‘. ~ «um-san-L‘Kw-mq 1 -\ 1 . v11; .f:.,:.t.'. . .“_‘ u f. 1. (31.13.11 "1 1. 11.1... 1 ~ 1 £1" 1: 1. ..~. - LL11. -1. -* 1 1.1.1.1".1 1 . .1.»n..~11\1\.»- v .1: 111:.1 3.1 1.7.1....“1'1. 1. '- 1_~ 131. 111.11.,_. 1 . m _ a; 13:52“ .1; 1 1 . 1.1,1 “1211‘ 1.1 , 11,. --;11..,.,1 ’ 111‘ 13;,“1111H ”" ’I"'.' .51;11..,1.1'1 :‘e'n 111-1 11 1.11:1: , r .1 1 111.1 1.", ,1.... u..." 1 111—1.”. . 1 1,.1'1’J11," 11‘ .1 . 1-11. 1.1 .1... .11....» -1-1--1 ... 1:11.111 in": H' : llllllllllllllllllllllllllglm 312930 This is to certify that the thesis entitled The Geochemistry and Isotopic Chemistry of Saline Ground Water Derived From Near-Surface Deposits of the Saginaw Lowland, Michigan Basin presented by Laura Santina Badalamenti has been accepted towards fulfillment of the requirements for M.S. degree in Geological Sciences WWWXM J Major professor Date gm/qy 9c; / 0-7639 MSU i: an Affirmative Action/Equal Opportunity Institution LIBRARY Michigan State Untversity PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE c:\ctrc\detedw.prn3-p.1 THE GEOCHEMISTRY AND ISOTOPIC CHEMISTRY OF SALINE GROUND WATER DERIVED FROM NEAR-SURFACE DEPOSITS OF THE SAGINAW LOWLAND, MICHIGAN BASIN BY Laura Santina Badalamenti A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Geological Sciences 1992 5a,?» 81/: '2’ ‘7 ABSTRACT THE GEOCHEMISTRY AND ISOTOPIC CHEMISTRY OF SALINE GROUND WATER DERIVED FROM NEAR-SURFACE DEPOSITS OF THE SAGINAW LOWLAND, MICHIGAN BASIN BY Laura Santina Badalamenti Near—surface ground water in the Saginaw Lowland is characterized by relatively high salinities and light isotopic values of del oxygen—18 and del deuterium. The source of the salinity and isotopically light water and the cause of their occurrence is unclear. It is hypothesized (1) that the source of the isotopically light water is meteoric water that recharged the system when the climate was much cooler (perhaps glacial meltwater of the Pleistocene Epoch); (2) that the source of salinity is the upward advection or diffusion of brines; and (3) that their occurrence is due to slow flushing of the water in argillaceous deposits by recent meteoric water. Isotopic ranges in the ground water were as follows: del oxygen—18, —8.7 0/00 to —l7.9l o/oo; del deuterium, - 56.5 o/oo to —126.7 0/00; and del carbon—13 —7.9 0/00 to — 15.4 0/00. Del oxygen—18 and del deuterium plot along the meteoric-water line, ranging from modern-day meteoric to extremely light values. The range of dissolved solids is 145 to 13115 mg/l. It is concluded that the argillaceous deposits, which act as aquitards, retain meteoric water formed at cooler temperatures and brines which have diffused or advected upward. These conclusions are consistent with the fact that the Saginaw Lowland regional ground—water system is a discharge zone. In addition, the outer perimeter of the Saginaw Lowland is a transition zone to fresher, modern—day meteoric recharge. ACKNOWLEDGEMENTS I wish to thank Dr. David Long, Dr. Grahame Larson and Dr. Michael Velbel for their guidance and support during this project and for reviewing this thesis. I thank Dr. Duncan Sibley for his assistance during the defense of the thesis. I express much appreciation to Rick Mandle and Gary Dannemiller, formerly of the United States Geological Survey, for their assistance with field work, analytical work, and data evaluation. I also acknowledge the United States Geological Survey for their financial support of this project. iv TABLE OF CONTENTS LIST OF TABLES Vii LIST OF FIGURES viii INTRODUCTION 1 SECTION 1: Definition of the Problem and Past Works 3 1.1 Definition of the Problem 3 1.2 Past Works 4 SECTION 2 : STUDY AREA ' 10 2.1 Location of Study Area 10 2.2 Geology 10 2.2.1 Geology of the Glacial Deposits 11 2.2.2 Geology of the Bedrock Deposits 17 2.3 Hydrogeology 21 SECTION 3: METHODOLOGY 34 3.1 Ground—Water Sampling Sites 34 3.2 Sample Collection 37 3.3 Analytic Techniques 41 3.4 Data Reduction 43 SECTION 4: RESULTS AND DISCUSSION 45 4.1 Chemical Characterization 45 4.1.1 Student's t—Tests for Comparison of Drift and Bedrock Geochemistry 45 4.1.2 Multivariate Factor Analysis of the Geochemical Data Set 61 v 4 . l . 3 Graphical Analyses 4.1.4 . Chemical Modeling 4.2 Oxygen and Deuterium Isotopes 4.3 Carbon Isotopes SECTION 5 : SUMMARY AND CONCLUSIONS 5 . 1 Summary 5.2 Conclusions APPENDIX Aluminum Speciation Investigation BIBLIOGRAPHY vi 68 76 87 96 102 102 107 110 119 Table Table Table Table Table Table LIST OF TABLES Parameters and methods of chemical analysis. Summary of analytical results of ground— water samples collected from drift deposits in the Saginaw Lowland. Summary of analytical results of ground— water samples collected from bedrock deposits in the Saginaw Lowland. Comparison of Saginaw Lowland and Bay County drift and bedrock ground—water chemistries via the Student's t-Test. Test was conducted at the 90% confidence level, with the null hypothesis:HO: difference = Laboratory analytical results of 1987 ground water sampling for aluminum Speciation investigation. Saturation indicies for various aluminosilicate minerals, 1987 aluminum Speciation investigation. 42 46 48 51 111 115 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 1.1 LIST OF FIGURES Surficial geology of the Michigan basin, with locations of the Saginaw Lowland, simcoe and Greater Lansing study areas. Location of the Saginaw Lowland study area, with generalized bedrock geology. Stratigraphic cross section A—A'. Stratigraphic cross section B—B'. Stratigraphic cross section C—C'. Locations of stratigraphic cross sections. Chloride distribution in the drift deposits of the Saginaw Lowland. Chloride distribution in the bedrock deposits of the Saginaw Lowland. Water table elevations of the Saginaw Lowland. Potentiometric surface of the drift deposits of the Saginaw Lowland. Potentiometric surface of the bedrock deposits of the Saginaw Lowland. Residual potentiometric surface. Locations of drift and bedrock ground- water sample sites. Locations of ground—water sample sites, with site numbers. Frequency histogram of calcium concentrations. Frequency histogram of magnesium concentrations. viii 12 13 14 15 23 24 25 27 28 32 35 36 52 52 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 4.3 4.4 4.5 4.6 4.7 fish- 4. 4. 4. 4. 4 4. .10 .11 .12 .13 14 .15 16 17 18 .19 20 Frequency histogram concentrations. Frequency histogram concentrations. Frequency histogram concentrations. Frequency histogram values. Frequency histogram concentrations. Frequency histogram Frequency histogram concentrations. Frequency histogram evaporation values. Frequency histogram concentrations. Frequency histogram concentrations. Frequency histogram concentrations. Frequency histogram concentrations. Frequency histogram concentrations. of of of of potassium sodium chloride alkalinity sulfate pH values. silica residuals of deuterium oxygen—18 dissolved oxygen bromide strontium Diagram of R-mode factor loadings. Ternary diagram of ground—water chemistry in drift deposits. Ternary diagram of ground-water chemistry in bedrock deposits. Histogram of chloride to bromide concentration ratios. Sodium versus Chloride molar concentrations. ix 53 53 54 54 55 55 56 56 57 57 58 59 6O 65 7O 71 75 77 Figure 0 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 4.21 4.22 4.25 4.26 4.27 4.30 4.31 4.32 4.33 4.34 Log chloride concentration versus oxygen-18 concentration in the drift deposits. Log chloride concentration versus oxygen-18 concentration in the bedrock deposits. Frequency histogram of calcite saturation indices. Frequency histogram saturation indices. Frequency histogram saturation indices. Frequency histogram indices. Frequency histogram indices. Frequency histogram indices. Frequency histogram saturation indices. of aragonite dolomite gypsum saturation halite saturation quartz saturation chalcedony Mineral stability diagram of potassium aluminosilicate minerals. Mineral stability diagram of sodium aluminosilicate minerals. Deuterium concentration versus oxygen—18 concentration. Distribution of oxygen—l8 concentrations in the drift deposits. Distribution of oxygen-18 concentrations in the bedrock deposits. Histogram of carbon-13 concentrations. Carbon-13 concentration versus depth. 78 78 81 81 82 82 83 83 84 86 86 88 93 94 98 100 Figure 9 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 4.21 4.22 4.23 4.24 4.25 4.26 4.29 4.30 4.31 4.32 4.33 4.34 4.35 4.36 Log chloride concentration versus oxygen~18 concentration in the drift deposits. Log chloride concentration versus oxygen-18 concentration in the bedrock deposits. Frequency histogram of calcite saturation indices. Frequency histogram of aragonite saturation indices. Frequency histogram of dolomite saturation indices. Frequency histogram of gypsum saturation indices. Frequency histogram of halite saturation indices. Frequency histogram of quartz saturation indices. Frequency histogram of chalcedony saturation indices. Mineral stability diagram of potassium aluminosilicate minerals. Mineral stability diagram of sodium aluminosilicate minerals. Deuterium concentration versus oxygen-18 concentration. Distribution of oxygen—18 concentrations in the drift deposits. Distribution of oxygen—18 concentrations in the bedrock deposits. Histogram of carbon—13 concentrations. Carbon—13 concentration versus depth. 78 78 81 81 82 82 83 83 84 86 86 88 93 94 98 100 INTRODUCTION This study examines the extent of isotopically—depleted ground water in the Saginaw Lowland of the Michigan basin and investigates several hypotheses based primarily on previous work of Long et al. (19885 in Bay County of the Saginaw Lowland. The occurrence of isotopically-depleted ground water is compared to regional lithology and ground- water geochemistry in order to test the following hypotheses regarding characteristics of the ground water system in the Saginaw Lowland: Hypothesis 1: the ground water system contains isotopically-depleted meteoric water that recharged the system when the climate was much cooler (perhaps glacial meltwater of the Pleistocene Epoch); Hypothesis 2: the ground water system contains saline water derived from brines deep in the Michigan basin; and Hypothesis 3: the isotopically-depleted water and saline water are retained in the ground water system by the combined affects of aquitards, low regional horizontal hydraulic gradients, and a regional ground—water discharge zone. This study also attempted to verify that the additional processes identified by Long et a1. (1986, 1988) in the Bay County ground water system are also affecting the Saginaw Lowland system. The additional processes include water-rock interactions including calcium carbonate equilibrium and stability with potassium aluminosilicate minerals; and biological reduction of sulfate. The study was also used to estimate whether the data reduction techniques used in the Bay County study (Long et al., 1986, 1988), will produce interpretable results on more widely-spaced data points. SECTION 1 — DEFINITION OF THE PROBLEM AND PAST WORKS 1.1 Definition of the Problem Long et al. (1986, 1988) hypothesized that the following processes are occurring in the near-surface ground water system of Bay County, Michigan: 1) upward diffusion of local formation brines into the near-surface ground water and mixing of the brines with surface water as suggested by ternary diagrams of major cations and anions; 2) water-rock interactions, including calcium carbonate equilibrium as suggested by calcite saturation indices, and stability with the potassium aluminosilicate minerals, as suggested by a potassium aluminosilicate mineral activity plot of ground— water samples with points clustered along the boundary of muscovite and potassium feldspar; 3) the biological reduction of sulfate within the drift and bedrock, as suggested by the correlation of decreased sulfate (due to consumption by microbes) and increased del sulfur—34 values (due to selective assimilation of sulfur-32 by microbes resulting in isotopic enrichment of sulfur-34 in the ground water); and 4) the mixing of present-day meteoric water with meteoric water formed at cooler temperatures (possibly Pleistocene in age) as suggested by oxygen and deuterium stable isotopic values. This study examines the extent of isotopically-depleted ground water in the Saginaw Lowland of the Michigan basin, which encompasses Bay County (investigated by Long et al., 1988). The occurrence of isotopically-depleted ground water is compared to regional lithology and ground-water geochemistry in order to test several hypotheses outlined in the Introduction. This study evaluates a large geochemical data set with the aid of statistical, graphical and computer modeling techniques of data reduction. The statistical data reduction techniques are the Student's t-Test and multivariate factor analysis. The graphical data reduction techniques include ternary diagrams, frequency histograms, and x-y plots. The computer modeling technique used is chemical modeling with WATEQF. The research allows the comparison of interpretations made on ground-water processes from a data set composed of a large number of samples (about four hundred) taken in a one-county area with a data set composed of a small number of samples (about one hundred) taken from a seven county area. The extent of sulfate reduction in the Saginaw Lowland is briefly examined. 1.2 Past Works Stable isotopes of oxygen and deuterium give insight into the origin of ground water, and reflect physical and Chemical processes which have affected the ground water. In low-salinity ground water, oxygen—18 and deuterium concentrations are conservative; they are unaffected by physical, chemical or geological processes. However, in the hydrosphere, oxygen-18 and deuterium concentrations are not conservative (Frape and Fritz, 1982). Craig (1961 a) demonstrated a linear relationship between concentrations of oxygen-18 and deuterium in meteoric waters which have not undergone large amounts of evaporation. Samples of meteoric water will plot along the meteoric water line (MWL) as described by Craig (1961 a). The shallow geologic deposits in the Saginaw Lowland area of the east-central Michigan basin (Figure 1.1) yield isotopically-depleted, saline ground water, as reported by Badalamenti et al. (1988) and Long et al. (1988). Long et al. (1988) found that ground water from the glacial deposits and near-surface bedrock in Bay County, Saginaw Lowland (see Figure 1.2 for location) has del oxygen—18 values ranging from -6.93 0/00 to -18.46 0/00; and del deuterium values ranging from -47.3 0/00 to -l37.4 0/00. The concentrations of oxygen-l8 and deuterium are reported according to the standard of Craig (1961 b). These values plot along the MWL of Craig (1961 a) indicating meteoric origin of the ground water. They interpret these data to indicate that during the Pleistocene epoch, isotopically-light meteoric water recharged the system, and has since mixed with modern-day meteoric water producing a variable isotopic composition in the ground water. These observations and conclusions are consistent with conclusions of Desaulniers et al. (1981) (study area location shown on Figure 1.1). I I r I I I I I \ L4 \ \ “In... <., .h Figure 1.1 .zwoaomm xooucmn beneaouocom cud; .mwum, >c=um ccoazoq 3ocfimmm ecu mo ceaumooq ~.H madman m I‘ll Oil d z Esmw _ 3.53 _ oiow Ex cow 22255 5:58 oo=E OOr SOOmDP 23m 0250:). >03._.m . Z_m_ The shallow geologic deposits in the Saginaw Lowland, primarily lacustrine sediments and glacial drift, are characterized as aquitards (see Figure 1.1 for distribution of the aquitards). The ground water in these deposits has total dissolved solids concentrations of up to 53,000 mg/L (Long et al., 1986). The ratios of bromide to chloride concentrations for ground water samples compared to samples of meteoric water and local brines suggest that the ground water samples have been geochemically influenced by brines. The upward migration of brines from deeper in the Michigan basin via diffusion or advection may be producing salinity in the ground water (Long et al., 1988). Isotopically-light ground water has been identified in aquifers by Clayton et al. (1966); Fritz et al. (1974); Perry et al. (1982); and Siegel and Mandle (1984). The origin of isotopically-light ground water in aquifers has been interpreted as recharge during a period when the climate was cooler than present day. Isotopically-light ground water is frequently found in association with near-surface aquitards. The Quaternary argillaceous deposits of the Great Lakes region, considered aquitards, produce ground water significantly depleted in isotopes compared with modern—day meteoric water (Desaulniers et al., 1981; and Bradbury, 1984). Although the role of aquitards in affecting the hydrogeochemistry of ground water is becoming understood (Back, 1986), their role in controlling the isotopic composition of ground water has been little studied. Aquitards have low hydraulic conductivity values, are composed of fine-grained reactive minerals, and have predominantly vertical ground-water flow direction (Back, 1986). They often produce water of poor quality. For example, the dissolution of reactive minerals in the aquitard deposits can produce mineralized water. Aquitards can control flow patterns of ground water, and thereby determine the distribution of different chemical types of ground water. Long et al. (1988) hypothesize that ground-water salinity and isotopic depletion are due to the effect of aquitards in the ground water system. Modern meteoric recharge to the system is inhibited by the presence of relatively thick glaciolacustrine deposits, impermeable glacial till, and shale; and a low regional ground—water flow gradient. Long et al. hypothesize that the "glacial" water is retained in these deposits, because modern—day recharge water is only slowly "flushing" the system. In addition, they hypothesize that because the ground—water flow is stagnant and because the deposits lie in a discharge zone (potentiometric surfaces indicate upward flow), salinity can build up in the system, as brines move upwards into the system via diffusion or advection. SECTION 2 --STUDY AREA 2.1 Location of Study Area The Saginaw Lowland study area is a region of approximately 14000 square kilometers, located in the east- central Michigan basin, as shown in Figure 1.1. The Saginaw Lowland encompasses the counties of Arenac, Gladwin, Midland, Gratiot, Saginaw, Bay and western Tuscola in the state of Michigan (Figure 1.2). The study area is partially bordered by Saginaw Bay of Lake Huron. 2.2 Geology The near-surface geologic deposits in the Saginaw Lowland are regarded as two units throughout this study. These units are referred to as "drift", corresponding to unconsolidated deposits of primarily glacial origin, and "bedrock", corresponding to the shallow bedrock formations which underlie the unconsolidated deposits or drift. The topography of the Saginaw Lowland is characterized by level plains of lacustrine clay and gently rolling hills of ancient sand bars, beaches and moraines (Leverett and Taylor, 1915; Pringle, 1937; Vanlier, 1963). Figure 1.1 illustrates the generalized surficial geology of the Great Lakes region. The lacustrine deposits denoted in Figure 1.1, which include the Saginaw Lowland, are known to produce saline ground water (Lane, 1899 a; 10 11 Lane, 1899 b; Twenter, 1987). The predominance of clay in regional glacial geology is illustrated in stratigraphic cross section A-A' (Figure 2.1), stratigraphic cross section B-B' (Figure 2.2) and stratigraphic cross section C-C' (Figure 2.3). The drift is primarily composed of clay-rich layers of great thicknesses. The well logs on which these cross—sections are based are provided in Appendix A. The locations of the stratigraphic cross sections are shown in Figure 2.4. 2.2.1 Geology of the Glacial Deposits The drift has a thickness of up to 170 meters in western Gratiot County (Vanlier, 1963) but is not present in portions of northeastern Arenac County (Pringle, 1937). The drift is described as a sequence of stratigraphic units (Vanlier, 1963; Stark and McDonald, 1980; Dow Chemical Company, 1986). These units include a shallow (surficial) sand, glaciolacustrine deposits of clay and sand, a glacial till (morainal or till plain), and an aquifer (primarily sand and gravel outwash) lying in bedrock valleys. The regional stratigraphy of the drift is highly variable. Thicker drift deposits may lie in buried erosional bedrock valleys with reliefs of 30 to 60 meters in which unconsolidated drift material has been deposited (Rhodehamel, 1951; Twenter and Cummings, 1985). The buried bedrock valleys trend northeast to northwest in Bay County Venice. Sc”o AI _ 550 l / 750f t r i-_650 / _450 // HorIzontaI seal. .pp'o‘lm-tol Z no."d a section A_A.. 22 c o O o O :8 m m m m m u h I .. m v n //////////g 14 «00w OMB.III atom .25.; ..otu coauoom mmouo ofinmmumaumuum n.~ ousmwm >nnummMMMm ucoaoa 7.25.3.3.— snum_n.c0u=oz .mcofiuoom mmouo candoumfiumuum uo mcofiumooq v.~ wusmam 23w 15 Z 16 (Twenter and Cummings, 1985). In Bay County, the basal water-bearing formation, which lies below the glacial till, reaches 40 meters in thickness in bedrock valleys and is thinner where the bedrock surface is topographically high (Stark and McDonald, 1980). The composition of the drift, although variable, is primarily composed of illite, kaolinite, chlorite, quartz, feldspar (unspecified), calcite and dolomite (Chittrayanont, 1978). In Bay County the water—bearing formations lying beneath the glacial till consist of poorly—sorted coarse sand, gravel and pebbles with some silt and clay. The sand is quartzose and the gravel consists of sandstone, carbonates, quartzite, chert, and igneous and metamorphic rocks (Stark and McDonald, 1980). The origin of the drift deposits is primarily the result of glacial processes during the Pleistocene age (Vanlier, 1963; Stark and McDonald, 1980). The lithologic description of glacial till in the drift includes unsorted clay, silt, sand,-grave1, cobbles and boulders (Dow Chemical Company, 1986). The glacial till is sometimes referred to as heavy boulder clay, according to Pringle (1937) and Vanlier (1963). Lenses of well—sorted sand and gravel, which are found in the glacial till, originated by glacial meltwater deposition. Lacustrine clays, deltaic deposits, Channel deposits, and beach and dune sands are also present in the drift (Vanlier, 1963; Dow Chemical Company, 1986). 17 The glacial till was deposited as moraines in Arenac, Tuscola, western Gratiot and northwestern Gladwin counties. In some portions of the Saginaw Lowland, glacial till was deposited subaqueously. Some moraines have distinctive reliefs (up to 60 meters in height). Other moraines have slight relief (Leverett and Taylor, 1915; Pringle, 1937; Vanlier, 1963). The moraines were laid in belts which parallel the Saginaw Bay shoreline, as shown in the generalized map of surficial geology (Figure 1.1). The moraines are paralleled by plains of sand and clay, which range from 15 to 30 meters in thickness (Leverett and Taylor, 1915; Pringle, 1937). The sand is typically well— rounded, sorted and quartzose, with a thickness of zero to six meters. The sand was typically deposited in a near— shore environment in proglacial lakes. The clay was typically deposited in glacial lakes and composed of mostly heavy blue clay and silt with some sand, gravel, hardpan (a clay and gravel mix) and pebbles of primarily quartzite, sandstone, siltstone, mafic igneous rocks, granite and carbonates (Pringle, 1937; Stark and McDonald, 1980). 2.2.2 Geology of the Bedrock Deposits The drift lies on top of bedrock formations of the Michigan basin. The Michigan basin is a structural basin, consisting of layers of Paleozoic and younger sedimentary rocks, underlain by Precambrian crystalline rock (see Figure 18 1.2 for generalized bedrock geology). Regional dip is generally southwestward into the basin, with a northwest- southeast strike (Martin, 1958). Smaller scale folds trending northwest-southeast are found in Arenac County (Pringle, 1937) and most likely occur elsewhere in the study area. Kelly (1936) believed that a bedrock low approximately parallel to the axis of Saginaw Bay is evidenced in the Pennsylvanian Saginaw Formation. Kelly postulated that this generally supports the idea of post- Devonian folding which differs in trend from that of pre- Devonian time (Newcombe, 1932). The Saginaw Lowland was developed primarily on the erosional surface of the underlying Pennsylvanian Saginaw Formation. Vanlier (1963) and Rhodehamel (1951) suggest that the Saginaw Lowland was base leveled due to the high resistance of the surrounding sandstone members of the Saginaw Formation and the low resistance of local shale members of the Saginaw Formation, which is predominantly shale. The neighboring bedrock uplands received thick glacial deposits due to the convergence of the Saginaw and Erie ice lobes, forming the Michigan—Saginaw and Huron- Saginaw Uplands as a result (Leverett and Taylor, 1915; Western Michigan University, 1981). Martin (1958) supports this, stating that the central lowland (which includes the Saginaw Lowland) falls between the highlands of the north— central southern peninsula and the highland extending from 19 the Thumb area to the south-central southern peninsula. The topography of the drift is similar to the bedrock topography on a large scale. The youngest bedrock underlying portions the drift is referred to as "red beds", which consist of sandy gypsiferous shale and shaley sandstone with an erosional upper surface. Western Michigan University (1981) states that the red beds are of Jurassic age. The red beds have been completely removed by erosion in major pre-Pleistocene valleys (Vanlier, 1963). The red beds are found in northwestern Gladwin, western Midland and southern Gratiot Counties (Martin, 1936). The red beds are approximately ten meters thick and yield little water (Vanlier, 1963). The major Pennsylvanian formation is the Saginaw Formation. This is the primary formation upon which the drift was deposited. The Saginaw Formation crops out in Arenac, Saginaw and Tuscola counties and underlies all of the region except northeastern Arenac County, where it is absent. The Saginaw Formation consists of lenticular beds of sandstone, shale, coal and limestone (Martin, 1936). The Saginaw Formation has a basin—wide thickness ranging from 0 to at least 163 meters with an average thickness of at least 122 meters (Kelly, 1936). In the Saginaw Lowland, the Saginaw Formation is composed primarily of shale, as shown in stratigraphic cross-sections A-A' (Figure 2.1), B—B' 20 (Figure 2.2) and C-C' (Figure 2.3). The locations of the stratigraphic cross sections are shown in Figure 2.4. Within the Saginaw Lowland, the Saginaw Formation is noted for bituminous coal deposits which were mined in the region from the mid-1800's to 1950 (Cohee et al., 1950; Stark and McDonald, 1980). In Arenac County, the sandstone of the Saginaw Formation is grey and micaceous; and the shale is gray-black and carbonaceous. Marcasite and clay- iron stone concretions with sphalerite are common (Pringle, 1937). Zircon and tourmaline are found in white Saginaw sandstone of Arenac County (Kelly, 1936). Although bedrock geology in the Saginaw Lowland is variable, Stark and McDonald (1980) define four bedrock units within the Saginaw Formation of Bay County. The deepest of these units is a fine-grained, quartzose, well— sorted and well—rounded sandstone of variable thickness, with some thin shale beds. This layer is overlain by a shale unit, consisting of thin beds of shaley quartz—rich sandstone, with some limestone. This shale unit averages 10 meters in thickness. On top of the shale unit is a coal unit of 0.5 meters average thickness. The uppermost bedrock unit is another shale unit, consisting of thin coal seams, quartzose sandstone, and mudstone and siltstone beds. The upper shale unit averages 15 meters in thickness (Stark and McDonald, 1980). 21 Unconformably underlying the Saginaw Formation are the Mississippian Bayport and Michigan Formations, respectively. The Bayport Formation consists of limestone, dolomite and calcareous or argillaceous sandstone. It crops out in Arenac and Saginaw Counties, and ranges in thickness from 12 to 30 meters basin—wide (Martin, 1936; Pringle, 1937). The Michigan Formation is composed primarily of sandstone and shale with beds of gypsum and anhydrite underlain by dolomitic limestone. The formation crops out in Arenac County and has been mined for gypsum. The thickness of the Michigan Formation ranges from 0 to 152 meters basin—wide (Martin 1936; Pringle, 1937; Martin, 1958). Formations deeper than the Michigan Formation were not investigated in this study. 2.3 Hydrogeology The drift and bedrock can each be characterized as aquitards enCompassing discontinuous aquifers. The drift aquitards are composed of thick sequences of lacustrine clay and argillaceous till. The bedrock aquitards are composed of shale. Stratigraphic cross sections A-A' (Figure 2.1), B-B' (Figure 2.2) and C-C' (Figure 2.3) illustrate this. The locations of the stratigraphic cross sections are shown in Figure 2.4. In the drift, the discontinuous aquifers are composed of isolated water-bearing sand and gravel lenses within the glaciolacustrine clay and clay-rich till; and of 22 sand and gravel outwash deposits (found at the base of the till) (Vanlier, 1963; Chittrayanont, 1978). Although the drift and bedrock are each characterized as aquitards, there are chemical and isotopic differences between the two units. These differences will be discussed in detail in subsequent sections. The ground water derived from the drift and bedrock in the Saginaw Lowland is noted for its salinity (Lane, 1899 a; Lane 1899 b; Vanlier, 1963; Twenter, 1987). Figure 2.5 and . Figure 2.6 illustrate the regional distribution of chloride concentrations in ground water derived from these deposits. High chloride concentrations are common in ground water derived from both drift and bedrock deposits in Saginaw Lowland (see Figure 2.5 and Figure 2.6). The Saginaw Lowland is noted for its artesian saline springs (Lane, 1899 a; Lane, 1899 b; Houghton, 1928; Martin, 1958); while the lower bedrock deposits in the region are noted for their brine content (Vanlier, 1963; Western Michigan University, 1981). Bedrock deposits located deeper within the Michigan basin produce brines concentrated to an unusually high degree (Case, 1945; Egleson and Querio, 1969; Sorenson and Segall, 1975). The deeper bedrock formations are also sources of oil and gas (Vanlier, 1963). Regional water table elevations are shown on a contour map (Figure 2.7) which is derived from a map by Mandle .ccmH3oq Suswmmm on» no muflmonoc uuwuv on» Ca :oHuaafiHumac wcfiuoflno m.N musmah 03:. ->. it.» £33.50 .52 fi\ - 3:20p —||_||l view /.> .74. 23' 24 .caoHZOA 3ocwmmm on» uo muwmoamc xooucon may :« newusafiuumfic wcfiuoHno .¢=::L_____ 2: m.m unease 03-...) . 1:8... 53:00 — \ E.— :— 23w .ccma3oq zucdacm 0:» ac macaum>oao canon noun: n.~ ouzmam «on. 00.. . 1:3:— 53:00 ooh \\11 h(\hll\II/Iéoo m4 8 0 W 00? ~ on 09 "UK or do .K/I/W\ap t\iJ \\\va@ 26 (1989). Elevations are based on water level measurements taken in water wells as recorded by well drillers. The measurements are reported on well logs dated 1969 through 1975. Figure 2.7 accounts for the influence of surface- water levels from major inland lakes and streams and the Great Lakes. Additionally, Figure 2.7 reflects the influence of topography and major stream systems on the water table. The water—bearing formations of the lower units in the drift and of the underlying bedrock are hydrologically confined (Vanlier, 1963; Stark and McDonald, 1980; Dow Chemical Company, 1986). Water levels in the confined formations average 10 meters below land surface (Stark and McDonald, 1980). Figure 2.8 is a contour map of the potentiometric surface of the confined portions of the drift. Figure 2.9 is a contour map of the potentiometric surface of the confined portions of the bedrock. These maps are based on water level measurements as reported on well logs which have been density-corrected. Water level measurements reported on well logs for all drift wells sampled were used and are provided in Appendix B. Water level measurements reported on Well logs for all bedrock wells sampled were used and are provided in Appendix C. Water level measurements reported on additional wells logs, which were not sampled, were 27 Ex 0' — oo=E OP — view .ccma3oq 3ocfiacm ecu no muAmOro uu«nc 02» mo oocuusm ofiuuoEoHucouom w.m ousmfim .03 or . 1:3:— uaoucou — / . w 6 o. 9 m. a... 9 .. © 0 ea . onm one x 9 w. A? 29 selected at random to increase the density of data points for construction of Figure 2.8. These well logs are provided in Appendix D. The density of highly saline ground water in a well may affect its hydraulic head and the interpretation of piezometric potential in a well (Long et al., 1986). Because ground water in the Saginaw Lowland is highly saline, a series of equations described by Gudramovics (1981) was used to calculate water density, accounting for variability due to temperature and salinity. Water level measurements for piezometric surface interpretations in the Saginaw Lowland were then adjusted using the calculated water density. A description of the calculations used to adjust water levels according to water density follows. At a constant temperature, the relationship between density and salinity is linear. Salinity was estimated from the chloride concentration by Equation 1 (Schopf, 1980) as follows: ' Salinity (o/oo) = 1.80655 * (Cl (o/oo)). To calculate the density of water at the measured temperature in a well, the value of the y-intercept and slope was calculated. The y-intercept (density of pure water at a specified temperature) at a temperature (T) was calculated by Equation 2 (Weast, 1979) as follows: 3O - yFintercept = [999.83952 + (16.945176*T) - (7.9870401*10-3*T2) - (46.170461*10-6*T3) + (105.56302*10-9*T4) — (280.54253*10-12*T5)]/{[1 + (0.01687985*T)]*1000}. The slope for the specified temperature was calculated by Equation 3 (Horne, 1969) as follows: ' slope = 8.300245*10-4 - 2.2274915*10-5 * In T. After calculations were completed for salinity, y- intercept and slope for a water sample at the temperature at which it was sampled, the density of the water was calculated by Equation 4 as follows: density = (salinity * slope) + y—intercept. The density for each groundewater sample was calculated using Equations 1 through 4. The density was then multiplied by static water level elevations, resulting in piezometric surface elevations which were contoured as shown in Figure 2.8 and in Figure 2.9. The maps were contoured using SurferTM (Golden Software). Each of the water level contour maps (water table, Figure 2.7; and piezometric surfaces, Figure 2.8 and Figure 2.9) indicate that within the Saginaw Lowland, the horizontal component of ground—water flow is towards Saginaw Bay. 31 The vertical component of ground-water flow is‘ indicated by a residual potentiometric surface map (Figure 2.10) (see also Thorstenson and Fisher, 1979 and Long et al., 1988). The residual potentiometric surface map is the result of subtracting the potentiometric surface of the drift from the potentiometric surface of the bedrock. A positive value of residual potentiometric surface results when the potentiometric surface of the bedrock is greater than that of the drift. Figure 2.10 indicates that ground water discharges from the bedrock into the drift in much of the Saginaw Lowland. Potentiometric surfaces are generally greater (higher) in the bedrock wells than in the drift wells. Additional evidence that ground water tends to discharge from the deeper units into the shallower units is that artesian flowing bedrock wells have been reported throughout the region from as long ago as 1899 (Lane a and b) and 1906 (Cooper, 1906). A horizontal hydraulic gradient can be estimated by observing the vertical change in water level elevation over a specific horizontal distance using Figure 2.7, Figure 2.8 and Figure 2.9. Horizontal hydraulic gradients are estimated to be 0.0001 to 0.006. The horizontal hydraulic gradient decreases in the direction of ground water flow (Chittrayanont, 1978; Stark and McDonald, 1980). Horizontal ground water flow is towards Lake Huron, and locally towards streams and drains. Stark and McDonald (1980) analyzed 33 potentiometric head data from Bay County which indicated a horizontal hydraulic gradient of 0.003. The horizontal hydraulic gradient in Bay County was reported as less than 0.0005 to less than 0.002 by Long et al. (1988). These values for horizontal hydraulic gradients support the statement that the Saginaw Lowland is poorly drained (Stark and McDonald,1980). The local vertical hydraulic gradient in Bay County had a downward component of 0.03 (Stark and McDonald, 1980). SECTION 3 - METHODOLOGY 3.1 Ground-Water Sampling Sites Ground-water samples were collected from 96 domestic and municipal wells during the summer of 1986. Fortyeone of these samples were taken from wells which derive water from drift deposits, referred to as drift wells. Fifty-five samples were obtained from wells which derive water from bedrock deposits, referred to as bedrock wells. The locations of ground-water sampling points (sampled wells) distinguished as drift or bedrock wells are shown in Figure 3.1. The locations of the ground-water sampling points distinguished by study identification number are shown in Figure 3.2. Well logs for all drift wells sampled are provided in Appendix B. Well logs for all bedrock wells sampled are provided in Appendix C. Sample sites were selected on the basis of two primary factors: (1) type of deposit the well was screened or completed in; and (2) location of the well. A well was sampled depending on the type of deposit from which water was derived (drift or bedrock). Locations of sample sites were selected to obtain a relatively even distribution of sample points. Geochemical data from nine wells reported Long et al. (1986) for Bay County were added to the data set, with selection based on location and type of deposits from which water was derived. 34 35 . EQUHW GHQ—flan .ku63 IUCDOHD XOOHUNQ nun—fl new dhmu HO MCOHHMOQH H . n Ohgfih - - - 85:ch =03 goo-wo- . 5:304 9:12.56 C 0 =25 £5 . £2.39.- uSKEcG I O O O I .86... .. - .00.: I I n - - 1.3.9.3 .092 .0045 omém .omdv .oo.vv. 36 .muonasc oufim cu“: .moufim caused nouaatucsoum no mcofiumooq ~.n ounmflm . en 3 shop . on . 1|.— 3 . kn . 5.... 3 me . .v . w... an or 4.. ow. an . 8 6 . 28m eo w... 3mm emu . mm. on nu Q I hf. mm D or. the «w. 3 u .5 .mN 4.“ em VON. a NO 0” QN non vn n” C . nu . n. . an N» mo. . an n. vm . an . o o o N: e on? e an 8 S x: .2 8 . 37 3.2 Sample Collection Wells were pumped and field measurements and water samples were taken after constant water temperature was detected (approximately twenty minutes). Care was taken to avoid the influence of any water treatment on the samples. Field measurements and sample collection were performed according to standard techniques of the United States Geological Survey (U.S.G.S.) as described in Skougstad et al. (1985). Sample bottles were shipped on ice daily by air freight to the U.S.G.S. Laboratory in Denver, Colorado for chemical analysis. The field measurements taken were specific conductance, temperature, pH, dissolved oxygen content and alkalinity. Specific conductance and temperature were recorded using a YSI Model 33 S-C-T meter. Specific conductance values were corrected for temperature by a correction factor equal to 1/{1 + [.02 (temperature - 25)]}. The pH was measured with a Beckman Model §21 pH meter. Dissolved oxygen was measured at the sampling location with a Hach chemical kit and digital titrator, using an azide modification of the Winkler method. A sample was collected by running a hose directly into a biochemical oxygen demand (BOD) bottle. Care was taken to ensure no air bubbles came through the hose. The bottle was immersed in 38 water while being allowed to overflow about three times its volume. The bottle was then glass-stoppered. A premeasured amount of manganous sulfate powder and alkaline iodide azide reagent were added to the BOD bottle. The mixture formed a flocculate precipitate. When this precipitate settled, a premeasured amount of sulfamic acid powder was added. This caused the floc to dissolve, resulting in a yellow colored solution. Two hundred milliliters of the yellow solution were measured into a beaker. Two milliliters of starch indicator solution was added to the beaker, causing a blue color to develop. A Hach digital titrator was used to titrate the blue solution with 0.200 N sodium thiosulfate. Titration was complete upon reaching a colorless solution. The dissolved oxygen value in mg/L oxygen was read directly from the Hach digital titrator, based on the strength of the sodium thiosulfate, the amount of water titrated, and the amount of sodium thiosulfate released with each rotation of the titrator. Alkalinity was measured at the sampling location. 'The determination of alkalinity as calcium carbonate was performed on one hundred milliliters of unfiltered ground water which was collected in a beaker. A Hach digital titrator was used to titrate the water with 1.6 normal sulfuric acid. Titration was complete upon reaching the calcium bicarbonate titration endpoint (approximately 4.5 pH), where all solutes contributing to alkalinity had 39 reacted with the sulfuric acid (Hem, 1985). The alkalinity value in mg/L calcium carbonate was read directly from the Hach digital titrator, based on the strength of the sulfuric acid, the amount of water titrated, and the amount of acid released with each rotation of the titrator. Six sample bottles were filled for laboratory analyses at each sampling site, because the various chemical constituents to be measured required different collection, preservation and analytical techniques, and large sample volumes were needed. All sample bottles were acid—washed and rinsed with distilled water. Additional sample bottles were filled for isotopic analyses, as will be discussed later. Four of the six sample bottles collected for chemical analyses were filled with water filtered with a Geotech backflushing filter assembly. The nitrocellulose filter membrane was 142 millimeters in diameter with a pore size of 0.45 micrometers. The four sample bottle types, constituents analyzed and preservation techniques are listed below: (1) A one—liter acid—washed white plastic high density polyethylene (HDPE) bottle was used to collect a sample for the analysis of dissolved calcium, dissolved magnesium, dissolved sodium, dissolved potassium, dissolved iron, dissolved lithium, dissolved manganese, dissolved strontium, 40 aluminum, arsenic, and zinc. This sample was preserved with two milliliters of nitric acid. (2) A one-liter white plastic bottle (HDPE) was used to collect a sample for the analyses of dissolved sulfate, dissolved chloride, dissolved fluoride, dissolved solids, dissolved silica, dissolved boron, and dissolved bromide. This sample did not require preservation. (3) A two-hundred milliliter brown plastic bottle (HDPE) was used to collect a Sample for the measurement of dissolved nitrite and nitrate nitrogen. This sample was chilled on ice. (4) A two—hundred and fifty milliliter brown plastic bottle (HDPE) was used to collect a sample for the measurement of dissolved ammonia nitrogen. This sample was preserved with 13 milligrams of mercuric chloride and chilled on ice. At each sample location, a fifth sample bottle was filled with water which had been filtered with a Gelman stainless steel filter assembly. The 45 millimeter diameter filter paper was silver-coated. A 125 milliliter glass bottle was used to collect a sample for analysis of dissolved organic carbon content. This sample was chilled on ice. 41 A sixth sample collected at each sampling location was not filtered. A 250 milliliter white plastic bottle (HDPE) was used to collect a sample for analysis of total sulfide. Zinc acetate (0.5 grams) was added to each bottle as a preservative. Additional samples for analysis of stable isotopes of oxygen, deuterium and carbon were collected at the same time as the other samples. Oxygen and deuterium isotope samples were collected, without filtering, in-125 milliliter glass bottles, and were preserved with 26 milligrams of mercuric chloride. Stable carbon isotope samples were collected, without filtering, in one-liter glass bottles, with the addition of 50 milliliters of ammoniacal strontium chloride as a preservative. 3.3 Analytic Techniques All chemical parameters were measured according to the standard methods chemical analysis of the U.S.G.S. as described in Skougstad et al. (1985) and summarized on Table 3.1. Oxygen and deuterium isotope samples were analyzed at the U.S.G.S. Reston Stable Isotope Laboratory in Reston, Virginia. The analytic techniques for analyses of oxygen isotopic ratios included equilibration with carbon dioxide, with a modification of the analytic technique of Epstein and Mayeda (1953). Deuterium analyses were performed with a 42 ’ Table 3 . 1 Parameters and methods of chemical analysis. Bannister AnalyfigMejm Alkalinity, as CaC03 Electrometric Titration Aluminum, Dissolved Chelation-Extraction Atomic Absorption, . Spectrometric Arsenic, Dissolved Hydride Atomic Absorption, Spectrometric, Boron, Dissolved Bromide, Dissolved Calcium, Dissolved Chloride, Dissolved Fluoride, Dissolved Hardness Hardness,Noncarbonated Iron, Dissolved Lithium, Dissolved Magnesium, Dissolved Manganese, Dissolved Nitrogen, Nitrite plus Nitrate, Dissolved Nitrogen, Ammonia, Dissolved Nitrogen, Ammonia, Total - Oxygen, Dissolved H . P Potassium, Dissolved Silica, Dissolved Sodium, Absorption Ratio Sodium, Dissolved Sodium, Percent Solids, Residual Dissolved Solids, Dissolved on Evaporation Specific Conductance Strontium, Dissolved Sulfate, Turbidimetn'c Dissolved Sulfide, Total Zinc, Dissolved Automated DC Plasma Atomic Emission Spectrometric Ion Exchange, Chromatographic-Electrochemical, Automated Direct Atomic Absorption, Spectrometric Ferric 'I'hiocyanate, Colorimetric, Automated- Discrete Ion-Selective Electrode, Electromctric, Automated Calculation Calculation Direct Atomic Absorption, Spectrometric Direct Atomic Absorption, Spectrometric Direct Atomic Absorption, Spectrometric Direct Atomic Absorption, Spectrometric Cadmium Reduction-Diazotization. Colorimetric, Automated-Segmented Flow Indophenol Colorimetric, Automated Indophenol, Colorimetric, Automated Winkler Method, Titrimctric Glass Electrode, Electrometric Direct Atomic Absorption, Spectrometric Molybdate Blue, Colorimetric, Automated Calculation Direct Atomic Absorption, Spectrometric Calculation Gravimetric, Residue on Evaporation at 180 C Electrometric, Wheatstone Bridge Direct Atomic Absorption, Spectrometric Barium Sulfate, Turbidimetn‘c, Automated-discrete Iodometric, Titrimetric Direct Atomic Absorption, Spectrometric 43 semi—automatic mass spectrometer after reaction with zinc according to the techniques of Kendall and Coplen (1985). Oxygen and deuterium isotopic ratio values are reported as del values, relative to values of standard mean ocean water (SMOW) (Craig, 1961 b). Analytic precision for oxygen is +0.15 o/oo and deuterium isotopic analyses are and + 2.0 o/oo, respectively. Carbon-13 samples were analyzed at Global Geochemistry of Los Angeles, California. The analyses employed a nuclide mass spectrometer to measure the carbon-l3 to carbon-12 ratio on carbon dioxide generated from a strontium carbonate precipitate, according to a modification of the technique of McCree (1950). Analytic precision for del carbon-13 values is +2.0 o/oo, relative to the PDB standard which has an assigned del carbon-13 value of O o/oo. 3.4 Data Reduction The methodology for chemical data reduction included statistical, graphical and computer modeling techniques. The statistical data reduction technique of Student's t— Tests is described and the results of this technique are discussed in Section 4.1.1. The multivariate statistical data reduction techniques are described and the results of these techniques are discussed in Section 4.1.2. The graphical data reduction techniques are described and the 44 results are discussed in Section 4.1.3. The computer modeling techniques are described and the results are discussed in Section 4.1.4. SECTION 4 - RESULTS AND DISCUSSION 4.1 Chemical Characterization The results of chemical analyses of ground-water samples collected from the drift are summarized in Table 4.1. The results of chemical analyses of ground-water samples collected from the wells screened within the bedrock are summarized in Table 4.2. The techniques and results of data reduction, which included statistical (Student's t— Tests and multivariate factor analysis), graphical and computer modeling techniques, are discussed below. 4.1.1 Student's t-Tests for Comparison of Drift and Bedrock Geochemistry Student's t-Tests were used to compare ground—water geochemistry of drift and bedrock for selected chemical parameters, to test for statistically significant differences in values between the drift and bedrock ground water. The t-tests were run with two different sets of equations, including the Cochran‘s Approximation to the Behrens-Fisher Student's t—Test (U.S. EPA, 1987) and Pooled Variance, which assumes that the populations have equal variances. Both sets of equations produced the same results with the Saginaw Lowland geochemical data. The population was defined as the concentrations of a chemical parameter in either drift or bedrock ground water samples. Each individual t-test for a given population was performed at 45 46 :munv Fumvnv anvnv —vv~n¢ .0 0.580 25...... 47 5. 3:30:28 .2 N05 .28. hhh‘ hh‘rx #745 h h h .h h .x. .h h ,. KB nhhhn to .— In cou>x0 3.500 :02”? Fvanv vanv 22.53 35 48 mnn—¢c vanvm Vnuwvw —onwna Oemfinv —vmnn¢ mmNNnc n—mnnv 3553 cam 49 I2 N05 cgom In coo>xo 0.55... 35 50 the 90 percent confidence level. When multiple t-tests are compared, the combined confidence level is less than 90 percent. The Student's t-Tests compared the null hypothesis to an alternate hypothesis to determine if the data supplied significant evidence to reject the null hypothesis in favor of the alternate hypothesis. The null hypothesis, HO, states that the average value of the parameter in the ground water in the drift was equal to the average value in the ground water in the bedrock. The alternate hypothesis, H1, states that the average value of the parameter in the ground water in the drift was not equal to the average value in the ground water in the bedrock. The results of this comparison are summarized in Table 4.3, along with the average values of the chemical parameters. The Student's t-Tests were run on log— normalized data. The data were found to be log-normally distributed as illustrated by frequency histograms of concentrations (Figure 4.1 through Figure 4.15). As shown on Table 4.3, the results of the Student's t— Tests indicate no significant difference between the average values of calcium and magnesium concentrations, and the average values of alkalinity, pH, and temperature for the drift and bedrock. Significant differences occur, however, between average concentrations of sodium, potassium, Table 4 . 3 Comparison of Saginaw Lowland and Bay County drift and bedrock ground-water chemistries via the Student's t-Test. Test was conducted at the 90% confidence level, with the null hypothesisd-Io: difference Average Average Average Value Value Value Drift Bedrock Significant Greater Variablel Deposits Deposits Difference? in Drift?Z Ca 99.3 93.2 NO Mg 30.86 25.64 NO Na 194.0 404.7 YES NO K 2.63 4.92 YES N0 804 191.24 260.19 YES NO Cl 282.9 537.2 YES NO Alkalinity3 232.2 256.] NO pH 7.55 7.55 NO Temp.('C) 11.47 11.45 NO Si 13.15 9.67 YES YES Dissolved Solids 1189.7 1744.1 YES NO Deuterium (del) ~63.50 -77.57 YES YES Oxygen-18 (del) - 10.65 -11.70 YES YES 1 2 3 Reported as CaC03. All components reported in mg/l unless otherwise noted. When the t-Test indicates no significant difference, then the difference in average values is not significant. 52 >1 0 s a o H I": {I 9C1 1543 23“] 3 n 450 Calcium (mg/1) Figure 4.1 Frequency histogram of calcium concentrations. a o c o s u o n m a 38 so 99 m 150 Magnesium (mg/1) Figure 4.2 Frequency histogram of magnesium concentrations. 53 Frequency 0 i 6 9 u 5 Potassium (mg/1) Figure 4.3 Frequency histogram of potassium concentrations. 5 '7. ”LT? ....... ETD J ......... '.mmmf ..... 7’: I? j ,- j.) w— a . - 0 ._. . s :U* ----- - o - . fl - . U1 . a .. m J . [av—j; : : 3 -— 0—_ 3 .flnnin l 1 Al 1 I AA A l l . 11A] . I 6 4 3 R H M M Sodium (mg/l) film} Figure 4.4 Frequency histogram of sodium concentrations. 54 >1 0 I: ii a u m [1 1000 21.133 3303 4800 SW Chloride (mg/l) Figure 4.5 Frequency histogram of chloride concentrations. a o c i o H m ‘3 2GB 400 600 803 Alkalinity (mg/l CaC03) Figure 4.6 Frequency histogram of alkalinity values. 55 h u c a o u h o 400 an 1:133 1.5m ' 200:3 Sulfate (mg/1) Figure 4.7 Frequency histogram of sulfate concentrations. a 0 c 2' o u m pH Frequency histogram of pH values. Figure 4.8 56 18 a 8 u c a o n h 1 2 8 o 4 3 12 1.5. 211 24 silica (mg/l) Figure 4.9 Frequency histogram of silica concentrations. 4L.) I ' f 1,7 I l T r r 1 31:1 - if ' L ..‘ 1 :>4 ~ / . 0 r {'7 5 - 13-, E ”"f '16.? o _i/ ‘ E - 5:11 16 -7“, , -/¢; . / ’,.- _/ . I 2??ij ; a / r l/ ‘ / XI. / n m n m l 1 l L l 1 1 l 1 l 0 3 c 9 12 15 (X 1008) Residuals on Evaporation (mg/1) Figure 4.10 Frequency histogram of residuals on evaporation values. Frequency 4% Figure 4.11 Frequency -19 Figure 4.12 57 -128 ‘156 -38 -EB -48 Del Deutrium (o/oo) Frequency histogram of deuterium concentrations. ~J -P '5 43 41 4 - 1 Del Oxygen-18 (o/oo) Frequency histogram of oxygen-18 concentrations. 58 Frequency c1 2 4 13 1e Dissolved Oxygen (mg/l) Frequency histogram of dissolved oxygen Figure 4.13 concentrations. 59 r1—rlrrrlvvlrlvr] Frequency 11.111511 111-1.; 1141i141liiiliiil 0 2 4 6 E Bromide (mg/l) Figure 4.14 Frequency histogram of bromide concentrations. 60 I I I I I I 1 I Tfi’] I I y m- .................... _ 16" ‘ fi 7 0 M' 5 1, g. 12* ...................... . 0 H h I ,__:.'- _ (X 11W) strontium (ug/l) Figure 4.15 Frequency histogram of strontium concentrations. 61 sulfate, chloride, and dissolved solids for the ground water samples collected from the drift and from the bedrock. Each of the chemical parameters has statistically greater values in the bedrock. Significant differences also occur for average concentrations of silica, deuterium and oxygen—18 concentrations, but the concentrations of these parameters are significantly greater in ground water from the drift than from the bedrock. Student's t-Tests were also performed by Long et all (1986) with the Bay County data set. These t—tests were performed at the 95 percent confidence level, but those done for this study were performed at the 90 percent confidence level for each t—test (for a combined confidence level of less than 90 percent). Despite the different confidence levels, the Bay County t-tests indicate similar trends between drift and bedrock ground—water chemistry as the Saginaw Lowland t—tests, with the exception of sulfate concentrations. Sulfate concentrations are significantly different between drift and bedrock sources in the Saginaw Lowland, but are not significantly different in Bay County. 4.1.2 Multivariate Factor Analysis of the Geochemical Data Set The statistical technique of multivariate factor analysis was used to examine the structure in the geochemical data set as a means of defining the number and 62 types of water masses represented in the data set, and the controlling geochemical processes affecting the ground-water system. Multivariate factor analysis was used to redefine the variables of this data set into a smaller number of variables. A description of factor analysis methodology and usage is discussed by Klovan (1975), Joreskog et al, (1976), Davis (1986) and Kim and Mueller (1978). Multivariate factor analysis has been applied to geochemical data sets by Hitchon et al. (1971); Bopp and Biggs (1981); Hull (1984); Davis (1986) and Long et al. (1988). The factors which result from factor analysis applied to geochemical data sets are interpreted as an indication of different water masses (Q—mode) or as specific processes affecting the ground water (R-mode). Q-mode factor analysis is a test of the homogeneity of a data set. Q-mode compares relationships among samples in terms of chemical variables. R-Mode compares relationships among chemical variables in terms of samples. R—mode attempts to find structure in geochemical data by grouping elements with similar chemical behaviors together into factors which represent simpler relations among fewer variables. The geochemical data set for the Saginaw Lowland study consists of 19 chemical parameters and 105 ground—water samples. The data were found to be log-normally distributed as illustrated by frequency histograms of concentrations (Figure 4.1 through Figure 4.15). Because factor analysis 63 assumes that data are normally distributed (Hitchon et al., 1971), log transformation of the geochemical data (except pH) was done prior to factor analysis. Only chemical parameters were included in the analysis, with the exception of specific conductance which has a strong linear correlation with dissolved solids and thus distinguishes water of similar ionic ratios, but with different ionic concentrations (Hitchon et al., 1971). The parameters included dissolved calcium, dissolved magnesium, dissolved sodium, dissolved potassium, dissolved strontium, dissolved lithium, dissolved chloride, dissolved sulfate, alkalinity, dissolved ammonia nitrogen, dissolved bromide, dissolved fluoride, dissolved silica, dissolved boron, dissolved manganese, dissolved iron, zinc, specific conductance and pH. Q—mode analysis was performed on the data set with the use of a main frame computer. The data was rotated according to Promax rotation. Q-mode revealed that one factor explains more than 90 percent of the variance in the data set, and almost 96 percent of the variance is explained by two factors. Homogeneity in the data set is indicated because one factor explains more than 90 percent of the variance in the data set. There is one major "population" or water mass in the data set (Hitchon et al., 1971; Bopp and Biggs, 1981; Davis, 1986; Long et al., 1988). Therefore, the data set can be treated as one geochemical 64 system. R—mode factor analysis, which requires a homogeneous data set, was then performed. R-mode factor analysis was applied with Promax rotation, which was found by Long et al. (1986) to be the most meaningful manner of rotation with this type of data. The application of R-mode factor analysis resulted in five geochemical factors which account for 97 percent of the variance in the data set. R-mode factor analysis was also performed by Long et al. (1986) on the Bay County data set. The analysis resulted in three geochemical factors. Figure 4.16 is a diagram which shows the factor loadings of the five factors resulting from Promax rotation. The factor loading of a variable (chemical parameter in this case) indicates the relative significance of the variable in defining a factor. The greater the factor loading (positive or negative) for a parameter, the more significance the parameter has on the factor. A factor loading near zero indicates that the parameter is of little significance to the factor. As a convention, parameters falling between +0.250 and —O.250 are considered to be of little significance to the factor's meaning (Long et al., 1986). The R—mode factors for the Saginaw Lowland data set are interpreted as follows. Factor One, which accounts for 42 percent of the variance in the data, has large positive loadings for chloride, sodium, bromide, potassium, specific 65 fixed m6- 1 mud- 1 Swans... . .u: .m .8. 66:2 .6. XQQ m Boom“. .mmcwcmoa uouomu oposlm no Ewummwo wa.v whomflm $16 :30: 552 some >2 .8 cmocaooom oocmtm> 38:88.". .. XXK. 2.? -2183. IV: .6 cu .9 var i... 4 e cocoon. *Qoow ---eN 62-. Q .vom: 1F 1 6 m5 oz .6 .2 .3. m 55mm Eozfiom .8894: 1x3.“ as? F1 4m. ll. IQ l' .I A II- m.°I on. em . ---a..-.c.....:... - mm o o ~ g £620.39- the .5: . 3.6 1.. o E! 1.. a .i . k 9.12 m o kin a! 60 9662.5 r N Boom“. F Baum”. m_m>._uzm3ammu Fl ‘1. -2 r: l _. - CALC= ': \ ~P/ LDC (IAP/K Frequency histogram of calcite saturation indices. Figure 4.23 1 f ..—.pr—.L._.. qua—‘u.—<‘4—--—.-. - .— 20 T 15 } 2 8 4 0 >mzmzammu TE LOG (IAP/Ksp) ARAGDNI Frequency histogram of aragonite saturation indices. Figure 4.24 82 N D l I ' H 0) III! FREQUENCY m K, .....,. # u l u . ““‘“““‘ c: ,r E 3 & ........................................................ u -4.5 -3.5 -2.5 ‘-1.5 -0.5 0 0.5 1.5 LDC (lAP/Ksp) DOLUMITE Figure 4.25 Frequency histogram of dolomite saturation indices. H FREQUENCY \\\\\\\\‘.\\\‘R\\\\\‘ : Loc (IAP/Ksp) GYPSUM Figure 4.26 Frequency histogram of gypsum saturation indices. 83 7COA//VOA//VQOA///QC//- T///OC///VCC//7 / r/VOOfl///O// _/fl/UCA//O//WC///mw/UW /u /, / _ _ _ — _ 0 8 5 4 2 U 1 >uzm30mmd -5 -3 p) HALITE -9 ~11 UG (iAF/Ks Frequency histogram of halite saturation indices. Figure 4.27 mwinn _/ /v. . //oo////yaa _vx_/7;//,//7/7 ///. _ //// / 76 w/ /,,/// anA/ ,/, m _ _ _ _ _ 1 S 0 S O 5 U 2 2 1 1 >uzm3ammd 1 D -0.5 "i311. ELI h—‘ cu) . LDC (lAF/K Figure 4.28 Frequency histogram of quartz saturation indices. FREQUENCY p—a 2 84 -0. 5 -D. 3 -D. 1 D. 3 0. 5 LEE (IAP/Ksp) CHALCEECNY Figure 4.29 Frequency histogram of chalcedony saturation indices. 85 1986; Long et al., 1988; and Rose and Long, 1989). These diagrams are based on activities calculated with chemical modeling and on thermodynamic stability fields. Using the results of chemical modeling on the ground water data, an activity-activity diagram of potassium aluminosilicate minerals is presented as Figure 4.30. The mineral stability fields in the diagram are based on Velbel (1992). The diagram was constructed from the data of Robie et al. (1979). The placement and extent of the illite stability field in Figure 4.30 is based on an idealized illite—surrogate as defined by Grim (1968). The idealization of the illite stability field is intended to simulate the approximate placement and extent of the illite stability field in relation to the placement of the ground water sampling points. The ground-water sample data in Figure 4.30 plot within the illite and kaolinite stability fields, straddling the illite/kaolinite stability field boundary. This indicates thermodynamic equilibrium of the ground water with respect to illite and kaolinite, which are present in the drift deposits in the Saginaw Lowland (Chittrayanont, 1978). Using the results of chemical modeling on the ground water data and mineral stability fields based on Drever (1982), an activity-activity diagram of sodium aluminosilicate minerals is presented as Figure 4.31. The 86 v 6 I luoeovtto IIYVVTI K~foldspar LOG [K/H] w Kaolinlte Pyro phyllite llLlllLJlllllllllIllllllllllll' ‘ITIUIIVIIIVVVIIIIIIII‘I D I a I A I a . A A I s . A . l -s -4 -3 -2 L05 [H4SIU4l Figure 4.30 Mineral stability diagram of potassium aluminosilicate minerals. 9 .-‘ I ' I‘ ' '. : | i L l : E3: 1 : : I 3 "a“ 7? 1 ‘: g 6 :Gibbsite 3: : D : ENE : b o _ - 53 5 7 EIa) 1 : :1 <' : 4r 3' 5° i E . 0| . . l . 3 L l Kaoltmte _- 1 . . . J . . . . 1 l . . l 1 -5 -4 -3 -2 LOB [H4SIU4] Figure 4.31 Mineral stability diagram of sodium aluminosilicate minerals. 87 ground-water sample data plot within the kaolinite stability field. This also indicates thermodynamic equilibrium of the ground water with respect to kaolinite. 4.2 Oxygen and Deuterium Isotopes Concentrations of the stable isotopes of oxygen and deuterium in ground water are used to indicate the origin and evolution of ground water. Isotopic ratios in the water reflect physical processes which have affected it. In low salinity ground water, del oxygen—18 and del deuterium values are conservative; they are unaffected by physical, chemical or geological processes. In the hydrosphere, del oxygen-18 and del deuterium values are not conservative (Frape and Fritz, 1982). A plot of del oxygen—18 values versus del deuterium values for ground-water and surface- water samples can be used to demonstrate that ground water is derived from local precipitation (Wood and Low, 1986). Oxygen—18 and deuterium concentrations were measured in the Saginaw Lowland ground-water samples. The results of the isotopic analyses are summarized in Table 4.1 for drift samples and Table 4.2 for bedrock samples. The concentrations of oxygen-18 and deuterium are reported according to the standard of Craig (1961 b), as del (ratios) values relative to values of standard mean ocean water (SMOW). Samples were collected as described in Section 3.2 and analyzed as described in Section 3.3. Figure 4.32 is a 88 v .cofluouucoocoo mH1com>xo msmuo> coaucuucoocoo Ecfiuouooo 323 chom>xO ED 0 v1 m1 NH1 mn.v magmas o .l a o n .I 0 .III. .- 1... .eo— 52.325 .00 new $753.55 .00 u noEEom .335 950.0 eo.< :55 .335 . we once: 533$ . . . D.......>«C300>mm...... — I 1 I — I I I — I I I I I I — d I I .— d I I — . >3 {orgasm «a .55: 3.3 n. H H H I H mm“: .253. .355 so: .83 u o H H u .. .. . noctm 53m 2320:). .1. I . . . 1 \ . . ~2< 0:355 .8920 . . I . . .ooacem j’. . . Icon-onooucncuooo.yon-oo-oo' *uol .. a...~anuo.ooooc.oo a... I] I o a O. n 4 o c v u . . o . . . . . at. . . . . I . I. . . . . o . a a I a o — I I I - I I I — D I I — I I I — I L I — I I I — Dmfit Dmfil Cat 001 om1 (00/0) Cl IBG 89 plot of del oxygen-18 vs. del deuterium for all ground—water samples from the Saginaw Lowland. Also plotted on Figure 4.32 are the following: (1) the meteoric water line (MWL) established by Craig (1961 a) as the relationship between oxygen-18 and deuterium concentrations in meteoric waters which have not undergone large amounts of evaporation; (2) the line resulting from oxygen versus deuterium concentrations for rain-water samples collected near Simcoe, Ontario (see Figure 1.1 for location of Simcoe study site) by Desaulniers et al. (1981); (3) the isotopic compositions of samples of Michigan basin formation brines (Wilson and Long, 1986); (4) the isotopic compositions of local rain water and Lake Huron at Saginaw Bay (Long et al., 1986); (5) the range of isotopic values from Bay County ground waters (Long et al., 1988); and (6) the range of isotopic values of ground water from the greater Lansing area (GLA) of Michigan (see Figure 1.1 for location) (Slayton, 1982; Long and Larson, 1981). The ground-water samples have del oxygen—18 isotopic values which range from -l7.91 to -8.70 o/oo and del deuterium isotopic values which range from -126.7 to -56.5 0/00. The ground-water samples plot along a line of slope 7.8 and y-intercept 10.86. This is similar to the MWL of Craig (1961 a) with a slope of 8 and y-intercept of 10. Figure 4.32 suggests that meteoric water (modern-day and 90 ancient) is the source of both drift and bedrock ground water in the Saginaw Lowland, because samples plot along the MWL. The linear relationship of the Saginaw Lowland ground- water samples is similar to the line resulting from isotopic values of rain-water collected at Simcoe, Ontario, located in a region geographically and hydrogeologically similar to the Saginaw Lowland (Desaulniers et al., 1981).‘ The Simcoe line has a slope of 7.5 and y—intercept of 12.6. This line also falls close to the MWL. The similarity of the Simcoe line to the Saginaw Lowland line suggests that modern—day precipitation is one of the sources of ground water in the Saginaw Lowland. The formation brine samples on Figure 4.32 have del oxygen-l8 values ranging from -3.02 to +1.10 o/oo and del deuterium values ranging from —28.6 to —10.7 o/oo. The ground-water isotopic data does not indicate mixing of Michigan basin formation brines with meteoric water. The mixing of brine and meteoric water would result in a linear relationship between the two end-members on a del oxygen—18 versus del deuterium plot (Frape and Fritz, 1982). Samples from the interface of brine and fresher shallow ground water w0uld allow better understanding of the role brines play in the near—surface hydrogeologic system. Long et al. (1988) 91 did not collect ground-water samples from depths great enough to detect the interface of brine and fresher shallow ground water. The isotopic values of ground—water samples from Bay County (Long et al., 1988) plot along the MWL. The samples range from -7.0 to -18.5 o/oo for del oxygen-18 and from -56 to -l37 o/oo for del deuterium. Figure 4.32 illustrates the range of these samples, but it does not illustrate their actual position along the MWL. The Saginaw Lowland samples plot densely in the region of heavier isotopic values, while the Bay County samples tend to be more isotopically depleted. A meteoric source for Bay County ground water is indicated by oxygen and deuterium isotopic data (Long et al., 1988). Del oxygen-18 and del deuterium values for the GLA ground water plot within a small range of values on the MWL (Figure 4.32). The GLA ground water is believed to be modern-day meteoric water (Long et al., 1988). The range of the GLA samples is similar to the heaviest values found in the ground water of the Saginaw Lowland, suggesting that modern-day meteoric water is one end-member of the ground water in the Saginaw Lowland. Isotopic values for the Saginaw Lowland samples range from isotopic depletion to values similar to modern—day precipitation. This wide range of values suggests either 92 the mixing of isotopically-depleted glacial meltwater with modern-day meteoric recharge, or the isotopic variability in annual precipitation. Long et al. (1988) investigated these possibilities for the Bay County samples and concluded that large isotopic variability in annual precipitation is unlikely. Yearly average isotopic values for local precipitation were estimated at -7.34 and -8.55 o/oo del oxygen-18; and -50.59 and -60.44 o/oo del deuterium.. Lake Huron water has del oxygen—18 and del deuterium values of —7.45 o/oo and —55.22 o/oo respectively, similar to calculated yearly average isotopic values for precipitation. These approximate isotopic values of regional annual precipitation are heavier than measured ground-water isotopic values. The distribution of del oxygen—l8 values for drift ground-water samples is shown in Figure 4.33. The distribution of del oxygen—18 values for bedrock ground- water samples is shown in Figure 4.34. The bedrock aquifer is slightly depleted isotopically relative to the drift (average del oxygen—18 value of -11.70 o/oo for bedrock and -10.65 o/oo for drift). The distribution of del oxygen—18 values is similar in the drift and bedrock systems (Figure 4.33 and Figure 4.34), in that isotopic values are depleted near the center of the Saginaw Lowland, and become more enriched towards the perimeter of the study area. This pattern indicates that the Saginaw Lowland encompasses two 93 Exo— 010m .mufimomwp uufiup on» a“ unawuouucmocoo $1.30th an. cofiusnfiuumwo nné musmfim .03.; .3 N . 138:. 53:00 . C >\\ Av Z 0.x 6.x it- .muamommc xooupma map :a mcofiumuucoocoo cancwomxo mo coauznauuwfio en.v wusmfih 25.2, .03 N . 138:. 53:00 \ A I . Ego, — 00...: Gr — \ oiom (1 n1 \ n7 A/ ., .v Q , 9x / 95 zones: (1) a region of isotopically—depleted water and (2) a transition zone from isotopic depletion to isotopic enrichment as in modern-day precipitation. Isotopic depletion indicates formation of water at cooler temperatures than present day (Dansgaard, 1964; Mook, 1972; Payne; 1972 Siegenthaler, 1979; Gat, 1981). Isotopic depletion in ground water has been attributed to glacial meltwater recharge in aquifers by Clayton et al. (1966); Fritz et al. (1974); Perry et al. (1982); and Siegel and Mandle, (1984). The extremely low isotopic values of the Saginaw Lowland ground water represent retention of glacial meltwater by aquitards, which can have considerably different geochemical and hydrologic properties than aquifers (Back, 1986). Desaulniers et al. (1981) investigated ground—water geochemistry of argillaceous Quaternary deposits (aquitards) in southwestern Ontario (location shown in Figure 1.1) which have a geologic setting similar to that of the Saginaw Lowland. They concluded that the pore water in the deposits is a mixture of late Pleistocene—aged and modern—day recharge water. Bradbury (1984) performed a similar investigation in argillaceous deposits of northwestern Wisconsin, and concluded that ground water was original pore water from time of till deposition 9,500-10,000 years BP. 96 Isotopic evidence suggests that Saginaw Lowland ground water is a mixture of modern-day precipitation and glacial meltwater from approximately 10,000 years before present (3?). This is the most likely time of the most recent glacial meltwater recharge in the east-central Michigan basin (Fritz et al., 1974) based on documented climatic changes in the Great Lakes region. .There is no isotopic evidence that other physical or chemical processes such as mixing with brine or intense evaporation have affected the ground water (Craig, 1961 a; Dansgaard, 1964; Hitchon and Friedman, 1969; Gat, 1981). 4.3 Carbon Isotopes The concentration of the stable carbon isotopes in ground water is used to indicate the origin and evolution of ground water. The carbon isotopic signature is the result of exchange of ground water with various sources of carbon (coal, carbon dioxide gas in soil zones, carbonate minerals, other carbon-bearing minerals, methane gas, plant material) each of which has a distinct range of del carbon-l3 values (Curtis et al., 1972; Mook, 1972; Payne, 1972; Rightmire and Hanshaw, 1973; Long et al., 1988). Sources of dissolved carbon and microbiological processes affecting the ground water can be determined from studies of the carbon-l3 isotope. Stable carbon isotope values in ground water can be used to characterize the origin of ground water (Plummer, 97 1977; Thorstensen and Fisher, 1979; Desaulniers et al., 1981; Wood and Low, 1986; Chapelle et al., 1988; and McMahon et al., 1990). Carbon—13 concentrations were measured in the Saginaw Lowland ground water on seven samples from drift deposits and seven samples from bedrock deposits. The results of isotopic analyses are summarized in Table 4.1 for drift samples and Table 4.2 for bedrock samples. The concentration of carbon-l3 is reported as a del (ratio) value relative to the PDB standard. Samples were collected as described in Section 3.2 and analyzed as described in Section 3.3. Figure 4.35 is a histogram of ground—water del carbon-13 values. Also plotted on Figure 4.35 are local rain water (-8.27 o/oo) and Saginaw Bay, Lake Huron, water (-9.93 o/oo) (Long et al., 1988). The ground—water samples have del carbon—13 values which range from —15.4 o/oo to -7.9 o/oo, with an average of —13.08 0/00. The Bay County ground-water samples exhibited a wide range of del carbon—13 values (-21.38 o/oo to —8.14 o/oo), with an average of -13.8 0/00 (Long et al., 1988). The del carbon—13 values of ground water in the Saginaw Lowland fall within this range. A much greater number of Bay County ground water samples were analyzed for carbon—13 than for Saginaw Lowland. Long et al. (1988) found no systematic distribution of del carbon—13 in the Bay County ground-water system. Frequency 98 4.5 y I Average 3.1333 ____.> < Lake Hugm Water 4 _ Saginaw Bay 3.5 " <"“—""‘ Local Rain Water 3 (August 1985) 2.5 — 2 — 1.5 — 1 — 0.5 " . l I i l .20 '15 ’10 -5 0 Del C 13 (0/00) Figure 4.35 Histogram of carbon-13 concentrations. ~ 99 A slight correlation between del carbon-13 values and depth of sample is suggested by Figure 4.36. Points on the graph represent the values of del carbon-l3 plotted at the midpoint depth of the well screens from which samples were withdrawn. Drift wells were screened in sand and/or gravel. Bedrock wells were screened in combinations of shale, sandstone and/or limestone. Based on the small number of samples analyzed, Figure 4.36 suggests that ground water becomes more enriched in carbon-13 at depth. The process which would result in this vertical distribution of carbon- 13 values is unclear. Photosynthesis causes preferential uptake of carbon-l2 in plants, producing depletion of carbon-l3 in plant matter. As a result, marine organic carbon and recent marine sediments have del carbon-l3 values of about -20 o/oo. Similarly, land organic carbon and recent freshwater deposits have del carbon—l3 values of about -25 0/00. (Stumm and Morgan, 1981). The amount of carbon-l3 in dissolved carbon reflects the relative amounts of several carbon sources (Wigley et al., 1978). For example, a combined source of carbon consisting of lignitic material (- 25 o/oo) and carbonate minerals (O o/oo) results in an actual average value of dissolved carbon-13 content in ground water of -12.1 o/oo (Thorstenson and Fisher, 1979). The average del carbon-13 values in Saginaw Lowland ground water of —13.08 o/oo may represent the combined V Del Carbon-13 loloo) -‘I 100 _9— Figure 4.36 Depth of Well Screen (feet) Carbon-13 concentration versus depth. 101 influence on dissolved carbon of lignitic material derived from local coal deposits (-25 o/oo) and of carbonate minerals such as calcite, aragonite and dolomite, found in local drift and bedrock deposits. The more enriched values of Saginaw Lowland del carbon—13 values (—7.9 o/oo) indicate a meteoric source. V SECTION 5 - SUMMARY AND CONCLUSIONS 5.1 Summary The following is a summary of the many results of the study and an evaluation of the several hypotheses proposed in the Introduction. Hypothesis 1, that the ground water system contains isotopically-depleted meteoric water that recharged the system when the climate was much.cooler (perhaps glacial meltwater of the Pleistocene Epoch), was investigated with the use of stable isotopic data, as described in Section 4.3 and Section 4.4. These data indicated that the isotopic geochemistry of the ground water is influenced by mixing of modern—day meteoric recharge water with meteoric water formed at cooler temperatures bearing isotopic depletion. The data suggest that the isotopically-depleted water was derived from glacial meltwater recharge during the late Pleistocene epoch, when the climate was much cooler than present day (Emiliani, 1971). Hypothesis 2, that the ground water system contains saline water derived from brines deep in the Michigan basin, was investigated with the use of chemical characterization, as described in Section 4.1. The chemical data suggest that salinity is a dominant factor of the ground—water geochemistry. The salinity is the result of both the diffusion or advection of highly concentrated Michigan basin 102 103 brines upward into the near—surface deposits of the Saginaw Lowlands and the dissolution of halite. The brine input is suggested by factor analysis, chloride to bromide ' concentration ratios and ternary diagrams of major ion chemistry. The halite input is suggested by a plot of molar sodium concentrations versus molar chloride concentrations. The chemical characterization also suggests geochemical interactions between ground water and surrounding geologic deposits, including calcium carbonate thermodynamic equilibrium. Hypothesis 3, that the isotopically-depleted water and saline water are retained in the ground water system through the combined affects of aquitards, low regional horizontal hydraulic gradients, and a regional ground-water discharge zone, was investigated with the use of geologic maps and water level and lithologic data for each well sampled, as described in Section 2.2 and Section 2.3. The maps and these data suggest that the Saginaw Lowland ground—water system has a low horizontal hydraulic gradient and regional ground-water discharge. The system is dominated by thick aquitards. The combined affects of these three characteristics are believed to result in salinity buildup, retention of isotopically—depleted ground water and retention of upwardly migrating basin brines. The different water types identified mix within the system under the influence of these characteristics. 104 This study also attempted to verify that the additional processes identified by Long et al. (1986, 1988) in the Bay County ground water system are also affecting the Saginaw Lowland system. Verification of the additional processes was evaluated by analyzing the ground water samples for similar chemical and isotopic constituents and by evaluating the geochemical data in a similar manner as in the Bay County study. The additional processes include water—rock interactions including calcium carbonate equilibrium and stability with potassium aluminosilicate minerals; and the biological reduction of sulfate. Evidence for water-rock interactions was provided by chemical modeling with WATEQF. Chemical modeling indicates that ground water is in equilibrium or slightly supersaturated with respect to calcite and is slightly undersaturated with respect to aragonite and dolomite. Activity-activity diagrams show that the ground-water data plot primarily within the illite and kaolinite stability fields, indicating thermodynamic equilibrium of the ground water with respect to illite and kaolinite. Evidence for the occurrence of the biological reduction of sulfate in Bay County was provided by trends in sulfate and bicarbonate concentrations and del carbon-13 and del sulfur—34 analyses (Long et al., 1988). Based on the results of factor analysis, the process of microbiological reduction of sulfate does not appear to be occurring to a 105 great extent in Saginaw Lowland. The process may be a phenomenon localized in Bay County. Further investigation including sulfur and carbon isotope analyses would be necessary to better characterize the process of microbiological reduction of sulfate in the Saginaw Lowland ground-water system. The study was also used to estimate whether the data reduction techniques used in the Bay County study (Long et al., 1986, 1988), would produce interpretable results on more widely-spaced data points. The statistical data reduction techniques used are the Student's t-Test and multivariate factor analysis. The graphical data reduction techniques include ternary diagrams, frequency histograms, and x-y plots. The computer modeling technique used is chemical modeling with WATEQF. The data reduction techniques produced similar and/or complimentary results for the Saginaw Lowland and Bay County studies, despite large differences of scale with regards to sampling sites. The Saginaw Lowland study used a total of 105 sampling sites over an area of about 14000 square kilometers. The Bay County study used a total of almost 400 sites over an area of about 800 square kilometers. The data reduction techniques have proven useful in data interpretation in a large scale ground-water study. 106 Despite the different confidence levels in the Student's t-Tests of this study and of the Bay County study, as discussed in Section 4.1.1, the t-tests indicated similar trends between drift and bedrock ground—water chemistry as the Saginaw Lowland t-tests, with the exception of sulfate concentrations. Sulfate concentrations were significantly different between drift and bedrock sources in the Saginaw Lowland, but were not significantly different in Bay County. R-mode multivariate factor analysis resulted in three factors representing processes affecting ground water in the Saginaw Lowland. These factors corresponded to three R—mode factors for the geochemical data from Bay County ground water of Long et al. (1986). However, a total of five factors were required to account for the same amount of variance in the Saginaw Lowland data as were only 3 factors in the data of the Bay County study. More factors are required to describe the many processes affecting the ground-water system on a regional scale. For example, the study area encompasses not only "Bay County-like" ground water, but also fresher, younger water on the perimeter of the study area. Ternary diagrams revealed stronger evidence for mixing between surface water and brine endmember waters than was found in the Bay County data (Long et al., 1986). This supports the hypothesis that the ground water of the Saginaw V 107 Lowland occurs in a transition zone between surface water and brine compositions. The histogram of chloride to bromide concentration ratios (in mg/l) for all ground water samples illustrated a range of brines similar to that of Bay County ground-water samples (Long et al., 1986). A brine influence on the ground water is clearly suggested. Halite dissolution is also suggested by the plot of sodium versus chloride molar concentrations which indicated that dissolution of halite provided sodium and chloride to the system. This is similar to the findings of Long et al. (1986). These findings suggest that the movement of deep basin brines upwards into the system and the process of halite dissolution both influence the ground-water geochemistry. These two processes have not yet been adequately explained and incorporated in a model of the ground-water system in the Saginaw Lowland. 5.2 Conclusions The goal of this study was to examine the extent of isotopically-depleted ground water in the Saginaw Lowland of the Michigan basin and investigate several hypotheses based primarily on the previous work of Long et al. (1986, 1988) in Bay County of the Saginaw Lowland. The occurrence of isotopically-depleted ground water was compared to regional 108 lithology and ground-water geochemistry in order to test the hypotheses. As summarized in Section 5, the results supported the hypotheses. Isotopic evidence is lacking for brine influence on the ground water system, but geochemical evidence suggests a brine influence on the system. This is consistent with the model of the system, because diffusion of brines into the system would have little affect on isotopic signatures in low concentrations, yet low concentrations of brines diffusing into the system will have strong impacts on geochemical signatures, due to the extremely high concentrations of ions in brines. Collection of ground-water samples from depths great enough to detect the interface of brine and fresher shallow ground water would allow for better understanding of the role brines play in the near—surface hydrogeologic system. This study found a correspondence in geographic distribution of low del oxygen—18 and del deuterium values (Figure 4.33 and Figure 4.34), high chloride concentrations (Figure 2.5 and Figure 2.6), surficial lacustrine and morainal clay deposits (Figure 1.1, Figure 2.1, Figure 2.2 and Figure 2.3), and the residual potentiometric surface (Figure 2.10). The isotopically-depleted, saline water occurs where argillaceous deposits are thick, in the central portion of the study area near Saginaw Bay. 109 The ground water system has not flushed out the isotopically-depleted meteoric water formed when the climate was much cooler than the present, nor the salinity in the system. The dynamics of the ground water flow system are subject to further interpretations of the past and present hydrodynamics of the system. APPENDIX APPENDIX ALUMINUM SPECIATION INVESTIGATION Aluminum is commonly analyzed as total dissolved aluminum. For example, the analytical data for the 1986 sampling for this study of ground water geochemistry in the Saginaw Lowland included total dissolved aluminum. In order to . determine to determine the equilbria cf aluminosilicate minerals via chemical modeling, concentrations of aluminum species are required. The equilibria of aluminosilicate minerals via chemical modeling was investigated during the ground water study in the Saginaw Lowland. In order to determine the concentrations of aluminum species, resampling of selected wells, based on preliminary chemical modeling, was done in the summer of 1987. Twenty—five wells were resampled and analyzed in the Geochemistry Laboratory at Michigan State University. Collection and analysis of the samples for aluminum speciation.was done according to the methods described below. The laboratory analytical data (total dissolved aluminum, monomeric and polymeric aluminum species, and major ions) resulting from this investigation of aluminum Speciation is provided in Table A.l. The concentrations of 110 111 .ouuo: omfisuonuo mmwaco .H\mfi cw oouuomou mcoHuauucoocoo Haé ”ouoz Hmn.e- and «.mvm on m.~ o- a.» an no wmo.no~ adm h.e~o coma m.» ooaa mo ov~ mm msm.ean omm ~H.¢¢ Hm v.~ we H~ em as oma.Hmn mew on.» me s.H mo HH an mo oa~.oma and m.¢no comm ~.m ooo~ we on” me msm.ean mm“ mm.oH om o.~ no HH an ac ma~.oflv nvn sn.ao cad v.n omN m.m an ov oan.vmn mam H~.ss can H.q cow «H am an sow.vov «an ~m.m ~H n.~ mm mm no an mmh.m~v omn nm.~n me «.n and o.a mm kn mca.enn «ma n.hsn can o.¢ com we one am mq>.omfl nod «.mao can o.o ode ea 65H m ham.s«~ non s.mnn om v.n me an aka a no~.non «mm n.spn an o.n nNH em and n "wHHmS ufimonon xoouomm HnH.n- and b.5wn ooow «.6 case v.0 ~.fl hm ~o~.asn “an mo.ns we H.~ as on me mo www.mnn ohm mn.mm HH s.H an pH no on «an.vsn nan m~.~m H.k ¢.H we «a me an nsm.mea mud m.mov comm s.m coma mo“ can on non.sn~ mad H.mmH oom o.~ 6mm mo ode mm own.nan smN ms.vo a.o a.” me Ha «v mm emo.moa one n.ssn h.m o.~ «a on com ma anm.m¢n sou mv.nm s.m m.a «N am mm «H nmm.m~a nod n.nom a.HH m.n em as ov~ nH HnH.n- no” . v.ns~ m.m m.a on an on m “made: uflmommo uufiun Anoumc Anoouo. nonssz auacaaaxae suaaaauxfl< «om an x .a2 a: no muam _ cofiuamwumu>CH scaumwoomm ascfiadaa you mafiamamm noun: ocsouu smafl no muaomwm Haneuxamcd auouauonaq u H.< manna 112 .oouo: oufi3hwnuo moods: .H\mfi cw oouuomwu mCOwuuuucoocoo Had "muoz 00.0 no.0 00.0 00.0 no.0 Ho.0 0.n nn.o n0 ov.o n0.H 0~.H 00.0 00.0 vn.0 o.v n.v N0 no.0 nNeo NN.o v0.0 00.0 0H.0 H.0 n.H H0 no.0 HH.o 00.0 00.0 00.0 NH.0 n.0 00.0 0o oH.o wn.a 0H.H 00.0 o~.o 0o.0 H.0 v.n mv v0.0 NHno 00.0 00.0 no.0 00.0 0.o oo.o Hv v0.0 «H.o 0H.o H0.0 no.0 No.0 H.o ~.H 0v 00.0 0n.o on.o 00.0 ov.o no.0 n.o n.H an 00.0 oN.o om.o 00.0 00.0 v0.0 «.0 00.0 0n No.0 «0.0 00.0 no.0 00.0 00.0 v.o n.H 0n n0.v no.0 00.0 no.0 00.0 on.0 0.n n.H Hm om.o 00.0 0n.o Ho.o mo.H 00.0 N.n o.v 0 0H.H 0v.~ HN.a 00.0 00.0 00.0 0.0 H.N v vu.fi vv.~ 0N.H 00.0 00.0 mo.0 0.0 0.0 n "maaoz ufimommo xoouoom od.o nN.0 oo.o no.0 oo.o n0.0 0.0 00.0 00 no.0 00.0 «0.0 H0.o no.0 . v0.0 «.0 Ho.o no 00.0 0v.o 00.0 00.0 00.0 00.0 OuNH 00.0 on 0H.o 00.0 00.0 00.0 ov.o 00.0 0.0 n0.o 0n 0N.o 00.0 on.n 00.0 00.0 vo.0 v.0 v.0 on oo.o no.0 00.0 no.0 00.0 Hm.0 0.0 0.0 mm oa.o o0.o oo.o no.0 oo.o 00.0 N.HH no.0 nn «N.o 00.0 nn.o Ho.o o~.o 00.0 0.0 0.0 0H 00.0 00.0 nn.o no.0 N¢.0 00.0 n.0H N.H 0H nH.0 o0.d 0n.n No.0 00.0 00.0 o.m o.v nu 00.0 00.0 on.o «0.0 00.0 vd.0 «.0 o.H n «mad»: ufimodoo among um mm mm commxo amass: cannon aauoa moouuom m oo>aommwo an an um auflm .Ao.ucouv H.< canoe 113 .0ouo: omfiahonuo mmuaco .H\OE cw omuuomou mcofiumuucoocou HH< "ouoz n~.v 0N.H am.a v.HH 000 00 0o.~ 00.0 00.0 0.NH oo0n N0 ao.n, «B.H Hm.d w.oa new as on.n nH.H n0.a N.HH 00o mo om.~ oo.o oo.o m.mH mnNm me eox~ _Hm.o Hm.o H.~H Hmm a. m~.~ nv.o oo.~ o.afi mmmfl ow 00.0 nH.0 00.0. o.HH o¢nH mm 0n.H 0d.0 0o.o n.~H ooo 0m 00.? nv.o nv.o n.HH 00o 0n n0.v nd.o 00.0 n.NH ~00H Hm om.” NN.H Ho.o 0.0a 0000 0 n0.v 00.d v0.0 n.~H voHH v an.n 00.0 00.0 o.oH «odd 0 “madmz uwmommn xooupmm m0.N no.d 00.0 0.0a vooo 00 NH." 00.0 n0.H 0.HH Nmo «0 00.0 00.0 ao.o 0.0a Nov 0n .NH.N «v.0 oH.n n.HH v0v 0n nn.n nn.~ 00.0 n.nH 000HH on H0.N 00.0 00.0 H.~H omHN nn 0N.n «0.0 00.0 N.HH nan nn o0.n no.0 no.0 n.HH 000 0a vo.n nv.H 00.d n.0 0nn 0H 00.0 v0.0 00.0 0.HH v0oH nu co.n 0a.~ 00.0 o.aH ovn n "mafia: ufimoaua hogan .mesc Aaxmso .H\m=0 Au mmmuomeo Asu\moaasc nonasz at a< at wuouoummama oucouooocoo oufim Hauoa u>wuooom Ouuofiocox ofiufioomm . 8.280 Ta 032. 114 monomeric aluminum and major ions were chemically modeled using WATEQF (Plummer et al, 1984), a FORTRAN IV version of WATEQ. Saturation indicies for various aluminosilicate minerals calculated by chemical modeling are provided in Table A.2. Analyses for aluminum species were conducted on a Perkin— Elmer atomic adsorption unit equipped with a graphite furnace and an autosampler. Analyses for major ions were conducted according to standard U. S. Geological Survey techniques (Skougstad et al, 1985). A ten-second extraction was performed in the field on the twenty-five ground water samples. The resulting extract contained monomeric aluminum species. The ten-second extraction sample was prepared by filtering the sample through a 0.2 micrometer pore size polycarbonate Nuclepore filter which had been acid rinsed, soaked in distilled deionized water and rinsed thoroughly with the sample water before use. A Nuclepore 0.47 millimeter diameter filtration apparatus was attached to an acid-rinsed 200 milliliter Erlenmeyer flask and a peristaltic pump was used to provide reverse pressure for filtration. Fifty milliliters of the filtrate was then carefully pipetted out and dispensed into a 200 milliliter acid-rinsed beaker. Distilled deionized water was added to produce a final volume of approximately 100 milliliter. Immediately 115 0HH.0I 00.HI 000.Hl 000.0I 00H.Hl 000.Hl 00H.HI H00.Hl 00H.Hl 000.Hl 00H.Hl 00fi.dl 00H.dl 000.Hl H00.0I 000.0! 000.0! 000.0I 000.0I 000.Hl 00.Hl 000.0! 000.0l 000.0I 000.0l 0000 Nahuac 000.0 00.0 000.0 000.0 000.0 000.0 00H.0 HH0.0 000.0 00H.0 000.0 000.0 00H.0 000.0 000.0 0H0.0 000.0 0H0.0 000.0 000.0 00.0 00.0 000.0 0H0.0 000.0 000.0 H00.0 000.0 000.0 HH0.0 000.0 00H.0 000.0 000.0 000.0 000.0 00.0 000.0 000.0 000.0 0H.0 H00.0 H00.0 000.0 000.0 000.0 0H0.0 H00.0 000.0 000.0 uuaaaann unqzax lou0m 000.0 .00H.HH 00.0H 000.HH H00.0 000.0H 00.HH 000.HH 000.0H 000.0 000.00 000.00 000.0 0H0.0H H00.fld 000.0 000.0 H00.0H 000.0 000.0 000.HH 00.0H 000.Hd 000.00 000.0 .mcfiawuoa Hmofifiozo CH 00m: mum: ennufioau uuoHODCOCOE uo mcofiunuucoocoo 000.0 000.0 000.0 000.0 00.0 000.0 000.0 H00.0 00H.0 H00.0 000.0 00H.0 000.0 000.0 000.0 0H0.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 000.0 3223. .5233 000.H HHO.~ 000.0 000.H 0H0.0 000.0 0H0.H 000.0 00.H 000.0 H00.H 000.0 000.0 00.H 000.H 000.H 0H0.0 00H.0 000.0 0H0.H 000.0 0HH.0 00.H 000.0 000.0 "QUOZ 000.0 H0a.0n n0 0nn.0 Hmn.0 00 000.0 00H.H H0 000.0 000.0 00 00H.0 000.0! 00 000.0 000.0 H0 000.0 000.0 00 000.0 000.0 00 000.0 0H.0I 00 ~00.H 000.0I 00 000.0 000.0 00 00.0 0.0 0 000.H 000.0I 0 000.0 000.0 0 "mafic: nanomoo xuouoom 000.0 00.0 00 H00.0 0nd.oc 00 000.0 000.0! 00 00.0 000.0 00 00.0 000.0! 00 000.0 000.0 00 000.0 000.0 00 00.0 000.0 0H 000.0 00.0 0H 0H.0 000.0 0H 0H0.0 0H0.0I 0 ”mafic: ufimomoo uumuo mafia caucaoo< amass: Iamnfiom ouflm coaummwumo>oH acaunfioomm Bocfiasaa 000a .mauuocw: ououwaamOCwaaad macauo> you mofiowocH cowuououwm I 0.< wanna 116 after this step, 2 milliliters of 8-hydroxyquinoline solution was added by repipette. The 8-hydroxyquinoline solution was prepared according to Barnes, 1975. The preparation technique is also described below. "Metal free" ammonium hydroxide was added dropwise to adjust the pH to 8.3. The ammonium hydroxide was prepared according to Barnes, 1975. The pH probe had been standardized with pH buffers 7 and 10, which were at ambient temperature. Five milliliters of buffer solution were then added with a repipette to keep the pH at 8.3. The buffer solution was prepared according to Barnes, 1975. The preparation technique is also described below. The mixture was poured into an acid—rinsed 250 milliliter volumetric flask. Ten milliliters of methyl isobutyl ketone (MIBK) were then added by pipette and swirled for 10 seconds. The layer of MIBK was allowed to separate out. The volumetric flask was carefully filled with water. Using an Eppendorf pipette, the MIBK extract (approximately 6 milliliter) was removed and stored in 30 milliliter high- density polyethylene (HDPE) bottles, which had been acid rinsed and soaked in distilled water. In addition to the ten-second extraction, a thirty—minute extraction was performed in the field on a freshly obtained ground water sample. The resulting extract contained polymeric aluminum species. The thirty-minute extraction was conducted in the same manner as the 10 second 117 extraction, except that the volumetric flask was shaken by hand for 30 minutes before the MIBK was added. The extract was then removed, in the same manner as for the 10 second extract . Several reagents were prepared for use in this laboratory work. The 8-hydroxyquinoline solution was prepared in the following manner. Two grams of 8-hydroxyquinoline were dissolved in 5 milliliters of distilled glacial acetic acid and diluted to 200 milliliters with distilled deionized water (Turner 1969; May et al., 1979; Okura et al., 1962; James et al. 1983; Bloom et al. 1978; Barnes 1975). The buffer solution was prepared by adding 223 milliliters of 10 M "metal free" ammonium hydroxide and 115 milliliters of distilled glacial acetic acid to 500 milliliters of distilled deionized water. The pH was then adjusted to 8.3 with 10 molar ammonium hydroxide and acetic acid. The solution was then diluted to one liter with distilled deionized water (Barnes, 1975). To collect samples for analysis of total aluminum concentration in the field, sample water was filtered with the same filtration equipment as the extractions, except that a 0.4 micrometer pore size polycarbonate Nuclepore filter was used. Two hundred and fifty (250) milliliters of filtered sample was poured into 250 milliliter HDPE bottles, which had been acid rinsed and soaked in distilled water prior to use. The sample was then acidified to pH 2 with 118 0.5 milliliter double distilled Ultrex* nitric acid. The acidified sample was left undisturbed for at least two weeks prior to analysis, in order to allow larger polynuclear aluminum species and colloidal suspended material to convert to a more reactive form of aluminum (Barnes, 1975). BIBLIOGRAPHY BIBLIOGRAPHY Back W. (1986). Role of aquitards in hydrogeochemical systems: a synopsis. Applied Geochemistry 1, 427-437. Badalamenti L. 3., Long, D. T., and Dannemiller, G. T. (1988). Geochemistry of highly saline near surface ground water, Saginaw Valley, Michigan Basin. Geol. Soc. Am. Abstracts with Programs 20:5. Barnes R. B. (1975). The determination of specific forms of aluminum in natural water. Chemical Geology 15, 177-191. Bloom P. R., Weaver, R. M., and McBride, M. B. (1978). The spectrophotometric and fluorometric determination of aluminum with 8-hydroxyquinoline and butyl acetate extraction. Soil Science Society of America Journal 42, 713-716. Bopp F., III, and Biggs, R. B. (1981). Metals in estuarine sediments: factor analysis and its environmental significance. Science 214, 441-443. Bradbury K. R. (1984). Major ion and isotope geochemistry of ground water in clayey till, northwestern Wisconsin, U.S.A. First Canadian/American Conference on Hydrogeolog , Editors: B. Hitchon and E. I. Wallick. Nat. Water Well Assoc., Worthington, OH. Case L. C. (1945). Exceptional Silurian brine near Bay City, Michigan.. Bulletin of the AAPG 29:5, 567-570. Chapelle F. H., Morris, J. T., McMahon P. B., and Zelibor, J. L., Jr. (1988). Bacterial metabolism and the del carbon-13 composition of ground water, Floridan aquifer system, South Carolina. Geology 16, 117-121. Chittrayanont S. (1978). A geohydrologic study of the Garfield Township coal basin area, Bay County, Michigan. Masters Thesis, Mich. Tech. Univ. Clayton, R. N., Friedman, 1., Graf, D. L., Mayeda, T. K., Meents, W. F., and Shimp, N. F. (1966). The origin of saline formation waters, I: isotopic composition. J; Geophysical Research 71:16, 3869-3882. 119 120 Cohee, G. V., Burns, R. N, Brown, A., Brant, R. A., and Wright, D. (1950). Coal resources of Michigan. Mich. Geol. Survey Circular 77. Cooper, W. F. (1906). Geological report on Bay County, Michigan. Mich. Geol. Survey Annual Report for 1905. Craig, H. (1961 a). Isotopic variations in meteoric waters. Science 133, 1702-1703. Craig, H. (1961 b). Standard for reporting concentrations of deuterium and oxygen-18 in natural waters. Science 133, 1833-1834. Curtis, C. D., Petrowski, C., and Oertel, G. (1972). Stable carbon isotope ratios within carbonate concretions: a clue to place and time of formation. Nature 235, 98-100. Dansgaard, W. (1964). Stable isotopes in precipitation. Tellus XVI, 4, Sweden, 436-468. Davis, J. C. (1986). Statistics and data analysis in geology. John Wiley & Sons, New York. Desaulniers, D. E., Cherry, J. A., and Fritz, P. (1981). Origin, age and movement of porewaters in argillaceous Quaternary deposits at four sites in southwestern Ontario. J. Hydrology 50, 231-257. Dow Chemical Company (1986). Salzburg landfill. RCRA, Part B Application, #MID 980617435, Section E. Drever, J. I. (1982). The geochemistry of natural waters. Prentice-Hall, Inc., Englewood Cliffs, New Jersey. Egleson, J. C. and Querio, C. W. (1969). Variation in the composition of brine from the Sylvania formation near Midland, Michigan. Environmental Science and Technology gig, 367-371. Emiliani, C. (1971). The amplitude of Pleistocene climatic cycles at low latitudes and the isotopic composition of glacial ice. The Late Cenozoic Glacial Ages, Editor: K. J. Turekian. Yale University Press, New Haven, 183-187. Epstein S. and Mayeda T., (1953). Variations in oxygen-18 content of waters from natural sources. G. C. Acta 27, 213. Frape, S. K. and Fritz, P. (1982). The chemistry and. isotopic composition of saline groundwaters from the Sudbury Basin, Ontario. Can. J. Earth Science 19, 645— 661. 121 Frape, S. K., Fritz, P., and McNutt, R. H. (1984). Water— rock interaction and chemistry of groundwaters from the Canadian shield. Ga_g*_Ag;a_4§, 1617-1627. Freeze, R. A. and Cherry, J. A. (1979). ngundwatar. Prentice—Hall, Inc., Englewood Cliffs, New Jersey. Fritz, P., Drimmie, R. J. and Render, F. W. (1974). Stable isotope contents of a major prairie aquifer in Central Manitoba, Canada. , Vienna, 379—398. Fritz, P. and Frape, S. K. (1982). Saline groundwaters in the Canadian shield— a first overview. 15, 179— 190. Gat, J. R. (1981). Groundwater. ' —l i h A A , Editors: Gat, J. R. and Gonfiantini, R. 223—240. Gudramovics, R. (1981). AW inyastigatigg Qf a deerg ggastal marina sabkna. Unpublished Masters Thesis, Mich. St. Univ. Hem J. D. (1985). Study and interpretation of the chemical characteristics of natural water. H, S. Gagl, Sgrygy — 1 P 22 4. Henderson, T. (1985). Geochemistry of ground-water in two sandstone aquifer systems in the Northern Great Plains in parts of Montana and Wyoming. U.§, Gagl. Sarvey Pr f i n 1 Pa 4 — Hendry, M. J. and Schwartz, F. W. (1990). The chemical evolution of ground water in the Milk River aquifer, Canada. Grgand Watar 2§;2, 253-261. Hitchon, B., Billings, G. K., and Klovan, J. E. (1971). Geochemistry and origin of formation waters in the western Canada sedimentary basin-III. Factors controlling chemical composition. G, g, Agta 35, 567-598. 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. Q, Q, Agta 33, 1321—1349. Horne, R. A. (1969). Marina ghamistgy. Wiley—Interscience, New York. 122 Houghton, D. (1928). W. ' ' ' ' 7- , Editor: Fuller, G. R., Lansing, the Michigan Historical Commission. Hull, L. C. (1984). Geochemistry of groundwater in Sacramento Valley, California. James B. R. Clark, C. J. and Riha, S. J. (1983). An 8- hydroxyquinoline method for labile and total aluminum in soil extracts W 4_, 893— 897. Jenne, E. A., Ball, J. W., Burchard, J. M., Vivit, D. V., and Barks, J. H. (1980). Geochemical modeling - Apparent solubility controls on Ba, Zn, Cd, Pb, and F in waters of the Missouri Tri—State mining area, in n W Editor: D. D. Hemphill. University of Missouri, 353-361. Joreskog, K. G., Klovan, J. E., and Reymont, R. A. (1976). Geochemical factor analysis. Mathgdsl 1a Gggmathamati 195 1. Elsevier Scientific, N. Y. Kehew, A. E. and Passero, R. N. (1990). pH and redox buffering mechanisms in a glacial drift aquifer contaminated by landfill leachate. r un a er 2 :5, 728—737. Kelly, W. A. (1936). Pennsylvania system in Michigan. Annual Rapgrt Qf thg Migh. Gagl, Sgrvay Div,, Dept. of n rv i n Pu 4. l 1 r1 4. Kendall and Coplen (1985). Analytigal thm, 57, 1437—1440. Kim, J. and Mueller, C. W. (1978). Introduction to factor analysis. Saga Uniygr sity Paper Ser l§§ on Quantitative Applications in the Social Sciences. Paper 13. Editor: E. M. Uslaner. Sage Publications, Beverly Hills. Klovan, J. E. (1975). R and Q— —Mode factor analysis. ggnggpts in Gagstatl istigs, Editor. R. McCammon. Springer—Verlag, New York, 21- 69. Lane, A. C. (1899 a). Water resources of the Lower Peninsula of Michigan. U 1 ur r—Su l Papgr 39. Lane, A. C. (1899 b). Lower Michigan mineral waters. U.S. 1 r r- l P er 1. 123 Leverett, F. and Taylor, F. B. (1915). The Pleistocene of Indiana and Michigan and the history of the Great Lakes. W. Long, D. T. and Larson, G. J. (1981). Field evidence for shale membrane filtration. W. 260. Long, D. T., Rezabek, D. H., Takacs, M. J., and Wilson, T. P. (1986). Geochemistry of groundwaters Bay County, Michigan. Mich. pant. Public Health; QED 38553. 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. 93911 I 1568- 1577. Mandle, R. J. (1989). Personal communication. Martin, H. (1936). The centennial geological map of the southern peninsula of Michigan. Martin, Helen M. (1958). Outline of the geologic history of Midland County. Migh. Dept. of ansatyatiga,, Gagl. sflryey 21 y. May H. M. Helmke, P. A. and Jackson, M. L. (1979). Determination of mononuclear dissolved aluminum in near- neutral waters. Chamita1_Gaglggy_24, 259- 269. McCree, J. M. (1950). The isotopic chemistry of carbonates and a paleotemperature scale. 5. Chem. Phy, 13, 849. McMahon, P. B., Williams, D. F. and Morris, J. T. (1990). Production and carbon isotopic composition of bacterial C02 in deep coastal plain sediments of South Carolina. Gr rgang Watgt 2§;5, 693- 702. Mook, W. G. (1972). Application of natural isotopes in groundwater hydrology. Qagl Qgi a an Mijntguw 51 Editor. D. Boschma. The Netherlands, 131— 136. National Atmospheric Deposition Program, (1981). National atmospheric deposition program report -precipitation chemistry. r ni r ' F ll'n l r Newcombe (1932). Unpublished manuscript. Okura T. Goto, K., and Yotuyanagi, T. (1962). Forms of aluminum determined by an 8- -quinolinolate extraction method. Malytigal Cheml istty 34 581- 582. 124 Payne, B. R. (1972). Isotope hydrology. Adyante§_;n ' . Editor: V. TeChow. Academic Press, New York, 95-138. Perry, E. C. Jr., Grundl, T., and Gilkeson, R. H. (1982). Hydrogen, oxygen and sulfur isotopic study of the groundwater in the Cambrian- Ordovician aquifer system of northern Illinois. leetepe Editors: E. G. Perry, Jr. and C. W. Montgomery. Northern Illinois Univ. Press, Dekalb, Illinois, 35—45. Plummer, L. N. (1977). Defining reactions and mass transfer in part of the Floridan aquifer. Watet_3e§etttee W, 801-812. Plummer, L. N., Jones, E. F. and Truesdell, A. H. (1984). WATEQF - A FORTRAN IV version of WATEQ, a computer program for calculating chemical equilibrium of natural waters. — R ' ' , 76—13. Pringle, G. H. (1937). Geology of Arenac County, Michigan. WWW- Rhodehamel, E. C. (1951). An igtetptetatien ef the pre— Pl 1 n m h l r i f h in Lewland. Masters Thesis, Mich. St. College of Agriculture and Applied Sciences. Rightmire, C. T. and Hanshaw, B. B (1973). Relationship between the carbon isotope composition of soil C02 and dissolved carbonate species in groundwater. AmI fieephysieal gnien, 958- 967. Rose, S. and Long, A. (1989). Dissolved inorganic carbon in the Tucson Basin aquifer. Qtetag Water 27;;, 43-49. Schopf, T. J. M. (1980). Paleegeanegraphy. Harvard Univ. Press, Cambridge. Schwartz, F. W. (1974). Origin of chemical variations in groundwaters from a small watershed in southwestern Ontario. Can. J. Earth Science 11. 893-904. Siegel, D. I. and Mandle, R. J. (1984). Isotopic evidence for glacial meltwater recharge to the Cambrian—Ordovician aquifer, North—Central United States. a rna Researgh 22, 328-335. Siegenthaler, U. (1979). Stable hydrogen and oxygen isotopes in the water cycle Leetetea__a_;eetepe . Editors. E. Jaeger, et al. Springer— Verlag, Berlin, Federal Republic of Germany, 264-273. 125 Skougstad, M. W., Fishman, M. J., Friedman, L. C., Erdmann, D. E., and Duncan, S. S. Editors (1985). Methods for determination of inorganic substances in water and fluvial sediments. — Slayton, D. F. (1982). We Masters Thesis, Mich. St. Univ. Sorenson, H. O. and Segall, R. T. (1975). Natural brines of the Detroit River Group, Michigan basin. Fettth Sympeaitm_ea_5alt Editor: A. H. Coogan. Northern Ohio Geol. Society, Inc., Cleveland, OH, 91—99. Stark, G. R. and McDonald, M. G. (1980). Ground water of coal deposits, Bay County, Michigan. - ' R - . Stoessell, R. K., Ward, W. C., Ford, B. H. and Schuffert, J. D. (1988). Water chemistry and calcium carbonate dissolution in the mixing zone, coastal Yucatan Peninsula, Mexico. Geel, See, Am, Balletln lQl, 159-169. Stumm, W. and Morgan, J. J. (1981). Agtatle Chemistty. Wiley-Interscience, New York. Thorstenson D. C., Fisher D. W. and Croft, M. G. (1979). The geochemistry of the Fox Hills— basal Hell Creek aquifer in southwestern North Dakota and northwestern South Dakota. Water Reseerees Reseatgh l5;§, 1479— 1498. Turner R. C. (1969). Three forms of aluminum in aqueous systems determined by 8— —quinolinolate extraction methods. Canadian geatn a1 ef th emistty 47, 2521— 2527. Twenter, F. R. (1987). Maps: Groundwater yields from bedrock. andi Groundwater yields from glacial deposits. An In M hi n' r R r . Institute of Water Research, Mich. St. Univ. Twenter. F. R. and Cummings, T. R. (1985). Quality of groundwater in Monitor and Williams Townships, Bay County, Michigan. 1 r r—R ur In i i n R r -411 U. S. Environmental Protection Agency (1987). 40 Code of Federal Regulations Chap. 1, Part 264, Appendix IV (July 1, 1987). Vanlier, K. E. (1963). Ground-water resources of the Alma area, Michigan. U, S. Qeel, Sarvey Water-Supply Paper 1519—E 126 Velbel, M. A. (1992). Seawater-freshwater mixing and clay— mineral stability in sandstones. Submitted to Applied Geochemistry, January 1992. Weast, R. C., Editor (1979). Handbook of chemistry and physics. CRC Press, Boca Raton, Florida. Western Michigan University (1981). Hydrogeologic Atlas of Michigan. Dept. of Geology, College of Arts and Sciences, Western Michigan University, Kalamazoo, MI. Wigley, T. M. L., Plummer,.L. N. and Pearson, F. J. Jr. (1978). Mass transfer and carbon isotope evolution in natural water systems. G. C. Acta 42, 1117-1139. Wilson, T. P. and Long, D. T. (1984). The behavior of bromide during the dissolution of halite at 25'C and 1 atmosphere pressure. Geol. Soc. Am. Abstracts with Programs, 697. Wilson, T. P. and Long, D. T. (1986). The geochemistry of Michigan basin brines. American Chemical Society Abstract and paper presented at the Annual Meeting, Anaheim, CA. Wood, W. W. and Low, W. H- (1986). Aqueous geochemistry and diagenesis in the eastern Snake River Plain aquifer system, Idaho. Geol. Soc. Am. Bulletin 97, 1456-1466. I’IICHIGQN STQTE UNIV. LIBRQRIES IIW ll H \H Hlll 31293 09 69679 .H.wu‘.. q»......~ .2". ,‘ ~ ‘ . ".. ‘ “““-‘-' « H “'w v w .'. ' . «.1. . ..,. ..... .. . 3:: llll O --- , . .n