.9
1

2400
2400
202
690
3340
32.8
1

1030
1050
63 .0
750
1430
15.7
1

Ca (411 mq/kq)
45100
A
B
45400
C
3660
D
38400
49500
E
SWR 110.
5.3
EF

29000
33300
4790
11600
67700
63.0
4.2

24500
27100
2020
7390
7330
59.6
4.0

65200
68600
6440
27800
96200
158.
5.0

88000
89700
17300
72400
107000
214.
4.20

74000
76900
4290
35200
124000
180.
5.5

22800
23300
1530
15800
35500
55. 5
3 .5

Mcr (1290 mq/kq)
A
8250
B
8270
C
355.
D
7600
E
8790
SWR 6.4
EF
0. 30

6600
6810
630.
3840
9700
4.7
0.31

5100
5360
284 .
1820
11300
3.9
0.26

8700
9220
939.
3370
14600
6.7
0.21

11800
11800
1450
10400
13300
9.1
0.18

11500
12100
763.
6080
20000
8.9
0.27

5130
5280
410.
3200
9110
3 .9
0.25

71000
72400
2340
38800
103000
6.6
0.44

34200
37100
4370
12400
61500
3.2
0.10

23100
23100
900.
22200
24000
2.1
0.04

31400
33500
2500
9820
48800
2.9
0. 088

47300
47700
1890
37200
59000
4.4
0.28

Na (10760 mq/kq)
A
63000
68200
B
63100
70100
C
2580
6250
D
58200
43000
E
66700
88700
SWR 5.9
5.8
0.39
EF
0.28

160

TABLE A-2 (cont'd.).
U. ).
DD

RF

DT. R

N/S

TBR

1660
1910
264.
440.
4750
3.8
0.26

1640
2010
220.
380.
8360
4 .1
0.27

7200
7930
1035
3570
13300
18.
0.58

13400
14300
5020
9270
19300
33.6
0.66

9250
11000
1400
2300
24100
23.2
0.71

3680
3780
272 .
2180
5440
9.2
0.59

1060
1260
195.
350
2920
117
7.9

780
911.
86.0
170.
2350
98
6.5

1970
2440
487 .
290.
3980
246
7 .8

2680
2710
365.
2340
3070
335
6.6

2190
2430
230.
800.
6030
274
8.4

750.
766.
43.0
540.
1040
94 .
6.0

Rb (0.12 mg/kg)
3.6
3.7
A
B
3 .6
5.6
C
0.2
1.7
D
3.4
0.4
E
17.2
3.9
SWR 30.
30.8
EF
1.4
2.1

3 .0
4 .2
0.7
0.4
11. 1
25.
1.7

12 .4
13.6
1.9
6.5
23.8
103
3.3

37 .3
37 .5
3.7
33.8
41.2
310
6.1

24. 1
30.1
5.3
10.7
77.7
200
6.1

7 .9
8.3
1.2
4 .5
11.
66.
4.2

Li (0.18 mg/kg)
8.0
29.3
A
8.0
B
33.7
*
C
4.6
D
8.0
9.2
E
8.0
81.0
SWR 44.
150
2.1
EF
10.

23 .9
31.8
4.2
2.0
68 .0
133
8.9

46.5
48.4
3.9
24.0
76.0
258
8.2

92 .9
96. 0
24 .0
72.0
102 .
516
10.

59.2
63.9
5.2
18.0
101.
330
10.

37 .5
38 .1
2 .1
26.0
47.0
210
13 .

B (4.5 mg/kg)
A
B
C
D
E
SWR EF

21
43
13
1
85
4 .6
0.31

119
150
37
39
377
27.
0.85

226
260
120
134
383
50.
1.0

127
170
34
25
498
28.3
0.9

13
14
2
6
21
2.9
0.20

BEREA
K. (399 mg/_kgl_
A
640.
640.
B
C
35.0
D
580.
700.
E
SWR 1.6
EF
0.075
sr (8 mg/kg)
A
2000
2000
B
114.
C
D
1780
E
2140
SWR 250
EF
12 .

TRV

28
40
13
18
117
6.4
0.43

161

TABLE A-2 (cont'd.).
BEREA
TRV
Si (0.5-10 ma/kg)
A
3.2
8
B
3.4
1
C
90
50
D
4
4
E
4
2
61
SWR
53
EF
031
036

DD________ RF________ DT. R
.34

2.6

2

2.9
0.30

2.,7
.60
0.
0.
.90

0.30

8.5
0. 4t>
0.015

29
228

228
230
240
115
400

31
51

1.2

.

11 6

0. 49
0. 033

N/S

TBR

2

.

3
37
.007

5
30
0.90
6.3
42
013

0.70
7
63
040

550
590
570
441
690

547
620
600
235
1087

80
80
85
45
126

160

180

210

210
110
100

90
150
42

38
41

1.6

n h 4n

A ~
B
C
D
E
HCOj
A ~
B
C
D
E
SWR
EF

74
77
80
50
126

125
129
130
76
193

(142 mq/kq)
11
85
20
45
6
16
10

25.
0.14
0. 007

110
120

0.022

43
45
390
1.4
0. 042

330
1.5
0.030

0. 032

180
300
57

1200
1200
1200

10

49

28
26

66
11

20

0

0

0

100

1130

350
0.44
0. 013

50
0.037
0.0007

150
0. 018

720
0.076
0 .005

25
28
3
7
65
420
13
345000
352000
14000
196000
444000
10.
0.28

10

0

0

160
0.32
0. 017

SO, (2710 ma/kq)
A
45
65
B
71
130
C
43
35
D
10
0
E
100
610
SWR 0.017
0.023
EF 0.001
0.001
I (.06 mg/kg)
A
23
B
23
C
3
D
17
E
28
SWR 380
EF
18

107

140
0.36

0.66

0. 004

11

11

13

13

23
28

2

1
0
28

6
10
68

4
71
170
11

TDS (35.000 ma/kg)
A
315000
282000
B
316000
289000
C
16400
26100
D
284000
21500
E
335000
326000
SWR 9.0
7.4
EF
0.43
0.50

180

380

12

12

38
41
16
25
57
630
13

278000
280000
5570
197000
349000
7.9
0.49

296000
315000
31900
125000
395000
8.5
0.27

355000
360000
61400
299000
421000
10.
0.20

0
610
1.1

.

0 001

6

25
90
0.29
0.018
205
250
51

11
12

1
6

26
180
12

207000
210000
10100
151000
273000
5.9
0.38

162

TABLE A-2 (cont'd

MC1„
A “
B
C
D
E
SWR
EF

TRV
BEREA
(69.2 mea/ka)
2980
1840
2220
2990
274 .
213.
2580
102 .
3240
4240
4 3.
26.6
2.0
1.7

DD

RF

DT. R

N/S

TBR

1660
1810
122.
573.
4640
23.9
1.6

4020
4240
403.
1680
6010
58.0
1.8

5420
5520
1010
4525
6500
78.3
1.6

4720
4890
262.
2290
7380
68.2
2.1

1580
1610
106.
1060
2540
23.
1.5

KEY;
All values in mg/l unless noted
A: Geometric mean
B: Minimum variance unbiased estimator of population mean
C: Unbiased estimator of standard eviation
D: Minimum value
E: Maximun value
SWR: Seawater ratio; average brine value(B)/seawater value
EF: SRW/ Br seawater ratio
MCL2 : Mg + Ca + Sr - O.SHCO^ - S04 (meq/1)
n: Number of analyses for major components
TRV: Traverse N/S: Niagara/Salina DD: Dundee
RF: Richfield
DT. R: Detroit River
TBR: Trenton Black River

163
function Xn is from statistical tables or calculated by the
following equation from Gilbert (1987):
Xn (x) = 1 + (n-1)x/n + (n-1)3x2/ (2!*n2 (n+1))
+ (n-1)5x 3/ (3!n3 (n+l)(n+3))...

,

COMPARISON WITH SEAWATER
The

average

compared

with

Figure A - 3 .

brine

the

compositions

chemistry

of

from

Table

evaporating

A-2

are

seawater

in

The seawater data from McCaffrey et a l . (1988)

are used in this figure.

Inspection of this diagram shows

the relative degree of agreement that exists between these
brine components and seawater.
be

compared

with

seawater

The brine chemistry can also
by

examining

concentration ratios listed in Table A - 2 .

the

seawater

These ratios help

to quantitatively illustrate the changes that seawater would
have

to undergo

brine.

For

average

brine

if

it were to

example,

based

salinity

on

evolve
brine

ranges

from

into Michigan
salinity
5X

to

basin

(TDS), the

10X

that

of

seawater depending upon formation, with a maximum of 1 2 .6x.
If the brine had originated from seawater and no mineralogic
reactions had affected them, then each element concentration
ratio

(brine

ratio,

at

clearly

not

value/seawater

least
the

various degrees

within
case,

value)

should

individual
as

some

(Br, Ca, Mg, Na,

equal

formations.

elements
K, Sr,Rb,

are
Li,

the

TDS

This

is

enriched

to

B,

I) while

other are depleted to various degrees (alkalinity, S04 , S i ) .
This

type

demonstrate

of

evidence

that

has

formation

been

brine

used
is

not

in

the

simply

past

to

unaltered

164

F i g u r e A-3.

Average Michigan formation brine compostion
(log mg/l) com pa re d with s e a w a t e r ( da sh ed line).
S e a w a t e r d a t a f r o m M c C a f f r e y et al. ( 1 9 8 8 ) .

165

5.6

.4.6

6 .5 1

Log No

O

#5°

5]

X
/

fXTv
4
JS
®n

4.5]

4
4.6 |

~<5
4

A

A
*o

^3.5
o 3
_l
|

2.5

5.5]
o

5

O
?4.5|

CP

4 .5 1

I

«no

4

o>
o

3.5
_l

u»3.5

u

O *

JE
k .

(/)
o»
j2-5!
_i
r*r

2

___1----- 1---- 3.5
25
3
Log Br ( mQ / l )

O
O

BEREA
TRAVERSE
• DUNDEE
■ R IC H F IE L D
A D E T R O I T R.
O N IA G A R A /
S A L IN A
A. O R D O V I C I A N

166

seawater (Chave, 1960; Collins, 1975).

However, salinity is

not the best reference to judge the alterations.
to

better

judge

depletions,

the

magnitude

1975),

concentration
factor:

EF=

is

enrichments

or

In this study, as in others

assumed

can

be

(concentration

conservative,

used

to

ratio

ratio of Br)

factor,

S r , Rb,

Br

ratio

concentration
excess

the

a concentration ratio of a conservative element

must be used as a reference.
(Collins,

of

In order

calculate

of

a

(Collins,

given

an

its

excess

constituent/

1975).

Based on the

elements enriched over seawater

Li, I .

and

include

Ca,

Elements depleted from seawater are C l , Mg,

Na, K ,

alkalinity (HC03) , S04 , and T D S .

Boron and MC12 (Ca

+

+

are

Mg

Sr

-

0.5HC03-S04

conservative with,
to

Br.

Both

enrichments

meq/1)

approximately

or only slightly enriched,

Chapter

and

in

1 and

depletions

chapter

with respect

2 explain

generally

that

reflect

these

water-rock

reactions.
FORMATION COMPARISON
Univariate

statistical

methods

are

used

to

test

equality of water chemistry between different aquifers,

from

different

areas

well

over time

(NCASI, 1985).

to

equality

test

formations.
similarities

within

This

of

an

aquifer,

in

from

a

single

The Student's t-test is used here
mean

comparison

found

or

brine

brine

concentration

is warranted
chemistry

between

in light
using

methods described earlier (Chapter 1, Chapter 2).

of the

graphical

167

Before

applying

the

Student's

t-test,

equality

of

variances must be demonstrated by the F-test, using a null
hypothesis

Ho:

components

meeting

significance
using

a

s22= s l2 '

or

null

populations.

with

the

better,

F-test
the

hypothesis

s2=variance.
at

the

Student's

For

0.01

t-test

Ho:u2=u1 , where

those

level
was

u=mean

of

applied

of

sample

Not rejecting H0 suggests that no statistical

evidence exists

to

suggest

the means

are different,

while

rejection of H Q suggests that there is statistical evidence
for a difference in m e a n s .
For this test, means of the Berea, Traverse, and Dundee
were

first compared to establish

if this group of

samples

could be combined. Generally, the t-test results showed that
means

appear

similar

of

the major

several

between

each

elements

formation.

in

the

However,

Traverse

samples

exhibit a high degree of variance, which caused Br, Cl, Na,
Ca,

and

Dundee

TDS,

to

samples

fail

were

the

used

to

formation water

chemistry

Table

summary

A - 3is

a

components

and

ratios

Generally,

most elements

F-test.

represent

for

of
that

Because
the

of
upper

intraformational

the

t-test

meet

passed

results

the
at

F-test

the

this,

the

Devonian
testing.
for

the

criteria.

0.05level

of

significance for the F- and T-test, respectively.
Inspection of Table A-3

shows the

following:

(1)

mean

concentrations in the Dundee and Traverse formation samples
can be considered statistically similar in the cases where
the

t-test

is

applicable.

Many

of

the

intraformation

TABLE A-3
STUDENT'S t-TEST 'RESULTS
FORMATION
DDTR

DDRF

DDNS

DDTBR

RFNS

R
R
R
R
R
R
R
R

R
R
R
R
R
R
R

NR
R
R
NR
NR
R
NR
R

R
R
NR
R
R
R
NR
NR
R

R
R
NR
R
R
R
R
NR
NR

NR
NR
NR
NR
R
NR
NR
R
NR
NR
NR
NR
R

R
R
R
R
NR
R
R
R
NR
R

NR

R

NR
NR

NR

RFTBR

TB!
NS

COMPONENT
Br
Cl
Na
Ca
Mg
K
Sr
Rb
Li

B
I
si
NH4N
so4
hco3
TDS
D
180

NR
NR
NR
NR
NR
NR
NR
NR
NR
R
R
NR
NR

NR
NR
NR
NR
NR
R
NR

R

R
R
R
R
R
R
R
R
R
R
R
NR
R
R
R
R
NR
NR

KEY:
H0 = mean-L = mean,
NR, R = not rejected, rejected at the level of significance of at
least 0.01 for the F-test, and the 95% to 99% confidence level for
the T-test, depending on F-test results
Blank = t-test not applicable due to rejection by F-test at
the 0.01 level of significance
FORMATION CODES:
DD= Dundee, TR = Traverse-Berea, RF = Richfield-Detroit River,
NS = Niagara-Salina, TBR = Trenton -Black River

169

comparisons

made

using

the

Dundee

samples

apply for Traverse samples as well,
in

the

Richfield-Detroit

River

brine can be considered similar,

were

found

to

(2) Mean concentrations

brine
(3)

and

Niagara/Salina

Mean concentrations in

the Richfield, Detroit River, and Niagara/Salina samples are
statistically
and

(4)

different

Mean

silica

from

(Si)

the

Traverse-Dundee

and deuterium

(D)

samples;

concentrations

are statistically similar between all formation samples.
These results suggests that two groups of brine exists,
the lesser concentrated water from the Berea,
Dundee

water

and

the

more

highly

Traverse, and

concentrated

Richfield,

Detroit River and Niagara/Salina formation waters
CORRELATION COEFFICIENTS
Correlation
relationship
correlation,
might be
and Cl
are

between

A

or negative,

the

degree

of

high

degree

of

between elements

suggestive of a common control.

For example,

Na

concentrations would correlate positively when both

Alternatively,

only

inversely

equilibria,

as

by

strong

equilibrium.

correlate

Ca

decrease in sulfate.
helpful

measure

variables.

either positive

supplied

mineral

coefficients

the

inverse
For

when
cannot

dissolution
correlations

example,

Ca2+

controlled

only

increase

of
might

and
by

without

halite.
suggest

S042=

would

CaS04 mineral
a

concomitant

Although correlation coefficients are

in reducing geochemical data,

elemental correlation

is not evidence of causation (Egleson and Querio,

1969).

170

Table

A-4

logtransformed

is

a

data

correlation
by

the

SAS

matrix

generated

program.

The

for

the

closer

a

correlation coefficient is to ^ 1 , the more the variables are
related.

The

following

is

a

summary

of

significant

correlations from Table A - 4 .

COMPONENT
pH
Br
Cl
Ca
HCO3
I
K
Li
Mg
Na
so4
Sr
Rb
Cs
B
Si
TDS

Several

+ CORRELATION
CORRELATION
Cl, so4
Mg
Rb, Ca, K
Na, S04
S04 , pH
TDS, Mg
Sr, K, Rb,Mg
S04 , Na
NONE
so4
Na, S04
Ca, Sr, Rb
Ca, Rb, B, Sr
Na, S04
B, K, Br, Rb
Na, S04
Ca, Sr, K, Br
S04 , pH
so4
Cs, Rb, K
Na
S r , Ca
Ca, K, Mg
S04 , Na
K, Ca, Cs, Sr,Br
Na
Rb, K, Br
Na
K, Rb, Li
Na
NONE
HCO3
Cl
pH, S04

correlations

are

noted

in

this

reoccurring inverse correlation exists between
many of the other elements.
clear,

but

may

reflect

concentrations

are

samples.

other

Two

found

notable

the

the
most

lowest
highly

correlations

are

A

(Na,S04 ) and

The meaning of this

that
in

table.

S04

is not
and

Na

concentrated
between

HCO3 , and TDS-C1, which might represent "basic" pairs,

Sithat

is, each are more highly correlated with each other than any

TABLE A-4
Correlation matrix
log-transformed data.

PH
BR
CL
CA
HC03
I
K
LI
MG
NA

SO*
S«

a
SI
TOS

MG

1.00000
-0.397*5
-0.51536
-0.36*38
0.35092
- C . 2*392
-0.29*16
-0.06310
-0.37018
0.09390
0.*7277
- C . 29159
- 0 . 3 l q23
-0.13211
-0.171*2
-0.16636
-0.*9935

-0.297*5
1.00000
0 . * 2 713
0.78716
-0.2551*
0 . 38 30 6
0.73060
0 . 7 1279
0.7*98*
-0.*6562
-0.4*378
0.7C307
0.83975
0.7570*
0.701*7
0.27602
0.27669

-0.51636
0.42713
1 .oooco
0.56032
-0.37858
0.47153
0.42915
0.395B7
0 . 6 6 5 0*
0.18757
-0.52 *1 *
0.51537
0.39905
0.0*299
0.21127
-0.05950
0.77501

- C . 36*33
0.78716
0.56032
1.00000
-0.2*165
0.63511
0 . 9*t 59 I
0.7111*
0.89699
-0.65677
-0.69076
0.95165
0.90696
0.7061 *
0.78612
0.29576
0.39155

0.35092
-0 •2 55i*
-0.37 85 8
-0.2*165
I . 00000
-0.10 2*2
-0.23808
-0.00 01 *
-0.32052
0.0*931
0.33815
-0.20803
-0.22176
-0.11275
0.03337
-0.51067
-0.36*76

-0.2*382
0.38306
0.47153
C . 63511
- 0 . 10 2* 2
1.00000
0 . 57 760
0.491*8
0.59988
-0.38150
-0.35*50
0.62891
0.61593
0.*0263
0.53619
0.01753
0.46821

-0.28*16
0.78060
0.42915
0.9*591
-0.23308
0.57760
l . COOOC
0.76103
0.76860
-0.71112
-0.62208
0.88565
0.9*146
0.77362
0.8792*
0.43*57
0.32191

-0.06310
0.71279
0.29597
0 . 7 111*
-0.0C 01*
0.*91*8
0.76103
1.00000
0.48781
-0.35199
-0.29753
0.65719
0.70070
0.5*22*
0.81678
0.02962
0.23*69

-0.57018
0.7498*
0.685C*
0.99699
- 0 . 3 2 n 52
0.59988
0.76960
0.48791
1.00000
-0.465*8
-0.69666
0.36333
0.77237
0.530*2
0.55*88
0.17223
0.53576

3

SI

TOS

-0.171*2
0.701*7
0.21127
0.78612
0.03337
0.53619
0.8792*
0.91678
0.55*88
-0.6*378
-0.33838
0.69*76
0.8*393
0.7*608
1.00000
0.238*1
0.12871

-0.16636
0.27602
-0.05950
0.29576
-0.51067
0.01758
0.43*57
0.02962
0.17223
-0.48192
-0.26296
0.29097
0.41108
0.48577
0.238*1
1.00000
-0.03510

-0.49935
0.27669
0.77501
C . 39155
-0.36*76
0.46821
0.32191
0.23*69
0.53576
0.1*07*
-0.48709
0.36*32
0.296*0
-0.01327
C . 1 2 97 1
-0.03510
l.OOOCC

NA
PH
BR
CL
CA
HC03
I
K
LI
MG
NA
SO*
SR
RB
CS
B
SI
TOS

0.09390
-0.46562
C . 10757
-0.65677
0.0*931
-0.38150
-0.71112
-0.35199
-0.465*8
I . 00000
0.5009b
-0.66066
-0.7160*
-0.76989
-0.6*379
-0.*9192
0.1*07*

SR

R9

- 0 . 2 3159
0.*7277
0.70307
- 0 . * * 37 8
0.51537
-0.52 *1 *
0.95165
-0.69076
-0.20803
0.3391 5
0.62891
-0.35*50
0.99565
-0.62203
0 . 6 5 7 L9
-0.29753
-0 .6 9 6 6 6 ‘ 0.96333
-0.66066
0.50096
-0.70*06
I . 00000
L.0C000
-0.70*06
0.9*562
- 0 . 5 * 0 16
0 . 6 7°53
-0.33165
0.69*76
-0.33838
0.29097
-0.26296
0.36*32
-0.48709

-0.31923
0.33875
0.39805
0.90696
-0.22176
0.61593
0.941*6
0.70070
0.77237
-0.7160*
-0.5*01 6
0 ®8■♦562
1 .CuOOO
0.89302
0.8*393
0.41108
0.296*0

SO* .

CS
-0.13211
0.7570*
0.0*288
0.7061*
- 0 . 11275
0.40263
0.77362
0.5*22*
0.530*2
-0.76989
-0.33165
0.67953
0.89302
I . 00000
0.7*608
0.48577
-0.01327

171

R9

CS

I

PH

HCO 3

LI

CL

CA

K

BR

172

of

the

remaining

variables.

The

meaning

of

these

relationships are explored further using the R-mode analysis
discussed
Egleson
pairs

below.

It

and Querio

in

the

correlation

is

(1969)

Sylvania

not

also

interesting

note

report KC1-B and Br-I

Formation

found

to

in

the

brine from

that

are basic

Michigan,

a

present study. Finally,

elements with similar geochemical behaviors are observed to
correlate together,

for example, Ca-Mg-Sr and K-Rb-Cs.

MULTIVARIATE STATISTICS
Multivariate

statistical

tests

of

Q-mode

and

R-mode

factor analysis were conducted on the data set in order to
evaluate

population

homogeneity

possible geochemical controls.
reduce

the

measured

interpretable
determine
together

that

possible.
of m

called

best-fit
describe

In other words,

variables

(chemical

to

help

much

a fewer

factors.

factors
as

to

elucidate

Factor analysis attempts to

variables

variables,

the

and

by
of

The

grouping
the data

number
goal

is

of
to

variables
variance as

factor analysis combines the set
components

or

wells)

into

the

p

number of best fit, mutually uncorrelated factors, with p<m.
This

is accomplished

in two

fashions,

R and Q mode

factor

analysis.
R-mode factor analysis determines relationships between
variables by extracting eigenvectors and eigenvalues from a
correlation or covariance matrix generated from the chemical
data.

The

amount

of

resulting
variance

R-mode
in

the

factors
chemical

describe
data.

the

maximum

Variables

173

(chemical
scores,

components)

which

factor.

show

Because

according to

assigned

their

the

their

are

factor

importance

variables

are

in

loadings,

describing

grouped

intercorrelations, the

into

factors

or
each

factors
may,

or

may not, be interpretable to reflect geochemical processes.
Principal component analysis
to R-mode

factor

manipulation
of

a

analysis,

(principal factors)
but

is strictly

a mathematical

involving the calculation of the

variance-covariance

or

correlation

is similar

eigenvectors

matrix

(Davis,

1986).
Q-mode factor analysis is concerned with relationships
between

individual

samples.

That

is,

it

attempts

to

discover patterns within a group of objects.

This test is

used to

a mixture

determine

if a sample population

is

various end-members, thereby testing data homogeneity.

of
Q-

mode extracts eigenvectors and eigenvalues from a similarity
matrix

between

analysis

all

also

possible

generates

interrelationships between
chemical characteristics.
help

deduce

end

pairs

member

of

samples.

factors

samples,

Q-mode
describing

based on their

similar

These factors then can be used to
water

compositions.

If

Q-mode

results show the data set is homogeneous, that is, if only 1
factor describes the relationships between samples,

then R-

mode can be applied to the entire data set.
Excellent reviews of R and Q-mode

factor analysis

are

found in Klovan (1975), Joreskog et a l . (1976), Davis (1986)

174

and Long et al.

(1986) .

No attempt will

be made here to

present further detailed explanation of these procedures.
INTERPRETING FACTORS
The
not,

factors

be

generated

interpretable

in R-mode

processes.

Factor makeup can change due to choice of data,

the number

of

rotate

extracted,

data to get the
discusses
factors,

some
and

and

"best
of the

the

reflect

or may

geochemical

factors

to

analysis may,

the

fit"

methods

used

to

to the variance.

difficulties

philosophy

in

Davis

the

(1986)

in selecting number

choosing

rotation

of

methods.

The changes on the R-mode factors generated in this study as
the result of several different rotation methods and choice
of

data

are

evaluated

interpretations

of

below.

factors

Additionally,

may

be

because

subjective,

it

is

instructive to determine if similarities exist in the factor
loading patterns

generated

The past work of Kramer

in other

(1969),

ground water

studies.

Egleson and Querio

(Figure A - 4 ) , Hitchon et a l . (1971)

(1969)

(Figure A - 5 ) , and Long

(Figure A - 6) are summarized next.

et a l . (1986)
PAST WORK
Kramer
2400

brine

region.

(1969)
samples,

applied

R-mode

separated

by

factor

analysis

formation

rock

common

over

type

and

The analysis used the squared multiple correlations

of the data and the varimax rotation method.
Kramer's

to

study

to

use

characteristics

of

discussed below.

was

factor

analysis

formation

The goal
to

waters,

of

illustrate
which

are

1
Mg

Fe

M g-P R I,C A

Sr
Ca/Mg

Depth, Fe

K,Na,Ca/Mg

Figure

A-4.
R-mode
A
rotation)
Sylvania
B a s in .

NHo
M g-P R I,S t> 4,l,B

U.K
Depth,Sr
Na
Spec. Gr.
Ca

NH3,I
Mg,Na,K,U,Sr
Br,B,SQ4,Rb, Depth
Mg-PRI,Ca/Mg
Ca
Sr.Spoc. Gr.

Mg-PRI,Ca

Br,NH3,l,U
K,8

Fe,Br,Rb,

Spec. Gr.

Ca/Mg

(V a r i m a x
results
factor
analysis
,,n c o \
9
for
- eson
and
Querio
( 1 9 66 99 )
from
Egl
Michigan
Sandstone
formation
waters,

1 75

Mg
Spec Gr.,S04
Na

Fe
Depth
Mg,U,Sr,Br,B
NH3,l,S04,Hb

176

Figure

A-5.

R - m o d e f a c to r a n a l y s i s r e s u l t s f r o m H i t c h o n et
a l . (1971), for W e s t C an a d a b as i n w a t e r s . TOP:
varimax
rotation,
BOTTOM:
bi-quartum
oblique
rotation .

177

Hitchon et al . (1971)

R-MODE VARIMAX ROTATION

K .U .R b
Cl.C a.M o .N a.S T
Br.TDS

Na.Sr

4
K ,U ,M g .R b
CI,Ca,N a,Sr,TD S
Br

Hitchon et al.

(1971) R-HODE B1QU ART I MIN
OBLIQUE ROTATION.

Figure A-5

178

Figure

A - 6 . R - m o d e f a c t o r a n a l y s i s r e s u l t s fro m L o n g et a l .
( 1 9 8 6 ) f o r n e a r - s u r f a c e s a l i n e g r o u n d w a t e r s in
M ichi g a n .

179

Long

a l.

et

1

R-MODE

2

N l.C I
COND
K

-

(1986)

NH,
—
Mn

-

-

-

Fe
Mn

_

ROTATION

3
Ca
Mg

-

VAR1MAX

ALK
SI

K
Co
Mg
Fe
F,pH

SI,Zn
s o 4,n h .
COND
Cl

s o 4 ,a l k ,si

nh

4

pH,Fe,Zn

Na.ALK

Na,CI,F,Mn
Mg,COND,Ca

PH

K

F

so4
----------------------—

1

2

Na,Cl,COND
K

Ca
Mg

NH.
Ca.Mn
Mg

3

Mn
Fe
K

ALK.Si

COND
Fe

NH4 ,Zn
Na.Cl
S 04

so4

-

nh4

Si
F.pH.Fe.Zn

ALK,F,pH
ALK

Mn
SI

PH
Mg.Na.CI
Ca.COND
k ,s o 4

F

-

•

Long

et

al.

-

(1986)

R-MODE

Figure »-6

-

-

PROMAX

ROTATION

180

Egleson

and

Querio

(1969)

conducted

R-mode

factor

analysis (varimax rotation) on the Sylvania Formation waters
from the Michigan basin.

Both chemical

and physical

data

were used, but it is not clear if their data were first logtransformed
found to
first

before

account

factor

had

analysis.

for

Although

100% of the

interpretable

data

ten

factors

variance,

loadings;

the

were

only

the

remaining

factors had high loadings on only a single element.

Their

first factor showed high loadings for the minor elements and
production

depth,

interpreted

to

reflect

a

common

source

related to paleotopography.
Hitchon et a l . (1971) used R and Q-mode factor analysis
to study West Canada sedimentary basin formation waters.
mode factor analysis showed their data are homogeneous.
variation

in

the

Q-mode

results

were

combined

with

QThe

flow

direction and depth and used as evidence for shale membrane
filtration.

R-mode

factor

analysis

was

without the partialling out of salinity,
oblique rotation methods.
was

conducted

preformed

both

using varimax and

A second R-mode factor analysis

combining the

results

from the

first

R-mode

test with physical parameters and stable isotopic data.

The

R-mode results were interpreted as evidence that the brine
originated

from

seawater

modified

by

shale

membrane

filtration.
Long

et

al.

(1986)

analysis to study saline,
mid-Michigan area.

used

both

Q

and

R-mode

factor

near-surface groundwaters

in the

Although the Q-mode results demonstrated

181

a single population
scores

suggested

of water

two

types

exists,

the

extracted

of waters, a normal

Q-mode

background

(regional) water and localized saline water.

Q-mode scores

were

saline

contoured

infiltration.

to

delineate

areas

of

water

Four different R-mode rotation methods were

used, with similar factors produced by each method.

Factors

were

factor,

interpreted

to

show:

(1)

a

saline

water

indicating brine input, (2 ) a water-rock factor,
carbonate,
halogen

oxide,

factor

and

clay

reflecting

mineral

reactions,

sulfate

reduction

reflecting
(3)

and

a

non­

silicate

interactions.
FACTOR PATTERNS
These
factor

loadings

processes.
and

the

loading

studies
can

Kramer

processes
between

suggests mineral
anion-anion
reactions,
single

or
and

factor

provide
be

anions

might
and

in

light

several
a

represent:
cations

of

(2) a direct

cation-cation

suggests

illustrations
of

of

how

geologic

outlined several general patterns

equilibria,

(3)

good

interpreted

(1969)
they

a

might

or

inverse
minerals

loading

loading

source

an

common

represent

variables
common

(1 )

between

ion-exchange
highly
origin.

in

a
In

addition to these patterns, the other studies have suggested
the following interpretations of factor pattern .
A common source:

Hitchon et a l . (1971) interpreted the

high loadings on Li, Rb, K, T D S , Mg, Ca, Na, Cl, Sr, Br, and
S04 in their first R-mode factor
common

origin

for

these

(Figure A-5)

elements,

seawater.

as showing a
A

similar

182

interpretation of many elements loading together in a single
factor was made by Egleson and Querio (1969)
Mixing of waters:

(Figure A - 4 ) .

The correlation between the first Q-

mode factor with decreasing salinity was used by Hitchon et
al.

(1971)

(1986)
Na,

to support dilution by freshwater.

Long et a l .

interpreted an R-mode factor showing high loadings on

Cl,

K,

conductivity, Ca,

Mg,

NH 3 to

and

reflect

input and mixing of saline water (Figure A - 6) .
of

their

first

Q-mode

factor,

having

a

the

Contour maps

high

loading

on

alkalinity, revealed the location of brine inputs.
Membrane filtration:
second

Q-mode

interpreted
membrane

by

factor

in

Hitchon

filtration.

The increasing importance of the
a
et

High

down-dip
al.

flow

(1971)

loadings

to

of Mn,

and high negative loadings on Cu, Na,

direction
reflect

Rb,

was
shale

S r , and

Zn,

C l , HCC>3 , and TDS in

this factor for downflow samples suggested that metal rich,
lower salinity ground waters exit the outflow side of shale
layers.

In

their

R-mode

analysis

combining

chemical

and

physical data, high positive loadings on the seawater factor
(factor 1, Figure A-5)

of temperature, depth,

pressure,

and

SD, opposed by high negative loadings on pH, was thought to
reflect increased membrane efficiency with dept h, pressure,
and temperature, as well as the isotope mixing hypothesis of
Hitchon and Friedman (1969).
Dissolution of halite or other minerals:
other

data,

positive

Hitchon

loading

on

et

al.

both

(1971)
Na

and

thought
Cl

In addition to
that

suggested

a

weak

halite

183

dissolution

(Figure A - 5 ) .

This pattern

is

in contrast to

factors interpreted to show mineral equilibria,
inverse loadings of common mineral components

which

show

(Kramer, 1969;

Long et al., 1986).
Dolomitization:
evidence
thought

in

their

results

that

high

positive

seawater

factor

with

dolomitization.
Kramer

Hitchon

This

to

al.

(1971)

suggest

loadings

TDS
is

et

little

dolomitization,

of

Ca

partialled

consistent

found

and

out

with

Mg

in

may

the

but
the

reflect

findings

of

(1969), who interpreted high positive loadings on Ca

and Mg coupled with a negative loading on HC03 as reflecting
dolomitization.
and

HC03 are

Feth

Similar bipolar factor loadings of Ca, Mg,

also

reported

by

Collins

^1967), and Long et a l . (1986).

Ca-Mg

pattern

is

associated

with

these

studies

supported

and

Although this typical
dolomitization,

studies have not found this result,
in

(1969a) , Dawdy

several

even though other data

dolomitization

(Collins,

1969 ;

Lee, 1969; Egleson and Querio, 1969).
Sulfate
alkalinity

reduction:

and

sulfide

Bacterial
was

reduction

thought

to

be

of

sulfate

to

represented

by

bipolar loadings on pH2S and S04 by Hitchon et a l . (1971) .
Long

et

between

al

(1986)

interpreted

alkalinity

reduction (Figure

and

a

sulfate

strong
to

bipolar

represent

loading
sulfate

A-6).

Formation of chlorite: An inverse loadings of Mg and Fe
was

interpreted

by

Hitchon

et

chlorite formation (Figure A-5).

al.

(1971)

to

represent

184

Cation exchange:

The bipolar loadings of Ca and Na

in

the R-mode results with salinity partialled out was thought
by Hitchon et al.
A-5).

Also,

decreasing

Rb>K>Sr>Li>Mg
because this
cations

on

(1971) to reflect cation exchange

were

factor

thought

to

loadings
represent

is exactly the order of
clays

(decreasing

in

(Figure

the

cation

order

exchange,

"replacing” power of

non-hydrated

ionic

radius,

increasing hydrated ionic radius, and polarizability).
Contribution from organic matter: Hitchon et al.
concluded

that

similar

factor

scores

represent an organic material source
negative

loadings

of

I , Br,

and

for

Br

and

(Figure A - 5 ) .

NH3 , and

the

(1971)

I

might
Highly

correlation

between I , B r , and production depth in a single factor were
interpreted
halogen

by

source

Egleson

and

from organic

Querio
debris

(1969)

to

reflect

in topographically

a
low

areas of the Sylvania Formation (Figure A - 4 ) .
Mineral

equilibria

Hitchon et a l . (1971)
S04

along

with

high

control:

Both

Kramer

(1969)

and

interpreted high loadings of HC03 and
negative

loadings

of

Ca,

Mg,

and

Sr

(with salinity partialled out) to reflect dolomite, calcite,
and celestite equilibrium.
loadings

of

Ba

and

S04

Kramer
reflect

(1969)
barite

suggested inverse
equilibrium.

An

inverse loadings on Sr and S04 was also found by Egleson and
Querio (1969)
inverse

(Figure A - 4 ) .

loadings

of

Si

Long et a l . (1986)
and

K

to

interpreted

represent

silicate

interactions (Figure A - 6) . Equilibrium was also supported by
chemical model calculations.

Kramer (1969) interpreted both

185

the

inverse

cations),

loading

and

of Si

positive

and Cl

pH-Si02

(thought to
loadings

represent

to

represent

aluminosilicate equilibrium.
Q-MODE METHODS AND RESULTS
Q-mode

factor

analysis

was

preformed

in

this

study

using the SAS statistical program on the IBM 4381 computer
at

Michigan

State

University.

transformed

similarity

the

theta

cosine

A

matrix

was

coefficient

of

non-standardized

used,

constructed

proportional

logusing

similarity

suggested by Davis (1986).
The principal factor solution for the first five Q-mode
factors,

their

eigenvalues,

total

variance,

and

the

cumulative percentages of variance are listed below.

Factor

1

2

3

4

5

Eigenvalue

140.7

0.558

0.354

0.202

0.714

Difference

14 0.2

0.204

0.151

0.132

0.033

Proportions

0.991

0. 004

0.003

0.001

0.001

Cumulative %

0.991

0.995

0.998

0.999

0.999

The

first

variance of the
from

a

analysis

single
can

be

eigenvector
data,

accounts

for over

strongly suggesting that

homogeneous
conducted

99% of

the data are

population, and that

using

the

the

entire data

R-mode
set.

A

single eigenvector explaining such a large percentage of the

186

data variance

implies that a single gradient characterizes

the

This

samples.

is

interpreted as a salinity gradient,

similar to the interpretations of Hitchon et al.

(1971)

and

Long et a l . (1986).
Davis

(1986)

suggests that a dominant

first factor

is

often obtained in Q-mode analysis, such as is the case here.
Such a dominant factor reveals little about the structure of
the

data,

second

and

and

magnitude

is

considered

third
and

a

"nuisance"

eigenvectors,

accounting

for

although
small

factor".

much

This was

the

case

for Hitchon et

al.

the

whose

variance.

(Davis,

and

factor accounted

The

amount

of

in
the

2.986) .
first

for Long et

first Q-mode
extreme

of

(1971) , whose

factor accounted for 55% of the variance,
a l . (1986)

smaller

proportions

variance, may reveal more useful information

The

for 95%

variance

of

(>99%)

explained by the first factor in the present study, however,
suggests that further Q-mode factoring may not be warranted.
R-MODE FACTOR METHODS
R-mode

factor

transformed
available

data
on

University.

the
The

analysis

set,

using

IBM

4381

review

of

was

preformed

the

SAS

computer
other

on

the

statistical
at

factor

package

Michigan
analysis

log-

State
studies

demonstrated three problems that might be encountered in the
R-mode factor analysis of brines,
this study.

Firstly,

which were considered

salinity might dominant the

in

factors,

as was suggested by the Q-mode analysis and by the results
of Hitchon et al.

(1971).

In order evaluate the effects of

187

salinity in the present study, TDS was partialled out of the
R-mode analysis.

A comparison of the effects

discussed below.

Secondly,

of this

are

missing or erroneous variables

might bias factors, especially if the variable represents an
important species such as pH.

Many of the samples obtained

from outside this study (M.D.N.R. and oil company open file
data)

did

study.

not

To

report

address

were evaluated,

minor

this

components

concern,

two

measured

subsets

of

(1) the entire data set was used

in

this

the

data

including

samples with missing data, and (2 ) only those samples having
complete data.
the resulting
A-5,

case

because

No difference were

factors using either data

la,

the

and

pH

A-5

case

lb) .

measurements

concentrated waters,
and without

found

are

In

in the makeup

set
a

suspect

(compare Table

similar
in

the R-mode analysis was

including the pH data.

of

fashion,

these

highly

run both with

Slight differences

in

the makeup of the extracted factors were found when pH was
included, which is discussed below.
R-MODE RESULTS
Table

A-5

reports

the

eigenvectors,

communalities, and

factor scores

combinations

data.

of

diagrammatically

in

for each of the different

The

Figures

eigenvalues,

results

A-7

to

are

A-10.

displayed
Components

plotting towards either end of these diagrams have
approaching
promax
shown;

— 1.

rotation

The
results

principal
of

the

component,
following

loading

varimax,

four

trials

and
are

188

Case

1

(results

samples having pH
study)

are

used,

shown

in

Figure

A-7):

only

those

(generally only samples collected in this
pH variable

is

included,

and TDS

is

not

partialled out.
Case lb
is

(not shown): all samples are used, pH variable

included,

and TDS

is not partialled

out.

Results

are

listed in Table A-5 for comparison with case 1.
Case 2 (Figure A - 8) : All samples are used, pH variable
is included, and TDS is partialled out.
Case 3 (Figure A - 9 ) : All samples are used, pH variable
is not included, and TDS is partialled out.
Case 4 (Figure A-1 0) : All samples are used, pH variable
is not included, TDS is not partialled out.
The eigenvalues,
that

has

been

representing the variance in the data

extracted

onto

each

factor,

factors account

for 90 to 92% of the variance

(Table

After

A-5).

the

fourth

factor,

show

that

4

in the data

the

amount

of

variance explained be each factor drops to less than 3% of
the variance,

well

below

the

estimated

experimental

error

for the components.

Communalities (Table A - 5 ) , a measure of

the efficiency that

each variable

and

show

the

factors

variability in the data.

are

is explained,

efficient

in

are

explaining

>0.75
the

Exceptions are pH, alkalinity, and

I, having communalities near 0.5.

This could signify that

additional factors need to be extracted in order to better
explain these later variables.

189

TABLE A-5
R-MODE FACTOR ANALYSIS RESULTS
Case 1) Only samples with pH are used
pH variable included
TDS not partialled out

1

Factor
Eigenvalue
Difference
Proportion
Cummlative

9.117
6.597
0.5960
0.5960

COMMUNALITIES
PH
.5177
Aik .5372
MG
.8931
RB
.9598
TDS .6799
TOTAL 13.8118

BR
I
NA
CS

2

2.5206
1.0682
0.1648
0.7608

.8189
.5404
.9797
.8789

A) PRINCIPAL FACTORS
FACTOR 1_______FACTOR 2
CA
.9761
PH
.9636
K
.9636
.4976
CS
RB
.9582
B
.3463
SR
.9279
Aik .3335
MG
.8810
S04 .3018
BR
.8352
SI
.2571
B
.8118
.1848
RB
CS
.7759
.1476
K
LI
.7254
LI
.1405
I
.6451
BR
.07602
CL
.5455
SR
-.0022
TDS .4224
-.0171
CA
SI
.3652
I
-.1473
Aik -.2861
MG
-.3128
PH
-.4216
NA
-.5646
NA
-.6708
TDS -.7060
S04 -.6839
CL
-.7767

CL
K
S04
B

3
1.4524
0.7109
0.0949
0.8557

0.74145
0.2657
0.0485
0.9042

.9469
.9520
.7611
.8604

CA
LI
SR
SI

FACTOR 3
Aik .5173
LI
.4984
PH
.2892
S04 .2765
B
.2737
NA
.2656
I
.2580
CL
.1573
BR
.0678
CA
.0539
TDS .0456
SR
.0366
K
.0075
RB
-.0014
MG
-.0249
CS
-.1103
SI
-.7181

5
0.4757
0.4570
0.0311
0.9353

.9640
.8583
.9001
.7831

FACTOR 4
NA
.3747
S04 .3548
BR
.3328
SI
.2605
LI
.2527
CL
.1460
CS
.1303
RB
.0864
B
.0811
K
.0408
PH
.0316
TDS .0300
CA
-.0900
MG
-.1358
I
-.1897
SR
-.1941
Aik -.2765

190

TABLE A-5 (Continued)
Case 1 (continued)
B) VARIMAX Rotation
FACTOR 1_______FACTOR 2_______FACTOR 3_______ FACTOR 4
B
.9063
CL
.9271
NA
.7551
SI
.8366
LI
.8919
TDS .8160
S04 .5831
CS
.3545
RB
.8682
MG
.6562
PH
.1753
BR
.2545
K
.8496
CA
.4251
CL
.1273
RB
.2399
BR
.8232
I
.3988
LI
.0946
I
.2087
CS
.8054
SR
.3930
TDS .0222
CA
.0942
CA
.7808
BR
.2762
BR
-.0087
MG
.0917
SR
.7117
K
.2704
Aik -.0220
SR
.0554
MG
.5555
RB
.2341
SI
-.1986
B
-.0109
I
.5118
NA
.2176
B
-.1986
TDS -.0252
CL
.2611
LI
.1591
I
-.2789
CL
-.0552
SI
.1999
B
.0177
CS
-.2988
LI
-.1685
TDS .1136
SI
-.0606
SR
-.3062
S04 -.1768
Aik -.0007
CS
-.1233
K
-.3369
I
-.2039
PH
-.0360
ALk -.4355
MG
-.3814
NA
-.2519
S04 -.2465
S04 -.5736
CA
-.4059
PH
-.2576
NA
-.5465
PH
-.6475
SR
-.4858
Aik -.5892
C) PROMAX Rotation
FACTOR 1
FACTOR 2
RB
.9455
CL
.9548
K
.9362
TDS .8225
B
.9113
MG
.7484
CA
.8988
CA
.5568
LI
.8686
SR
.5174
BR
.8666
I
.4858
SR
.8353
K
.4117
CS
.8287
BR
.4010
MG
.7168
RB
.3775
I
.5939
LI
.2954
CL
.3936
B
.1677
SI
.2700
NA
.1028
TDS .2484
CS
.0139
Aik -.1141
SI
-.0266
PH
-.1918
Aik -.4251
S04 -.4377
S04 -.6249
NA
-.6208
PH
-.6493

FACTOR 3
NA
.9182
S04 .7093
PH
.2914
Aik .1577
CL
-.0648
TDS -.0974
LI
-.2366
BR
-.4100
I
-.4372
SI
-.4390
B
-.5231
MG
-.6377
CS
-.6451
RB
-.6843
K
-.7008
CA
-.7245
SR
-.7659

FACTOR 4
SI
.8684
CS
.4515
RB
.3245
BR
.2998
K
.2939
CA
.1749
SR
.1422
MG
.1404
B
.0784
TDS -.0626
CL
-.0991
LI
-.1190
I
-.1531
S04 -.2288
PH
-.2428
NA
-.3874
Aik -.5571

191

TABLE A-5 (Coi -ined)
Case 1-b) Sane as case 1 except all samples are used.

1

Factor
Eigenvalue
Difference
Proportion
Cumulative

9.1174
6.5968
0.5960
0.5960

COMMUNALITIES
.5176
PH
Aik .5372
MG
.8931
RB
.9598
TDS .6799
TOTAL 13.832

BR
I
NA
CS

2

2.5206
1.0681
0.1648
0.7608

.8189
.5404
.9797
.8788

A) PRICXPAL FACTORS
FACTOR 1_______FACTOR 2
CA
.9761
PH
.5126
K
.9636
CS
.4976
RB
.9582
B
.3463
SR
.9279
ALK .3335
MG
.8810
S04 .3018
BR
.8353
SI
.2571
B
.8118
RB
.1848
CS
.7759
K
.1476
LI
.7254
LI
.1405
I
.6451
BR
.0760
CL
.5455
CA
-.0171
TDS .4224
SR
-.0022
SI
.3652
I
-.1473
ALK -.2862
MG
-.3128
PH
-.4126
NA
-.5646
NA
TDS -.7060
-.6708
S04 -.6839
CL
-.7767

CL
K
S04
B

3
1.4524
0.7109
0.0949
0.8557

4
0.7414
0.2658
0.0485
0.9042

.9469
.9520
.7611
.8605

CA
LI
SR
SI

.9641
.8583
.9000
.7831

FACTOR 3_______ FACTOR 4
ALK .5173
NA
.3747
LI
.4984
S04 .3548
PH
.2892
BR
.3329
S04 .2765
SI
.2605
B
.2737
LI
.2528
NA
.2655
CL
.1460
I
.2580
CS
.1303
CL
.1573
RB
.0864
BR
.0678
B
.0811
CA
.0539
K
.0408
TDS .0455
PH
.0316
SR
.0366
TDS .0300
K
.0075
ALK -.0900
RB
-.0014
MG
-.1358
MG
-.0249
I
-.1897
CS
-.1103
SR
-.1941
SI
-.7181
HC03 -.2765

5
4757
6457
0311
9353

192

TABLE A-5 (Continued)
Case lb (continued)
B) VARIMAX Rotation
FACTOR 1
FACTOR 2
B
.9063
CL
.9271
LI
.8919
TDS .8160
RB
.8682
MG
.6562
K
.8496
CA
.4251
BR
.8233
I
.3988
CS
.8054
SR
.3930
CA
.7808
BR
.2762
SR
.7117
K
.2704
MG
.5555
NA
.2176
I
.5118
RB
.2341
CL
.2612
LI
.1591
SI
.1999
B
.0177
TDS .1136
SI
-.0606
ALK -.0008
CS
-.1233
PH
-.0360
ALK -.4355
S04 -.2465
S04 -.5736
NA
-.54654
PH
-.6475

FACTOR 3
NA
.7552
S04 .5832
PH
.1753
CL
.1273
LI
.0947
TDS .0222
ALK -.0221
BR
-.0087
B
-.1964
SI
-.1986
I
-.2789
CS
-.2988
RB
-.3062
K
-.3369
MG
-.3814
CA
-.4059
SR
-.4858

FACTOR 4
SI
.8366
CS
.3544
BR
.2545
RB
.2399
K
.2087
CA
.0942
MG
.0917
SR
.0555
B
-.0109
TDS -.0252
CL
-.0553
LI
-.1685
S04 -.1768
I
-.2039
NA
-.2519
PH
-.2576
ALK- .5892

C) PROMAX Rotation
FACTOR 1
FACTOR 2
RB
.9455
CL
.9518
K
.9362
TDS .8225
B
.9114
MG
.7484
CA
.8988
CA
.5568
LI
.8686
SR
.5174
BR
.8666
I
.4858
SR
.8353
K
.4117
CS
.8287
BR
.4010
MG
.7168
RB
.3775
I
.5939
LI
.2954
CL
.3936
B
.1677
SI
.2700
NA
.1028
TDS .2484
CS
.0139
ALK -.1141
SI
-.0266
PH
-.1918
ALK -.4251
S04 -.4377
S04 -.6249
NA
-.6208
PH
-.6493

FACTOR 3
NA
.9182
S04 .7093
PH
.2914
ALK .1577
C'L
-.0648
TDS -.0974
LI
-.2366
BR
-.4100
I
-.4372
SI
-.4390
B
-.5231
MG
-.6377
CS
-.6451
RB
-.6842
K
-.7008
CA
-.7245
SR
-.7569

FACTOR 4
SI
.8684
CS
.4515
RB
.3245
BR
.2998
K
.2939
CA
.1749
SR
.1422
MG
.1405
B
.0785
TDS -.0626
CL
-.0991
LI
-.1190
I
-.1531
S04 -.2288
PH
-.2428
NA
-.3874
ALK -.5571

193

TABLE A-5 (Contined)
Case 2) All samples are used,
pH variable included
TDS partialled out.
Factor
Eigenvalue
Difference
Proportion
Cummlative

1
8.775
7.1462
0.6141
0.6141

COMMUNALITIES
PH
.3776
ALK .4500
MG
.8876
RB
.9579
TOTAL 12.66

BR
I
NA
CS

2
1.6290
0.2249
0.1140
0.7281
8018
4500
9889
8727

A)PRINCIPAL FACTORS
FACTOR 1_______FACTOR 2
CA
.9724
ALK .5664
K
.9671
LI
.5240
PH
RB
.9652
.5240
SR
.9197
S04 .3956
CS
.8713
B
.3156
B
.8487
I
.2953
MG
.8484
NA
.2046
BR
.8236
CL
.0746
.7121
BR
.0624
LI
I
.5581
CA
.0355
SI
.4191
K
.0308
CL
.3540
RB
.0296
ALK -.1447
SR
.0248
PH
-.2390
CS
-.0345
S04 -.5985
MG
-.1209
-.6408
NA
-.8268
SI

CL
K
S04
B

3
1.4041
0.5441
0.0983
0.8263

4
0.8579
0.2715
0.0602
0.8865

.8955
.9481
.6978
.8552

CA
LI
SR
SI

.9690
.8401
.8927
.7837

FACTOR 3_______ FACTOR 4
CL
.8568
S04 .3980
NA
.3913
BR
.3343
MG
.3484
NA
.3320
CA
.1187
SI
.3303
SR
.0892
LI
.2394
BR
.0887
CL
.1742
LI
.0347
CS
.1644
I
-.0561
RB
.1085
ALK -.0799
B
.0824
K
-.0896
PH
.0731
RB
-.1166
K
.0626
S04 -.1570
CA
-.0900
B
-.1700
MG
-.1786
CS
-.2919
SR
-.1943
SI
-.2973
I
-.2177
PH
-.3253
ALK -.3192

5
5884
1331
0412
9277

TABLE A-5 (Continued)
Case 2 (continued)
B) VARIMAX Rotation
FACTOR 1_______FACTOR 2_______FACTOR 3_______ FACTOR 4
B
.9165
S04 .7701
SI
.8019
CL
.9086
RB
.9083
NA
.6064
CS
.3015
NA
.3911
K
.8949
PH
.4422
BR
.2732
MG
.3634
LI
.8627
LI
.1101
RB
.2166
BR
.2812
CS
.8511
ALK .1062
K
.1871
LI
.2393
.8249
CA
BR
-.094 0
MG
.1056
CA
.2027
BR
.7996
B
-.1159
CA
.0761
SR
.1376
SR
.7586
SI
-.1648
SR
.0198
K
.0482
MG
.5905
CL
-.1795
CL
.0123
RB
.0355
I
.5655
CS
-.2000
B
-.0405
B
-.0056
SI
.2661
I
-.2512
NA
-.1577
I
-.0199
CL
.1935
RB
-.2911
LI
-.1628
S04 -.0616
PH
.0426
K
-.3314
S04 -.1768
ALK -.1232
ALK .0307
CA
-.4947
I
-.2584
CS
-.1320
S04 -.2641
SR
-.5452
PH
SI
-.3313
-.2065
NA
-.6659
MG
-.6289
ALK -.6501
PH
-.2655
PROMAX Rotation
C) :
FACTOR 1
FACTOR 2
RB
.9642
S04 .8334
K
.9582
NA
.7999
CA
.9239
PH
.4589
B
.9124
ALK .2018
CS
.8866
LI
-.1582
SR
.8629
CL
-.2522
BR
.8345
SI
-.3722
LI
.8158
I
-.3744
MG
.7365
B
-.4013
I
.5819
BR
-.4021
SI
.3455
CS
-.5133
CL
.2776
RB
-.6059
ALK -.0508
K
-.6343
PH
-.0936
CA
-.7447
S04 -.4381
SR
-.7616
NA
-.7674
MG
-.8039

FACTOR 3
SI
.8471
CS
.4149
BR
.3514
RB
.3405
K
.3141
MG
.2230
CA
.2092
SR
.1558
B
.0690
CL
.0138
LI
-.0942
I
-.1661
S04 -.2887
NA
-.3123
PH
-.3625
ALK -.6454

FACTOR 4
CL
.9082
NA
.3936
MG
.3615
BR
.2758
LI
.2589
CA
.2062
SR
.1432
K
.0477
RB
.0339
B
.0075
I
-.0003
S04 -.0516
ALK -.0875
CS
-.1382
PH
-.2448
SI
-.2463

195

TABLE A-5 (Confined)
Case 3) All samples are used,
pH variable not included
TDS partialled out

1

Factor
Eigenvalue
Difference
Proportion
Cummulative

8.7132
7.2014
0.6513
0.6413

Communalities
Br
.7983
I
.4440
NA
.9898
CS
.8827
T0TAL=12.3466

CL
K
S04
B

1.2784
0.4353
0.0956
0.8599

4
0.8432
0.4057
0.0630
0.9229

.9755
.8138
.9097
.7611

ALK
MG
RB

2

1.5118
0.2333
0.1130
0.7643

.9167
.9479
.7365
.8822

A) PRINCIPAL FACTORS
FACTOR 1_______FACTOR 2
.9730
CA
ALK .4809
K
.9696
LI
.4636
RB
.9663
CL
.4534
SR
.9209
NA
.3669
CS
.8729
S04 .2924
B
.8508
I
.2323
MG
.8399
B
.2119
BR
BR
.8219
.0918
LI
.7177
CA
.0740
I
.5604
MG
.0625
SI
.4131
SR
.0364
CL
.3491
K
-.0283
ALK -.1389
RB
-.0291
S04 -.5933
CS
-.1670
NA
-.8271
SI
-.7017

CA
LI
SR
SI

.4789
.8474
.9624

FACTOR 3_______ FACTOR 4
ALK .3747
S04 .4286
S04 .3395
BR
.3327
B
.3198
NA
.3225
CS
.2421
SI
.3043
I
.1858
LI
.2336
LI
.1711
CS
.1849
RB
.1449
CL
.1344
K
.0597
RB
.1207
BR
-.0592
B
.1051
SI
-.0742
K
.0579
CA
-.1139
CA
-.1011
SR
-.1216
MG
-.1769
NA
-.2588
I
-.2035
MG
-.3261
SR
-.2132
CL
-.7557
ALK -.2966

5
4375
0403
0327
9556

196

TABLE A-5 (Continued)
Case 3 (continued)
B) VARIMAX Rotation
FACTOR 1_______FACTOR 2_______ FACTOR 3_______FACTOR 4
B
.9077
SR
.6731
SI
.7988
CL
.9314
RB
.8650
MG
.6473
CS
.3057
MG
.3877
LI
.8531
CA
BR
.6035
.2685
NA
.3441
K
.8249
K
.4648
RB
.2237
BR
.2782
CS
.8238
RB
.4016
K
.2089
CA
.2517
BR
.7863
I
.3421
MG
.1179
LI
.2413
CA
.7328
CS
.3086
CA
.1042
SR
.2097
SR
.6386
B
.2338
SR
.0687
K
.0867
I
.5180
SI
.2001
CL
.0237
RB
.0524
MG
.5137
BR
.1747
B
-.0576
I
.0044
SI
.2288
CL
.1513
LI
-.1631
B
-.0150
CL
.1604
LI
.0341
NA
-.1858
CS
-.1236
ALK .0543
ALK -.0757
S04 -.2227
S04 -.1349
S04 -.1269
NA
-.7202
I
-.2421
ALK -.1448
NA
-.5641
S04 -.8078
ALK -.6703
SI
-.1748
C) PROMAX Rotation
FACTOR 1
FACTOR 2
RB
.9600
SR
.8671
K
.9411
CA
.8454
B
.9247
MG
.8176
CA
.8967
K
.7609
CS
.8876
RB
.7179
BR
.8422
CS
.6231
LI
.8276
B
.5251
SR
.8204
BR
.5028
MG
.7120
I
.4534
I
.5703
SI
.3584
SI
.3271
LI
.3184
CL
.2699
CL
.2514
ALK -.0366
ALK -.1786
S04 -.3779
S04 -.8324
NA
-.7293
NA
-.8741

FACTOR 3
SI
.8506
CS
.4594
RB
.3913
K
.3806
BR
.3775
CA
.2814
MG
.2676
SR
.2482
B
.0975
CL
.0245
LI
-.0571
I
-.1153
S04 -.3594
NA
-.3869
ALK -.6538

FACTOR 4
CL
.9388
MG
.4203
NA
.3089
BR
.3052
CA
.2948
LI
.2907
SR
.2519
K
.1259
RB
.0914
I
.0498
B
.0369
CS
-.0926
ALK -.1086
S04 -.1480
SI
-.2004

197

TABLE A-5 (Contined)
Case 4) All samples are used,
pH variable not included
TDS not partialled out.
Eigenvalue
£
Difference
K?
Z
Proportion
A
4
Cummlative
6
Communality
BR
.8152
I
.5377
NA
.9814
B

8.9521

2.2999

1.3506

0.7291

6.6522

0.9493

0.6215

0.3075

0.6246

0.1605

0.0942

0.0509

0.6246

0.7850

0.8793

0.9302

CL
9594
K
9536
S04
7765
.8768
SI

CA
.9676
LI
.8355
RB
.9617
.7595
TDS

ALK .5646
MG
.8722
CS
.8864
.6761
:

9085
TOTAL=13.33
A) PRINCIPAL FACTORS
FACTOR 1_______FACTOR 2_______ FACTOR 3_______ FACTOR 4
CA
.9773
CL
.8135
ALK .5786
NA
.3694
K
.9694
TDS .7152
LI
.4831
S04 .3642
RB
.9626
NA
.5824
S04 .3321
BR
.3308
SR
.9308
MG
.3144
B
.3227
SI
.2474
MG
.8675
I
.1990
• I
.2198
LI
.2416
BR
.8359
CA
.0554
NA
.2004
CS
.1422
B
.8209
SR
.0502
BR
.0687
CL
.1355
CS
.7873
LI
-.0458
CL
.0416
RB
.0923
LI
.7359
BR
-.0474
CA
.0280
B
.0864
I
.6439
K
-.1117
RB
.0124
K
.0384
CL
.5268
RB
-.1623
K
.0001
TDS .0225
TDS .4023
S04 -.2856
SR
-.0047
CA
-.0932
SI
.3630
ALK -.2860
TDS -.0479
MG
-.1311
ALK -.2722
B
-.3020
CS
-.0512
I
-.1877
S04 -.6723
SI
-.3194
MG
-.0599
SR
-.1987
NA
-.6824
CS
-.4935
SI
-.6808
ALK -.2717

198

TABLE A-5 (Continued)
Case 4 (continued)
B) VARIMAX Rotation
FACTOR 1_______ FACTOR 2_______FACTOR 3_______ FACTOR 4
B
.9037
CL
.9546
NA
.7449
SI
.8107
LI
.8695
TDS .8119
S04 .6721
CS
.3237
RB
.8498
MG
.5947
ALK .0331
BR
.2588
K
.8158
I
.4009
LI
.0039
RB
.2435
BR
.8151
CA
.3943
CL
-.0177
K
.2263
CS
.8098
SR
.3662
BR
-.0889
MG
.1344
CA
.7348
NA
.3213
TDS -.0933
CA
.1269
SR
.6512
BR
.2753
SI
-.1961
SR
.1008
MG
.5167
K
.2446
B
-.2414
TDS .0493
I
.4687
LI
.2314
CS
-.3119
CL
.0342
CL
.2161
RB
.1984
I
-.3604
B
-.0415
SI
.2043
B
.0144
RB
-.3753
LI
-.1610
TDS .0759
SI
-.1453
K
-.4208
I
-.1659
ALK .0144
CS
-.1689
MG
-.4832
NA
-.2208
S04 -.1854
ALK -.3370
CA
-.5062
S04 -.2341
NA
-.5240
S04 -.4854
SR
-.5832
ALK -.6431
C) PROMAX Rotation
FACTOR 1
FACTOR 2
RB
.9478
CL
.9709
K
.9299
TDS .8183
B
MG
.9201
.6632
CA
.8833
.4870
CA
BR
.8619
I
.4632
LI
.8549
SR
.4544
CS
.8502
BR
.3573
SR
.8133
K
.3417
MG
.6976
LI
.3188
I
.5717
RB
.2971
CL
.3506
NA
.2356
SI
.2877
B
.1164
TDS .2129
CS
-.0755
ALK -.1036
SI
-.1226
S04 -.4126
ALK -.3788
NA
S04 -.5286
-.6382

FACTOR 3
NA
.8789
S04 .7738
ALK .2075
TDS -.2172
CL
-.2204
LI
-.3395
SI
-.4145
BR
-.4956
I
-.5172
B
-.5704
CS
-.6528
MG
-.7322
RB
-.7497
K
-.7777
CA
-.8158
SR
-.8411

FACTOR 4
SI
.8522
CS
.4534
RB
.3707
K
.3541
BR
.3406
CA
.2523
SR
.2292
MG
.2245
B
.0863
TDS .0326
CL
.0161
LI
-.0772
I
-.0833
S04 -.3174
NA
-.3861
ALK -.6186

199

Figure

A 7 . R - m o d e f a c t o r a n a l y s i s r e s u l t s , c a s e 1. O n l y
s a m p l e s w ith pH v a r i a b l e are used, pH v a r i a b l e
I n c l u d e d , TDS n ot p a r t i a l l e d out.

Ca, K, Rb, Sr
Mg.Br.B
Cs,U

Cs
b i ,a l k ,s o 4

Na,S04,Br
SI,U

SI
pH,S04lB,Na
Rb,K,U
Br

1

CI.Cs.Rb
K,pH,TDS,B

Rb,Mg
Cs

Mg.I.Sr

Sr.Ca

Na,S04
TDS,Cl

proportion

OP VAR IAHCE

9.5J

16. 51

59. 61

CASE

1 - PRINCIPAL

Figure A-7

COMPONENTS

4. 9Z

200

Cl
Br.Ca.TDS.Sr.K

201
CASE 1 - VARIMAX ROTATION

B
U .R b.K .B r.C s

CI.TDS
SI

Ca.Sr

-

-

-

-

Na

—

-------

-

Mg
Mg.i

S 04
C a.I.Sr

Cl,SI

“■
TDS

-

-

Br,K ,R b,N a

-

-

pH ,C l

-------

U ,B

C s.B r
l,R b

-

II,TDS
A LK .pH

-

so4

B r.A LK

Cs

C a.M g.S r

-------

-

SI,B

_

C s.S r.K .I

ALK

-

-

ALK

pH

1

1

_

4

3

Cl
TDS

Na

Mg

so4

SI

C a .S r
l,K ,B r,R b

Cl
S I,T D S

—

2

R b .K .B
C a .U .B r .S r
C s ,M g

-

Ca,Sr,M g

so„

Na

B,TDS,Cl
U ,S 0 4 ,l
Na.pH

Cs
U
—

-

B ,N a

pH

—

R b ,B i,K
_

_

C a .S r ,M g

'

ALK

Cs
ALK
pH
-

B

SI
_

T D S .C l
U ,l,p H ,S 0 4

C l,T D S

_

-

~

-

-

-

U
S04

Na

ALK

B r.I.S I
B

S 0 4 ,pH

M g ,C s ,R b

Na

ALK
K .C a
Sr

CASE

1 - PROMAX R O T A T I O N

Figure A-7

202

Figure

A-8 . R - m o d e f a c t o r a n a l y s i s r e s u l t s , c as e 2.
samples
are
used,
pH
variable
included,
partialled o u t .

All
TDS

3

2

1

4

Cl

Ca.K.Rb.Sr
C$,8,Mg,8r
U

ALK,U,pH

1

so4

Si
Cl

Mg.Na

B.l
Na

U

Cs
Mg

pH

-

LALK.K
Rb,S04,B
-

---

«

Cs.SI
pH

so4

203

Ca
Sr,Br,U

CI,Br,Ca,K
Rb.Sr
ALK

S04,Br,Na
SI
CI.Cs
Rb,B,pH,K
Ca
Mg.Sr.l

ALK

SI

Na
P R O P O S T IO N
OF V A R IA N C E

1 U 4 I

9 .8 1

01.**

CASE

2 -

P R IN CI PAL
Figure A-8

I

C O MP ON E N T S

204

CASE

2 - VAR1MAX ROTATION

1

2

3

B

Cl

SI

B b .U .K .C a
B f .C a
Sr

4

m

*

S04

*

““

“

—

-

Na
I.M g
pH
M g .N a .B r
SI

C t.B r
-

-

«.

-

_
K .R b
M g .C a
S r .C I

U
ALK

A L K ,p H

B ,S I

so4

-

K .B b

B

B t .C I
-

U
C s .S r

«

-

Cl

B. 1
C s .S O ^ .A L K

U ,N a

K .R b
C s .l

_

_

-

-

s o 4,i

-

PH

pH
Ca
Sr

Na

ALK

Mg
-

---------- -

1

-

-

2
S 0 4 .N a

-

-

3

B b .K .C a .B
C s .S r . B r . U

Mg

-

—

4
Cl

SI

- —

-

- —

-

1
SI

-

N a .M g
B r .U

Cs

PH
Cl

B r .R b .K
-

~

-

-

ALK

Mg
C a .S r

-

-

B .C I

U

A K .p H
-

-

_

Cl
S l.l

U .l
-

-

1,S O , , A L K
Cs
-

-

so4

so4

Ca
Sr
K .R b .B

p H ,S I

N a .p H

___ E L ! l _
R b .K
_

-

-

_

_

C a .S r

Mg

Na

CASE

_

ALK

2 - PROMAX R O T A T I O N
Figure A--8

_

-

205

F i g u r e A - 9 . R - m o d e f a c t o r a n a l y s i s r e s u l t s , c a s e 3.
s a m pl es are used, pH v a r i a b l e not i n c l u d e d ,
p a r t i a l l e d out.

All
TDS

C a .K .B b .s r
C »,B ,M g ,B r

AUCU .C I

A L K ,S 0 4,B

Br.Na.SI

N a ,S 0 4
C s.C l.R b.B

206

l,B
Br.Ca.Mg.Sr
K ,R D

p r o p o r t io n
of

65. 12

11.3:

C a .S r

Ca.Mg

9 .6 £

6.3£

v a r ia n c e

CASE 3 -

PRINCIPAL COMPONENTS
Figure A-9

207

CASE

1

3

-

VAR1MAX

2

ROTATION

3

e
R b .U .K .C *
B r.C a

4
Cl

SI

Sr
S r . M g .C a

________ i , M g _ _

---------

----------

K ,R b
C a ,l
SI
Cl
ALK

so4

B .S I

C n .B r

B r .C I
U

K .R b
M g .C a
S r.C I

ALK

B

■“

““

M g .N a . B r
C a ,K J ,S r

—

K ,R b ,t
B
C s .S O ^ .A L K

U .N a
S 0 4 ,l

----------

----------

Na

—

Na

ALK

s.

M

—

-

so 4

2

1

3

R b ,K ,B ,C a
C s ,B r ,U ,S r

S r ('C a ,M g

Mg

Rb
Cs

I

B ,B r

SI

1
S I ,U

4

SI

Cl

C s .R b .K .B r

Mg
N a ,B r , C a ,U

K

C a .M g .S r

Cl

Sr

Cl

ALK
ALK

B ,C I

K .R b
LB

U
1

C s .A L K

so4

£»

s

SI

1

1

I1

1

1

1

S O jj.tla

Na

-

-

---

ALK

-

S 0 4 ,N a

CASE

3 - PROMAX

R OTATION

Figure A - 9

- _

-

208

Figure

A-10.

R - m o d e f a c t o r a n a l y s i s r e s u l t s , c a s e A.
s a m p l e s are u s e d , p H v a r i a b l e n ot i n c l u d e d ,
not p a r t i a l l e d out.

All
TDS

Rb.K.B
Ca,U,Br,Sr
Cs.Mg
-

m

«M

2

3

Cl
TDS

Na
-

so4

-

SI
-

m

Mg
Ca.Sr

1

l,K,Br,Rb
U

pH

Cs
Rb.Br.K

Cl
SI,TDS

-

-

Ca.Sr,Mg
B,Na
Cs

B

Cl,TDS

TDS,Cl
U.l

SI
.

pH,S04

PH
U

-

so4

ALK

Na

S 04,pH

Br.I.Sl
B ____
Mg.Cs.Rb
K.Ca
Sr

— “

CASE 4 - PRINCIPAL COMPONENTS
Figure A-1Q

Na

ALK

209

ALK

ALK

210

CASE A - VAR1MAX ROTATION

1

2

3

4

Cl
TOS

SI
«z

U.Rb^Br.Cs
Ca

“

Cl,SI

l,Ca,Na

“

-

Br

-

--

ALK
LI

K,LI,Rb
t d s .a l k

B

so4

----

Na,S04
-

K,Mg

so4

Ca
Sr

Na

1

2

Rb.K.B,Cs
Ca,Br,U,Sr

-

B
U,1

—

ALK

Cs,Br,Rb,K
Mg,Ca.Sr
TDS,Cl

Br.TDS.CI
SI.B
Cs.I.Rb

SI.Cs
-

-------

1
1
1

1
1

_____ M g ____
1

I

Sr
Mg

ALK

3

4

Na

Cl,TDS

SI

so4

Mg,I
Mg
Cl,SI
-

TDS

-------

Ca.Sr,1

-

“

Rr.K
LI,fib
Na

Cs
-

--

-

Ll,l

SI

TDS.CI.LI

----

-

----

-

ALK

-

--

S 0 4,Na

Si.Bf
1

so4
Na

so4

--

Aik

B
Cs
Mg.Rb.K
Ca.Sr

CASE

-

B,TDS,Cl

ALK

B
Cs

ALK

nb,K,Br
Ca.Sr,Mg

A - PROHAX

Figure A-10

ROTATION

-

-

211

EFFECT OF pH VARIABLE
Figures A-7 and A-10 shows the effect of including pH
in the data set on factor makeup when TDS is partialled out.
No significant changes in the makeup of the 4 factors occurs
with

the

varimax

rotation

when

pH

is

not

included.

Likewise, no significant changes are observed in the promax
rotation, although slight changes in the loadings of Ca, Mg,
and

Sr

in

factor

2

occur

when

pH

is

removed.

In

both

rotations, pH loads in the factors near ALK and S04 .
Removing pH also does not change the results when TDS
is partialled out

(compare Figure A-7 and A-1 0) .

However,

factor 2 is found to invert upon removal of pH, although the
relative factor loadings of the variables remains constant.
Also,

including pH apparently reduces the loading of Ca, Mg,

and Sr

in factor 3,

but the relative positioning of these

elements remains constant.
significantly

alter

the

In sum,

factor

pH apparently does not

makeup

or

relative

factor

loadings.
EFFECT OF PARTIALLING OUT SALINITY
As discussed above, variations

in sample salinity may

mask identification of other processes affecting the brine
chemistry.
TDS

In order to remove the influence,

variable)

correlation

is

partialled

matrix

in

the

out

R-mode

by

using

brine salinity,

(except S04

and therefore,

a

calculations.

important to determine the effect of salinity,
because most elements

salinity

(the

partial
It

is

for example,

and Na) increase with the

variables may intercorrelate

212

with salinity in this respect.

The R-mode analysis results

preformed with and without TDS partialled out,

are compared

A -8 and A-9, respectively.

in Figures A-7 and A-10, and

The most apparent change resulting from partially

out

of salinity is found in the relative makeup and importance
of factors 2, 3, and 4:

Factor______________ 1__________2__________ 3__________ 4
% variance with
59.6
TDS not partialled out

16.5

9.5

4.9

% variance with
TDS partialled out

11.4

9.8

6.0

61.4

While the amount of variance explained by each

factor

does not dramatically change the partialling out of salinity
affects

the

factors

switch

TDS.

Factors

make-up

of

factors

2 to

in

their

importance

that

might

represent

4.

In

upon

essence,

partially

water-rock

the

out

of

interactions

(factor 2 and 3) increase in their relative importance over
a factor that might represent salinity (factor 4)

.

EFFECT OF ROTATION METHOD
Both
study.

varimax

and

promax

rotations

Varimax is an orthogonal

an oblique rotation method.

were

rotation,

used

in

this

while promax

In the varimax solution,

is

factor

axes are kept orthogonal during rotation until the best fit
is obtained,

maximizing

the

variance

each

factor

explains

213

(Hitchon et a l ., 1971).

Promax differs in that the axes are

not constrained to be orthogonal.

Comparing the results of

the cases demonstrates that the rotation method chosen,

for

this

the

example,

factors.

generally

However,

separated

using

result was

does

the

the

not

alter

variables

promax

the makeup

appear

oblique

to

be

rotation.

found by Long et a l . (1986)

A

of

better
similar

and Hitchon et a l .

(1971) .
A problem with oblique rotation methods is the possible
correlation

between

uncorrelated

factors,

which

1986).

When

(Davis,

by

design,

factors

correlate,

relationships between the variables and the
highly

complex,

remain hidden.
factor

and

the

true

occurred

using

or if pH was included).
clear,

be
the

factors may be

controlling

processes

may

A correlation between the first and second
the

regardless of the test case

not

should

but

the

promax

rotation

(r =

-.53),

(if salinity was partialled out

The meaning of this correlation is

observation

that

both

orthogonal

and

oblique rotations produce nearly identical factors suggests
that factor correlation does not effect factor makeup.
INTERPRETATION OP R-MODE RESULTS
The

varimax

and

promax

rotations

with

salinity

partialled out and pH not included (case 2, Figure A-9) best
illustrate

the

factor

interpretations
Factor

loadings

can

be

analysis
made

between

results,
from

— 0.25

each
were

although

similar

case

presented.

not

considered

significant when making the interpretations.

214

Factor

1

is

characterized

loadings on B, Rb,
highly

negative

common

source

elements,
It

is

other

loading

of

factor,

but

seawater

Na.

eg.,

It

a

very

Na

loads

is

this

may

source

as

for

a

the

such as Rb and B.

strongly

reflect

evaporation.

positive

interpreted

seawater

so

high

Sr, Mg, and I, and a

for minor elements

clear why

elements,

during

Li, K, Cs, Br, Ca,

especially

not

by

inverse

the

to

behavior

Alternatively,

the

of

the
Na

inverse

relationship between Na and Cl in this factor may represent
halite equilibrium.
This

first

factor

Hitchon et a l . (1971)
loadings

on minor

earth elements.
salinity

factor

as

to

the

first

factor

(Figure A-5) , as both have high
alkaline

metals,

and

of
(+)

alkaline

The negative loading of Na in factor 1 with

Egleson and Querio

mode

similar

elements,

partialled

minor elements

is

out

is

(1969)

also

similar

interpreted

to

the

each

high

study.

loading

of

(B, K, Li, I, B r , and N H 3) in their first Rshowing

a common mechanism affecting

these

elements.
Factor 2 is characterized by high
Mg,

and Ca,

and

high

negative

interpreted
perhaps

intermediate

as

also

characteristic

loadings

representing
indicating
of

this

different trials studied,
S04 .

(+)

(+)

loadings on Sr,

loadings on K, Rb,
Na
a

CaS04
factor,

and

S04 .

carbonate
mineral
found

This

I,

and Cs,

factor

mineral

factor,

equilibrium.
in

each

is

of

A
the

is the strong co-loading of Na and

A possible explanation is that this loading represents

215

the decrease

in Na

in evapo-concentrating seawater

halite precipitation)
samples

contain

are possible

and the fact that more highly saline

little or

include:

(1)

no S04 . Other explanations

that

CaS04 and NaCl equilibrium,

both

common minerals in the basin,
or
al.
and

(3)

(2) NaS04 mineral dissolution,

strong Na~S04 ion pairing

(1971).
the

(due to

suggested by Hitchon et

NaS04 minerals have not been reported in basin,
ion-pairing

theoretical

explanation

considerations.

is

Pitzer

not
(1973)

supported
reports

by
that

single ion electrolyte interaction parameters between Na and
S04

are

quite

small,

thus

ion-pairing

between

these

components is not expected to be significant.
Factor

3

is

single ion, Si,
B r , and
factor

a

might

Carpenter

(-)

represent

(+) loading

on

a

loading

on

alkalinity.

silicate-carbonate

This

buffering

or

in essence, the balance between aluminosilicate

carbonate

represent

high

intermediate (+) loadings on Cs, Rb, K, and

significant

equilibrium,
and

characterized by

mineral

the

reactions.

illite-carbonate

(1978):

2K+

+

CaC03 +

Such

a processes

reactions

might

suggested

3Al2Si20 5 (OH)4 +

by

4Si02

Ca2+ + 2KA12 (AlSi3 )O10 (OH) 2 + H 20 + C02 alkalinity) .

=

This

is further supported by the positive loadings of K and Rb in
this factor, particularly in the promax rotation.
Factor

4

loading on C l ,
Li,

Ca,

loading.

and Sr,
This

is

characterized by

a

single high

positive

and intermediate (+)loadings on Na, Mg,

Br,

and no variable showing a strong negative
is interpreted to be a salinity

factor,

or

216

alternatively, because a similar factor is found when TDS is
partialled out,

it might represent that Cl

is the dominant

anion in the brine samples.
DISCUSSION
The

R-mode

results

appear

interpretations made using graphical
techniques

in

Chapter

1

and

consistent

with

and chemical modeling

Chapter

2.

Most

important,

factor analysis supports a seawater origin for the Michigan
basin brine, as a common source is suggested for many of the
elements, especially the minor elements.
because once the minor elements,
enriched

by

evaporation,

significantly reduce

This is reasonable

such as Br,

very

few

B and Rb,

reactions

their concentrations.

Q-mode

suggest that this characteristic, and origin,
throughout

the

entire

Furthermore,

the

differences

between

(Berea,

Traverse,

Student's
the
and

River-Niagara/Salina
represent

the

Michigan

upper
Dundee)

basin

t-test

will
results

reflected

sample

population.

results

suggesting

Devonian

formation

and

Richfield-Detroit

the

samples, might

different

is

are

degrees

be
of

in

waters

interpreted

to

evapo-concentration

reached by these waters in the basin.
The importance of carbonate reactions

(Chapter

1)

may

be supported by the combined loading of Ca-Mg-Sr in a single
factor.

The third factor which is interpreted to represent

silicate-alkalinity

buffering,

tenuous,

consistent

is

also

affecting K (Chapter 1).

although
with

somewhat

proposed

more

reactions

217

One

notable

difference

lack of a halogen

from

previous

factor where Br,I , and Cl

(Hitchon et al., 1971; Egleson and Querio,
1969).

studies

is

the

load together

1969; and Kramer,

It is not apparent why this factor is not found, but

it may reflect that the I enrichment occurs by processes not
related to evapo-concentration, such as from organic matter
decomposition (Egleson and Querio,

1969).

CONCLUSIONS
(1)
Michigan

Q-mode

factor

basin

brine

homogeneous

analysis
data

population,

demonstrates

are

described

interpreted as a salinity gradient.

that

the

characteristic

of

by

factor

a

single

a

R-mode factor analysis

of the entire data set is therefore, justified.
(2)

R-mode

factor

analysis

was

orthogonal and

promax oblique rotation

trial

determine,

cases to

including

pH

resulting
similar

and

factor

the

interpretations

was

out

what
of

varimax

the effect

salinity

evaluated.

can be made

using

methods. Different

for example,

partialling

makeup,

run

In

from all

on

of
the

generally,

cases.

Four

factors were found to account for 90 to 92% of the variance
in

the

data.

Communal it ies

were

high

(>.75)

for

most

variables.
(3) The factors are interpreted as follows:
Factor

1:

minor elements,

common

source for

suggested to be

most

elements including

a seawater source.

Factor 2: A carbonate-sulfate mineral factor.
Factor 3: Silicate-carbonate buffering or equilibrium.

218

Factor 4: A salinity
(4)

The

interpretations
results,

and

multivariate

made

suggest

rock reactions.

or dominant anion (Cl) factor.

from

statistical

graphical

and

methods

chemical

support

modeling

a seawater origin modified by water-

the

APPENDIX B

APPENDIX B

METHODS
BRINE SAMPLING
The

brine

samples

directly

from

several

Niagara/Salina

distribution
each site,

the

used

lines

well

or

in

head

this

whenever

samples

the

bottom

stand and

were

collected

possible,

although

were
of

collected

separator

from

tanks.

At

sampling valves at the well head were opened and

the brine-hydrocarbon mixture was
pre-rinsed,

study

plastic
separate

carboy.
for

The

collected
mixture

several minutes,

brine was withdrawn through the bottom

in a
was

5 gallon

allowed

after which
spigot.

to

clean

The brine

was passed through glass-fiber w o o l , filtered through Watnum
#1 filter paper,

and collected

in a plastic beaker.

Some

samples would not separate in-field because of hydrocarbon
viscosity and in these instances,
brine

mix

was

collected

in

the combined hydrocarbon-

sealed

plastic

bottles

and

allowed to separate in the laboratory over several days to
several weeks.
the bottle and

The brine was withdrawn from the bottom of
filtered in the same manner as above.

The

following splits were taken from each sample:
1) 25ml were pipeted in to a capped widemouth plastic
bottle for pH, Eh, and alkalinity analysis.
2) 125ml

were

carefully

pipeted

into

plastic

containing 125ml of 5% HN03 acid for cation analysis.

219

bottles

220

3)

125ml

were

preserved

with

1ml

of

reagent

grade

formaldehyde for S04 analysis.
4) 500ml were collected in boro-silicate glass bottles
with teflon sealed caps for stable isotope analysis.
it

was

necessary

to

open

these

bottles

to

Often
release

accumulating gas pressure.
5)

500ml

analysis.
case

for

observed

were

collected

for Cl

and

If salt precipitation was
Niagara/Salina
in

the

replaced by

brines,

filtered

diluting

125ml

or

brine,
of

strontium

expected,
if

salt

then

brine

as was

crystals

this

with

isotope
the
were

sample

12 5ml

of

was

double

distilled H 20.
The

field

dilutions

were

made

using

a

single

100ml and 25ml Class A glass volumetric pipets.

set

of

All samples

collected for chemical analysis were stored in high density
polyethelyne

bottles.

The

samples

were

labeled

and

transported to the laboratory in a covered plastic pail that
was iced on particularly warm days.

FIELD ANALYSIS
Field analysis consisted of measuring temperature,
Eh,

and

when

measured

possible,

using

a

alkalinity.

standard

Temperature

laboratory

pH was measured using an Orion model
Orion

91-05

Standardization
reference

gel
of

standards

temperature

of

the

filled
pH

electrode

of

pH

4

was

and

was

thermometer.

407 pH meter with an

combination

the

brine

mercury

(°C)

pH,

7.

normally

was

pH
made

Because
between

electrode.
using
well
9 to

NBS
head
10°C,

221

the pH standards were chilled on ice during the summer field
work.

Electrodes

detergent.

were

An

cleaned

Orion

model

between

use

with Alconox

96-78 combination

platinum

electrode was used for measuring Eh, which was normally done
concurrent with pH.

Alkalinity was measured following the

procedure in Brown et al.
with

H 2S04

standardized

buret.

(1970) by titrating 25.0ml samples
acid

delivered

from a

50ml

glass

The acid was standardized daily using a standardized

HC03 solution.
DISCUSSION
There
between
Most

are

several

sampling

brine

importantly

in

and

is

the

temperature that occurs
brine

aspects

Michigan

normal

differ

considerably

salinity

groundwater.

potential change

during

is

that

brine

produced

in

pressure-

production.
from depths

Because

of

several

thousand feet of depth, brine undergoes considerable changes
in

temperature

changes

result

precipitation

and
in

of

pressure
the

salts

during

exsolution
and

production.

of

paraffin in

gases
the

These

and in
well

the

bore,

a

common problem faced during Michigan oil production.

There

is no

these

method

for

surface

sampling

that

will

reduce

changes, short of using down-hole pressurized samplers.
chemical

changes

temperature changes

in

brines

that

are difficult

well understood at this time.

result

from

to evaluate

The

pressure-

and

are

not

Several

aspects

of

the

deserve

some

measurements

also

measurements

involve

the

use

of

pH

Eh,

and

discussion,

alkalinity

because

electrodes.

these

Sources

of

error that become significant in using electrodes to measure
high salinity waters include the changes in liquid junction
potentials

caused

composition

of

by

the

differences

standards

specificity of electrodes.

and

in

samples,

salinity
and

the

and
non­

Liquid junction potential errors

could be minimized if standards had compositions that match
the brines,

which is not possible for highly saline brines

such as the Michigan basin brines.
standards

suggested

by

For example, the TRISH+

Millero

(1979)

for

seawater have a salinity some 9 to 10 times
average Michigan

brines.

The

commercial

measuring

less than the

standards

used

in

this study were dilute solutions, therefore, liquid junction
error may have been a significant error in this study.
electrodes may also respond to monovalent

ions

and Li when present at high concentrations
Similar

problems

affect

electrodes

responding

Kharaka

a l .,

were

not

et

stable

to

1980) .
over

Eh
Fe

measurements

Often the

short

and

pH

long

(Levine,

as

or S species
and

such

as Na
1978).

w e l l , with

Eh

(Langmuir, 1971;
Eh measurements

periods

of

time,

could be arbitrary adjusted by slowly stirring the
Considering these problems,

pH

or

sample.

the pH and Eh measurements are

probably not representative of true solution conditions, and
the true pH-Eh values of the brines at subsurface conditions
remain an unknown at this time.

223

It was also observed during the alkalinity titrations
that the pH would not respond to the addition of acid,
upon the addition of very large amounts of acid.

even

This was

observed most often in the more highly concentrated Detroit
River,
This

Richfield,

pH

response

and
may

phenomenon

described

measurement

may

acids present
1975).

Niagara/Salina
be

due,

in

above,

formation

part,

however,

have

been

affected

by

in the

brine

( Case,

1945;

Attempts

at

to

measuring

the

samples.

the

the

electrode
alkalinity

dissolved
Willey

dissolved

organic
et

a l .,

organic

acid

content of these brines using liquid-ion-chromatography were
not successful due to salinity interferences
personal communication).

However,

For example,

preformed

on

titration

a

near

carbonate

is

thought

buffered,

and

a

pH

of 3.5,

not

to

reflect the

near

the

a

non-saline

inflection points are observed

observed in the ground water sample.
PH

samples,

Assuming that the pH measurements

accurate in the brines,
brines

and Niagara/Salina

typical

groundwater sample.

these

in the brines was

Figure B-l shows alkalinity titrations

Richfield

of

Fischer,

evidence suggesting that

dissolved organic acids might be present
found.

(B.

pH

of

are
for
4.5

An inflection at this

presence

of

acetate,

a

principal organic acid in brines (Willey et a l ., 1975).

ANALYTIC METHODS
Cations

were

measured using

emission spectroscopy methods,
instrument.

Anion

atomic

adsorption

and

on a Perkin-Elmer 506 A.A.S.

analysis

was

by

titrimetric,

224

pH

I

2 3 4 5 6
7 6 9
ml of a c i d

F i g u r e B - l . Alk al in i t y ti tr at i o n curves, pH v e . v o l u m e of
added (ml).
Curve A, Ri c hf i e l d brine s am pl e,
curve B, a N i a g a r a / S a l i n a f o r m a ti o n b ri n e ,
curve C, a n e a r - s u r f a c e ground water from
Michigan .

ecid

225

colorimetric,

and

gravimetric

methods.

Individual

procedures and their references are listed in Table B-l, and
selected methods are discussed in detail here.
Sample Preparation
Suitable working dilutions were prepared based on the
concentrations of species of interest.
and K analysis,
pipeted

1ml

For Ca, Mg, Sr, Na,

a 1:2000 dilution was prepared by carefully

of

the

volumetric flask,

50%

field

dilution

diluted

a

1000ml

adding 1ml of concentrated HN03 , and then

bringing the volume up with de-ionized,
field

into

samples

were

hydrocarbons, aliquots
of

found

distilled water.
to

contain

If

visible

of non-diluted brine were withdrawn

from the

bottom

the

sample

filtered,

and allowed to sit.

bottles

in

the

laboratory,

This procedure was

repeated

until clean brine was obtained,

which was then handled in a

similar

described abo v e .

fashion to the

Class A pipet

samples

(1.0ml)

and volumetric

flask

A

(11)

single

were used

for all samples in order to minimize error between samples
caused

by

glassware.

introduces

Although

considerable

such

error

in

a

large

the

analysis

discussion below), this dilution was necessary
bring

concentrations

minimize

matrix

analysis.

All

polyethylene
followed

for

and

within
other

diluted

plastic
other

measurable

bottles.
dilutions

were

stored

Similar
in

in

this

(see

in order to

ranges

interferences

samples

dilution

and

the

in

A.A . S .

pre-rinsed

procedures
study.

to

were

Working

standards for most components were prepared from commercial

226

TABLE B-l
Components measured and analytic methods.
COMPONENT
PH
Ca, Mg, Sr
Na, K
Rb, Cs
Li.

Cl
Br
I
B
Si
n h 4n

S04
Alkalinity
Density

TDS

METHOD
electrometric
flame emmision
w/1:10 of 87g/l
LaCl ^
flame emmision
w/1:10 of 25.4g/l
NaCl or KCl
flame emmision
w/1:10 of 25.4g/l
Na-K-Cl
flame emmision
w/1:10 of 25.4g/l
Na-K-Cl
Mohr titration
colorimetric
bromide oxidation
colorimetric
w/carminic acid
colorimetric
following extraction
potentiometric titration
gravimetric
potentiometric titration
pyconometer @2 5°C
and by calculation
SPG=log TDS * 7.102x10 -7
calculation

DILUTION
none
1:2000

REFERENCE
1
1

1:2000

1:26
1:200

none
1 :40
1 :2
1 :10

1
2

1
1

1:200

1:2

none
none
none

4
5
1
5

+ 0.996

References:
1: Brown et al.,
(1979) Methods for Determination of
Inorganic Substances in Water and Fluvial Sediments.
U.S.G.S. Water Resources Investigation Book 5, Chapter
A-l, Washington, D.C.
2: Presely, B.J. (1971), Part I: Determination of selected
minor and major inorganic constituents, in: Ewing, J.I.
and others, Initial Report of the Deep Sea Drilling
Project, Volume VII, Part 2: Washington, D.C., U.S.
Govt. Printing Office, p. 1749-1755.
3: Schrink, D.R., (1965) Determination of silica in sea
water using solvent extraction, Anal. Chem., 37,
764-765.
4: Collins, A.G., Cassaggno, J.L., and Macy, V.W. (1969)
Potentiometric determination of ammonium nitrogen in
oil-field brines, Environmental Science and Technology,
3, 274-275.
5: A.P.I. (1968) Recommended Practice for Analysis of
Oil-field Waters, American Petroleum Institute,
Washington, D.C., 2nd edition, 58p..

227

1000mg/l stock standards.
Br,

and

Stock standards of Rb, Cs,

B, I,

S04 , were prepared from reagent grade salts

that

were oven-dried overnight and cooled in a desiccator before
weighing.

All standards were made in reagent grade Na-K-Cl

solutions

having

Cl

concentrations

similar

to

the

diluted

sample.

Ca, Mg, and Sr
The alkaline earth elements were measured on a single
aliquot

of

the

ionization,

1ml

sample

and

analysis,

1:2000
of

87g/l

standard.

following

dilution.

In

LaCl2 was

Flame

added

emission

Brown et al.

order
to

was

to

suppress
10ml

each
used

for

of
the

(1970).

Na, and K
The

alkali

dilution.
25.4

metals

were

also

In order to reduce

measured

on

the

ionization effects,

1:2000
1 ml

of

g/1 NaCl or KCl was added to each 10ml

of sample and

standard for K and Na analysis, respectively.

A red filter

was

used

for K analysis,

analysis, following

and

flame

Brown et al.

emission was

used

for

(1970).

Rb, Cs, and Li
These metals were particularly difficult to analyze due
to their low concentrations compared with the alkali metals.
For Rb

and

Cs,

3ml

of

the

50% field

diluted

diluted with 10ml of distilled water, while

sample

were

Li was measured

228

in a 1:100 dilution of the
was

not

added

to

these

field diluted

samples,

but

sample.

was

Na-K-Cl

added

to

the

standards by adding 1 to 2ml of 25.4 g/1 Na-K-Cl solution to
each

10ml

against

of

standard.

distilled

measurable Rb,

The

water

Cs,

Na-K-Cl

and

was

solution

found

checked

to

contain

not

or Li at the dilutions used.

was by flame emission using a red filter,
et al.

was

Analysis

following

Brown

Schrink

(1965)

(1970).

Silica
Silica was measured by the method
and

Collins

(1975).

mixed ammonium

in

The method involves adding

molybdate

sulfuric

acid

52g/l ammonium-molybdate mixed with

1ml of a

solution

50ml of

(100ml

of

1 molar H2S04 )

to 50ml of 1:100 dilution of the field diluted sample.

The

mixture is stirred well and allowed to stand for 20 minutes.
15ml

of

1:1

H 2S0 4 acid

is

then

added

and

the

mixture

cooled to room temperature for several minutes on ice.
100ml

pre-rinsed

separatory

funnel

(pre-rinse

and

the

sample-reagent

In a
ethyl-

1 0 .0ml of ethyl-acetate

acetate and wash dox^n with acetone)
is pipeted

with

is

mixture

is

added.

The

funnel is shaken vigorously for 1 minute, and the mixture is
allowed

to

separated.

The

ester phase

is

measured

at

1 minute

or

aqueous

phase

then

collected

in a

separate

a

for

wavelength

of

is

until

cuvette
335nm

all

aster

discarded
and

using

its
a

and

has
the

absorbance

B&L

Spec-20

229

spectrophotometer.

Standards were prepared using

25.6

g/1

(lml/100ml) Na-K-Cl solution and treated in the same manner.

Boron
Boron was measured using the
by

Lico et al.

using

(1982).

recrystallized

distilled water).

carmine method

outlined

Stock B standard was first prepared
Na2B40*10H20

(place

overnight

in

0.2ml of the 50% field dilution is mixed

with 2ml of distilled H 20 and 2 drops of concentrated H C l ,
to which 10ml of concentrated H 2S04 was slowly add e d .

The

mixture is iced for 30 minutes to c o o l . Then 10ml of carmine
solution

(0.5g carmine dissolved

H 2S04 ) is added,

mixed

and allowed to

stand

absorbance

blanks,

measured

of

at

a

in 1000ml

very well

for

of concentrated

using a vibrating mixer,

1 hour mixing occasionally.
standards,

wavelength

of

and the

600nm

The

samples

using

a

B&L

were
Spec-20

spectrophotometer.

n h 4n

The
procedure

NH4N
in

analysis

followed

Collins (1975).

an

adaptation

10.0ml of

the

of

50%

the
field

dilution is pipeted into a 250ml flask, brought up to 100ml
with

distilled

minutes.

H 20,

The mixture

and

boiled

a

hot

plate

for

5

is cooled on ice and the pH adjusted

to 7.00 with

IN and 0. IN NaOH.

formaldehyde

solution

mixed well,

on

is

then

5.0ml of reagent grade
pipeted

into

the

37

solution,

and the mixture heated to 40°C on a hot plate.

230

The mixture is quickly cooled to room temperature using icewater

bath,

and

titrated

8.6

to pH

using

0.02N

NaOH.

A

blank was carried along with each batch of samples.

Chloride
Chloride

was

measured

on

a

using the Mohr titration method

50ul

non-diluted

(Brown et al.,

sample

1970).

This

method involved the addition of 50ml of distilled water to
several drops of K 2Cr04 reagent, and titration

the sample,
with

standardize

prepared

from

silver-nitrate

Na

A.A.S

solution.

standard

that

was

Standards

were

check with

NaCl

salt standard.

Bromide
Bromide was
Presely (1971)
For

this

measured using

as modified by

analysis,

50ul

the

procedure

Long and Gudramovics

of

acidified

50%

field

sample was added to 950ul of distilled water.
of phenol red-acetate buffer
.IN

NaOH

brought

up

to

outlined

by

(1982).
diluted

To this,

2 5ml

(0.016g phenol red plus 2ml of

100ml

mixed 1:4 with acetate buffer

with

distilled

water,

then

(18.lg anhydrous Na-Acetate,

7ml glacial acetic acid brought up to 1000ml with distilled
H 20 ) )

and

4ml

of

adjustable repipets.

Chloromine-T

(. 56g/l)

are

added

using

After mixing for exactly 30 seconds,

10ml of Na-thiosulfate is added (0.955g/1000ml anhydrous Nathiosulfate), mixed,

and

the

absorbance

measured

using a B&L Spec-20 student grade spectrophotometer.

at

592nm
Trials

231

showed that HN03 did not affect the development of
but

the

method

However,

the

is

sensitive

to

the

presence

of

color,
iodine.

I concentration of these waters are generally

less than 5% of the Br concentration, with in the estimated
error

of

the

analysis.

grade

K B r , and

along

Standards
with

a

were

blank,

made

were

from

reagent

measured

in

an

identical fashion.

Iodine
Iodine was measured using an adaptation of the method
in

Brown

et

overnight

al.

with

(1970).

lg

of

Non-acidified

CaO

to

remove

sample

iron.

was

The

treated

sample

is

filtered and 25.0ml are added to 75ml of distilled H 20.

The

pH is adjusted to 6.2 with 1M H 2S04 and/or 0.1M NH40H.
an

iodine

flask,

glacial

acetic

mixed,

and

added until

acid,

the
6ml

Approximately

15ml

of
and

flask
of

273g/l

Na-acetate,

5ml

of

is

stoppered

Na-formate

saturated

5ml
Br2

for

solution

the yellow color disappears,

of

In
2. 2M

water

5

are

minutes.

(50g/100ml)

the

is

Br vapors

in

the flask are blown out with air, and the sides washed down
with distilled H 20.

Then lg of KI and 3ml of 3.6M H 2S04 are

added, the solution is well stirred,

tightly stoppered,

allowed

5

minutes,

to
1

stand
to

2ml

in
of

the

dark

starch

for

indicator

minutes.
is

added

and

After
and

5

the

solution titrated with standardized 0.01N Na2S 20 3 until the
endpoint

is

procedure.

reached.

A

blank

is

carried

through

this

232

Sulfate
Sulfate

was

outlined in
used,

measured

A.P.I

using

(1968).

to which 50ml

the

gravimetric

50ml of non-diluted sample were

of distilled water were added,

concentrated H C 1 , and 10ml of BaCl solution
mixture
hours,
is

is loosely covered and heated to

through

Watman

#42

filter

30°C

paper

of
The

for several
The mixture

which

is

then

dried for

1 hour at

30°C,

crucible,

and

at

minutes.

The crucibles were cooled in a desiccator and re­

weighed.

A blank and a series of standards were measured in

a similar

ashed

fashion,

placed

1ml

(100g/l) .

cooled and allowed to stand overnight.

washed

procedure

850°C

in a pre-weighed ceramic

in

a muffle

furnace

for

10

and a calibration curve was prepared.

Two calibration curves were required,
one for higher concentrations.

one below 25mg/l, and

The results reported in this

study are calculated from the calibration curve.

TDS and Density
TDS

values

components.

were

calculated

by

summing

This procedure was used because

measured

it was

found

very difficult to get repeatable weights by evaporation and
weighing

salts

hydroscopic
determined

directly,

nature
on

25°C.

The

before

cooling

25

of

most

the

samples

brine

25°C

in

warmed
a

the

result

components.

by weighing

pyconometer was
to

likely

in

the

Density

was

a pyconometer

slightly

water

of

bath.

to

at

remove

gas

Density

was

233

calculated

for

the

remaining

samples

from

the

least squares line relating TDS to density:
TDS

*

7.102xl0"7

relationship,

+

brine

0.996

(r2

analyses

=

0.9)

obtained

were put on the same density scale.

best-fit

Density = log
.

Using

from

other

this

sources

Adjustment to formation

temperature and pressure can be made using:
p* = p * exp [B* (P—0.1)
(kg/m3)

at

density at

temperature
25oC and

T

- A* (T-25) ] where p* = density

oC

1 MPa,

and

pressure

P

(MPa) , p

=

A = 5xl0-4 oC- 1 , B = 4.3xl0-4

MPa- 1 .

Oxygen and Hydrogen Isotopes
The oxygen (180/160) and hydrogen (D/H) stable isotopic
ratios

were

Laboratory,
Several

measured
University

18/ 16o

(Princeton,

at
of

analyses

New

the

Waterloo,

were

Jersey),

Environmental

made

and

Waterloo,

at Teledyne

were

checked

replicate samples measured at Waterloo.

Isotope
Ontario.

Laboratory,

with

Identical

several
results

within, analytic error and variance, were obtained from each
laboratory.

Oxygen

isotope

analysis

utilized

72

hour

equilibration with C02 , and deuterium analysis used complete
distillation

techniques.

detail

Fritz

in

et

al.

All

methods

(1986).

are

discussed

Measured

ratios

in
are

normalized to SMOW and have a precision of + 2°/oo for D and
+ 0.2°/oo for oxygen.

However,

repeat analysis of samples

show larger ranges of measured values,
lsO and 3.9°/oo for

D:

up to

1.2 2°/oo

for

234

TABLE B-2
Variability in 180/160 and D/H

SAMPLE
3081
3082
1085
1086
2092
2097
2100
8040
3042
2020
1004
10043
3007
3082
3086

-6.19,-5.86
-3.50,-4.77
-2.58,-2.16
-5.81,-5.02
-3.39
0.45
0.23,0.37
-2.90,-3.03
-2.03
0.75
-0.60,—0.80
-0.93,-1.17
-4.51
-3.50,-4.77
-5.81,-5.02

D/H
-52.7
-46.5,-46.1
-43.6
-55.8
-48.2,-51.4
-38.6,-41.4
-40.8
-46.5
-40.5,-37.9
-37.4,-39.9
-31.5
-32.9,-29.0
-53.9,-50.1
-46.5,-46.1
-55.8

Deuterium values were corrected for activity following
Sofer and Gat (1975) :
del aD = del CD - [6.1MCa + 5.lMMg + 2.4MK + 0.4MN a ]
and

oxygen

isotope

activities

were

corrected

to

concentration using:
del cl80 = del al80 - l . H M Mg + 0.47MCa
Sofer and Gat (1972).

- 0.16MR , from

235

Strontium Isotopes
Strontium
using the

(87Sr/86Sr)

procedures

isotopic

Throughout this procedure, distilled HC1 was used.

lml of

non-diluted

column

containing

first

Stueber

loaded

100ml of Dowex resin

onto

et

measured
(1984).

is

by

were
al.

sample

outlined

ratios

a

(8x-200).

large

The Sr was then

eluted from the column using standardized 2N distilled H C 1 .
A calibration curve prepared for this column showed that Sr
was

separated

1150ml to
eluting

from

1400ml
a

Ca

and

of acid.

Ca-Sr

eluted

after

passing

The column was

standard

(having

a

through

standardized by

similar

absolute

concentration to typical samples), collecting the eluent in
25ml

sub-samples,

methods.
new,

and

analyzed

acid

washed,

scratch-free

overnight

beaker was

loosely covered and

on

a

periodically with H C l .
2N

HCl,

completely,

making

and

is

eluted

from

slightly

using

A.A.S

its

beaker,
warm

and

hot

slowly

plate.

The

sides were washed

down

The residue is then taken up in 5ml
certain

loaded

(8X-200, 50ml volume)
Sr

Ca-Sr

The 1150-1400ml eluent volume was collected in a

evaporated

of

for

onto

to

dissolve

the

a small column of

residue

Dowex

resin

along with 10ml of distilled H 20.
this

standardized 2N H C l .

column

using

450

to

600ml

The
of

The eluent is collected in an acid-

washed, scratch free beaker, and a 5ml aliquot of the eluent
is taken and analyzed for Sr and Ca.
was

dried

slowly

on

a

hotplate

The remaining eluent

overnight,

periodically

236
washing down
cooled,

the

the

beaker

residue

aliquot containing

sides

with acid.

is taken up
10-15ug of

in 10ml

Sr

Once

dried

of 2N HCl

is transferred,

and

and an

based

on

the measured Sr in the eluent, into a 10ml acid washed glass
beaker,
the

that

beaker

parafilm
o7

Sr/

o fi

is slowly evaporated on a hot plate,

sides

for

with

transport

HCl.

The

sample

to Argonne

was

National

washing

covered

with

Laboratory

for

,

Sr analysis.

samples

with

Both

50ml

of

columns

are

concentrated

regenerated

distilled

between

HCl

washed

through with at least 5 column volumes of distilled water.
Each
water

column

is

after

periodically

several

back-flushed

samples.

At

the

with

deionized

laboratory,

the

Sr

residue is taken up in 50ul of distilled HCl and loaded onto
the tungsten-rhenium filament and loaded into a VG-Isotopes
Inc.

MM54R

instrument

90°

arc

mass

spectrometer

is calibrated to N.B.S.

for

analysis.

standard

#987

The

having

a

87Sr/86Sr of 0.710201 ± 0.040/mil

Subsurface Temperature and Pressure
Subsurface temperatures
measured

in

the

Michigan

conditions

were

Vugranovich

(1986).

used
33m.

in this

wells

temperatures

A

study,

Geophysical

sampled

estimated

logs were

for a few

basin,

are very

therefore,

using

geothermal

starting

and

and pressures

methods
gradient

subsurface
outlined

of

at a temperature
obtained

indicated
locations.

These

in

2 3°C/km

is

of

at

for a number

approximate

seldom

10°C
of

the

bottom-hole

temperatures

are

237

suspect however,
this data

as there are no means for determining from

if the geophysical probe had equilibrated before

temperature

was

recorded

(Vugranovich,

personal

communications).
Variations in measured temperature,
equilibration

or

real

variations

due either to non­

in

basin

gradients are illustrated by Figure B - 2 .

temperature

This figure shows

formation temperatures that Vugranovich (1986) thought to be
reliable, and are plotted versus production d e p t h .

As this

figure shows, measured temperatures vary considerably around
the assumed geothermal gradient.
Subsurface

pressures

hydrostatic

( Wilson

Changes

subsurface

in

in

the

and Long,

basin

1985;

pressures

are

are

thought

to

Vugranovich,
therefore,

be

1986).

dependent

upon brine salinity as a function of formation depth in the
basin.
individual

The

characteristic

formation

brine

of

the

basin

compositions,

and

shape,

the

location

of

production within unique areas of the basin result in a poor
correlation between salinity,

brine density,

and production

elevation

least

best-fit

or

depth.

The

squares

relating brine density to production elevation is:

line

density

= 1.171 -4.264xl0-4 * (production elevation in met e r s ) .
poor correlation

(r2 = -0.367)

of this

line

suggests

The
that

predicting pressure versus depth in the basin based on brine
salinity or density would not provide useful information.

TEMPERATURE C
25

30

•

5

Q j T

DEPTH

a^
a aa

50

55

■

^

®

-■

-

A

0A*a!NcPb*'*d?A0 ~
a

“a

^

A

DUNDEE

RICHFIELD

A

A

q

O

" ' v°

0

X

°

A

^ ° 0
O

AA

A

NIAGARA/
SALINA
ORDOVICIAN

o
'0% 0
*o''^
*

&

2000

-

2500 -

Figure B - 2 . F o r m a t i o n t e m p e r a t u r e vs. p r o d u c t i o n e l e v a t i o n
(meters).
A l s o s h o w n is a g e o t h e r m a l g r a d i e n t
of 23 C / k m , s t a r t i n g at 10° at 3 3m.
Data from V u g r a n o v i c h (1986).

238

(meters)

O

A

TRAVERSE

a

■
0^4"

1500 -

60
•

'.i-Arftf

A
a

45
A

• •

^6'a ®
A

40

■
.•

>0

A
■

*«4 •

A .*

A

1000

35

239

ANALYTIC ERRORS-CHARGE BALANCE
One method available for determining the reliability of
chemical

water

balances.
as:

analyses

In this

is

study,

the

calculation

of

charge balances were

charge

calculated

(CATIONS-ANIONS)*100/(CATIONS+ANIONS), where CATIONS and

ANIONS represent the sum of cation and anion equivalences,
respectively.
5% error.

Most charge balances were found to be under

It should be realized that charge balances

for

highly saline formation waters can be misleading and do not
represent
salinity
First,

the

same

waters.
charge

associated
single

degree
This

is

balance

with

of
the

For

several

example,

the

brines is C l , balanced by Ca,
analytic

errors

represent

the

are

sum

of

result

calculations

determining

anion.

reliability

cations

dominant

and

two

cations,

these

error

normally

considered

salinity waters, 5% for example,
error

in

salinities

a

single

or

in

error

offset
in

large,

a

these

Because

calculations

their

by

really

associated

error

The combined

however, and may
Secondly,

acceptable

for

charge
normal

can represent a very large

several

of these waters.

the

K, and Sr.

account for some of the charge imbalance.
balance

normal

phenomenon.

anion

offset by the concentration and error in C l .
error in the cations may be quite

for

include

Na, Mg,

additive,

of

as

components

For example,

at

a +5%

the

high

error

in

charge in a typical Michigan basin water with 350,000 mg/1
TDS,

can

represent

21,000mg/l S04o

an

error

in

measurement

of

about

240

ERROR ESTIMATE
An

important

estimation

of

of

analytic

particularly
brines

part

chemical

precision

important

requiring

any

in

large

and

studying

dilutions

analysis

accuracy.
highly

for

is

an

This

is

concentrated

analysis.

Error

in

this study results from non-determinate sources, and can be
estimated or calculated using propagation of error methods.
Precision of the different component analyses was determined
by measuring
and

is

ten

listed by

replications
component

of the analysis mean,

in Table

and compared

(SD* 10 0/MEAN)

concentrations.

a single

sample

B- 3

(#2078),

Generally,

the

as represented by 1 standard deviation

analytic precision,

variation

of

It

is

should

less
be

by the coefficient

than

noted

5%

of

that

the

these

of

measured
precision

values do not reflect human error in the field dilution of
the brines.

Also

listed

in this

table

are

the

estimated

accuracy in the analyses.
The propagation of error method can be used to estimate
the minimum error (maximum accuracy)
employed in this study
were generally
(except
taken

for
into

(see below).

in the analytic methods
The calculated errors

less than 5% of the measured concentrations

S04 ).
account.

However, other
For

example,

considerations
it

was

the

must

be

authors

experience that the brines were difficult to work with both
in the field and laboratory,
viscosity,

due to their high salinities,

and presence of large amounts of species such as

iron which are known to interfere with many of the analytic

241

TABLE B-3
ANALYTIC PRECISION AND SAMPLE COMPARISONS
Zimmerman #19-A
#2078
Niagara/Salina
Mean
Stnd.
mg/1
Dev.
*
*
*
*
*
*
*

Ca 109160
Mg
11000
Na
27760
24100
K
3586
Sr
Cl 274750
Br
3010
HC03
10
S04
22
* Rb
77.7
* Cs
14.3
Si
1.32

1160
152
135
260
24
4250
80
-

Coeff. of
variance
1.06
1.38
0.48
1.08
0.67
1 .60
2.66
-

0.37
0.43

0.48
3.00

-

-

-

Charge
Balance
2.91%
* = average of 10 analyses, other species n=3.
- = Standard deviation out of range.
Coeff. of variance= Stnd. deviation *100/mean.

Wm. Schmidt B-l
#1024
Traverse brine
This study
Ca
30200
Mg
5560
Na
78100
K
928
Sr
1140
Cl
194000
Br
1280
HC03
5.4
S04
106
Cs
2.2
I
12
Li
32
Rb
3.3
Si
4.28
NH4
106
SPG 1.2263
Charge
balance
-0.73%

SAMPLE COMPARISION
Wm. Schmidt B-l
M.D.N.R.
(1976)
31500
7460
81500
1760
1210
182000
1260
NA
NA
2.36
5
33
NA
NA
NA
1.225
+5.97%

Sample

242

methods used. Considering this, the accuracy of the analysis
were

estimated

(Table

concentrations.

For

alkalinity,

and

B,

B-l)

to

some

error

be

10%

of

components

cannot

be

the

measured

such

as

predicted

S04 ,

and

is

estimated to be much higher ( A.P.I, 1968) .
Maximum accuracy of the analyses were computed by error
propagation
here.

methods.

Two

Propagation

of

typical
error

examples

for

are

presented

procedures

following

X=AaBbCc ... is: Sx/X = [ (SA/A)2 + (Sg/B)2 + (SQ/ C ) 2 + ...],
(a,b,c

=

-1,0,1)

when

the

procedure

follows

X=A+B+C,

the

error is calculated as: S2X = S2A + s 2B + S 2C , and when the
procedure follows F = Ax + By + C z , the error is calculated
as: S 2 (X) = A 2S2 ( x )
SA ,

SB ,

and

calculated

Sq

value

respectively.
the pipets,

+ B2S2 (y) + C2S2 (z).

are

the

(X) , and

variances
measured

For all cases, Sx ,

in

the

values

unknown
A,

B,

or

and

C,

Table B- 4 lists the glassware tolerances for

volumetric,

standards, and the balance used

in

this work.
Example 1) Error in Ca (40,000 mg/1) analysis.
This
this

error

study,

and

is

typical

includes

of

all

error

dilutions, and in the measurement.

A.A.S
in

analysis

field

made

dilution,

in
lab

Field dilution consisted

of making a 50% dilution by adding 125ml of sample to 125ml
of acidified water (termed the dilution factor, df= 2 ) .

This

results in the following error: Sx = [ 0.082 + 0.032 + 0.082
+ 0.032 ]1/ 2 = 0.12ml, where glassware tolerances from Table
B-2

are

used

as

the

variance.

A

1:1000

dilution

of

the

243
TABLE B-4
GLASSWARE TOLERANCES

Class A pipets
Volume
Accuracy
ml
ml
1
0.006
5
0.01
10
0.02
20
0.03
25
0.03
50
0.05
100
0.08

Volumetries
Volume
Tolerance
ml
ml
50
0.05
100
0.08
200
0.10
250
0.12
500
0.20
1000
0.30

Eppendorf pipets
Volume
Accuracy

Precision

ul

ul

ul

10
50
100
200
250
500
1000

0.06
0.30
0.60
1.20
1.50
3.00
6.00

0.04
0.10
0.20
0.40
0.50
1.00
2.00

244

field dilution is next prepared

in the

lab,

(1ml

of

field

dilution brought up to 1000ml). The error in this mixture is
calculated as:

[Q.0022 +

0.302 ]1/ 2 =

0.30ml,

bringing

the

total error at this point to [0.122ml + 0. 302m l ] 1//2=0. 32ml.
10ml

of

the

sample

(and

LaCl2 before analysis,
ml.

The

final

standards)

were

with an error to

sample

volume

error

added

.022 +

is

1ml

to

.0022 =

calculated

to

of
.02
be

0.32ml.
The next error to consider is in the concentration of
the

standards.

1000mg/l+10%.

Stock

standards

have

a

concentration

of

Preparing 100ml of 20mg/l standard results in

an error of Sx = 10* [(10/ 1000)2 + (0.006/1)2 + (0.006/1)2 +
(0.08/100)2 ]1/ 2 ~ 0.014 = 0.14 m g / 1 .
estimated

by

repeated

measurements

Instrument error was

of

standards,

and

was

found to be approximately 5% of the measured concentration,
or

in

this

example

0.5mg/l.

The

resulting

error

measuring a 20mg/l concentration is calculated to be

in

[0.152

+ 0.52 ]1/ 2 = 0.52mg/l.
With these calculated errors, the error in measuring

a

brine having 40,000 mg/1 of Ca is:
N 1 = N 2V 2*df / V x= 20*1000*2/1 = 40,000mg/l
Sni/Nj. -

[<SN 2/N2 )2 + (Sv 2 / v 2 ) 2 +

(Sv 1 / V 1 ) 2 ] 1/ 2

Snl = N l* C (0.52/20)2 + (0.32/2000)2 + (0.006/1)2 J1/ 2
=40,000 * 0.03 mg/1, so SN1

= 1200.

mg/1,

245

Example 2) Error in bromide measurement.
In

this

analysis

the

dilution

factor

is

40x,

which

results from adding 50ul of the 50% field diluted sample to
950ul

of distilled water.

weighing

out

dissolving

1.4943g

in

The

( +

1000ml

(

Br standard was prepared by

0.0005g)
+

0.30ml)

of

oven

dried

distilled

KBr

H 20.

and
The

calculated error in the stock standard is SBr = 1000mg/l*[
(0.0005/1.4943)2 + (0.30/1000)2 ]1/ 2 = 0.45mg/l.
The standards were prepared by diluting this stock into
100ml of water.

The error,

for example,

in a 30mg/l

standard is:
N 1V 1/V2 = 1000mg/l * 3ml / 100ml = 30mg/l
s30 = 30*t (0.45/1000)2 + 3*(.01/1)2 + (0.08/100)2 ]V 2
30*0.02 = 0.60 mg/1

(0.OOlmg/ml)

(3-lml aliquots of stock standard were used)
This error increases to 1.47 mg/1 for a 60mg/l standard.
A typical standardization curve for Br is presented below:

mg Br

SBr (mg)

0

%T

ABS

100

0

SABS

0.03

0.001

54.0

0,270

0.004

0.04

0.001

42.5

0.375

0.005

0.05

0.001

33.0

0.480

0.006

0.06

0.0015

27.0

0.570

0.008

246
Repeated measurements of standards would typically vary
by 0.5%T, which is converted to the variance
(SABS ' listec* above)
decimal

form

of

%T,

by SABS/A =
ABS

in adsorbance

-ST/ (T In T) , where

= absorbance,

Sabs

T =

= variance

in

absorbance in %T.
For
line

is

this

typical

calculated

r =0.998.

standardization

as

Graphical

Br

(mg)

and

=

curve

the

0.0006

numerical

+

methods

best-fit

A B S * 0 .1044
exist

for

calculating the variance in the Br content when the error in
the

best-fit

However,

line

slope

considering

standardization curve,

the

and

intercept

high

are

degree

of

considered.
fit

in

the

the error in the measured absorbance

is estimated by the best-fit line to the calculated error in
the absorbances listed above:
SABS= °-0001 + 0.0134*ABS r2=0.987
The

bromide

analysis

follows

concentration of about 60mg/l
so that

concentrations

spectrophotometer cell

law

up

to

a

(Long and Gudramovics, 1982),

can be
used

Beer's

calculated

in this

from c=A/eb.

analysis

is round

The
and

has a path length of 1.00cm ± 0.001. Neglecting the error in
the molar absorbtivity (e) and the error in adding reagents,
the

error

in

measurement

of

a

50ul

sample

containing

1200mg/l Br (a typical Br concentration for Michigan brines
equal to 0.06mg of Br) is: 0.06mg = 0.57/e*l.00
with a calculated error of:
SBr = 0.06 *[ (0.008/0.57)2 + (0.01/1.00 )2 ]X/ 2 = 0.001 mg
Thus, the brine concentration and error is:

247
(0.06)/(0.05) * 1000ml/l = 1200 xng/1
and:
SBr = 1200 * [(0.001/0.06)2 + (0.003/0.05)2 ] 1/ 2 = 74.7mg/l
r

or 6.2% error.

APPENDIX C

248

APPENDIX C
DATA

Key: ID: Identification number
GROUND ELEV.: Ground elevation in feet above sea
level
PROD. ELEV.: Production elevation in feet above sea
level
TWN: Tier number, + north, - south
RNG: Range number, + west, - east
DEPTH: Depth of well
TDS: Total dissolved solids, in mg/1
SPEC. GRAV: Specific gravity, estimated
TEMP: Estimated down hole temperature, °C
IMBALANCE: Charge balance error, in percent
Chemical data reported in mg/1.
■"O, D : del
0 and del D values, in °/oo vs SMOW
Sr:
Sr/ Sr ratio
MC12 : MCI 2 = Ca + Mg + Sr - S04 - 0.5HCO3 (in meq/1)

249

ID

GROUND
ELEV.
Ft.
BEREA
7001 643.0
7005 650.0
7006 661.7

PROD.
ELEV.
Ft.

TWN RNG SEC

-1778
-1811
-1739

14
15
14

-3
-2
-2

7
36
1

TRAVERSE
1101 694.3
1102 824.0
1103 648.0
1104 670.0
1105 710.0
1107 640.0
1109 877.4
1110 647.3
1111 1175.1
1112 990.0
1113 913.0
1114 836.0
1115 1042.9
1116 835. 0
1117 1230.0
1118 719.0
1119 751.3
1008 972.2
1009 972.4
1011 993.4
1014 913.0
1017 1221.3
1028 718.2
1024 640.0
1048 1142.0
1049 1222.0
1085 888.0
106d 631.2

-644
-960
-823
-850
-1235
-1586
-1460
-2478
-1428
-2113
-1874
-1492
-1737
-1265
-1880
-1204
-484
-1888
-1891
-1885
-1874
-1879
-1296
-1586
-1822
-2268
-1407
-953

1
3
4
4
19
17
9
15
24
12
12
12
14
14
18
7
-1
16
16
16
12
18
19
17
22
19
19
16

13
11
14
15
-4
-4
11
1
5
6
10
13
11
14
9
13
14
10
10
10
10
9
-3
-4
4
10
13
17

10
8
34
2
21
18
35
13
16
10
1
11
8
5
6
36
17
27
27
28
1
6
36
18
17
36
18
6

DUNDEE
3301 725. 3
3302 594 .5
3304 1104 .0
3305 1136.0
3306 1132 .0
3307 723.2
3309 754 .5
3311 665.2
3312 1185.5
3313 796.0
3314 600. 0
3315 840. 3
3002 703.4
3003 601.6

-2200
-2276
-2792
-2640
-1778
-2797
-2881
-3114
-263
-2382
-1468
-1645
-3347
-2201

19
14
19
20
27
17
16
16
22
10
16
21
17
15

-3
-4
6
6
1
-2
3
1
6
5
18
-3
-1
-4

26
2
33
12
28
18
28
26
26
3
1
16
30
30

DEPTH
Ft.
FORMATION
7
7
7

3
3
3
3
3
3
3
3
3
3
3
3
3
3

TDS
SPEC. TEMP
mg/1 GRAV. °C

IMBALANCE
%
SOURCE

2421
2461
2550

328650
284278
335300

1.229 26.3
1.198 26.6
1.230 26.4

1
1
1

4.25
1.43
0.44

1338
1784
1471
1520
1945
2226
2337
3125
2603
3103
2787
2328
2780
2100
3110
1923
1235
2860
2863
2878
2787
3100
2014
2226
2964
3490
2295
1584

273504
258184
268000
268468
295324
306690
326440
325110
286836
315300
308690
278481
303560
264686
260692
272433
263852
254280
249883
301253
316363
235642
299466
298312
297052
294026
282070
215250

1.190
1.179
1.186
1.186
1.206
1.214
1.228
1.227
1.199
1.220
1.215
1.194
1.211
1.184
1.181
1. 189
1.183
1.176
1.173
1.210
1.220
1.163
1.208
1.208
1.207
1.205
1.196
1.149

18.7
21.8
19.6
20.0
22.9
24.9
25.7
31.2
27.5
31.1
28.8
25.6
28.8
24 .0
31.1
22.8
18.0
29.3
29.4
29.5
28.8
31.0
23.4
24.9
30.1
33.8
25.4
20.4

2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1

6.21
3.12
0.88
4.99
5.10
5.95
6.21
6.48
5.51
6.52
3.66
4.03
5.51
4 .90
3.70
4 .19
-0.51
1.22
1. 13
0. 59
0.74
5.18
-1.13
2 .46
0.82
3.78
0.02
-3 .16

2925
2870
3896
3776
2910
3520
3635
3779
381
3178
2068
2485
4050
2803

273921
301830
264050
291750
258517
220063
300650
348500
287310
196796
252806
271217
296662
251291

1.190
1 .210
1.183
1.203
1.179
1.152
1.209
1.243
1.200
1. 136
1.175
1. 188
1.206
1.174

29 .8
29 .4
36 .6
35.8
29.7
34.0
34.8
35.8
36.1
31.6
23.8
26.7
37.7
28.9

2
2
2
2
2
2
2
2
2
2
2
2
1
1

6.22
6 .57
3 .32
0.71
3.26
6.08
-2.92
4.22
3.27
2 .62
3 .57
2 .63
-1.95
0.70

250

ID

GROUND PROD. TWN RNG SEC
DEPTH TDS
SPEC. TEMP
IMBALANCE
ELEV. ELEV.
Ft.
wg/1 GRAV. °C
%
Ft.
Ft.
FORMATION
SOURCE
DUNDEE (Contined)
3004 597.9 -2220 15 -4 26
3 2818 267592 1.186 29.1 1
0.773007
804.3 -1605 20 -4
3
3 2409 283407 1.197 26.2 1
3.99
3012
3
298387 1.208
1
0.37
3015 976.5 -2567 13
9 36
3 3543 328379 1.229 34.1 1
1.50
3016 973.0 -2587 13
9 36
3 3560 328880 1.229 34.3 1
1.70
3022 600.0 -2122 14 -8 29
3 2722 286155 1.199 28.4 1
2.81
3029 746.0 -2835 18
1 14
3 3581 298884 1.208 34.4 1 -0.21
3033 787.6 -2768 15
4
2
3 3556 315206 1.220 34.2 1
2.69
3042 1183.0 -2728 21
7 22
3 3911 299414 1.208 36.7 1
2.74
3051 808.6 -1598 20 -4 10
3 2407 283464 1.197 26.2 1
4.13
3052 590.3 -2145 14 -4 12
3 2735 218059 1.151 28.5 1 -9.59
3070 1123.0 -3869 20
6 31
3 4992 298290 1.208 44.3 1 -2.83
3071 1109.0 -2629 20
6 29
3 3738 309790 1.216 35.5 1 -4 .18
3027 736.4 -2790 18
1 23
3 3526 313898 1.219 34.0 1
4.94
3034 754.3 -2899 15
4 13
3 3653 301726 1.210 34.9 1
0.35
3025 787.5 -2094 19 -3 22
3 2881 277895 1.193 29.5 1 -0.77
3023 645.5 -2310 17 -4 18
3 2955 240643 1.167 30.0 1 -0.01
3047 1135.8 -2802 21
3 30
3 3938 276568 1.192 36.9 1
2.26
3039 1195.0 -2653 22
7 25
3 3848 269534 1.187 36. 3 1 -2.34
3026 781.6 -2104 19 -3 15
3 2886 292162 1.203 29.5 1
0. 32
3063 587.7 -2097 14 -8 29
3 2684 269155 1.187 28.1 1
3.50
3070 1123.0 -3869 20
6 31
3 4992 298290 1.208 44.3 1 -2.83
3087 692.3 -1550 17 16 26
3 2242 315100 1.220 25.0 1
0. 59
3043
21
7 30
3
298021 1.207
1
0.33
3082 591.3 -1474 16 18
1
3 2065 261758 1.182 23.8 1
0.64
3086 700.0 -1549 17 16 25
3 2249 232360 1.161 25.1 1 -0.32
3081 643.7 -1462 16 17
8
3 2106 213550 1.147 24.1 1
0. 67
RICHFIELD
4010 899.5
4013 911.1
4018 912.2
4035 1395.5
4036 772.1
4037 786.5
4068 1135.8
4069 1138.0
4046 1142.0
4050 1259.7
4403 1111.8
4406 1010.0
4407 1252.0

-2404
-2421
-2393
-2747
-3871
-3993
-2802
-3918
-3975
-3004
-3840
-3037
-2928

12
12
12
24
15
16
20
19
22
24
20
22
24

10
10
10
-2
4
4
6
6
4
-1
5
-2
-1

1
1
12
33
11
35
30
19
8
26
8
18
14

4
4
4
4
4
4
4
4
4
4
4
4
4

3303
3332
3305
4142
4643
4779
3938
5056
5117
4264
4952
4047
4180

326500
327135
146374
320218
125260
294832
379680
386780
334670
317460
315240
395297
375940

1.228
1.228
1.100
1.223
1.085
1.205
1.265
1.270
1.233
1.221
1.220
1.277
1.263

32.5
32.7
32.5
38.3
41.8
42.8
36.9
44.7
45.2
39.2
44.0
37.7
38.6

1
1
1
1
1
1
1
1
1
1
2
2
2

2.23
0.51
1. 52
0.55
3.62
1.97
-0.66
0.93
1.31
5.04
7.40
7.27
5.58

DETROIT RIVER
5019 647.0 -3245
5030 735.5 -3777

14
18

-8
1

36
36

5
5

3892
4512

298670
421420

1.208 36.6
1.295 40.9

1
1

5.84
6.01

251

ID

GROUND PROD.
ELEV. ELEV.
Ft.
Ft.
NIAGARA/SALINA
2206 1233.0 -5229
2207 1255.0 -4297
2209 616.0 -1988
2210
745.0 -2563
2211
652.7 -2252
2212
966.8 -2084
2213 848.0 -2323
2214 586.4 -1570
2215 938.0 -3248
2020
607.0 -6403
2072 902.3 -3165
2077 915.0 -5275
2078 1067 .0 -4843
2083 838
-3872
2084 762 .
-4474
2089 808.
-4269
2090 849.
-5497
2091 915.
-5866
2092 808.
-2923
2093 812.
-2854
2100
964.0 -3140
2101
970.0 -3094
2097 925.0 -2256
2096 930.2 -2257
2099 917.0 -2727
2098

TWN RNG SEC

DEPTH
Ft.
FORMATION

TDS
SPEC. TEMP
mg/1 GRAV.
C

IMBALA1
%
SOURCE

29
30
4
5
5
-1
3
3
2
14
2
26
25
23
23
25
26
26
33
33
2
2
-1
-1
1
2

4
2
-15
-15
-15
-3
11
-16
-1
-8
-3
10
11
15
14
13
10
10
-3
-3
-4
-4
4
4
-7
-3

32
5
32
5
24
6
17
10
12
31
35
28
9
14
10
31
10
12
3
3
19
36
27
34
18
17

2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2

6462
5552
2604
3308
2905
3051
3171
2156
4186
7010
4067
6190
5910
4710
5236
5078
6346
6782
3731
3666
4104
4064
3181
3187
3644

441270
406980
195724
284420
272320
326460
343114
282530
391890
349730
363330
219944
443700
349035
355080
407930
422130
280140
343690
399790
393590
392300
365700
350500
370460
380126

1. 309
1.285
1. 135
1. 198
1.189
1.228
1.239
1. 196
1. 274
1. 244
1. 254
1. 152
1. 311
1. 244
1. 248
1. 285
1. 296
1. 195
1.240
1. 280
1.275
1. 274
1.256
1. 245
1. 2 59
1. 266

54.6
48. 2
27. 6
32. 5
29. 7
30.7
31.5
24.4
38. 6
58.4
37. 8
52.7
50. 7
42. 3
46.0
44 .9
53 .8
56. 8
35.5
35. 0
38. 1
37.8
31. 6
31. 6
34 .8

2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
2
2
1
1
1
1

5.78
2.91
6.91
5.99
2.39
2.04
4.26
7.47
5.45
4.95
-1.38
4.85
4.73
4.38
4.02
4.80
4.70
0.41
3.90
4 .12
6. 38
5.99
3.76
5.31
2 .83
6.08

TRENTON-BLACK RIVER
6601 1086.2 -3289
6602 949.5 -3761
6603 1044.6 -3155
6604 3177.0 -2885
6605 1125.2 -2907
6606 1149.0 -2921
6073 1179.0 -2893
6074 1133.0 -2891
6075 1071.0 -2875
6076 1071.0 -2875
6095 1113.4 -3304
6094 1085.1 -3260

-4
-1
-3
-5
-5
-6
-5
-5
-5
-5
-4
-4

2
5
4
2
3
2
2
3
3
3
2
2

9
25
36
31
25
5
32
23
3
3
18
7

6
6
6
0
6
6
6
6
6
6
6
6

4375
4710
4200
4062
4032
4070
4072
4024
3946
3946
4417
4345

229560
273040
201140
210375
189679
170930
150672
213336
209660
209424
231572
226794

1. 159
1. 190
1. 139
1. 145
1. 131
1. 117
1. 103
1. 147
1. 145
1. 145
1. 160
1. 157

40.0
42 .3
38 .7
37 .8
37. 6
37 .8
37 .8
37 .5
37 .0
37 .0
40. 3
39.8

2
2
2
2
2
2
1
1
1
1
1
1

0.24
6.04
3 .13
8.99
1.84
2.32
2.93
2 .01
2.83
2.60
0. 00
0.10

St. PETER
8040 1196.0

22

7

36

8 10586

403770

1. 283 83 .5

1

0.01

-9390

252

ID
BEREA
7001
7005
7006

Ca

K

49500
38400
48300

580
700
650

8
8
8

8790
7600
8400

66700
58200
64500

TRAVERSE
1101
23400
1102
20600
1103
17800
1104
21900
1105
20200
1107
31500
1109
45400
1110
67700
1111
21200
1112
44500
1113
43500
1114
28900
1115
38100
1116
20000
1117
28000
1118
26700
1119
23000
1008
40600
1009
39600
1011
49400
1014
41400
1017
21400
1028
20400
1024
30200
1048
34600
1049
35600
1085
23100
1088
11600

1440
1220
1330
1620
1480
1760
2350
4750
1680
3130
3320
1630
2050
1300
2050
1630
1520
2330
2310
2930
2410
1160
480
930
1920
1620
1260
440

12
9
12
18
27
33
40
81
41
46
50
35
40
29
33
16
18
38
38
48
48
30
18
32
40
24
34
19

6640
7140
5430
5430
7120
7460
7830
9700
8320
8290
7430
6650
8920
5840
6930
8450
6250
6930
6800
8510
6720
4400
4800
5560
5540
6300
5870
3840

78600
70200
85200
78400
88700
81500
73900
43700
82400
69400
64200
71300
69700
78200
63800
68500
66800
43720
43000
48500
67000
67100
78000
78100
70200
73500
75000
62300

DUNDEE
3301
3302
3304
3305
3306
3307 •
3309
3311
3312
3313
3314
3315
3002
3003

1980
2950
2690
3350
1100
1880
4380
8360
2360
2180
1220
1930
1880
980

20
33
32
41
10
17
41
68
41
27
29
22
36
14

7310
7040
6350
5950
2640
7010
6530
11300
8330
5030
4340
5030
4720
5000

80600
76300
61800
66300
92900
56500
57600
38800
67000
41000
73600
82000
76600
62500

20500
35000
31000
33700
7390
22100
38500
73300
32100
26700
20500
17200
25600
25600

Li

Mg

Na

PH

Sr

4.6
5.1
5.0

2140
1780
2080

5.4
5.8

710
734
650
470
940
1210
1540
2920
1330
1500
1690
960
1460
610
910
1160
577
1540
1460
1880
1620
800
900
1140
1250
980
940
350

4.8
5.3

675
1290
1020
1230
170
760
1180
2350
1200
910
470
460
780
946

5.3
4.8
4.9
5.9
5.4
4.5
5.8

Rb
3.4
3.9
3.5

Cs

Si

2.9

4.0
4.4
1.4

6.7
3.1
8.3
6.8
3.3
2.2
3 .3
17.2
1.7
0.4

5.2
3.6

2.9
3.2
3.0
2.1
2.9
1.4
4.3
2.9

0.2
0.6
0.6
0.2
0.3
0.2
0.3
1.4
0.2
0.5
0.2
0.4
2.5
2.4

5.2
4.0

1.8
2.1

253

ID
Ca
K
DUNDEE (Continued)
3004
14400
380
3007
19800
1014
3012
33600
2150
3015
51000
4040
3016
51800
4160
3022
21400
600
3029
28000
1880
3033
36400
3010
3042
26000
1870
3051
21800
1220
3052
11000
800
3070
29900
1680
3071
25600
1120
3027
25400
1910
3034
44900
2820
3025
15900
1120
3023
23000
920
3047
29600
3224
3039
32000
2540
3026
21700
1120
3063
16000
520
3070
29900
1680
3087
10300
12 20
3043
33100
1650
3082
20900
900
3086
19800
960
3081
12500
640

Li

Mg

Na

PH

Sr

6
26
40
54
52
10
52
44
33
24
2
18
18
32
44
24
24
38
44
22
2
18
29
41
33
27
20

4200
4020
6140
7040
6880
6810
5460
5790
5070
4000
3000
5220
5660
4660
7200
3540
4420
4900
5080
4000
5320
5220
1820
6060
4110
3860
3750

83500
87900
69300
60500
60400
81600
75300
76100
83500
86100
68000
70400
77200
94200
55800
84400
61300
68800
68400
92400
81100
70400
103000
69900
73600
62400
64700

4.5
4.6
5.2
5.2
5.0
3.5
5.3
5.5
5.2

420
600
1180
1740
1720
730
2000
1340
780
440
280
1170
1020
740
1340
460
750
850
1020
480
580
1170
340
1620
560
580
340

1.8
2.9
6.0
10. 2
11.1
2.6
5.4
8.1
5.2
3.4
0.4
2.5
2.4
5.3
7.7
3 .3
2.9
2.9
7.1
3.0
1.7
2 .5
1.5
4.9
0.9
1.0
0.6

RICHFIELD
4010
53000
4013
51800
4018
50500
4035
61400
4036
27800
4037
65100
4068
82100
4069
86100
4046
77600
4050
69600
4403
73800
4406
96200
4407
92000

4070
4020
3570
8320
3800
7700
8660
8700
9300
9020
9110
12800
13300

54
54
50
42
24
32
44
46
76
40
65
47
54

7180
6960
7090
8830
3370
7700
9740
10700
10700
9840
9700
14600
12600

58600
57800
61500
38400
12400
28600
34500
34100
31300
33000
31000
32200
25600

5.1
5.0
5.0
4.2
5.9
4.6
4.7
4 .3
3.7

1720
1640
1640
2130
1040
2580
3480
3980
2800
2300
2940
290
3390

10.4
10.9
9.2
23 .8
8.9
21.7
10.5
11.0
23.4
6.5

3.9
3 .9
3 .6
5.5
8 .9
5.5
2 .2
2 .2
6. 4
8.8

DETROIT RIVER
5019
72400
5030
107000

9270
19300

72
120

10400
13300

22200
24000

4.9
3.9

2340
3070

33.8
41.2

6.4

5.3
4.8
4.3
5.2
4.1
4.7
5.9
4.8
5.3
4.8
5.1
5.6
3.9
5.1
6.2
6.0

Rb

Cs

Si

1.3
1.6
2.8
4 .3
4.4
1.6
2.9
3.1
2.3
1.8
1.0
0.9
1.0
2.1
3.1
1.6
1.7
1. 1
2.8
1.6
1.4
0.9

2.4
2.0
2.5
1.5
2.3
2.1
2.0
3.3
1.2
8.1
4.3
3.3
11.6
1.9
2.1
1.7
1.2
3 .7
2 .6
1.4
2 .5
3 .3
3 .3
1. 2
2 .3
3 .7
2 .3
4 .2
1.6
3.3

1.6
2.1
2.3
8.5
3.2
2 .9
2 .4
1 .4
0.9

1. 6
2.3

254

ID
Ca
NIAGARA/SALINA
2205
35400
2206
124000
2207
102000
2208
40500
2209
35200
2210
55900
2211
49900
2212
65100
2213
65100
2214
48000
2215
85400
2020
65200
2072
73400
2077
60000
2078
109200
2083
61100
2084
80300
2089
99100
2090
106000
2091
67700
2092
72900
2093
89800
2100
86500
2101
86700
2097
76600
2096
74200
2099
77000
2098
79600

K

Li

9560
22600
21200
2110
2300
4070
3320
6470
9340
4460
11800
11200
8070
9500
24100
11400
19200
19000
21200
14200
5890
6700
8800
7920
7320
6960
8610
8780

18
61
85
17
18
39
29
50
55
36
83
60
56
34
70
68
64
93
74
45
82
101
74
68
69
67
89
74

6970
12800
12600
14000
6230
7800
8350
10300
15100
11300
14300
10400
15500
6080
11000
7700
11300
11600
12600
8670
14600
17500
20000
19300
10700
10300
10700
16700

8010
9820
16100
34300
33500
43100
39400
38300
40300
47800
38300
48800
28400
23000
27800
27600
23600
26200
21200
21000
38000
34000
33600
33900
43400
43400
42200
40200

TRENTON-BLACK RIVER
6601
25600
5440
6602
35500
3960
6603
23200
4360
6604
23100
3820
6605
19200
4010
6606
17400
3490
6073
15800
2180
6074
23400
3200
6075
23000
3220
6076
22800
3220
6095
25800
4300
6094
24600
4120
St. PETER
8040
90100 19500

42
47
36
37
33
29
26
40
40
40
43
43

6150
9110
6020
5300
5360
4780
3200
4700
4800
4600
4750
4740

48400
59000
43800
55500
43600
39900
37200
50200
49800
50000
51600
51300

44

9710

28200

Mg

Na

Sr

Rb

Cs

1960
6030
4010
650
800
1330
1120
1860
1630
910
2870
1790
2230
1800
3590
2040
3400
3440
4330
2750
1800
2140
2580
2500
1980
1940
1840
2290

23.4
14.3
27.7
77.7
35.2
49.5
60.5
71.1
48.5
10.7
12.3
16.1
13.0
13.1
13.2
14.2
17.9

5.3

6.3
0.9

5.4
5.4
5.4
5.8
6.1

840
1040
740
680
640
600
540
760
760
760
870
870

4.5
7.4
7.4
7.4
11.3
11.4

3.8

3160

58.1

11.9

19.9

PH

4.3
5.0
4.8

4.5
5.0
5.1
4.8
4.7
4.7
5.1
3.5
5.0

J

>*

Si

255

ID

Br

Cl

ALK I
as h c o 3

so4

B

n 18o

nd

° /0 0

°/oo

BEREA
7001 940
7005 1600
7006 1840

200000
176000
209000

25 28 100
25 25 85
10 17
11

TRAVERSE
1101 710
1102 2290
1103 590
1104 650
1105 880
1107 1260
1109 1420
1110 2340
1111 910
1112 1480
1113 1550
1114 1040
1115 1330
1116 730
1117 1000
1118 990
1119 710
1008 1560
1009 1510
1011 1930
1014 1510
1017 880
1028 890
1024 1260
1048 1247
1049 1050
1085 900
1088 720

162000
156000
177000
160000
176000
182000
194000
194000
171000
187000
187000
168000
182000
158000
158000
165000
165000
158000
155000
188000
196000
140000
175000
182000
184600
182000
175000
136000

13
4 22
2 24 110
16
4 31
32
6 215
1
9 17
5 85
2
7 10 100
2 28 30
73
5
2
20 12 160
27 18 60
68
7 73
1 11 80
24
6
9
44
10 70
7
5
5
4 155
33
38 18 208 22
36 20 297 24
31 25 86 68 -0.57 -23.80
11 18 59 27 -0.70 -31.50
59 12 109 18 -3.60 -40.00
3 12 67
-0.49 -32.50
5 12 106
0.56 -29.10
18 10 162 117
18 12 150
61 71 28
-1.83 -43.60
156 10 607
-7.06 -61.20

DUNDEE
3016 1920
3301 860
3302 1250
3304 1190
3305 1220
3306 315
3307 820
3309 1460
3311 2390
3312 1320
3313 980
3314 ' 680
3315 590
3002 1180
3003 1070

202000 59
162000 63
178000 59
160000 43
180000 63
154000
2
131000 52
191000 39
212000 139
175000 49
120000 71
152000 54
164000
9
186000 26
155000 23

20
6
5
10
12
5
4
12
26
11
8
3
7
15
18

1.10 -10.70
0.29 -24.10

103 84 1.31 -35.50
317
150
410
122
810
676
140
40
33
1130
364
660
141 39 0.20 -30.80
295 11 -2.38 -28.10

87Sr
0.70915
0.70909

NH,N

NaD
°/oo

MCI,
jneq/I

50
1.04 3240.4
64 -14.80 2580.6
126
3148.7

193
180
121
146
83
0.70940 76
0.70909 106

-12.43
-21.52
-34.16
-26.45
-21.21

0.70845 174 -36.89
76 -57.26

1729.9
1630.0
1349.4
1546.2
1615.0
2211.6
2942.9
4242.5
1773.5
2933.8
2819.6
2010.3
2666.9
1492.5
1987 .1
2054.1
1672.2
2627.3
2563.1
3206.7
2654.7
1446.4
1432 .1
1988.4
2207.7
2314.3
1657.1
891. 5

0.70813 216 -23.69 3188.6
1633.8
2352.6
2084.5
2197.3
573.0
1683.3
2482.8
4641.!
2314. 1744.1
1383.7
1269.0
119 -23.84 1680.9
29 -21.50 1704.5

256

ID

Br

ALK
as h c o 3
DUNDEE (continued)
3004 690 164000 20
3007 750 169000 64
3012 1420 185000 28
3015 1860 202000 51
3022 1020 174000 37
3029 1240 185000 43
3033 1570 191000 27
3042 1140 181000 11
*3043 1090 185000 35
3051 1500 169000 10
3052 480 162000 17
3070 1220 189000
2
3071 1290 198000 29
3027 1190 186000 28
3034 1570 188000 27
3025 780 172000 75
3023 950 149000 21
3047 1090 168000 40
3039 1190 188000 42
3026 790 192000 86
3063 650 158000 27
3070 1220 189000
2
3087 490 181000 11
3082 690 161000 84
3086 760 144000 98
3081 620 131000 106
RICHFIELD
4010 1930
4017\ 1920
4018 1870
4035 1240
4036 1050
4037 2610
4068 1200
4069 1200
4046 3200
4050 2700
440'1 2690
4406 4210
4407 4050

Cl

200000 56
203000 48
202200 45
200000 114
75800 229
182000 213
240000 240
242000 240
220000 244
191000 240
186000 278
235000 385
225000 207

I

so4

B

n 18o

°/oo
13
28
13
19
8
19
17
13
14
20
21
15
15
10
20
10
9
9
17
12
8
10
5
7
8
12

382
319
112
109
166
139
160
137
94
200
314
31
30
220
96
295
189
67
118
429
20
31
491
364
510
745

1
20
52
80
2
19
31
8
8
38

29 112
17 74
11 109
40 33
8 348
25 161
28 0
30 0
13 0
10 0
19 0
4
66
68 34

75
81
65
371
137
180
39
82
377
137

ND
°/oo

-3.02 -28.60
-4.51 -52.00
-0.40 -34.20
0.43
-0.53
-2.03
-1.05

-26.10
-30.80
-39.20
-30.95

-6.11 -47.60
31
66
37
85
26
6
53
47
28
1
31
11

DETROIT RIVER
5019 3060 179000 326 57
2 383
5030 3750 251000 100 25 54 134

87Sr

n h 4n

NaD
°/oo

64 -23.65
-46.04
228 -25.69
200
45
162 -18.58
185 -21.66
0.70831 107 -31.95
174 -22.56
0.70838
-43.81

-3.39 -41.60

125
164
71 -50.26
0.70898 81 -33.91
161
135 -22.06
73
-36.16

-4.02 -46.30
-5.71 -55.80
-6.58 -52.70

109
73 -40.45
115 -50.45
78 -48.57

-4.99 -55.40
-2.27 -39.90
-0.38 -30.00

1.49 -27.00
4.31 -51.70

2.93 -41.60

2.58 -51.70
4.36 -16.70

mc12
meq/I
1066.0
1326.4
2206.7
3162.1
1641.8
1889.6
2320.3
1729.7
2185.7
1422.9
795.8
1947.6
1766.1
1663.5
2861.8
1089.7
1524.8
1898.6
2036.0
1414.7
1249.2
1947.6
661.4
1387.0
1309.1
925.5

214
3272.9
0.70819 203 -15.24 3193.9
198
3138.9
0.70809 401 -37.99 3839.4
115
1682.9
242
3939.4
4979.7
-22.96 5269.7
333
4818.7
4337.3
4550.2
6011.5
5706.0
441 -36.13 4524.6
0.70782 689
6.97 6503.5

Br

Cl

ALK
as h c o 3

NIAGARA/SALINA
2206 3020 263000 614
2207 3070 248000 258
2208 1150 140000 103:
2209 690 117000 52
2210 2220 170000 31
2211 2230 168000
3
2212 2430 202000 108
2213 2640 211000 49
2214 2060 168000 32
2215 3220 236000 53
2020 3340 209000 50
2072 2730 233000 150
2077 1270 153000 150
2078 3010 265000 150
2083 1800 169000 150
2084 2360 213000 150
2089 3060 244000 150
2090 2940 253000 374
2091 1820 190000 207
2092 2500 218000 17
2093 3300 248000 36
2100 3110 239000 39
2101 2980 239000 240
2097 2700 223000 123
2096 2700 211000 84
2099 3110 227000 88
2098 2560 230000 352

(ft
Oc
Ef

ID

H

257

28 10
6
27
22 :
152
7 71
19 70
23 30
22 29
13 87
14 154
22 18
19 110
9 56
32 119
50 22
41 106
64 131
50 75
31 85
29 28
17 34
19 25
22 42
21 52
27 75
38 82
54 75
28 34

TRENTON-BLACK RIVER
6601 1130 142000 27
6602 1430 163000 39
6603 1020 122000 40
6604 980 121000 43
6605 870 116000 55
6606 760 104000 90
6073 750
91000 25
6074 1080 130000 25
6075 1080 127000 25
6076 1040 127000 25
6095 1250 143000 67
6094 1160 140000 39

9
11
8
8
7
6
10
13
13
10
15
26

St. PETER
8040 3100

20

250000

10

369
113
393
530
249
717
250
238
125
125
100
100

B

n 18o

nd

°/oo

°/oo

220 0.75 -37.40
81
106
239 2.87 -48.70
354
346
498
226
152
25 -3.39 -49.80
123
98 0.30 -40.80
72
81 -1.99 -24.30
72
61
86

87Sr

NHaN

0.70799 437
235
283
0.70879 56
672
784
1087
1053
667
0.70830 356
415
549
476
451
431
641
588

NaD
°/oo

-22.15
-24.61

-32.86
-19.84
-7.32

MCI,
meq/I
7383.4
6220.2.
3185.5
2286.4
3460.4
3202.1
4138.8
4526.9
3342.9
5503.8
4148.3
4989.1
3534.2
6437.0
3728.1
5014.3
5977 .9
6426.4
4155.6
4879.7
5969.7
6020.5
5972 .5
4747.5
4593.4
4764.0
5400.7

6
13 -1.85 -26.40
13
15
15 -1.76 -26.70
21 -1.99 -24.30

1795.2
2542.8
1662.1
1593.7
1409.0
1261.1
45
1059.1
73 -20.39 1567 .0
70
1557.6
78
1531.2
120 -20.15 1696.6
0.71031 126 -17.97 1635.7

0 107 -2.96 -46.50

0.70920 247 -26.66 5367.3

BIBLIOGRAPHY

258

BIBLIOGRAPHY

Ahrens L.H.
(1965) Distribution of the
Planet. McGraw-Hill, New York,
llOp.

Elements

in

Our

American Petroleum Institute (1968) Recommended practice for
Analysis
of
Oil-field
Waters. American
Petroleum
Institute, Washington, D.C., 2nd. edition, 58p.
Anderson R.J., Graf D.L. and Jones B.F. (1966) Calcium and
bromide contents of natural waters. Science, 153, 16371638.
Back W. and Barnes I. (1965) Relation of electrochemical
potentials and iron content to ground water flow
patterns. U.S. Geol. Surv. Prof. Paper 498-C, 16p.
Bassett R.L. and Bentley M.E. (1983) Deep Brine Aquifers in
The Palo Duro Basin: Regional Flow and Geochemical
Constraints.
Texas Bur. Econ. Geol. Rep. Invest. 130,
59p.
Beecker R. (194 0)
Chemical variations in the Dundee
Formation waters.
Unpub. M.Sc. thesis, Univ.
of
Michigan,
68p.
Bein A. and Land L.S. (1983)
Carbonate sedimentation and
diagenesis associated with Mg-Ca-Cl brines: the Permian
San Andres Formation in the Texas panhandle. J. Sed.
Pet. 53, 243-260.
Bethke C.M. (1980) Compation-driven groundwater flow and
heat transfer in intercratonic sedimentary basins and
genesis
of the
upper Mississippi
valley
mineral
district, unpubl. Ph.D. thesis, Univ. of Illinois,
Urbana, 111., 124p.
Billings G.K., Hitchon B. and Shaw D.R. (1969) Geochemistry
and origin of formation waters in the western Canada
sedimentary basin, 2. Alkali metals. Chem. Geol. 4,
211-233.

259

Borchert H. and Muir R.O. (1964) Salt Deposits. The Origin.
Metamorphism
and
Deformation
of
Evaporites.
Van
Nostrand, London, 338p.
Braitsch 0.
(1971)
Salt Deposits;
Their origin
Composition. Springer-Verlag, New York, 297p.
Brand

and

U. and Veizer J. (1980) Chemical diagenesis of a
multi-component system: I. Trace elements. J. Sed. Pet.
51, 987-998.

Brewer P.G. (1975) Minor elements in sea water. In: Riley
J.P. and Skirrow G. (eds.) Chemical Oceanography. 2nd
ed. vol. 1, Acedemic Press, New York, pp. 415-496.
Briggs L.I., Gill D. , Briggs D.Z. and Elmore R.D. (1980)
Transition from open marine to evaporite deposition in
the Silurian Michigan basin.
In: A Nissenbaum (ed) ,
Hvpersaline
Brines
and
Evaporitic
Environments.
Developments in Sedimentology 28, Elsevier, pp. 253296.
Brown E . , Skougstad M.W. and Fishman M.J. (1970) Methods for
collection and analysis of water samples for dissolved
minerals and gases. U.S. Geol. Sur. Tech. of Water-Res.
Inv., book 5, chapter A-l, 160p.
Burke, W.H., Denison R.E., Hetherington E.A., Koepnick R.B.,
Nelson H.F. and Otto J.B. (1982) Variation of seawater
Sr/86Sr throughout Phanerozoic time. Geology 10, 516519.
Burst J.F. (1969)
Diagenesis of Gulf Coast clayey sediments
and its possible relation to petroleum migration. Bull.
Am. Ass. Petrol. Geol 53, 73-93.
Bush

P.R.
(1970)
Chloride-rich
brines
from
shabka
environments and their role in ore formation. Inst.
Min. Metal1. Trans. 79, B137-B144.

Carpenter A.B. (1978)
Origin and chemical evolution of
brines in sedimentary basins. Okl. Geol. Sur. Cir. 79,
60-77.

260

Carpenter A.B. (1979)
Interm report on lead and zinc in
oil-field brines in the central Gulf Coast and in
southern Michigan. Society of Mining Engineers of AIME
reprint number 79-95.
Carpenter A.B. (1980)
The chemistry of dolomite formation
I : the stability of dolomite. In: Concepts and Models
of Dolomitization. SEPM special publication 28. Soc. of
Econ. Paleo. and Mineral., Tulsa, Ok., pp. 111-121.
Carpenter
A.B.,
Trout
M.L.
and
Pickett E.
(1974)
Preliminary report on the origin and chemical evolution
of lead-and Zinc-rich oil field brines in central
Mississippi.
Econ. Geol. 69, 1191-1206.
Case L.C. (1945)
Exceptional Silurian brine near Bay City,
Michigan. Bull. Am. Ass. Petrol. Geol. 29, 567-570.
Case L.C., Heck E.T., Fash R.H., Minor H.E., Crawford J.G.,
Jessen F.W., Foster M.D.
and Ginter R.L.
(1942)
Selected annotated bibliography on oil-field waters.
Bull. Am. Ass. Petrol. Geol. 26, 865-881.
Cercone K.R. (1984)
Thermal history of Michigan
Bull. Am. Ass. Petrol. Geol. 68, 130-136.

basin.

Cercone K.R. and Lohmann K.C. (1987)
Late burial diagenesis
of Niagaran
(Middle
Silurian)
pinnacle
reefs
in
Michigan basin. Bull. Am. Ass. Petrol. Geol. 71, 156166.
Chave

K.E. (1960)
Evidence on history of seawater from
chemistry of deeper subsurface waters of ancient
basins. Bull. Am. Ass. Petrol. Geol. 44, 357-370.

Clayton R . , Friedman I., Graf D. , Mayeda P., Meets W. and
Shimp. N.F. (1966)
The origin of saline formation
waters: I . Isotopic composition. J . Geophy. Res. 71.
3869-3882.
Cohee

G. V. and Landes K.K. (1958)
Oil in the Michigan
basin.
In: Habitat of Oil. Am. Ass. Petrol. Geol.,
Tulsa, Ok., pp.473-493.

261
Collins A.G.
(1969a)
Chemistry of some Anadarko Basin
brines containing high concentrations of iodide. Chem.
Geol. 4, 169-187.
Collins A.G.
(1975)
Elsiever, 496p.

Geochemistry

of

oilfield

brines.

Collins,
A.G., Cassaggno J.L.,
and Macy V.W.
(1969b)
Potentiometric determination of ammonium nitrogen in
oil-field brines, Envir. Sci. Tech. 3, 274-275.
Cook L. (1914)
Salt and brine deposits in Michigan.
Geol. Sur. Publ. 15, lOOp.
Craig H. (1961)
Isotopic variations
Science 133, 1702.

Mich.

in meteoric waters.

Das N. and Holland H.D. (1980)
Fluid inclusions in halite
from the Salina group of the Michigan basin, and the
chemistry
of
Late
Silurian
seawater.
unpublished
manuscript.
Davis J.C. (1986) Statistical and Data Analysis in Geology.
John Wiley and Sons, New York,
646p.
Dawdy

D.R. and Feth J.H. (1967)
Application of factor
analysis in study of chemistry of groundwater quality,
Mojave River Valley, California. Water Res. Res. 3,
505-510.

Dellwig L.F. (1955)
Origin of the Salina salt of Michigan.
J. Sed. Pet. 25, 83-110.
Degens E.T., Hunt J.M., Reuter J.H., and Reed W.E. (1964)
Data on the distribution of amino acids and oxygen
isotopes in petroleum brine waters of various geologic
ages. Sedimentology 3, 199-225.
De Sitter L.U. (1947)
Diagenesis of oilfield brines. Bull.
Am. Ass. Petrol. Geol. 31, 2030-2046.
Dickey P.A. (1966)
Increasing concentration of subsurface
brines with depth. Chem. Geol. 4, 361-370.

262
Dickson A.G. (1984) pH scales and proton-transfer reactions
in saline media such as seawater. Geochim. Cosmochim.
Acta. 48, 2299-2308.
Domenico P.A. and Robbins G.A. (1985)
The displacement of
connate waters from aquifers. Geol. Goc. Amer. Bull.
96, 328-335.
Dorr J.A. and Eschman D.F. (1984) Geology of Michigan. The
University of Michigan Press, Ann Arbor, Michigan,
497p.
Dresel P.E. and Rose A.W. (1982) Chemistry and origin of oil
and gas well brines in Western Pennsylvania, unpubl.
manuscript.
Drever J. I . (1988)
The Geochemistry
Prentice-Hall, New Jersey, 436p.

of

Natural

Waters.

Dutton A. R. (1987)
Origin of brine in the San Andres
formation, evaporite confining system, Texas Panhandle
and eastern New Mexico. Geol. Soc. Amer. Bull. 99, 103112.

Egleson G.C. and Querio C.W. (1969)
Variations in the
composition of brines from the Sylvania Formation near
Midland Michigan. Envir. Sci. Tech. 3, 367-371.
Ells G.D. (1967)
Michigan's Silurian Oil and Gas Pools.
Mich. Dept. Nat. Res. Geol. Sur. Rep. Invest. 2, 49p.
Eugster H.P., Harvie C.E. and Weare J.H. (1980) Mineral
equilibria in a six-component seawater system, Na-K-MgCa~S04-Cl-H20, at 25°C.
Geochim. Cosmochim. Acta 44,
1335-1347.
Faure G . , Hurley P.M. and Fairbairn H.W. (1963) An estimate
of the isotopic composition of Sr in rocks of the
Precambrian Shield of North America. J. Geophy. Res.
68, 2323-2329.
Friedman G.M. and Sanders J.E. (1967) Origin and occurances
of dolostones.
In: G.V. Chilingar, H.J. Bissel, and
R.W.
Fairbridge
(eds.) Carbonate Rocks
- Origin.
Occurrence, and Classification. Elsevier, pp. 267-348.

263

Fritz P. (1977)
Geochemical characteristics of dolomites
and the 180 content of Middle Devonian oceans. Earth
Planet. Sci. Lett. 11, 277-282.
Fritz P., Frape S.K., Drimmie R.J. and Heemskerk A.R. (1986)
Reply to comments by Grabczak et al. on "Water-rock
interaction and chemistry of groundwaters from the
Canadian Shield." Geochim. Cosmochim. Acta. 50, 15611563.
Gardner W.C.
(1974)
Middle Devonian stratigraphy and
depositional
environments
in
the
Michigan
basin.
Michigan Basin Geological Society Special Paper l,
Lansing, Michigan, 135p.
Gilbert R.O. (1987) Statistical Methods for Environmental
Pollution Monitoring. Van Nostrand Reinhold Co. , New
York, 32Op.
Gill

D. (1977)
Salina A-l sabkha cycles and the Late
Silurian paleogeography of the Michigan basin. J. Sed.
Pet. 47, 979-1017.

Gill D. (1979)
Differential entrapment of oil and gas in
Niagaran pinnacle-reef belt of northern Michigan. Bull.
Am. Ass. Petrol. Geol. 63, 608-620.
. Graf D.L., Meents W.F., Friedman I . and Shimp N.F. (1966)
The Origin of Saline Formation Waters: 3. CaCl Waters.
111. St. Geol. Sur. Cir. 397, 3Op.
Garrett D.E. (1970)
The chemistry and origin of potash
deposits. In: Rau, J.L. and Dellwig L.F. (eds.) 3rd
Symposium on Salt, Northern Ohio Geological Society,
Cleveland, Ohio, pp. 211-222.
Gussow W.C. (1954)
Differential entrapment of oil and gas:
a fundamental principle. Bull. Am. Ass. Petrol. Geol.
38, 816-853.
Hanor J.S. (1982)
Fifty years of development of thought on
the origin and evolution of subsurface sedimentary
brines.
In: S.J. Boardman (ed.) Revolution in Earth
Sciences:
Advances
in
the
Past
Half-Centurv.
Kendall/Hunt, Dubuque, IA. pp. 99-111.

264
Hanor J.S. (1984) Variations in the chemical composition of
oil-field brines with depth in northern Lousiana and
southern Arkansas:
implications for mechanisms and
rates of mass transport and diagenetic reaction.
Transactions-Gulf
Coast
Association
of
Geological
Societies XXXIV, 55-61.
Harvie C.E. and Weare J.H. (1980) The prediction of mineral
solubilities in natural waters: the Na-K-Mg-Ca-Cl-S04H20 system from zero to high concentrations at 25°C.
Geochim. Cosmochim. Acta. 44, 981-997.
Hawley J.E. and Pytkowicz R.M. (1973)
Interpretation of pH
measurements in concentrated electrolyte solutions.
Marine Chem. 1, 245-250.
Hem J.D. and Crooper W.H. (1959)
Survey of Ferrous-Ferric
Chemical Equilibria and Redox Potentials. U.S.Geol.
Surv. Water-Supply Paper 1459-A.
Hitchon B. and Friedman I . (1969)
Geochemistry and origin
of formation waters in the Western Canada sedimentary
basin:I . Stable isotopes of hydrogen and oxygen.
Geochim. Cosmochim. Acta. 33, 1321-1349.
Hitchon
B.M.,
Billings
G.K.
and
Klovan
J.E.
(1971)
Geochemistry and origin of fromation waters in the
western
Candada
sedimentary
basin-III.
Factors
controlling chemical composition.
Geochim.
Cosmochim.
Acta. 35, 567-598.
Hogarth C.G. (1985)
Conodont color alteration, organic
metamorphism,
and thermal history of the Trenton
Formation,
Michigan
basin.
unpub1.
M.Sc.
thesis,
Michigan State University,
East Lansing,
Michigan,
125p.
Holser W.T.
(1979)
Trace elements
and
isotopes
in
evaporites.
In: R.G. Burns (ed.) Marine Minerals.
Miner. Assoc. Amer. Short Course Notes Volume 6, pp.
295-346.
Hower J . , Eslinger E.V., Hower M.E. and Perry E .A. (1976)
Mechanism
of
burial
metamorphism
of
argillaceous
sediment: 1. mineralogical and chemical evidence. Geol.
Soc. Amer. Bull. 87, 725-737.

265

Hull

A.B.
(1987)
Kinetics of evaporite mineral-brine
interactions: mathematical modeling and experimental
determination of the effect of gamma radiation and
threshold crystallization ii hibition on Permian basin
brine
composition.
unpub.
Ph.D.
dissertation,
Northwestern University, Evanston 111., 221p.

Joreskog
K.G.,
Klovan
J.E.
and
Reymont R.A.
(1976)
Geochemical
factor
Analysis.
In: Methods
in
Geomathematics. vol.l, Elsevier Scientific, New York.
Kastner
M.
(1981)
Authigenic
silicates
in
deep-sea
sediments— formation and diagenesis. In: C. Emiliani
(ed.) The Sea, volume 7: The Oceanic Lithosphere. John
Wiley and Sons, pp.
Kharaka Y.K. and Berry F.A.F. (1973)
Simultaneous flow of
water and solutes through geological membranes - I .
experimental investigation. Geochem. Cosmochim. Acta
37, 2577-2603.
Kharaka Y.K., Berry F.A.F. and Friedman I . (1973)
Isotopic
composition of oil-field brines from Kettleman North
Dome, California,
and their geologic implications.
Geochim. Cosmochim. Acta. 37, 1899-1908.
Kharaka Y.K. and Berry
geological membranes
waters from Miocene
in California. Water

F.A.F. (1974)
The influence of
on the geochemistry of subsurface
sediments at Kettleman North Dome
Resour. Res. 10, 313-327.

Kharaka Y.K., Lico M.S. and Carothers W.W. (1980)
Predicted
corrosion
and
scale
formation
properties
of
geopressurred geothermal waters from Northern Gulf of
Mexico basin. J. Petrol. Tech. 45, 319-324.
Klovan J.E. (1975) R and Q-mode factor analysis. In: R.B.
McCammon (ed.)
Concepts in Geostatistics. SpringerVerlag, New York, pp. 21-69.
Knauth L.P. and Beeunas M.A. (1986)
Isotope geochemistry of
fluid inclusions in Permian halite with implications
for the isotopic history of ocean water and origin of
saline formation waters. Geochim. Cosmochim. Acta. 50,
419-434.

266

Koch

G.S. and Link R.F. (1971) Statistical Analysis of
Geological Data. Dover Publications, New York, 438p.

Kozin A.N. (1960)
Geochemistry of bromide and iodine of
formation waters in the Kuybyshev area of the Volga.
Pet. Geol. 4, 110-113.
Kramer
J.R.
(1969)
Subsurface
brines
equilibria. Chem. Geol. 4, 37-50.

and

mineral

Krauskopf K.B. (1979) Introduction To Geochemistry. McGrawHill, New York,617p.
Krejci-Graf K. (1963) Uber rumanische olfeldwasser. Geol.
Mitt. Hydrolgeol. Hydrochem. 2, 351-392.
Kunasz I.A. (1970)
Significance of lamination in the Upper
Silurian Evaporite deposits of the Michigan basin.
In:
Rau J.L. and Dellwig L.F (eds.) 3rd Symposium on Salt.
Northern Ohio Geological Society, Cleveland, Ohio, pp.
67-77.
Land

L.S.
(1980)
The
isotopic
and
trace
element
geochem ist ry of dolomite: the state of the art. In:
Concepts and Models of Dolomitization. SEPM special
publication 28. Society of Economic Paleontologists and
Mineralogist, Tulsa Ok., pp. 87-110.

Land L.S. (1987)
The major ion chemistry of saline brines
in sedimentary basins. Second international symposium
on physics and chemistry of porous media, SchlumbergerDoll, Ridgefield, Connecticut: American Institute of
Physics conference proceedings.
Land

L.S. and Prezbindowski D.R. (1981) The origin and
evolution of saline formation water, Lower Cretaceous
carbonates, south-central Texas. J. Hydrol. 54, 51-74.

Landes KK., Ehlers G.M. and Stanley G. (1945)
Geology of
the Mackinaw Straights area and the subsurfce geology
of Northern Lower Michigan.
Mich. Geol. Sur. Pub. 44,
Geological Series 30, 204p.

267
Lane A.C. (1899)
Lower Michigan Mineral Waters: A study
into the connection between their chemical composition
and mode of occurence.
U.S. Geol. Surv. Water-Supply
Paper 30.
Langmuir D. (1971)
Eh-pH Determinations.
In: R.E. Carver
fed.) Procedures in Sedimentary Petrology. John Wiley
& Sons, pp. 597-635.
Lerman A.
(1970)
Chemical equilibria and evolution
chloride brines. Mineral. Soc. Amer. Spec. Paper
291-306.
Lico

of
3,

M.S., Kharaka Y.F., Carothers W. W. and Wright V. A.
(1982)
Methods
for
collection
and
analysis
of
geopressured geothermal and oil field waters. U.S.
Geol. Sur. Water-Supply Paper 2194, 2Op.

Lloyd R.M. (1966)
Oxygen isotope enrichment of seawater by
evaporation. Geochim. Cosmochim. Acta. 30, 801-814.
Long D.T and Angino E.E. (1982) The mobilization of selected
trace metals from shales by aqueous solutions: Effects
of temperature and ionic strength. Econ. Geol. 77, 646652.
Long D.T., Rezabek D.H., Takacs M.J. and Wilson, T.P. (1986)
Geochemistry of Groundwaters: Bay County, Michigan.
Michigan Department of Public Health: ORD 38553, 3
volumes, 30Op.
Long D.T., Wilson T.P., Takacs M.J. and Rezabek D.H. (1988)
Stable-isotope geochemistry
of saline
near-surface
ground water: East-central Michigan basin.
Geol. Soc.
Amer. Bull. 100, 1568-1577.
Lounsbury R.W.
(1963)
Clay mineralogy of the Salina
formation, Detroit Michigan. In: Rau J.L. and Dellwig
L.F.
(eds.) 2nd Symposium on Salt. Northern Ohio
Geological Society, Cleveland, Ohio, pp. 56-63.
Matthews R.D. (1970) The distribution of Silurian potash in
the Michigan basin.
In:Proceedings of 6th Annual Forum
on Geology of Industrial Minerals: Michigan Geological
Survey, Misc. 1, pp.20-33.

268

Matthews R.D. and Egleson G.C. (1974) Origin and implication
of a mid-basin potash facies in the Salina salt of
Michigan. In: Rau J.L. and Dellwig L.F. (eds.) Fourth
Symposium on Salt. Northern Ohio Geological Society,
Cleveland, Ohio, pp. 15-34.
Matthews R.D. (1977)
Evaporite cycles and lithofacies in
Lucas
Formation,
Detroit
River
Group,
Devonian,
Midland, Michigan.
In: J. H Fisher (ed.) Reefs and
Evaporites-Concepts and Depositional Models. Studies in
Geology no. 5, Amer. Assoc. Petrol. Geol., Tulsa, Ok.,
150p.
McCaffrey M.A., Lazar B. and Holland H.D.
(1988)
The
evaporation path of seawater and the coprecipitation of
Br and K with halite, unpub1. manuscript.
McNutt R.H., Frape S.K. and Dollar P.
(1987)
A strontium,
oxygen and hydrogen isotopic composition of brines,
Michigan and Appalachian Basins, Ontario and Michigan.
Appl. Geochem. 2, 495-505.
Merino E. (1975)
Diagenesis in teriary sandstones from
Kettleman
North
Dome,
California-II.
Interstitial
solutions: distribution of aqueous species at 100°C and
chemical relation to diagenetic mineralogy. Geochim.
Cosmochim. Acta. 39, 1629-1645.
Millero F. J. (1979)
The thermodynamics of the carbonate
system in seawater. Geochim. Cosmochim. Acta. 43, 16511661.
Montgomery S.L. (ed.) (1984) Michigan basin:
deep frontier. Petroleum Frontiers, 80p.
NACSI

Exploring the

(1985) Groundwater Quality Data Analysis, Technical
Bulletin no.
462.
National Council of the Paper
Industry for Air and Strem Improvement, Inc. New York,
197p.

Nordstrom D.K., Jenne E. A. and Ball J.W. (1979)
REDOX
equilibria of iron in acid mine waters.
In E.A. Jenne
(ed .) Chemical Modeling in Aqueous Systems: Speciation.
Sorption. Solubility and Kinetics. Am. Chem. Soc. Symp.
Ser. 93, pp. 51-79.

269
Nowak

R.P.
(1978)
Clay mineralogy
of
pre-Coldwater
(Mississippian) argillaceous sediments in the StateFoster number 1 well, Ogemaw County, Michigan.
unpub.
M.Sc. thesis, Michigan State University, East Lansing,
Michigan, 66p.

Nunn

J.A., Sleep N.H. and Moore W.E.
(1984)
Thermal
subsidence and generation of hydrocarbons in Michigan
basin.
Bull. Am. Ass. Petrol. Geol. 68, 296-315.

Nurmi

R.D. and Friedman G.M. (1977)
Sedimentology and
depositional environments of basin-center evaporites,
Lower Salina Group (Upper Silurian), Michigan basin.
In: Fisher J.H. ed.. Reefs and Evaoorites-Concepts and
Depostional Models. Studies in Geology no. 5, The Am.
Ass. Petrol. Geol,, Tulsa Ok., pp. 150-155.

O'Neil
J.R.
and
Epstein
S.
(1966)
Oxygen
isotope
fractionation in the system dolomite-calcite-carbon
dioxide. Science 152, 198-201.
Perry E.A. and Hower J. (1970)
Burial diagenesis in Gulf
Coast pelitic sediments. Clays and Clay Minerals 18,
165-177.
Pierre C. (1982) Teneurs en isotopes stables (180, 2H, 13C,
34S) et conditions de gene'se des e'vaporites marines:
applications
a'
quelques
milieux
actuels
et
au
Messinien de las M e 'diterrane'e . unpub. French State
doct. thesis, Univ. paris-Sud, Orsay, 266p.
Pierre C . , Ortlieb L. and Person A.
(1984) Supratidal
evaporitic
do;omite
at
Ojo
de
Liebre
lagoon:
mineralogical
and
isotopic
arguments
for
primary
crystallization. J. Sed. Pet. 54, 1049-1061.
Pitzer K.S.
(1973)
Thermodynamics of Electrolytes I :
Theoretical basis and general equations. Jour. Phys.
Chem. 77, 268-277.
Plummer L.N., Parkhurst D.L., Fleming G.W. and Dunkle S .A.
(1989) PHRQPITZ, A computer program for geochemical
calculations in brines.
U.S. Geol. Sur. Water-Res.
Inv. Report 88-4153, 310p.

270

Popp B.N., Anderson T.F. and Sandberg P.H. (1983) Textural,
elemental, and isotopic variations among constituents
in middle Devonian limestones, North America. J. Sed.
Pet. 56, 715-727.
Presely B.J. (1971), Part I : Determination of selected minor
and major inorganic constituents, in:. Ewing, J.I. and
others, Initial Report of the Deep Sea Drilling
Project, Volume III, Part 2. U.S. Govt.
Printing
Office Washington, D.C., pp. 1749-1755.
Rittenhouse G. (1967)
Bromide in oilfield waters and its
use in determining possibilities of origins of these
waters. Bull. Am. Ass. Petrol. Geol. 51, 2430-2440.
Roedder E. (1967)
Fluid incluisions as samples of ore
fluids.
In:
H.L.
Barnes
(ed.)
Geochemistry
of
Hydrothermal Ore Deposits. Holt, Rinehart and Winston,
Inc., pp. 515-574.
Russell W.L. (1933)
Subsurface concentration of chloride
brines. Bull. Am. Ass. Petrol. Geol. 17, 1213-1228.
Sass E. and Starinsky A. (1979)
Behaviour of strontium in
subsurface calcium chloride brines: southern Israel and
Dead Sea rift valley. Geochim. Cosmochim. Acta. 43,
885-895.
Sawtelle E.R. (1958)
The origin of the Berea sandstone in
the Michigan basin.
unpub. M.Sc. thesis, Michigan
State University. East Lansing, Michigan, 92p.
Schrink D.R. (1965) Determination of silica in sea water
using solvent extraction. Anal. Chem. 37, 764-765.
Shaw B. (1975) Geology of the Ablion-Scipio Trend, southern
Michigan.
unpub. M.Sc. Thesis, University of Michigan.
Ann Arbor, Michigan, 63p.
Sofer Z . and Gat J.R. (1972)
Activities and concentrations
of oxygen-18 in concentrated aqueous salt solutions:
analytical and geophysical implications. Earth Planet.
Sci. Lett. 15, 232-238.

271

Sofer Z. and Gat J.R. (1975)
The isotopic composition of
evaporating brines: effects of the isotopic activity
ratio in saline solutions. Earth Planet. Sci. Lett. 26,
179-186.
Sorensen H.O. and Segall R.T. (1975)
Natural brines of the
Detroit River group, Michigan basin. In: Rau J.L. and
Dellwig L.F. (eds.) Fourth symposium on salt. Nothern
Ohio Geological Society, Inc., Clevand Ohio, pp. 91-99.
Spencer R.J. (1987)
Origin of
formations, western Canada
Geochem. 2, 373-384.

Ca-Cl brines in Devonian
sedimentary basin. Appl.

Stoessell R.K. and Carpenter A.B. (1986)
Stoichiometric
saturation tests of NaCl1_xBrx and KCl1_xBrx . Geochim.
Cosmochim. Acta. 50, 1465-1474.
Stoessell R.K. and Moore C.K. (1983) Chemical constraints
and origins of four groups of gulf coast reservoir
fluids. Bull. Am. Ass. Petrol. Geol. 31, 391-396.
Stueber A.M., Pushkar P. and Hetherington E.A. (1984)
A
strontium isotopic study of Smackover brines and
associated
solids,
southern
Arkansas.
Geochim.
Cosmochim. Acta. 48, 1637-1649.
Stumm W.S. and Morgan J.J (1981)
Aquatic Chemistry; An
Introduction Emphasizing Chemical Equilibria in Natural
Waters. Wiley, 78Op.
Taylor
T.R.
(1982)
Petrographic
and
geochemical
characteristics of dolomite types and the origin of
ferroan dolomite in the Trenton Formation, Michigan
Basin.
unpub. Ph.D.
Dissertation, Michigan State
University. East Lansing, Michigan, 75p.
Torrey P.D. (1966)
The discovery of interstitial water.
Production Monthly 30, 8-12.
Van Everdingen R.O. (1968)
Studies of foramtion waters in
western Canada: geochemistry and hydrodynamics. Can. J.
Earth Sci. 5, 523-543.

272

Vugrinovich R. (1986)
Patterns of regional subsurface fluid
movements in the Michigan basin.
Michigan Geological
Survey Open File report OFR-86-6, 27p.
White

D.E.
(1957)
Magmatic,
connate, and metamorphic
waters. Geol. Soc. Amer. Bull. 68, 1659-1682.

White D.E. (1965)
Saline waters of sedimentary rocks.
In:
(A. Young and J.E. Galley, eds.)
Fluids in Subsurface
Environments. Amer. Assoc. Petrol. Geol. Mem. 4, pp.
342-336.
Willey L.M., Kharaka Y.K., Presser T .S .e Rapp J.B. and
Barnes I. (1975) Short chained aliphatic anions in oil
field waters and their contribution to the measured
alkalinity. Geochim. Cosmochim. Acta. 39, 1707-1711.
Wilson T.P. and Long D.T. (1984) The behavior of bromide
during the dissolution of halite at 25°C and 1 atm.
(abs.) G.S.A. Abstracts with Programs, 677p.
Wilson T.P. and Long D.T.
(1986)
Constraints on the
evolution of the Michigan Basin Brines. Geol. Soc.
Amer. Abs. with Program 18, 791.
Wood W.W. (1981) Guidlines for collection and field analysis
of
ground-water
samples
for
selected
unstable
constituents. U.S. Geol. Surv. Tech. of Water-Res.
Inv., chapter D-2, book 1, 48p.
Zherebtsova I.K. and Volkova N.N. (1966) Experimental study
of behavior of trace elements in the process of natural
solar evaporation of Black Sea water and Sasyk-sivash
brine. Geochem. Intern. 3, 656-670.