Stratigraphic Framework and Landsystem Correlation for Deposits of the Saginaw Lobe, Michigan, USA

Transcription

1 Western Michigan University ScholarWorks at WMU Master's Theses Graduate College Stratigraphic Framework and Landsystem Correlation for Deposits of the Saginaw Lobe, Michigan, USA Ivan R. Guzman Western Michigan University, Follow this and additional works at: Part of the Geomorphology Commons, Glaciology Commons, and the Sedimentology Commons Recommended Citation Guzman, Ivan R., "Stratigraphic Framework and Landsystem Correlation for Deposits of the Saginaw Lobe, Michigan, USA" (2014). Master's Theses This Masters Thesis-Open Access is brought to you for free and open access by the Graduate College at ScholarWorks at WMU. It has been accepted for inclusion in Master's Theses by an authorized administrator of ScholarWorks at WMU. For more information, please contact

2 STRATIGRAPHIC FRAMEWORK AND LANDSYSTEM CORRELATION FOR DEPOSITS OF THE SAGINAW LOBE, MICHIGAN, USA Ivan R. Guzman, M.S. Western Michigan University, 2014 Since the time of the Last Glacial Maximum (LGM) the south-central portion of the Lower Michigan Peninsula has been subject to several glacial advances and retreats by the Saginaw lobe. As part of the U.S Geological Survey Great Lakes Geological Mapping Coalition projects, several rotosonic borings were drilled between 2006 and 2013 in Barry, Kalamazoo and Calhoun Counties. Gamma ray logs and textural analyses were completed for each core. Five of these borings were selected according to their diamicton (till) content and correlated using water well logs and surficial geology maps. Glacial deposits such as diamicton serve as evidence of glacial advance/retreat, and are usually present as nearly continuous layers of sediments. Analysis of these layers affords the ability to accurately correlate these types of sediments across an area. Three cores, BA and BA-09-02, KA were drilled along the Kalamazoo moraine, each one containing 1 to 3 diamicton units separated by lacustrine sediments. The last two cores, CA and KA were drilled on a drumlinized till plain; both contain 2 to 4 diamicton units separated by outwash sediments. These diamicton units indicate the presence of at least one major and two minor advances/retreats of the Saginaw Lobe.

3 STRATIGRAPHIC FRAMEWORK AND LANDSYSTEM CORRELATION FOR DEPOSITS OF THE SAGINAW LOBE, MICHIGAN, USA by Ivan R.Guzman A thesis submitted to the Graduate College in partial fulfillment of the requirements for the degree of Master of Science Geosciences Western Michigan University June 2014 Thesis Committee: Alan E. Kehew, Ph.D., Advisor Rama V. Krishnamurthy, Ph.D. Upul B. Attanayake, Ph.D.

4 Copyright by Ivan R. Guzman 2014

5 ACKNOWLEDGMENTS I would like to thank Dr. Alan Kehew for his guidance, advice and patience during the realization of this project. I am grateful for the help of my committee members, Dr. Rama Krishnamurthy for his continuous help and support in the early stages of my project, and Dr. Upul B. Attanayake for collaborating with me and offering me advice. I also offer my sincere appreciation to my country, the Dominican Republic for providing me with a scholarship as well as the necessary funding to live and study in the United States. I would also like to thank the WMU Department of Geosciences, the WMU Graduate College, the W. David Kuenzie Research Fund and the U.S Geological Survey Great Lakes Geological Mapping Coalition from whom I received funding for this project. I would like to express my gratitude to Stephanie Ewald, Abdou Mohammed, Derrick Lingle, Sita Karki, Todd White and many others for their various contributions to this project. I thank Racha El Kadiri for her support and encouragement, which helped me to overcome many obstacles during my graduate career. Lastly, I would like to thank my family for their unconditional support and patience. They always believed in me and kept me motivated to stay on track and reach my goals. Ivan R. Guzman ii

6 TABLE OF CONTENTS ACKNOWLEDGMENTS... LIST OF TABLES... LIST OF FIGURES... ii vi vii CHAPTER I. INTRODUCTION... 1 Previous Investigations... 2 Site Description... 4 II. GEOLOGY... 7 Bedrock Geology... 7 Wisconsin Glaciation in Michigan Saginaw Lobe Landsystems III. GEOCHRONOLOGY Glacial Geochronology in Michigan Radiocarbon Dating on Glacial Till: Main Concerns IV. METHODS Particle Size Analysis Atterberg Limits Bulk Organic Carbon Assay Inorganic Carbon Assay Map and Cross Section Data iii

7 Table of Contents Continued CHAPTER V. RESULTS Landsystem 3: North of the Thornapple Valley BA Landsystem 2: South of the Thornapple and North of the Kalamazoo Valleys BA KA Landsystem 1: South of the Kalamazoo Valley CA KA Diamicton Clay Consistency Cross Section A-A Radiocarbon and δ 13 C Analyses VI. DISCUSSION Core Interpretations Landsystem Correlation across the Saginaw Lobe VII. CONCLUSIONS APPENDICES A. Particle Size Analysis Results B. Atterberg Limits Results C. Bulk Organic Carbon Results iv

8 Table of Contents Continued D. δ 13 C Results BIBLIOGRAPHY v

9 LIST OF TABLES 1. Boreholes: Coordinates, Depths and Elevations Interpreted Water Well Lithology Diamicton Clay Consistency Data Results Diamicton Bulk Organic Carbon Data Diamicton Carbonates Data Carbon 14 Data Results vi

10 LIST OF FIGURES 1. Study area and borehole locations in southern Michigan Bedrock geology, south-central portion of the Lower Michigan Peninsula Interaction between the Saginaw, Huron-Erie, and Lake Michigan Lobes Landsystems of the Saginaw Lobe, Southern Peninsula of Michigan Wentworth grain-size scale for sediments Clay consistency and Particle size analysis tools Logplot Diagram of BA showing lithology, gamma ray signature, and grain size distribution Matrix texture (<2.00 mm) of all samples in BA Logplot Diagram of BA showing lithology, gamma ray signature, and grain size distribution Matrix texture (<2.00 mm) of all samples in BA Logplot Diagram of KA showing lithology, gamma ray signature, and grain size distribution Matrix texture (<2.00 mm) of all samples in KA Logplot Diagram of CA showing lithology, gamma ray signature, and grain size distribution Matrix texture (<2.00 mm) of all samples in CA Logplot Diagram of KA showing lithology, gamma ray signature, and grain size distribution Matrix texture (<2.00 mm) of all samples in KA Location of cross section A-A vii

11 List of Figures Continued 18. Cross Section A-A Matrix texture (<2.00 mm) of diamicton samples in BA Matrix texture (<2.00 mm) of diamicton samples in BA Matrix texture (<2.00 mm) of diamicton samples in KA Matrix texture (<2.00 mm) of diamicton samples in CA Matrix texture (<2.00 mm) of diamicton samples in KA Matrix texture (<2.00 mm) of diamicton samples from cores: BA-09-02, BA-10-02, KA-12-02, CA and KA Cross Section A-A : Proposed Correlation of Sediments viii

12 1 CHAPTER I INTRODUCTION Since the time of the Last Glacial Maximum (LGM), sediments from three lobes of the Laurentide Ice Sheet have dominated the Lower Peninsula of Michigan. After the LGM, the first lobe to retreat was the Saginaw lobe, but not before experiencing a series of small advances and retreats (Kehew et al, 2012a). Today, only small portions of the surface landscape of the Saginaw Lobe has been mapped, and very little is known about the characteristic of the region s subsurface deposits as well as how these deposits correlate with the advance and retreat of the ice. Recently, studies have examined the subsurface deposits of the Saginaw Lobe, but only a few borings have reached bedrock and fewer still have mapped the subsurface stratigraphy of these glacial deposits. This study is intended to work out the subsurface stratigraphy of the Saginaw Lobe. For that purpose, five rotosonic boreholes were drilled in the counties of Barry, Kalamazoo and Calhoun, Michigan. These borings were selected according to their thickness and stratigraphy. The principal objective of this research is to identify and map major stratigraphic units across the south-central portion of the Lower Peninsula of Michigan related to glacial advances/retreats and investigate if these units can be correlated through the area. A second objective of the study is to characterize the glacial drift stratigraphy by interpretation and documentation of lithological units in the Saginaw Lobe. Cores drilled along the Saginaw Lobe have shown diamicton a

13 2 few meters above the bedrock. This study will demonstrate if the diamicton units correlate with each other and if the units are related to glacial advances/retreats. Water well logs and a bedrock topographic map are used, as well as textural, consistency and geochronology analysis to more clearly understand the glacial deposition of the Saginaw Lobe. Previous Investigations The first studies to map the surficial deposits of the Saginaw Lobe, were performed by Leverett and Taylor (1915). Based mostly on topographic analysis, these studies describe Michigan surficial deposits by identifying and mapping various glacial landforms, including the end moraines of the Saginaw Lobe, which they associated with ice marginal positions. Martin (1955) compiled the first revised map of the surficial geology of Pleistocene glacial deposits in Michigan. The map shows the general distribution of glacial landforms across the southern Peninsula of Michigan. Later Farrand and Bell (1982) published a revised map of the Quaternary geology of southern Michigan that includes textural descriptions derived from the previous soil surveys. Monaghan and Larson (1986) identified and correlated two upper till units in south central Michigan, the Bedford and Fulton tills, using grain size distribution and clay mineralogy analyses. Using only six 9-meter boreholes and other surface samples, the authors traced the Bedford till from the Lansing Moraine to the Kalamazoo Moraine and the Fulton till from the Lansing Moraine to the Tekonsha Moraine. From this investigation they reach the conclusion that the Kalamazoo

14 3 Moraine of the Saginaw Lobe could be correlated with the Kalamazoo Moraine of the Lake Michigan Lobe, and to the Powell Moraine of the Huron-Erie Lobe. In the last fifteen years, new research has incorporated stratigraphy, morphology, and clay mineral content of diamicton units into glacial studies (Taylor et al., 1998; Fisher and Taylor, 1999; Kozlowski, 1999; Kehew et al., 1999; Kozlowski et al., 2001; Fisher et al., 2003; Kozlowski et al., 2004; Kozlowski et al. 2005; Kehew et al 2012a). Woolever (2008) focused on the surface geology created by subglacial meltwater of the Saginaw lobe. He concluded that several linear valleys containing eskers in Barry County were tunnel valleys produced by meltwater erosion at the base of the glacier. This finding helped to further understand the dynamic in drainage systems of the ice sheet and how tunnel valley and eskers were formed. Barnes (2010) analyzed till samples looking for systematic variation in the organic matter content that could potentially explain the high iron concentration in southwestern Michigan groundwater. He found that high organic carbon content was most likely producing the higher iron concentration in the groundwater. Recently Kehew et al (2012a) proposed a four landsystems approach to classify the surface terrain of the Saginaw Lobe in Michigan according to the sediment/landform relation in the lobe. This made it possible to interpret the glacial dynamic and interactions between the ice and the substrate in the Saginaw Lobe. Kehew et al (2012b) also described the role of subglacial meltwater flow systems in the formation of tunnel valleys and concluded that subglacial water flow in tunnel valleys played a crucial role on the drainage and stability of the ice sheets. Ewald

15 4 (2012) using data from six cores, reconstructed the depositional environment for the glacial and lacustrine sediments present in Barry and Calhoun counties. She concluded that several intervals within these borings were associated with a proglacial or subglacial lake not previously identified in the Lower Peninsula of Michigan. Site Description The area of study in this analysis is located in the south-central portion of the Lower Peninsula of Michigan between the counties of Barry, Kalamazoo and Calhoun (Figure 1). Between 2009 and 2013, five rotosonic boring were drilled in several sites along these areas. The boring sites were selected to develop a generalized stratigraphic framework and to determine the stratigraphy of tunnel valleys in the study area.. The first group consists of two boreholes drilled in Barry County; BA drilled in 2009 and BA drilled in The second group was drilled in Calhoun County and consists of one borehole; CA drilled in The third group consists of two rotosonic boreholes drilled in Kalamazoo County; KA was drilled on 2012 and KA drilled in Most of the boreholes are located within tunnel valleys or glacial uplands to aid us in better understanding the overall drift stratigraphy in the area (Table 1).

16 Figure 1. Study area and borehole locations in southern Michigan. Borehole are labeled and marked with red circles. 5

17 6 Table1 Boreholes: Coordinates, Depth and Elevations Core ID Latitude Longitude Depth (m) Elevation (m) BA BA CA KA KA

18 7 CHAPTER II GEOLOGY Bedrock Geology The glacial deposits across the Michigan Southern Peninsula are mainly composed of material eroded from bedrock or previous sediment along the path of ice movement from ice-sheet centers near Hudson Bay (Monaghan and Larson, 1986; Dodson, 1993; Gardner, 1997; Flint 1999). Understanding the bedrock geology is crucial towards gaining a fuller understanding of the stratigraphic settings of these glacial sediments. The Michigan Basin is formed of sedimentary rock units of which 1% comes from Pennsylvanian age, 5% from Mississippian, 16% Devonian, 30% Silurian, 21% Ordovician and 27% Cambrian (Cohee 1965; Dorr and Eschman 1970). These sedimentary rocks can be grouped in a series of formations (Figure 2). The bedrock formations are part of the Michigan Basin, which is an elliptical intracratonic basin, located against the southern margin of the Canadian Shield (Gillespie et al. 2008). Bedrock formations subcrop beneath the glacial deposits in a series of irregular concentric rings (Figure 2). The ages of these formations range from Cambrian at the margins of the basin to Pennsylvanian in the center, capped by a small area of Jurassic rocks. The basin strata are mainly dominated by dolomite and limestone, with a significant presence of siliciclastics (shale, sandstone and siltstones)

19 8 and evaporites (gypsum, halite) (Dorr and Eschman, 1970; Howell and Van der Pluijm, 1999). During the Pleistocene, bedrock topography played a crucial role in the advance of the Laurentide Ice Sheet sub lobes by controlling their boundaries (Kehew et al, 2012a; Ewald 2012). In the course of this time period, bedrock formations across the Michigan basin were buried beneath thick unconsolidated glacial sediment carried by the continental ice mass. Studies made by Dorr and Eschman, (1970) and Harrell et al, (1991) have estimated that the thickness of these glacial deposits varies in different locations ranging from 0 to more than 305 m (1000 ft). The area of study is located in the southwestern part of the Southern Peninsula of Michigan and includes the counties of Barry, Calhoun and Kalamazoo (Figure 2). The uppermost bedrock formations in this area are part of the Mississippian System, which extends northward from northern Indiana and northwestern Ohio to cover most of the northern counties of Michigan s Southern Peninsula. The system is largely dominated by shallow marine terrigenous detritus (mostly shale), followed by sandstones, and then carbonates and evaporates (Harrell et al, 1991). Several bedrock formation subcrops within the study area such as the Marshall Sandstone, Coldwater Shale, Bayport Limestone and Michigan Formation (Figure 2). Coldwater Shale. The Coldwater Shale is located in the southwest part of the study area. The Coldwater Shale subcrops predominantly in Kalamazoo County. However, Coldwater Shale also exists in Calhoun County, as well as a small portion in Barry County (Figure 2). The Coldwater Shale is mostly gray to bluish shale and

20 Figure 2. Bedrock geology, south-central portion of the Lower Michigan Peninsula. Counties are outlined in black and labeled. 9

21 10 can be found interbedded with limestone and dolostone (Harrell et al, 1991). According to Dorr and Eschman (1970) the fine-grained mud of the Coldwater Shale was deposited at the beginning of the Early Mississippian after the Lower Peninsula became an offshore marine environment. Marshall Sandstone. The Marshall Formation is located directly northeast of the Coldwater Shale and it crosses through Calhoun and Barry Counties (Figure 2). The Marshall Sandstone is mostly formed by gray, pink and red sandstones and siltstones, with an abundant clay matrix (Dorr and Eschman 1970; Harrell et al, 1991). The Marshall Sandstone was deposited after a major regression in the seas at the closure of Early Mississippian (Dorr and Eschman 1970). Michigan Formation. The Michigan Formation directly overlies the Marshall Sandstone. In the area of study, this formation can be found in Barry County (Figure 2). According to Dorr and Eschman (1970), the Michigan Formation is a marine deposit of shale, gypsum, dolomite, limestone and small intervals of sandstone, which was formed in the Late Mississippian by a transgression of the seas in Michigan. The gray shale of the Michigan Formation is usually intebedded with sandstone in the southern and central part of the basin, and carbonates and evaporate interbeds in the west and north section (Harrell et al, 1991). Bayport Limestone. The Bayport Limestone is the youngest of the Mississippian rocks and is located to the northeast between the Michigan Formation and the Saginaw Formation (Pennsylvanian). In the area studied, it can be found in Barry and Calhoun Counties (Figure 2). It is comprised mainly of Gypsiferous,

22 cherty, sparsely fossiliferous dolostone interbedded with some sandstone (Harrell et al, 1991). 11 Wisconsin Glaciation in Michigan The last glaciation had a tremendous impact on modern topography and glacial landforms of southwestern Michigan. This process began in the Pleistocene approximately 2 million year ago, when the climate in the northern part of the continent changed. During this period, glaciers advanced and retreated about twenty times until their last major advance and retreat, called the Wisconsinan Glaciation (Farrand, 1988). During this period, the Laurentide Ice Sheet experienced its biggest expansion in North America. The Wisconsin glaciation is divided into three sub episodes: the early (Ontario), Middle (Elgin) and the Late (Michigan), based on the extent of the ice margins (Johnson et al., 1997). During the early sub episode of the Wisconsin glaciation between 65,000 79,000 yr. BP, ice advanced from the northeast and dammed a lake in the Ontario Basin (Karrow 1984; Larson et al., 2001). Later during the middle sub episode, between 65,000 and 35,000 yr. BP, the ice sheet extended from the Ontario Basin to somewhere near the Finger Lakes region of New York, where it terminated in a proglacial lake (Karrow 1984; Larson et al., 2001). Finally came the late Wisconsin sub episode, between 35,000 and 10,000 yr. BP (Larson et al., 2001).

23 12 The late Wisconsin was characterized by a series of major and minor advances and retreats of the ice sheet (Grimley 2000; Larson et al., 2001). This period was one of the most significant episodes because it is when all or most of the drift deposits in the Lower Peninsula of Michigan were deposited during and following the Late Glacial Maximum. After the Late Glacial Maximum, approximately 20, C BP, the margin of the Laurentide Ice Sheet advanced in a series of sub lobes which at some point covered the entire Great Lakes watershed (Dyke et al., 2002, Larson et al., 2001, Ewald 2012). Three major lobes developed over Michigan: the Lake Michigan, Saginaw, and Huron- Erie Lobes (Figure 3). After the Late Glacial Maximum, the southern margin of the Laurentide Ice Sheet began a general retreat northward into the Great Lakes watershed (Larson et al., 2001). The Saginaw Lobe was the first of the three major lobes to readvance into Michigan and northern Indiana (Kehew et al., 2005). The lobes were asynchronous, and when the Saginaw Lobe began to wane or retreat the Lake Michigan and Huron- Erie lobes advanced (Kehew et al., 2005) (Figure 3). The Kalamazoo Moraine is a result of this interaction between the Saginaw and Lake Michigan Lobes. This moraine appears to represent a prominent ice-marginal position and is attributed to the Saginaw and Lake Michigan Lobe (Kehew et al., 2005). This current study is concerned with glacial landforms and sediment deposited by the Saginaw Lobe in Lower Peninsula of Michigan.

24 Figure 3. Interaction between the Saginaw, Huron-Erie, and Lake Michigan Lobes (Kehew et al., 2005). Re-advance of the Saginaw Lobe after LGM approximately 21,000 yr BP (left). Retreat of the Saginaw Lobe approximately 15,000 16,000 yr BP (right). 13

25 14 Saginaw Lobe Landsystems Few studies have examined the Saginaw Lobe glacial landforms and even fewer have attempted to characterize its subsurface deposits. A general classification was made by Colgan et al (2005), who mapped the landsystems of the entire southern Laurentide Ice Sheet margins. Recently, based on this classification, Kehew et al (2012a) divided the surficial deposits of the Saginaw Lobe into four distinct landsystems according to their morphology and depositional relationships (Figure 4). Cores analyzed in this research will serve to correlate the glacial deposits with glacial advances/retreat through these landsystems. Landsystem 1 is composed of the Sturgis Moraine, which is a terminal/recessional moraine, and a drumlinized till plain to the northeast. The Sturgis Moraine is composed of glaciofluvial sediment with thick alluvial fans that slope off the moraine (Kehew et al. 2012a). Tunnel valleys are present in this landsystem, cutting and extending beyond the moraine. The drumlinized till plain is mostly formed from sandy diamicton and is bounded by the bedrock contact of the Coldwater Shale and overlying Marshall Sandstone to the north (Dodson, 1985; Kozlowski, 1999, Kehew et al 2012a, Ewald 2012). Cores CA and KA were both drilled south of the Kalamazoo Valley on drumlins. Landsystem 2 is bounded by of the Thornapple Valley in the north, the Kalamazoo Valley in the south and contains the Kalamazoo Moraine. The Thornapple Valley, a west flowing river valley, served as a channel to carry meltwater from the

26 15 Huron-Erie lobe (Kehew et al. 2012a). The Kalamazoo Valley, a major trench like valley, begins as a network of tunnel valleys incised into the limestone bedrock with a floor covered by numerous glacial boulders which are a product of the down cutting of the overlying glacial drift (Kehew et al. 2012a; Kozlowski et al. 2005). Landsystem 2 is mainly dominated by the Kalamazoo Moraine, which was first described by Leverett and Taylor (1915), and it includes in its topography a subglacial element, like tunnel valleys and eskers, which are covered by supraglacial sediment. The landforms and sediments are believed to be the product of ice stagnation and collapse, and include hummocks, kames and ice walled lake plains (Kehew et al. 2012a). Cores BA and KA were drilled in landsystem 2. Landsystem 3 is located north of Thornapple Valley. The area is mainly composed of open tunnel valleys and eskers. The Thornapple Valley is interpreted to have carried meltwater from the Huron-Erie to the east (Kehew et al. 2012a). Core BA was drilled in this landsystem north of the Thornapple River Valley. Landsystem 4 is mainly composed of recessional moraines formed from backwasting of the Saginaw Lobe. For this research the area studied only covers landsystems 1, 2 and 3.

27 Figure 4. Landsystems of the Saginaw Lobe according to Kehew et al (2012a), Southern Peninsula of Michigan. Red lines represent the western boundary of the Saginaw Lobe. Boundaries between landsystems are represented by black lines. 16

28 17 CHAPTER IV GEOCHRONOLOGY Glacial Geochronology in Michigan Geochronology of glacial events in Michigan was the subject of much speculation before the development of Carbon 14 ( 14 C) radioactive dating methods in the mid 1940 s. The 14 C method enabled the dating of organic material up to a maximum of about 50,000 to 70,000 years old. As result of this technology, portions of the Wisconsinan glacial time scale can be accurately dated. Sadly, organic material from older glacial time like, the Illinoian deposits can t be dated because the age limit goes beyond the reach for 14 C dating (Dorr and Eschman 1970). Some studies have been conducted in an attempt to accurately date glacial advances/retreats from different glacial landforms. Organic remains, buried by till or outwash, have assisted in dating some of these glacial advances (Dreimanis, 1977). Moraines or sediments associated with moraines have been extremely useful for radiocarbon dating. The 14 C methods have provided both maximum and minimum ages to sediments below, within, and above the moraines (Briner, 2011). One of the oldest glacial deposits dated in Michigan thus far corresponds to an unweathered, unnamed till unit beneath the John Ball State Park organic bed in Grand Rapids. Samples from this organic bed yielded 14 C ages between the ranges of 39,900 to 51,000 yr. BP (Zumberge and Benninghoff, 1969; Eschman and Mickelson 1986).

29 18 The unnamed till was thought to be from the early to middle Wisconsin age and is believed to antedate deposition of the organic sediment by a short interval of time (Eschman and Mickelson, 1986). According to Dreimanis (1977) such dates are rare, especially for the interval between 13,000 and 17,000 B.P. which is close to the Late Glacial Maximum. Recent studies from Colgan (2013) and Lingle (2013) dated a new organic deposit beneath Hemlock Crossing Park in Ottawa. The organic sands were located between two till unit from core OT and yielded 14 C ages between the ranges of 41,920 to 42,950 yr. BP. This implies, according to Colgan (2013), that there is a significant amount of glacial sediment older that ~42,000 yr. BP around the area of Ottawa and surrounding counties. Most of the radiocarbon dating made on glacial landforms has been conducted on wood surfaces or other plant remains. Diamicton (till) unit have only been directly dated recently (Kehew et al, 2009). This study will obtain some radiocarbon dates on till units from three rotosonic boreholes; BA-09-01, BA and OT-12-01, and use these dates to correlate the till units. Core BA was analyzed by Ewald (2012) and OT by Lingle (2013), for their master s thesis. Radiocarbon Dating on Glacial Till: Main Concerns Glacial deposits like diamicton document glacial advances and/or retreats. Therefore, performing studies obtaining radiocarbon dates of diamicton units represents a potential source of information about glacial events. However, there are several potential problems regarding dating organic matter in soils. One of the main

30 19 problems in dating soil bulk organic carbon is that the 14 C ages obtained are too young due to contamination by recent contribution of carbon. The formation of soil organic matter is an ongoing process, in which fresh carbon is continuously incorporated at different rates and in any size fraction (Wang et al. 1996). Because of this, 14 C dates have been interpreted as minimum ages (Perrin et al., 1964; Scharpenseel, 1971a,b, 1972, 1976; Cherkinsky and Brovkin, 1991, Wang et al. 1996). The landscape in the Lower Peninsula of Michigan consists mainly in glacial drift, and organic carbon in these sediments is concentrated in the diamicton units (Kehew et al, 2009). The Lower Peninsula of Michigan has been subject to several glacial advances/retreat during the late Wisconsinan and multiple sources of organic carbon are present in the glacial drift; mid-wisconsin wood, late-wisconsin soil and vegetation as well as clasts of coal are also disseminated in these deposits (Kehew et al, 2009). Mid-Wisconsin and coal organic carbon could result in bias toward older 14 C ages. Radiocarbon dates from the bulk organic carbon of diamicton could serve as an important source of information to correlate and estimate ages in different areas, and assist in determining past glacial events. Core KAL analyzed by Barnes (2007), were used to correlate dates from the boreholes analyzed in this study.

31 20 CHAPTER IV METHODS The data used in this research comes from five rotosonic cores. The cores were drilled as parts of the Michigan Geological Survey projects funded by the U.S Geological Survey Great Lakes Geological Mapping Coalition with the purpose of providing more detailed analysis of glacial geology throughout the region. These cores were then taken to the Soil Laboratory at Western Michigan University (WMU) for grain size distribution analysis. Diamicton samples from some of the cores were collected and taken to the WMU Engineering College Geotechnical Laboratory for consistency analysis (Atterberg Limits). Samples for 14 C dating were also collected to be pretreated in the Geosciences Dept. Isotope Laboratory and have then sent to the DirectAMS Laboratory in Seattle to be dated. The following methods and procedures were used to accomplish the proposed research: Particle Size Analysis Five borings were chosen for textural/particle size analysis to determine the grain size distribution. These tests were done according to the method modified from Bowles (1978) to separate the 2µm clay particles. Sieves were selected according to the ASTM protocol E 11. Similar techniques of analysis were used by Gardner (1997),

32 21 Flint (1999), Wong (2002), Beukema (2003), Barnes (2007), Woolever (2008) and Ewald (2012). Results are presented in Appendix A. Soil samples were taken about every 2 to 3 foot interval or when a change in the soil layer was noted. The sample weight varies depending on the water content at the time of collection and grain size distribution, but approximately 400 to 500 g of sample was collected. As stated above, two sieving methods were used depending on the type of samples: for coarse samples, the dry sieve method and for fine samples, the wet method. The following procedure was used to analyze the grain size distribution in coarse samples: 1. Approximately 400 to 500 grams of the sample was collected, placed in an aluminum pan and dried in the oven at 105 o C for at least 24 hours. 2. A pestle and a porcelain mortar were used to disaggregate the dried samples. A rubber tip pestle was used to gently disaggregate the sample (porcelain pestle was used for sample with high clay content). After this the sample was weighed. 3. Seven sieves were stacked in order from: #5, #10, #18, #35, #60, #120, #230 and bottom pan (order from the coarsest to the finest). Approximately 400 grams of the disaggregated sample was poured into the stack of sieves and covered with a top pan. 4. The stack of sieves was placed in the Ro-tap mechanical sieving device to be agitated for 10 minutes.

33 22 5. After 10 minutes, the amount of particles retained in each sieve was weighed and recorded. 6. The fines (silt and clay) from the bottom pan were stored in the oven and then separated using the gravitational separation method based on Stokes Law (Hillel, 1998). The following sieving procedure was used to analyze the grain size distribution in samples with high content of clay and silt: 1. Approximately 400 to 500 grams of sample was collected, placed in an aluminum pan and dried in the oven at 105 o C for at least 24 hours. 2. The sample was disaggregated with a porcelain pestle and about 450 grams were taken and poured in a metal cup with tap water. 3. The sample in the metal cup was then agitated in a sediment stirrer machine for 1 minute, then with a stir rod for another minute and washed through a #230 sieve with a bottom pan to separate sand and gravel from clay and silt. 4. The sample remaining in the #230 sieve was the sand and gravel. The content was then rinsed in a separate container and allowed to dry in the oven to be sieved again using the dry method explained above. 5. The sample captured in the bottom pan was the clay and silt. The content was then rinsed into a separate container and dried in the oven at 105 o C for at least 24 hours. After drying, the sample was then mixed with the

34 23 fines that remained from the sieving of the sand and gravel portion. From this mix 10 grams was collected for the silt and clay separation. The following procedure was used for the silt and clay separation (modified by Ewald, 2012): 1. Approximately 10 grams of the sample fines were collected and placed in a 1000 ml beaker (if the sample was less than 10 grams, then the entire sample was used). 2. An alkaline solution of 0.5% sodium hexametaphosphate (Na 6 O 18 P 6 ) was made to act as a deflocculant, and 700 ml of this solution was poured in the beaker already containing the 10 gram sample. 3. The sample with the solution was then agitated in a sediment stirrer machine for 10 seconds and then placed in an ultrasonic vibration device for 20 minutes. 4. After 20 minutes, the sample was allowed to settle for 2 hours (120 min), during this time the clay particles remained in suspension while the silt settled to the bottom of the 1000 ml beaker. 5. After 2 hours the suspended clay and the settled silt were poured into separate weighed aluminum pans and placed in the oven at105 o C for 24 hours. The next day the weight of the aluminum pans was recorded.

35 24 Particle Length Sieve Size ASTM No. Grade Class (mm) (Ø) (U.S. Standard) Pebble Gravel Granule Very Coarse Coarse Medium Sand Fine Very Fine Coarse P Medium a Fine Silt n Very Fine Clay Figure 5. The Wentworth grain-size scale for sediments: Wentwoth size classes, phi (Ø) units and U.S Standard sieve (modified from USGS, 2006)

36 25 Atterberg Limits The Atterberg Limits were used to determine the clay consistency in 18 diamicton samples from three of the five borings. These borings were chosen based upon location (landsystem), and relevance to the study. The test was performed according to the American Society for Testing and Material (ASTM, 2010) protocols to measure moisture content at which the sample changes from semi-solid to plastic state (Plastic Limit) and from plastic to liquid state (Liquid Limit). Engineers have been using this test since the 1900s for correlations of physical soil parameters and soil identification (Das, 2010). Casagrande (1932) conducted several studies using the liquid limit to correlate the plasticity index (PI) of different soil types. The following procedure was used to prepare the samples for the liquid and plastic limit test: 1. Approximately 200 grams of the sample was taken and disaggregated using a porcelain pestle and mortar. 2. The sample was poured in a #40 sieve with a bottom pan and sieved. This process was repeated until about 120 to 200 grams of sample was retained in the bottom pan. 3. The sample was weighed, recorded and then poured into a porcelain dish for liquid and plastic limit test. The liquid limit (LL) according to the ASTM (2010) is determined by performing trials in which a portion of the soil sample is spread in a brass cup and a groove is cut at the center with a grooving tool. Then with a mechanical device the

37 26 cup is lifted and then dropped until the groove is closed (Figure 5). The number of drops and the moisture content is then recorded. The method used was the ASTM One-Point Liquid Limit - Method B. The following procedure (ASTM, 2010) was used to prepare the samples for the liquid limit (LL): 1. Deionized water was poured into the sample then thoroughly mixed until the consistency to close the groove was between 20 and 30 blows. If the number of blows exceeded 30 or was lower than 20, the sample was removed from the brass cup and the water content was adjusted. 2. After getting a number of blows between 20 and 30, a portion of the samples (from the closed groove) was removed to measure the water content, then the soil from the brass cup was removed, then remixed in the dish and a new sample is placed in the cup. 3. A second test was then made until the sample required the same number of blows to close the groove as the first test or the difference in the number of blows was equal to two. A portion of the sample was then removed to measure the water content. 4. The Liquid Limit (LL) is the average of the two tests (to the nearest whole number). 5. If the difference in values equal 1% the test had to be repeated. The plastic limit (PL) according to the ASTM (2010) is determined by rolling (in a ground - glass plate) a small portion of soil into a 3.2 mm diameter thread until its water content is reduced to a point at which the thread crumbles and can no longer

38 27 be pressed together and rerolled. The following procedure was used to prepare the samples for the plastic limit (PL): 1. From the soil prepared for the liquid limit test, a 20 gram sample was selected, reducing the water content to a consistency in which it could be rolled without sticking to the hand or the glass plate. 2. From the 20 grams, 2 grams were extracted and turned, by hand, into an ellipsoidal mass. The ellipsoidal mass was then placed in the glass plate and rolled until its diameter reached 3.2 mm. 3. When the samples had a 3.2 mm diameter, they were then broken into three pieces and squeezed together, reformed and turned into an ellipsoidal mass and rerolled again until the thread crumbled and could not be reformed into an ellipsoidal mass and rolled again. 4. The Plastic Limit (LL) is the average of two tests (to the nearest whole number). If the difference in values is equal or greater than 1.4% the test had to be repeated.

39 Figure 6. Clay consistency and Particle size analysis tools. Brass cup and a groove used for liquid limit (Left). Stack of sieves used for particle size distribution analysis (Right). 28

40 29 Bulk Organic Carbon Assay The bulk organic carbon assay was used to extract the carbon dioxide from the diamicton units. The amount of organic carbon was then calculated and the carbon dioxide was sent to DirectAMS Laboratory in Seattle for Carbon 14 dating. About 14 organic carbon samples were analyzed in the Stable Isotope Laboratory of the Geosciences Department at Western Michigan University, from which 6 were selected for Carbon 14 dating. The preparation procedure is listed below: Inorganic carbon removal: 1. Approximately 7 grams of the sample was taken and disaggregated using a porcelain pestle and mortar. 2. The sample was poured into a 15 ml plastic centrifuge tube, filled with 6N Hydrochloric acid (HCl), agitated and allowed to sit for 24 hours. 3. The next day, the sample was centrifuged for a total of 20 minutes; the first 10 minutes at ½ speed and the next 10 minutes at ¾ speed. 4. After centrifugation, the acid was decanted into a waste container, replaced with new acid and allowed to rest for 24 hours. This step was repeated 4 to 5 times. 5. After the last centrifugation, the acid was decanted and replaced with deionized water to remove the acid from the samples.

41 30 6. The samples were then agitated and centrifuged for a total of 20 minutes; 10 minutes at ½ speed and 10 minutes at ¾ speed. The deionized water was then replaced. 7. Step 6 was repeated until the samples ph turned neutral (around 7), using a ph indicator. 8. Once the samples ph turned neutral, the water was decanted in a waste container and the sample were left drying in an oven at a temperature of 38 0 C. Sample Combustion and CO 2 extraction: 1. Approximately 500 mg of the dry sample was taken and disaggregated. 2. The dry sample was poured into a 6 mm Quartz tube. The 6mm tube was then placed inside a 9 mm Quartz tube with 1 gram of cupric oxide. 3. Air inside the 9 mm tubes was extracted on one of the vacuum lines. The 9 mm tube was then sealed with a blowtorch. 4. The 9 mm tube was then placed in the furnace and combusted for 3 hours at a temperature of C. 5. The combusted sample was mounted onto an extraction unit in a vacuum extraction line. The 9 mm was then broken releasing the CO 2 gas into the extraction unit and then to the vacuum line. 6. Liquid nitrogen was used to capture the CO 2 gas in the U-shaped tube of the vacuum line and release other gases like Nitrogen.

42 31 7. The liquid nitrogen was removed from the U-shaped tube and replaced with a slush made with dry ice and alcohol (at 70 o C). This released the CO 2 gas trapping the moisture and water. The CO 2 gas was then capture in a 10 cm tube of the vacuum line with liquid nitrogen. 8. The liquid nitrogen was removed, releasing the CO 2 gas inside the 10 cm tube. The reading in the pressure gauge was then recorded to calculate the amount of CO 2 gas in the tube (micromoles). 9. A sample tube was added to collect the CO 2 gas. Liquid nitrogen was used in the sample tube to capture the CO 2 gas. 10. The δ 13 C of gas was then measured in a Mass Spectrometer. The gas was returned to the vacuum line, sealed in a 9 mm Pyrex tube with a blowtorch and set to DirectAMS for Carbon 14 dating. Inorganic Carbon Assay The inorganic carbon assay was used to extract the carbonates from the diamicton units. Six samples were pretreated in the Stable Isotope Laboratory of the Geosciences Department at Western Michigan University and then sent to the Stable Isotope Laboratory from Oklahoma State University for isotope analysis. The preparation procedure is listed below: 1. Approximately 20 mg of the sample was taken and disaggregated using a porcelain pestle and mortar.

43 32 2. The sample was then poured into blood tubes. In addition, a 9 mm Pyrex tube was glued inside the blood tube. 3. After the glue between the tubes was dry, Phosphoric acid (H 3 PO 4 ) was injected in the 9 mm tube glued to the blood tube. The blood tube was then sealed with a plastic cap. 4. Air inside the blood tube was extracted on one of the vacuum lines and then was taken to the water bath. The water bath was used to avoid 18 O fractionation. 5. Then the next day, acid and sediment sample were mixed inside the blood tube and the CO 2 gas was extracted following the same procedure as the bulk organic carbon assay. Maps and Cross Section Data Maps and cross sections were created using ArcGIS software program ArcMap 10. Data used to create the maps, including water well logs, were imported from the State of Michigan s Geographic Data Library and wellogic data base, with the exception of the rotosonic cores drilled in the field. This database is accessible on the State of Michigan website. The data was translated into uniform lithological terms. All of the lithological terms including the borehole data were combined into three categories based on the grain size distribution (Table 2). Lithologies were grouped as Sand & Gravel (yellow), Silt & Clay (blue) and Diamicton (green). Clayey units mixed with gravel and/or sand were

44 33 interpreted as diamicton. This procedure is also used by the U.S Geological Survey and recently by Ewald (2012). Bedrock topography data was obtained from Mr. John Esch of the Michigan Department of Environmental Quality. Table 2 Interpreted Water Well Lithology Water Well Logs Uniform Lithology Color Clay Silt & Clay Blue Clay & Sand Silt & Clay Blue Clay & Silt Silt & Clay Blue Sand & Silt Silt & Clay Blue Muck Silt & Clay Blue Marl Silt & Clay Blue Gravel Sand & Gravel Yellow Gravel & Boulders Sand & Gravel Yellow Gravel & Clay Sand & Gravel Yellow Gravel & Sand Sand & Gravel Yellow Sand Sand & Gravel Yellow Hardpan Diamicton Green Clay & Stones Diamicton Green Clay & Boulders Diamicton Green Clay & Gravel Diamicton Green Clay Gravel Sand Diamicton Green Clay Gravel Stones Diamicton Green Clay Sand Gravel Diamicton Green

45 34 CHAPTER V RESULTS Several cores were drilled within the Saginaw lobe terrain in the Michigan Southern Peninsula. These cores are described according to the landsystem in which they were drilled. A textural classification system is used to express the general characteristics in the borehole soils. Sediment textures for the cores BA-09-02, BA , CA-11-01, KA and KA are classified using the U.S. Department of Agriculture (USDA) textural classification method. Landsystem 3: North of the Thornapple Valley Cores BA was drilled north of the Thornapple River Valley in a northeast -southwest trending tunnel valley in Barry County. The area is located within the range of landsystem 3 (Figure 4). Textural analysis was completed by Ewald (2012) and is replotted in Figure 6. BA The stratigraphy in the core is composed mainly of diamicton and fine sediments. The total depth of the core was meters (207 feet) and it reached bedrock at -57 meters (187 feet). Bedrock, in this area, comes from the Michigan Formation (Figure 2), which is mainly shale. Three diamicton units are present below a depth of 16.3 meters, separated by two thick layers of silt and clay. The predominant particle fraction in this core is silt and clay (Figure 7).

46 35 The interval between 0 (surface) to -7 meters has two layers, one of sand and one of gravelly sand (Figures 7, 8). The first layer (0.4 m) is a sandy loam consisting mainly of sand, with an average normalized texture of 51.5% of sand, 34.1% silt and 14.4% clay. The gravelly sand unit is mostly coarse; clay is almost nonexistent in this layer; it is composed of about 95% sand, 3.5% silt and 1% clay. The sand unit is made up of mostly fine sand. Between -7 to -16 meters is located a silty clay bed with an average normalized texture of 1.2% sand, 66.6% silt and 32.2% clay, underlain by a sand and silt unit between two gravelly sand layers (Figures 7, 8). The first gravelly sand layer is formed by very coarse material with some sand, while the second is a mixture of gravel, sand and silt. The interval between -16 to -46 meters has two diamicton units separated by two layers; a silt unit and a silty clay unit (Figures 7, 8). The upper till unit (Unit A-1) is a uniform/compact clay loam diamicton with an average texture of 32.3% sand, 41.6% silt and 26.1% clay. The middle till (Unit A-2) has an average texture of 27.7% sand, 43.4% silt and 28.9% clay. The unit is very similar to the previous layer, being comprised of a clay loam diamicton, uniform and compact, but with a higher percentage of fines and a lower percentage of coarse particles. At -46 m, there is a bed of sand separating part of the middle till unit, which could mean that the till unit below the sand bed is not part of the middle till unit (Figures 6, 7).

47 Figure 7. Logplot Diagram of BA showing lithology, gamma ray signature, and grain size distribution. The core contains two thick diamicton units separated by lacustrine sequences. The bedrock is shale from the Michigan Formation. 36

48 Figure 8. Matrix texture (<2.00 mm) of all samples in BA Green circles represent diamicton, blue circles represent silt/clay and yellow circles represent sand/gravel. The upper diamicton (Unit A-1) consists of sand and silt, with an average texture of 32.3% sand, 41.6% silt and 26.1% clay. The middle diamicton (Unit A-2) is dominated by silt and clay, with an average texture of 27.7% sand, 43.4% silt and 28.9% clay. 37

49 38 The till below the sand bed is a clay loam diamicton similar to the previous till units. The last layer is between -47 and 54 meters and consists of a silty clay unit. This silty clay layer is formed of an equal amount of fines with an average texture of 0.7% sand, 51.4% silt and 48% clay. Landsystem 2: South of the Thornapple and North of the Kalamazoo Valleys Two cores BA and KA were drilled south of the Thornapple River Valley and north of the Kalamazoo River Valley. The boring BA was drilled in Barry County and KA in Kalamazoo County. The area is located within the range of landsystem 2 (Figure 4). The glacial features in this area include the Kalamazoo Moraine between the Thornapple River Valley to the north and the Kalamazoo Valley to the south. BA The stratigraphy in the core is composed mainly of diamicton, sand and silt (Figure 9, 10). The boring was drilled to a depth of -85 meters (279 feet) and reached bedrock at meters (240 feet). The bedrock in this area is shale and comes from the Michigan Formation (Figure 2). About three thick diamicton units are present in the core. The first two units are coarse grained and the deepest one is finer. Between 0 (surface) to -13 meters are interbedded sand and silt/clay in the first four meters, followed by a unit of sandy diamicton interbedded with gravel (Figures 9, 10). The diamicton (Unit B-1) is a sandy loam with an average texture of

50 % sand, 25.1% silt and 11.1% clay. Sand and gravel lenses are present in the middle and base of this till unit. The interval between -13 to -36 meters is mostly dominated by sandy sediments. The first six meters consist of a sandy loam, followed by a silt layer (Figures 9, 10). The sandy loam has an average texture of 82.6% sand, 15% silt and 2.4% clay. The next five meters is comprised mainly of sand with a thin bed of silty clay. Below this bed is the second diamicton unit (Unit B-2), a compact sandy loam with gravel. This unit has an average texture of 77.6% sand, 18.1% silt and 4.3% clay. Four meters of sand with some gravel are present below the diamicton. In the interval between -36 to -66 silt fractions start to dominate (Figures 9, 10). A sandy loam unit is present in the first 8 meters, followed by a silt layer in the next 10 meters. The sandy loam has an average texture of 55.7% sand, 38.7% silt and 5.6% clay. Interbedded diamicton (Unit B-3) with variable textures is present in the interval -54 to -66 meters. A gravelly/sandy loam diamicton is located between -54 and -58 meters, with an average texture of 58.4% sand, 30.7% silt and 11% clay. A gravel layer separates the second diamicton bed into upper and lower units. This part is a clay loam diamicton formed mostly by 30.7% clay and 47% silt. The third diamicton bed is separated by a silt layer. This bed consists mainly in sand and silt with an average texture of 40.5% sand, 40.4% silt and 19% clay. The last interval between -66 to -73 meters is mainly sand and diamicton (Figures 9, 10). The first 5 meters are composed of interbeded sand and silt.

51 Figure 9. Logplot Diagram of BA showing lithology, gamma ray signature, and grain size distribution. The core contains three thick diamicton units separated by sandy and silty lacustrine sequences. The bedrock is shale from the Michigan Formation. 40

52 Figure 10. Matrix texture (<2.00 mm) of all samples in BA Green circles represent diamicton, blue circles represent silt/clay and yellow circles represent sand/gravel. Upper diamicton units (Unit B-1, B-2) consist mainly in sand and silt with an average texture ranging from 58.4% % sand, 18.1% % silt, 4.3% % clay. The lower diamicton unit (Unit B-3) is mainly formed by silt and clay with an average texture ranging from 22.3% % sand, 40.4% % silt, 19.0% % clay. 41

53 42 Underneath this layer is another diamicton bed. This bed is mainly a clay loam diamicton, with 22.5% of gravel. The average normalized texture for this unit is 30% sand, 42.1% silt and 27.9% clay. This unit is believed to be part of the shale bedrock. KA The boring was drilled to a depth of -81 meters (266 feet) and reached bedrock at -79 meters (259 feet). The bedrock in this area is shale and comes from the Michigan Formation (Figure 2). The stratigraphy in this core is diverse, but mostly sand and gravel followed by intercalated silt and clay (Figures 11, 12). A diamicton unit is located between thin sand lens in the intervals of -16 and -30 meters. The unit is mainly fine graineds. Interval 0 (surface) to -14 meters. The first 5 meters in this core are mainly sand, followed by 6.5 meters of gravel. A thin bed of diamicton is present below the gravel. This unit has an average texture of 50.3% sand, 22.5% silt and 27.2% clay. The final interval is mostly made up fines; a silt layer is below the diamicton unit, followed by a clay loam. The gravel percentage in these layers is below 1% (Figures 11, 12). Interval -14 to -45 meters. A thick diamicton unit (Unit C-1) is lies between two small layers of sand. The unit is 13 meters thick and has a loam texture with an average normalized texture of 49.5% sand, 34.9% silt and 15.5% clay. The content of gravel is less than 5 %, and is the only diamicton unit in this core.

54 Figure 11. Logplot Diagram of KA showing lithology, gamma ray signature, and grain size distribution. The core contains one thick diamicton unit. Lacustrine sequences are present above the diamicton unit and between 49 and 65 meters. The shale bedrock underlies thick coarse sediments. 43

55 Figure 12. Matrix texture (<2.00 mm) of all samples in KA Green circles represent diamicton, blue circles represent silt/clay and yellow circles represent sand/gravel. The diamicton unit (Unit C-1) in this core has a mean texture of 49.5% sand, 34% silt and 15.5% clay. 44

56 45 From a depth of -30 to -45 meters the core shifts from silty and clayey to sandy and gravelly. Intercalated sand and gravel with some silts dominates this interval (Figures 11, 12). The stratigraphy changes two more times between the depths -45 and -79 meters, first from sand to silt/clay and then, to gravelly sand followed by 3 meters of gravel. The interval begins with a silt unit followed by a fine sand layer, at -48 meter this setting changes to interbedded of silt/clay, and then to interbedded sand/silt. Finally the core becomes coarser at -65 meters, with 10 meters of gravelly sand, followed by 3 meters of gravel (Figures 11, 12). Landsystem 1: South of the Kalamazoo Valley Two cores, CA and KA-13-01, were drilled south of the Kalamazoo River valley. Core CA was drilled on a drumlin, within the city of Battle Creek in Calhoun County. Core KA was also drilled on a drumlin, but south of the Kalamazoo Valley and southeast of the city of Portage, Kalamazoo. Both cores fall within the area of landsystem 1 (Figure 4). CA The stratigraphy in this core is composed mainly of diamicton. The boring was drilled to a depth of meters (179 feet) and reached bedrock at -54 meters (177 feet). Bedrock in the area is siltstone of the Marshall Formation (Figure 2). Four

57 46 diamicton units are present in this core, separated by thick units of coarse sediments. The predominant particle size fraction is mainly coarse (Figures 13, 14). The first interval goes from 0 (surface) to meters, and is mainly a sandy loam diamicton. The diamicton unit (Unit D-1) has an average normalized texture of 68.9% sand, 22.1% silt, and 9.0% clay. The unit becomes more gravelly at greater depths, with an average of 7.5% gravel in the upper section and 16.2% between the middle and bottom sections. An interbedded unit of sand and gravel is present between -3 and -6 meters. Below the diamicton, 1.5 meters of fine sand is present and its marks a shift from sand to silt (Figures 13, 14). Between and -31 meters is the second diamicton unit (Unit D-2). One meter above the diamicton, two beds, one of sandy loam and one of silt are present. The sandy loam has an average normalized texture of 59.2% sand, 35.2% silt and 5.6% clay. A small bed of silt lies directly below the diamicton, followed by 4 meters of gravel and another small bed of silt. The dimicton in this interval is a sandy loam unit, with about 11.1% gravel and an average normalized texture of 60.1% sand, 28.5% silt and 11.4% clay (Figures 13, 14). The interval from -31 to -41 meters is dominated mainly by gravel and sand. Interbedded diamicton (Unit D-3) and gravelly sand is present in the first 5 meters, followed by 4 meters of gravel. The last meter is occupied by a silt/sand layer. Diamicton in this interval is a sandy loam unit with an average normalized texture of 66.5% sand, 21.6% silt and 11.9% clay. The amount of gravel in this diamicton unit is an average of 11.1% (Figure 13, 14).

58 Figure 13. Logplot Diagram of CA showing lithology, gamma ray signature, and grain size distribution. The core contains four thick diamicton units separated by sand and gravel. Sand and gravel between diamictons are poorly sorted, and are interpreted as outwash deposits. The bedrock is siltstone from the Marshall Formation. 47

59 Figure 14. Matrix texture (<2.00 mm) of all samples in CA Green circles represent diamicton, blue circles represent silt/clay and yellow circles represent sand/gravel. Diamicton units in this core have a high amount of gravel and sand. The mean texture of the upper two diamictons (Unit D-1, D-2) are 68.9% sand, 22.1% silt, 9.0% clay, and of 59.2% sand, 35.2% silt, 5.6% clay. The lower two diamictons (Unit D-3, D-4) have a mean texture of 66.5% sand, 21.6% silt, 11.9% clay and 40.8% sand, 40.4% silt, 18.8% clay. 48

60 49 The last interval, between -41 to -54 meters, is mainly diamicton finer than the upper till unit (Figure 13, 14). The diamicton unit (Unit D-4) is divided at -50 meters by 1.5 meter of silt, and above the unit, 4 meters of gravelly sand. The first diamicton in this interval is a loam unit with an average normalized texture of 40.8% sand, 40.4% silt and 18.8% clay. Gravel percentage is low in this unit, about 1.8%. The second diamicton is another sandy loam unit, located beneath the silt layer. The average normalized texture is about 71.3% sand, 23.6% silt and 5.2% clay. KA The stratigraphy in this core is composed of more than 70% sand/gravel, and sandy diamicton. The boring was drilled to a depth of meters (159 feet) and reached bedrock at -47 meters (154 feet). Bedrock in the area is shale of the Coldwater Shale (Figure 2). At least two diamicton units are present in this core, separated by thick layers of coarse sediments. The predominant particle fraction is gravel, followed by sand (Figures 15, 16). A diamicton unit (Unit E-1) is located between 0 and -15 meters and it split by a small sand bed at -6 meters. The first unit is mainly formed by loamy sand with 13% gravel and an average normalized texture of 64.0% sand, 28.4% silt and 7.6% clay. Interbedded diamicton and gravel/sand is present beneath the small sand bed at - 6 meters. The unit is a sandy loam diamicton with an average texture of 63.5% sand, 30.6% silt and 5.8% clay. A thick gravel unit is present below the diamicton,

61 50 followed by another unit of sand. The gravel is about 3.5 meters thick (Figures 15, 16). The interval between -15 and -37 meters is mainly composed of very coarse sediment (Figures 15, 16). The first bed is composed of 6 meters of gravel, which is very coarse with some sand; clay/silt is very low in this unit. The next 5 meters include intercalated gravelly sands with a gravel bed between them. Interbedded sand/ gravel occupy the next 9 meters, followed by 3 meters of gravel. The last diamicton unit (Unit E-2) is located between two sand units in the interval that goes from -37 to -47 meters. The unit is a sandy loam diamicton, uniform and compact, low on gravel (about 3.8%), with an average texture of 59.1% sand, 29.0% silt and 11.9% clay. The sand units above and below the diamicton are form mostly by fine sand and some silt. Gravel is present only in a small bed above the diamicton unit (Figures 15, 16).

62 Figure 15. Logplot Diagram of KA showing lithology, gamma ray signature, and grain size distribution. The core contains two diamicton unit separated mainly by gravel with sand. The bedrock in this area is shale of the Coldwater Shale Formation. 51

63 Figure 16. Matrix texture (<2.00 mm) of all samples in KA Green circles represent diamicton, blue circles represent silt/clay and yellow circles represent sand/gravel. The two diamicton in this core are mainly sandy units with gravel. The mean texture of the upper diamicton (Unit E-1) is 64.0% sand, 28.4% silt, 7.6% clay, and the lower diamicton (Unit E-2) of 59.1% sand, 29.0% silt and 11.9% clay. 52

64 53 Diamicton Clay Consistency The clay consistency of diamicton samples from cores BA-09-02, BA and CA-11-01was tested with the plastic and liquid limit test (Table 3). Samples from these boreholes are mostly inorganic clay with low to medium plasticity. The result indicates a strong correlation between plasticity indexes in diamicton between the depths of -46 to -72 meters. The following are the results of the plastic and liquid limit test: BA-09-02: According to the plasticity index, the clays in these diamictons range from low to medium plasticity. Samples from the upper and lower diamicton consist of low plasticity clay. Diamictons samples between the intervals -22 to -46 meters are formed by medium plasticity clay, close to high plasticity. Clay color varies between brown and gray. The diamictons in this core contains higher clay content. Bedrock in the area is shale, which is a possible clay source in the diamicton (Table 3). BA-10-02: According to the plasticity index, clay in these diamictons varies from slight to low plasticity, close to medium plasticity in samples between -10 and -55 meters. The clay between intervals -12 to -32 meters is slightly plastic to almost non plastic, provably due to the high silt content in these diamictons. Silt content becomes lower in diamictons samples close to bedrock. Clay color varies from orange to brown and dark gray. Bedrock in the area is also shale, and this explains the dark

65 54 Sample # Depth (meters) Table 3 Diamicton Clay Consistency Data Results Depth (feet) Liquid Limit (LL) Plastic Limit (PL) Plasticity Index (LL - PL) Plasticity Classification BA B Low 2-B Medium 3-B Medium 4-B Medium 5-B Medium 6-B Low BA C Low 2-C Slightly 3-C Slightly 4-C Slightly 5-C Low 6-C Low CA D Slightly 2-D Slightly 3-D Slightly 4-D Slightly 5-D Low 6-D Medium

66 55 gray color in the clays. Diamicton below -55 meters, near the shale bedrock has low plasticity (Table 3). CA-11-01: According to the plasticity index, the clays in these diamictons are slightly plastic, approaching a value of zero plasticity between the intervals of -9 to -25 meters. The plasticity increases from low to medium between -46 and -47 meters depth. Clay color in these diamictons varies from light brown to gray, with gray diamicton having the lowest plasticity. The bedrock in the area is mainly siltstone, which doesn t contain a clay source (Table 3). Cross Section A-A Cross section A-A has been constructed between Barry, Kalamazoo and Calhoun Counties and it shows the general distribution of glacial deposits from the Saginaw Lobe (Figure 17). The cross section includes water well logs within 826 meters of the cross section line, as well as cores BA-09-02, BA-10-02, KA-12-02, CA and KA Red dashed lines represent the boundaries between sedimentary packages. Water well logs are grouped into three categories according to grain size. Boreholes BA and BA Water well logs between the cores BA and BA indicate that uplands are mainly underlain by fine sediments such as silt/clay and fine - grained sediment mixed with diamicton (Figures 17, 18). Ewald (2012) interpreted the high silt and clay content as a possible lacustrine origin. A few deposits of coarse sediment occur in the middle section of the profile, followed

67 Figure 17. Location of cross section A-A. Water wells are marked with smaller blue circles. 56

68 57 by sandy diamicton mixed with clay above the bedrock. According to the water well logs, sediments in the Thornapple valley consist mostly of sand in the first -24 meters and become mixed with diamicton and clay/silt until it reaches the bedrock. Boreholes BA and KA Water well logs between the cores BA and KA indicates diversity in the uplands stratigraphy (Figures 17, 18). The northern section, close to core BA-10-02, consists of thick intervals of silt/clay and fine grained sediment mixed with diamicton. The middle section contains a layer of thick coarse sediments, which is mainly sand and gravel. The silt/clay continues below the sand, followed by a thick interval of sandy diamicton. Uplands on the southern section, close to core KA-12-02, are comprised of about 50 meters of coarse sediment, followed by a thick interval of diamicton mixed with silt/clay and sand. A small layer of sand separates some of the silt/clay from the diamicton. Boreholes KA and CA Coarse and fine sediments intercalate between cores KA and CA According to the water well logs, uplands near core KA-12-02, consist primarily of coarse sediments and sandy diamicton, then a few meters below silt and clay start to dominate the middle section, which is dominated by an interval of sand (Figures 17, 18). The bottom section has a layer of silt/clay, followed by sand. Water well logs, show that The Kalamazoo Valley is underlain mainly by coarse sediment, with silt and clay are present below the sand. It is possible that erosion removed part of the fine sediments. Stratigraphy near core

69 Figure 18. Cross Section A-A. Water wells and cores lithologies in landsystem 1, 2 and 3. Lithologies were grouped in 4 different categories (Silt&Clay, Sand&Gravel, Diamicton and Bedrock) for better interpretations. Red dashed lines marks the interpreted stratum and solid gray lines the bedrock elevation. 58

70 59 CA consists mainly of diamicton. Around -37 meters below surface coarse sediments start to appear, followed by diamicton and clay. Boreholes CA and KA Uplands in the area consist mainly in sandy diamicton and coarse sediments deposits (Figures 17, 18). Stratigraphy near core CA is mainly diamicton in the upper and bottom parts. The middle part is mostly silt, clay and fine sands. The middle-bottom section between cores CA and KA is mainly sand. More data is needed to draw interpretations regarding the extent of sand in the layer in this section. Near core KA diamicton, silt and clay appear again in the uplands, followed by a small layer of interbedded clay and sand. Close to the bedrock, in the bottom part, is a sandy layer with some traces of diamicton. Radiocarbon and δ 13 C Analyses Samples from four boreholes were analyzed in the Geosciences Dept. Stable Isotope Laboratory, and then sent to different laboratories for Carbon 14 dating. Samples from cores BA-09-01, BA and OT were dated at DirectAMS laboratory. The last core, KA was analyzed by Barnes (2007) (Table 6). Organic carbon content, along with δ 13 C values, were also determined in the Stable Isotope Laboratory in the Geosciences Department (Table 4). Inorganic carbon (carbonates), along with δ 13 C and δ 18 O were analyzed by the Stable Isotope Laboratory at Oklahoma State University (Table 5).

71 60 δ 18 O (VSMOW) analyses on carbonates collected from the samples of diamicton ranged between 23.8 and 25.0 per mil, and δ 13 C (VPDB) analyses ranged between -2.4 and 1.3 per mil (Table 5). These δ 13 C values are commonly seen in marine sediments, which indicate carbonate deposits of marine origin. This result is consistent with the enriched δ 18 O values, which are mainly seen in the dolomite/ limestone deposits. In addition glacial deposits around southwestern Michigan overlie Mississippian bedrock, which consists of 47% carbonates (limestone, dolomite) (Dorr and Eschman 1970). δ 13 C analyses on bulk organic carbon collected from the diamicton samples ranged between and per mil (Table 4). These values suggest the presence of C-3 plants, which is consistent with the climate and vegetation of southwestern Michigan. In addition, radiocarbon ages extracted from the diamicton units bulk organic carbon give similar dates between 9 and 11 meters in three of the core samples (Table 6). Radiocarbon ages tend to get younger between 29 and 57 meter. The bulk organic carbon in these sediments appears to be late Wisconsin in age.

72 61 Table 4 Diamicton Bulk Organic Carbon Data Footage Core ID (meters) (feet) δ13c Mean organic carbon (%) BA OT KA OT BA BA Table 5 Diamicton Carbonates Data Footage δ 13 C δ 18 O Core ID (meter) (feet) (VPDB) (VSMOW) BA OT OT OT BA BA

73 62 Table 6 Carbon 14 Data Results Footage Radiocarbon age Lab ID Core ID (meters) (feet) BP Error D-AMS BA , D-AMS OT , UGAMS KA , D-AMS OT , D-AMS BA , D-AMS BA ,799 58

74 63 CHAPTER VI DISCUSSION Geotechnical analyses, grain size, textural classification, consistency, along with glacial lithologies described in previous chapters, were used to identify the stratigraphic framework, to interpret lithologic units and to correlate sediments of glacial advances/retreat. Correlation of units was focused on the diamicton units based on their stratigraphy and grain size distribution. Glacial deposits such as diamicton serve as evidence of glacial advance/retreat, and are usually present as nearly continuous layers of sediments. Analysis of these layers affords the ability to accurately correlate these types of sediments across an area. The five cores analyzed; BA-09-02, BA-10-02, KA-12-02, CA and KA-13-01, each have between one and four diamicton units. Core Interpretations Core BA was drilled in landsystem 3. The area represents stagnation and tunnel valleys in the area are assumed to have been formed by subglacial meltwater (Kehew et al. 2012a). The core BA contains two thick diamicton units separated by silt/clay units. Ewald (2012) interprets these silt/clay units as lacustrine sequences based on their thickness, lamination and texture. Diamicton in this core consists of thick clay loam units, one near the surface (Unit A-1), and the other close to the middle section (Unit A-2) just above a silty clay unit (Figure 19).

75 Figure 19. Matrix texture (<2.00 mm) of diamicton samples in BA Red circles represent upper diamicton (Unit A-1), orange circles represent middle diamicton (Unit A-2). The lower diamicton samples have more silt and clay when compared to the upper diamicton samples. 64

76 Figure 20. Matrix texture (<2.00 mm) of diamicton samples in BA Red circles represent upper diamicton (Unit B-1), orange circles represent middle diamicton (Unit B-2), black circles represent lower diamicton (Unit B-3). The middle diamicton samples have more sand than the upper and lower units. In the lower diamicton samples, sand content decreases and clay content increase with depth. 65

77 Figure 21. Matrix texture (<2.00 mm) of diamicton samples in KA Red circles represent upper diamicton (Unit C-1). Samples in this unit are mainly sandy and silty, with an increase in clay content in the upper and bottom sections. 66

78 67 Clay consistency in these diamictons varies from low plasticity in the upper and bottom to medium plasticity in the middle section. Cores BA and KA were drilled in landsystem 2. Laminated silts and clays sediments in the area are assumed to have been deposited in a glaciolacustrine environment (Ewald, 2012). Clay consistency in the diamictons from this boring varies from slight to low plasticity. Core BA contains three main diamicton units and is separated by lacustrine sequences. The upper (Unit B-1) and middle (Unit B-2) diamictons are mainly sandy loam units. The lower diamicton (Unit B-3) is a very coarse sandy loam unit with high gravel content (Figure 20). Core KA contains one thick diamicton unit of about 13 meters between silt/clay sediments probably deposited by meltwater during deglaciation. The diamicton (Unit C-1) is a loam unit composed mostly of sand and silt (Figure 21). Core KA is located in an ice marginal zone; sediments in this area are mostly a product of the stagnation and melting of the Saginaw lobe. Lacustrine sequences are present above the diamicton unit and between -49 and -65 meters. The shale bedrock is overlain by thick coarse sediments. Both cores BA and KA seem to correlate in at least one diamicton unit. Cores CA and KA were drilled in landsystem 1. The surficial geology of the area consists of drumlins and outwash deposits. Diamictons in this area are mostly sandy units, with a clay consistency ranging from slight in the first -30 meters, to low/medium below the -40 meters. Core CA contains four thick

79 Figure 22. Matrix texture (<2.00 mm) of diamicton samples in CA Red circles represent upper diamicton (Unit D-1), orange circles represent upper middle diamicton (Unit D-2), black circles represent lower middle diamicton (Unit D-3), green circles represent lower diamicton (Unit D-3). Samples from the upper and middle diamictons are mainly sandy units. The lower diamicton samples have more silt and clay than the upper units. 68

80 Figure 23. Matrix texture (<2.00 mm) of diamicton samples in KA Red circles represent upper diamicton (Unit E-1), orange circles represent middle diamicton (Unit E-2). Diamicton samples in this core are mainly sandy and silty, with and average clay content of 10%. 69

81 70 diamicton units separated by sand and gravel. Sand and gravel between diamictons are poorly sorted, and are interpreted as outwash deposits. The first three diamicton (Unit D-1, D-2, and D-3) are sandy loam units. The forth diamicton (Unit D-4) lies above the bedrock and consists of a loam unit rich in silt (Figure 22). The bedrock is siltstone, which explains the high silt content in the lowest diamicton unit. Core KA contains two diamicton unit separated mainly by gravel and sand. Both diamictons are sandy units; the upper diamicton (Unit E-1) has a thickness of 11 meters, the lower diamicton (Unit E-2) has significantly more clay than the first one and a thickness of 1.5 meters (Figure 23). Gravel followed by sand are the predominant fractions in this core. Diamicton in drumlins were the product of subglacial deposition and as the ice retreated outwash and other coarse sediments were deposited along the area. Landsystem Correlation across the Saginaw Lobe Sediments in the uplands between landsystem 3, 2 and 1, consist mainly of diamicton and lacustrine deposits. For this thesis sedimentary packages along the south-central portion of the Lower Peninsula of Michigan have been tentatively correlated in cross section A-A (Figure 25). Diamicton units indicate the presence of at least one major and two minor advances/retreats of the Saginaw Lobe. Cores CA and KA were collected from a drumlinized till plain across landsystem 1. The bedrock elevation is shallow in this area, and decreases to the north and west of the study area. Sediments in KA vary from sandy diamicton, of which the

82 Figure 24. Matrix texture (<2.00 mm) of diamicton samples from cores: BA-09-02, BA-10-02, KA-12-02, CA and KA Red circles represent upper diamicton samples from depth between 0 to -31 meters; orange circles represent lower diamicton samples from depth between -32 to -74 meters. The upper diamicton samples have more sand when compared to the lower diamicton samples. 71

83 Figure 25. Cross Section A-A : Proposed Correlation of Sediments. Correlated sedimentary packages in landsystem 1, 2 and 3. Glacial sediments were grouped in 3 different categories (Silt, Clay & Fine sand, Sand&Gravel and Diamicton) for better interpretations. Solid gray lines the bedrock elevation. Sedimentary packages in the cross section indicate at least one major glacial retreat and two minor advances/retreats. 72